Oceans '88 a partnership of marine interests : proceedings : Baltimore, Maryland, October 31-November 2, 1988

[From the U.S. Government Printing Office, www.gpo.gov]

"v"
MR,

7
6
-RAWN E RS Hl P -'@O R,, I T
=Eli REIN ELRESTS",

'YZ

-88
-CH2585-8,

TROC EDINGS

APIP

till,

GC
1001
.025
1988
v. 2

COASTAL ZONE INFORMATION'.@CENTER-.,

OCE-AN-S-88-
CONFERENCE AND EXPOSITION PRESENTED BY MTS-OES-IEEE IN COOPERATION WITH THE PORT OF'BXLTIMORE
ALTIMORE CONVENTION CENTER, BALTIMORE, MARYLAND OCTOBER 31 -NOVEMBER 2, 1988
DPNALD SCHAEFER, GOVERNOR OF MARYLAND, HONORARY CHAIRMAN
A,pMIRAL PAUL A. YOST. COMMANDANT UNITED STATES COAST GUARD, GENERAL CHAIRMAN

88-CH2585-&'

___too

-,@@PROQEEDINGS
VOL 2

I'A

Ul.

GC 'o,
1001
.025
1988
v. 2 v.

COASTAL ZONE INFORMATIOW'CENTER.I.

FERENCE AND EXPOSITION PRESENTED BY MTS-OES-IEEE IN COOPERATION WITH THE PORT OF'BALTIMORE
IMORE CONVENTION CENTER, BALTIMORE, MARYLAND OCTOBER 31 -NOVEMBER 2, 1988
LID SCHAEFER, GOVERNOR OF MARYLAND, HONORARY CHAIRMAN
IRAL PAUL A. YOST, COMMANDANT UNITED STATES COAST GUARD, GENERAL CHAIRMAN

OCEANS 88

A Partnership of Marine Interests

Property Of CSC Library

PROCEEDINGS

Conference Sponsored by
Marine Technology Society
IEEE

Baltimore, Maryland
October 31-November 2, 1988

U.S. DEPARTMENT OF COMMERCE NOAA
COASTAL SERVICES CENTER
2234 SOUTH HOBSON AVENUE
CHARLESTON., SC 29405-2413

IEEE Catalog Number 88-CH2585-8

Oceans '88 Proceedings

Volume 1: Pages 1 to 2 74
Volume 2: Pages 2 75 to 718
Volume 3: Pages 719 to 1086
Volume 4: Pages 1087 to 1732

Copies of the Oceans '88 Proceedings are available from:

The IEEE Service Center
445 Hoes Lane
Piscataway, NJ. 0885

and
The Marine Technology Society
2000 Florida Avenue, N.W., Suite 500
Washington, D.C. 20006

Copyright and Reprint Permissions: Abstracting is permitted with credit to the source. Libraries
are permitted to photocopy beyond the limits of U.S. copyright law for private use of patrons
those articles in this collection that carry a code at the bottom of the first page, provided
the per-copy fee indicated in the code is paid through the Copyright Clearance Center,
29 Congress Street, Salem, Mass. 01970. Instructors are permitted to photocopy isolated articles
for noncommercial classroom use without fee. For other copying, reprint or re-publication
permission, write to Director, Publishing Services, IEEE, 345 E. 47th Street, New York,
NY 10017.

IEEE Catalog Number 88-CH2585-8

OCEANS '88

Proceedings

Volume Two

Table of Contents

PLASTICS IN OUR OCEANS: WHAT ARE WE DOING MESOCOSMS AS TOOLS FOR COASTAL AND ESTUARINE
ABOUTIT? ENVIRONMENTAL RESEARCH-1

Chairman: Chairman:
B. Griswold G. F. Mayer
OAR, National Oceanic and Atmospheric National Oceanic and Atmospheric Administration
Administration
E. Klos 1529
J. M. Coe and A. R. Bunn I An Experimental Estuarine Salinity Gradient
Marine Debris and the Solid Waste Disposal Crisis
S. W. Nixon and S. W. Granger 1604
D. Cottingham 6 Development of Experimental Ecosystems for the
Federal Programs and Plastics in the Oceans Study of Coastal Lagoons

X. Augerot .1711 J. G. Sanders and G. F. Riedel 23
Sea Grant Faces Oceans of Plastic The Use of Enclosed Ecosystemsfor the Study of
Cycling and Impact of Trace Elements
K. J. O'Hara 12
Education and Awareness: Keys to Solving the Marine S. J. Cibik, J..G. Sanders and C. F. D'Elia 29
Debris Problem Interactions Between Insolation and Nutrient Loading
and the Response of Estuarine Phytoplankton
J. R. Whitehead 1507
Reducing Plastic Pollution in the Mar 'ine Environment- MESOCOSMS AS TOOLS FOR COASTAL AND ESTUARINE
The U.S. Coast Guardand Implementation. of Annex V ENVIRONMENTAL RESEARCH-11
of MARPOL 73178

Chairman:
CONTINENTAL SHELF ENVIRONMENTAL RESEARCH R. E. Turner
Center for Wetlands Resources,
Chairman: Louisiana State University
W. W. Schroeder
University of Alabama A. G. Chalmers 1652
Experimental Manipulations of Drainage in a Georgia
R. Rezak and D. W. McGrail 1602 Saltmarsh: Lessons Learned
Geology and Hydrology of Reefs and Banks Offshore
Texas and Louisiana. M. R. DeVoe, M. E. Tompkins and J. M. Dean 35
South Carolina's Coastal Wetland Impoundment
W. W. Schroeder, M. R. Dardeau, J. J. Dindo, P. Project (CWIP): Relationship of Large-Scale Research
Fleischer, K. L. Heck, Jr. and A. W. Schultz 17 to Policy and Management
Geological and Biological Aspects of Hardbottom
Environments on the LMAFLA Shelf, Northern Gulf of R. E. Turner 41
Mexico Experimental Marsh Management Systems in Louisiana

W. W. Schroeder, R. Rezak and T. J. Bright 22 A. G. van der Valk, B. D. J. Batt, H. R. Murkin, P. J.
Video Documentation of Hardbottom Environments Caldwell and J. A. Kadlec 46
The Marsh Ecology Research Program (MERP): The
Organization and Administration of a Long-Term
Mesocosm Study

iv.

TECHNICAL ADVANCES IN SEAFOOD TECHNOLOGY AND WATER REUSE ON ONSHORE MARICULTURE AND
SAFETY PROCESSING FACILITIES

Chairman: Chairman-
D. Attaway R. Becker
Sea Grant, National Oceanic and Atmospheric Louisiana State University
Administration
J. M. Fox and A. L. Chauvin 1536
R. R. Colwell 1606 Deputation of Oysters in a Closed Recirculating
New Approaches for IndiceslMonitoring Microbial System
Pathogens in Seafood
M. P. Thomasson, D. G. Burden and R. F. Malone 70
J. Liston 52 Micro-Computer Based Design of Recirculating
Microorgam .sms as a Cause of Economic Loss to the Systems for the Production of Soft-shell Blue Crabs
Seafood Industry (Callinectes sapidus)

G. J. Flick, Jr. 56 G. E. Kaiser and F. W. Wheaton 76
Sea Grant Advances in Seafood Science and Computerized Rapid Measurement of Ammonia
Technology Concentration in Aquaculture Systems

S. Garrett and M. Meyburn. K. Rausch, W. H. Zachritz II, T. C. T. Y-Hsieh and
Development of New Approaches to Seafood R. F. Malone 84
Inspection Use of Automated Holding Systems for Initial Off-
Flavor Purging of the Rangia Clam, Rangia cuneata
TECHNICAL ADVANCES IN SEAFOOD TECHNOLOGY AND
SAFETY GULF OF MEXICO CHEMOSYNTHETIC PETROLEUM SEEP
COMMUNITIES
Chairman:
G. J. Flick, Jr. Chairman:
Virginia Polytechnic Institute R. Carney'
Louisiana State University
R. C. Lindsay 61
Flavor Chemistry and Seafood Quality Factors 1. MacDonald, R. Carney and D. Wilkinson 90
Gulf of Mexico Chemosynthetic Communities at Oil
H. 0. Hultin 66 Seeps: Estimating Total Density
Technical Problems and Opportunities Related to
Utilization of Our Seafood Resources R. S. Carney 96
Emerging Issues of Environmental Impact to Deep-Sea
J. P. Zikakis 1608 Chemosynthetic Petroleum Seep Communities
A Biotechnological System for the Utilization of Waste
Products of the Seafood and Cheese Minufqctuiffig H. H. Roberts, R. Sassen and P. Abaron 101
Industries Petroleum -Derived Authigenic Carbonates of the
Louisiana Continental Slope
A. P. Bimbo 1513
The Production of Menhaden Surimi

UNDERSEA VEHICLES AND PLATFORMS FOR SCIENCE OIL AND GAS INDUSTRIES CONFLICT
APPLICATIONS
Chairmen:
Chairman: R. W. Middleton
A. N. Kalvaitis Minerals Management Service
National Undersea Research Program, M. Holliday
National Oceanic and Atmospheric Administration National Marine Fisheries Service
G. A. Smith and R. S. Rounds io6 J. Brashier 136
Scientific, Technological and Social Impact of NOAA's Coexistence of Fishing and Oil and Gas Industries in
Mobile Undersea Research Habitat the Gulf of Mexico
P. J. Auster, L. L. Stewart and H. Sprunk 1286 B. R. Clark 143
Scientific Imaging Problems and Solutions for ROVs
Potential Conflicts Between Oil and Gas Industry
L. L. Stewart and P. J. Auster 1610 Activities and Commercial Fishing
Low Cost ROVs for Science R. M. Meyer 146
R. A. Cooper and 1. G. Babb 112 Information on Fisheries Risk Assessment in the Alaska
Manned Submersibles Support 2 Wide Range of OCS Region
Underwater Research in New England and the Great R. C. Wingert 150
Lakes Geophysical Survey and Commercial Fishing Conflicts,
R. 1. Wicklund and B. L. Olla 119 Environmental Studies and Conflict Mitigation in the
Field Research Programs at the Caribbean Marine Minerals Management Service Pacific OCS Region
Research Center-National Undersea Research Program A. S. Knaster 156
The Use of Alternative Dispute Resolution in
OCS FISHERIES AND RESOURCES Resolving Outer Continental Shelf Disputes

Chairmen: CUMULATIVE ENVIRONMENTAL EFFECTS OF THE OIL AND
R. W. Middleton GAS LEASING PROGRAM-I
Minerals Management Service
M. Holliday Chairmen:
National Marine Fisheries Service J. Goll
Minerals Management Service
R. W. Middleton 123 J. M. Teal
Oil and Gas Industry Conflicts on the Outer Woods Hole Oceanographic Institute
Continental Shelf
D. Christensen 1624 D. V. Aurand 161
Outer Continental Shelf Fisheries and Resources in the The Future of the Department of the Interior OCS
Northeast Region Studies Program
R. J. Essig 127 T. Chico 166
Outer Continental Shelf Fishery Resources of the Air Quality Issues, Environmental Studies, and
South Atlantic Cumulative Impacts in the Pacific OCS Region
B. G. Thompson 1613 R. E. Miller 172
Outer Continental Shelf Fisheries and Resources in the Georges Bank Monitoring Program: A Summary
Gulf of Mexico J. M. Teal 177
S. Koplin 132 The Role of the Scientific Advisory Committee, Outer
Continental Shelf Program of Minerals Management
The Outer Continental Shelf Fishery Resources of the Service
Pacific Coast

vi.

CUMULATIVE ENVIRONMENTAL EFFECTS OF THE OIL AND OIL AND GAS EXPLORATION-H
GAS LEASING PROGRAM-11
Chairmen:
Chairmen: J. R. Pearcy
J. Goll Minerals Management Service
Minerals Management Service C. Welling
J. M. Teal Ocean Minerals Co.
Woods Hole Oceanographic Institute
C. A. Dunkel 208
R. M. Rogers 953 A Qu2litative Assessment of the Hydrocarbon Potential
Factors Contributing to Wetland Loss in the Coastal of the Washington and Oregon Continental Shelf
Central Gulf of Mexico
J. M. Galloway and M. R. Brickey 1611
S. D. Treacy 180 The Hydrocarbon Potential of the Federal OCS,
The Minerals Management Service Bowhead Whale Offshore Northern C21ifomi2
Monitoring Program and Its Applications
J. Kennedy and C. Grant 213
R. B. Clark 184 Impact of the Oil-be2ring Monterey Formation on
Impact of Offshore Oil Operations in the North Sea Undiscovered Resources of Offshore C21ifornia

J. P. Zippin 1615 S. Sorenson, C. Alonzo and M. Ibrahim 1612
Cumulative Environmental Effects of the Department Wilson Rock Field: A Case History
of the Interior's Offshore Oil and Gas Program: 1987
Report to Congress OIL AND GAS RESOURCE MANAGEMENT

OIL AND GAS EXPLORATION-1 Chairmen:
R. V. Amato
Chairman: Minerals Management Service
J. R. Pearcy C. Welling
Minerals Management Service Ocean Minerals Co.
F. R. Keer 188 G. M. Edson 219
Geologic Characteristics of an Atlantic OCS Gas The Ancient Atlantic Reef Trend
Discovery and Its Implications
P. K. Ray 193 B. J. Bascle 223
Hydrocarbon Potential of the Deepwiter (600 Feet) The Effect of Exploration on Resource Estimates for
Gulf of Mexico the Alaska Outer Continental Shelf
W. E. Sweet and J. C. Reed 202 D. Mayerson 229
Correlation of Cenozoic Sediments-Gulf of Mexico Pre-lease Geophysical Permitting for the Pacific OCS:
Outer Continental Shelf Procedures, Problems, and Solutions

D. A. Steffy 235
Post-Lease Sale Exploration of the Navarin Basin,
Bering Sea, Alaska

vii.

OFFSHORE DRILLING-ENVIRONMENTAL STUDIES ACOUSTIC APPLICATIONS-II

Chairman: Chairman:
D. Cottingham A. 1. Eller
National Oceanic and Atmospheric Administration Science Applications International Corp.

D. K. Fran(;ois 241 L. C. Haines, 'W. W. Renner and A. I. Eller 295
Environmental Studies and Impact Assessment on the Prediction System for Acoustic Returns from Ocean
Atlantic Outer Continental Shelf Bathymetry

R. B. Krahl and C. E. Smith 250 G,,. P. Vellemarette 298
Developing Technologies for Offshore Oiland Gas P@Ogrammable Subsurface Acoustic Recording System
Structures in Frontier and Hazardous Areas
D. F. McCammon 3o4
OCEAN LEASING AND DEVELOPMENT The Relationship Between Acoustic Bottom Loss and
the Geoacoustic Properties of the Sediment
Chairman:
G. Pettrazzulo ACOUSTICS-NOISE
Technical Resources Inc.
Chairmen:
S. Ashmore 259 D. J. Ramsdale
Offshore Leasing Boundaries Along the Receding Naval Ocean R&D Activity
Alaskan Coastline N. Miller
T. J. Mac Gillvray 262 West Sound Association
Development and Analysis of DCF Computer Models W. S. Hodgkiss 310
for EEZ Marine Mining Source Ship Contamination Removal in a Broadband
M. E. Dunaway and P. Schroeder 268 Vertical Array Experiment
Minimizing Anchoring Impacts During Construction of R. J. Lataitis, G. B. Crawford and S. F. Clifford 315
Offshore Oil and Gas Facilities A New Acoustic Technique for Remote Measurement
ACOUSTIC APPLICATIONS-1 of the Temporal Ocean Wave Spectrum

Chairman: ACOUSTICS-PROPAGATION
A. I. Eller
Science Applications International Corp. Chairman:
D. G. Browning
W. Hill, G. Chaplin and D. Nergaard 275 Naval Underwater System Center
Deep-Ocean Tests of an Acoustic Modem Insensitive
to Multipath Distortion D. G. Browning, P. M. Schiefele and R. H. Mellen 318
Attenuation of Low Frequency Sound in Ocean
A. Novick 1617 Surface Ducts: Implications for Surface Loss Values
A Shallow Water Sonar Performance Prediction
System W. J. Vetter 1540
On Ray Trajectories and Pathtimes for Acoustic
R. L. Spooner 283 Propagation in a Medium with Velocity Gradients
Signal Processing Using Spreadsheet Software
D. K. Roderick 1619
J. M. Tattersall, J. A. Mingrone and P. C. King 1618 An Introduction to the Physics of Underwater Sound
A VCR Based Digital Data Recorder for Underwater and Their Application to Passive Anti-Submarine
Acoustics Multipath Measurements Warfare

L. Wu and A. Zielinski 287
Multipath Rejection Using Narrow Beam Acoustic Link

viii.

ACOUSTICS-SIDE SCAN SEA BOTTOM PROPERTIES

Chairman: Chairman:
R. Walker M. Cruckshank
USCG R&D Center University of Hawaii

A. St. C. Wright 323 R. B. Perry 366
The Wide Swath, Deep Towed SeaMARC Mapping the Slopes of Expanding Continental Margins

R. G. Asplin and C. G. Christensson 329 C. de Moustier, T. Hylas and J. C. Phillips 372
A New Generation Side Scan Sonar Modifications and Improvements to the Sea Beam
System On Board R/V Thomas Washington
E. Kristof, A. Chandler and D. Schomette 335
Using a Sector-Scan Sonar to Hunt for Shipwrecks D. E. Pryor 379
Through Ice Theory and Test of Bathymetric Side Scan Sonar

J. W. Nicholson and J. S, Jaffe 338 S. M. Smith, J. S. Charters and J. M. Moore 385
Side Scan Sonar Acoustic Variability Processing and Management of Underway Marine
Geophysical Data at Scripps
R. Gandy and S. Paulet 1620
Realtime Side Scan Sonar Target Analysis R. L. Cloet 1636
Implications of Using a Wide SWATH Sounding
W. R. Abrams 344 System
A Practical High Tech Advance in Side Scan Sonar
Target Positioning and Analysis SEDIMENT STUDIES-1

AC OUSTIC DOPPLER CURRENT PROFILING Chairman:
A. G. Young
Chairman: FUGRO-McClelland
H. R. Frey
Office of Oceanography and Marine Assessments, S. K. Breeding and D. Lavoie 391
National Oceanic and Atmospheric Administration Duomorph Sensing for Laboratory Measurement of
Shear Modulus
G. F. Appell, J. Gast, G. Williams and P. D. Bass 346
Calibration of Acoustic Doppler Current Proftlers D. Lavoie, E. Mozley, R. Corwin, D. Lambert and
P. Valent 397
Y. Kuroda, G. Kai and K. Okuno 353 The Use of 2 Towed, Direct-Current, Electrical
Development of a Shipboard Acoustic Doppler Resistivity Array for the Classification of Marine
Current Profiler Sediments

D. Wilson, D. Bitterman and C. Roffer 359 P. F. Wainwright, B. Humphrey and G. Stewart 405
The Acoustic Doppler Current Profiling System at Sediment Contamination by Heavy Metals and
AOML Hydrocarbons

ix.

SEDIMENT STUDIES-11 SATELLITE REMOTE SENSING

Chairman: Chairmen:
H. G. Herrmann III D. E. Weissman
Naval Facilities Engineering Command Hofstra University
A. E. Hay, L. Huang, E. B. Colbourne, J. Sheng and J. Gallagher
A. J. Bowen 413 Naval Underwater Systems Center
A High Speed Multi-Channel Data Acquisition System M. R. Willard 162
for Remote Acoustic Sediment Transport Studies Ocean Sensing Capabilities on Landsat 6
D. G. Hazen, A. E. Hay and A. J. Bowen 419 S. W. McCandless, Jr. and J. Curlander 479
Design Considerations for RASTRAN-System 2 The Influence of Packing Technologies on
A. G. Young, L. V. Babb and R. L. Boggess 423 Environmental Application of Space-Based Synthetic
Mini-Probes: A New Dimension in Offshore In Situ Aperture Radar
Testing J. R. Benada, D. T. Cuddy andB. H. jai 473
K. L. Williams and L. J. Satkowiak 428 Adapting the NSCAT Data System to Changing
Bounded Beam Transmission Across a WaterlSind Requirements
Interface, Experimentand Theory W. B. Campbell and M. L. Weaks 1626
An Inexpensive Interactive Processing System for
L. J. Satkowiak 433 NOAA Satellite Images
Remote Sea Bottom Classification Utilizing the
Ulvertech Bottom Profiler Parametric Source D. S. Bryant, A. M. Ponsford and S. K. Srivastava 485
A Computer Package for the Parameter Optimization
THE GREAT LAKES AS AN OCEANIC MICROCOSM of Groundwave Radar

Chairman:
L. Pittman
Merchant Marine and Fisheries Committee,
U.S. Congress

J. R. Krezoski 437
Particle Reworking in Great Lakes Sediments: In-Situ
Tracer Studies Using Rare Earth Elements

J. R. Krezoski 442
In-Situ Tracer Studies of Surficial Sediment Transport
in the Great Lakes Using a Manned Submersible

L. F. Boyer 443
Video-Sediment-Proffie Camera Imagery in Marine and
Freshwater Benthic Environments

L. F. Boyer, R. J. Diaz and J. D. Hedrick 448
Computer Image-Analysis Techniques and Video-
Sediment-Profile Camera Enhancements Provide a
Unique and Quantitative View of Life at or Beneath
the Sediment-Waterface interface

OCEAN APPLICATIONS OF REMOTELY SENSED MICROWAVE WATER COLUMN MEASUREMENTS-1
TECHNIQUES
Chairmen:
Chairmen: R. S. Mesecar
D. E. Weissman Oregon State University
Hofstra University T. M. Dauphinee
J. Gallagher National Research Council, Canada
Naval Underwater Systems Center
W. Kroebel 491
C. Bostater and V. Klemas 462 Results of Exact Investigations About the
Remote Sensing of Physical and Biological Properties Characteristics of the Extremely Fast and Accurately
of Estuaries Measuring Kiel Multisonde and Representations About
Its Newest Performance
D. E. Weissman 1546
The Dependence of the Microwave Radar Cross K.-H. Mahrt and C. Waldmann 497
Section on Ocean Surface Variables During the Field Proven High Speed Micro Optical Density
Fasinex Experiment Profller Sampling 1000 Times Per Second with 10-6
Precision
W.-M. Boerner, A. B. Kostinski, B. D. James and
M. Walther 454 R. Mesecar and C. Moser 505
Application of the Polarimetric Matched Image Filter Multi-Sample Particle Flux Collector
(PMIF) Technique to Clutter Removal in POL-SAR
Images of the Ocean Environment J. M. Moore, C. de Moustier and J. S. Charters 509
Multi-Sensof Real-Time Data Acquisition and
D. L. Murphy 467 Preprocessing at Sea
Radar Detection of Oceanic Fronts
L. S. Fedor and E. J. Walsh 1697 WATER COLUMN MEASUREME NTS-11
Interpretation of SEASAT Radar Altimeter Returns Chairmen:
from an Overflight of Ice in the Beaufort Sea J.Jaeger
L. S. Fedor, G. S. Hayne and E. J. Walsh 1704 Honeywell Hydro Products
Airborne Pulse-Limited Radar Altimeter Return K. Hill
Waveform Characteristics over Ice in the Beaufort Sea Honeywell Hydro Products

J. Wagner and R. Mesecar 518
A Common XBTIPersonal Computer Interface

D. 1. Nebert, H. Saklad and G. Mimken 1627
CTD Data Acquisition Package

H. Tremblay 522
Hydroball-A New Expendable: Uses and Issues

Xi.

COMMUNICATIONS OCEAN ENGINEERING-I

Chairman: Chairmen:
R. A. Buddenberg, USCG C. A. Kohler
Office of Command and Control USCG R&D Center
R. A. Buddenberg and A. Givens 526 R. Geminder
Shipboard Tactical Computer: The Coast Guard's Mechanic Research Inc.
Combat Information Center Modernization R. L. Benedict 577
R. L. Moe 532 Destruction of Offshore Platforms by Accelerated
Networking and Ship-to-Shore Ship-to-Ship Galvanic Corrosion
Communication C. A. Kohler 582
S. C. Hall 537 -Corrosive-Wear of Buoy Chain
The Defense Mapping Agency's Navigation J. Larsen-Basse, B. E. Liebert, K. M. Htun and
Information Network A. Tadjvar 1628
Long-Term Abrasion and Corrosion Damage to the
COLD REGIONS OPERATIONS Hawaii Deep Water Power Cable

Chairmen: M. Briere, K. C. Baldwin and M. R. Swift 588
S. Smith Collision Tolerant Pile Structures: Design Analysis
U.S. Coast Guard Software
E. Early T. Dowd 595
University of Washington United States Naval Experience with Antifouling Paints
J. D. Crowley 543
Cold Weather Effects upon Marine Operations OCEAN ENGINEERING-II

S. M. Smith and D. Strahl 549 Chairmen:
Articulated Lights in Ice J. R. Vadus
Office of Oceanography and Marine Assessments,
M. Gorveatt and M. C. Yee 555 National Oceanic and Atmospheric Administration
Arctic Ice Island Coring Facility K. Okamura
Special Assistant to the Minister of Science and
COLD REGIONS MEASUREMENTS Technology, Japan

Chairmen: A. Bertaux 598
S. R. Osmer Tapered Interface in Harsh Environment Connectors
USCG International Ice Patrol
W. E. Hanson J. F. Legrand, A. Echardour, L. Floury, H. Floch, J.
USCG International Ice Patrol Kerdoncuff, T. Le Moign, G. Loaec and Y. Raer 602
Nadia: Wireline Re-Entry in Deep Sea Boreholes
W. E. Hanson 561 P. K. Sullivan and B. E. Liebert 606
Operational Iceberg Forecasting Concerns Impedance Measurements of Biofouling in Seawater
G. Steeves and S. Grant 567 Condensers: An Update
An Autonomous Atmospheric Pressure Recorder for E. A. Fisher and H. P. Hackett 607
Establishing Polar Sea Surface Height World's First Rigid Free-Standing Production Riser
T. K. Newbury and A. J. Adams 573 F. EI-Hawary 291
Estimated Ice-Gouge Rates on a Manmade Shoal in the
Beaufort Sea Compensation of Vertical Displacement Components
in Marine Seismic Applications Using the Coupled
Heave and Pitch Model

Xii.

INFORMATION SYSTEMS-I INFORMATION SYSTEMS-III

Chairmen: Chairmen:
J. A. Smith C. D. Kearse
USCG R&D Center office of Marine Operations,
G. Williams National Oceanic and Atmospheric Administration
Texas A&M University D. White
General Instrument Corp.
H. Bhargava and S. 0. Kimbrough 1554
Oona: An Intelligent Decision Support System for the G. Samuels 648
U.S. Coast Guard A Shipboard Data Acquisition, Logging and Display
% System
T. F. Pfeiffer 612
A Single Board Computer Based Sail Controller C. V. Baker and W. T. Whelan 650
Offshore Oceanographic Applications for Battery-
M. R. Nayak 615 Powered, High-End Microprocessors
On the Knowledge-Based Expert System for Marine
Instrumentation R. Findley 655
CIDS-A Shipboard Centralized Integrated Data
R. J. Smith 618 System
OPDIN-One Way the Ocean Community Informs
M 'Reynolds, R. Hendershot, M. jungck and
INFORMATION SYSTEMS-II B. Reid 1560
The Zeno Alliance Network: A Dual-Loop Fiber Optic
Chairman: Instrumentation Network for Ships
P. Topoly
Systems Planning NESDIS, MOORING
National Oceanic and Atmospheric Administration
Chairmen:
D. Stamulis and M. P. Shevenell 623 K. R. Bitting
The Use of WORM Optical Disks in Ocean Systems USCG R&D Center
W. B. Wilson 629 R. Swenson
A Method for Optimizing Environmental Observing Neptune Ocean Engineering,
Networks
J. D. Babb 660
W. C. Sutherland 632 Validation of Computer Model Predictions of the
A User-Friendly Multi-Functional CTD Software L2rge-Sc2le Transient Dynamic Towing Response of
Package Flexible Cables
D. Hamilton and J. Ward 637 H. 0. Berteaux, D. E. Frye, P. R. Clay and
On-line Access to NODC Information Services E. C. Mellinger 670
Surface Telemetry Engineering Mooring (STEM)
E. Voudouri and L. Kurz 641 D. R. May 681
Robust Sequential m-Interval Approximation Detectors New Technologies and Developments in NDBC Buoy
with Q-Dependent Sampling and Mooring Design

xiii

OPERATIONAL OCEANOGRAPHY INTERNATIONAL COUNCIL FOR EXPLORATION OF THE SEA

Chairman: Chairmen:
S. R. Osmer J. B. Pearce
USCG International Ice Patrol ICES Marine Environment Quality Committee
S. R. Osmer and D. L. Murphy 687 J. N. Moore
International Ice Patrol Applied Oceanography Center Ocean Law and Policy,
University of Virginia
R. L. Tuxhorn 691 J. F. Pawlak 719
Oceanography on EAGLE Australia '88 Cruise A Review of the Origins, Responsibilities, Composition
J. A. McNitt 696 and Main Activities of the International Council for
United States Navy Operational Oceanography: the Exploration of the Sea (ICES)
Fighting Smart with Oceanography Intelligence S. A. Murawski 726
An Evaluation of Shellfish Research in the
OCEANOGRAPHY-MEASUREMENTS AND ANALYSIS International Council for the Exploration of the Sea

Chairman: J. B. Pearce 732
W. D. Scherer The ICES Marine Environmental Quality Committee
Office of Oceanography and Marine Assessments, (MEQC): Its History and Activities
National Oceanic and Atmospheric Administration
F. P. Thurberg 736
P. Clemente-Cdlon and J. Zaitzeff 1629 The ICES Working Group on Biological Effects of
Upwelling Monitoring Off Western Sahara Contaminants: A Case Study
K. Monkelien and T. L. Murrell 699 SEWAGE SLUDGE DISPOSAL AND MONITORING
Windrose, PC Software for Wind Data Analysis
M. Enomoto, T. Kawanishi and W. Kato 703 Chairmen:
Measurement of Luminance Distribution on the Sea C. Dougherty
Surface for Comfortable Living Space Environmental Protection Agency
G. Lotzic
UNDERWATER PHOTOGRAPHY New York City Department of Environmental
Protection
Chairman: J. C, Swanson and K. jayko 740
E. Kristof Modeling the Impacts of CSO Treatment Alternatives
National Geographic Society on Narragansett Bay
E. Kristof, J. Stancampianc, and A. Chandler 709 H. M. Stanford and D. R. Young 745
Use of a Mac Iro-Hybrid Camera at National Geographic Pollutant Loadings to the New York Bight Apex
E. Kristof, A. Chandler and W. Hamner 713 S. E. McDowell, C. S. Albro, W. R. Trulli,
3-D as an Underwater Too] W. G. Steinhauer and F. G. Csulak 1630
Optimum Techniques for Tracking Plumes in the
Ocean: A Case Study of Sludge Plume Dispersion at
the 106-Mile Site

C. E. Werme, P. D. Boehm, W. G. Steinhauer and
F. G. Csulak 1631
A Monitoring Plan for Disposal of Sewage Sludge at
the 106-Mile Site

C. D. Hunt, W. G. Steinhauer, C. E. Werme, P. D.
Boehm and F. G. Csulak 1632
Monitoring Water Quality Characteristics During
Dispoasl of Sewage Sludge at the 106-Mile Site

xiv.

MARINE MINERAL RESOURCES PROBLEMS IN OUR BAYS AND ESTUARIES

Chairman: Chairman:
B. Haynes V. K. Tipple
Environmental Protection Agency Estuarine Program Office,
National Oceanic and Atmospheric Administration
R. J. Greenwald and H. F. Hennigar, Jr. 752
Designation of an Ocean Mining Stable Reference Area E. M. Burreson and J. D. Andrews 799
Unusual Intensification of Chesapeake Bay Oyster
R. M. Mink. B. L. Bearden and E. A. Mancini 762 Diseases During Recent Drought Conditions
Regional Geologic Framework of the Norphlet
Formation of the Onshore and Offshore Mississippi, C. F. D'Elia and P. R. Taylor 803
Alabama, and Florida Area Disturbances in Coral Reefs: Lessons from Diadema
Mass Mortality and Coral Bleaching
T. J. Rowland 768
Availability of Minerals Offshore Virginia P. A. Tester, P. K. Fowler and R. P. Stumpf 808
Red Tide, the First Occurrence in North Carolina
C. E. McLain 777 Waters: An Overview
Ocean Mining: An Opportunity for Public-Private
Partnership B. L. Welsh 1633
Hypoxia in Long Island Sound (LIS), Summer of 1987
R. V. Amato 783
Recent Nonenergy Mineral Activity in the Atlantic P. Molinari 1609
Outer Continental Shelf EPA's Response to the Flotables Incidents of the
Summer of 1987

TRASH ALONG THE COAST
THE DOLPHIN DIE-OFF
Chairman:
L. Swanson Chairman:
State University of New York N. M. Foster
National Marine Fisheries Service
J. B. Pearce 786
Events of th6 Summer of '87 D. R. Cassidy, A. J. Davis. A. L. Jenny and
D. A. Saari 812
L. Schmidt 790 Pathology of the Diseased Dolphins
Impacts and Implications of the Summer of 1987,
New Jersey Flotable Incidents J. Geraci 1634
Epidemiology of Bottlenose Dolphin Disease-U.S.
R. E, Dennis, R. P. Stumpf and M. C. Predoehl 1569 Atlantic Coast, 1987-1988
Environmental Conditions in New York Bight, July-
August, 1987 J. G. Meade, C. W. Potter and W. A. McLellan 815
Statistical Characteristics of the 1987 Bottlenose
R. L. Swanson, R. Zimmer and C. A. Parker 794 Dolphin Die-Off in Virginia
Meteorological Conditions Leading to the 1987
Washup of Floatable Wastes on Newjersey Beaches W. Medway 818
and Comparison of These Conditions with the Results of the Dolphin Epidemic Investigation as the
Historical Record Disease was Presented in New Jersey Specimens of
Bottlenose Dolphins in 1987

G. P. Scott, D. M. Burn and L. J. Hansen 819
The Dolphin Dieoff. Long-Term Effects and Recovery
of the Population

XV.

SHIPWRECK ARCHEOLOGY OIL SPILL MOVEMENT

Chairmen: Chairman:
W. C. Phoel D. F. Paskausky
NMFS, Sandy Hook Laboratory, USCG R&D Center
National Oceanic and Atmospheric Administration
J. Bonclareff I. M. Lissauer 842
House Merchant Marine and Fisheries Committee A.Verifled Model for Oil Spill Movement, Beaufort
Sea, Alaska
J. D. Broadwater 824 M. Reed and E. R. Gundlach 847
Historic Shipwrecks: Resources Worth Protecting Hindcast of the Amoco Cadiz Oil Spill
A. G. Giesecke 827 E. J. Tennyson and H. Whittaker 853
The Abandoned Shipwreck Act: A Context The 1987 Newfoundland Oil Spill Experiment: An
P. J. A. Waddell 833 Overview
Reburial of a l6th Century Galleon E. J. Tennyson 857
J. D. Broadwater 837 Shipboard Navigational Radar as an Oil Spill Tracking
Supporting Underwater Archaeology with Ocean Too]: A Preliminary Assessment
Technology C. M. Anderson and R. P. LaBelle 1673
R. W. Lawrence 1627 Update of Occurrence Rates for Accidental Oil Spills
Consequences of the Abandoned Shipwreck Act: The on the U.S. Outer Continental Shelf
North Carolina Example
DRIFT MEASUREMENT
J. Fullmer 1677
Myth and Management-The Shipwreck Management Chairmen:
Act L M. Lissauer
USCG R&D Center
ART R. Q. Robe
USCG R&D Center
Chairman:
H. B. Stewart, Jr. A. A. Allen and C. B. Billing 860
Old Dominion University Spatial Objective Analysis of Small Numbers of
Lagrangian Drifters
C. Olsen 1576
Art and Technology, on 20th-Century Vessels M. J. Lewandowski 865
A Minicomputer Application to Graphically Display
H. B. Stewart, Jr. 840 Tidal Current Drift
Artists on Oceanographic Expeditions: A Neglected
Partnership E. A. Meindl 871
Drifting Buoy Data Quality and Performance
Assessment at the National Data Buoy Center

P. J. Hendricks 1635
Drift Current Measurements from a Submarine

Xvi.

ECONOMICS OF MARINE OPERATIONS ENVIRONMENTAL POLICY

Chairmen: Chairmen:
F. Olson S. Bolton
Environmental Consultant Office of Legislative Affairs,
D. M. King National Oceanic and Atmospheric Administration
ICF Inc. R. Dye
House Merchant Marine and Fisheries Committee
M. D. Aspinwall 876
Commercial Vessel Operations in the Exclusive R. W. Zeller 905
Economic Zone: Will the Jones Act Keep Up? Resolving the Environmental Decisionmaking and
Research Dilemma
M. W. Clark, Jr., D. P. Robinson and L. G. Antle 1690
Economic Impacts from Coal Exports: Through the C. A. Crampton and R. C. Helland 910
Port of Baltimore and the Port of Norfolk A Strategy for Program Implementation

C. D. MacDonald and H. E. Deese 880 J. N. Leonard 914
Opportunities for DevelopInent: A Growth Scenario Updating the Stratton Commission: A Proposal for the
and Situation Analysis of Hawaii's Ocean Industries U.S. Coast Guard Ocean Survey Corps

D. L. Soden, J. D. Reighard and W. H. Hester 891 H. E. Schultz 920
Outside Influence on Port Operations: The Insider's National Response Mechanism
Perspectives
J. S. Hawkins 925
M. G. Johnson 896 Satellite Ocean Monitoring at Ten Years: Perceptions
Use of Systems Analysis Techniques in Ocean and Realities
Resources Development
P. Stang and E. Turner 1616
EDUCATION AND TRAINING Legal and Policy Issues at Stake in the Current 5-Year
Program
Chairmen:
Richard Asaro, USCG ESTUARINE STUDIES-1
Office of Marine Safety, Security and
Environmental Protection Chairman:
Thbyer Shafer D. J. Basta
Office of Oceanography and Marine Assessments,
J. Morton 899 National Oceanic and Atmospheric Administration
Marine Field Projects: Teaching is the Easy Part S. E. McCoy 930
S. Teel 1582 Monitoring the Estuary
Maritime Training and Ocean Education 1. C. Sheifer 937
H. F. Trutneff 9.02 Climate, Weather, and Coastal Recreational Growth in
The Impact of Marine Technology on Education and the Southeast U.S. in 1986
Training in Marine Transportation
A. Stoddard 942
An Innovative Approach for the Synthesis of Large
Oceanographic Data Sets with Pre-Processing and
Post-Processing of an Ecosystem Model of the New
York Bight

J. Gerritsen 948
Biological Control of Water Quality in Estuaries:
Removal of Particulate Matter by Filter Feeders

xvii.

ESTUARINE STUDIES-H OCEAN POLICY-A MATRIX OF FEDERAL, STATE AND
INTERNATIONAL ISSUES
Chairmen:
S. E. McCoy Chairmen:
Estuarine Program Office, E. W. Cannon
National Oceanic and Atmospheric Administration USCG Governmental Affairs Staff
D. Ashe K. U. Wolniakowski
House Merchant Marine and Fisheries Committee State of Oregon

Oral only E. W. Cannon 1717
The USCG: A Prototype for National and International
FISHERIES AND RESOURCE ASSESSMENTS Ocean Policy Implementation
Chairman: L. A. Berney
R. Smolowitz The Unspoken Yet Vital Partnership Between the
National Marine Fisheries, USCG Reserve and the Civilian Community at Large
National Oceanic and Atmospheric Administration C. R. Corbett 992
F. L. Ames 961 International Oil Spill Liability and Compensation
Improved U.S. Strategy for Fisheries Law Enforcement E. Hout, R. Bailey and K. U. Wolniakowski 994
G. Reetz 966 Ocean Resource Management in Oregon: Pushing
California Sea Otter: Impact Assessment and Mitigation Open the Window of Opportunity
D. Luo 972 J. S. S. Lakshminarayana 1000
Theoretical Analysis of Fish School Density Overview and Analysis of Coastal Zone Management
in the Atlantic Provinces, Canada
R. J. Smolowitz and F. M. Serchuk 975 D. C. Slade 1006
Marine Fisheries Technology in the United States: Coastal States and Marine Resource Development
Status, Trends and Future Directions Within the United States Exclusive Economic Zone

B. F. Beal 980
Public Aquaculture in Downeast Maine: The Soft-She]] OCEAN DRILLING PROGRAM
Clam Story
Chairman:
P. H. Averill 1637 J. H. Clotworthy
Development of Separator Trawl Technology Consultant
FISHERIES-IMPACT STUDIES P. Brown, K. Lighty, R. Met Irill and
P. D. Rabinowitz 1012
Chairman: Collection and Quality Control of Marine Geological
J. Chambers Data by the Ocean Drilling Program
National Marine Fisheries Service D. Graham, B. Hamlin, B. julson, W. Mills, A. Meyer,
H. A. Carr 984 R. Olivas, P. D. Rabinowitz, D. Bontempo and
Long Term Assessment of a Derelict Gillnet Found in J. Tauxe 1018
the Gulf of Maine Shipboard Laboratory Support: Ocean Drilling -
Program
A. E. Pinkney, L. L. Matteson and D. A. Wright 987 P. Weiss, G. Bode, C. Mato, R. Merrill, P. D.
Effects of Tributyltin on Survival, Growth, Rabinowitz, M. Angell, J. Miller, P. Myre, S. Prinz,
Morphometry and RNA-DNA Ratio of Larval Striped D. Quoidbach and R. Wilcox 1025
Bass, Morone saxatilis Core Curation: Ocean Drilling Program

xviii.

OCEAN ENERGY-I MARINE MAMMALS RESEARCH AND MANAGEMENT

Chairman: Chairman:
D. Cotter D. Swanson
CBI Industries National Marine Fisheries Service,
National Oceanic and Atmospheric Administration
P. Vauthier 1029
The Underwater Electric Kite: East River Deployment H.. H. Armstrong and K. R. Banks 1073
Modern Eskimo Whaling in the Alaskan Arctic
D. E. Lermard and F. A. Johnson 1034
British OTEC Programmes-I OMW Floating and G. H. Allen 1079
0.5MW Land Based Observations on the 1987 Subsistence Harvest of
Northern Fur Seals on St. Paul Island, Pribilof Islands,
R. K. Jensen 1039 Alaska
Hydro Power from the Ocean
R. T. Bennett 1083
A. Thomas and D. Hillis 1045 Endangered Species and Marine Mammal Protection
First Production of Potable Water by OTEC and Its During Offshore Structure Removals in the Gulf of
Potential Applications Mexico

OCEAN ENERGY-H SHIP DESIGN AND REPAIR

Chairman: Chairmen:
L. Lewis M. S. Canavan
Department of Energy USCG Office of Engineering and Development
T. Colton
D. C. Hicks, C. M. Pleass and G. R. Mitcheson 1049 Colton Company
DELBUOY. Wave-Powered Seawater Des2lination
System M. S. Canavan and M. D. Noll 1087
K. P. Melvin 1055 U.S. Coast Guard's New Polar Icebreaker Design
A Wave Energy Engine and Proposals for its D. W. Yu and J. H. Devletian 1098
Development and Usage Electroslag Surfacing for Construction, Restoration,
L. Claeson 1638 and Repair of Ship Structures
Recent Wave Energy Research in Sweden RESEARCH VESSELS
K. Kudo, T. Tsuzuku, K. Imai and Y. Akiyama 1061 Chairman:
Wave Focusing by a Submerged Plate W. Barbee
Y. Masuda, M. E. McCormick, T. Yamazaki and University-National Oceanographic Laboratory
Y. Outa 1067 System
The Backward Bent Duct Buoy-An Improved J. A. Chance 1107
Floating Type Wave Power Device Conversion of Surplus Ofifield Supply Vessels to
Research Vessels

C. Hamlin 1111
Research Vessels: A Systems Engineering Approach

B. L. Hutchison and S. Jagannathan 1117
Monohull Research Vessel Seakeeping and Criteria

R. J. Wilber, C. E. Lea and S. E. Humphris 1639
The SSV Corwith Cramer: Sea Education Association's
New Sailing Research Vessel

xix@

SALVAGE AND TOWING SWATH SHIPS-I

Chairman: Chairman:
J. H. Boyd R. Dinsmore
Booz, Allen & Hamilton, Inc. Woods Hole Oceanographic Institution

J. K. Edgar 1640 G. R. Lamb 1131
Hazardous Materials in Marine Salvage Operations Relationship Between Seakeeping Requirements and
SWATH Ship Geometry
C. M. Kalro 1603
Launch and Retrieval of a 1,000 Ton Barge Shaped M. Rice, E. Craig, S. Drummond and
Vessel from a Conventional Tanker C. junemann 1144
Conceptual Design of an Intermediate Size
J. Strandquist 1124 Oceanographic Research Ship for the University-
Removal of the Wreck of the Ex-USS TORTUGA National Oceanographic Laboratory System
THE SMALL PASSENGER VESSEL INDUSTRY-1 R. D. Gaul, A. C. McClure and F. E. Shumaker 1149
Design of a Semisubmerged SWATH Research and
Chairmen: Survey Ship
E. G, Sharf C. Kennell 1157
National Association of Passenger Vessels Tankage Arrangement for SWATH Ships
H. Parker
National Association of Passenger Vessels SWATH SHIPS-II
W. B. Hamner 1641 Chairman:
The Future of the Tourist Submarine Industry K. W. Kaulum
Oral only Office of Naval Research
T. G. Lang, C. B. Bishop and W. J. Sturgeon 1163
THE SMALL PASSENGER VESSEL INDUSTRY-H SWATH Ship Designs for Oceanographic Research

Chairman: E. Craig and S. E. Drummond 1169
E. G. Scharf SWATH CHARWIN-Range Support Ship
National Association of Passenger Vessels
A. Galerne 1573
T. MacRae 1125 Development of Deep Water Technology as It Relates
The Realities of Bareboat Chartering to Future Salvage

Oral only E. Craig
Real World Experience with SWATH Design

XX.

SHIPBOARD TECHNICAL SUPPORT AN INTERIM STATUS REPORT ON ORGANOTINS-II

Chairman: Chairmen:
H. L. Clark P. F. Seligman
National Science Foundation Naval Ocean Systems Center
H. L. Clark 1644 M. A. Champ
Shipboard Technician Program of the National Science National Science Foundation
Foundation K. W. M. Siu 1716
J. D. Guffy, M. A. Spears and D. C. Biggs 1173 Analytical Chemistry of Butyltins
Automated Analyses of Nutrients in Seawater with the T. L. Wade, B. Garcia-Romero and J. M. Brooks 1198
Technicon TrAAcs-800 A utoanalyzer System Tributyltin Analyses in Association with NOAA's
National Status and Trends Mussel Watch Program
D. J. Murphy, E. Wilson and E. Powell 1178
An Application of a Low Flow Current Meter to Broad C. M. Adema, W. M. Thomas, Jr., and
Temperature Range Estuarine Current Measurements S. R. Mangum 1656
M. Maccio and C. Langdon 1181 Butyltin Releases to Harbor Water from Ship Painting
Description of Conversion of an EG&G VMCM into a in a Dry Dock
MVMS (Multi-Variable Moored Sensor) B. Cool
Summary and Status Report of EPA's Special Review
AN INTERIM STATUS REPORT ON ORGANOTINS-1 (PD14)

Chairmen: WAVE MOTION
P. F. Seligman
Naval Ocean Systems Center Chairman:
M. A. Champ R. H. Canada
National Science Foundation National Data Buoy Center,
National Oceanic and Atmospheric Administration
P. F. Seligman, J. G. Grouhoug and C. M. Adema
Field Monitoring of TBT Concentrations in Pearl L. J, Ladner, W. B. Wilson and P. J. Kies 1202
Harbor Correlated with Model Simulation Studies Lake Superior Winter Weather Station

R. S. Henderson 1645 N. Lang
Chronic Exposure Effects of Tributyltin on Pearl The Linear Properties of Spectra from a PitchlRoll
Harbor Organisms Buoy

M. H. Salazar and S. M. Salazar 1188 E. D. Michelena and R. Dagnall
Tributyltin and Mussel Growth in San Diego Bay A Computer Controlled Signal Simulator for Buoy
Motion Sensors
M. H. Salazar and M. A. Champ 1497
Tributyltin and Water Quality: A Question of D. Smith and F. Remond
EnvironmentQ] Significance 3-Meter Directional PitchlRoll Buoy

W. R. Blair, G. J. Olson, T. K. Trout, K. L. Jewett
and F. E. Brinckman 166.1;
Accumulation and Fate of Tributyltin Species in
Microbial Biofilms

Xxi.

WAVE MEASUREMENTS UNDERWATER VEHICLES

Chairman: Chairman:
L. Baer R. Blidberg
Office of Oceanography and Marine Assessments, University of New Hampshire
National Oceanic and Atmospheric Administration
M. Higgins and R. Whyte 1646
H. Brown 1205 Controlled Depressor Towed Sensor Platform-The
Infrared Loser Wave Height Sensor U.S. Navy's Mk28 Search System

G. Kontopidis and G. Bowers 1207 M. Higgins, B. Lawson and B. Field 1647
WavePro: An Autonomous Wave Processor with Long- Development and Testing of a Heavy-Duty Work ROV
Range Telemetry for 10,000 Foot Service

F. Ziemer, H. Giinther and E. Stockdreher 1212 H. Momma, K. Ohtsuka and H. Hotta 1253
Measured Transfer Functions for Shipmotions in JAMSTECIDeep Tow System
Natural Seaways
J. jalbert, M. Shevenell, S. Chappel, R. Welsh and
D. W. Farrell 1587 R. Blidberg 1259
The Next Generation Water Level Measurement EAVE III Untethered AUV Submersible
System: The Next Step in Real-Time Data for
Navigation M. E. Cooke, S. Gittings, J. M. Brooks and
D. C. Biggs
WAVE ACTION ON SEA SHORES Texas A&M University Remotely Operated
Oceanographic Vehicle (TAMU-ROOV)
Chairman:
M. Earle UNDERWATER VEHICLE SYSTEMS AND EQUIPMENT
MEC Systems Corp.
Chairmen:
M. J. Briggs and P. J. Grace 1218 S. B. Cable
Influence of Frequency and Directional Spreading on Naval Civil Engineering Laboratory
Wave Transformation in the Nearshore Region R. Wernh
D. D. McGehee and J. P. McKinney 1224 Naval Ocean Systems Center
Tidal Circulation Data from the Los AngeleslLong F. Dougherty, T. Sherman, G. Woolweaver and
Beach Harbors G. Lovell 1265
S. L. Da Costa and J. L. Scott 1231 An Autonomous Underwater Vehicle (AUV) Flight
Wave Impact Forces on the Jones Island East Dock, Control System Using Sliding Mode Control
Milwaukee, Wisconsin M, L. Nuckols, J. Kreider and W. Feild 1271
J.Rosati III and G. L. Howell 1239 Thermal Modelling of Electro-Mechanical Cables for
A Hierarchical Multiprocessor Data Acquisition System ROV Applications
for Field Measurement of Structural Response in M. P. Shevenell and C. Millett 1276
Breakwater Concrete Armor Units A LISP Environment for Real-Time Ocean Systems
J. P. Ahrens and E. T. Fulford 1244 S. B. Cable 1280
Wave Energy Dissipation by Reef Breakwaters A Guideline System for the Navy's Submarine Rescue
E. H. Harlow 1250 Ship (ASR) Class
Why Breakwaters Break W. J. Herr 1290
AUV Technology: Development and Demonstration
Program

Xxii.

MANNED SUBMERSIBLES HAZARDOUS CHEMICAL IDENTIFICATION AND MANAGEMENT

Chairman: Chairmen:
R. W. Cook L. H. Gibson
Harbor Branch Oceanographic Institute USCG Central Oil Identification Laboratory
Oral only E. F. Batutis
Phasesep Corp.

DIVING OPERATIONS AND SYSTEMS W. R. Cunningham 1321
NOAA Fleet Hazardous Materials and Hazardous Waste
Chairmen: Management
W. C. Phoel
National Marine Fisheries Service, L. H. Gibson and M. S. Hendrick 1326
National Oceanic and Atmospheric Administration U.S. Coast Guard Oil Identification System
J. M. Wells T. J. Haas, J.. J. Kichner and T. J. Chuba 1332
Office of Marine Operations, Course in Hazardous Materials
National Oceanic and Atmospheric Administration

J. K. Jeffries BUOY-BASED METEOROLOGY
Standards and Procedures for Dry Suit Diving
Education Chairman:
R. Canada
J. K. Jeffries National Buoy Center
Thermal Guidelines for Diving Operations
S. P. Burke and D. G. Martinson 1335
R. 1. Wicklund 1614 An ARGOS Meteorological Oceanographic Spir Buoy
An Inexpensive Mobile Self-Contained Habitat System for Antarctic Deployments
for Marine Research
D. B. Gilhousen 1341
J. W. Blackwell and C. D. Newell 1300 Methods of Obtaining Weather Data in Real Time
Diving in Hazardous and Polluted Waters
E. D. Michelena 1649
J. M. Wells 1305 The Measurement of Precipitation at National Data
The Use of Nitrogen-Oxygen Mixtures as Diver's Buoy Center Stations
Breathing Gas
R. R. Miller and R. Canada 1650
J. P. Fish and H. A. Carr 1309 Mini-Drifter Test Deployment Data-Gulf of Mexico
Integrated Remote Sensing of Dive Sites Spring 1988

S. L. Merry, S. L. Sendlein and A. P. jenkin 1315 P. M. Friday, J. S. Lynch and F. S. Long 1344
Human Power Generation in an Underwater Interactive Marine Analysis and Forecast System
Environment (IMAFS): The Oceanographic Workstation of the
Future

D. A. Storey and W. E. Woodward 1348
The Global Ocean Platform Inventory

VESSELS OF THE 80s AND BEYOND

Chairmen:
E. K. Pentimonti
American President Lines, Ltd.

P. Mentz
Advanced Ship Operations, MARAD

Oral only

Xxiii.

FACILITIES IN SUPPORT OF MARINE FREIGHT NAVIGATION CHARTING
TRANSPORTATION
Chairmen:
Chairmen: R. Vorthman
M. J. Vickerman, Jr. USCG Operations Control Center, Atlantic Area
Vickerman, Zachary, Miller M. Kumar
R. Katims Defense Mapping Agency
Container Transport Technology
W. M. Maynard 1371
Oral only Cooperative Electronic Chart Development: The
GAADS Project
NAVIGATION SYSTEMS AND OPERATIONS P. W. Mushkat and C. Lamson 1589
Chairmen: Electronic Chart Display Information Systems:
G. R. Perreault Operational, Policy and Legal Issues
Office of Navigation, N. D. Smith 1374
U.S. Coast Guard Automated Nautical Data and Charting Development
J. Illgen
E. A. Soluri
A. F. E. Fuentes 1651 Defense Mapping Agency's Navigational Information
A Survey of Radionavigation System Users System
L. Mehrkam 1352 NAVIGATION SYSTEMS
Leading Lines for the Nineties
G. R. Perreault 1356 Chairmen:
Contract Service of Federal Aids to Navigation C. D. Kearse
Office of Marine Operations,
R. J. Weaver and R. M. Piccioni. 1362 National Oceanic and Atmospheric Administration
Marine Radionavigation of the Future P. Stutes
John E. Chance & Assoc. Inc.
L. V. Grant 1365
Federal Radionavigation Plan Overview J. L. Hammer III and W. R. Hoyle 1379
The Continuing Need for Accurate Positioning in
J: Hammer III and W. R. Hoyle 1684 Naval Tactics
The Continuing Need for Accurate Positioning in
Naval Tactics E. F. Carter and J. Lewkowicz 1594
A Computer Navigation System Using Kalmaro Filter
Smoothing

R. Gandy and S. Paulet 1648
Design and Applications of SEATRAC, an Integrated
Navigation and Data Management System

D. C. Slade
Solar Navigation

A. E. Shaw III and T. E. Bryan 1384
Oceanographic Applications of the ARGOS System

xxiv.

AIDS TO NAVIGATION SYSTEMS AND EQUIPMENT SEARCH AND RESCUE-SURVEILLANCE EXPERIMENTS

Chairman: Chairman:
T. S. Winslow W. H. Reynolds
Office of Engineering, USCG R&D Center
U.S. Coast Guard
D. Finlayson, D. Bryant, B. R. Dawe and
T. S. Winslow, M. D. Dawe, K. R. Schroeder and A. J. Armstrong 1417
W. A. Fisher 1390 Results of an Experiment to Examine Certain Human
High-Voltage Solar-Powered Navigation Range Design Factors Relating to Searches Conducted with Marine
Radar
J. McCaffrey
An Alternative Hull Design for the U.S. Coast Guard D. Bryant, B. R. Dawe, D. Finlayson, W. Reynolds
Bell Buoy and M. J. Lewandowski 1422
Results of Canadian Shipbome Night Search
GLOBAL POSITIONING SYSTEM Experiments
Chairmen: B. R. Dawe, D. Finlayson and D. Bryant 1427
K. Nakamura Results of a Canadian Shipbome Radar Search and
Office of the Assistant Secretary of Defense Rescue Detection Experiment
R. S. Warren D. Finlayson, B. R. Dawe and D. Bryant 1433
TASC Results of a Canadian Visual Search and Rescue
Detection Experiment
L. D. Hottram
Relative GPS Kinemetric Surveying and Applications F. Replogle, Jr. 1436
for Marine Positioning A New Coast Guard Search Technique
M. J. Mes 1395 SEARCH OPERATIONS
Accuracy of Satellite Survey Measurements on
Offshore Platforms for Monitoring Subsidence Chairmen:
E. M. Geyer and R. S. Warren R. Q, Robe
USCG R&D Center
Mission Planning Issues and Answers for GPS Users
B. Dawe
SHIP OPERATIONS AND SCHEDULING Nordco Ltd.
Chairman: R. W. Berwin 1439
C. Pritchett Alaska SAR Facility Archive and Operations System
USCG R&D Center M. K. Kutzleb 1444
S. Cook, R. Benway, W. Krug, M. Nestlebush. A. The Search for South African Airways Flight 295
Picciolo, W. Richardson, P. Stevens and D. R. Paskausky, W. Reynolds, R. Gaines and
V. Zegowitz 1400 R. Q. Robe 1605
Volunteer Observing Ships and the U.S. Improving Search Success; Real-Time Collection and
Government-A Winning Partnership Transfer to User
L. C. Kingsley, K. S. Klesczewski, J. A. Smith and R. Q. Robe, D. F. Paskausky and G. L. Hover 1448
R. A. Walters 1405 Performance of Coast Guard Medium Range
Comparing the U.S. Coast Guard Buoy Tender Surveillance (MRS) Aircraft Radars in Search and
Performance Using Simulation Rescue (SAR) Missions
K. S. Klesczewski 1411 J. B. Brewster 1454
Using Spacefilling Curve to Generate the Feasible Sea Based Aerostats (SBA): Effective *Surveillance for
Routes for the Set Partitioning Problem Maritime Interdiction
S. F. Roehrig 1643
Scheduling Patrols Using a Hybrid Iriteger
Programmingl Rule-Based System Approach

-xXv.

PORT MANAGEMENT AND SECURITY

Chairmen:
T. Robinson
Port Safety and Security Division,
U.S. Coast Guard

D. Smith
House Merchant Marine and Fisheries Committee

D. J. Evans, R. W. Owen and P. R. Farragut 1457
Innov3tive Technology Applied to Maximize a Port's
Lifeline: A Case History for the Sea Lanes of the
Chesapeake Bay

N. A. Marziani 1463
The Multi-Agency MOU on Port Security: A Model for
Conflict Resolution

D. J. Sheehy and S. F. Vik 1470
Mitigation Planning for Port Development

J. J. Zagel, R. T. Kilgore and S. M. Stein 1642
Hydrodynamic and Mass Transport Modeling of Navy
Harbors

MARINE SAFETY

Chairmen:
C. L. Hervey
USCG R&D Center

S. Steele
House Merchant Marine and Fisheries Committee

F. H. Anderson 1598
Awakening the Consciousness of the Boating Public
Regarding Pollution, Intoxication, and Common Sense
Safety of the Nation's Waterways

A. Colihan 1476
Coast Guard Recreational Boating Product Assurance
Program

C. L. Hervey 1482
Determining Horsepower Limits on Recreational Boats

S. Johnson and J. Veentjer 1487
Regulation of Passenger Carrying Submersibles

G. L. Traub 1493
Recreational Boating Accidents in Ocean Waters

Manuscript unavailable for publication

Xxvi.

Authors List

Abrams, W. R ............... 344 Boehm, P. D ......... 1631, 1632 Clark, M. W., jr ............ 1688
Adams, A. j ................. 573 Boerner, W.-M .............. 454 Clark, R. B ................. 184
Adema, C. M ............... 1656 Boggess, R. L ............... 423 Clay, P. R .................. 670
Aharon, P .................. 101 Bonetempo, D ............. 1018 Clemente-Colon, P .......... 1629
Ahrens, J. P ................ 1244 Bostater, C ................. 462 Clifford, S. F ................ 315
Akiyama, Y ................ 1061 Bowen, A. j ............ 413, 419 Cloet, R. L .................. 1636
Albro, C. S ................ 1630 Bowers, G ................ 1207 Coe, J. M ..................... I
Allen, A. A ................. 86o Boyer, L. F ............. 443, 448 Colbourne, E. B ............. 413
Allen, G. H ................ 1079 Brashier, j .................. 136 Colihan, A ................ 1476
Alonzo, C ................. 1612 Breeding, S. K ............... 391 Colwell, R. R .............. 16o6
Amato, R. V ................ 783 Brewster, J. B .............. 1454 Cook, S ................... 1400
Ames, F. L .................. 961 Brickey, M. R .............. 1611 Cooper, R. A ................ 112
Anderson, C. M ............ 1673 Briere, M ................... 588 Corbett, C. R ............... 992
Anderson, F. H ............. 1598 Briggs, M. j ................ 1218 Corwin, R .................. 397
Andrews, J. D ............... 799 Bright, T. j .................. 22 Cottingham, D ................ 6
Angell, M ................. 1025 Brinckman, F. E ............ 1668 Craig, E ............. 1144, 1169
Antle, L. G ................ 1688 Broadwater, J. D ........ 824, 837 Crampton, C. A ............. 910
Appell, G. F ................ 346 Brooks, J. M ............... 1198 Crawford, G. B .............. 315
Armstrong, A. j ............. 1417 Brown, H ................. 1205 Crowley, J. D ............... 543
Armstrong, H. H ............ 1073 Brown, P ................. 1012 Csulak, F. G ..... 1630,1631,1632
Ashmore, S ................. 259 Browning, D. G ............. 318 Cuddy, D. T ................ 473
Aspinwall, M. K ............. 876 Bryan, T. E ................ 1384 Cunningham, W. R .......... 1321
Asplin, R. G ................ 329 Bryant, D ........ 485, 1417, 1422, Curlander, j ................ 479
Augerot, X ................ 1711 1427,1433 Da Costa, S. L ............... 1231
Aurand, D. V ............... 161 Buddenberg, R. A ............ 526 Dardeau, M. R ................ 17
Auster, P. j ........... 1286, 161o Bunn, A. R ................... I Davis, A. j .................. 812
Averill, P. H ............... 1637 Burden, D. G ................ 70 Dawe, B. R ........... 1417, 1422,
Babb, 1. G .................. 112 Burke, S. P ................ 1335 1427,1433
Babb, J. D .................. 660 Burn, D. M ................. 819 Dawe, M. D ............... 1390
Babb, L. V .................. 423 Burreson, E. M .............. 799 Dean, J. M ................... 35
Bailey, R ................... 994 Burroughs, R. H ............ 1607 Deese, H. E ................. 880
Baker, C. V ................. 650 Cable, S. B ................ 1280 D'Elia, C. F .............. 29,803
Baldwin, K. C ............... 588 Caldwell, P. j ................ 46 de Moustier, C .......... 372,509
Banks, K. R ................ 1073 Campbell, W. B ............ 1626 Dennis, R. E ............... 1569
Bascle, B. j ................. 223 Canada, R ................. 1650 Devletian, J. H ............. 1098
Bass, P. D .................. 346 Canavan, M. S .............. 1087 DeVoe, M. R ................. 35
Batt, B. D. j ................. 46 Cannon, E. W .............. 1717 Diaz, R. j ................... 448
Beal, B. F ................... 980 Carney, R. S .............. 90,96 Dindo, J. j ................... 17
Bearden, B. L ............... 762 Carr, H. A ............. 984,1309 Dougherty, F .............. 1265
Benada, J. R ................ 473 Carter, E. F ................ 1594 Dowd, T ................... 595
Benedict, R. L ............... 577 Cassidy, D. R ............... 812 Drummond, S ........ 1144,1169
Bennett, R. T .............. 1083 Chalmers, A. G ............. 1605 Dunaway, M. E .............. 268
Benway, R ................ 1400 Champ, M. A .............. 1497 Dunkel, C. A ................ 208
Berney, A ................. 1725 Chance, J. A ............... 1107 Echardour, A ............... 602
Bertaux, A .................. 598 Chandler, A ........ 335, 709, 713 Edgar, J. K ................ 164o
Berteaux, H. 0 .............. 670 Chaplin, G ................. 275 Edson, G. M ................ 219
Berwin, R. W .............. 1439 Chappel, S ................ 1259 EI-Hawary, F ................ 291
Bhargava, H ............... 1554 Charters, J. S ........... 385,509 Eller, A. I .................. 295
Biggs, D. C ................ 1173 Chauvin, A. L .............. 1536 Enomoto, M ................ 703
Billing, C. B ................ 86o Chico, T ................... 166 Essig, R. j .................. 127
Bimbo, A. P ............... 1513 Christensen, C. G ............ 329 Evans, D. j ............. - . 1457
Bishop, C. B ............... 1163 Christensen, D ............. 1624 Farragut, P. R .............. 1457
Bitterman, D .......... ..... 359 Chuba, T. j ................ 1332 Farrell, D. W ............... 1587
Blackwell, J. W ............. 1300 Cibik, S. j ................... 29 Fedor, L. S ........... 1697,1704
Blair, W. R ................ 1668 Claeson, L ................. 1638 Feild, W .................. 1271
Blidberg, R ................ 1259 Clark, B. R ................. 143 Field, B ................... 1647
Bode, G .................. 1025 Clark, H. L ................ 1644 Findley, R .................. 655

xxvii.

Finlayson, D .......... 1417, 1422, Hazen, D. G ................. 419 Kerdoncuff,J ............... 602
1427,1433 Heck, K. L., Jr ............... 17 Kichner, J. J ............... 1332
Fish, J. P .................. 1309 Hedrick, J. D ............... 448 Kies, P. J .................. 1202
Fisher, E. A ................. 607 Helland, R. C ............... 910 Kilgore, R. T ............... 1642
Fisher, W. A ............... 1390 Hendershot, R ............. 1560. Kimbrough, S. 0 ........... 1554
Fleischer, P .................. 17 Henderson, R. S ............ 1645 King, P. C ................. 1618
Flick, G. J., Jr ................ 56 Hendrick, M. S ............. 1326 Kingsley, L. C ............... 1405
Floch, H ................... 602 Hendricks, P. J ............. 1635 Klemas, V .................. 462
Floury, L ................... 602 Hennigar, H. F., Jr ........... 752 Klesczewski, K ........ 1405, 1411
Fowler, P. K ................ 808 Herr, W. J ................. 1290 Klos, E ................... 1529
Fox, J. M .................. 1536 Hervey, C. L ............... 1482 Knaster, A. S ................ 156
Fran@ois, D. K .............. 241 Hester, W. H ................ 891 Kohler, C. A ................ 582
Friday, P. M ............... 1344 Hicks, D. C ................ 1049 Kontopidis, G .............. 1207
Frye, D. E .................. 670 Higgins, M ........... 1646, 1647 Koplin, S ................... 132
Fuentes, A. F. E ............ 1651 Hill, W .................... 275 Kostinski, A. B .............. 454
Fulford, E. T ............... 1244 Hillis, D .................. 1045 Krahl, R. B ................. 250
Fullmer, J ................. 1677 Hodgkiss, W. S .............. 310 Kreider, J ................. 1271
Gaines, R ................. 1605 Hotta, H .................. 1253 Krezoski, J. R ........... 437,442
Galerne, A ................ 1573 Hout, E .................... 994 Kristof, E .......... 335,709,713
Galloway, J. M ............. 1611 Hover, G. L ............... 1448 Kroebel, W .................. 491
Gandy, R ............ 1620,1648 Howell, G. L ............... 1239 Krug, W .................. 1400
Garcia-Romero, B ........... 1198 Hoyle, W. R .......... 1379, 1684 Kudo, K .................. 1061
Gast, J ..................... 346 Htun, K. M ................ 1628 Kuroda, Y .................. 353
Gaul, R. D ................. 1149 Huang, L ................... 413 Kurz, L .................... 641
Geraci, J .................. 1634 Hultin, H. 0 ................. 66 Kutzleb, M ................ 1444
Gerritsen, J ................. 948 Humphrey, B ............... 405 LaBelle, R. P ................ 1673
Gibson, L. H ............... 1326 Humphris, S. E ............. 1639 Ladner, L. J ................ 1202
Giesecke, A. G .............. 827 Hunt, C. D ................ 1632 Lakshminarayana, J. S. S ..... 1000
Gilhousen, D. B ............ 1341 Hutchison, B. L ............ 1117 Lamb, G. R ................ 1131
Givens, A .................. 526 Hylas, T ................... 372 Lambert, D ................. 397
Gorveatt, M ................ 555 Ibrahim, M ................ 1612 Lamson, C ................ 1589
Grace, P. J ................ 1218 Imai, K ................... 1061 Lang, T. G ................ 1163
Graham, D ................ 1018 Jaffe, J. S ................... 338 Langdon, C ................ 1181
Granger, S. W .............. 16o4 Jagannathan, S ............. 1117 Larsen-Basse, J ............. 1628
Grant, C ................... 213 jai, B. H .................... 473 Lataitis, R. J ................. 315
Grant, L. V ................ 1365 Jalbert, J .................. 1259 Lavoie, D .............. 391,397
Grant, S ................... 567 James, B. D ................. 454 Lawrence, R. W ............ 1627
Greenwald, R. J ............. 752 Jayko, K ................... 740 Lawson, B ................. 1647
Guffy, J. D ................ 1173 Jenkin, A. P ............... 1315 Lea, C. E .................. 1639
Gundlach, E. R .............. 847 Jenny, A. L ................. 812 Legrand, J. F ................ 602
Gunther, H ................ 1212 Jensen, R. K ............... 1039 Le Moign, T ................ 602
Haas, T. J ................. 1332 Jewett, K. L ................ 1668 Lennard, D. E .............. 1034
Hackett, H. P ............... 607 Johnson, F. A .............. 1034 Leonard, J. N ............... 914
Haines, L. C ................ 295 Johnson, M. G .............. 896 Lewandowski, M. J ..... 865,1422
Hall, S. C ................... 537 Johnson, S ................ 1487 Lewkowicz, J .............. 1594
Hamilton, D ................ 637 Julson, B .................. 1018 Liebert, B. E ........... 6o6,i628
Hamlin, B ................. 1018 Junemann, C ............... 1144 Lighty, K .................. 1012
Hamlin, C ................. 1111 Jungck, M ................. 1560 Lindsay, R. C ................ 61
Hammer, J. L., III ..... 1379,1684 Kadlec, J. A .................. 46 Lissauer, M ................. 842
Hamner, W ................. 713 Kai, G ..................... 353 Liston, J .................... 52
Hamner, W. B ............. 1641 Kaiser, G. E ................. 76 Loaec, G ................... 602
Hansen, L. J ................ 819 Kalro, C .................. 1603 Long,F.S ................. 1344
Hanson, W. E ............... 561 Kato, W ................... 703 Lovell, G .................. 1265
Harlow, E. H .............. 1250 Kawanishi, T ............... 703 Luo, D .................... 972
Hawkins, J. S ............... 925 Keer, F. R .................. 188 Lynch,J.S ................ 1344
Hay, A. E .............. 413,419 Kennedy,J .................. 213 Maccio, M ................. 1181
Hayne, G. S ............... 1702 Kennell, C ................ 1157 MacDonald, C. D ............ 880

xxviii.

MacDonald, I ................ 90 Moser, C ................... 505 Reetz, G ................... 966
Mac Gillvray, T. J ............ 262 Mozley, E .................. 397 Reid, B ................... 156o
MacRae, T ................. 1125 Murawski, S. A .............. 726 Reighard, J. D ............... 891
Mahrt, K.-H ................. 497 Murkin, H. R ................. 46 Renner, W. W ............... 295
Malone, R. F .............. 70, 84 Murphy, D. J ............... 1178 Replogle, F., Jr ............. 1436
Mancini, E. A ............... 762 Murphy, D. L ........... 467, 687 Reynolds, M ............... 1560
Mangum, S. R .............. 1656 Murrell, T. L ................ 699 Reynolds, W ......... 1422, 1605
Martinson, D. G ............ 1335 Mushkat, P. W ............. 1589 Rezak, R ............... 22, 1602
Marziani, N. A .............. 1463 Myre, P ................... 1025 Rice, M ................... 1144
Masuda, Y ................. 1067 Nayak, M. R ................. 615 Richardson, W ......I ........ 1400
Mato, C ................... 1025 Nebert, D. L ............... 1627 Riedel, G. F ................. 23
Matteson, L. L ............... 987 Nergaard, D ................ 275 Robe, R. Q ........... 1448, 1605
May, D. R .................. 681 Nestlebush, M .............. 1400 Roberts, H. H ............... 101
Mayerson, D ................ 229 Newbury, T ................ 573 Robinson, D. P ............. 1688
Maynard, W. M ............. 1371 Newell, C. D ............... 1300 Roderick, D. K ............. 1619
McCammon, D. F ............ 304 Nicholson, J. W ............. 338 Roehrig, S. F ............... 1643
McCandless, S. W,, Jr. ........ 479 Nixon, S. W ............... 1604 Roffer, C ................... 359
McClure, A. C .............. 1149 Noll, A D ................. 1087 Rogers, R. M ................ 953
McCormick, M. E ........... 1067 Novick, A ................. 1617 Rosati, J., III .............. 1239
McCoy, S. E ................ 930 Nuckols, M. L .............. 1271 Rounds,R.S ................ 106
McDowell, S. E ............. 1630 O'Hara, K. J ................. 12 Rowland, T. J ............... 768
McGehee, D. D ............. 1224 Ohtsuka, K ................ 1253 Rusch, K ................. @ . .84
McGrail, D. W ............. 1602 Okuno, K .................. 353 Saari, D. A .................. 812
McKinney, J. P ............. 1224 Olivas, R .................. 1018 Saklad, H ................. 1627
McLain, C. E ................ 777 Olla, B. L .................. 119 Salazar, M. H ......... 1188, 1497
McLellan, W. A .............. 815 Olsen, C .................. 1576 Salazar, S. M ............... 1188
McNitt, J. A ................. 696 Olson, G. J ................ 1668 Samuels, G ................. 648
Meade, J. G ................. 815 Osmer, S. R ................ 687 Sanders, J. G ............. 23, 29
Medway, W ................ 818 Outa, Y ................... 1067 Sassen, R ................... 101
Mehrkam, L ............... 1352 Owen, R. W ............... 1457 Satkowiak, L. J .......... 428, 433
Meindl, E. A ............ 629, 871 Parker, C. A ................ 794 Schiefele, P. M .............. 318
Mellen, R. H ................ 318 Paskausky, D. F ....... 1448, 1605 Schmidt, L ................. 790
Mellinger, E. C .............. 670 Paulet, S ............. 1620, 1648 Schomette, D ............... 335
Melvin, K. P ............... @ 1055 Pawlak, J. F ................. 719 Schroeder, K. R ............ 1390
Merrill, R ............ 1012, 1025 Pearce, J. B ............. 732, 786 Schroeder,P ................ 268
Merry, S. L ................ 1315 Perreault, G. R ............. 1356 Schroeder, W., W .......... 17,22
Mes, M. J .................. 1395 Perry, R. B ................. 366 Schultz, A. W ................ 17
Mesecar, R ............. 505,518 Pfeiffer, T. F ................ 612 Schultz, H. E ................ 920
Meyer, A .................. 1018 Phillips, J. C ................ 372 Scott, G. P ................. 819
Meyer, R. M ................ 146 Picciolo, A ................ 1400 Scott, J. L ................. 1231
Michelena, E. D ............ 1649 Piccioni, R. M .............. 1362 Sendlein, S. L .............. 1315
Middleton, R. W ............. 123 Pinkney, A. E ............... 987 Serchuk, F. M ............... 975
Miller, J ................... 1025 Pleass, C. M ............... 1049 Shaw, A. R., III ............ 1384
Miller, R. E ................. 172 Ponsford, A. M .............. 485 Sheehy, D. J ............... 1470
Miller, R. R ................ 1650 Potter, C. W ................ 815 Sheifer, I. C ................ 937
Millett, C .................. 1276 Powell, E ................. 1178 Sheng,j ................... 413
Mills, W .................. 1018 Predoehl, M. C ............. 1569 Sherman, T ................ 1265
Mimken, G ................ 1627 Prinz, S ................... 1025 Shevenell, A P. . . 623, 1259, 1276
Mingrone, J. A ............. 1618 Pryor, D. E ................. 379 Shumaker, F. E ............. 1149
Mink, R. M ................. 762 Quoidbach, D .............. 1025 Siu, K. W. M ............... 1716
Mitcheson, G. R ............ 1049 Rabinowitz, P. D ........... 1012, Slade, D. C ................ 1006
Moe, R. L .................. 532 1018, 1025 Smith, C. E .................. 250
Molinari, P ................ 1609 Raer, Y .................... 602 Smith, G. A ................. 106
Momma, H ................ 1253 Rausch, K ......... ** ... *- .84 Smith, J. A ................ 1405
Monkelien, K ............... 699 Ray, P. K ................... 193 Smith, N. D ................ 1374
Moore, J. M ............ 385,509 Reed,J.C .................. 202 Smith, R. J ................. 618
Morton, J .................. 899 Reed, M ................... 847 Smith, s. m ............. 385, 549

xxix.

Smolowitz, R. j .............. 975 Valent, P ................... 397 Zaitzeff, j ................. 1629
Soden, D. L ................ 891 van der Valk, A. G ............ 46 Zegowitz, V ............... 1400
Sorenson, S ............... 1612 Vauthier, P ................ 1029 Zeller, R. W ................ 905
Spears, A A ............... 1173 Veentjer, j ................. 1487 Zielinski, A ................. 287
Spooner, R. L ............... 283 Vetter, W. j ................ 1540 Ziemer, F ................. 1212
Sprunk, H ................. 1286 Vik, S. F .................. 1470 Zikakis, J. P ................ 1608
Srivastava, S. K .............. 485 Villemarette, G. P ............ 298 Zimmer, R ................. 794
Stamulis, D ................. 623 Voudouri, E ........... I .... 641 Zippin, J. P ................ 1615
Stancampiano, J ............. 709 Waddell, P. J. A ............. 833
Stanford, H. M .............. 745 Wade, T. L ................ 1198
Stang, P. R ................ 1616 Wagner, j .................. 518
Steeves, G .................. 567 Wainwright, P. F ............ 405
Steffy, D. A ................. 235 Waldmann, C ............... 497
Stein, S. M ................ 1642 Walsh, E. j ........... 1697, 1704
Steinhauer, W. G ........... 1630, Walters, R. A ............... 1405
1631, 1632 Walther, M .............. I ... 454
Stevens, P ................. 1400 Ward, j .................... 637
Stewart, G ................. 405 Weaks, M. L ............... 1626
Stewart, H. B., jr ............ 840 Weaver, R. j ............... 1362
Stewart, L. L .......... 1286, 1610 Weiss, P .................. 1025
Stockdreher, E ............. 1212 Weissman, D. E ............ 1546
Stoddard, A ................ 942 Wells, J. M ................ 1305
Storey, D. A ............... 1348 Welsh, B. L ................ 1633
Strahl, D ................... 549 Welsh, R .................. 1259
Strandquist, J .............. 1124 Werme, C. E .......... 1631, 1632
Stumpf, R. P ........... 808, 1569 Wheaton, F. W ............... 76
Sturgeon, W. j ............. 1163 Whelan, W. T ............... 650
Sullivan, P. K ................ 606 Whitehead, J. R ............ 1507
Sutherland, W. C ............ 632 Whittaker, H ................ 853
Swanson, J. C ............... 740 Whyte, R ................. 1646
Swanson, R. L ............... 794 Wicklund, R. I ......... 119, 1614
Sweet, W. E ................ 202 Wilber, R. j ................ 1639
Swift, M. R .................. 588 Wilcox, R ................. 1025
Tadjvar, A ................. 1628 Wilkinson, D ................ 90
Tattersall, J. M ............. 1618 Willard, M. R .............. 1625
Tauxe, j .................. 1018 Williams, K. L ............... 428
Taylor, P. R ................ 803 Williams, R. G .............. 346
Teal, J. M .................. 177 Wilson, B ................. 1202
Teel, S .................... 1582 Wilson, D .................. 359
Tennyson, E. j .......... 853) 857 Wilson, E ................. 1178
Tester, P. A ................. 808 Wilson, W. B ............... 629
Thomas, A ................ 1045 Wingert, R. C ................ 150
Thomas, W. M., jr .......... 1656 Winslow, T. S .............. 1390
Thomasson, A P ............. 70 Wolniakowski, K. U .......... 994
Thompson, B. G ........... 1613 Woodward, W. E ........... 1348
Thurberg, F. P .............. 736 Woolweaver, G ............ 1265
Tompkins, A E .............. 35 Wright, A. St.C .............. 323
Traub, G. L ................ 1493 Wright, D. A ................ 987
Treacy, S. D ................ 180 Wu, L ..................... 287
Tremblay, H ................. 522 Yamazaki, T ............... 1067
Trout, T. K ................ 1668 Yee, M. C .................. 555
Trulli, W. R ................ 1630 Y-Hsieh, T. C. T ............. 84
Trutneff, H. F ............... 902 Young, A. G ................ 423
Tsuzuku, T ................ 1061 Young, D. R ................ 745
Turner, E ................. 1616 Yu, D. W ................. 1098
Turner, R. E ................. 41 Zachritz, W. H. ,II ............ 84
Tuxhorn, R. L ............... 691 Zagel, J. j ................. 1642

xxx.

Deep-Ocean Tests of an Acoustic Modem
Insensitive to Muleipath Distortion

Winfield Hill, Gerald Chaplin, David Nergaaxd

Sea Data, Inc., A Pacer Systems Company
1 Bridge Street, Newton, Massachusetts, 02158

Abstract by multiple sound pathways, bottom and surface scattering
and moving inhomogeneities in the ocean('). However the
The concept, design and ocean testing of a new low-power repeated observation of such degradation has obscured the
acoustic modem is presented. The telemetry system em- fact that sound transmission quality over direct vertical or
ploys a novel "chirp" frequency sweep and has other fea, slanted pathways (other than in a sound channel) may be
tures to allow operation in the presence of multipath inter- quite good(').
ference. The chirp system uses fsk data modulation and Kearney and Laufer(") demonstrated this point whil e de-
performs a carrier sweep starting at 9 or 31kHz, depend-
ing upon the model, to obtain the benefits of frequency livering a paper at Oceans '84, by playing a cassette tape
diversity without requiring a frequency synthesizer, multi- recording of voice and music transmitted from 1500 me-
ple filters or a FFT analyzer. Intended for retrofittable to ters depth to a shipboard recorder; my memory is that the
existing instruments, the new system is designed for use in primary degradation was due to the use of a very poor cas-
the deep ocean and the continental shelf over distances to sette recorder. Designers of acoustic high-resolution pos-
6km. Ocean tests were performed in about 4000 meters of tioning systems have long taken advantage of good direct-
water using the low frequency version. Additional shallow- transmission paths by detecting the axrival of short acoustic
water tests are planned, including a typical harbor. pulses using narrow-band (Q > 30) filters('). When used
at low frequencies (10kHz), these positioning systems re-
ocean acoustic telemetry multipath quire several milliseconds of phase coherence in the leading
frequency diversity modem chirp edge of the pulse. Short-range (< 400m) acoustic telemetry
systems have been constructed (6,7) using simple frequency-
shift keying (fsk) modulation in the expectation of a reliable
acoustic path, with some success. One system transmits at
.1. Introduction the v*ery slow rate of I bit-per-second(') to achieve up to a
1000m range.
Real-time observation of data is a commonly desired capa, 2.1 Multipath. Despite the good quality of a direct path
bility which is not commonly available in undersea oceano- or reliable-acoustic-path signal transmission channel, most
graphic instruments. Although convenience and peace-of-
mind may be occaisional motivations for these desires, strong practical underwater acoustic systems must contend with
axguments have been made for the value of this capability('). strong undesired signals scattered from the surface or the
These include use in real-time operational systems, multi- bottom. This is especially true when one of the acoustic
year deployments (where it's impractical to wait until the transducers is near the ocean surface. Superimposing the
end for the data), performance monitoring, repair flexibil- surface-scattered signals upon the direct-path signal causes
ity and expendible instrumentation. We report here on the fading and phase instabilities, possibly including complete
design and initial ocean tests of a new chirp acoustic teleme- cancellation of the desired signal for a few milliseconds from
try method, which has simplicity and reliability properties destructive interference. Therefore, a pulsed one- or two-
desirable for fitting acoustic telemetry data links to existing frequency signal, which begins with good receive quality,
deteriorates as multipath interference arrives.
:undersea instrument designs.
As an example of simple two-frequency fsk telemetry per-
formance when surface and bottom scatter have a strong
2. Background - Ocean Acoustic Telemetry influence, consider the experiences of Ryerson at Sandia
Labs('). Transmission from a 10m subsurface buoy with
Underwater sound travelling a substantial distance in the a slant range of 180 to 280 meters to a surface buoy was
sea suffers from severe amplitude fluctuations and phase desired. Water depth was 200 meters. Optimum perfor-
distortion. Acoustic temporal incoherence may by caused' mance was obtained only after a variety of system-tuning

CH2585-8/88/0000-275 %1 @1988 IEEE

changes were made. Operating frequencies were selected When the transmitter is straight below (A = 0 and we
(near 50kHz) to reduce transducer backside and side-lobe 0), equations (2) above simplify to:
response and to attenuate long, multiple-reflection paths. d
In mid-experiment, the receive transducer depth was in- Pj + Vh2 + -(dt.B)2 - (h - d) (2c)
creased by one meter. Also, lower error rates were achieved cos B
with a -12dB power change (0.6 watts instead of the design
level of 10 watts). An 85 to 90% success rate was achieved. The first delayed surface-scatter multipath arrival occurs
at Td = 2d/c, when the axrival angle B = 0 (surface angle
= 90*), followed by more sound arriving for B > 0. The
WC B first arrival delay is about 27ms for a receive hydrophone
depth of 20m.

For slanted sound paths (A > 0) equation (2a) shows that
T the surface-scattered first-axrival delay time is slightly faster
d than for the direct overhead case; the shortest path occurs
for equal angles of incidence and scatter at the surface. As
an example, for a transmitter in 3000m of water, to a 60m
h deep hydrophone at a 2000m watch circle distance, A = 34 0
so the surface incident angle (given by 90'-A) is about 55'.
Sound scattered at 55* from the surface (B = 351 will ar-
rive with a 67ms delay (compaxed to 80ms for the straight-
below case). Straight-line sound travel has been assumed
throughout, even though for a slanted direct path sound
travel is actually slightly curved, due to refraction by the
sound-speed depth profile; This does not affect our conclu-
sions.

After the first multipath axrival, sound travelling longer
paths continues to arrive for a substantial period of time;
2.2 Surface Multipath. A common surface-path sit- this additional sound constitutes most of the multipath in-
uation is illustrated in figure 1. The offending surface- terference. Some of the sound has travelled very complex
scattered (backside arrival) signals clearly travel a longer pathways, involving volume scatter as well. Although the
path than the direct path signal and therefore take a longer multipath signals suffer surface-scattering losses(16) of 10 to
time to arrive. The earliest-arriving scattered signals take 20dB, the beneficial effects of these losses are reduced by
a,n extra delay time (Td) to travel an extra path delay (Pd) the large area of the surface. The desirable losses are fur-
as follows: ther reduced during high sea-state conditions, when acous-
tic surface scattering increases (e.g. see the backscattering
curves in ref 15 p. 264). However, for frequencies above
Td = PdIC (1) 20kHz (e.g. 33kHz), wind velocities above 10 to 15 m/s
may actually cause reduced surface-scatter sound due to
sound attenuation by small-bubble populations in the top
5 meters of the ocean(17).
Pd = d +,lh2+[(h-d)tanA-dtan -h-d
@os_B cos A 2.3 Fighting Multipath. Several methods have been sug-
(2a) gested to reduce signal degradation by multipath interfer-
ence. One is to use a transducer with high back rejection
where d is the receive transducer depth, h is the transmitter (or use a baffle). In the 3000m example above, the first
depth, A is the transmitting slant angle, B is the scattered- offending sound arrived at an angle of 145* from the trans-
signal receive angle (both angles axe measured from the .ducer forward direction (given by 180' - A, assuming the
vertical) and c is the speed of sound in seawater, about transducer is pointed down). A second method is to cre-
1.5m/ms. The surface watch-circle radius (we) i s related ate a highly-directive receive transducer array("'). These
to the slant angle by we = (h - d) tan A. If the watch circle approaches increase the cost of the system, are painful to
radius is known instead of the slant angle A, equation (2a) implement at low frequencies and have limited utility for a
can be written: variety of reasons. Furthermore, in shallow water, directive
sensors may not be very helpful.
d + Vrh -2
Pd + (we - dtan By - Vr(h - d)2 + WC2 Acoustic transmission in shallow water is much more dif
cos B ficult than in deep water, since it suffers from the exis-
(2b) tance of many strong sound pathways to the destination,

276

iinvolving varying numbers of surface and bottom reflec- where TS is the duration of the sweep and f, and fs are
tions. Computer modelling(7,") indicates that for 10kHz the starting and ending sweep frequency and M(t) r= 0
transmission in 200m deep water, the Direct-to-Multipath or 1 according to the data bits. The modulation ampli-
signal Ratio (DMR) may be as poor as 6 dB at ranges of tude, f2, is chosen large enough, e.g. > 15OHz, to eliminate
less than 1000m. Actual measurements in the ocean may doppler-shift spreading problems, which will be less than
give even poorer DMR. Higher-frequency transmissions will 40Hz (0.33Hz/kt per kHz).
experience increased sea-water absorption attenuation for If the receive frequency is accurately swept to match the
the longer multiple-bounce pathways, but less than 5dB of
improvement is calculated at 50kHz, due to this effect. transmitter, a small receive bandwidth (constrained by the
data rate and the fsk 0,1 frequency shift) can be used, just
Frequency-diversity sound transmission methods have been as in a conventional fsk system. A small bandwidth will im-
developed("') to solve the multipath problem. Systems prove the signal-to-noise ratio (SNR) not only by rejecting
with many frequency channels(",",") have been proposed, ambient noise but also by rejecting the (delayed) multipath
even up to 32 frequencies (14), so that the system can switch energy from the "old-channel" frequencies.
to a new frequency before the multipath interference ar-
rives. In a common approach, the frequencies in use are In a chirp telemetry system, the effective frequency-diversity
Ichanged every 50 to looms, allowing the multipath energy channel usage time (T,,) can be equal to the time required
to decay on the old channel. Since the decay time allowed for the carrier frequency sweep to change by more than the
before a channel can be reused is proportional to the num- receive bandwidth (BW), as follows:
ber of available frequency channels, this may well be a true
case of "more is better". Of course the telemetry system TS
will become more complex, but the improved results that T,, = BW f3 - A (6)
.Can be obtained in all environments are very attractive.
The usage time can be easily set at under 50ms (e.g. BW
= 30OHz, sweep 400OHz in 650ms), allowing excellent re-
f3 jection of multipath signals.

As an added benefit, the new chirp telemetry approach
can be inexpensive, compared to other frequency-diversity
START BIT methods, since multiple frequencies are not required (i.e.
fp no synthesizer) and receive decoding can be simplified (Le
,no multiple filters or FFT analyzer). To understand our
f, approach and the role of all the elements in the sweep wave-
form of figure 2, we'll start by considering how the receiver
f 0 1+- to -*I t, works (see figure 3).
3.1 Signal Description. Because a telemetry receiver
contains many circuit elements that consume electrical power,
3. A New Chirp Telemetry Method it's desirable to switch the power to a portion of these cir-
cuits. In the receiver design above, a number of components
The new Sea Data chirp acoustic telemetry system is based are continuously powered in order to detect the arrival of
upon a variation of the frequency-diversity idea: use an "in- an alert signal. These are the preamp, Al, a bandpass fil-
finite number" of frequencies. This is achieved by sweep- ter, BP, the fo detector and a power-control circuit (for our
ing the telemetry carrier frequency while applying fsk data experiment, fo = 9.0kHz). The bandpass filter is designed
modulation (fig 2). The transmitted signal P(t) is a single to pass signals over the entire fo to f3 range of the system
frequency, starting at fl, and changing at a smooth rate (a double- or triple-tuned filter) and to reject intense low-
df Idt, plus fsk frequency shifts with amplitude f2: frequency noise from shipping, etc (extra LF cutoffs). The
fo energy detector operates on a principle similar to that
used by many high-resolution acoustic positioning systems
P(t) = cos[w(t)t] (3) (5): an amplitude limiter (to establish constant power), a
sharp fo filter and a comparator with a time constant, work
together to determine if the fo energy present is above the
f background noise adjacent to fo by a threshold amount.
W(t) == 27r h + �-t + f2M(t) (4)
( dt When the fo alert-tone energy is detected, the remainder
START B@IT@
fp
fl

of the receiver, including the microprocessor, is turned on.
df f3 - fi (5) After a short time, to, the transmitter shifts its frequency to
dt Ts f, (for our test f, = fo + 60OHz), creating the data-trigger

277

BP
Al F0 POWER
_iR DETECTOR CONTROL
SWITCHED
POWER
LIMITER F, SWEEP
tone. This trigger tone "start pulse" is detected by the DETECTOR GEN
f, energy detector (sirnilar to the fo energy detector) and@.
.is used to start a sweep generator and voltage-controlled fL0
oscillator (VCO). The resulting frequency ramp is designed MIXER
to precisely track the transmitter's sweep with a fixed offset, +5
frp, where f1F > (f3 - fo)/2 to avoid images. This ramp
frequency is the local oscillator (LO) input to a mixer, and. LP R2
.has a frequency, Ao , similar to equations (3) to (5) except LF
as follows: VC0
A0 = At) = h + @f t + hp (7) Cl RI
dt
The resulting intermediate frequency (IF) output from the CLAMP
mixer after IF-stage filtering is:
v(t) =zz sin[27rf (t)t] + n(t) (8) manner to maintain the VCO output frequency near w(t)
in formula (9). Further filtering of the input to the VCO -
a varying dc voltage v(t) oc w(t) - along with ac-coupling
f (t) = fIF + AM(t) (9) and clamping, yields the original data-stream signal, M(t).
The PLL loop filter and low-pass filtering of M(t) set the
where n(t) is the received noise, with a noise bandwidth noise bandwidth, BW, of the telemetry receiver. The re-
given by the IF-stage bandpass. This signal is limited and ceiver should be able to operate with very low SNRs, al-
applied to a frequency discriminator to track f2 and deter- though the data error rate may not then be zero.
mine whether M = 0 or 1. A data precursor time delay, t,
in fig 2, allows the circuits to settle before data discrimina-
tion must start. Also a warmup time, tw = to + t1, less the 4. Mransmitter
fo detect time, is available for-the crystal in the receiver's The transmitter (figure 4) helps illustrate the simplicity
microprocessor to s .tart, etc. of the chirp telemetry scheme. A few low-data-rate con-
The frequency discriminator in fig 3 is a phase-locked loop trollable outputs from the instrument's microprocessor are
(PLL) circuit, which forms a tracking filter to further nar- sufficient to operate the transmitter. These outputs include
@row the noise bandwidth of the receiver. The input stage the sweep generator power control, an enable for the output
of the PLL is a limiter that responds to the strongest sig- driver as soon as the VCO is stable, a start pulse (SP) shift-
nal within the IF bandpass and acts to reject any weaker ing the frequency for a data trigger, a sweep enable (SE)::
signals, thereby further rejecting (quieting) unwanted mul-@' and the data bit (DB) modulation signal. Another line sets :
tipath signals. A full-wave mixer phase-detector circuit the sweep rate (SR) to allow optimizating the system for
(exclusive-OR) and the PLL loop filter act in a deepsea or shallow-water use.
noise-insensitve

+5
POWE

BATTERY LI
V8
SYSTEM SWEEP RI
PP GEN/ OUTPUT
MODULATOR DRIVER X1

ENABLE TI
3R

S
P
P OF
S
SE
S
TD

278

At low frequencies, e.g. lOkHz, obtaining a transmit op- 4.2 Design Simplicity. Because I always miss the ab-
erating range of 5kHz is a challenge, due to the narrow- sence of electronic-circuit schematics at IEEE conferences,
band nature of a tuned acoustic transducer. In figure 4,' I'll be sure to include one here. Figure 5 shows details of
the reactive component of transducer X1 is removed us- the transmitter sweep generator and serves to further illus-
ing tuning coil L1, with a series resistor R1 to increase the trate the simplicity of our new approach, while giving me
frequency range. In the 9 to 14kHz experiment to be de- a chance to dispel any concerns over drifts, tuning, etc.
scribed, a modified ITC type 3013 transducer (which nor-
mally has transmit-voltage-response peaks at 9 and 14kHz) The most important component is the voltage-controlled
was used with a 22mHy choke and a 50 ohm damping resis- oscillator (VCO) chip U4, an Analog Devices AD537, which
tor. A very acceptable calculated network output flatness operates at twice the transmitter output frequency. This
(+140�2dB/V) was obtained over a 7.5 to 14kHz range, VCO chip creates a very stable frequency and has low
and verified with pulsed measurements in the local YMCA power-supply and temperature drift coefficients (0.01%/volt
swimming pool. If necessary, a more complex network could and 0.03%/10 degrees C). When used with stable compo-
be devised. When operating the system with 5kHz sweeps nents (capacitor C1 and resistor RIO are low-tc compo-
at 33kHz, using a custom-designed transducer, a damping nents), the AD537 may allow a circuit with lifetime factory
resistor is less important. calibration. The VCO follows the formula f = V3/[IO(R9
+ RIO) C11. Here R9 sets the exact coefficient for the VCO
4.1 Power. In the deep ocean test, the output stage con- frequency-programming voltage, V3, which comes from am-
sisted of a pair of VMOS transistors driving a center-tapped plifier A3 (LM10, chosen for ImA sink capability when
transformer with a regulated 12V input. This provided V,,,,t = 0.2V at the end of the sweep). This amplifier's
about 20 watts of power into the transducer network and summing junction allows the telemetry system operating
yielded a modest calculated source level of +179dB re 1jUPa parameters to be exactly ratiometrically determined by pre-
at 1m, confirmed in the pool test. The current drain from cision resistors RO, R1 and R2 according to the following
the instrument battery was less than 2A during transmis- formula:
sion, a very acceptable level for any instrument with several
3tacks of alkaline batteries.
Although lower power levels may be used in practise, our 2f = 2k F, + dFj (10)
thought was to get good quality data on the experiment I )
DAT tape and subsequently degrade it with noise when we where F0 = I/RO sets the fo alert frequency, dF, = 1/R1
tested transmit codes and receiver designs in the lab. How-: sets the f, - fo data trigger frequency shift, dF2 = 1/R2
aver the higher-power energy usage is not unattractive: At sets the f2 fsk modulation level and dF3 = 1/R3 sets the
300 baud, less than 0.1 Joules per bit is required, including sweep rate (and hence f3). Amplifier Al (OP-20, chosen
alert tone, etc. Since a single stack of alkaline D-cells con- for low offset voltage) creates a reference 1-volt above the
tains about O.5MJ of energy, it could power about 100,0M amplifier-reference signal (also 1 volt), so that k ='R8. In
transmissions of 50-bit data blocks. ithe experiment, an electronic switch selected two values of'
R2 to allow-two fsk modulation levels.
IK ... C2 X44( Amplifier A2 (OP-90, chosen for low input current and off-
,SE set voltage) is a ramp, which operates (when switch SE
R4 R3 opens) with an integration constant r =(R4+R5)C2. The
integrator uses voltage source trim R6 to allow two cali-
+1 28-OK R8
OP-90 6 YA brated sweep rates according to the resistor ratio R4/R5
SR R5 4.99K M,
R6 R7 +2V +1 >_ 0 C13nFX440
2 166K 1% DLMIO IR9 12 +5
+1 + RO 200:1 F -'I
Al --- Emitter - Coupled fIOK
RIO MULTIVIBRATOR 14
OP-2 R 1 826 :
3
2f
li@. 140*AOK SID 4 1
RZ 5
1_@@ -
+1 D13 7
+1 1.00 volt
j___ once U4 AD537
R4 @SE VR 0.1
-I@
np_.
R -
r
R@
R(
20K 46K
I%

1% 1OOK I- SI
+1 5
$7 1.00 volt
'T8

279

and switch SR. The sweep generator operates on 5.0 volts signal strength o +99dB re near the surface, is about
(�5%) supply, and the entire circuit requires only two sim- 16dB louder than the wind noise for 20m/s (NSL = +59dB
ple calibration points, yet we're able to get our "infinite" at 10kHz), assuming a 30OHz receiver bandwidth (+25dB)
frequency channels. and an isotropic receive transducer (DI = OdB).

5.1 Shallow water. Using 33kHz in shallow water at
5. Signal Propagation Loss, Noise, SNR 10'C, 3km of range will result in about -22dB of absorption
loss, assuming the actual (scattered and reflected) path is
The expected signal-to-ambient-noise ratio (SNR) can be 30% longer than the range. Since the sound is in a channel,
,calculated (in dB) by subtracting the speading and attenu- the spreading loss may be less than the -69dB value from
ation losses and the background noise level from the trans- formula (11), say +10dB for 30m water depth (2). The
mitter source level: nal system SNR is similar to the case above since NSL is a
bit lower at 33kHz. Because 3km of range in shallow water
will be subjected to severe multipath interference, the sweep
SNR = SL - 20 log(r) - ce r _ NSL-10log(BW) (11) rate may increased and the data rate may be decreased to
1000 combat this. Also, the telemetry system's processor can
easily allow using slower data rates, with 25ms dead peri-
where SL is the source level (dB re ljzPa at 1 yaxd), cor-, ods in between each bit, to allow the immediate multipath
rected for the transducer directivity index, r is the range energy to decay.
(yards - not kni), ce is the seawater attenuation coefficient
(less than IdB/km for frequencies below 15kHA, NSL is the 6. Sea Mial using a DAT Recorder
ambient-noise spectral level (dB re 1,uPa/N/Hz), and BW
is the receiver bandwidth (Hz). The equation assumes a The experiment was performed on 16 to 17 June 1988 dur-
low-noise receive preamp and does not include the improve- ing cruise OC200 of the WHOI vessel RN. Oceanus, at a
ment a directive receive transducer will provide in rejecting site approximately 400 miles east of Cape Hatteras, just
(wind-generated) surface ambient noise, which could exceed north of the Gulf Stream, in 3775m of water. The undersea
.6dB. transmitter for the experiment operated over a range of 9
The seawater attenuation is due to magnesium-sulfate ionic to 14.5kHz and was installed in a Sea Data model 1665 In-
relaxation with an absorption coefficient of about 0.7 and verted Echo Sounder (IES), deployed on the bottom. The
5.5 dB/kyd (at 100C and zero depth) for 12 and 35Hz, re-: :receive hydrophone was the standard EG&G acoustic re-
spectively (see ref. 15, page 109 and ref. 22). Over the lease deck-set sensor (an ITC 3013 transducer), suspended
range of 8 to 5OkHz, the absortion coefficient increases by over the side of the ship about 18m below the surface. A
the square'of the frequency, decreases about 7% for each custom-built preamp with a 5kHz 2nd-order bandpass filter
1000in of depth and increases about 2% for each 'C of tem- was used with the hydrophone.
perature decrease. The latter two effects tend to cancel
To test the new telemetry system, we elected to transmit
each other out in the top half of the deep ocean. Applying
test signals from the ocean bottom to various lab-based re-
the formulas to expected ocean conditions yields the values ceiver circuits via a shipboard precision audio recorder. In
below, which can be integrated over 'the sound propaga-' this way we could perform receiver tests in the lab with var-
tion pathways to determine the absoption loss for various ious, noise levels and different types of interference, using a
telemetry applications.

Depth (m) Temp (*C) Attenuation (dB/kyd)
@10kHz @33kHz
0 20 0.65 4.5
3000 4 0.50 3'.3
6000 4 0.38 2.4 d

SURFACE
When using the system in deep water, with a 5krn path, we SCATTERED RECEIVE.
PATH
can calculate a 16dB expected SNR for (poor) 20m/s wind
conditions, as follows: Given the TVR of the transducer DIRECT
at +141 dB per volt, and considering a 1.5dB; loss for the PATH
TEST
50ohm. tuning resistor, we can calculate an output acoustic ROGRA
DAT

RECORDER

intensity of +178dB, for 20 watts (this was 'confirmed in
the pool test). We lose -74dB from 5km spreading and- 6
5dB from attenuation (at 10kHz). The resulting calculated TRANSMIT
77777-7.77" Pr7-r777

280

large variety of transmission types as they were actually. 7. System Considerations
,received in. the- ocean. We used a small portable 16-bit dig-
ital audio tape (DAT) recorder (Technics mo, del SV-MD1, The receiver and transmitter of the acoustic modem each
complete with a manual entirely in Japanese). Thus we occupy one card, as does the processor. The 33kHz trans-
were able to obtain "perfect" (90dB dynamic range, flat to ducer is very small, 1.6-in (4cm) in diameter, and is con-
18kHz, 0.01% time stability) digital analog recordings of structed with an O-ring groove and 3/4-16 stud with em-
the received hydrophone signals. bedded wires, to allow it to be screwed directly into an
endcap. Thus, the system can easily be added to many
During the experiment the transmitter variables were cy- existing designs. A standalone version mounted in a small
cled through a variety of combinations using a parameter housing with a battery is planned as well.
table in the microprocessor's program. The parameters in-
cluded: data rates (100 to 360 baud), modulation index The final telemetry system software will employ a data
(300 and 50OHz), chirp sweep rate (10 and 40kHz/s), chan- transmission protocol suited for systems applications, and
nel decay "quiet times" (62ms to 14s), transmit duration (1. error checking features. A unique code can be sent from
to 10 bytes/record and 3 to 80 records) and the transmitted each transmitter for identification. Controlled redundancy
data patterns. can greatly reduce the eff or rate: block error-correction
codes such as the Reed-Solornon code (21) can allow for cor-
During the experiment the receive variables included slant rection (after reception) of up to 15 errors within a 155-bit
range and Dolphin activity level. Winds of 10kts and a
.block while achieving an 80% code rate (125 data bits).
steady rainfall both occurred at various times during the
experiment. The experiment was performed with the low- Although both receiver and transmitter cards will often be
frequency version (fo = 9kHz), since all the available com- located at both ends of a system, creating a full underwater
ponents (IES transmit stage, receive hydrophone and DAT MODEM, telemetry systems can be substantially simplified
recorder) weighed against the 33kHz version. if one-way data transmission is used. If stable timebase os-
.cillators are employed("), offset time-slot channels may be,
Initial oscilloscope examination of the DAT tapes shows 3 established so that many undersea instruments can trans-
to 10dB of fading after the sweep was under way (due to mit to a central receiver (20) on a single frequency, without
multipath?), -6 to -10dB of delayed (obvious) multipath requiring a command for the transmission. Furthermore,
interference and a +10 to +20dB SNR (4kHz noise band- studies have shown(') that a one way acoustic data trans-
width), depending upon surface conditions. mission system can be optimal, e.g. "Analysis of the ADTL
data indicated that command and retry provided only min-
The concept of using a DAT recording to provide receiver imal improvement in the amount of data passed without
test signals has proven to be very useful. At this writing, errors."
excellent performance has been obtained playing back the
tapes into our prototype receiver. In this fashion we will
easily be able to optimize the performance of the receiver
design with tests using bench instruments, e.g. the SNR can .8. Conclusion
be degraded with noise generators. Already, we were able to
painlessly test the improvement that a CMOS-switch ana- It is our expectation that considerable improvement over
log mixer provided over a limiter/XOR-gate mixer. Further other traditional methods will be experienced with our new
DAT recorder ocean experiments are planned in shallow swept-frequency telemetry, at a reduced cost. It is our hope
water. that our work will help lead to a greater and happier use
of acoustic telemetry in the ocean.
6.1 Dolphins. We experienced considerable interference
from dolphins, who were curious about the ship and enjoyed
playing with the hydrophone. A few dolphins used their 9. Acknowlegments
variable-rate pulse sonar to locate and "ping" the trans-
ducer; at closest approach they increased the ping rate to One of us (GC) wrote some of the transmitter software
buzz.- Like our hydrophone, the IdolphinsIcould.hear the and singlehandedly (!) performed the undersea experiment,
transmitter on the ocean bottom. Amazingly, they did a during the wee hours when the rest of the ship was asleep,
good job of mimicking the 9 to 14kHz sweep signal of the while another (DN) modified the IES undersea transmitter
telemetry! But we haven't yet decoded their transmissions and constructed preamps and prototype receivers to ana-
(Does anyone know, do they use ASCII code? And if so, lyze the DAT tapes. Special,thanks are due to Kevin Boyce
is it Isb first?). It was necessary to move the ship several for creating a major portion of the original IES micropro-
times, and to turn off the fantail lights. This may be an cessor code, to Dan Frye and others at WHOI for their,
axgument in favor of higher frequencies, such as our 33kHz suggestions and review of telemetry system goals and to'
version. Prof. Randy Watts at URI for his encouragement.

281

(11) Zielinski, A. and Caldera (1985). Digital Acoustic Communications in Multi-
path Underwater Channels, Oceans '85, pp 12W1301.
10. References
(12) Caldera, M.K. (1987). A Multi-Frequency Digital Communication technique
for Acoustic Channel with Multipath, Oceans '87, pp 140-145.
(1) Parker, B.B. (1985) Real-Time Oceanographic Model Systems: Present and (13) Catipavic, J., Baggeroer, Heydt and Koelsch (1984). Design and Performance
Future Applications, Oceans '85, Proc. IEEE-MTS Conf., pp 204-214. Analysis of a Digital Acoustic Telemetry System for the Short Range Underwater
(2) Urich, R.J. (1982) Sound Propagation in the Sea, Peninsula Publishing, Los Channel, IEEE J. Oceanic Engineering, OE-9(4), pp. 242-252.
Altos, CA, chapter 10-12. (14) Gastounioties, C. and Moropoulos (1985). Programmable Deep Ocean Tran-
ceiver, Oceans '83, pp 145-149.
(3) Coffey, D.M. and Paquett@-flggS).--A-nuracy-ofA:cousticMultipath Timing-and-
Ranging Predictions over Extended Ranges, Oceans '85, pp 480-489. (15) Urich, R. (1983). Principles of Underwater Sound, 3rd Ed, McGraw-Hill.
(4) Kearney, P.O. and Laufer (1984). Sonarlink - A Deep Ocean High Data Rate, (16) Eller, A.I. (1985). Implementation of Rough Surface Loss in Sonar Performance
Adaptive Telemetry System, Oceans '84, pp 49-53. Models, Oceans '85, pp 494-498.
(5) Vijayakumar, G. (1982). Acoustic Navigation - New Microprocessor Generation, (17) Farmer, D. and Lemon (1984). The Influence of Bubbles on Ambient Noise in
Oceans '82, pp 100-105. the Ocean at High Wind Speeds, J. Physical Oceanography, 14 (11), pp.1762-1778.
(6) Scally, D.R., Ryerson amd Towles (1984). Acoustic Telemetry in an Automated (18) Zielinski, A. and Wu (1988). Data Retrieval from Bottom Instrumentation
System for Long-Term Ocean Data in Real Time, Oceans '84, pp 748-752. Using Acoustic Link, Instrumentation and Measurements in the Polar Regions,
Proc IEEE-MTS Workshop, MTS, Berkeley, CA 94702, pp 283-294.
(7) Towles, T.L. and Hauser (1986). ATZ - An Acoustic Telemetry System for
Collection of Subsurface Temperature Data from a Moored Buoy, MDS '86, Proc. (19) Hill, W. (1988). Engineering Considerations for Underwater Remote-Sensing
MTS Conf. on Marine Data Systems, Marine Tech. Soc., Wash. D.C., pp 178-181. Instruments, rbid, pp 325-329.
(8) Cronan, P.H. and Gonsalves (19i33). ENDECO Type 1033 Directional Wave & (20) Bowers, G. and Lanza (1986). Data Telemetry System for Wide Array Tem-
Current Telemetry System, Proc 1983 Symposium on Buoy Technology, pp 292-298. perature Sensing, MDS '86, pp 216-220.

(9) Jacobsen, H.P., Vestgard and Knudsen (1982). Acoustic Control System, Oceans (21) Backes, J.L., Bell and Miller (1983). Implementation of Error Detection and
'82, pp 106-110. Correction Codes for Acoustic Data Telemetry, Oceans '83, pp 167-175.

(10) Garrood, D.J. and Miller (1982). Acoustic Telemetry for Underwater Control, (22) Fisher, F.H. and Simmons (1977). Sound Absorption in Sea Water, J. Acous-
Oceans '82, pp 111-114. tical Society of Amexica, 62, p.558.

=.mm# frequency
10 Hz 20 50 100 HL euv 300 1k- z 5 10 kHz 20 50 kHz
heavy HAIL 80
SHIPPING
0m-o, d,e"r,at"e NSL
......... ....... ..... 30 m/s
qc- N - 70dB
20 light' rain
I Whr % 113,15 kHz peak
light 1

10
ro.0.51 in/hr%%,,
60
'RAIN
0.1
0.1 ln1hr-

50
A
IN10
15
L,nlhr

-40
0.0 in1hr.
NOISE SPECTRAL LEVEL 1 M -S ---12 30
typical curves as compiled by Sea Data 1 kt)
5

OdB I pPa /Hz Y2
e n s Nit i v i t \y
20

282

SIGNAL PROCESSING USING SPREADSHEET SOFTWARE

Ronald L. Spooner

Marymount University
2870 North Glebe Road
Arlington, Virginia 22207

ABSTRACT FIGURE 1

This paper demonstrates that LOTUS 123, a B8: @RAND+@RAND+@RAND+@RAND+@RAND
trademark of the LOTUS Development
Company, is an extremely useful tool for
active sonar signal design. Three features A B C
of LOTUS are important to the effort: (1) 1 SIN WAVE PLUS NOISE
the lotus random number generator, (2) the 2
LOTUS cell relational copy feature and (3) 3
the excellent LOTUS graphing capabilities. 4 DATA NOISE
5 SAMPLE DATA
The paper demonstrates that Spreadsheet 6 NUMBER SAMPLE
Software is a powerful tool not only for 7
accounting and business, but also for 8 1 0.774688
sonar signal processing. -

The result is a simulation of a white
gaussian noise waveform of as many noise
i. INTRODUCTION samples as we desire. The first 13 samples
of a 200 sample noise waveform are shown
The purpose of this paper is to show the in column B of Figure 2. The full noise
usefulness of Spreadsheet Software for wavoform is plotted in Figure 3 using
sonar signal processing. The steps consist LOTUS graphics.
of simulating a random noise waveform,
constructing a signal design, and then
performing correlation processing to find FIGURE 2
and evaluate performance.
B8: @RAND+@RAND+@RAND+@RAND+@RAND-
2. THE NOISE WAVEFORM

We create an independent (approximately) A B C
gaussian noise waveform using the LOTUS I SIN WAVE PLUS NOISE
LaRAND function which generates a uniform 2
random number between zero and one. By 3
summing twelve of these random numbers and 4 DATA NOISE
normalizing, we can create a zero mean 5 SAMPLE DATA
unit variance (approximately) gaussian 6 NUMBER SAMPLE
random variable (1). 7
8 1-0.32606
Figure I shows the procedure for the first 9 2-2.17980
noise sample data cell where the samples 10 3-1.42729
are numbered down column A and the noise 11 40.378947
data samples are in column B. 12 52.122943
13 60.026867
If we had to type the equation in cell B8 14 70.212809
for each of the noise data samples the Is 8-0.76279
process would be prohibitively laborious. 16 90.910396
Fortunately, however, all we have to do is 17 10 -0.28609
to use the LOTUS copy feature and copy the 18 11 2.127792
function to the desired number of noise 19 12 0.180714
sample cells down column B. 20 13 1.535139
19-Jul-88 01:2S PM

CH2585-8/88/0000-283 $1 @1988 IEEE

FIGURE 3 the remaining signal sample ce I I s. For
_____IIOIIE WAVEFORM this example we used two cycles of the
sine wave as shown in Figure 5. Also we
3- included a variable amplitude gain ($T$2)
so that the signal-to-noise ratio can be
2- easily changed.

FIGURE 5
SIGNAL WAVEFORM
n
D - @4

0.6

0.2
. . ........................ . ..................
G 100 150 20D
DATA SAMPLE -0.2
--0.4
Note that non-white noise can be produced __O.G
by defining relationships between samples, -0.0-
such as a single step markov process, and -Ij
then copying the relationship- down the [email protected]
.4
noise sample column.
0 50 100 ISO 200
3. THE SIGNAL WAVEFORM DATA SAMPLE

Generation 'of the signal depends on the
desired signal waveshape. For our first 4. THE SIGNAL + NOISE WAVEFORM
example we use a sine wave. The signal
samples are placed down column C and in Once the noise and signal wavforms have
time sequence are placed in the center of been generated they may be summed in
the noise waveform samples.The functional column D. The result is plotted in Figure
form of the sine wave is shown in Figure 6 for a signal-to-noise ratio (SNR) of
4. OdB.

FIGURE 6
FIGURE 4 SIGNAL + NOISE WAVFFORM SNR-OdB

C83:' +$T$2*@SIN(@Pl*(A83-$A$83)/iO) 3

A C 2
I SIN WAVE PLUS NOISE W
2
3
1 0
4 DATA NOISE SIN WAVE
5 SAMPLE DATA SIGNAL
6 NUMBER SAMPLE SAMPLE
7 -2
83 76 0.621254 0
84 77 -1. 89134 1. 3108SO
85 78 -0.79248 2.493385 0 so 100 ISO 2130
86 79 -1.60264 3.431850 DATA SAMPLE
87 80 -1.06422 4.034381
88 81 0.178056 4.242 S. CROSS-CORRELATION PROCESSING
89 82 -1.50562 4.034381
90 83 0.097219 3.431850 Cross-correlation processing of the signal
91 84 1.091418 2.49338S and signal plus noise is the process of:
9-, 85 -1.94063 1.310850 (1) overlaying the two waveforms, (2)
93 86 -0.50615 1.4E-18 multiplying the samples, (3) adding the
94 87 0.391116 -1.31085 products (integrating), (4) dumping the
95 88 0.316510 -2.49338 result as a single number and (4) shifting
19-Jul-88 01:29 PM @one sample and repeating the process.

The function was created by typing the Starting with the first noise sample, the
equation in cell C83 and then copying to crosscorrelation function is typed in the

284

corresponding cell of column E as shown in FIGURE 9
Figure 7. Typing this function is an error CORRELATION OUTPUT- SNR-OdB
prone process which if it had to be 50 -_ 7
repeated for each data cell would prove 40-
Spreadsheet Software useless for signal
processing analysis. 30-

20-

FIGURE 7 1

E8: +$C$83*D8+$C$84*D9+$C$85*DI04
-10

A E F -20
1 SIN WAVE -;R)

-40
3 0 50 100 150 200
4 DATA SIGNAL AND DATA SAMPLE
S SAMPLE SIGNAL PLUS NOISE
6 NUMBER CORRELATION
7 6. SIGNAL-TO-NOISE RATIO VARIATIONS
a 1 0.771082
T-he effects of signal-to-noise ratio
changes on processing performance may be
However, because of the relational copy examined by simply varying the amplitude
features of the software we only have to on the sine wave signal. Figure 10 shows
type the function once and then just copy the signal plus' noise waveform and Figure
it for the rest of the data cells., The 11 shows the crosscarrelation results for
software automatically indexes each a -3dB SNR.
function by one sample cell as shown in
Figure 8. (Note that the use of the $ FIGURE 10
SIGNAL NOISE ORM SNR--34B
prevents the signal waveform values from 3
being indexed).

FIGURE 8

E9: +$C$83*Dg+$C$84*DIO+$C$85*Dll-

A E F
I SIN WAVE
2
3
4 DATA SIGNAL AND
5 SAMPLE SIGNAL PLUS NOISE
200
6 NUMBER CORRELATION DO
7 0 DATA SAMPLE
8 1 2.867470
9 2 1.646376 FIGURE 11
10 3 0.644179 CORRELA'nONq.SNR--3d5
11 4 -0.34991
12 5 -0.62400 141
13 6 -0.43120 1
14 7 0.092879
15 8 0.797706
16 9 1.292642
17 10 1.487631 2-
18 11 1.662129
19 12 1.690333
20 13 1.700339
19-Jul-88 03:26 PM

The crosscorrelation processing results -121
for the sine wave example with OdB SNR are -14
shown in Figure 9. 0 60 100 1;0 200
DATA SAMPU
0
2
4

285

7. SIGNAL WAVESHAPE VARIATIONS FIGURE 14
CHIRP CORRELATIOMSM-Uo
The effects of signal waveshape variations 2a
(increased time-bandwidth products) can be 20
considered within the overall system
simulation structure by simply changing la-
the signal waveshape. For example, Figure
12 shows a chirp sine wave substituted for 10-
the conventional sine wave of the previous
example. a
IL

0
FIGURE 12
CHIRP SIGNAL WAYEFORM

a's-
0.8- -10
0.7-
0.0 -15 ................ ......
0.5 0 so 100 150 200
0.4 -
0.3- DATA SAMPLE
W 02
0.1
'a
-0.1 9. REFERENCES
-0.2
-0.3
-0.4 (1). W. Mendenhall, Introduction to
-0.5 Probability and Statistics, Duxbury
-0.7 Press, 4th. ed., 1975.
-0.8
i.. . ......... .. . ..... .....
0 50 100 180 200
DATA SAMPLE

Figure 13 illustrates the signal plus
noise waveform for an SNR of -3dB. The
crossoorrelation results are shown in
Figure 14.

FIGURE 13
CHIRP SIGNAL + NOISE SNR--3dB
3

2 -

-2

-4
0 so 100 150 200

DATA SAMPLE

8. CONCLUSIONS & CONTINUING WORK

The results presented above demonstrate
that Spreadsheet Software can be usefully
applied to sonar signal processing.

These results, however, pertain to the
noise limited case of sonar signal
processing. Future work is directed at the
more important and difficult case of what
to do when faced with reverberation
iimiting noise.

286

MULTIPATH REJECTION USING NARROW BEAM ACOUSTIC LINK

L. Wu and A. Zielinski

Department of Electrical and Computer Engineering
University of Victoria, Victoria, BC, CANADA V8W 2Y2

If, is intensity of n-th multipath (in dB)
ABSTRACT Id is intensity of the direct path (in dB)

Multipath suppression using narrowbeam link is dis-
cussed. Two types of narrowbeam systems are com- Surface Mullipath Surface
pared: umbrella-type beam produced by linear axray
and conical, searchlight-type beam produced by planar Recai@e
array. Specular and non-specular reflections are consid-
ered. It was shown that in both uses, linear array offers Bottom Multipath
better performance in rejecting multipath than the pla- Transmitter
nar array with the same number of elements.

INTRODUCTION Figure 1: Multipath Structure
The presence of multipath in underwater acoustic com- It has been shown [5] that the DMR increases with cen-
munication channels is the major limitation to reliable, tral frequency employed and ocean depth, and decreases
high-rate data transmission needed in diverse applica- with the range. For this reason, high frequency, near
tions. vertical transmission from deep ocean is least affected by
the multipath. For example, vertical transmission from
The multipath conditions can vary significantly depend- a 200m depth encounters D21R = 28dB at frequency
ing on sea state, ocean depth, type of bottom, sound f = 100kHz which drops to 15dB at f = 10kHz.
velocity profile, transmitter-receiver configuration and The onmidirectional transducers, smooth surfaces and 5
their respective radiation patterns. Because of a rel- Mrayls bottom acoustic impedance have been assumed.
atively high frequency carrier used for acoustic com- On the other hand, DMR can drop as low as 5dB for
munication the ray theory can be applied to determine very shallow water (50m) and long propagation range
the propagating paths linking a transmitter and a re- (1000m).
ceiver. Shown in Figure 1 are primary possible paths Since multi-path signals generally arrive from an .gular
linking transmitter and receiver in an isospeed acous- directions different from that of the direct path signal,
tic channel. The intensity of a particular multipath can
be calculated taking into account the propagation dis- they can be suppressed by utilizing suitable radiation
tance, number of surface and bottom reflections and patterns at the transmitter and the receiver. Such pat-
associated losses, and transmitting/receiving radiation terns will consist of a narrow mainlobe and small side-
patterns. The channel quality can be quantitatively ac- lobes. The transmitting/receiving beams should be kept
cessed by direct signal-to-multipath signal ratio. As- aligned along the direct path. Suppression of some mul-
suming noncoherent multipath signals, this ratio (in dB) tipath can lead to a substantial increase in the DMR
can be expressed as [5]. [5]. In order to ease the beam's al -ignment task, it is
also possible to employ a broadbeam transmitter and a
narrow beam receiver.
DMR log @ 1 (1)
Flo'-/10
n
Where

CH2585-8/88/0000-287 $1 @1988 IEEE

MULTIPATH SUPPRESSION The configuration shown in Figure 2b uses a relatively
broad conical fixed transmitting beam while the steer-
Receiver able receiving beam is narrow [3]. The beam alignment
for this configuration is less complicated. It will be
demonstrated that the system performance is not criti-
cally dependent on a transmitting bearnwidth.

Shown in Figure 2c is a steerable umbrella-type receiv-
ing beam, produced by a vertically suspended linear ar-
ray. Beam alignment and tracking for this configuration
is particularly simple. Subsequently it will be assumed
that a fixed, relatively broad conical transmitting beam
and steerable narrow (searchlight or umbrella-type) re-
Transmitter (a) ceiving beam are utilized.
Receiver The multipath signals can be admitted to the receiver
through the main lobe beam or through the sidelobes of
the radiation pattern. Sidelobe levels can be controlled
by proper weighting coefficients applied to each element
of the array. The Dolph-Chebyshev weights, for exam-
ple, allow for theoretically arbitrary suppression of the
sidelobe level [4,6,71.
P
The multipath signal can also be admitted to the re-
ceiver together with the direct-path signal through the
properly steered mainlobe. This will occur whenever
Transmitter the difference between the angles of arrival of the di-
(b) rect and multipath signals is less than the bearnwidth
Receiver of the mainlobe. In order to reduce the probability of
this occurrence, the mainlobe should be as narrow as
possible.

SPECULAR REFLECTIONS

L Let us assume that the multipath is caused by speculax
reflections from smooth bottom and ocean surfaces. In
such a situation, angles Op and f1L indicated in Figure
Tra nsmitIter 2b and 2c provide the correct measure of array ability
to reject multipath signals which might otherwise be
W admitted through the main lobe.
Figure 2: Configurations of acoustic beams Shown in Figure 3 is the ratio flP/nL vs the pointing
angle (off vertical) for arrays employing the same num-
ber of elements N separated by a half wavelength. 3dB
Three basic beam configurations depicted in Figure 2 bearnwidth has been assumed for calculations. It can be
a) b) and c) are considered. Shown in Figure 2a are seen that the linear array has a greater capacity of mul-
narrow searchlight-type beams for both transmitter and tipath suppression than the planar array for practically
receiver [2]. Both beam are produced by steerable mill- all ranges of pointing angles and array sizes.
tielement arrays. Although the configuration offers the
best rejection of ambient noise, a relatively complex Subsequently, we will consider a situation where a multi-
scheme is required to steer and to track the beams. Ad- path signal is caused by reflections from rough surfaces.
ditional difficulties arise from the fact that the steerable
transmitting beam must be able to handle the neces- NONSPECULAR REFLECTIONS
sary transmission power. This necessitates a number of
power amplifiers driving each element of the transmit-
ting array. To simplify the analysis we will assume zero sensitivity

288

of the receiving array in the upper hemisphere. This Since the receiving array is directional, only a certain
will allow us to neglect multipath arriving directly from amount of scattered power will be admitted to the sys-
the ocean surface. The ocean surface is assumed rough tem. This amount can be calculated by considering an
while the bottom will be considered to be a perfect re- intersection of the receiving beam (conical or.umbrella-
flection as indicated in Figure 1. For such a situation type) and the image surface as shown in Figure 5.
the scattering from the sea surface can be represented
by an image source as illustrated in Figure 4. umbralla S1
beam image
source

S 2

C
onical
&Tray size beam

N 64 49 36 25 Jr,

C:
Figure 5: Intersection of the Receiving Beam
and the Image Scattering Surface
The total mean power received from the image surface
is then obtained

0. 20* 40* 60* so*
Beam pointing angle P,. = 8 - f f G(O, 0, x, y, Oi, 8i) ds; i = 1, 2 (2)
Figure 3: Compaxison of beamwidths for planar 'i
and linear arrays where G (0, 0, x, y, 0j, Oj) is the intensity distribution func-
Scattedng Area tion of the mean power scattered in the direction (0, 0)
from each source point (X, Y) for a given incident an-
_.: . . . . . . . . . . . .
gle (0i, 60, si is the shaded region as shown in Figure 5,
ii: is
... ......
and is transmission loss along surface-bottom-receiver
path.

The intensity distribution function of the scattered field
in various directions for a given angle of incidence has
been discussed in[l] For the slight surface roughness
(standard deviation a = O.1A) and the long correla-
tion distance (T = 200k here T is the distance in which
the spatial auto correlation coefficient C(t) drops to the
value e-1) only 46.8% 0.468) of the incident en-
ergy is reflected in the speculax direction, the remaining
energy is diffusely scattered in all other directions.
Image Scaftedng Area
For rough surfaces with a = 10A and T = 200A, the
Figure 4: Applicatio n of Image Principle mean power reflected in the specular direction exceeds
to Rough Sea 37%. This means that for laxge T (gently rolling surface
Part of the incid-ent upon surface power Pi is reflected with long distance between hills and valleys) the spec-
in the specular direction while the remaining power P, ular direction remains privileged even if the surface is
is diffusely scattered in all directions as given by. [1]: very rough.
We will compare the average multipath power received
P. = (1 - ii)pi (1 - P)apt from the image surface by the linear array Pg and the
planar array P@, that is
where Pt is transmitted power P,1 f f., G(O 0, x, y, Oi, Oi) da
a is transmission loss from the projector (3)
to the surface PIT f L2 G (0, 0, x, y, Oi, Oi) ds
IV = 60 3. 2i M

and is the variance of the scattering coefficient p.

289

For simplicity we will assume that the intensity distri- CONCLUSIONS
.bution function G(-) is uniform on the image surface,
and therefore Reduction of multipath interference in a shallow water
P't S, channel has been discussed. It is suggested that a ver-
P', S2 tical suspended linear array is a useful tool for shallow
water, middle range acoustic communications and its
2 tan(%L1 performance is superior to that of a planax array with
2 1 sin 200 the same number of elements.

@50 tan 2Oo tan 2 - 9 tan4 0. - tan 4
22'_ 20) ACKNOWLEDGEMENTS
x tan-' 5 tan2 00 - tanZ
CA21
Nrlor __r2 This work was supported by Natural Sciences and En-
sin 28o X tan-' gineering Research Council of Canada Operating Grant
02
rp 5 - r (4) A6830.

where Ot and Op are the beamwidth of the linear and
planax array, respectively, 00 is the steering angle, -yt is
the beamwidth of transmitter, and References

tan 2(a) [11 Berkmann, P. and A. Spizzichino, The Scatter-
2
r tan 200 (5) ing of Electromagnetic Waves from Rough Surface,
Macmillan, New York, 1963, pp. 137-146.
120' [21 Catipovic, J., A. B. Baggeroer, K. Van Der Heyat
N 1 x64, 8x8 100* 110, and D. Koelsch, Design and Performance Analysis
001 of a Digital Telemetry System for the Short Range
Underwater Channel, IEEE J. of Oceanic Engineer-
70* ing, Vol. OE-9, No. 4, Oct. 1984, pp. @242-252.
[31 -Collins, J. S., and K. H. V. Booth, Simulation of
Multipath Channels Terminated by Narrow Beam
Non-parametric Transducers, Fourth International
Symposium on Unmanned Untethered Submersible
Technology, June 1985, pp. 124-139.
[4] Dolph, C. L., A Current Distribution of Broadside
Arrays which Optimizes the Relationship between-
Beamwidth and Side Lobe Levels, Pro*c. I.R.E.,
20* 30* 40* so* 60, June 1946, pp. 335-348; May 1947, pp. 489-492.
Beam pointing angle [51 :Howse, D., and A. Zielinski, Multipath Modelling
Figure 6: Comparison of Multipath Rejection for for Acoustic Communication, in Proc. Oceans '82,
Planar and linear axrays pp. 217-222.
The ratio PelPp changes only slightly when the num- [6] Stegen, R. J., Excitation
ber of elements N varies. Figure 6 shows the perfor- Coefficients and Beamwidth of Tschebyscheff Ar-
mance of two types of arrays; lineax and planar with rays, Proc. I.R.E., pp. 1671-1674, Nov. 1953.
the same number of elementn (N = 1 x 64 and 8 x 8)
separated by A/2. We can see that. at steering angle [7) Zielinski, A., Matrix Formulation for Dolph-
Oo = 45' and for transmitting beamwidth -Yt = 100* lin- Chebyshev Beamforming, Proc. of IEEE, vol. 74,
ear array gives 7.4dB improvement over planar array in no. 12, Dec. 1986, pp. 1799-1800.
rejecting the multipath.

290

COMPENSATION OF VERTICAL DISPLACEMENT COMPONENTS

IN MARINE SEISMIC APPLICATIONS

USING THE COUPLED HEAVE AND PITCH MODEL

FERIAL EL-HAWARY

TECHNICAL UNIVERSITY OF NOVA SCOTIA P.O.BOX 1000,
HALIFAX, NOVA SCOTIA, CANADA B3J 2X4

The motion's effects appear on the reflection records along
the ship track as additional undulations of the sea floor and of
the sub-bottom reflectors. When pitching is neglected this is
ABSTRACT commonly called heave effects. Removal of the vertical
In underwater seismic exploration using acoustic arrays the displacement component is an important preprocessing task for
aim is to determine the structure of the underwater layered improving displays of the raw and filtered reflection data, for
media in terms of significant geometric and physical properties extracting media parameters such as reflection coefficients
. In this method a towed body is employed to house both and reflector depths. Partial removal of these effects can be
acoustic energy source and the receiver array. The reflection done by use of estimates of the vertical source motions from
records received using this method, contain, undesirable noise hydrostatic pressure and motion sensors to delay or advance
components due to the vertical motion of the towed body. A the pulse firing instants relative to a clock pulse reference.
major portion of this component is compensated for usinj the This is done such that the source- bottom- sensor pulse travel
well known heave compensation technique. The intent 0 time corresponds to that of a source and sensor located at a
this constant depth relative to the mean water surface level.
paper is to address the issue of incorporating a more realistic
model for the vertical motions involved. Marine hydrodynamics references such as Bhattacharya [8],
The towed body experiences six different kinds of motion, Price and Bishops [9], and McCormick [10] discuss details of
three of which are linear and the other three are rotational modeling the vertical displacement hydrodynamics. We high-
light a model of the process, which is designed for subsequent
about the three principal axes. None of the motions occur Kalman filtering implementation. The model is derived in
singly, as they are all coupled in a manner that depends on
current and wave directions. For the purposes of this paper, transfer function form, and is then cast in state-space form.
the relevant motions are in the vertical direction which is The model of the vertical displacement dynamics is consistent
controlled by both heave and pitch effects. We highlight a with those found in the area of hydrodynamics. Spectral
model of the process, which is designed for subsequent analysis of field records provides the basis for modeling the
Kalman filtering implementation. The model is introduced in process via the transfer function, and parameter estimation
conventional form, and is then cast in state-space form. techniques.
Spectral analysis of field records provides the basis for This model provides the basis for formulating the vertical
modeling the process via the transfer function, and parameter displacement extraction problem as one of Kalman filtering in
estimation. techniques. The paper also discusses issues in the theory of optimal linear estimation. The design of the
preprocessing requirements as a result of incorporating the Kalman filter enables reducing the residual vertical dis-
effects of the pitch component. placement effects, i.e., for delaying and advancing the
recording trigger on successive firings so as to effect a
1. INTRODUCTION smoothing or removal of such undulations. The filtering can
be applied to post-experiment data records, or preferably in a
Acoustic arrays are employed in underwater seismic real-time mode during acquisition of the reflection responses.
exploration to determine the structure of the layered media in The Kalman filter applications are well known for some
terms of significant properties that aid in the exploration task different physical situations as reported in Gelb [11] and
of identifying hydrocarbon formations. Conventional acoustic Meditch [12]. Heave Component Extraction details have been
techniques have gained increasing use over the past several presented in [13]. The heave motion is also of interest in buoy
years for marine layer identification@ and classification of wave data analysis [14].
sediments. Mendel [1] gives a, tutorial introduction to the Compensation for source vertical displacement using
subject area, while [2-41 are classical treatments dealing with Kalman filtering involves two steps. In the first a model of the
the theoretical foundations of the estimation process. An vertical displacement phenomenon is obtained on the basis of
approach to extracting the subsurface features using a system available data. The type of model as well as its order are
theoretic treatment can be found in [5]. Using the enhanced important considerations. In early applications [15] a linear
acoustic signals for extracting information about the media second order model of the heave phenomenon was found
properties has also been treated in [6-7]. satisfactory. In this step the estimation of the model
A deep towed array of acousti signal sources and parameters is an important aspect. The resulting estimation
hydrophone receivers are used in conducting software uses an iterative procedure [16]. Once a model and
shallow marine
seismic explorations. Each source imparts energy to the water its parameters have been identified, the process of Kalman
and underlying media. The source signal then undergoes filtering is performed. Two versions exist. In the early
multiple transmissions and reflections at the layers' boundaries. implementation conventional Kalman filtering (14-15] is used
The towed body experiences six different kinds of motion, satisfactorily in a majority of cases. Kalman filtering theory
three of which are linear and the other three are rotational and parallel Kalman filters are applied to the exploration
about the three principal axes. None of the motions occur problem using a second order model in a recent paper [17].
singly, as they are all coupled in a manner that depends on Particular emphasis is given to the multi-receiver case as an
current and wave directions. These dynamics cause vertical important application of array processing methodology. Parallel
motions of the source and sensor which are controlled by both Kalman filtering is designed to take advantage of systolic
heave and pitch effects. array implementation. investigations concerning higher order
These vertical components have the effect of a varying model implementations are of current interest [18].
acoustic wave travel path to the sea floor and to the The paper discusses issues in preprocessing requirements as
sub-bottom reflectors between successive pings of the source. a result of incorporating the effects of the pitch component in
compensating for the vertical displacements.

CH2585-8/88/0000- 291 $1 @1988 IEEE

2. VERTICAL DISPLACEMENT MOTION MODELING (m -,- a .) 2 + b;@ + cz + A + e6 + hO F(t) (3)

The towed body experiences six different kinds of motion, Here we have:
three of which are translational (linear), and the other three m= mass of towed body.
are rotational about the three principal axes. The direction of
travel is taken as the x-axis, and the motion in this direction a. =added mass
is called the surge . The translational motion along the y-axis b= hydrodynamic damping force associated with the
is termed the sway, and the vertical or z-axis translational heave velocity.
motion is known as the heave. The rotational motion about the
x-axis is referred to as the roll, while that about the y-axis is c=F heave restoring force associated with a unit heave
the pitch, and finally the term yaw refers to motion about the displacement z(t).
z-axis. d= moment of inertia coefficient associated with the
None of the motions occurs singly, as they are coupled in angular pitch acceleration's effect on the vertical
a way that depends on the wave and current directions. motion's force balance.
Decoupling the motions simplifies the analysis and is useful e= hydrodynamic damping coefficient associated with the
when motion in a plane is of interest. For the purposes of this angular pitch velocity's effect on the vertical motion's
study the relevant motion is in the vertical direction which is force balance.
controlled by heave and pitch. In this section we outline the
various models available for representing the vertical dis- h= hydrodynamic restoring coefficient associated with the
placement components. angular pitch's effect on the vertical motion's force
Model for Heave Qn!L balance.
There are four kinds of forces acting on a submerged body F(t)= exciting force in the heave direction.
in the heave direction.
I- Inertial force, given as the product of the virtual mass For the pitching dynamics, we have:
(body's mass plus the added mass term) and the vertical
acceleration. (I,Y+AYY)6+B6+C0+D2+E_@+Hz=M(t) (4)
2- Damping force, given as the product of the damping Here we define:
constant and the vertical velocity. This force always resists the
motion. I.,= Principal axis moment of inertia of towed body.
3- Restoring force, which tends to bring the body back to its AYY = added principal axis moment of inertia of towed
equilibrium position and is given as the product of restoring or body.
spring constant and the vertical displacement. B= hydrodynamic damping torque associated with the
4- Exciting force which acts on the body. This is an unknown pitch velocity.
random function of time. C= Pitch restoring torque associated with a unit pitch.
The heave-only model is a second order linear time D= Mass coefficient associated with the heave's effect on
invariant system. This type of model has been widely accepted the angular torque balance.
to reasonably represent the vertical displacement dynamics of
the towed body. E= hydrodynamic damping coefficient associated with the
heave velocity's effect on the angular pitch torque
Model for Pitching QILIL balance.
A submerged body will experience a pitching motion H= hydrodynamic restoring coefficient associated with the
under the effect of four categories of moments, which are heave effect on the angular pitch torque balance.
similar to forces acting in the heave-only case.
I- Inertial moment, given as the, product of the virtual mass M(t)= exciting moment in the pitch direction.
moment of inertia (body's mass moment of inertia plus the The various coefficients in the hydrodynamic model are
added mass moment of inertia term) and the angular pitching evaluated either experimentally or using strip theory results
acceleration. [9]. The next section treats the problem of developing a state
space model of the dynamic vertical displacement process of
2- Damping moment, given as the product of the damping the towed body.
moment coefficient and the angular pitching velocity. This
moment always resists the motion. 3. VERTICAL DISPLACEMENT STATE SPACE MODEL
3- Restorin$ moment, which tends to bring the body back to
its equilibrium position and is given as the product of Equations (3) and (4) can be manipulated to obtain a state
restoring or spring moment coefficient and the angular pitch. space model by defining the following state variables:
4- Exciting moment which acts on the body. This is an IX, = Z
unknown random function of time.
The pitch-only model is a second order linear time
invariant system. X2=0
Model for Counled Heave and EiJ& X-3
Starting with a heave-only perspective, one can stipulate
that pitching will add a component to the vertical position at X,=6
any point on the submersed body. The vertical position zP of
strip at point P which is located at a distance @ from the
body's center of gravity is given by: As a result we have:
z P = z - @ sin 0 1 -@ , = X3 (5)

For small pitch angle 0, we have: -)@2 m X4 (6)
Z, = z (2) Equations (3) and (4) are now written as:
Two general dynamic equations of motion in z and 0 can (m,aje,+bX,+cX, +d)@,"X4+hx, F(t) (7)
.be written. The equations are coupled due to the interaction
between the two variables. For the heave dynamics, we have:
(1YY + Ayy)X4 + BX4 + CX2 + D,@, + Ex, + Hx, = M(tX8)

292

Equations (7) and (8) couple the derivatives of the state 4. COMPENSATION USING KALMAN FILTERING
variables and further manipulations are required to obtain a
canonical state space form of the heave and pitch model. The vertical displacement motion model of Eq. (14), can
It is convenient to define the state vector X using the be written in discrete state space form as:
partitioned form: x(k + I ) = 0(k + I , k)x(k) + T(k + I , k)w(k) (16)
X = X 12 (9-1) The state transition matrix 0 (k+l,k) and the matrix r
IX341 (k+l,k) are constants and we therefore write:
Here we have: x(k + I Ox(k) + Fw(k) (17)
(9-2) The input sequence w(k) is assumed to be a Gaussian
X 12 X I white sequence with zero mean and a covariance matrix Q (k),
[X2 being positive semi-definite. The initial state is assumed to be
a Gaussian random vector with zero mean and known
covariance matrix P (0). It is further assumed that w(k) is
X34 [Xj (9-3) independent of x(O). The record of vertical displacement
X, component is assumed to be the basis for the measurement
model given by:
zp(k + I Hx(k + I ) + v(k + 1 (18)
We now define the following matrices:
The measurement error sequence v is assumed to be
M34 rn+ a. d 10 _ I Gaussian with zero mean and a covariance matrix R(k).
D IYY + AYY Assume that measurements z(l), z(2), ..., z(j) are available,
from which we like to estimate x(k), denoted by x (kjj). In
filtering we have j=k, and we therefore wish to ind x (k1k).
N c h] (10-2) We utilize the standard predictor-corrector form of a Kalman
H C filter given by:
N14 b e (10-3) PREDICTOR
E B In the predictor stage we obtain a prediction of the state
As a result, we write equations (7) and (8) in the vector form: based on the previous optimal estimate:
M 34'@ 34 + N 12X 12 + N34X*34 @ Fm I I - 1 X'(_ ) = Ok- I X k- I (+) (19)
Here we have In addition, we obtain for the error covariance matrix
[ F ] )OT_,
Fm= M (11-2) PJ1 0 k- I P J+ k (20)
Equation (11-1), can be rewritten as:
'@34 1 X 12 + '4 2 X34 ' M -IF (12-1) CORRECTOR
- 34
Equations (5) and (6) are now written as: In the corrector stage we obtain an updated state estimate
i
'@ 12 = X34 (12-2) Xk(+) = Xk(-) + Kk(yk - HkXk(-)) (21)

Equations (12-1) and (12-2) express the derivatives of the
states explicitly in terms of the state. In addition,we obtain an update of the covariance matrix
In equation (12-1) we have defined: as:
A, = -M341 N 12 (13-1) Pk(+) = (I - KkH@)PJ_) (22)
A2 = _M3" N34 (13-2) Here K is the Kalman gain matrix given by:
We now combine equations (12-1) and (12-2) into the T @:p@'(_)HT
single state space equation model given by: Kk = P,(-)Hk [H k + Rk] (23)
Ax + Gw
(14)
We note here that adopting a higher order model to represent
The matrix A is given by: the vertical displacement process involves an increase in the
order of computation required to perform the Kalman filtering
A=[ 0, . Ij (15) process.
A A The optimal estimate of the vertical displacement is given
The matrix 0 is 2x2 with zero entries, and I is a unity matrix. according to equation (2), by:
The effects of the excitation forces and moments are lumped ZP=Z-@o (24)
into the matrix product G w. Equation (14) is the desired state
space model. In terms of the state variables we have:
ZP X I - @X2 (25)

293

We note here that we have to deal with a fourth order model [7] Ferial, El-Hawary, and W.J. Vetter, "Event Enhancement
as contrasted with a second order model if one considers only on Reflections from Subsurface Layered Media", IEEE
the heave dynamics. Knowledge of the value of @ representing Journal of Oceanic Engineering, Vol. OE-7, No. "Pp.
the distance of the source-receiver point from the towed 51-58, IT82 .
body's center of gravity is also important. 18] R. Bhattacharya, Dynamics of Marine Vehicles, New York
It is clear that implementing this compensation approach Wiley- Interscience, 1978
requires knowledge of the model parameters. In a practical
situation one can use a transfer function approach combined [9] W.G.,Price, and R.E.E.Bishops, Probabilistic Theory 2f Shit)
with a least squares parameter estimator to arrive at the Dynamics , New York: Wiley-Interscience, 1974
required values. [10] M.E. McCormick, Ocean Engineering Wave Mechanics,
5. CONCLUSIONS New York: Wiley-Interscience, 1973
[11] @. Gelb, Applied optimal estimation, The Analytic
Sciences Corp., 1974.
In this paper we reviewed the basic problem of [121 J.S. Meditch, Stochastic Optimal Linear Estimation and
compensating for the vertical displacement effects. Here the
inclusion of the pitch dynamics results in two coupled second Control, New York: McGraw-Hill, 1969
order equations in the heave and pitch components. The [13] F. El-Hawary, and W.J. Vetter, "Heave compensation of
conventional hydrodynamic equations were reviewed and a shallow marine seismic reflection records by Kalman
state space model for representing the vertical displacement filtering," presented at IEEE Oceans'81, Boston, MA,
dynamics was derived. The model is of the fourth order and is September 1981
of a form suitable for Kalman filtering, implementation. The
application of a standard Kalman filter to estimate the vertical [14] R.W. Severance, " Optimum Filtering and Smoothing of
displacement component on the basis of this higher order Buoy Wave Data% Journal gf Hydronautics, Vol. 9, pp.
model was also discussed in the paper. 69-74, April 1975.
[15] Ferial, El-Hawary, " Compensation for Source Heave by
6.REFERENCES use of Kalman Filter", IEEE Journal of Oceanic Engineer-
ing, Vol. OE-7, No. 2, pp. 89-96, 1982
[1] J.M. Mendel, "Some Modeling Problems in Reflection [16] Ferial, El-Hawary, " An Approach to Extract the
Seismology", IEEE ASSP, Vol.3, No.2, pp.4-17, April 1986 Parameters of Source Heave Dynamics", Canadian Electri-
[2] E.A. Robinson, Multichannel Time atrjU Analysis with cal Engineering Journal, pp. 19-23, January 1987.
Digital Comi)uter Programs, Holden Day, San Francisco, [17] Ferial, El-Hawary, K.M. and Ravindranath, " Application
1967. of Array Processing for Parallel Linear Recursive Kalman
Filtering in Underwater Acoustic Exploration", Proceedings
[3] J.M. Mendel, Or)timal Seismic Deconvolution: An Estima- gf IEEE Oceans '86, Washington, D.C., Vol. 1, pp.
tion-Based Armroach, Academic Press, New Y 1983. -
336-340, September 1986.
[4] E. A. Robinson, and T.S. Durrani, Geor)hvsical Signal [18] Ferial, El-Hawary, and T. Richards, " Heave Response
Processing,Prentice- Hall, Englewood Cliffs, N.J.,1986 Modeling using Higher Order Models', presented at the
[5] Ferial, El-Hawary, " An Approach to Seismic Information International Symposium on' Simulation and Modeling,
Extraction ", in Time atEipa Analysis 1 Theory and Practice Santa Barbara, CA, May 1987.
!k, O.D. Anderson, J.K. Ord and EA. Robinson (editors), [19) Ferial, EI-Hawary, " Image Analysis Methods From Sea-
Elsevier Science Publishers, Amsterdam, pp.223-238, 1985. bed Reflections and Multiple Reflections International
[6] Ferial, El-Hawary, and W.J. Vetter, "Spatial Parameter Journal gf Pattern Recognition and Artificial Intelligence,
Estimation for Ocean Subsurface Layered Media", Cana- Vol. 1, No. 2, July 1987.
dian Electrical Enpineering Journal, Vol.5, N0.1,pp.
28-31,1980

204

PREDICTION SYSTER FUR ACOUSTIC 1011URZ FRCK OCEAN BATHYNETRY

L.C. Haines, W.W. Renner, and A.I. Eller

Science Applications International Corporation
1710 Goodridge Drive
Mclean, Virginia 22102

ABS'T13ACT The primary engineering challenge to this
developuent effort was to balance the trade-
A real-time capability to predict ocean bottom offs between accuracy and execution time. In
acoustic returns (reflections and backscatter) the numerical progranming careful attention was
has been developed for at-sea support of paid to the sequence in which acoustic param-
acoustical ocean surveys. The systean operates eters were calculated,. the manner in which
on a HP-9.02o desk top ocaputer and provides results were stored for later access, and the
color screen displays of predicted bottom step sizes used for range and azimuth varia-
reverberation as well as echoes from seamounts. tion.
Predictions are based on gridded archival data
bases for bathymetry and sound speed profiles, As part of the ccffnprcmise between resolution
both of which may vary with location. A and conpitaticn time, a special dual resolution
special dual resolution approach was devised to approach was devised to provide a high-
provide a high-resolution depiction of the resolution depiction of the seamount returns in
seamount returns in conjunction with a lower- conjunction with a ladez-resolution presenta-
resolution presentation of the reverberation tion of the reverberation from the slowly
from the slowly varying bottom features. This varying bottom features. Separate calculations
approach allows large ocean areas to be covered are made for each of these contributions.
in reasonable cmqxitaticn time.
APPROACH

The prediction system consists of four parts:
Data Bases, two sub-models ASEW and PMERB
INTRODUCTION that address the pkWsics involved, and the
graphics output. There also are an auxiliary
A real-time ccaputer based prediction System bathymetric processing sub-model that prepares
has been developed for caV&ing acoustic a semmft data base and a user-friendly driver
returns from the ocean bottom. An intended menu routine. These are represented in the
application of such a system would be to overview flow diagram in Fig. 1. The bathy@
provide b-t-situ support to acoustical surveys metric data base presently used in the system
of ocean bathymetry, seamount locations, and is SYNBAPs, which provides bottom depth on a
bottom materials as indicated by scattering 1/6-degree grid.
properties. The system operates on a HP-9020
desk top calculator and provides color screen A. Transmission loss
displays of predicted bottom reverberation as
well as echoes from seamounts. It operates in The ptWsics in the system is addressed by two
real time in the sense that calculations for sub-models: ASERr and REVERB. The ASTRAL
large ocean areas, covering 300 nmi in any 5@rstem for the Estimation of Radial Trans-
direction, can be ocupleted in less than 30 mission Loss (AsEPT) is, by itself, a system
min; thus, revised predictions can be made for predicting acoustic transmission loss in
during the course of a survey as ship locations all directions from a fixed point location.
change. The currently operating version of ASERT automatically extracts the needed bathy@
this prediction system can acommiodate single metry, bottom loss, and sound speed profiles
ship surveys, in which the acoustic source and along a 3600 fan of radial tracks about a
receiving array are located essentially at the fixed location. The fixed angular step size
same place, as well as multi-ship or similar between radials may be assigned a value from 50
configurations where source and array may be to goo, depending on desired resolution and
separated by a substantial distance in a so- accuracy-
called bistatic arrangement. The system
predictions are based on gridded archival data Ttansmission loss calculations are made to
bases for bathymetry and sound speed profiles, support both the contribution from the low-
-both of which my vary with location. resolution bathymetry as well as from the

CH2585-8/88/0000-,295 $1 @1988 IEEE

seancurts. For the bathymetric contribution, missed. To overcome this problem the seamounts
transmission loss and propagation time to a are extracted from the bathymetric data base
moving "target" point along the bottom are and handled separately.
computed along each radial track by means of
the ASTRAL model. The new ASTRAL version rib address the seamounts a portion of the
containing CZ structure has been included here. bathymetry data base is extracted to cover the
Surface contributions are cmVuted also, in a entire area of interest. It is then filtered
similar way, so that they may be couPared to with a low pass filter, and a pattern recog-
the bottom contributions. Transmission loss to nition algorithm is used to extract the sea-
surface and bottom are modified to incorporate mounts. Each seamount is modeled individually
the enexgy arrival angles as combined with the using a set of parameters that are related to
appropriate angle dependence of the boundary its target strength. The large scale rever-
scattering kernels. - In order to estimate the beration is =nputed from the low pass filtered
acoustic returns from sem=unts, transmission version of the bathymetry using equally spaced
losses are canputed, simultaneously with the radial sampling about the source and receiver.
above calculations, from the same fixed loca- The radial estimates of the smoothly varying
tion to three additional "target" depths along long range effects are then combined with the
each radial. it is assumed here that an discrete returns from seamounts. This process
effective depth for each seffimunt can be is illustrated conceptually in Fig. 2.
approximated by one of the three values RESUIMS
selected.
For bistatic arrangements AsERr is called twice Representative output plots of the bottom
with the acoustic source and then the receiving acoustic returns from a bistatic survey
array occupying the appropriate fixed point arrangement are illustrated in Figures 3 and 4.
location. The Output graphics are designed for display on
the color screen of the HP--9020. By a separate
B. Reverberation routine they also may be redrawn in color on
paper or transparencies by an offline pen
in REvERB the modified surface and bottom platter. Figure 3 is a black and white Xerox
transmission losses are used separately to copy of a color plat from the pen plotter.
compute surface and bottom reverberation Figure 4, representing a different plot tech-
densities. The bottom contribution represents nique, is a shaded black and white drawing.
only the portion attributed to the slowly
varying bathymetry. The term reverberation In Figure 3 the receiver is located at 50ON and
density is used to indicate a normalized 1610W, which point corresponds to the center
reverberation resulting from a ore-second of the drawing. Me source is at 500 N, 1600
pulse, as received by an ideal beam, Ora-degree W, directly east of the receiver. The effects
wide in azimuth. The common Practice is to of several small seamounts can be Observed.
describe bottom scattering by Lmnbertls law The acoustic returns in Fig. 4, also for a
with a bottom scattering coefficient of -27 Pacific location, correspond to a source at 390
dB/sq yd and surface scattering by the Chapman-- N, 1280 W and a receiver 10 due west. The
Harris results. The seamoant contributions are receiver is located at the origin of the plot.
handjLed as distinct echoes from each of the The drawing here. shows returns coming from the
seamounts. Mendocino escarpment, running east and west,
north of the receiver. (This work was jointly
The ambient noise field may be ccuputed by the sponsored by the U.S. Naval Ocean Research and
Ambient Noise DirectlGrAlity Estimation System Develcpnent Activity and by the U.S. Office of
(ANDES), which also is a part of the Prediction Naval Research.)
system, using its customary shipping density
and errvirormiental acoustic data base. Ambient
noise and surface returns are carputed so that
one can identify tames when the bottom returns
dominate the total received field.

C. Rationale for Seamount Approach

Programing decisions were driven by the
objective to achieve reasonable accuracy within
a run time consistent with at-sea applications-
If there were no limitations an run time and
core storage space, then calculations could be
made along as many equally-spaced radials, from
source and receiver, as needed to include
seamounts. For practical reasons, however, the
least angular spacing allowed, at least at
present, is 50, and for this spacing most
seamounts would fall between radials and be

296

ES BATHYMETRY PROCESSOR BOTTOM RE-VERBERATION DENSITY

LOW PASS FILTERED
ASERT BATHYMETRY

E;;..'T loll 11

10 -

F-RAPHIC OUTPUTS
oil

FIG. 1. FIA)W DIAGRAM OF
PREDICTION SYSTEM 0
co

0 LS.4*W
16T2*W 1ST01W 158-W
200 -160 -120 -80 -40 0 40 @O 120 160 200
RANGE (NM)

FIG. 3. BOTTOM ACOUSTIC RETURNS FOR
BISTATIC SURVEY GEOMETRY

RnTTnm RFVFRRFRAT-TnN nFINSITY

0
.... .......

FILTER
@uo
ki. ... .... iii

FIG. 2. REPRESENTATION OF HATHYMETRIC
.... . . .....
PROCESSING CONCEPT

7771
200 1'60 1'20 40 -40 6 @0 I;i, I ISO 200
RHNGE fNM)

FIG. 4. BOTTOM ACOUSTIC RETURNS FOR
BISTATIC SURVEY GEOMETRY
4.A:TAF:B-A:-SES @"BA'WI'H
Y'
E

P
A S FILTERED
' y
[email protected]

297

Programmable Subsurface
Acoustic Recording System

Glenn P. Villemarette

U.S. Naval Oceanographic Office
Stennis Space Center, MS 39522-5001

ABSTRACT signal inputs from the hydrophones and power from
the external battery package through watertight
This paper describes the upgrading of the Naval connectors.
Oceanographic Office's electronics package for the
Moored Acoustic Vertical Array (MAVA) buoy system. The IRIG-B TCG produces a time reference. it
The MAYA can simultaneously record the outputs of provides a real-time clock output for use by the
as many as 12 hydrophones for 32 hours. A low- control system. It also provides an IRIG-B output
power mi crop rocess or-based design was chosen for which is recorded on a dedicated track of the
versatility. The effective recording time is put tape.
to maximum use with the ability to program start
and stop times on an event-by-event basis. A separate battery provides isolated power to the
Amplifier gains can be automatically selected hydrophones. This ensures a low noise source for
based on maximum probable peak amplitudes using this high-gain system.
the automatic gain select function. The effective
dynamic range of ambient-noise measurements is A 14-track analog tape recorder is used for
greatly increased using the prewhitening filter. storage of the acoustic data. It contains 9200
Long deployment times are possible due to greatly feet of tape and operates at 15/16 inch per
reduced power consumption between events. second. This provides about 32 hours of recording
Automatic testing, status reporting, and a user time. Two different record amplifiers are
friendly interface provide user confidence in the available. The intermediate band direct-record
automated system operation. amplifier provides response from about 100 hertz
to 3900 hertz. More reliable performance is
available using the Frequency Modulation (FM) Wide
INTRODUCTION Band group 1 (WB1) record amplifier. The FM WB1
record amplifiers will accurately record data from
The Moored Acoustic Vertical Array (MAVA) upgrade direct current to 625 hertz.
is a subsurface, bottom-moored, automatic acoustic
data acquisition and recording system consisting The microprocessor-based control system performs a
of a subsurface instrumentation package and variety of system functions including hydrophone
buoyancy assembly supporting a multihydrophone interface, signal conditioning and amplification,
array and the necessary bottom mooring cable. A and automatic system operation. It utilizes user-
diagram of the MAVA is shown in figure 1. A entered parameters for start time, stop time,
magnetic tape recorder in the instrumentation amplifier gains, and synchronizing tone frequency.
package records acoustic signals simultaneously In addition, diagnostic features are included to
from each of up to 12 installed hydrophones. The aid in calibration and troubleshooting. it
internal electronics is microprocessor controlled consists of 14 type 'STD-BUS' cards enclosed in a
and may be programmed for a variety of prescribed card cage. All cards use Complementary Metal
recording intervals. During all recording cycles Oxide Semiconductor (CMOS) components for low
an Inter-Range Instrumentation Group-B (IRIG-B ) power consumption. It includes a microprocessor
time-code signal and a synchronizing tone are card, a memory card, a parallel input card, a
recorded on two tracks of the recorder for future serial input/output cardi six amplifier cards, two
synoptic analysis of acoustic data. The major hydrophone interface cards, a number-controlled
subsystems of which the MAYA is composed are the oscillator (NCO) card, and a power-regulator card.
Instrumented Pressure Vessel (IPV), battery All cards are contained in one card cage.
package, flotation buoy, and a hydrophone array.
This paper limits itself to the upgraded IPV Automatic system operation is provided by the
electronics portion of the MAVA. microprocessor card, the memory card, the parallel
input card, the serial input/output card, and the
power-regulator card. The microprocessor card
IPV ELECTRONICS utilizes a National Semiconductor NSC-800
microprocessor. It executes a custom program
The IPV electronics are contained in a pres- which occupies over 8000 bytes of Read Only Memory
surized vessel. It consists of an IRIG-B (ROM). The memory card contains volatile, Random
Time-Code Generator (TCG), a hydrophone battery, a Access Memory (RAM), and nonvolatile, Electrically
14-track reel-to-reel tape recorder, and a Erasable Programmable Read Only Memory (EEPROM).
microprocessor-based control system. It accepts The parallel input card allows the microprocessor

298 United States Government work not protected by copyright

to access the actual (REAL) time by reading the in recording levels which are outside the dynamic
?utput of the time-code generator. The serial range of the system. Amplifier gain can also be
input/output card permits user interface with a selected automatically using AGS programming. AGS
terminal. The power-regulator card converts the will be discussed in the next section.
raw battery power to the required voltages for
STD-BUS operation. In addition, it controls The Number-Controlled Oscillator (NCO) card
switching operations and provides status signals provides a synchronizing tone for recovery of
to the microprocessor. Another feature of the data. This card provides a sine wave output to a
power-regulator card is its ability to reduce dedicated track of the recorder. The frequency of
power consumption drastically by switching to a operation can range from 0 hertz (off) to 32767
low-power, SLEEP, mode when not in use. hertz in steps of 1 hertz. The frequency is
controlled by the microprocessor. The sine wave
Interface to the hydrophones is provided by the is used with a phase-locked loop to synchronize
the hydrophone interface cards. They provide the analog to digital-conversion sample rate.
power to the hydrophones and modify their outputs This synchronous recovery is performed when the
to be compatible with the system. Two types of recorded data are processed.
hydrophone interface cards are available which
permit the use of a variety of hydrophones. OPERATION

The amplifier cards provide signal conditioning Operation of the MAVA is controlled by the
and amplification. Signals are patched from the integral microprocessor which accepts status
hydrophone interface cards to the amplifier cards inputs (Ref. table 1) and provides control
using small cables. A single hydrophone signal functions (Ref. table 2). The operation of the
may be patched to several amplifier cards for MAVA is controlled by its custom program. The
recording on multiple tracks of the recorder. MAVA program consists of two parts: the MONITOR
Outputs of the amplifier cards are applied to the program and the MISSION program. The MONITOR
tape recorder. program provides a user interface for entering
programmed parameters, testing, and determining
Signal conditioning consists of one status. The MISSION program performs the data
microprocessor-controlled and three manually collection process. The NAVA normally operates in
controlled functions. Manual activation of the MISSION program mode and only enters the
switches provides control of a 20-dB attenuator, a MONITOR program as a result of manual interrupt by
20-dB preamplifier, and a prewhitening filter for the user.
each channel. The position of the switches is
monitored by the microprocessor and displayed to Table 1. Status Inputs to Microprocessor
the user with the STATUS command. The
microprocessor can control the gain of a Type Source
programmable gain amplifier (PGA) for each -----------------------------------------------
channel. The range of gain selection is 0 dB to Actual (REAL) Time Time-Code Generator
6U dB in 6-dB steps. Fault Indicator Tape recorder
Recorder Indicator Tape recorder
Manual control of a 20-dB attenuator and a 20-dB Relay Driver Set Power-Regulator Card
preamplifier is available to extend the range of Relay Driver Reset Power-Regulator Card
operation for various signal levels and hydrophone Recorder Power Main Battery
sensitivi 'ties. A prewhitening filter may also be Hydrophone Power Hydrophone Battery
manually controlled. The prewhitening filter is Analog Power, +12V Power-Regulator Card
used to attenuate the large signals present at low Analog Power, -12V Power-Regulator Card
frequencies to a level similar to middle and Peak Detector Amplifier Cards
higher audio frequencies. A simple design was Attenuator Ainplifier Cards
chosen and precision components were used to allow Preamplifier Amplifier Cards
precise estimation of the transfer function. It Prewhitening Filter Amplifier Cards
consists of one resistor and one capacitor. It is
a high-pass filter which provides a 6-dB/octave Table 2. Control Outputs of Microprocessor
transfer function with its corner frequency at Type Controlled Device
3959 hertz. Its transfer function is shown below.
Vout = Vin * 1/(1 + j(3959/F)) SLEEP Control Power-Regulator Card
Recorder Power Recorder Power Relay
where: Hydrophone Power Hydrophone Power Relay
NCO Frequency NCO Card
j = square root of -1, F = frequency in hertz. Amplifier Gain Amplifier Cards
MONITOR PROGRAM
The PGA varies the gains on an event-to-event
basis. For each event, the PGA gain is selected Interface to the user is provided by the MONITOR
as a result of fixed gain (FG) or automatic gain program. The user accesses the MONITOR program by
selection (AGS) programming. For fixed-gain manual activation of an interrupt signal. A
programming, the operator enters the exact gain terminal provides the human interface. Commands
desired for each amplifier when programming the for programming, testing, and status checking are
scheduled events. Because of the uncertainty of available. The commands are: CLEAR, DEPLOY, EDIT,
signal levels, fixed-gain programming may result HELP, LOAD, SAVE, STATUS, and TESTx.

299

The HELP command enables a text message which During any point of the MONITOR program execution,
describes the available commands and tests. failure to activate a keystroke for a time period
in excess of 30 minutes results in an automatic
The CLEAR command is used to initialize the branch to the MISSION program. This will result
working copy of programmed EVEMTs which are stored in a power-off of the MAVA electronics unless an
in volatile RAM. The CLEAR command turns all EVENT is to be performed. This feature is
events off. provided to prevent an undesired mission failure
if the system is accidently left operating the
The EDIT command provides an interactive means of MONITOR program.
programming parameters of the EVENTs in volatile
memory (RAM). Up to 64 entries of EVENTs are
allowable. The parameters include times to begin MISSION PROGRAM
and end each event, the frequency of the NCO card,
and the amplifier gains for all channels. The The MISSION program performs the primary function
START and STOP times are entered as a Day, Hour, of the MAVA, data acquisiton. It acquires data in
and Minute. These times will be compared with a predetermined method as defined in the user-
actual (REAL) times read from the TCG. The programmed EVENTs. The MISSION program is
frequency of operation for the NCO card is also executed as result of a normal power-on reset.
entered by the user. The range of operation for This power-on reset occurs as a result of the MAVA
the NCO card is 0 hertz to 32767 hertz in 1-hertz SLEEP circuit timing out and applying power to the
steps. All twelve amplifier gains are entered STD-BUS. The MISSION program performs the pre-
independently. Fixed gains may be selected in the programmed functions and returns status in
range 0 dB to 60 dB in 6-dB steps. Automatic Gain nonvolatile, EEPROM. Normal operation of the MAVA
Select (AGS) for the amplifiers may be also results in the following sequence, which begins
entered. upon timeout of the SLEEP circuit on the power-
regulator card. This sequence is repeated over
The SAVE command stores the programmed EVENTs of and over again.
the volatile memory (RAM) into nonvolatile memory
(EEPROM). The configuration of the signal Power-Up Mission Sleep
conditioning is also saved. Finally a checksum, Reset Program --> Circuit
modulo 256 addition of programmed EVENTs, is Execution Activate
performed and saved.
The MISSION program begins by reading the actual
The LOAD command provides a means of reediting (REAL) time from the Time Code Generator (TCG).
previously SAVEd EVENTs. The LOAD command copies It then compares the REAL time with the START time
the EVENTs from nonvolatile memory (EEPROM) into for each of the PENDING events in sequential
volatile memory (RAM) where they can be EDITed. A order. A particular event will be performed if
checksum is performed and compared to a previously the START time is in the range indicated below.
stored checksum to verify the proper storage in This prevents a START time from being missed due
the EEPROM. An error message is displayed if the to back-to-back event programming.
checksums don't match.
The DEPLOY command is used to discontinue START <= REAL < ( START + 1 hour
operation of the MONITOR and vector to the MISSION
program. The DEPLOY command is normally performed If no events match up, the MISSION program will
after the EVENT program has been generated (using conserve power by de-engerizing the MAVA
EDIT) and stored in nonvolatile memory (using electronics. This is performed using the sleep-
SAVE). activate function of the power-regulator card.
The MAVA electronics will wake up after a
The STATUS command initiates a display of the predetermined sleep interval has expired. This
system status. The display includes the MAVA will result in re-engerizing the MAVA. The MAVA
identification message, actual (REAL) time as read will again resume operation of the MISSION program
from the time-code generator, actual amplifier as a result of a normal power-on reset.
switch configurations, stored EVENT program, and
any accompanying error messages. A typical When a PENDING event is performed, its status in
display of the STATUS might appear as shown in nonvolatile memory is immediately changed to
figure 3. COMPLETED. The actual START time is also saved in
EEPROM. The recorder power and hydrophone power
The TESTx command performs any of 13 are turned on by activating the SET relay driver
troubleshooting procedures as determined by the circuitry. Successful switching is checked for
suffix X. Tests include reading status, test using the status register on the power regulator
pattern generation, and manual control. The card. The status is then recorded in the
status of each of the amplifier cards, the nonvolatile memory. Automatic Gain Select (AGS)
parallel I/F (TCG I/F), and power-regulator, card is performed if it is required and the actual
can be displayed to the user. Each amplifier card amplifier gains are stored in nonvolatile memory.
can also be injected with a test pattern from the The amplifiers are then set to the determined
microprocessor, which allows digital trouble- gains. The NCO card is set to its programmed
shooting. The relays can be manually controlled frequency. The MAVA then waits for the stop time
to allow testing of the associated driver as determined by:
circuitry. The power-reduction circuitry, SLEEP,
can also be manually activated. STOP <= REAL < ( STOP + 1 hour

300

This 1-hour window was used in case of temporary entire event execution. The selected gains will@
time-code generator failure. Next the actual stop also be stored in nonvolatile EEPROM for user
time is saved in nonvolatile memory. Then the verification at recovery.
MISSION program turns off the recorder power and
hydrophone power, checks and saves status, and POWER CONSERVATION
performs a new checksum for the nonvolatile
memory. It then branches back to the beginning of Most of the time in a typical deployment of the
the MISSION program to recheck for any pending MAYA is spent waiting for the REAL time to match a
events at the new time. If the REAL time doesn't programmed event start time. Care was taken to
meet the START TIME requirements, the MAYA will go reduce power consumption when the MAYA is in
to sleep to conserve energy. standby. Power to the recorder and hydrophones is
switched out when not operating using latching
AUTOMATIC GAIN SELECT relays. Power to the electronics is greatly
reduced by de-energizing most circuits. This
Automatic Gain Selection (AGS) provides a means of removal of power is called sleeping.
accurately selecting the amplifier gain for a
stationary process with known distribution. AGS The MAYA MISSION program periodically reduces
results in a gain selection based on two standby power by sleeping. When asleep all power
parameters, the peak amplitude of the actual is removed from the STD-BUS. A low-power timer
output of the particular amplifier channel and a circuit on the power-regulator card remains
relative fixed offset entered at the EVENT energized and reapplies power to the STD-BUS after
program. The MISSION program performs the AGS expiration of a preset Sleep Period. Two sleep
function in the first minute of execution for each periods are available; SHORT NAP = 45 seconds and
event. The corresponding programmable amplifier normal = 11 minutes. The sleep period is
gain is first set to 0 dB. The signal at the initially set to normal. The sleep period will be
output of the amplifier is tested if outside the SHORT NAP if the microprocessor determines that a
range of +/- 0.5 Volts. This test is performed PENDING event is less than one hour away. This
about 18000 times a second for 9000 samples. This dual time approach provides efficient standby
gain remains unchanged and exits from the AGS operation and accurate turn-on times.
module if the number of samples exceeding this
@ange is greater than 5. Otherwise, the gain is About 300 milliamperes (mA) from the main battery
incremented by 6 dB and the process is repeated are consumed when the STD-BUS is energized. When
until it exits or the gain equals 60 dB. The asleep, about 30 mA are required for the TCG and
selected AGS value is then scaled by a value sleep timer circuit. When in standby. the MAYA
entered by the user. The scale factor, Bias, can executes the MISSION program (about 0.5 seconds)
be -12 dB, -6 db, 0 dB9 +6 dB, or +12 dB. The at periodic intervals. For long deployments it
actual gain is limited to the range of the will operate in the normal sleep period most of
amplifier, 0 dB to 60 dB. The expected peak the time. Current consumed will be 30 mA + (0.5
voltages of signals to the recorder are shown seconds * 300 mA /11 minutes) or about 30.2 mA.
below. Sleeping reduces the standby current to about 10%
of the amount required for STD-BUS operation.
Bias Expected Peak Voltages This permits long deployment times for the MAYA.
to Recorder.
Probability [8995/90001 BUILT-IN TESTING AND DIAGNOSTICS
AGS-12 dB +/- 0.125V to +/- 0.25V The MAVA performs a number of different types of
AGS-06 dB +/- 0.25V to +/- 0.5V automatic and manually activated tests. Automatic
AGS+00 dB +/- 0.5V to +/- 1.OV tests are performed during operation of MISSION
AGS+06 dB +/- I.OV to +/- 2.OV and MONITOR programs. Manually activated tests
AGS+12 dB +/- 2.OV to +/- 4.OV are available with the TEST command described
earlier.
If a gaussian distribution is assumed [Ref.
URICKI, we can translate the table above for other The MAYA performs two automatic self-tests upon
probabilities. A probability of 8995/9000 entry into the MONITOR program. They consist of a
corresponds to a deviation of +/- 3.4 standard RAM test and a ROM test. The RAM test performs
deviation units from the mean. Expected peak write/read operations for several different
voltages for a deviation of +/- 1 standard patterns which are written to all Random Access
deviation are shown below. Memory (RAM). The ROM test performs a checksum of
the Read Only Memory (ROM) and compares the result
Bias Expected Peak Voltages with the known value. The operator is notified
to Recorder. with an appropriate message if either test has
Probability [0.683] failed.

AGS-12 dB +/- 0.036V to +/- 0.074V The NAVA checks the operator inputs during
AGS-06 dB +/- 0.074V to +/- 0.147V operation of the MONITOR program. An erroneous
AGS+00 dB +/- 0.147V to +/- 0.294V keyboard entry from the operator results in an
AGS+06 dB +/- 0.294V to +/- 0.588V audible indication (bell) at the data terminal.
AGS+12 dB +/- 0.588V to +/- 1.176V This can arise from the operator entering an
incorrect command or incorrect keystroke sequence
The selected gains will remain constant for the while operating in the MONITOR program.

301

There i.s a variety of other automatic tests which result of improper feedback during event turn-on/
are performed during the operation of the MISSION off or event execution.
program. They can be described in three
categories: event turn-on/off and execution, The configuration of attenuators, preamplifiers,
configuration, and EEPROM checksum. They result and prewhitening filters used for signal
from operator error or component failure. Failure. conditioning is stored in EEPROM during the SAVE
of any test results in notification to the command. The STATUS command compares the act7ual
operator. configuration to the SAVEd configuration and
displays a warning message if different. This is
Automatic tests are performed during the turn-on, used as a reminder to the operator. The operator
turn-off, and execution periods for each EVENT. should check if the configuration is different
Faults detected during actual event turn-on and from that.planned.
turn-off sequence are recorded, saved in EEPROM,
and indicated under control of the STATUS command. For all write operations to the EEPROM, a checksum
Faults detected during actual event execution are is generated and stored along with the EVENT
separately recorded, saved in EEPROM, and, program data. Any operation that reads EEPROM
indicated under control of the STATUS command. If such as actual event execution, LOAD command, or
any faults are detected they are displayed along STATUS command compares actual checksum to saved
with the remaining status for each event. Such checksum. An incorrect checksum results in the
faults are recorded for each event separately. setting of an EEPROM failure flag. This flag is
Items tested include analog power, hydrophone also set if a write to EEPROM can not be verified.
power, relay drivers, and recorder status-fault The status of this flag is tested during the LOAD
indicators. Items are tested at the appropriate command and the STATUS command. If it is set, an
instances before, during, or after the actual error message is displayed to the user.
switching has occurred.

Analog power failure is set as a result of analog CONCLUSION
power not being detected. Hydrophone Power
failure is set as a result of inappropriate The upgraded MAVA electronics provides versatility
absence or presence of power applied to the to a proven data acquisition system. The
hydrophones. Relay driver failure is set as a microprocessor-based control system can adapt to a
result of incorrect feedback from transistor variety of requirements. The modular card design
drivers during event turn-on/off in either can be easily upgraded as requirements change.
energized or de-engerized states. Recorder fault These are features which are desired for all new
indicator and record indicator failure is set as a data acquisition systems.

BUOY WITH STROBE,
TRANSMITTER r-----?.5 nEr

teirr MID

VM j i

SYNTAM
HYDROPHONE CABLE ARRAY FUM
WERY

AIR WEIGHT POSITIVE ROTATION
5750 LOS. 1270 LOS.

TANDEM ACOUSTIC RELEASES

MOORING CABLE

ANCHOR

Figure 1. Diagram of the Moored Acoustic Vertical Array (MAVA).

TERMINAL
HYDROPNONES POWER CONTROL/STATUS (USER INTERFACE)

I/F YDROPNONEI/F SERIAL 1/0
CARD POINEN-REGUIATOR
1 2 3 4 6 1 7 1 1101112 CARD CARD
0000 0 000000 1
I I I I @t It]
STU - iu-s
I F_
20AW 2 CH A
2 IN AMP 2 CH AMP 2 CN AMP CUD
101 1 '112 CN AMP -CODE
a" #0 CARD #1 0 #2 CARD #3 CARD 5 NCO PARALLEL INPUT BCD TIME
cm 1 M2 CK 3 04 CNP a 07 C1 I CH I I CN 12 CARD CARD GENERATOR
00 0 0 00 00 00 0.1. 00 00 00 00 00 00

C1 12 6"
CN
010
CH
CK 8
CK7
CN I
CN 5COD

CN 3

CN I

Figure 2. MAVA Upgrade.

Moored Acoustic Vertical Array Event 01 Status: COMPLETE
Version 1.03, 26 October 1987 Start Time: 001:10:00
U.S. Naval Oceanographic Office Stop Time: 001:12:30
Stennis Space Center ELAPSED TIME= 000:02:30
Mississippi, USA Tone (Hz): 00000
Time=016:09:32 Programmed Gains
20-dB Attenuators CHOI CH02 CH03 CH04 CH05 CH06
CH01 CH02 CH03 CH04 CH05 CH06 OdB 6dB OdB 24dB 30dB 36dB
IN OUT IN IN IN OUT CH07 CH08 CH09 CH10 CH11 CH12
CH07 CH08 CH09 CH10 CH11 CH12 OdB 6dB OdB 12dB OdB 24dB
IN IN OUT IN IN IN Event 02 Status: PENDING
20-dB Preamplifier Start Time: 021:10:00
Stop Time: 021:12:30
CH01 CH02 CH03 CH04 CH05 CH06 ELAPSED TIME= 000:02:30
IN IN OUT OUT OUT OUT Tone (Hz): 00000
CH07 CH08 CH09 CH10 CH11 CH12 Programed Gains
IN IN OUT IN IN OUT CHOI CH02 CH03 CH04 CH05 CH06
Prewhitening Filter OdB 6dB OdB A+00 A+06 A+12
CH01 CH02 CH03 CH04 CH05 CH06 CH07 CH08 CH09 CH10 CH11 CH12
IN IN IN OUT IN IN OdB 6dB OdB 12dB OdB 24dB
CH07 CH08 CH09 CH10 CH11 CH12 Event 03 Status: OFF
IN OUT OUT OUT IN OUT
Event 04 Status: OFF
through . . .
Event 64 Status: OFF
Figure 3. Status Display.
REFERENCES
1. Urick, Robert i Principles of Underwater Proceedings 1983 Symposium on Buoy Technology,
Sound, 3rd ed., McGra;-1H1II, Inc., 1983. Marine Technology Society, Gulf Coast S 'ection.
2. Swenson, R., J. Selleck, and N. Dennis, "High 3. Beyer, William H., CRC Standard Mathematical
Performance Deep Sea Subsurface Buoy Mooring," Tables, CRC Press, 19A.
SIR

P
-57

303

The Relationship Between Acoustic Bottom Loss and the Geoacoustic
Properties of the Sediment

Diana F. McCammon

Applied Research Laboratory
P. 0. Box 30 State College, PA 16804

ABSTRACT The layer model chosen for this study is a
relatively simple three fluid layer model that
A study of the functional variations between uses 10 geoacoustic parameters and makes several
acoustic bottom loss from 50-1500 Hz and the assumptions[2]. It relies upon the "hidden
sediment geoacoustic parameters of density, sound depths" concept which states that the regions well
speed, thickness and attenuation has shown that below the turning point do not affect the ray.
simple relationships exist between certain of The sediments are assumed isotropic, and the
these parameters and specific angular regions of roughness of the sediment and basement interfaces
the loss curves. These relationships permit as well as multiple scattering within the layers
bottom losses to be estimated directly from the is neglected. The neglect of roughness of the
geoacoustic parameters without the need to sediment may be justified because this is a total
exercise complex layer modeling. energy model and the roughness primarily
redistributes the energy in angle about the
specular without decreasing the total energy.
Finally, shear wave propagation is ignored. While
I. INTRODUCTION the effects of shear wave conversion at the water-
sediment layer have been shown to be negligible in
Geoacoustic modeling of the interaction between Vidmar(31 the process is substantially more
acoustic energy and the ocean floor has gained important at the basement interface. the neglect
considerable favor among researchers because it is of this loss mechanism at low frequencies and in
presently the best method of incorporating the thin sediments may be the primary weakness of this
physics of sound propagation, refraction, model.
scattering, and boundary wave conversion into the
bottom loss process. This type of modeling The required parameters for this layer model are
consists of developing a layered profile of the sediment density, thickness, sound speed gradient
sediment with depth dependent attenuation, sound and curvature, attenuation and attenuation
velocity and density over-laying a basement gradient, thin layer density and thickness,
halfspace. Ray tracing techniques can then be basement reflectivity and water-sediment velocity
employed to compute the total energy returned to ratio.
the water column. This approach has been shown by
Knobles and Vidmarl] to be quite successful at For this model, the total energy bottom loss in
predicting the bottom returning waveform. the form of plane wave reflection at grazing angle
Certainly the emergence of this type of modeling Ow is computed by two path contributions; the
,has greatly improved the quality of bottom loss reflected path and the sediment transmitted path.
calculations for those areas of the world's oceans 2 2
whose geoacoustic profiles have been constructed. BL(Ow 10 log,,[Rw+ (Tl 3T31 RBe

One drawback to the efficient operational usage of Where Rw is the reflection coefficient from a
the geoacoustic layer model of bottom loss is that three-layer fluid sediment(41,
the geoacoustic parameters do not, of themselves,
convey a sense of the loss to be expected. T13 is the transmission coefficient,
Presently, given the area of interest and its
geoacoustic parameters, a model must be exercised RB is the basement reflection
on a computer before bottom loss can be known. coefficient,
Because of the complex interaction of acoustic
energy in this layered environment, an association a is the attenuation of the
between bottom loss and the geoacoustic profile is compressional wave, linear in frequency
not immediately obvious. However, the research and depth,
presented in this paper will demonstrate that
there are clear relationships between the r is the ray path length in the
geoacoustic parameters and certain angular regions sediment.
of the loss curves. These relationships permit
the assignment of a loss ranking system based on II MODEL PREDICTIONS-DOMINANT EFFECTS
the values of the geoacoustic parameters
themselves. To study Eqn. (1), the bottom types should first

CH2585-8/88/0000- 304 $1 @1988 IEEE

be divided into two groups: those with a velocity 1.0
ratio greater than unity and those less than :-- f = 1000 Hz
unity. A velocity ratio greater than unity CS/CW= 1.015
predicts a critical angle Oc below which most of Z 0.8- K = 0.0062 dBfm/KHz
the energy is reflected; cosOc - cw/cs. A Q . - 3
velocity ratio less than unity will cause the FE PS VARIATION FROM 1.5 TO 5.0 gmkm
z. 0.6-
energy to be driven more deeply into the sediment a -
making the transmitted path dominant. 5.0 REFLECTED PATH
0-4- 1.5
Hip-h Frequencies

0.2-
The study will begin with velocity 'ratios greater fr- 2.0
than unity, frequencies, about 1 kHz and thick 1.5
sediments. In Figure 1, the upper curve displays 0
11 0 30 40 50 60 70 80 90
the energy variations in the two paths vs angle 10 L2 ANSMITTED PATH
with the high velocity ratio, cs/cw - X as a
parameter. The lower curve displays the resulting 15-
bottom loss. At frequencies above 200 Hz the 1.5
reflected path carries most of the energy and the, ---------
C@
position of the critical angle controls the onset 10- 2.0
of loss. The major characteristics of reflection 2.5
influence the remainder of the curve and are seen 3.0
to be independent of the velocity ratio. In Fig. 5-
2, Variations in the sediment density ps are shown
with a fixed high velocity ratio. Here the 5.0
impedance characteristics of the sediment have a 0 -------
direct effect on the level of reflected energy 0 10 20 30 40 50 60 7a 80 90
above the critical angle of 10*. GRAZING ANGLE (degi
1.0 1.0 Fig. 2 Density variation, high X, 1000 Hz.

F5 0.8 1.1326 1.0 005
I = 10003
m1cm
PS = 2.4 g
0.6
REFUCTED PATH t 0.8 -.06 C IC = 1.015
S W
Cr
.0 Z
0
0.4
.4- Z 0.6 - K VARIATION FROM .06 TO .005 dBImIKHz

REFLECTED PATH
0.2
< 0.4-
FF TRANSMITTED PATH
J -a---- 1. 1326 - .005
0.2-
15 - f = 10M HZ 3 -TRANSMITTED PATH
PS = 2.4 gm/cm
.06
K= 0.0062 dB/MIKHz 0- - - - - - -- 6.
10 - C S/CWVARIATION FROM 1.0 TO 1.1326 10 -
ta-
5 1.0 5 0.06
0.005

1.1326

0 0
0 10 20 30 40 50 60 80 90 0 10 20 30 40 50 6D 7'0 80 90
GRAZING ANGLE (deg) GRAZING ANGLE (deg)
Fig. 1 Sensitivity to velocity ratio at 1000 Hz.- Fig. 3 Attenuation, high X, 1000 Hz.

The more dense sediments reflect more energy and the transmitted path is important from 0. to 20.*
consequently show lower losses as expected. The grazing, but at angles above this region,
thin layer remains acoustically transparent- at reflection dominates just as in the high ratio
thickness less than a quarter wavelength; at cases shown above, as shown, for example, in Fig.
'@reater thicknesses it replaces the sediment as 4, in comparison with Fig. 2.
t:@e primary reflector. Fig. 3 contains examples
of the predicted losses when the sediment Thus we can conclude that high angle bottom losses
attenuation is varied for a fixed density and in thick sediments are entirely governed by the
velocity ratio. impedance characteristics of the upper layers, and
the major characteristics are predictable by
In the low velocity ratio cases, (less than unity) density ranking alone.
,uu.7
I
P'2
"0 'm"M'
K

P
HZ
5
Si 'M' K"Z
VAR ATI ON FROM 5 05 -m3
I I- T .0 gmir

REFLECTED PATH
5-0
115
2.0
M
.5
t`12

.0
2.5
30

0
LRE-F-CTEDITPrATH '06 *05
10
KVAR IAT N FROM .06 TO .005 dBImIKH7

REFLECTED PATH

A ED PATH
.0@
TR P
ANSMI TTED A
06
0. @005
/1-1326

305

1.0 1.0 f = 1000 Hz3
I = 1000 Hz PS = 2.4 gm1cm
CS1CW = 0.990 C I C = 0.990
0.8 0.8 S W
K = .0062 dB/m/KHz Z
3 K VARIATION FROM .06 TO .005 dB/m1KHz
PS VARIATI ON FROM 1.5 TO 4.0 gm1cm
C9 0.6 Z U.0 P TRANSMITTED PATH

005
@5 0.4 TRANSMITTED PATH 0.4
.5 4.0 REFLECTED PATH
4.0 REFLECTED PATH
0.2 0.2
.01

0 0 Meg
20-
15 1.5 0.06

kA
15-0.04
10 2.0 0.031
2.5 K 1O.M02
0.021 0.0001
cc 10

0.6
4.0 Cd 5 0.005

a 10 20 310 40 10 60 70 M 90
GRAZING ANGLE (deg) 0 L
10 20 30 40 50 60 70 80 90
Fig. 4 Density variation, low X, 1000 Hz. GRAZING ANGLE (deg)
Fig. 5 Attenuation variation, low X, 1000 Hz.
At shallow grazing angles and high frequencies, Attenuation gradient variation is also
shown for two cases.
the most important loss factor in the transmitted
path is the dissipation loss e-,r. This damping
is dependent upon the sediment attenuation and the Low Freguencies
sediment sound speed profile. Variations of the
two paths vs angle and the resulting bottom loss Next, consider the low frequency (50 Hz)
are shown in Fig. 5. Here, the parameter changes sensitivity in thick sediments. Most of the
of the absorption coefficient x create a profound energy that penetrates the sediment will be
change in the low angle losses. returned because the attenuation is very small,
making the transmitted path contribution much
The variation of the absorption gradient is also larger. Examples are given in Figs. 6 and 7 for
demonstrated for 2 values and is shown to have a high and low velocity ratio cases with sediment
minor role. However, variations of the sediment density as a parameter.
sound gradient curvature and velocity ratio that
are not shown here also have a major effect on the 1.0 TRANSMITTED PATH
To account for all >- f = 50 Hz
low angle curve behavior. 1 5 C 1CW= 0.990
these sensitivities and obtain a suitable factor S
for ranking the low angle losses, dissipation loss 0.81 K = D062 dB/M/kHZ
1 3
exponent or is normalized and approximated by VARIATION FROM 1.5 TO 4.0 gm/cm
z 0.6 S
@2 4.0\\
KXcos (X)/gs, where x - cs/cw. (2)
0.4 REFLECTED PATH
Z
By' this equation, low angle losses are found to be 4.0
0.2 - - - - - - - - - - - -
directly proportional to the sediment attenuation,
inversely proportional to the sound speed .5
gradient, and functionally dependent on the 0
velocity ratio. Specific details of the profiles 10
such as the curvature and absorption gradient
produce only second order effects. 1.5
5 .0
Thus far, the preliminary conclusions are that,
for frequencies about 1 kHz and thick sediments, 4.0
high angle losses can be predicted based on a 0
knowledge of the sediment density while the a la 20 30 40 50 60 70 80 90
velocity ratio, velocity gradient and sediment GRAZING ANGLE Ideq)
absorption will determine the major
'S
'C

S
W

V
AR
5
P

T
274

10
9
g
0
H
M
'Cm

OM
LK VA R I=A N FR3.06 TO .005LdCl
MNSMIT70 PATH
J\
005

REF E
.0
2'- 0- 04V
0. 03
2 0 0'=
.5
K' 0.000,
021

0. 005

40\\-

15
4.0

characteristics below 20*. Fig. 6 Density variation, low X, 50 Hz.

306

1.0 loss. It is primarily employed in regions where
>. anomalously high returns were recorded that could
U
not be predicted using the measured geoacoustic
0.8 - parameters of the sediment. Densities of the thin
TRANSMITTED PATH
2 0 layer are higher than the sediment to account for
0.6 -
these anomalies. The principal effect of this
REFLECTED PA extra layer is to replace the sediment as a
TH 5.0 reflecting surface when the layer thickness is
It 0.4- 4.5
greater than 10 cm.
4.0

Q 0.2 - Table I summarizes all these trends for high and
- - - - - - - - - - - low frequencies in thick sediments.

0
15 Table I
f= 50 Hz Principle Sources of Energy Predicted
C /C = 1.015
S W by the Thin Layer Model
K = .0062 d8imUz
10 PS VARIATION FROM 1.5 TO 5 gm/cm3 T transmission R - reflection
13 Angular Frequency Region
5 2.0
Region f < 200 z f > 200 Hz
-5.0
Loss Loss
0 0 20* Rath mechanism path mechanism
0 10 20 30 40 50 60 70 80 90
GRAZING ANGLE Ideq) Low X T dissipation T dissipation
Fig. 7 Density variation, high X, 50 Hz.
High X R+T critical angle R critical
+ dissipation angle +
reflection
The gradual decrease in transmitted energy shown losses
in both figures from the peak to, 30* is' caused by
an increase in the dissipation ar as a result of Loss Loss
increasing r. The abrupt flattening above 33* 30-90* path mechanism i)ath mechanism
results because the transmitted ray is now
basement limited. no thin T+R basement reflec- R reflection
layer tion + reflec- from
Despite the increased importance of the tion from sediment sediment
transmitted path, the low frequency,' low angle
behavior is caused by the same mechanisms as at thin T+R basement reflec- R reflection
high frequencies so that the@ 1000 Hz ranking layer tion + reflection from thin
-system will still be valid (by velocity ratio or from thin layer layer
dissipation loss):.

The high angle behavior is now; however, -clearly Sediment Thickness
influenced by the transmitted path, primarily by
the loss combination T3lTl3RB- Since T31T13 - I- As the sediment thins, the basement will be
R2 13, the factors that affect the reflection encountered at shallower and shallower angles.
coefficient will inversely affect the transmission Below 200 Hz where the transmitted path is
coefficient, that is, low sediment densities important this will cause the angle where the
correspond to increased energy in the transmitted curve levels out, (here referred to as the knee of
path. Fortunately, this reversal is not the curve), to be shifted toward lower angles. In
sufficiently strong at 50 Hz, to reorder the effect, the sediment thickness determines the
ranking system. While the bottom losses for dividing point in angle between the two separate
changes in ps are much closer in value as compared low angle and high angle loss mechanisms. These
to the 1000 Hz case, they can still be ranked by shifts can be seen clearly in Fig. 8, where only
order of decreasing sediment density -as -in the the sediment thickness is varied.
1000 Hz case. Changes in RB as great as 3 dB
create at most a 1. 5 dB, variation in the high A secondary effect is the lowered level of high
angle section of the bottom loss. Thus, the high angle loss as the sediment thins. This is a
angle behavior can be ranked using the same system consequence of the shorter path through the
as at 1000 Hz _- namely by sediment density, even thinner sediments that engenders less attenuation
.though reflection is no longer the dominant loss of the transmitted energy. Note that the slope of
mechanism. the curve rising to the knee is independent of
sediment thickness since it is associated
Thin LAyer primarily with the transmission coefficient
properties. Only the very thinnest sediments (<80
The thin layer is an artificially introduced upper m) show these effects above 200 Hz because, at
sediment layer that serves to decrease the bottom higher frequencies, the transmitted portion of the

307

15 f = 50 HZ angle) of the average curve using the loss ranking
CS) CW = 0.990 scheme.
K =.D4 dSImikHz
ps = 3.0 Ist PARAMETER 3rd PARAMETER
10 LOW ANGLE LOSS HIGH ANGLE LOSS
SEDIMENT THICKNESS VARIATION FROM 0.5 TO 1.0 15 - 4
SEC. TWO WAY TRAVEL TIME.
5 10- 3
0.5 BOTTOM LOSS 6 2
@ 5,
1.0 AT I kHz 1
00 10 20 30 40 50 60 70 so 90
GRAZING ANGLE (deg)
Fig. 8 Various sediment thicknesses at 50 Hz. 10 2nd PARAMETER
SEDIMENT THICKNESS-LOW FREQUENCIES
BOTTOM LOSS
energy is much smaller and penetration to the AT 50 Hz 5
basement is more severely attenuated.
0
III LOSS CIASSIFICATION 0 10 20 30 40 50 60 70 80 90
GRAZING ANGLE Ideg)
Based upon the recognized dominant trends of the
thin layer model, the data base parameters can be Fig. 9 Schematic diagram of loss classification.
placed in a loss ranking system that indicates the
relative amount of bottom loss to be expected.
Each set of geoacoustic parameters can be assigned IV SUMMARY
a designator giving a direct link to the amount of
loss in different grazing angle regions. A thin layer of acoustic energy interacting with
the sea floor has been utilized to study the
For example, a ranking valid from 50-1000 Hz can sensitivity of the predicted bottom loss with the
be constructed using a three digit designator in geoacoustic parameters. Parameters for inputs to
the form of a number, letter, number (ie. 4B2). this model are the densities of each layer.
The first digit is a number corresponding to the Ceoacoustic modeling represents a clear,
low angle losses, with the lowest number the improvement in the theoretical computation of
lowest loss. This category is assigned based on bottom loss because it contains the physics of
the critical angle or dissipation loss, where the layer interaction. However, there is one
highest velocity ratio yields the lowest loss, and disadvantage to the geoacousticapproach: there is
for velocity ratios less than unity, the highest no direct relationship between the geo-acoustic
dissipation loss yields the highest loss. The parameters and the resulting losses. For
third digit, a number, corresponds to the high operational use, this disadvantage is severe. The
angle loss level, in which low numbers indicate analysis in this study reveals that the low angle
low losses. This category is assigned based on loss mechanisms are independent from the high
the sediment density, thin layer thickness and angle loss mechanisms. The low angle losses are
thin layer density, where dense sediments yield caused by losses of the transmitted ray that
lower losses. The second digit, a letter, depend upon the sound speed ratio, sound speed
indicates the sediment thickness class. An A gradient and the. attenuation within the sediment
corresponds to thick sediments while a D whereas high angle losses are caused by reflection
corresponds to very thin sediments. The effect of from the water/sediment interface and depend
this parameter is to cause a shift in the knee of primarily on the sediment density. The sediment
the loss curves toward the origin so that class D thickness affects the position in angle of the
curves have more propagation loss at low angles change between these this regimes. These
than class A curves, however, this shift only generalizations enable the formulation of a loss
becomes noticeable at frequencies below 200 Hz. classification scheme that links the geoacoustic
An example of the ranking system is shown in Fig. parameters directly with the amount of loss to be
9, where 6 low angle and 4 high angle classes have expected in specific angular regions.
been def ined. The dashed lines demonstrate the
cross linking between low and high angle. The Comparisons made between the predictions f the
lower figure illustrates the effect of sediment thin layer model and the group averages using the
thickness classes. As an accuracy check, some loss classification scheme have shown exceptional
3000 geoacoustic parameter and frequency agreement. Better than 88% of 3000 cases fall
combinations were chosen at random and comparisons within +2 dB of the average when reorganized using
were made between the full model result and the the loss classification.
classification scheme with simplified model
result*. The result was that an overwhelming
majority of bottom loss curves (better than 88%) The simplified model is a least squares fit to
fell with +2 dB (everywhere in frequency and the average loss curve of each loss category.

308

REFERENCES

1. Knobles, D. P. and P. J. Vidmar, "Simulation
of Bottom Interacting Waveforms," J. Acoust.
Soc, Am. 79(6), pp. 1761-1766 (1986).

2. Vidmar, P. J., "Ray Path Analysis of the
Sediment Shear Wave Effects in Bottom
Reflection Loss," J. Acoust. Soc, Am. 68, pp
639-648 (1980).

3. Spofford, C. W., Greene, R. R., and J. B.
Hersey, "The Estimation of Geoacoustic Ocean
Sediment Parameters from Measured Bottom Loss
Data," Science Applications Inc., McLean,
22102, Report No. SAI-83-879-wa (1983).

4. Spofford, C. W. , "Inference of Geo-Acoustic
Parameters from Bottom-loss Data," Bottom
Interacting Ocean Acoustics, W. A. Kuperman
and F. B. Jensen Ed. Plenum Press, New York,
p. 159-171 (1980).

309

SOURCE SHIP CONTAMINATION REMOVAL IN A BROADBAND
VERTICAL ARRAY EXPERIMENT

W.S. Hodgkiss

Marine Physical Laboratory
Scripps Institution of Oceanography
San Diego, CA 92152

A13STRACT
In March 1987, the Marine Physical Laboratory conducted a kn, the HX-90 source was at approximately 105 in (350') in
broadband experiment 400 nmi west of San Diego in 2200 fm depth.
water. A 55 element vertical array was deployed from FLIP The presence of the tow ship at close range presented a
to a depth of 305 in. During the experiment, several high-level, broadband contaminant in the data. The focus of
broadband source tows were made at close range to.FLIP this paper is on the use of adaptive cancellation techniques to
with typical CPA's on the order of 1 rimi. These data were remove the tow ship contamination.
collected in order to perform cross-correlations between
hydrophone pairs and between beams receiving the direct and
surface reflected paths. The presence of the tow ship at close
range presented a high-level, broadband contaminant in the H. Adaptive Interference Rejection
data. The focus of this paper is on the use of adaptive noise
cancellation techniques to remove the tow ship contamination. One approach to the problem of interference rejection is
The adaptive processing is accomplished by treating the time the use of an adaptive noise cancelling structure. In this case,
series obtained at the output of the beam pointing at the tow a broadband beam pointing at the source ship is used as an
ship as the reference input to a simple adaptive noise interference reference. The output of this reference beam is
cancelling structure with each hydrophone time series (in adaptively filtered, then subtracted from the output of one of
turn) treated as the primary input. After 'noise cancellation, the array elements. This procedure is repeated for each
the contamination-free individual element signals then are element of the array. The contaminant-free element signals
available for processing as originally intended. then are available for processing as originally intended.
The all-zero, joint process, least-squares lat tice (JCLSL)
structure used to accomplish the noise cancelling operation is
I. Introduction shown in Figure L, The reference channel process x(n) is
filtered to form an estimate of the negative of the primary
The Marine Physical Laborat Iory conducted a broadband channel process d(n) (often known as the desired response).
Of particular interest in noise cancelling applications is the
vertical array experiment in March 1987 approximately 400
residual ed(n) obtained by adding the filtered reference
nmi west of San Diego. The FLIP (FLoating Instrument channel to the primary channel.' In the application being
Platform) was towed out to 32 * 49. 1'N, 124' 59.2'W (water considered in this paper, d(n) is the output of one'of the array
depth approximately 2200 fm) and drifted in the vertical in elements contaminated by the source ship signature and x(n)
the vicinity of these coordinates for the remainder of the is the output of the reference beam which is pointing directly
experiment. at the source ship. The reference beam is prevented from
Shortly after arriving on station, the MPL vertical array passing signals propagating from directions of interest (i.e. the
was deployed to a depth of approximately 305 in (top surface reflected and direct paths of the source) due to its low
element). Of the 55 available elements across the array sidelobe response in those directions. The adaptive filter is
aperture, 50 functioned properly. Array element spacing was attempting to generate a good estimate of the contaminant
3.75 in (half-wavelength at 200 Hz). present in d(n).
The FLIP tow ship provided sound source tow services The equations describing the JCLSL implemented here
for the experiment. During the active periods, the source ship have been discussed in detail elsewhere and will not be
towed a HX-90 sound source emitting an index CW signal for repeated [1-21. Additional discussion on adaptive least-
approximately 2 minutes at the beginning of each magnetic squares lattice structures can be found in [3-51.
tape recording the digitized hydrophone data followed by
broadband (nominally 100-300 Hz) for the remainder of the
tape (approximately 16 minutes).
I Typically, the constant-course runs began and finished at
a range of 5 nmi on opposite sides of FLIP with CPA's of
approximately 1 nmi. With a typical source ship speed of 5.5

CH2585-8/88/0060-310 @$1 @1988 IEEE

-d 0 (n) -^dP(n) .-d i (n)
d
a (n) ad
d (n) + 0 d (n) P + N
1 151 Stage P 1h Stage d
K d K
0 r. (n) 'p (n) r,, 1(n) + r(n)
r
K J
a (n) P + a(n)
X (n) L

Figure Ia. All-zero, joint process, least-squares lattice. Figure 1b. ilk stage of the lattice.

M. Data Amalysis Figure 6 displays power Spectra at the output of time-
delay beamformer beams pointing at the surface reflected
Figures 2-8 are from an analysis of Tape 07052 which arrival (-16.6 * ) (top panel), tow ship arrival (-13.1 * ) (middle
encompasses the CPA of the first source tow. Figure 2 panel), and direct path arrival (-10.2 * ) (bottom panel)
displays power spectra from hydrophones *, 20, 40, and 55 calculated during the fifth segment of Tape #7052 (16384
calculated during the first tape segment (65536 data points or points or 13.1 s starting after 262144 points or 209.6 s with
52.4 s with f =1250 Hz). The line near 200 Hz is the CW fs=1250 Hz). The broadband signal being projected by the
8 source (approximately 100-300 Hz) is easily seen in the surface
signal projected by the source at the beginning of each tape. reflected and direct path beam spectra. The additional
Figure 3 is vertical arrival structure of this CW signal as bandwidth apparent (300-400 Hz) is due to a resonant peak in
calculated by a FFT beamformer. The direct path arrives at the HX-90 compensating for the lowpass roll-off
approximately -10 and the surface reflected path arrives at characteristics in the source signal generation network.
approximately -16 Note that the surface reflected path
arrival is on the order of 1.5 dB lower in level than the direct Similar to Figure 6, Figures 7-8 also display power
path arrival. spectra at the output of time-delay beamformer beams. In
Figure 4 displays power spectra from hydrophones both cases, the top panel corresponds to the tow ship arrival
20, 40, and 55 calculated during the fifth segment of Tape (-13.1 ' ) which is used as the reference input to the adaptive
07052 (65536 points or 52.4 s starting after 262144 points or filter. The bottom panel in Figure 7 corresponds to the direct
209.6 s with f =4250 Hz). A broadband signal is being path arrival (-10.2' ) beam formed from the array element
projected by tshe source (approximately 100-300 Hz) at this time series after adaptive cancellation. It differs little from
time but it is not easily seen in the single hydrophone spectra. the same beam in Figure 6 which has not had adaptive
cancellation applied to the element time series. In contrast,
Figure 5(a) displays broadband vertical arrival structure the bottom panel in Figure 8 corresponds to the time series
as calculated by a time-delay beamformer where the input from array element * after adaptive cancellation. Note the
element signals have been bandpass filtered to 100-300 Hz. A significant difference between this power spectrum and the
slice in the waterfall represents 256 beam power estimates power spectrum from the same element prior to adaptive
where the output of each beam has been averaged over 64 cancellation which is displayed in Figure 4 (top panel).
points in time. The entire waterfall represents only 1.6 s in
time. Note the severe contamination of the display due to the
tow ship arrival (-13.1 *). Figure 5(b) displays the same data
after adaptive removal of the tow ship contamination. The M Summary
adaptive processing was accomplished by treating the time
series obtained at the output of the beam pointing at the tow This paper has focussed on the use of adaptive n Ioise
ship (properly delayed to account for propagation delay across
the array) as the reference input to a simple adaptive noise cancellation techniques to remove tow ship contamination
cancelling structure with each hydrophone time series (in from array element time series. A broadband beam pointing
turn) treated as the primary input. In this case, the JCLSL at the tow ship was formed and used as the reference input to
parameters p (number of adaptive stages) and a (adaptation a least-squares lattice adaptive noise canceller. Each
rate parameter) were set to p = 8 and a = 0.02. The hydrophone time series (in turn) was treated as the primary
contamination-free individual element signals then were input. The contamination-free individual element signals then
bearnformed yielding the results in Figure 5(b). The ability to were beamformed. The results indicate removal of the
remove the tow ship contamination from the individual contaminating tow ship signature while preserving the
element signals is important in order to calculate meaningful broadband transmission from the towed source.
cross-correlations between pairs of hydrophones.
d
0 n

d
n

d d
tK
@7K

311

Acknowledgernents References
This work was supported by the Naval Air Systems
Command and the Office of Naval Research under contract [11 W.S. Hodgkiss and D. Alexandrou "An Adaptive
N00014-87-C-0127. Dr. Fred Fisher was chief scientist during Algorithm for Array Processing," IEEE Trans. Antennas
the experiment. and Propagation, AP-34: 454-458 (1986)
[21 D. Alexandrou, "Boundary reverberation rejection via
constrained adaptive beamforming," J. Acoust. Soc. Am.
82(4): 1274-1290 (1987).
[3] D. Lee, M. Morf, and B. Friedlander, "Recursive Least
Squares Ladder Estimation Algorithms," IEEE Trans.
Acoust. Speech. Signal Proc., ASSP-29: 627-641 (1981)
[41 B. Friedlander, "Lattice Filters for Adaptive
Processing," Proc. IEEE 70: 829-867 (1982).
[51 B. Friedlander, "Lattice Methods for Spectral
Estimation," Proc. IEEE 70: 990-1017 (1982)-

rum

40
0 50 100 150 200 250 300 350 400 450 5@ 5@ 6@

1 20 Fr.* -y ft)
80 Figure 2. Power Spectra: Tape 07052.1.
60 FFT 13in Width = 153 mHz.
40 Calibration: dB//,uPa/*%/Hz.
Channels #1, 20, 40, and 55.
a 50 13D 150 2W 2% 3W Y10 400 450 sm 550 bw
1 40 F-q-V N.)

6
4:

0 50 100 150 280 250 300 350 400 450 500 550 6W
el 55 Freq@y (Hz)
80 lit

40-

0 50 100 150 20e 250 300 Y-0 400 450 W 550 600
FreqxmN (Hz)

Array RErsponGe 87052. L Bfn #5416
1 201.26 Hz, KB indow alpha 1.5 (above), rect wmdow (be]

Figure 3. OW Signal Vertical Arrival
S - 20 Structure: Tape 07052.1. Sea state
5. Wind speed 18 kn. 19 March
1087, 03:50 GMT. FFT Bin Width
153 inHz. Kaiser-Bessel (a = 1.5)
-40 and rectangular shading functions.
-90 -80 -70 -60 -50 -40 -30 -20 -10 e 10 20 30 40 50 60 70 60 90 Positive angles refer to downward
looking beams. Calibration:
dB//IiPa/'\/HzDeg.

9-20
P Spect ?052

-40
-90 -80 -70 -60 -50 -40 -30 -20 -10 8 10 213 30 40 Se 60 7e 80 90

ANI. (dg)

312

@er Spectrm 8?052.5
a] O'l

60
40

0 50 100 200 250 300 350 400 450 0 550 600
el f2O Fre*wry IHz)

0 50 lee 150 200 250 UO . 350 400 450 500 550 600
el 40 F-M-N (HA

0 50 100 150 200 250 3M 350 40D 450 500 550 600
al -55 Freq,-y (M

0 50 lee 150 200 250 300 35e 400 450 5M 5W 6M
Fr.q.n (Hz)

Figure 4. Power Spectra: Tape M7052.5. FFT Bin Width = 153
mEz. Calibration: dB//,uPa/\/Hz. Channels #1, 20, 40,
and 55.

Rr-roy Response 87052.5 t 0 E; ArroV Pesponse - 87052. 5 1 z 0 s
Bundpuss FEI-tered. le@-300 Hz Bondposs Filtered, 100-300 Hz

'F
-40 -35 -30 -25 -15 L 40 0
35 -301 -25 0' 45' 40' '-5
Angle (deg) Angle (deg)

Figure 5a. Broadband Vertical Arrival Structure: Tape jW7052.5. Figure 5b. Broadband Vertical Arrival Structure: Tape *87052.5.
Bandwidth: IOG-300 Hz . -Positive angles refer to downward Bandwidth: 100-300 Hz. Positive angles refer to downward
looking beams. Uncalibrated. Before removal of tow ship looking beams. Uncalibrated. After adaptive removal of,
contamination (-13.1 * arrival angle). tow ship contamination (-13.1 * arrival angle).

313

r trum 87052.5
1BE! [0 014, A A
80 IT P w-r- I' I -r-

60 --

50 1 0 150 200 250 300 350 400 450 500 550 600

Z)
j LIP'., Wj J,,J LA A M
Be NIV PIT 17 1 f, ITV I N, IM 11' V Y07 re 6. Power Spectra: Tape 07052.5.
FFT Bin Width = 153 mHz.
60 Calibration: dB//IiPa/%/Hz.
40 Beams -16.6 * (source surface
reflected path), -13.1 * (tow ship),
510 100 150 200 250 300 350 400 450 500 550 @O and -10.2 (source direct path).
Bearn -9 @(H?)

0 50 100 150 200 250 300 350 400 450 500 550 600

Frequency (HA

S ctrum ?0 2.5

60
40 Figure 7. Power Spectra: Tape 07052.5.
FFT Bin Width = 153 mHz.
50 100 150 200 250 300 350 400 450 500 550 600 Calibration: dB//pPa/N/Hz.
Beams -13.1 * (tow ship) and -10.2
(source direct path) (after removal
80. of tow ship contamination)-

0 50 100 150 200 250 300 350 400 450 500 550 600

Frequency (Hz)

S ctrum - P705g.5

80 FTVPJ IRRIVr I NI 44" 7"1 YN

60
40 T Figure 8. Power Spectra: Tape M7052.5.
0 50 100 150 200 250 300 350 400 450 500 550 600 FFT Bin Width = 153 mHz.
Calibration: dB//,uPa/\/Hz. Beam
Frequency (Hz) -13.1 * (tow ship) and Channel #1
i i (after removal of tow ship
80 contamination).
& 60

40
@r 4tru- 8 @11125

S, 4S@r t r u 052.
7"

4S&j,.t;u. 705 -5

0 50 100 150 200 250 300 350 400 450 500 550 600

Frequency (Hz)

314

A NEW ACOUSTIC TECHNIQUE FOR REMOTE MEASUREMENT OF THE TEMPORAL OCEAN WAVE SPECTRUM
R.J. Lataitis, G.B. Crawfordl, and S.F. Clifford

Wave Propagation Laboratory
National Oceanic and Atmospheric Administration
Environmental Research Laboratories
325 Broadway
Boulder, CO 80303

ABSTRACT

A method is described that uses low-frequency
sound (< 200 Hz) to measure the waveheight
variancZ and the nondirectional temporal
waveheight spectrum of a random rough surface
such as the sea surface. The technique requires
a vertically pointing broad-beam acoustic source
and a colocated receiver that records the
amplitude and phase fluctations of the back-
scattered field. When the rms surface
waveheight is much smaller than the acoustic
wavelength, the temporal spectrum of the
backscattered amplitude and phase fluctuations
can be directly related to the nondirectional
temporal surface waveheight spectrum.
Typically, surface waves with wavelengths of a
few meters or more can be probed, and the
corresponding temporal waveheight spectrum out
to frequencies of roughly I Hz can be retrieved.

1. INTRODUCTION
Figure 1. A decomposition of the sea surface
The temporal structure of the sea surface is into its spatial Fourier components. (Pierson,
usually measured using buoys and/or wave et al., 19551).
staffs. A variety of acoustic techniques are
also gaining popularity. These usually involve The actual sea surface is the superposition of
measurements of the backscattered power or all surface wave components of different
Doppler spectra, both of which contain surface amplitudes, wavelengths, and propagation
wave information. We describe an alternative directions weighted by the directional surface
method in which a single broad-beam source is waveheight spectrum. Provided a @ << X, each
used to insonify the sea surface from below and surface wave component produces a sinusoidal
a colocated receiver is used to detect the perturbation in the amplitude and phase of the
amplitude and phase fluctuations of the back- backscattered field2 as shown in Fig. 2.
scattered field. The temporal statistics of the
amplitude and phase can then be used to extract
the temporal statistics of the sea surface pro-
vided the rms surface waveheight ce << X, where
X is the acoustic wavelength. The method is
simple, it requires little signal processing,
and it should provide an accurate measure of the
nondirectional temporal waveheight spectrum over
a broad range of frequencies.

This method is best described by first
considering the Fourier decomposition of the sea 2V'
surface illustrated in Fig. 1. Each surface 2A
wave component is a two-dimensional sinusoidal
waveheight perturbation of a given amplitude and Figure 2. A point source of sound insonifying a
wavelength propagating in a particular direction. single Fourier component of the sea surface.
2V

1G.B. Crawford is currently with the Department of Oceanography, University of British Columbia, Vancouver,
British Columbia, Canada V6T IW5.

CH2585-8/88/0000- 315 $1 @1988 IEEE

For a sinusoidal surface wave component with a where w = 2wf, Se(w) is the nond recti nal
spatial wavelength A and propagation velocity temporal surface wave spectrum, tF(K)i 2 is the
V', the backscattered amplitude and phase transfer function relating the amplitude of a
perturbation pattern in an observation plane at surfne wave component with wavenumber
the source is also sinusoidal and has a spatial K = IKI - 2w/A to the amplitude of the
wavelength 2A and a propagation velocity 2V1. correspondingFourier component of the
The factor of two is due to the diverging nature backscattered wave perturbation pattern, and g
of the spherical wave field. The perturbation is the gravitational acceleration at sea
in the backscatter field is therefore merely the level. The details of the derivation leading to
projection of the surface wave onto the Eq. (1) are presented by Lataitis et al.,3 and
observing plane. The amplitude of the are not repeated here. Their approach is based
perturbation is proportional to the amplitude of on a perturbation expansion for the back-
the surface wave. The full backscattered scattered field presented by Labianca and
amplitude and phase perturbation pattern is the Harper4, and a method for relating the acoustic
superposition of the fields scattered from all wave to the surface wave fluctuations presented
of the surface wave components. In the same way by Fuks.5 Equation (1) requires that I << kz,
as a buoy or wave staff measures the temporal 14a @ A, that the surface be in the Fraunhofer
structure of the surface by detecting the regfon of the source I and that the first order
waveheight fluctuations of each surface wave dispersion relation 12 gK hold for the surface
component as it propagates by, a single receiver waves.
measures a proportional quantity by detecting
the corresponding amplitude and phase
fluctuations in the backscattered field as they
propagate by. We note that the frequency f
detected by a buoy at the surface (i.e., f - V'/A)
is the same as that observed at the receiver
(i.e., f = 2V'/2A), so that there is a one-to-
one correspondence between the frequency induced
at the surface and the frequency observed by a
remote receiver.

This paper examines the theoretical limitations
of this technique. In particular we present an
expression for the transfer function that
relates the spatial structure of the surface to
the spatial structure of the backscattered 466:0
amplitude and phase perturbation pattern. By
using the first order surface wave dispersion
relation we determine the range of surface wave 01
frequencies that can be observed with this Z
technique.

2. THEORETICAL RESULTS

Consider the backscatter geometry shown in
Fig. 3. A point source of sound is located a
distance z below the plane of the mean
surface. The surface waveheight E describes Figure 3. Scattering geometry
excursions of the surface about its mean level
and is a tunction of time t and the transverse Equation (1) relatts the observed log-amplitude
position pl in the mean plane defined from the and phase spectra C(w) of the backscattered
origin of coordinates located directly above the field to the surface wave temp ral srctrum
source. We assume that the maximum frequency of SESw). The transfer function IF(K)I is given
oscillation of the surface is much less than the by
source frequency wo = kco, where k = 27/X and co
is the sound speed. We also assume that the IF(K)I' = IF (K)I 2 + IF (K)I 2, (2a)
waveheights are sufficiently small so that the X S
scattered acoustic signal will be a narrow-band where
process about %. The backscattered amplitude 3 J (Kp')
and phase fluctuations are recorded by a FX(K) = 2kz 0
f dP'P' 2 _2
colocated receiver. The sum ZM of the power S 0 (p, + z )2 1(p')
spectra of the log-amplitude and phase Cos
P7@2
fluctuations, where the log-amplitude is the x [2 (-2b)
natural logarithm of the amplitude normalized to sin
the mean amplitude, is given by3 In Eq. (2b) the integration is over the mean
2 2 12, urface plane, Jo is the zero order Bessel
CM - 4k S (W)IF(K - W /g) Iflunction, and I is a function describing the
acoustic beam pattern at the surface. For a
simple Gaussian beam pattern function
66" 0

316

_P' 2/(2z2 tan 29)
I(p') = e (3) between the temporal surface wave spectrum and
the measured power spectra of the amplitude and
where 0 is the half-beam width of the source or phase fluctuations of.the backscattered field
receiver whichever is narrower, Eq. (2) can be was presented. Our results indicate that the
evaluated using straightforward stationary phase surface wave temporal spectrum can be retrieved
techniques. For kz >> I we obtain 3 for angular freque2cies W sajisf Mg
2 2 2 0 < W < Y"2-kg [2tan e/(1+2tan 0)] . As an
IF(a) 12 (1-a 2) 2e- a /[(l-a ) tan a < 1, (4a) examplZ we consider a situation where
0 a > 1, (4b) a - 0.5 m typical ofa bay, inlet, or coastal
rigion. A system with X - 7m mounted at a depth
where a K/2k. Equations (4) are plotted in of say 10 or 20 meters could continuously
Fig. 4 for kz = 60 and 8 = 90* (i.e., an monitor the surface wave temporal spectrum over
isotropic source), 0 45% and 0 15% frequencies 0 < f < 0.6 Hz provided a beam width
of 45' was used. We feel that this technique is
a simple, attractive alternative to existing
100 methods for monitoring the temporal behavior of
the sea surface.

10
4. ACKNOWLEDGMENTS

0=90* This work was supported by the Office of
0=45' Naval Research under Contract NOO014-0090.
0.1-
0 =150 5. REFERENCES
0.01- 1 . Pierson, W.J., Jr., Neumann, G., and James,
R.W. (1955). Practical Methods for
0.0011 Observing and Forecasting Ocean Waves by
0.01 0.1 10 Means of Wave Spectra and Statistics, (U.S.
a Navy Hydrographic Office, Pub. No. 603)
Figure 4. The transfer function IF(o,)12 defined p. 24.
in Eq. (2) plotted as a function of a = K/2k for 2. Lee, R.W., and Harp, S.C. (1969). "Weak
kz = 60 and various values of the half-beam width O@ scattering in random media, with appli-
cations to remote probing," Proc. IEEE, 57,
The significance of the cut-off at K = 2k in 375-406.
Fig. 4 for an isotropic source is that surface
waves with higher wavenumbers produce scattered 3. Lataitis, R.J., Crawford, G.B., and
waves that are evanescent and decay within a few Clifford, S.F. (1988). "A simple low
wavelengths of the surface. Therefore there is frequency acoustic technique for remote
no information about these shorter surface waves measurement of the temporal ocean wave
in the backscattered field at the source. We spectrum," (submitted to Jour. Acoust. Soc.
note also that the cut-off frequency decreases Am.).
as the beam width narrows. This is due to the
reduction of the effective field of view of the 4. Labianca, F.M., and Harper, E.Y. (1977).
system which filters the contribution of the "Connection between various small-waveheight
shorter surface waves. This filtering occurs solutions of the problem of scattering from
because the shorter surface waves scatter the the ocean surface," J. Acoust. Soc. Am., 62,
incident acoustic field into larger angles and 1144-1157.
eventually out of the systems,field of view. We
can identify the shortest surface wave from 5. Fuks, I.M. (1975). "Determination of the
which we can extract useful information by parameters of ocean waves from amplitude and
determining where the curves in Fig. 4 fall to phase fluctuations of reflected radiowaves,"
say e 2 of their value at a = K/2k = 0. From Izv., Atmospheric and Ocean Physics,lb
Eq. (2b) we find that this occurs when 1038-1046.
K = 2k[2tan20/(1+2tan20)]1/2. Using the first
order surface wave dispersion relation, this
translates to an Iservable 5 reqy7@cy band
0 < w < /2kg [2tan 0/(1+2tan 0)]

3. DISCUSSION

We have described a simple acoustic technique
that uses a single broad-beam source and a
7 e

colocated receiver to remotely monitor the
nondirectional temporal surface wave spectrum.
A transfer function describing the relationship

317

ATTENUATION OF LOW FREQUENCY SOUND IN OCEAN SURFACE DUCTS:
IMPLICATIONS FOR SURFACE LOSS VALUES

David G. Browning and Peter M. Scheifele
Naval Underwater Systems Center
New London Laboratory
New London, Connecticut 06320

Robert H. Mellen
Kildare Corporation
New London, Connecticut 06320

ABSTRACT BACKGROUND

The attenuation of low frequency sound in the sea For a typical deep ocean sound speed profile,
is pH-dependent, i.e., the higher the pH, the various modes of propagation will dominate
greater the attenuation. In most ocean areas, the depending on the depth of the sound source (Fig.
value of PH changes significantly with depth so 1). sound fixing and ranging (SOFAR) propagation
this must be included in an accurate attenuation centers along the principal sound channel axis,
computation. Previous determinations of other surface duct propagation along a surface layer,
parameters, such as surface loss which were based while convergence zone (CZ) propagation covers a
on older attenuation formulae, should be major portion of the water column.
reexamined. An analysis of a previously reported
surface loss formula indicates the predicted Previously the same value(s) of attenuation had
values are too high. been used for all propagation modes [7], but as
can be seen from a representative deep water PH
INTRODUCTION profile, different propagation modes will
experience different values of pH (Fig. 2).
Nothing is more fundamental to sound propagation Generally, the surface duct mode will see the
in the ocean than the attenuation of sound in sea highest PH and, hence, the highest attenuation;
water. Depending on the propagation path, some- the pH minimum is near the sound channel axis,
times there is bottom loss, sometimes there is thus, SOFAR propagation will have the lowest
surface loss, and sometimes there is reverbera- attenuation. CZ propagation will be somewhere in
tion, but every time there is attenuation. between..

Since the discovery of a new low frequency COMPUTATION OF ATTENUATION
attenuation mechanism in 1965 (1], our knowledge
has rapidly evolved (2] to a comprehensive The comprehensive attenuation formula for sea
attenuation formula [3] based on three relaxation water is based on three chemical relaxation
processes of which two are pH-dependent. As a processes: boric acid (B(OH3), magnesium
result, the higher the value of pH, the higher carbonate (MgC03), and magnesium sulfate
the attenuation [4]. (MgS04) (see Fig. 3). The first two mechanisms,
which are significant at low frequencies, have
The purpose of this paper is to present the been recently discovered and are both
impact of the latest attenuation formula on pH-dependent.
surface duct propagation loss predictions and to
revise values of low frequency surface loss [5) The pH-dependency can be expressed in terms of a
that were previously obtained from surface duct "K" factor, and plots of K contours will indicate
propagation loss measurements and are based on the relative change of attenuation. Two examples
older attenuation formulae [6]. are given in Fig. 4 showing K contours at the

SOFAR MODE

SOUND CONVERGENCE ZONE
SPEED (CZ) MODE
PROFILE
SURFACE DUCT MOD

Fig. 1. Sound speed profile, propagation modes

318 United States Government work not protected by copyright

2 0

3 0 N. PACIFIC
0 N. ATLANTIC
Z@ IVIED. SEA
ARCTIC
4 t
7.6 7.8 8.0 8.2 8.4

pH
Fig. 2. Typical deep ocean pH profile

A Al (M9SO4) + A2(B(OH)3) + A3(M9CO3)

f2f /(f2 + f2)
A, = (S/35) an n n

a, = 0.5 x 10-D(krn)/20 1 50 X 1 OT160
a2 = 0-1 X 10(pH-8) 1`2 = 0.9 X 10T170
a3 = 0.03 x 1 &H-8) f3 = 4.5 x 1 OT130

I K = 10(pH-8)

3-COMPONENT ABSORPTION MODEL

DEFINE K FACTOR AS: K=10(pH.8)

Fig. 3. httenuation formula

319

1:? ISOUnd-Channel Confou ISurface onfou
so >1.2 __j 60

1.9

1.0
30 0.9 30 --\40 <2.0 T\ -

0.8 >_0'6 @1.6
0.7 1.7
0 a 1.9
>0.6 0.8
30 0.7 30

1.8 1.7
so I [ @=i6
90 60 30 0 30 60 90 UO 90 .1 __30 0 30 60 90 120

60 @0.7 60 4__ -1.6 -

0.6 1.8

30 >01 0
[email protected]
L

0
V 0.e 0.7 N-E-j7'
0.8 I-N
30 K

0.9

so so
120 ISO 160 150 120 90.. so 90 120 150 ISO ISO 120 90 so
Fig. 4. Attenuation contours: Sound channel axis Fig. 5. Attenuation contours: Surface

sound channel axis and Fig. 5 shows K contours at. attenuation that, at that time, only included the
the surface. The previously used Thorp Formula magnesium sulphate,relaxation.
corresponds - to a value of K=l, so, as can be
seen, this would be a low estimate for most The original propagation loss data from AMOS
near-surface calculation, high for some SOFAR appears to have faded into history, but if we
propagation. assume a duct thickness of 60 m and a corre-
sponding skip distance (limiting ray) of 6.3 km,
To illustrate the application of the new we can work backwards. By adding the old
attenuation formula to surface duct propagation, attenuation loss to the surface loss formula to
a calculation is made for a duct In the Bismark get the total loss, we can present a rate-of-loss
Sea (8] and the result compared to experimental versus frequency comparison (Fig. 7) which shows
measurements (Fig. 6). The agreement is good the surface loss rate for various wave heights on
whereas calculations using an average value the top, then the new attenuation, and below that
(Thorp formula) would be low, and calculations the original attenuation. You can visualize the
done before approximately 1970, which used correction as the difference between these last
attenuation formulae without any low frequency two attenuation curves, and, very interestingly,
mechanism at all, would be significantly too low. this difference curve has the same shape as the
rate-of-surface loss curves.
RECALCULATION OF SURFACE LOSS
We can, therefore, revise the formula for surface
in 1968 Schulkin [6] developed an expression for scattering as follows:
surface loss based on surface duct measurements
taken during the Acoustic, Meteorological, and As = 1.0 (Fh)1/2 dB/bounce.
oceanographic survey (AMOS) cruises, i.e,
Again we cannot claim that this is definitive
As = 1.6 (Fh)1/2 dB/bounce, because we do not have the original data but it
is certainly a reasonable estimate of the impact
which was determined using the Marsh-Schulkin that the new attenuation formula has on the
calculation.

320

(dB/kM)
A BISMARK SEA pH 8.3 300C
SURFACE DUCT,&z 90m

DIFFRACTION A2

Al 10- --K A3

.01
10
FREOUENCY (kHz)
Fig. 6. Bismark Sea attenuation: surface duct

(dB/km) SURFACE DUCT
Az = 60 m
pH 8.3 250C
h(m)
2
E 1 oo

0
_j
U.
0
w

NEW a OLD a

.011
10
FREOUENCY (kHz)
Fig. 1. Rate-of -loss plot
A@/

00-

321

CONCLUSIONS 3. R. H. Mellen, P. M. Scheifele, and, D. G.
Browning, "Global Model for sound Absorption
The attenuation of low frequency sound is in Sea Water," NUSC Scientific and Engineering
pH-dependent which is depth-dependent. Studies, Naval Underwater Systems Center, New
London, CT (1987).
For accurate predictions, you must include
a depth-dependent attenuation variation. 4. R. H. Mellen and D. 0. Browning, "Variability
some propagation modes are more critical of Low-Frequency Sound Absorption in the
than others. Ocean: PH Dependence," J. Acoust. Soc. Am.,
vol. 61, pp. 704-706 (1977).
The computation of other parameters such as
surface loss should be reexamined. 5. M. Schulkin, "Surface-Coupled Losses in
Surface Sound Channels," J. Acoust. Soc. Am.,
vol. 44, pp. 1152-1154 (1962).
REFERENCES
6. M. Schulkin and H. W. Marsh, "Absorption of
1. W. H. Thorp, "Deep-Ocean Sound Attenuation in Sound in Sea Water," J. Acoust. Soc. Am., vol.
the Sub- and Low-Kilocycle-Per-Second Region," 34, pp. 864-865 (1962).
J. Acoust. Soc. Am., vol. 38, pp. 648-654
(1965). 7. W. H. Thorp, "Analytic Description of the Low
Frequency Attenuation coefficient," J. Acoust.
2. J. R. Lovett, "Geographic Variation Of LOW- Soc. Am., vol. 42, p. 270 (1961).
Frequency Sound Absorption in the Atlantic,
Indian, and Pacific oceans," J. Acoust. Soc. 8. R. H. Mellen and D. G. Browning, "Attenuation
Am., vol. 67, pp. 338-340 (1980)@- in Surface Ducts," J. ACOUSt. Soc. Am., vol.
63, pp. 1624-1626 (1978).

322

THE WIDE SWATH, DEEP TOWED SEAMARC

.Arthur St. C. Wright

Williamson & Associates, Inc.
Seattle, Washington 98109

ABSTRACT SeaMARC I was a deep-towed device capable of full
ocean depth with operating frequencies of 27 and
The advantages of sea bottom 'imagery collected 30 kHz. It was operated by Lamont-Doherty
with a deep towed system'to support geophysical Geological Observatory and was lost at sea in
studies of the deep seafloor for commercial and 1984.
research applications are frequently not
appreciated. Of the SeaMARC family, no two are
alike in terms of operating capabilities and, data SeaMARC II is a shallow-towed device capable of 10
output. Only the privately owned Sea MARC.IA cuts kilometer bathymetric swaths. Operating at 11 and
across and supports all regimes. The paper 12 kHz, it obtains off track depths by a phase
discusses the use of deep towed wide swath bottom angle measuring technique. This sonar is operated
imaging sonar and subbottom profiler. Data by Hawaii Institute of Geophysics.
examples are presented from SeaMARC IA, a sidescan
sonar operating at 27-30 kHz to image the seafloor SeaMARC IA is a full ocean depth system with oper-
in swath widths up to 5000 meters. Acoustic ating frequencies of 27 and 30 kHz, swath capabil-
response of the seafloor to the sidescan and 4.5 ities from 500 to 5000 meters and incorporating a
kHz subbottom profiler from a wide range of envir- sub-bottom profiler and a digital processing
onments are examined and the -image processing system. It is owned by International Deep Sea
techniques are shown. Survey, Inc. and operated by Williamson &
Associates, Inc.

SeaMARC IB was similar to SeaMARC I and was
1. INTRODUCTION initially purchased by Lamont-Doherty but was
subsequently transferred to Woods Hole
The many sidescan sonars in use today for mapping Oceanographic Institution. It was lost at sea this
and imaging fall into two groups; one group for year.
working the shallower waters, and the second for
working to full ocean depth for a variety of The SeaMARC CL-X is a smaller system which
purposes. The SeaMARC (Sea Mapping and Remote operates at 150 kHz with a maximum swath width of
Characterization) family was one of the early 800 meters. It was developed as an experimental
entries in the deep ocean field. Today, the name model in 1985 and is operated by Williamson &
SeaMARC is almost synonymous with towed sidescan Associates, Inc.
deep ocean imagery or swath bathymetry. These
devices are descendants of the HSES or High Speed The SeaMARC S is a medium depth, production
Exploration System which was developed by a conso- follow-on to the CL-X manufactured for Seafloor
rtium for manganese nodule exploration in the late Surveys International, Inc. in 1986. SSI has deve-
seventies. (The U.S. member of this consortium was loped a phase angle measuring system for swath
the International Nickel Company which is not an bathymetry which has been integrated into this
active participant in offshore exploration at this system.
time.) many of the personnel associated with the
development of the HSES subsequently formed assoc- In addition to complete systems, IST has also
iations that led to the development and operation provided arrays and primary electronics to
of the SeaMARC line. customers for their own further development. In
early 1988, International Submarine Technology
All SeaMARC's have been manufactured by ceased operations and sold the rights to the
International Submarine Technology, Ltd. of, SeaMARC line to Honeywell Inc. Reportedly, more
Redmond, Washington. There have been two general SeaMARC's are in process at the 146neywell plant
types; a deep-towed imaging system, and a near near Everett, Washington.
surface towed swath bathymetry system.

CH2585-8/88/00oo- 323 $1 @1988 IEEE

. . . .... . .. . ...- A variety of navigation systems have been used
with SeaMARC IA. Two types of positioning must be
considered; vessel positioning, and towfish to
vessel positioning. For vessel navigation,
advances in electronic and satellite navigation in
recent years have improved the accuracy in
positioning. The increase in the number of GPS
satellites will further increase the accuracy of
vessel positioning. Determining the towfish
position relative to the ship can be done with a
short baseline system, or in the situation where
the cable length is over 10,000 feet, a manual
calculation has to suffice. The most accurate
method is a long baseline system in which an inte-
rrogator on the SeaMARC towfish queries a
calibrated transponder array. This removes the
I L1@11 ship from the calculation and has provided results
to a two meter accuracy.
Figure 1. SeaMARC 1A towfish SeaMARC IA has been to sea on vessels ranging in
length from 120 to 350 feet. The basic
requirements are that the vessel accommodate the
personnel and equipment and be able to maintain
track in the sea conditions expected in the survey
SYSTEM DESCRIPTION area. Frequently other activities in addition to
SeaMARC operations are planned which mandates
The major components of the SeaMARC IA system are additional equipment and space integration. The
the towfish, cable and depressor, topside handling ship must have sufficient room for the handling
equipment, surface electronics package, and data equipment, and working area for the SeaMARC IA
.acquisition and processing systems. The towfish electronics, processing equipment, and navigation
(Figure 1) carries the port and starboard sidescan computers. The personnel count varies according to
arrays and associated electronics, the toroidal the requirement for on-board -processing and
,beam pattern 4.5 kHz subbottom profiler, and evaluation, but an average SeaMARC and navigation
various sensors and recovery beacons. The towfish scientific party numbers about ten. While the
weighs 1600 pounds in air and is slightly buoyant; SeaMARC IA has operated in gale conditions, the
it tows well at speeds up to five knots. A limiting factor in such adverse weather has been
standard 0.68 inch marine coax of up to 30,000 the ability of the ship to maintain track. Bow
feet connects the towfish to the winch. A 1600 thrusters have been used to advantage in some
pound depressor is positioned three hundred feet circumstances and are a desirable feature.
ahead of the towfish to eliminate towing vessel All the equipment except the winch and crane fit
heave and to mitigate the effects of pilot error. into a twenty foot long container that is easily
The topside handling components are a towpoint, road transportable and can be welded to the deck
which resembles a half A-frame with a gallows of the support vessel for a support and
block, a hydraulic winch with level wind, and a maintenance center at sea. If necessary the system
crane for towfish launch and recovery. If the is also air transportable (but it is in the
support vessel doesn't have a hydraulic deck client's better interest to find a crane locally).
outlet, a diesel power pack is also necessary. The Mobilization requires two to three days with the
surface electronics package contains the sonar critical path generally involving welding.
controls and the analog-to-digital processors.
Transmit power, receiver gain, swath width
selection, pulse burst length, and bandwidth can SEAMARC IA OPERATING FUNDAMENTALS
be controlled for each sonar. The data acquisition
and image processing systems consist of the recor- Tfi&key advantage of deep-towed sidescan sonars is
ders, computers, video monitors and tape and disk that all the errors associated with the water
decks which are used to simultaneously record and column effects and beam spreading are eliminated
process the data. and the Arrays are transmitting and receiving in
the immediate vicinity of the "target".
A critical component of the SeaMARC IA operation Inaccuracies in depth measurements caused by ray
is the availability of tools and spare parts so path bending and bottom penetration by the
that any casualty short of towfish-loss can be relatively low frequencies of surface systems and
overcome. Once offshore, clients take a dim view modulation 'of return signal characteristics by the
of having to suspend an operation to return to water column are also avoided. The philosophy of
port for repairs. So, included in the support this sonar is to present the best imagery possible
equipment are all tools that could conceivably be and the@ operating parameters have been selected
required, a replacement board for surface and accdrding'ly. In the design phase, two desired
subsurface systems board, many spare hardware, capabilities for SeaMARC IA were the ability to
electrical, and electronic items and a complete image large areas of seafloor in a wide swath mode
library of technical publications and drawings. and the ability to find small targets in the

324

narrow swath, high resolution mode. The Swath Altitude Pulse Length Bandwidth
flexibility offered by the variable swath widths meters meters milliseconds Hertz
from 500 to 5000 meters provides the seafloor
mission planner with powerful capability. In any 500 25-125 .15, .2, .3 5000
areal survey, the ability to obtain an initial 1000 25-250 .3, .4, .6 2000
wide swath look at the target area precludes navi- 2000 50-500 .8, 1.2 1000, 2000
gational surprises and indicates those areas which 5000 125-1250 1.2, 1.6, 2.4 500, 1000
deserve a closer look. The ability to accomplish
both wide area and high resolution missions with The bandwidths tabulated above are related to
the same machine saves capital costs and hours at tuning of circuits in the electronics signal proc-
sea in recovering and restreaming equipment. essing section and are not the same as the mechan-
ical transducer array bandwidth referred to
The design selection ol, frequency for a sidescan earlier.
sonar is dependent on the maximum swath width
desired. Once the frequency has been selected, the As opposed to the resolution of a sidescan sonar
mechanics of the transducer array construction in the cross-track direction being set by its
determine the characteristic of the resonance operating characteristics, the resolution and
curve which is assigned a value known as "Q". A consequently the pixel dimension along track is
lower Q means that a wider bandwidth about the set by vessel speed. For a given beam width, if
center frequency is available for signal the vessel moves too fast a small target may be
processing. An important point concerning resolut- missed. But, if the vessel moves such that each
ion of a sonar is that high resolution is not point is imaged several times, and the towfish
dependent on high frequency but rather on pulse is stable in pitch, roll, and heading. Then a
duration. A sonar puts power in the water in small target whose return signal is slightly
pulses, the length or duration of which determines higher than background will appear as a short,
the minimum size of a discernible object and the sharp vertical line at the appropriate range.
shape of which determines the sharpness of the (Figure 2) Therefore, the along track dimension of
return signals. The key to a short, well shaped a target or feature can be calculated from
pulse burst is wide bandwidth. The best resolution knowledge of the angular dimension of the beam
,.of an analog signal is equal to the velocity of width, the range to a target, and the ship's
sound in water times the pulse width expressed as speed.
time with due regard for units. In a digital pres-
entation, the best resolution attainable is the
physical size of a pixel in the cross-track
direction and is equal to the swath width divided
by the number of pixels in the scan line. For
optimum results digital and analog resolution
should be matched.

In the case of SeaMARC IA, the desired maximum
swath of 5000 meters mandated a maximum frequency
of 30 kHz. (The lower 27 kHz on the port side
array is for the purpose of eliminating cross-
talk.) This selection of 30 kHz also set the mini-
mum pulse width obtainable because the lowest "Q"
achievable with the transducer arrays dictated a
maximum bandwidth of 6 kHz. This converts to a
pulse duration of 0.15 milliseconds which equates
to a pulse width of 22.5 centimeters in seawater.
This 22.5 centimeters is the smallest dimension
that can be discriminated by the analog system. In
the SeaMARC IA operating procedures the 0.15 mill-
isecond pulse duration is for highest resolution
on the 500 meter swath. Since there are 2000
pixels in a scan line, the digital resolution of Figure 2. Zoomed image of a small, hard target
the system is 25 centimeters which essentially (Note the 25 cm pixels)
matches the analog resolution.

The altitude and control settings normally used
for the various swath widths for the SeaMARC IA For good imagery, returning intensity levels must
are: be displayed in as many shades or colors as the
human eye can discern. The SeaMARC IA Image Proce-
ssing System sorts incoming signals into 256
intensity levels and displays them in 16 shades of
gray or other pseudo-color. The number of
intensity levels assigned to any of the 16 colors
can be varied or "thresholded" so that specific
portions of the returning intensity level spectrum

325

PROCESSED SEAMARC IA IMAGE OF THE
_'n":: AXIAL CALDERA
A

JUAN OF FUCA RIDGE -NE PACIFIC

I km
-W-a. @&_-

7Z,

@_Z
%

. .... . .. . ......

Figure 3. SeaMARC 1A mosaic of the Axial Seamount Caldera in the Northeast Pacific.
The image represents 2 tracklines run with a 2 km swath width.

can be assessed. In signal processing terms, Figure 3 is a SeaMARC IA mosaic of the caldera on
twelve bits are reserved for each pixel, but only top of Axial Seamount in the northeast Pacific.
eight bits are being utilized. Although the rim wall, which is up to 150 meters
above the caldera floor, dominates the image,
other geomorphological features are well
SeaMARC IA EMPLOYMENT expressed. The high reflectivity of the caldera
floor on the left-most part of the image is due to
As offshore oil development, cable route surveys, fresh basalt which has probably emanated from the
mineral exploration, installation of sensor large fissure intersecting the rim wall in that
platforms, salvage operations and other area. The small cone in the lower left may be
engineering projects occur in progressively deeper still building, while the collapsed remnant of an
water, traditional methods of obtaining seafloor older one can be seen just above it.
data are proving inadequate. The high cost of
ocean engineering projects and the financial Figure 4 is a two kilometer swath image showing a
penalties associated with an engineering failure seafloor fault running upwards and to the right.
place a high premium on finding the optimum In the lower left portion, sharp spires of
location for a project and evaluating all the basaltic material which project several tens of
potential geo-hazards. meters from the seabed. Both the fault and spires
Conventional hull mounted multi-beam and shallow should be considered significant hazards to. a
pipeline installation, cable route or a mining
towed sidescan interferometric sonar systems may operation.
measure water depth to an accuracy of one percent
and can generate contour maps of the seafloor Another geo-hazard, shallow gas, is depicted in
which provide general indications of depths and Figure 5. Taken on a 500 meter swath, this image
slopes but little information on seafloor shows gassy sediment at the seafloor and free gas
properties. venting from the bottom. Much of the subbottom
detail is masked in the gassy area due to differe-
nces in acoustic velocity.

326

9@1
Towfish path

@iAq 1,

@t,

Figure 4. Image of a fault in
2500 m of water,
recorded on the 2 km
swath width

2 km swath width

Subbottom profile of methane gas

Towfish path

A, ax

M
-A
@Yt
P
P U.
14 1
@A j
.@O i
j" ! -- @ @ f
'r "A 44@ @Z' @r'
I T

'p
ey a R
-4 W M 44, A,
'F -ff

F Jr A V,
CA

-JA
zj' T V

500 m swath
"N"

erir;_. t-r-ii

Figure 5. 500 meter swath of shallow methane gas. Also shown is the 4.5 kHz subbottom profile
as the towfish crossed above the degassing area.

327

CONCLUSIONS

Records from a deep-towed sidescan sonar and
subbottom profiler such as SeaMARC IA supplement
the information available from surface systems and
provide the quality data necessary to plan modern
seafloor engineering projects. Where additional
geotechnical information is necessary, the SeaMARC
records will indicate the most promising positions
for taking core samples or conducting ROV
operations. The greater costs associated with deep
towed operations are minor compared to the risks
involved in the loss of a pipeline or cable route
because a fault or other geo-hazard was not
detected and evaluated.

ACKNOWLEDGMENT

The authors wish to thank the National Atmospheric
and Oceanic Administration (NOAA), Amoco Prod-
uction Company, and Exxon for permission to use
data samples presented in this paper.

REFERENCES

1. Kosalos, J.G. and Chayes, D.N., A Portable
System for Ocean Bottom Imaging and Charting;
published in Proceedings of OCEANS '83, Aug. 29 -
Sept. 1, 1983; p. 649-656.

2. Kosalos, J.G., Ocean Bottom Imaging,; Presented
at Offshore Technology conference, May 7-9, 1984

3. Williamson, M.E., Application of Wide Swath,
Deeply Towed Bottom Imaging Sonar in Geotechnical
Evaluation of the Seafloor; published in Current
Practices and New Technology in Ocean Engineering
- 1988 OED-Vol. 13 by A.S.M.E.

328

A New Generation side Scan Sonar

Robert G. Asplin
Carl G. Christensson

Mesotech Systems Ltd.
Port Coquitlam, B.C., Canada
ABSTRACT SOME DESIGN CONSIDERATIONS
Side Scan sonar has traditionally used a Operating Controls
graphic paper recording system for the There is a tendency to use keypads and
display and annotation of the data. A new displays as the operating controls for
Side Scan system specifically designed to much of todays electronic equipment.
use color video as the primary data Several reasons for this approach are:
display is described. Several design
considerations and trade-offs are - there are often too many variables to
discussed, some of which are unique to a control to allow the use of a dedicated
system using a video display. control for each function,
A parametric Sub Bottom Profiler and - invalid or illogical control setting
towfish orientation sensors are also combinations can be easily prevented or
integrated into the system. warned against,

- changes or special features can be
added requiring only software
modifications,

- it makes for a simple, clean looking
control panel,
INTRODUCTION - it is inexpensive.
The main objection to this scheme is that
The Mesotech Model 972 is a new Color Side it can be very frustrating to use,
Scan Sonar with an optional Sub Bottom especially when the control to be set is
Profiler function. Some of the main buried under several layers of menus.
features of the system are:
- High resolution full color display This sonar system uses dedicated controls
(1280 x 1024 pixels x 7 bits/pixel). in the form of conventional rotary
switches and potentiometers for all of the
- Tape recording/playback using standard most often used controls. Secondary
VHS cassettes. controls are adjusted using a software
driven menu system and a simple keypad.
- Neutrally buoyant towfish de-coupled Choice of a Color Display
from ship motion.
- Sensors in the towfish to monitor The most desirable characteristic of the
color video display is its ability to
depth, altitude, pitch, roll and speed. display data with a much wider dynamic
- Dual frequency Side Scan operation (100 range than is possible with a paper
kHz or 330 kHz). recorder system.
- Display corrections for speed and Images are formed on the display because
altitude of the towfish. the intensity of the reflected acoustic
return varies continuously within each
- Only two conductors required in the shot. If the return is too low in
cable to the towfish. amplitude, it will be shown in black and
will not be visible. If the return is too
- Operation from 110/220 VAC or 20 to 50 high in amplitude, it will be clipped to
VDC. the maximum display level and information
will be lost.

CH2585-8/88/0000-329 $1 @1988 IEEE

Therefore, there is a limited range of PITCH: over a range of 20 degrees.
reflected signal levels which can be
displayed. This is the dynamic range of ROLL: over a range of 20 degrees.
the display.
COMPASS (optional): magnetic heading to a
The system used here is known as an analog resolution of 1.4 degrees.
RGB video system, meaning that the
intensity of the three primary colors, SPEED: to a resolution of 0.1 knot.
red, green, and blue can be independently
controlled in very small steps. A total DEPTH: as measured by a pressure sensor to
of 128 return levels can be distinguished a maximum range of 300 meters.
by controlling both color and intensity of
each pixel on the display. This is ALTITUDE: as measured by a downward
considerably more than the 8 or 16 levels looking echo sounder in the towfish to a
which can be distinguished on a typical maximum range of 250 meters.
black and white paper recorder.
Only the speed and altitude sensor data is
The Neutrally Buoyant Towfish used to make corrections.to the display.
The other sensors are provided only to
Ideally, the towfish would be towed verify that the towfish is level and at
through the water at a constant speed and the expected depth.
depth with no pitch, roll, or yaw.
However, the towing vessel is normally Gain Control System
subject to surface wave action which will
result in continuously changing tension on Reflected return signals will vary over a
the tow cable. range of more than a million to one, much
wider than the display dynamic range.
Figure 1 illustrates the method used in Much of this variation is due to signal
this system to de-couple the towfish from loss with increasing range. This loss is
ship motion. The weighted depressor on fairly predictable and can be compensated
the end of the armored cable will tend to for with TVG (Time Varying Gain)
move up or down as cable tension changes. amplifiers. There are two main factors
The neutrally buoyant towfish is not causing the loss with range:
subject to these upwards or downwards
forces since it is being pulled 1) Spreading Loss
horizontally at the end of a neutrally
buoyant tether. 2) Absorption Loss

SPREADING LOSS
WINCH/TOW CABLE ASSBABLY When the sonar pulse is transmitted from
the transducer, a fixed amount of energy
is spread over a small area of water equal
to the area of the transducer face. As
the pulse travels through the water, it
spreads out over a wider area, resulting
in less energy per unit area. The energy
NEUTRALLY BOUYANT per unit area will decrease as the square
SIDESCAN SONAR of the distance traveled. By the time
EQUIPPED TOWFISH some of this energy is reflected back to
the transducer, the energy level will be
reduced by a predictable amount
NEUTRALLY BOUYANT according to how far it has traveled (how
TETHER long it. took).
DEPRESSOR
FIGURE 1: TOWFISH DECOUPUNG METHOD This loss is compensated in the towfish by
a TVG amplifier which increases its gain
at the predicted rate of loss due to
spreading. There is only one operator
To,wfish orientation Sensors adjustement for this gain correction, the
START gain. The rate of increase is
The towfish includes several orientation pre-set to a 20*log(range) function and is
sensors. Data from these sensors is not adjustable.
displayed on the screen and stored on
tape.

The sensors measure the following:

330

ABSORPTION LOSS This is accomplished as follows:

As the sound pulse travels through the a) A fixed length command block is sent
water, additional energy is lost to the from the surface to the towfish through
water due to an effect called absorption. the 2-wire telemetry system. This
The rate of loss varies with frequency as block includes parameters to specify
well as with salinity and temperature. the frequency (120 kHz or 330 kHz) and
This loss is expressed in db/m (decibels the desired shot interval time.
per meter) since it is linearly
proportional to the distance traveled. It b) The towfish will continuously repeat a
is compensated for in the sonar by a cycle of sending a sync tone to
second TVG amplifier whose gain is the surface, firing the transmitter,
designed to increase at a constant rate or receiving the echoed signals, and
slope. This SLOPE gain can be adjusted by simultaneously sending a TVG corrected
the operator either up or down from the and frequency shifted replica to the
default values of 0.035 db/m for 120 kHz surface.
and 0.065 db/m for 330 kHz. The default
values used are representative of data c) A new command block is sent only if@
from several published sources. They tend one of the operating conditions has
to vary from source to source since they changed.
are empirically derived values.
The towfish is fitted with sensors to
SYSTEM ORGANIZATION measure pitch, roll, depth, speed through
the water, and optionally, magnetic
The Mesotech Model 972 Side Scan Sonar heading. The information from these
System consists of the following major sensors is sent to the surface in fixed
electronic components: length data blocks through the telemetry
system. This occurs approximately every 2
a) Towfish Electronics Module seconds.

b) Left and Right Side Scan Transducers SUB BOTTTOM PROFILER SYSTEM

c) Echo Sounder Transducer (and optional The sub bottom profiler system uses a
Sub Bottom Transducers) technique known as parametric mixing to
generate a directional low frequency
d) Two-Conductor Cable to surface pulse. A relatively low frequency is
required to acheive penetration into the
e) Surface processor module bottom material. The higher frequencies
used by the side scan system are reflected
f) RGB color monitor from the bottom surface.

g) Tape recording system consisting of a The sub bottom system consists of two
PCM encoder and a VHS video tape transmitters, each with its own
recorder. transducer, and a receiver with a third
transducer. One transmitter/transducer
TOWFISH MODULE operates at 200 kHz while the second one
operates at 210 kHz.
The main function of the Towfish Module is
to collect the raw sonar data and send it The two frequencies mix in the water
to the surface for further processing and producing two more frequencies, the sum
display. and difference frequencies. The sum
frequency of 410 kHz is quickly absorbed,
The Towfish Electronics Module contains but the difference frequency of 10 kHz is
the acoustic transmitters, receivers with able to penetrate the typical bottom for a
20 log R TVG function, telemetry receiver few meters. The receiver is tuned to 10
and transmitter, control processors and kHz and receives the reflected 10 kHz
power supplies. It is designed to operate signals from the bottom and layers below
on a cyclic basis at a rate given by the the bottom.
surface unit rather than having to be
triggered from the surface for each shot. The method is very inefficient
electrically, but relatively little energy
is required. A 10 kHz transmit transducer
with the same directional characteristics
would be prohibitively large for a
towfish.

331

TELEMETRY SYSTEM g) The video display controller converts
its memory to analog RGB video
The telemetry system between the Surface signals which drive the color display.
Processor and the Towfish module uses a
full duplex 300 baud modem system for the h) Sensor data from the Towf ish
down-link command block and up-link data Module is received and displayed
block. Frequency division multiplexing is numerically on video screen.
used to send the side scan, sync, and sub
bottom signals to the surface. i) If the Speed Correction mode is
Frequencies used are as follows: enabled, the firing rate and display
scrolling rate are modif ied as
Digital Down Link :1170Hz required.
Digital Up Link :2125Hz
Right Side Scan (Up) : 80kHz j) If the Water Column Removal or Slant
Left Side Scan (Up) : 5OkHz Range Correction mode is enabled, the
Sub Bottom (Up) : 65kHz digitizing rates are adjusted as
Sync (Up) : 100kHz required.
The telemetry system signals are In addition to its basic function of
superimposed on the 48 volt DC power displaying the realtime sonar data from
supplied from the surface to the towfish, the Towfish Moduler the Surface Processor
thus requiring a total of only two will also display data from the tape
conductors in the tow cable. recording system.

RECORDING SYSTEM
SURFACE PROCESSOR
The recording system is basically a 5
The basic function of the Surface channel audio recorder using standard VHS
Processor is to convert the sonar signals video tapes as the recording media.
from the Towfish Module into an image on
the color display. Several processes are A Sony PCM encoder is used to digitize two
involved in this function: channels and record them on the video
track. Two more channels use the stereo
a) The panel controls are sensed to HI FI tracks on the tape. The fifth
determine the desired operation mode. channel uses the linear tracks recorded in
mono. Normally, a standard video recorder
b) Appropriate commands are sen-t,-t-o the will allow use of either the HI FI track
Tovufish--Module "ffo establish the correct or the linear tracks for audio, but not
data collection rate. both at the same time. The recorder must
be modified to allow their simultaneous
c) When the sonar data is received at the use.
surface, a second level of TVG
correction is applied to compensate on playback, the PCM encoder converts the
for absorption loss. digitized data on the video track back to
d) The sonar and sync signals are two analog channels.
converted to the audio range The 5 channels are assigned to the sonar
to allow them to be recorded data as follows:
on the external recording system.
Left Side Scan Left PCM
e) The sonar signals, now in the audio Right Side Scan Right PCM
range, are amplitude detected, Sub Bottom Left HI Fi
digitized, and stored in a buffer Sync Right HI FI
memory. Serial Data Linear
f) The digitized data is copied from the The serial data consists of digital sensor
buffer memory to the video display data, Surface Processor control settings,
controller by the main CPU. and externally input RS-232 navigation
data.

332

DISPLAY CORRECTIONS Water Column Removal

Speed Correction In an un-corrected display, the data
plotting is started synchronously with the
The purpose of the speed correction mode transmit pulse. This means that the first
is to control the scrolling rate of the
display such that the vertical and section of the display will be showing
horizontal scales are equal. This means returns only from the water column. If the
that the scrolling rate (and the 'shot' start of data plotting is delayed by an
rate) must change with the selected range amount equal to the height of the towfish
scale and also with the speed of the 'above the bottom (altitude), the water
towfish over the bottom. column will not be displayed. This is the
water column correction mode. It can be
In a system using a paper recorder, it is enabled for the side scan display and
possible, at least theoretically, to vary independently for the sub bottom display.
the paper advance speed continuously to There is one small difference with the
compensate for towing speed variation. water column removal in sub bottom: one
With a video system, however, it is only meter of water column is deliberately left
possible to increment the display by an in the sub bottom display. This allows
integer number of pixels. Even with a the water-bottom interface to be clearly
high resolution display such as is used visible.
here, increments of one pixel do not
provide fine enough control over the Water column removal requires the altitude
advance speed to make an accurate speed of the towfish to be known. This can be
correction with a fixed 'shot' rate. The taken automatically from the towfish echo
solution used in this system is to vary sounder or it can be entered manually with
the 'shot' rate in very small increments a front panel control.
while adjusting the display advance by
replicating an entire row of pixels where Water Column Removal Limitations
necessary.
With water column correction enabled, the
Towfish speed through the water is water column and any targets in the
measured by a paddle wheel sensor in the water column will be removed from the
nose of the towfish. Alternatively, speed display. They can be recovered by playing
can be entered by means of a front panel the tape back in the un-corrected mode.
control, or through the RS-232 interface.
A second limitation occurs when the
Speed Correction Limitations towfish altitude changes rapidly. This
can be caused by a change in towfish depth
The towfish sensor measures speed through or bottom depth. The altitude data from
the water, not the desired speed over the the towfish has a four second update rate
bottom. Using speed through the water, due to telemetry. This means that the
however, will tend to compensate for speed display can have jagged edges near the
changes due to winching in or out. Speed bottom due to instantaneous errors in
set from the front panel or externally altitude.
could be set to match an average speed
over the ground, if known, but it would Similarly, water column correction using
not compensate for winching. The speed altitude from the panel control can easily
correction is thus only a good have an altitude error. The effect will
approximation. be to remove too much or not enough water
column. This can be used to advantage in
A second consideration is the fact that some cases if it is desired to simply
the transmitting rate is varied along with remove most of the water column and still
the scrolling rate to match the speed. see the water-bottom interface.
The transmitting rate can only be
decreased from the maximum rate for a
given range. This means there will be
fewer 'hits' on a target with speed
correction enabled. Un-corrected speed
will always give a maximum transmitting
rate, and thus maximum 'hits' on a target
for any speed.

333

Slant Range Correction MASTER/SLAVE MODES

The side scan image plotted in the It is possible to connect two model 972
un-corrected mode is an image created by processor units and two displays together
plotting each point at its slant range in a master/slave arrangement with a
distance from the towfish. This can single towfish. In this configuration,
result in considerable distortion in the the master system is the processor
plotted size of objects very close to the connected to the towfish. The second
towfish relative to those further away. system is set to the slave mode (by means
The slant range correction mode applies a of a front panel switch) and is connected
geometric correction in an attempt to to the master by means of a master/slave
minimize this distortion. The slant range connector on the rear panel. The slave
correction takes place in two stages: system can then be set independently to a
First, the water column is removed as different display mode or even to a
described above. Second, the sample rate different range.
of the side scan is varied within each
'shot' as a function of altitude and The master/slave configuration thus allows
range. The result is a side scan image many more display combinations. For
with the horizontal scale representing example, the master can be set to show
true horizontal distance on the bottom. left side scan data while the slave shows
right side scan data, each using the full
Slant Range Correction Limitations display width.

The basic assumption of the slant range The master/slave arrangement can also be
correction is that the bottom is flat and used in the tape play back mode.
level. Since this is not always the case,
the correction can only be considered an
approximation.

The slant range correction is also
dependent on an accurate and up-to-date
altitude from the towfish. This is needed
for both phases of the correction: water
column removal and sample interval
control. Furthermore, the altitude must
be less than 50% of the range setting for
the correction to work properly. This is
seldom a limitation except on extremely
short ranges.

DISPLAY FORMAT

The display can be set to any one of seven
different formats showing any combination
of left side scan, right side scan and sub
bottom data. The full width of the
display is used in each case. All
displays are in the 'waterfall' form
whereby new data is plotted at the top of
the display and the older data is scrolled
down. Scaled range markers can be
automatically written over the data. In
addition, a movable cursor can be placed
anywhere on the display area and used to
mark targets with a unique target number.

The bottom area of the display is used for
the menu control system and to show the
towfish sensor data. The current settings
of most of the menu controlled parameters
are also shown continuously in this area,
even if the menu for that parameter is
'buried'.

334

USING A SECTOR-SCAN SONAR TO HUNT FOR SHIPWRECKS THROUGH ICE

EMORY KRISTOF & ALVIN CHANDLER, DONALD SHOMETTE

NATIONAL GEOGRAPHIC SOCIETY, LIBRARY OF CONGRESS

ABSTRACT conventional boat towed side-scan search, we
decided to combine the search with a new search
Many high tech tools are used to hunt for technique. We used a land survey team to layout
shipwrecks, side-scan sonars and magnetometers a search grid on the ice, holes were drilled
being the main ones. These are usually towed through the ice and then a sector-scan sonar was
behind boats. The Geographic is trying another lowered down to within 10 feet of the bottom for a
tool from another platform. In Lake Champlain look. If a target looked worthy a larger hole was
this winter we ran a search through the ice using cut and a MiniRover ROV deployed.
sector-scan sonar to look for one of the Benedict
Arnold 1776 Flotilla. This was done as a trial to There are several underwater archeology pro-
gain knowledge for $everal searches to be con- jects that the Geographic is considering in the
ducted in the high Canadian Arctic. In the remote high Canadian Arctic and in Antarctica where use
frozen terrain of the Arctic the cost of running of a ship as a search platform is out of the
a search from one of the few available ships can question. The search from the ice of Lake
be very dear. The.cost effectiveness of combining Champlain was a good trial run for this impending
a flying magnetometer search with through the ice work. The help of local sportsmen, headed by John
small area searches using sector-scan sonar and LeClair, complete with Honda all terrain vehicles,
LCROVs is being looked at by the Geographic and ice fishing shacks, ice augers and knowledge
Dr. Joseph MacInnis. This approach is evolving supplied the framework for the expedition. A
out of what was learned from the success of the local surveyor, John Deming, laid out the grid
BREADALBANE PROJECT. The Lake Champlain experi- with search holes spaced at every 200 feet.
ment, though not producing a shipwreck, was a
technical success that produces another method of Two sector-scan sonars were used, a Mesotech
search that will prove of use in future applica- 671 supplied and operated by Lt. Robert Gwalchmai
tions. and PO Robert Rheel from the Navy EOD Center at
Indian Head, Maryland, and a UDI 500 KHZ provided
by Chris Nicholson of Deep Sea Systems. The
search was conducted in water up to 300 feet deep.
In February of '88 the National Geographic Given the debris on the bottom, including logs,
conducted a search for one of the gunboats used by and the small size of the target, the 200 foot
Benedict Arnold to slow up the British advance spacing on the holes provided for a lot of
down Lake Champlain during the fall of 1776. The overlapping. Both sonars performed well in the
presence of Arnold's small fleet on the Lake had February cold. The topside electronics were kept
forced the British to take the time to assemble a out of the weather and provided some heat. The
larger counter fleet. Though Arnold's fleet was Mesotech would fit through a six inch hole, and
destroyed it occupied the British efforts until the UDI required a ten inch hole. This difference
the winter storms of 1776 forced them to retire to didn't matter with the one foot or less lake ice,
Montreal till the next spring. The time bought by but would have to be considered when drilling
the actions on Lake Champlain was crucial to the through six or more feet of solid sea ice.
success of the American Revolution. The 50 foot,
one-masted rowing boat that we were hunting for, We ran a three-man survey party out in front
the PROVIDENCE, is a sister of the PHILADELPHIA of two two-man ice drilling parties and two three-
that is on display in the Smithsonian. The man sonar parties. When on a roll across the ice,
PHILADELPHIA was sunk during the Battle of Valcour holes could be imaged with less than five minutes
Island and raised in the 1930's. The badly between holes. A useful way of recording the ima-
damaged PROVIDENCE was scuttled by Arnold the day ges was the Sony video printer supplied by Chris
after the loss of the PHILADELPHIA as the American Nicholson. Half of the search area was covered
forces retreated to the South. before the ice broke up.
Several side-scan searches have been con- A side trip was made to Burlington Bay to
ducted in the last few years in the area check out the wreck of a horse ferry from 1821.
surrounding Schuyler Island. Don Shomette through The wreck was relocated using the sector-scan
his research felt that South of Schuyler was an technique and examined by the MKII MiniRover. Ice
area of high probability. Instead of running a can be a wonderful platform from which to work.

CH2585-mmooo_ 335 $1 91988 IEEE

The experience gained from Lake Champlain will
soon be put to use in much colder and remote loca-
tions.

jr
>

(Top): Surveyor
I runs a grid line
South from
Schuyler Island
on Lake
Champlain.
Ir "IF (Left): UDI
AAj sector-scan sonar
is lowered
through ice.
Cable, topside
electronics, and
generator are all
hauled on a
x"
05 trailer behind an
6
ATV.
77
I__J
'Ali -

336

is!

A,
7777"a =77
A0
07"77777777"

44b
@q" M,
1@4 @' I
za@

fd'

Ilk-

"lop

Vol

(Top): An ice fishing shack is home for the Mesotech 671 operation, being hauled behind its ATV.
(Bottom left): Chris Nicholson reads and records the bottom signal from the UDI sectorscan sonar inside
an ice fishing shack. (Bottom right): Sector scan image of horse ferry on bottom of Lake Champlain.
Bow is pointed toward center of circle.

337

SIDE SCAN SONAR ACOUSTIC VARIABILITY

John W. Nicholson Jules S. Jaffe

Woods Hole Oceanographic Institution
Woods Hole MA, 02543

ABSTRACT of a given area. Variability must also be quantified
if meaningful comparisons of two images of the same
bottom taken at different times are to be conducted.
This paper reports the results of research conducted Many sources of side scan sonar image variability are
on the inherent variability of side scan sonar imagery already recognized. Towfish instability is a well un-
in order to determine the magnitude and nature of derstood aspect of side scan sonar which causes image
image intensity fluctuations. Two experiments are distortion. The departure of the towfish from constant
presented in which a Klein 100 KHz system is op- velocity and attitude result in a misdirection of the
erated under controlled conditions which remove all sonar beams. This can result in a complicated shuf-
but purely instrumental and acoustic causes of image fling of the acquired data which, when displayed,nor-
variability. The result of one experiment in a test tank mally, produces image distortion [Flemming, 1982]. In
is a statistical analysis of the transmitted waveform. this paper we will consider acoustic fluctuations. Pre-
A second experiment conducted from the Woods Hole vious studies have not included side scan sonar [Urick,
Oceanographic Institution Pier allows similar analy- 1982]. The resulting changes in side scan images as a
sis for the combined side scan sonar transmit/receive result of these fluctuations would still remain after sys-
signal path. The results indicate that intensity fluctu- tematic measures had been taken to minimize towfish
ations are multiplicative in nature and spatially and instability and variable imaging geometry. As such,*
temporally uncorrelated. they represent a fundamental limit to image repeata-
bility.
In order to quantify the acoustic variability of side
1 INTRODUCTION scan sonar returns and exclude other potential sources
of fluctuation it was necessary to operate a sonar sys-
Considering the various optical or acoustic methods tem in a manner that permitted a large degree of con-
for imaging large areas of the seafloor, side scan sonar trol. The investigation of image variability was first
is presently the most common. Throughout three subdivided into two experiments. In the first exper-
decades of commercial use it has been a tool for ob- iment the acoustic transmission variability was mea-
taining pictures of the bottom. Although images ob- sured so that the resultant amount of fluctuation due
tained by side scan sonar have always been subject to insonification could be determined. The second ex-
to much scrutiny, conjecture, and interpretation, ana- periment was devised so that repeated images of the
lytic treatment of the imagery has been limited. How- same bottom were obtained.This allowed examination
ever with increasing use of the technique, most notably of numerous images for fluctuations. Results of the
as the primary means of mapping the Exclusive Eco- first experiment indicate what fraction of overall im-
nomic Zone (EEZ) of the United States [Paluzzi et al., age variability is due solely to transmission variability.
1979], technical aspects which may influence side scan The remaining fraction is intrinsic to the acoustic en-
sonar imagery are receiving increased attention. vironment.
One aspect which has received little attention in the
past but warrants analysis is the inherent variability of 2 METHODS
the side scan sonar process. Knowledge of this is im-
portant in understanding how individual sonar targets In the test tank experiment a Klein model 422s-10lef
may appear then disappear during repeated surveys 100 KHz towfish was suspended in a test tank approx-

CH2585-8188/0000- 338 $1 @1988 IEEE

150
RESULTS

100-
The test tank experiment data was analyzed in order
to quantify variability in the transmitted waveform.
50
Figure 1 is a plot of the pressure of the transmitted
waveform versus time. The observed waveform con-
0 sists of a 122 KHz carrier modulated by an envelope
which rises linearly then decays exponentially in 100
usec. A noticeable feature of the waveform is that it is
50 -
clipped at negative pressures. This is most likely due
transducer cavitation caused by the 228 dB re I uPa
-100
0 50 100 150 200 250 300 sonar source level [Clay and Medwin, 1977).
t (microseconds) Of particular interest is the variation of total energy
in each transmission. Knowledge of this allows deter-
Figure 1: Klein 100 KHz waveform mination of expected image intensity variation due to
transmission power variability. The energy of each
waveform in the set was calculated as
imately 5 meters from a reference hydrophone. Ge- 128
ometry of the tank, towfish, and hydrophone were ad- E E An? (1)
justed to prevent multipath interference and assure a n=1
fixed relative orientation of all components. Only one
of two towfish acoustic channels was operated in or- where the p[n] are the pressure samples of each wave-
der to prevent mutual interference. Towfish power and form. The statistics of this distribution are shown in
transmit key signal were supplied by a Klein model 521 table 1.
sonar recorder. The transmitted waveform was sensed
by the hydrophone and sent to the data gathering sys- mean: 77300
tern, which consisted of an IBM PC-AT personal com- standard deviation: 4900
puter containing a Data Translation DT-2851 frame standard deviation/mean - 0.063
grabber card. The frame grabber digitized each trans-
mitted waveform sensed by the hydrophone to 8 bits
at 500KHz and recorded the waveforms to hard disk Table 1: 100 KHz towfish transmit energy fluctua-
files for further processing. tions.
In the second, or pierside, experiment the same
sonar system was deployed from the Woods Hole Having characterized the temporal variability of the
Oceanographic Institution pier. To eliminate tow- sonar power level, the sonar transmission was decom-
fish motion the towfish was mounted on a steel box posed into component frequencies for spectral analy-
beam which spanned two concrete dock pilings 10.5 sis. The mean power in each frequency bin taken as
M above the bottom in 17 M of water. The towfish an ensemble average over all 3500 recorded waveforms
was mounted level with the transducer axis oriented density is shown in Figure 2. The predominant fre-
20 degrees below the horizontal to minimize surface quency is the carrier at 122 KHz with a half power
scattering. A portion of the bottom extending 100 bandwidth of 9.2 KHz. This bandwidth corresponds
M outward from the pier was insonified. The experi- favorably with the manufacturers advertized specifica-
ment consisted of recording consecutive returns over a tion of a 0.1 msec pulselength which implies a 10 KHz
six minute period during slack tide. It was presumed bandwidth. A significant amount of power is seen at
that the sonar and bottom orientation remained con- the extreme ends of the spectrum, with local maxima
stant during this time. The image obtained from each at 226 KHz, 241 KHz, and at the extreme frequency
transmission was recorded on tape and later digitized 250 KHz. This region is probably the result of the
using the same data gathering equipment as the test generation of harmonics of the fundamental frequency
tank experiment. due to the non-linear response of water to high am-
plitude pressure fluctuations [McDaniel, 1965]. Lesser
energy is seen at frequencies below 122 KHz. The local
maximum at 65 KHz is- probably a cavitation gener-
ated subharmonic [Desantis et al., 1967]. Broadband

339

14 400

350 o Gaussian Distribution
12

300-
10

250-

200

150-

4- 100

2 50

-250 -200 -150 -100 -50 0 50 100 150 200 250 0 50 100 150 200 250 300
frequency (KHz) intensity

Figure 2: Mean transmission power spectral density Figure 4: Histogram of pixel intensity for range bin
200, pierside experiment

0.35
0.3 sonar image of 3066 rows and 1024 columns. Here each
row corresponds to a separate transmission and each
0.25 column a fixed range bin. In this representation the
image coordinates (x, r) correspond to transmission
0.2 number and range bin, respectively. The n" column
i(x, n) of the matrix therefore represents a time series
0.15 of the image pixel intensity or acoustic pressure am-
0.1 - plitude from one independent, non-overlapping region
of the bottom, while the n 1h row i(n, r) corresponds
0.05 to the same quantity for the n1h transmission.
0 n A - The probability density function of pixel intensity
-250 -200 -150 -100 -50 0 50 100 150 200 25( for a given column describes the echo stability of a
frequency (KHZ) single object or portion of the bottom. A represen-
tative estimated probability density function for this
Figure 3: Variance of transmission power spectral deri- data set is shown in Figure 4, a histogram of the 3066
sity image intensity values contained in column 200. At
this range a strong return is received from the bottom.
redistribution of energy across the spectrum is also an The figure also displays a Gaussian fit to the observed
effect of cavitation and is observed here. pdf. The mean and standard deviation of the Gaussiaii
The variance of the individual spectral components were assumed to be equal to the mean and standard
is shown in Figure 3. It shows a form similar to the deviation of the data. Compared to the Gaussian dis-
mean power spectral density in Figure 2, with peaks tribution, the histogram contains more points at the
in the same areas in both graphs. However the re- center. In general, the shape of the distribution for ar-
gion corresponding to the carrier frequency is propor- bitrary columns is Gaussian with a slight skew towards
tionately smaller than the regions of significant energy intensity values below the mean. The distributions, are
outside the carrier. This indicates that the majority sufficiently unlike the Gaussian distribution that they
of observed variability in total transmitted energy is fail the chi-square goodness of fit test [Bendat and
found at frequencies of no use to the sonar system, Piersol, 1971].
since these frequencies are filtered out in the record- The mean value of each range bin n, 1(x, n), is shown
ing process. The ratio of power standard deviation to in Figure 5. It should be noted that time varying
mean power in this bandwidth is 0.0201. This is ap- gain (TVG), a side scan sonar feature that applies in-
proximately one third the amount of variation found creasing gain with increasing range, was disabled in
previously in the total energy distribution. this experiment. Empirically Figure 5 can be divided
The digitized data set obtained from the pierside into three regions. The region nearest the towfish,
experiment may be looked upon as a single side scan approximately the first 100 range bins, corresponds to

340

300 0.25

250 0.2-

200-
0.15

0.1

100

a J 0
0 200 400 000 800 1000 1200 0 200 400 600 800 1000 1200

range bin range bin

Figure 5: Pixel intensity mean value vs. range bin, Figure 7: Pixel intensity Coefficient of Variation (V)
pierside experiment vs. range bin, pierside experiment

30 is defined as
V = O'i(x,n) (2)
25- %'(x, n)
20 - The mean value of V is approximately 0.08, with a
range from 0.04 to a peak of 0.14 which occurs at bin
15- 400.
To further evaluate the nature of image intensity
10 fluctuations the joint statistics of the 1024 columns
i(x, n) were evaluated. Of interest is the spatial cor-
5 relation of fluctuations in different range bins. One
'Li technique for evaluating this correlation consists of
0 200 400 600 800 1000 1200 computing the correlation coefficient [Papuolis, 1984]
range bin between pairs of range bins n, and n2. The correlation
coefficient
Orn, n2
CnIn2 - @ (3)
Figure 6: Pixel intensity standard deviation vs. range 0-n 1 0-n ,
bin, pierside experiment is the ratio of the covariance of the two range bin in-
tensities to the product of the variances of the two
ranges of 10 M or less and is generated by volume re- range bins. *
verberation. This is because the towfish was mounted Figure 8 is a plot of the co rrelation coefficient
10.5 M above the bottom. The region between range for range bin 300 versus all range bins. Note that
bins 100 and 600 contains the portion of the bottom C300,300 = 1 as Would be exvected, while all othee val-
with the most intense returns. The spiky nature of this ues fall between +/- 0.4. This shows that the inten-
region is due to the differences in backscatter strength sity fluctuations of data set range .bin 300 are weakly
between the various bottom subregions represented by correlated with the fluctuations in other bins.
i(x, n). After column 600 the signal is greatly attenu-
ated and increasingly noisy. The acoustic transmission power fluctuations ob-
.A plot of pixel intensity standard deviation o-,,, ver- served in the test tank experiment are a possible cause
sus range bin number is shown in Figure 6 and is seen of the weak but wide range correlation of intensity fluc-
to follow the same trend as Figure 5. This indicates tuations between the various range bins in the data
that the fluctuations in i(x, n) are a constant fraction set. Since scattered echo intensity is directly propor-
of the mean. This is seen to be approximately true in tional to the intensity of the insonifying transmission
Figure 7, a plot of the coefficient of variation V [Urick, and all range bins during one transmission cycle are
1982] versus bin number. The coefficient of variation insonified by the same acoustic transmission, pixel in-

341

0.8 0.8

0.6 0.6

0.4--
0.4 0

0.2

0 0 4

0.2 -0.2

-0.4 0.4
0 200 400 Boo 800 1000 1200 0 200 400 600 800 1000 1200
range bin range bin

Figure 8: Correlation coefficients, range bin 300, pier- Figure 9: Correlation coefficients, range bin 300, pier-
side experiment side experiment, compensated data

tensity variations due to transmission power fluctua, for any row n. 7 is the mean intensity of the, im-
tions can be expected to be correlated. Total'energy age as a whole. After removing row-wise image inten-
in each transmission was not measured during this ex- sity fluctuations the previous analyses were performed
periment, but an estimate based on the total energy again. Figure 9 shows the effect of this compensa-
contained in each row i(n, r) provides- satisfactory re- tion on C300,n- Compared to the uncorrected case the
sults. The total energy of each row is calculated as degree of correlation between column 300 and other
3066 7 columns is significantly reduced, indicating that trans-
0(n, r) (4) mission fluctuations are a probable cause of the weak
E,, = E correlation of pixel intensity fluctuations for a given
n=1 transmission or side scan sonar image line. This lack
The statistics'for this energy distribution are shown in of correlation leads.to the conclusion that the inten-
Table 2. The ratio of standard deviation to'mean is sity fluctuations observed at various ranges in a side
scan sonar image are essentially independent, in the
absence of change in the imaged topography.
mean: 3,74" 'Range bin mean, standard deviation, and coefficient
standard deviation: 127 of variation analyses were repeated on the row equal-
standard deviation/mean 0.0'140 i:zed matrix , however the results of these analyses
showed that the equalization process did not change
Table 2: Row energy statistics, pierside experiment these parameters to the same degree that it changed
C.In2. This lack of perceptible change indicates that
larger than the 0.0201 measured in the test tank ex- the temporal fluctuations in pixel intensity for a side
periment. The increased fluctuation is not unexpected scan sonar image are largely independent of transmis-
considering the round-trip pathlength to the region of sion energy fluctuations and cannot be attributed to
maximum intensity in this experiment was as much as them.
200 M as compared to 5 meters in the test tank ex- One final analysis of the pierside data was the com-
periment. The additional interaction of the acoustic putation of the power spectral density (PSD) of the
transmission with the medium is the probable cause. fluctuations of the column intensity sequences i(x, n).
Information about the transmission fluctuations can The power spectral density of fluctuations in range bin
now be applied to the data set in order to equalike it 400 are shown in Figure 10. Outside of a narrow spike
and remove the effects of these fluctuations. The data centered around DC which was caused by an overall
matrix is scaled on a row by row basis, creating a new increasing trend in intensity with time, the spectrum
matrix with elements P(x, r) such that is fairly flat. This implies that the intensity fluctua-
1024 tions are uncorrelated temporally.
1: V(n, r)
r=1

842

12
REFERENCES
10-

8
[1] J. S. Bendat and A. G. Piersol. Random Data:
Analysis and Measurement Procedures. Wiley-
Interscience, New York, 1971.

[2] C. S. Clay and H. Medwin. Acoustical Oceanogra-
2- phy. John Wiley and Sons, New York, 1977.
0. [3] P. Desantis et al. The use of subharmonics. J.
-2.5 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5 Acoust. Soc. Am., (46):514, 1967.
frequency (Hz)
[4] B. W. Flemming. Causes and effects of sonograph
Figure 10: Power spectral density, range bin 400 distortion and some graphical methods for their
manual correction. In W.G.A. Russell-Cargill,
editor, Recent Developments in Side Scan Sonar
Techniques, Central Acoustics Laboratory, Univer-
4 CONCLUSIONS sity of Cape Town, Cape Town, South.Africa, 1982.

[5] 0. H. McDaniel. Harmonic distortion of spheri-
This research indicates that there is a significant cal sound waves in water. J. Acoust. Soc. Am.,
amount of fluctuation in side scan sonar acoustics and (38):644, 1965.
therefore in the images obtained. The lack of tem-
poral or spatial correlation of the observed fluctua- [6] Peter R. Paluzzi et al. Computer rectification
tions shows that the fluctuations are not due primar- and mosaicking of side-looking sonar images. In
ily to transmit power fluctuations. In the absence of Eleventh Annual Offshore Technology Conference
other influences such as towfish instability or changes Proceedings, Houston, Texas, May 1979.
in imaging geometry, fluctuations of up to 14% were (7] A. Papoulis. Probability, Random Variablesi and
observed for images of identical bottom features. The Stochastic Processes. McGraw-Hill, New York,
fluctuations are approximately Gaussian distributed. 1984.
The fluctuations are nearly a constant percentage of
mean intensity and indicating that this is a multiplica- [8] R. J. Urick. Sound Propagation in the Sea. Penin-
tive rather than additive effect. This implies that if sula, Los Altos, California, 1982.
intensities in a side scan image are compensated to
produce the same mean intensity throughout a nearly
constant amount of fluctuation is observed throughout
the image.

ACKNOWLEDGEMENTS

We would like to acknowledge those who provided
fu'nding, equipment, and facilities for this research; in-
cluding Marine Imaging Systems, The Massachusetts
Commonwealth Centers for Excellence, The National
Science Foundation, Martin Klein, and The United
States Geologic Survey Division of Marine Geology.

343

A PRACTICAL HIGH TECH ADVANCE IN SIDE SCAN SONAR
TARGET POSITIONING AND ANALYSIS

William R. Abrams

Seaquest Associates
.Division of Structured Technology Corp.
Box 312, Mystic, Connecticut 06355

ABSTRACT InTAC Description
it is the intent of this paper to Acoustic digitizer technology, forms the
present and discuss a revolutionary basis for this patented method of target
system for analyzing Side Scan Sonar positioning and analysis. Two high
records. This system is called InTAC resolution transducers mounted on the
(Instant Target Analysis Computer). side scan sonar recorder, an acoustic
InTAC enhances target analysis in two pen source and a microprocessor are the
ways. The first speeds up the analysis system elements.
procedure by avoiding tedious manual
measurements of a sonar contact. The The InTAC system is compatible with any
second enhancement is the ability of commercially available side scan sonar.
InTAC to determine the position in With several "clicks" of its pointing
latitude and longitude of a sonar devic.e (the acoustic pen) one may de-
target. This position can then be used termine a target's position only or its
to guide submersibles or divers to the position and size.
exact location of the. target for
investigation or recovery purposes. The InTAC transducers (microphones)
This instant target analysis is per- mounting bracket is installed on the
formed realtime, thus avoiding costly side scan. sonar recorder and does not
project delays. interfere in any way with normal opera-
tion of the side scan sonar. Installa-
tion averages less than an hour for most
OLD ANALYSIS TECHNIQUE systems.
The software for each InTAC is designed
Historically Side Scan Sonar (SSS) for a specific model and make of side
target position and analysis could only scan on which it will operate. Specific
be done during a post-processing opera- model characteristics are taken into
tion after the SSS record had been consideration such as size of paper,
removed from the recorder. Each target altitude correction, speed correction,
of interest required mechanical scale and number of channels.
measurements of the anomaly, as well as
SSS fish position and heading at time of When the InTAC is used in the "on-line"
contact for resolving position or posi- mode, a navigation computer input is
tion and size. Analysis of multiple required. The navigation computer
contacts was a time consuming task and supplies realtime side scan position,
subject to error. heading and altitude at specific inter-
vals to the InTAC via a serial inter-
NEW ANALYSIS TECHNIQUE face. This data is stored by InTAC for
Background use when a target is marked.
Marking of a target is accomplished by
Just prior to the space shuttle touching the acoustic pen to a selected
Challenger explosion, Seaquest engineers target on the sonar record. Touching
(now a division of Structured Technology the acoustic pen to the target generates
Corp.) developed a SSS target analysis an acoustic burst and at the same time
system which may be used on-line or starts a clock in the microprocessor.
during post-processing. The new devel- The microphones detect the short
opment, named InTAC, was first used acoustic burst ("click") and measure
during the search for Challenger debris
at Cape Canaveral, Florida.

CH2585-8/88/0000- 344 $1 @1988 IEEE

the range to the pen. The pen's x,y
position on the sonar trace is then
calculated by the microprocessor using
triangulation with a temperature com-
pensated resolution of 0.01 cm. The
microprocessor then using a "history" of
the position data previously sent by the
navigation computer calculates the exact
real world position of the marked tar-
get. This InTAC calculated position
takes into consideration the time lag
between the time the target was detected
by the sonar and the later time it was
marked by the operator. It also takes
into consideration the altitude and
heading of the sonar at the time of
detection by the sonar. The target
height, width and length may also be
calculated with a few more pen "clicks".

The InTAC provides the option of saving
marked target data to the floppy disk
and/or outputting it to a printer. In
addition to this permanent storage, when
a target is marked, InTAC will transmit
the target data back to the navigation
computer via the RS-232 interface. This
allows a marked. target to be instanta-
neously stored by the navigation com-
puter and plotted on a tactical display.

In summation, InTAC provides a unique
way to accurately position and analyze
side scan sonar targets realtime thereby
making target data immediately available
in a printed and/or tactical display
presentation.

345

CALIBRATION OF ACOUSTIC DOPPLER CURRENT PROFILERS

Gerald F. Appelll, Joel Gast2 . Robert G. Williams1, and Patricia D. Bassl

1. NOAA National Ocean Service 2. RD Instruments
6001 Executive Blvd. 9855 Businesspark Ave.
Rockville, MD 20852 San Diego, CA 92131

ABSTRACT tow tank calibrated EG&G Vector Measuring Current
Meters (VMCM). Intercomparison results provided
Acoustic Doppler current profilers are widely used unacceptable uncertainties during the Charleston
within the oceanographic community for scientific Harbor survey and other methods were sought.
research, engineering analysis and routine current
surveys. They have the ability to profile the The requirement for RADS calibration is not unique
water column over a relatively large range and to NOS and was recently reviewed at the World Ocean
provide high resolution of the current field. Circulation Experiment (WOCE) workshop in early
However, an effective and practical calibration 1988. The conclusion drawn at the workshop was
technique for the determination of current measure- that a complete calibration and characterization is
ment accuracy of these instruments is not an involved process. It requires a combination of
established. Field intercomparisons are used as theoretical analysis, laboratory tow tank testing,
crude methods of achieving confidence in the and in-water testing in operational environments.
measurements but yield mixed results depending on An important part of this overall RADS calibration
the environmental conditions. and characterization is to develop the capability
for routine velocity calibration of RADS under
The National Ocean Service of NOAA has recently controlled conditions. This paper discusses
conducted tests and established a laboratory preliminary results of our efforts to evaluate the
calibration technique for Doppler current pro- feasibility of accomplishing in water DCP cali-
filers. A cooperative effort established with the bration to a precision consistent with the DCP long
manufacturer, RD Instruments, assessed the tech- term accuracy specification of 0.2 percent +/-0.5
nique and the instruments' performance. Several cm/s. Our approach.was:
systems ranging in operating frequencies from 300
to 1200 kHz were tested and calibrated at the David Perform calibration tests on several RADS systems
Taylor Research Center (DTRC) tow carriage faci- in the DTRC tow tank. Assess the degree to which
lity. A 1200 kHz system was characterized by the accuracy can be determined and use the results
manufacturer through a series of laboratory tests as a baseline for evaluating a lake calibration
and navigation runs on a calibrated lake course. facility.
This unit was then tested at DTRC for comparison
purposes. Calibrations of both bottom track and Develop and evaluate a lake calibration facility
water track velocity were performed at speeds from to assess the accuracy to which RADS can be
0 to 300 cm/s. Sources of error were investigated calibrated,using a moving barge navigation tech-
by NOAA and the manufacturer. Test results, cali- nique.
bration data, methods and procedures are discussed.
Develop and evaluate techniques of measuring
critical DCP parameters that affect accuracy in a
laboratory environment - specifically, transducer
1. INTRODUCTION beam parameters such as angle and geometry.

The use of Remote Acoustic Doppler Sensing (RADS) Compare the test results of these three tech-
systems by the National Ocean Service for current niques.
surveys of harbors and estuaries required the
development of calibration methods. The RADS The main results to date are of the initial DTRC
consists of a Doppler current profiler (DCP) and calibration tests, therefore, this paper will focus
all of the associated signal processing, data on presenting the results and conclusions from
transmission and recording systems. It had been these tests. Also, it is not within the scope of
generally believed that DCP's could not be cali- this paper to discuss RADS system operating
brated in tow tank facilities because of acoustic characteristics and terminology. The reference
reflections from the walls and bottom of the basin. section cites background literature for further
The initial attempt to satisfy NOS's calibration reading.
requirements was by in-situ intercomparisons with

346 United States Government work not protected by copyright

2. DTRC CALIBRATION TESTS transferred to a second Compaq 286 for analysis and
production of graphical displays. Adjustments were
A test of an RD Instruments 1200 kHz DCP was made to the test plan based on the interpretation
conducted in the Circulating Water Channel at DTRC of the results.
in mid 1987. The system was bottom mounted at 3-
meter depth in an upward looking mode. Although a These tests were successful in establishing the
rigorous test was not performed, the system feasibility of the tow basin calibration concept.
provided a good indication of velocity in two A formal program was established to calibrate all
1-meter depth bins. The results suggested that DCP's used by NOS for surveys. A total of nine
laboratory calibrations may be possible in a large systems all manufactured by RD Instruments were
facility. Thus, a test plan was prepared to involved: six 1200 kHz units, two 600 kHz units and
conduct experimental tests in the towing basin at one 300 kHz unit.
DTRC. This is a much larger facility and offered
an accurate speed measurement and control system. Initial DCP parameter settings were explored to
optimize the system performance and obtain high
The first DTRC tow basin tests were conducted on a quality data. The DCP was initially set up to
RD Instruments 1200 kHz DCP, serial number 0217, at record single ping ensembles for the determination
DTRC during the week of November 10-13, 1987. The of standard deviation (standard deviation is not
system was attached to the vertical rails of the computed in the DCP). The bin width was 1 meter,
carriage and the transducers were submerged to a the blanking distance (The programmable distance
depth of about 30 cm (Figure 1). The DCP is between the transducer and the first bin) was set
typically configured with 4 beams oriented 30 to 0.5 meter and the number of bins of data
degrees off vertical in 90-degree azimuth incre- recorded was 14. The time between ensembles was 1
ments. In the azimuth plane, beams 1, 2, 3, and 4 second and a minimum of 100 ensembles was recorded
are positioned at 0, 180, go, and 270 degrees, per test speed. Radial beam coordinates were
respectively. Mechanical alignment of the trans- recorded plus header data, echo amplitude and
ducers was performed to assure that tilt errors spectral width. Tilt attitude was adjusted and
were less than 1 degree. The characteristics of mechanically measured to within +/-1 degree.
the towing basin are: Single ping ensembles require that data be trans-
ferred after each-ping. This limits the ping rate
basin length 275 meters that can,be achieved. The low ping rate resulted
basin width 15 meters in a slow response of the DCP to tow carriage
basin depth 6 meters velocity changes.
fresh water environment at ambient temperature
carriage speed range 2.5 to 750 cm/s The second series of tests was conducted with the
estimated speed measurement uncertainty +/-0.1 DCP set up to record 20 ping ensembles at a ping
cm/3 rate of 5 per second. The other parameters
RADS CALIBRATION remained the same. This set up improved the
response of the DCP to carriage velocity changes.
A minimum of 12 ensembles were collected per test
run, depending on tow speed.

Tests were performed with the beam 3 and beam 4
pair aligned into the tow direction. Data were
first acquired under no-tow conditions in still
water. The tow carriage was then accelerated to
RADS the desired speed and then DCP sampling was
started. After 12 ensembles we stopped sampling
ARRIAGE and changed the carriage speed for the next test
run. We then repeated the procedure. Tests were
conducted in both east and west tow directions at
tow speeds up to 200 cm/s. We also conducted low
speed threshold tests at 2.5 cm/s. Repeatability
runs were made in each tow direc tion. Tests were
WATER
CHANNEL repeated with the beam 1&2 pair aligned into the
tow and at a 45-degree angle to the tow.
Figure i. Sketch of DTRC Tow carriage facility. snowing two
beams from a PADS unit. RD Instruments personnel participated in a second
series of tests conducted in February of 1988.
A Compaq 286 personnel computer recorded data and Another 1200 kHz unit, serial number 229, was used
controlled the DCP. Software was developed to for these tests. The test procedures were further
communicate with the DCP to set operating para- optimized to get bottom track data and include
meters and control sampling. DCP data and tow percent good and status in the data stream. We
carriage speed were recorded on the 20 Mbyte hard changed the ensemble size to 100 pings and the time
disk in ASCII format. Software then unpacked the between pings to 0.11 seconds, or 9 pings per
data, applied algorithms to transform data into second. Four profile pings were obtained for every
engineering units, and set up specified files for bottom track ping.
analysis. After each test run, data files were

347

The next series of tests were conducted in April of
1988. We used a tow basin at DTRC that provides an
additional 183 meters of towing length. Once
again, RD Instruments personnel participated.
Seven instruments were calibrated during a five day
period. We fixed the transducers at a 45-degree
tow angle for all the tests. This procedure saved
time and allowed calibration of both beam pairs.

3. LAKE AND TANK CALIBRATION TESTS

Laboratory tests were performed on a 1200 kHz DCP,
serial number 0181, at the tank facilities of RD
Instruments in March of 1988. The tests were ........
designed to measure the effective beam angles of
the system. Beam misalignment will result in
velocity scale factor errors. The angle of peak
power of each beam was measured as well as an
integration of each beam through angles of 5
degrees (in 0.2 degree increments) on either side *A
of the peak power point. Finally, the included
angles between beams, 1&2 and 3&4 were measured.
This facility is a tank of fresh water that has a
transducer positioning (azimuth and elevation)
system at one end and a calibrated receiving/trans-
mitting transducer at the other. The transmit path
is 2.4 meters long. The measurement accuracy of f
this facility was determined to be +/-0.15 degrees
in measuring the absolute pointing angle of an RD Instruments pontoon boat for lake
individual beam. Measurement accuracy of included calibration with a DCP protruding
angles was determined to be +/-0.10 degrees. through the test well.

A navigation course on a San Diego lake was Calibration of a DCP at one velocity consists of
surveyed by RD Instruments in March of 1988. The two traverses of the course at a constant velocity
purpose of this course was to determine whether a on reciprocal headings. Water velocity in the lake
DCP could be calibrated using a moving barge varies from near zero to 2 cm/s, depending on depth
navigation technique. The lake is a fresh water and time of day. The reciprocal runs average out
reservoir that is about 1500 meters long by 500 this water movement. Serial number 0181, a 1200
meters at its widest point. Its depth varies to a kHz unit, was used to establish procedures and
maximum of 30 meters. Two courses were surveyed, determine the best operating parameters for the
one 600 meters long and the other 1025 meters long. lake. Initially there were differences in distance
Conventional surveying methods were used in laying traveled between bottom track and water track of up
out the course and existing landmarks (trees, to 0.5 percent. This was traced to a signal "skew"
telephone poles, etc.) were used to provide error in the system that affected only the water
starting and ending markers. Flags were installed track performance. The system's processing filter
at either end of the course as bearing aids. The was altering the center of power of the returned
length of each course was checked with theodolite/ Doppler backscattered signal. Signal processing
infrared distance measurement equipment (DME). The algorithms were modified and the result was an
theodolite was fixed on land and a cluster of agreement between bottom track and water track of
corner reflectors mounted on the barge. The within 0.15 percent. The modification is described
distance to each marker was determined by posi- in a pending paper authored by Terry Chereskin of
tioning the barge at the markers. Repeated Scripps Institute of Oceanography, Eric Firing of
measurements confirmed the length of each course to the University of Hawaii, and Joel Gast of RD
be within 1 meter of the surveyed values. Instruments.

RD Instruments bought a pontoon boat (the barge) 4. CALIBRATION RESULTS
and rigged it with the necessary power, computers,
and fixtures to accommodate a DCP (see photograph). The initial DTRC tests conducted in November of
Specialized software was written to integrate DCP 1987 focused on the performance of the system in
velocity, resolve it into earth coordinates and the test facility. The decay in signal echo
compute the distance between the starting and end amplitude was investigated through 128 bins to be
points. This distance is computed for each bin of sure that succeeding pings were not being con-
the water profile as well as for bottom track. taminated by reverberation echoes from the previous
ping. Figure 2 shows a profile of the first 14
bins as the signal bounces within the facility
basin. The first 4 bins yield clean data and the
fifth apparently is contaminated from side lobes.
The sixth bin shows a large increase in echo
amplitude caused by reflection of the beams

348

directly from the concrete bottom. The remaining kHz DCP's in November, 1987, and again in February,
bins of data are reflected from the bottom bounce 1988, showed first bin echo amplitude levels of
of the signal. Measured velocities in the first from 35 dB to 40 dB. These measurements are
four bins were relatively similar. Bin 5, however, referenced to the receiver noise level in the DCP
showed significantly lower velocities due to side processing bandwidth. The normal levels measured
lobe contamination. It was determined that we in field operations are in the range of 60 dB to 90
could ping at a rate of 9 times per second and dB. The DCP has a receiver that includes an AGC
receive clean profile data in the first 4 bins. function to keep the signal at a constant level.
0 The AGC is preset, at the beginning of a ping, to
an echo amplitude of 90 dB to initialize the
I receiver gain for a nominal echo amplitude level.
2 The AGC then adjusts throughout the ping as the
Side lobe signal falls off with increasing range (depth).
3 from bottc, The rate of change of AGC reflects the rate of
-j; 4 change of the echo. This is normally proportional
L to range squared in the first few bins. The AGC
R 5- rate was observed to be higher than expected in the
6 __99iiqti basin. This indicates that it was "recovering"
from the difference between the preset value and
7 the actual echo amplitude.
c
a We decided to increase the scattering level of the
Reflected Side lobe basin to closely approximate ocean scattering
Z 9 Signal f am wall conditions. This was accomplished by dispersing
io pulverized limestone in the water. The idea was
il that the limestone would disperse slowly over the
SURFACE basin but eventually would dissolve and not con-
12 - ---- ------------------------- ------------ taminate the facility. Fifty pounds of pulverized
limestone was scattered over 270 meters of basin
13 length during the initial tests. The echo ampli-
14 - tude, as measured by serial number 0181, increased
0 20 40 66 80 to nearly 60 dB. The velocity calibration results
with this unit improved dramatically. A comparison
Mean Backscatter (dB) of this data to that taken by RD Instruments is
covered in a later section.
Figure 2. DTRC DCP Tests - Nov.12-13, 1987. Backscatter
profile of 22 test runs. Additional units were tested using the limestone to
increase echo amplitudes. Serial number 0262, a
Performance anomalies were noticed in the initial 600 kHz system, improved also, but still had speed
November tests. In still water conditions without errors approaching 6 percent. This was an improve-
tow carriage movement the DCP indicated a velocity ment from the previous 60 percent error. The echo
of 4 to 5 cm/s. Movement of the tow carriage to a amplitude for this system was lower than that for
threshold velocity of 2.5 cm/s resulted in a the 1200 kHz systems (48 dB vs. 60 dB). The
reduction of indicated velocity tonear the actual limestone scatterers are apparently better seen at
velocity. Continued increase in speed resulted in higher frequencies. We do not know if the 600 kHz
the DCP speeds being lower than the actual tow system would improve at higher echo amplitudes.
carriage speed. The 300 kHz system did not respond to the limestone
and backscatter amplitudes remained low.
Tests conducted in February of 1988 confirmed the
results of the November tests. It also showed that ERROR VELOCITY
different 1200 kHz systems exhibited similar
performance characteristics. Beam differences and error velocity was explored as
an indicator of system performance. Beam dif-
In April of 1988 seven systems were tested in the ferences are computed by the addition of opposing
DTRC facility. Speed calibration data on the 12DO beam velocities. Ideally, for horizontal motion
kHz DCP's was worse than the previous 2 systems. and no DCP tilt, opposing beams see the same
Serial number 0181, the unit calibrated by RD velocity, opposite in sign. Therefore, in the tow
Instruments, performed in a similar manner in that basin beam differences should be near zero.
the measured speed was significantly lower than the Although the DCP was mechanically adjusted to
actual tow carriage speed. Attempts at calibrating within 1 degree of tilt and there was no vertical
the 600 kHz and 300 kHz units resulted in poor velocity, beam differences were significant. On
velocity performance in all bins. serial number 0181, average opposing beam dif-
ferences were from 4 to 6 cm/s. Expected beam
ECHO AMPLITUDE differences for a 1 degree tilt and 30 degree beam
angles are approximately 6 percent of the hori-
We speculated that the poor velocity performance zontal velocity component along the beam pair
might be due to abnormally low echo amplitude direction. Thus, because of the sensitivity to
levels. The basin water is very clear with few tilt, beam differences are not good indicators of
acoustic scatterers. Tests conducted on the 1200 system performance.

349

Error velocity, which is computed by subtracting TABLE 2
the opposing beam differences of the orthogonal
beam pairs, should be a much better velocity Barge Velocity
quality indicator. For symmetrical beams, error Bin 80 cm/s 120 cm/s 200 cm/s
velocity should always be zero, independent of
tilt, as long as the water flow across the four Average Error in cm/s
beams is horizontally homogeneous. Observed error
velocities when forward velocity errors were small 1 -0.5 -0.1
ranged from -2.5 cm/s to +2.5 cm/s, and were not 2 0.0 -0.2 -o.6
correlated with speed. When forward velocity 3 -0.2 0.0 -0.4
errors were larger than 5 cm/s, error velocity was 4 -0.1 0.1 -0.5
generally larger, Thus, error velocity showed 5 -0.4 -0.1 -0.4
promise as an indicator of the errors. However, 6 -0.2 -0.2 -0.2
more analysis is required to understand and 7 -0.2 -0.1 -0.4
calibrate "error velocity" as an indicator of the 8 0.0 0.0 -0.4
actual velocity error. 9 -0.1 0.1 -0.1
10 -0.1 0.0 0.0
PERCENT GOOD Bottom track -0.2 -0.2 -0.3

The percent good criteria was not always a good Errors indicated are in cm/s for serial
indicator of the quality of the data. It would number 0181, 1200 kHz unit, on calibrated
generally flag problems in bin 1 but would not lake course. Each data point consists of
indicate bad data in bins 3 or 4 despite low signal averaged reciprocal runs on the course to
and large speed errors. average out water movement.

SPEED These data are plotted along with the DTRC data in
Figure 4. Figure 5 shows serial number 0181
Calibrations of both bottom track and water track results with and without limestone.
in 3 bins were performed on all systems. Figure 3
is a plot of the residual errors in the measurement
of speed from serial number 0181 and 0268, 1200 kHz
systems in bins 2, 3, 4 and bottom track. Table 1 0
summarizes the error statistics from the plot data.
This data is for the condition where echo amplitude 4 ---% -------- ----------------- -----------------
had been increased with limestone. 0 8 0 0
TABLE 1 Z +

ERROR WATER TRACK BOTTOM TRACK L +
STATISTIC BINS 2,3,4 OL + +
L
cm/s cm/s W F +
AVERAGE -1.1 -0.2
+
STD. DEV. 1.0 0.6 CL + +
MAXIMUM 0.5 0.5 -2-
+ +
MINIMUM -3.8 -1.7 L)
+

For comparison purposes an EG&G VMCM was calibrated +
at the DTRC facility acquiring data simultaneously
with a DCP. The results of calibrations on all 4 -3-
rotors at speeds of from 20 to 200 cm/s revealed an +
average error of -3.3 cm/s and a standard deviation DCP minus CARRIAGE SPEED DATAI
of 2.2 cm/s.

Serial number 0181 was tested by RD Instruments and -4-
by NOAA to determine the feasibility of calibrating 0 50 100 150 200 250 300
an instrument in a lake environment. The instru-
ment was set up the same in both the lake tests and Tow Carriage Speed (cm/s)
the DTRC tests of April. The orientation of the
beams was, however, different. For the lake tests, + BINS 2-4
beam 3 was oriented forward while the beams were 0 Bottom Track
oriented at 45 degrees to the direction of motion ---------- Zero Error
for the DTRC tests. The lake results on this unit
are summarized in Table 2. This data was taken Figure 3. DTRC 1200 KHz calibration errors. Data is from
with beam 3 forward. two systems were limestone was used to keep the echo amp-
litude above 40 dB. Zero tow speed data is not displayed.
Data points represent and average approximately 10 ensembles.

350

0.2- Several bottom track runs indicated a scale factor
error of approximately -0.16 percent. The DTRC
data indicated a bottom track scale factor error
0.0 ------------ ----------------------------------------- of approximately -0.21 percent. We expected a
larger scale factor at DTRC because of the dif-
+ ferent beam orientation (see Velocity Scale Factor
below). The average of the water track errors in
-0.2- a 0 Table 2 is -0.18 cm/s. The average of the water
track errors measured at DTRC was -0.66 cm/s for
five tests ranging in velocity from 1.5 cm/s to 102
r
-0.4- cm/s.
cc
0
a: The two 600 kHz systems exhibited larger errors
-0.6- with a -4.9 emls worst case bottom track error and
-10.7 cm/s worst case water track error. The echo
amplitude for these systems with limestone was less
than 50 dB. It is not known whether the accuracy
-0. a- K- would improve if the echo amplitude were increased.
No attempt was made to calibrate a 600 kHz unit on
the lake. The significant result is that all
-1.0 systems display a negative error at all flow speeds
6 36 e6 96 120 150 180 210 240 and a positive zero offset.

SPEED (cm/s) VELOCITY SCALE FACTOR

+ Lake Water Track The DCP error consists of two components: scale
0 Lake Bottom Track factor and bias. The scale factor is the slope of
OTRC Water Track the least-squares curve fit to the data. Bias is
A OTRC Bottom Track the intercept of the best straight line to the
---------- Zero Error Y-axis. The primary cause of velocity scale factor
errors is believed to be due to the effective beam
pointing angles. The beam pointing angles of
Figure 4. Comparison of lake and DTRC test results on serial number 0181 were measured by integrating
serial number 0181. 1200 KHz system. Lake data is the each beam through +/-5 degrees on either side of
average of 10 bins; OTRC is the average of 3 bins. the beam peak. The beam width is about 2 degrees.
The results of this test show an effective beam
4 pointing angle that was inward (toward the group
center) of the peak by an average of 0.02 degrees
3- for the four beams. This is small compared to the
angle measurement capability of the facility
2- (+/-0.15 degrees), but it does indicate that errors
due to non-symmetrical beams is probably small
(<O.l percent).

C_ The included angles, beams 1 to 2 and beams 3 to 4
0 0- ----- ---------- ----------- ----------
C_ were measured. The difference between these
C_
LU angles was measured to be 0.31 degrees with the
CL -I-
U beam 1-2 pair having the larger angle. Although
M the absolute measurement accuracy is believed to be
-2- +/-0.10 degrees, the difference measurement is
0 expected to be more accurate. As a result, we
-3 would expect to see a scale factor change of 0.47
0 percent between beam pair 1-2 in the direction of
motion and pair 3-4 in the direction of motion.
0 50 160 This was tested by taking the unit to the lake
course and performing two calibrations, rotating
Tow Carriage Speed (cm/s) the transducer 90 degrees between the two. The
difference in calibrations was measured to be 0.56
* BIN 2 + Limestone added' percent using bottom track. As expected, the beam
* BIN 3 X Limestone added 1-2 pair had the higher scale factor.
11 BIN 4 Limestone added
0 Bottom Track 0 Limestone added In the tests conducted at DTRC, the beams were
---------- ZERO ERROR oriented 45 degrees to the direction of motion.
The larger included angle between beams 1 and 2
should result in a larger scale factor of
Figure 5. OTRC calibration of RD S/N Oi9l: with and with- approximately 0.28 percent when compared to the
out limestone. Errors shown are the ensemble averages for lake navigation data that was run with beams 1 and
each speed minus the tow carriage speed. 2 at 90 degrees to the direction of motion.

351

UNRESOLVED DCP PROBLEMS overall calibration. Accuracy of facilities must
be improved. Work on DCP's with frequencies other
There are several problems that remain unresolved. than 1200 kHz must be done. Finally, when scale
First, each DCP tested exhibited positive biases at factors are calibrated in controlled circumstances,
zero velocity. This bias disappeared with the the effects of different ocean environments must be
slightest velocity (2.5 cm/s). Laboratory and lake determined through analysis, simulation, and
testing by the manufacturer didn't duplicate this in-water testing.
problem. It may be related to the low echo
amplitude and that is being investigated at the
present time. 6. REFERENCES

Second, the errors experienced when the echo 1. Mero, T. N., G. F. Appell, and D. L. Porter,
amplitude was low is not well understood. It is "Sea-Truth Experiments on Acoustic Doppler
reasonable to expect that the low levels we Current Profiling Systems," Proceedings of
experienced (30 dB below normal) might cause the Oceans 183, Marine Technology Society, San
DCP some difficulty. However, we expected the Francisco, California, August 1983.
percent good output to eliminate the data that was
in significant error. We found this to be the case 2. Appell, G. F., "A Real-Time Current Measure-
for Bin 1 where the data good output flagged ment System," Sea Technology, February 1984,
problems. Bins 2 through 4, however, were not Volume 25, Number 2.
flagged as bad despite large speed errors.
3. Magnell, B. A., "Fall 1983 Acoustic Doppler
CALIBRATION PROBLEMS Profile Measurements and Sea Truth Inter-
comparison Experiment at Ambrose Light, New
Scale factor and bias differences between the lake York," Technical Report to NOAA, EG&G Washing-
and DTRC need to be resolved. The different beam ton Analytical Services Center, Inc., 1984.
orientations between the two tests on serial number
0181 do not account for all the differences. The 4. Magnell, B. A., "Delaware Bay Intercomparison
accuracies of the calibration methods must be Experiment, Fall 1984,"Technical Report to
improved in order to allow a calibration to the DCP NOAA, EG&G Washington Analytical Services
specification of 0.2 percent +/-0.5 cm1s. The lake Center, Inc., 1985.
course is believed to be accurate within 1 meter
(0.16 percent). There is an additional error of 5. Appell, G. F., T. N. Mero, J. J. Sprenke, and
positioning on the starting and ending marks for D. R. Schmidt, "An Intercomparison of Two
the navigation technique. Repeated runs suggest Acoustic Doppler Current Profilers," Pro-
that this is also on the order of 1 meter. The ceedings of Oceans 185 Marine Technology
velocity uncertainty at DTRC is believed to be on Society, San Diego, California, November 1988.
the order of 0.1 cm/s. The DTRC error is a per
sample error, the lake error is for the entire 6. Woodward, W. E. and G. F. Appell, "Current
distance travelled. Error velocity data needs to Velocity Measurements Using Acoustic Doppler
be re-examined and correlated with beam angle dif- Backscatter: A Review", IEEE Journal of
ferences. Other frequency DCP's need to be Oceanic Engineering, January 1986, Vol.OE-11,
compared between the two facilities. Future No.1,PP-3-6.
testing, both by RD Instruments and by NOAA at DTRC
will be geared to resolving these problems. 7. Woodward, W. E., D. L. Porter, G. F. Appell,
Proceedings of the Acoustic Doppler Current
5. CONCLUSIONS Profiling Symposium, U.S. Department of
Commerce, NOAA Report, 1984
Significant progress has been made toward per-
forming velocity calibration of DCP's to precisions 8. Appell, G. F., T. N. Mero, R. G. Williams,
approaching instrument specifications. Tow tank W. E. Woodward, "Remote Acoustic Doppler
facilities like DTRC can be used in those cases Sensing: Its Application To Environmental
where the echo amplitude can simulate the ocean Measurements," Proceedings of ASME's Current
environment. The tow tank can also be used to Practices and New Technology
verify other procedures such as calibration on lake in Ocean Engineering Symposium, New Orleans,
facilities. The feasibility of performing calibra- February 23-28, 1986.
tions on a lake having small currents has been
established. It has been shown that such a 9. Chereskin, T., E. Firing, J. Gast, pending
facility has the capability of detecting small paper dealing with the identification and
changes in transducer beam pointing angles. screening of filter skew and noise bias
Laboratory measurements on transducer beam para- errors in acoustic Doppler current profiler
meters can provide meaningful calibration data. measurements.
This data can then be used to provide scale factor
corrections to the DCP. Bottom track can be used 10, Firing, E. (1988) WOCE Tropical Workshop
to calibrate velocity scale factors. This has the Report, in press.
practical advantage of not having water current
variability. It also has inherently lower
variances that result in more accurate measure-
ments. There is, however, much to do to achieve

352

DEVELOPMENT OF A SHIPBOARD ACOUSTIC
DOPPLER CURRENT PROFILER

1 1 2
Yoshifumi Kuroda, Gentaro Kai and Kiyonori Okuno
1 2
Japan Marine Science and Technology Center Japan Radio Co., Ltd. Laboratory
2-15,Natsushima,Yokosuka,237,Japan 1-1,Shinorenjaku,5, Mitaka, Tokyo, Japan

1 ABSTRACT transmitter-receiver unit was carried out, then comparison
between the ADCP and RCM was performed. These results
A shipboard long range acoustic Doppler current profiler of the open sea tests are discussed in Section 5.
(ADCP) was developed, which can measure vertical pro-
files of ocean currents and acoustic backscattering strength 3 SYSTEMDESCRIPTION
in 32 layers up to a depth of 400m, and track the sea bot-
tom up to a depth of 1000m. In the development of this The ADCP's principles is well known [2],[3]. The acoustic
ADCP, we aimed to develop an acoustic transducer mod- pulses are scattered by particles drifting in the water, and
ule(70kHz, phased array type) for the measurement of the the Doppler shift frequency which is proportional to the
open ocean currents, and introduce all advanced signal pro- relative velocity between the ship and the water occurs in
cessing method based on the auto-correlation function to a received signal. Conversely speaking, the Doppler shift
derive the mean Doppler shift frequency from the received frequency call be converted into the relative velocity. The
acoustic signal. And then, in order to confirm the per- depth of each measuring layer is determined by the delay
formance of the ADCP, the transducer and transmitter- time between pulse transmission and pulse receiving.
receiver unit were tested in the open sea. Furthermore,
the comparison between the ADCP and Aanderaa current TABLE 1 Fundamental specifications of the developed
meters (RCM) was carried out in the Kuroshio current on system.
the east side of the lzu Ridge. These results showed that measurable items current direction, velocity
the system has capability of measuring the ocean current scattering strength
velocity and the scattering strength up to a depth of 400m. range of detectable current 400 meters
measurable layers number 32 layers
range of detectable sea floor 1000 meters
2 INTRODUCTION velocity resolution 1 cm/sec
acoustic frequency 70 kHz
Shipboard acoustic Doppler current profilers (ADCP) which beam configuration 4 beams (JANUS type)
beam angle to the vertical � 300
can measure remotely vertical current velocity profiles un- beam width 50
der a ship have been developed recently in several countries. pulse width 7.5-30 msec
In Japan, conventional ADCPs are used as ship navigation pulse intermittence 1-5 sec
and fishing instruments which measure the currents up to measuring layer thickness 5-50 meters
one or two hundred meters in depth at several measuring averaging time 30-600 sec
layers. And a multi-layer ADCP for the coastal region has
been developed and used to measure a tidal current around
a narrow strait[l]. And as a next step, a development of an The specifications of the ADCP are shown in TABLE 1, and
ADCP for the open sea was required to reveal the current the block diagram of the ADCP system is shown in Fig.l.
structure of the Kuroshio and other deeper currents. The ADCP system consists of six units. The transducer
unit has two modules which transmit 70kHz four acous-
The development of a long range ADCP for the open ocean tic beams in a JANUS configuration. The ship navigation
is presented in this paper. The ADCP system is described unit has LORAN-C and GPS receivers which input the ship
in Section 3. The developments of the 70kHz transducer speed to the earth into the signal processor unit. This ship
module of phased array type and an advanced signal pro- speed enables the ADCP to calculate the current velocity
cessing method based on the auto-correlation are presented to the earth when the ADCP fails to track the sea bottom.
in Section 4. The performance test of the transducer and The signal processor unit derives the current velocity and

CH2585-8/88/0000- 353 $1 @1988 IEEE

SHIP NAVIGATION UNIT DISPLAY UNIT
- - - - - - - - - - - - - - - - -
GYRO GPS 147
CRT
COMPA S LORAN-C

MOUSE
L J

MICRO- . . . . . .
COMPUTER KEYBOARD
SIGNA PROCE T

-7
FEARTRI-67GE
RECEIVER
_L_M_T __j
...... [T@RANSMIT
TER
TRANSDUCER UNIT L - - - - - - - - - J L J
TRANSMITTER-RECEIVER UNIT RECORDING UNIT

Fig.1 Block diagram of the ADCP system.

the scattering strength profiles from the received acoustic
backscattering signals. The display unit consists of a iiii-
cro computer, Which is used to input measuring parameters
with a, niouge tool and display measured data on a CRT in
real time. The recording unit has a cartridge magnetic tape Fig-2 Transducer unit consisting of two modules.
recorder(1/4 inch,38. Mbytes).
4 SYSTEM DEVELOPMENT 4.2 SIGNAL PROCESSING
Figure 4 shows a, raw Doppler shift signal of the Fore/Aft
4.1 TRANSDUCER MODULE beanis in the open sea getting directly from the receiver
A transducer module of phased atray type which can trans- unit of the ADCP. And the auto-correlation function cal-
5
init a, pair of acoustic bearns of 70kHz was developed. The
ADCP uses two transducer modules to transmit two pairs
El A1\q v
of beanis of For'e/Aft and Port/ Starboard. The transducer _\j _V
'unit Consisting of two modules is shown in Fig.2. The >
square shaped transducer module consists of 800 pieces U@ G24
104.) C TIM11,(-)
25 x 32 elements) of transducer elementg, whose dimension (a) rmpw j W
is 40 x 33 x 14 cin. The beani pattern of a module is shown
in Fig.3. The beani axigles are � 301 to the vertical and the
beani width is 5o. And the side lobes are depressed under
15dB to the two rilain lobes.

...02 lool o-6 ooos D.,
TIME LAG

MH 102 iz6
[42.
(b) (637.) E Don i

o.oz Qoo4 0.o6 okos

TIME LAG(-)
Fig.4 Doppler shift signal in the open sea, and the auto@
Fig.3 Acoustic beam pattern of the phased array type correlation function. (a) signal from 104-156in depth, (b)
transducer module. signal from 637-702m depth.

354

QUADRATURE I A/D COMPLEX MAIN CALCULATOR
PHASE CONVERTER AUTO-CORRELATOR ,Re(R(AT))
DETECTER
(R(T))tan-1(:I@-/R],-
SIG. x LPF A/D DELAY f x2 MEAN
CONJUGATE FREQ.
MULTIPLIER +
x LPF A/D j X2 JR(
REF. INTEGRATOR
900 f AT 2
IR 0)
MULTIPLIER VARIANCE
x2 JR(O)j

Fig.5 Flow diagram of the signal processing based oil the correlation is shown in Fig.5. The resolution of the ad-
auto-correlation function. valiced signal processing was compared with the zero cross
culated Loin the signal is shown together, At the measuring iiiethod. The results are shown in Fig.G. The zero cross
time the ship was kept straight at a speed of about 6knots. method is successful in deriving the Doppler mean frequency
Generally the Doppler shift frequency of 111z corresponds when the SIN is 6 dB, but it fails ill that when the SIN
to 2.14ciii/sec. The frequency of the signal ill Fig.4a is decreased. Oil the other hand, the advanced signal process-
about 140Hz, therefore, the relative current velocity is esti- ing is successful up to the value of -6dB, furthermore, the
inated as about 300ciii/sec. That seems to be a reasonable variance of the derived values at S/N=GdB is smaller than
value. In that case, the signal scattered at the shallower that by the zero cross method. Therefore, it call be said
ra.lige (104-156111 depth), and the SIN of the signal seems that the method used in this ADCP shows all filiprovernelit
to be very large. Therefore, it is easy to decide the Doppler of SIN by 12dB.
frequency from the received signal. On the other hand the
depth becanie deeper, the signal became more complicated 5 OPEN SEA TEST
as shown in Fig.4b (637-702m depth). Thus it is difficult
to estimate the velocity from these low SIN signal. 5.1 TEST OF TRANSDUCER AND
TRANSMITTER-RECEIVER UNIT
In order to derive the mean frequency from these signals, all
advanced processing method based oil the auto-correlation 139E 140E
function [4] was introduced instead of the zero cross method
[email protected] Bay
used ill [1]. The signal processing flow using the auto-
Soso Pen.,

35N Sagami Bay
0.75

(a)

0.25
0,hima Is.

0.00.
,a. I.50 200 'H21 250 Sao
FREDUENCY Mooring Site

2.00.
A
zero-CrOss
0.75. auto- correlation

Miyake I
(b)

0.25 14N Traverse section

ikura Is.
0.001-
i0o 16 -00 no Sao
FMMNC@ (HZI Fig.7 Test sea, area.
Fig.6 Comparison between the resolution by the auto- The hatched square shows'the area where the transducer
correlation method and that by the zero-cross iliethod. The system tested, the triangle shows the mooring site of the
model signals consisted of 20OHz Doppler shift frequency R-CM, and the thick solid line shows traverse section of the
with white noise, (a) S/N=6dB, (b) S/N=-6dB. Kuroshio current.

355

The transducer unit and trallsinitter-receiver unit were tested 5.2 COMPARISON BETWEEN ADCP
ill 1984 ill the Sagami Bay near Tokyo) (Fig.7). The result is AND RCM
shown ill Fig.8. There was a large value in the receiver out-
put near the depth of 1000m, which was thought to show The deck units and transducer unit of the ADCP ill the test
a, strong sea bottom reflection. It shows that the ADCP are shown ill Fig.9,10 and 11. Before the comparison, the
has, the ability to track the sea bottom up to 1000in or dependence of the standard deviation of the Doppler shift
more. Furthermore, the receiver output level of -15dB at frequency (i.e. breadth of the spectrum of the received sig-
the water depth of 500m was sufficiently large to derive the lial) on the pulse width and measuring layer thickness was
Doppler frequency, because it was much greater than the measured using the ADCP ill the open sea, under the vari-
background noise level of -25dB. Thus, it was confirmed ous values of the two parameters. Figure 12 shows that the
that the transducer systern has stifficient capability to de- ,standard deviation was increased when the pulse width or
tect the water backscattering signals. the measuring layer thickness was decreased. Especially,
the standard deviation was rapidly increased where the
ineasuring layer thickness was 51n, and the velocity profiles
(not became complicated than the other. Conse-
Pulse 30rnsec
quently, the measuring layer thickness shorter than 10ill is
M Freq. 70KHz
not appropriate for this ADCP.

In order to confirm the performance of the ADCP, the corn-
parison between the shipboard ADCP and the moored Aan-
-2 sea bottom reflect.ion deraa current ineters (RCXI) was carried out ill 1987 at Izu
Ridge. The mooring site is shown in Fig.7, at which the
-1 5dB depth was 800111 and the strong current of the Kuroshio
was expected. The mooring consisted of five current me-
-2'5dB
backgroun noise level ters at 30, 140, 250, 375 and 500m depth. The experimental
conditions of the four test cases are shown in Table 2. In
R5) ( 520) (U ('e5) (1iZ the Case 1,2 and 3, the ship ran by the surface buoy of the
Depth (m) ' mooring within a distance of 50-100in at a speed of 5knots
Fig.8 Received signal level ill the open sea.

17777"M

Fig. 11 Fish shaped housing containing
two transducer modules, which was Exed
oil the port side of a ship at the same
5d @@

Fig-9 Ship navigation, signal Fig.10 Display and receiver unit. depth as the ship bottom.
processing and transmitter-
receiver unit.

356

and measured the current profiles. In the Case 4 the ship 0 RCM - . ROCP - x RCM - o ADCP - x
IiiiiiIIIIIIII Jill[ 11 11
was drifted near the site. In these cases, the reference ship
x C, eb
x x
x x
speed was inputted from LORAN-C. The sampling time of x x
x x
the RCM was 1 minute, and that of the ADCP was also 1 125 xx x
ininute with the pulse intermittence of 2,5sec. The results .0 10
x x
of Case 2 and Case 3 are shown in Fig.13. The values of '0 xx Ix
x x
x
the current velocity obtained by the ADCP and the RCM 0 250 -0 x x
r@ x x
at shallower depth than 140m were greater than 60cm/sec X, x
x x
and coincided with each other. And the both direction also x x xx 20
c- x x
coincided with each other. Thus, the coincidence of the two LU x x
375 -b, y
X0
vertical profiles is an evidence that the ADCP system sat- x x
x x
x x
isfies the initial ahn as shown in Table 1. The results of the x x
x x
x x 30
four cases are drawn together in FigA4. That figure shows 500 , I'-X' ''-, I...I I I 1 14,
the some variance around the solid line. It is difficult to ex- 0so 100 0 180 360
plain the cause clearly, but we can consider some reasons as VELOCITY CM/e DIRECTION deg
follows. The accuracy of the reference ship speed obtained
by the LORAN-C which was worse than the ADCP's; the
space averaged area by the ADCP was wider than that by 0 CM q0CP x RCM HDCP x
the R-CM as the measured value by ADCP was averaged rTT7-r-r
over the area among the four beams and the area along the xx x
x
x x
ship track during the mean time, on the other hand that 175 xx0 C@
ky the RCM was only averaged around the mooring; or the x
x x 10
x x
direction sensor of the RCM might respond slower than the L x 0 X@
M x x
ADCP to weak currents. Therefore the variance did not -0 x )e x
0S 350 x x
show necessarily the accuracy of the ADCP. Lastly the tra- 0 xx
x x
verse of the Kuroshio current using ADCP was successfully x x 20
-x
measured and depicted in Fig.15. x xx
U-1 525 xb x
x x
X, x
x
xx x 30
30 700 [....1'1114,11111,11 1 ... xxlx I I 1 11 ....IA
0 0-50 100 0 180 360
pulse width (@c) VELOCITY CM/8 DIRECTION deg
E3
0 20 - 7.5
0 015 a
A
A 22.5 Fig. 13 Comparison between the velocity profiles by the
A 0
13 30 ADCP and the RCM. (a) Case 2, (b) Case 3
10 -
45

LO
0 1 1 (a) so velocity (b) direction
0 10 20 30
layer thickness (m)
60 - 70 a
Fig.12 Standard deviation of received Doppler shift fre-

so
quency versus rneasuring layer thickness for various pulse U 40 -
widths. d4C

20 go
TABLE 2 . Parameters of the four test cases for conipari-
son between the ADCP and the RCM. 20 40 60 so LO 100 270 360
RCM RCM
parameter Casel Case2 Case3 Case4
ship speed (knot) 5 5 5 drift Fig.14 Scatter plots oftheADCP and the RCM data,. (a)
pulse width (rnsec) 15 22.5 30 30 velocity (b) direction, solid circles show larger value than
layer thickness (m) 10 15 20 15 15clillsee obtained by the R.CM and triangles show smaller
depth range (M) 15-335 22-502 30-670 32-512 than that.

357

40
80
120
160
200
240
280 N
E- 320 Mi
a4 E
360
400
50CM S
440
m I I I I I /A I I.
480

330055 46 N 34 0 22 96 N
139 42 , 1 E Fig.15 Traverse section-of the Kuroshio current over the 1350 39 18 E
Izu Ridge, the, solid lilies represent velocity vector. The pa-
railieters were used the same with Case3 in Table2.
6 CONCLUSIONS 7 ACKNOWLEDGMENTS

A long range ADCP has been developed, which can measure The authors wish to thank to Mr. M.Gondou, Mr. Y.Asada
ocean current profiles and acoustic backscattering strength and Mr.S.Maisuzawa for their great contribution to com-
in 32 layers tip to a depth of 400111. In the development the plete the work. Mr. T.Tsuchya and Mr. Y.Ainitani are
followings were performed. thanked for their assistance testing the transducer inodule,
(1) The 7OkHz transducer module of phased array type has and Mr. M.Katou is also thanked for his computer draw-
been developed for measuring the open sea current. ing. And the members of the system design committee
(2) The advanced signal processing method based oil the for this project are acknowledged for their valuable discus-
axito-correlation function was introduced. The method iiii- sions and advises. This work was supported by the Science
proved the capability to derive the mean Doppler shift fre- aild Technology. Agency in Japan under the special coor-
quency. It was confirmed the S/N become better l2dB than dination funds for "Research oil the development of new
the zero cross method which we had been used. oceanographic research systems"
(3) The capability of the transducer and transinitter-receiver
unit were tested in the open sea. The received signal level
at the 500m depth was lOdB greater than the background
noise level. 8 REFERENCE
(4) The comparison between the ADCP and the RCM was
carried out in the Kuroshio current. The vertical current (1) K.Okiilio,Y.Tstiji,S.Hisaiiioto,M.Okino and T.Einura,
profiles by the ADCP coincided with that by the RCM. And 1983, Three-Dilliensional Current and Scattering Strength
the traverse measurement of the Kuroshio was performed Distribution Mapping System, Proceedings of Oceans'83,
successfully. 301-305.
(2) F.D.Rowe and J.W.Young, 1979, An Ocean Current
From these results we consider that the ADCP has capabil- Profiler Using Doppler Sonar, Proceedings of Oceans'79,
ity of measuring the open sea current. And the ADCP will 292-297.
be made use of by JAMSTEC as follows, the R/V NAT- (3) T.M.Joyce,D.S.Bitteriiiaii,Jr. and K.E.Prada, 1982,
SUSHIMA will be equipped with the ADCP in Dec.1988, Shipboard Acoustic Profiling of Upper Ocean Curreiits,Deep-
and the R/V YOKOSUKA which is a new mother ship of Sea Research, 29, 903-913.
a submersible research vessel for deep ocean has launched (4) G.Kai,Y.Kuroda,S.Matsuzawa and K. Okuno, 1987,
July. 1988 and been equipped with the advanced ADCP. A Ship-niounted Acoustic Doppler Current Profiler, Pro-
The ADCPs are hoped to make clear the currents around ceedings of Toyohashi International Conference on Ultra-
Japan and in the Pacific., sonic Technology, 195-206.

358

THE ACOUSTIC DOPPLER CURRENT PROFILING SYSTEM AT AOML

Doug Wilson, David Bitterman, and Carol Roffer
National oceanic and Atmospheric Administration (NOAA)
Atlantic Oceanographic and Meteorological Laboratory (AOML)
4301 Rickenbacker Causeway, Miami, FL 33149

ABSTRACT doppler shift in frequency of the
backscattered acoustic energy to estimate
The design and performance of the data water speed relative to the ship in each of
acquisition and logging system of the 63 depth bins. The Ametek DCP 4400 115 kHz
Acoustic Doppler Current Profiler on the ADCP installation on the Researcher (now
NOAA Ship Malcom Baldrige are described. the Malcom Baldrige) is located in a sea
The accuracy of absolute current chest on a flat section of the hull, near
measurements is limited by our ability to the keel about 40% aft. This installation
determine ship movement and orientation. has shown little reduction in signal
Methods of processing TRANSIT satellite and quality at high speeds or in rough seas.
Global Positioning System navigational data The original DEC PDP 11/23 instrument
are evaluated. methods of precisely controller/acquisition system has since
determining the ADCP transducer alignment been replaced by a DEC microVAX II. A
relative to the ship are described and schematic block diagram of the system is
results shown. shown in figure 1.
Operator interaction with the system is
INTRODUCTION through the CRT, which also displays
operating parameters, single ping data, or
AI paper in OCEANS 183 described.the averaged profiles. The system is normally
initial installation and early performance run with a pulse length of 9.6 msec ( a
assessments of an Ametek-Straza DCP 4400 nominal depth bin width of 6.4 meters) and
3-beam acoustic doppler current profiler a ping rate of 1.76 sec. The major
(ADCP) on the NOAA Ship Researcher Since departure from suggested system default
that time the profiler has been run roughly parameters (based on early sea trials) is
180 sea days per year, primarily in@ the that the profile detection signal to noise
equatorial Pacific ocean and the Caribbean ratio is kept low (2 db), so that very few
Sea. This paper will focus on techniques profiles are rejected by the system.
developed to convert the basic ADCP product Originally, doppler data from each ping
(very many noisy single-ping profiles of were stored. Analysis of this data
water velocity relative to a moving ship) confirmed the manufacturer's error
into a useful oceanographic data set. ACIML
scientists and other oceanographers have
' t CRT 500 MEGABYTE 9TRACK
used these data for large scale Curren TERMINAL DISK
mapping, volume transport studies, DRIVE TAPE DRIVES
horizontal momentum and heat flux
estimates, and bulk studies. of eddy
2 DEC
viscosity coefficients (Morrison (1987) DECNET 0
88)3 MICROVAX 11 DECI
Wilson and Leetmaa (19 ; Leetmaa and CPU TO K UNIT
Wilson (1985 )4 ). The uses planned for the
data set determine the processing methods
CESIUM
to some extent; we generally require CLOCK AMETEK STRAZA TEMPERATURE
current estimates on moderately large DCP 4400 SENSOR
spatial scales, so that ADCP and ship PROCESSOR
navigation data can be averaged for 10 to SHIP'S
30 minutes and mapped to grids of 5 to 25 GYROCOMPASS
kilometers by 10'meters. AMPLIFIER (HEADING)
TT
ADCP INSTALLATION AND DATA ACQUISITION 3 BEAM
TRANSDUCER

Basic ADCP theory and operation are
SATNAV

GPS

I t

well described by Pinkel (1@80)5. The Figure 1. Schematic diagram of the ADCP
hull-mounted unit transmits a 115 kHz pulse controller acquisition system on the NOAA
downward and outward, and measures :the Ship Malcom Baldrige.

350, @ United States Government work not protected by copyright

estimates (ping-to-ping standard errors of between satellite fix locations. The
32 cm s-1 at 6.4 m bin width). Data from difference between locations is attributed
each ping is now rotated from the ship's to a mean current at the reference level
reference frame to a fixed geographic frame (figure 2). These currents are added to
using an instantaneous heading value from the shear profiles to give one minute
the ship's gyrocompass, and 36 pings (63 averaged absolute currents, which are dead
seconds) are averaged for storage to tape, reckoned again to get accurate profile
lowering the expected error for the positions.
ensemble mean to about 5 cm s-'. This approach works well but has
I Each output record contains the limitations. A bad satellite fix,
one-minute averaged ADCP data (north and especially if it closely follows another
east components for 63 bins and bottom fix, can produce large errors in reference
track), number of good samples per bin ' velocity. Satellite fixes are not used if
output time, parameter settings, average they have low or high elevations, require
ocean temperature, and all navigation excessive iterations for position
information received during the averaging determination, or occur within 20 minutes
interval. This can include up to 7 LORAN C of a previous good fix. Satellite fixes
or GPS fixes. An output record is 1152 producing 'unreasonable, reference level
bytes and up to five days may be stored on velocities are also deleted. The smaller
range of realistic velocities at 150 m
a 1200 foot, 1600 BPI, 9.track magnetic makes this a much easier field to edit than
tape. ship speed itself. Reference level
On the present system, the one minute velocities are spline fit or averaged in
,averaged data are also written to a disk space or time to produce a smooth field. A
every two hours for near-real time fore/aft transducer misalignment can cause
calculations of longer averages of absolute a cross-track bias in reference level
currents. With the ADCP controller / velocities and must be corrected before
acquisition program locked in memory, the final calculations are performed (see
microVAX can also process the ADCP data Calibration). Figure 3 shows an example of
from disk, log and plot data from the CTD reference level velocities from a cruise in
deck unit, and run other routine data the Eastern Pacific.
analysis and display programs. Satnav position accuracy is dependent
on a proper estimation of ship speed during
ADCP DATA PROCESSING a fix. The ADCP provides speed input
(relative to some subsurface reference
Principal ADCP data processing tasks level) to the satellite navigator for this
include further averaging to reduce the purpose. An alternative would be to store
error in shear profiles, and estimating actual satellite Doppler data and reprocess
ship's speed to correct shear profiles to it using ADCP measured ship speed.
absolute velocity. The latter is dependent The major drawback to using satellite
on the type of navigation data available. navigation only is that the horizontal
very little of our work takes place in resolution of the absolute velocity field
is limited by the distance between
water shallow enough to estimate ship's satellite fixes. while this can make
speed by ADCP bottom tracking. When bottom accurate surveys of high horizontal shear
tracking is possible, the system is run in regions (fronts, for instance) impossible,
alternate mode, where alternate pings it is adequate for larger scale current
sample bottom and water doppler shifts. mapping applications.
information from each TRANSIT satellite fix
is stored, as are the most recent GPS fix
and a mean LORAN-C position for each
one-minute ensemble. Additionally, all GPS
fixes (up to seven per ensemble) are stored
when available. When LORAN C data is ........... TRUE COURSE DEAD RECKONED COURSE
available, it may be substituted for GPS.
Since most of our work is done outside
LORAN C coverage areas, and its accuracy is DEAD
generally less than that of GPS, we will ................. RECKONED
discuss only techniques for estimating ship SATHAV POS
POS I
speed from GPS and satellite navigation. DISPLACEMENT
The primary source of ship navigation VECTOR
information is the TRANSIT (Satnav) MEAN REFERENCE ..................
satellite system. Satnav fixes are stored CURRENT VECTOR SATNAV
in the ADCP data record at their output POS2
time. Besides time and position, satellite
elevation and number of iterations required
to converge on a position are stored. One
minute averaged ADCP velocities relative to Figure 2. Difference between ADCP dead
some deep reference level (around 150 m) referenced and satnav positions is
are integrated to dead reckon ship position attributed to current at reference level.
DEAD
RECKONE1
POS

360

The ADCP System includes a Trimble averaged over each ADCP output interval;
4000A Global Positioning System (GPS) this speed is used for referencing when
locator as a navigation input. When three available.
or four satellites are in view, the system Preliminary, near real-time data
provides continuous position, time, and processing can be performed on the two-hour
,speed values. Presently this is 8 to 12 disk files at sea, although satnav speed
hours per day, but will ultimately be at determination can only be updated to the
all times. The unit is often augmented last satellite fix. Generally all data
with a cesium frequency standard to provide tapes from a cruise are consolidated into a
accurate time input when only two sin,(jle disk file. for final processing.
satellites are in sight. This can increase During this phase, ADCP output times (based
GPS coverage by 4 to 6 hours per day. on the controller's clock) are corrected to
GPS output to the ADCP system is every agree with GPS times. This is important
10 to 15 seconds, and the following are for dead reckoning between satellite fixes,
stored in the GPS data block on tape: and for referencing profiles when the ship
time, latitude, longitude, altitude, speed ' @is. maneuvering. After ship speeds are
heading, number of satellites, and a determined for each one minute ADCP
dilution of precision (DOP) value. Best average, they are used to reference the
estimates of position accuracy are 7 to 15 shear profiles - GPS or LORAN C take
meters (for optimum constellation precedence over satnav when available. The
configuration and satellite elevations, and absolute velocities are further averaged
accurate altitude and time values in 2 and over 10 to 30 minutes, depending on the
3 satellite modes); accuracy is reflected ultimate spatial resolution desired. These
in the DOP value. Dockside tests have -are the 'basic units, of processed ADCP
shown that 20 meters is a reasonable RMS data. They are often mapped onto regular
value. Mean ship speed estimates based on horizontal / vertical grids for display
positions 10 minutes apart would have an (see figure 5).
estimated error of 4 cm s This accuracy 220-
can be approached by editing position fixes
GPS POSITION
during processing to eliminate very bad
A
positions, and by estimating the mean 10 minute differen
position of a 1 minute ADCP record by 200 . .......... 2 minute
differen
ce
fitting a curve to all fixes obtained
during that interval.
Ship speed and heading output have .01
180-
knot and 0.1 degree resolution. Velocity
accuracy (at zero acceleration) is quoted
at 10 cm s rms error. Thus 2 minute
averaging (at 5 fixes / minute) lowers the 160
expected speed error to 3 cm s-1, better
than the speed based on position 220-
differentiation over the same interval. GPS VELOCITY
For the standard secondary averages of ten 10 minute mean
or fifteen minutes, the expected errorsare 200 . ..........2minute mean
about the same, but speed averages are less
affected by outliers (figure, 4). In
practice, the instantaneous velocities are
edited based on acceleration and DOP, and 180-

10 N- . . . . . . . 160 - F'. ...... ......... ......... ............
220-
<
ADCP BOTTOM TRACK

0 10 minute mean
200 . ..........2minute mean
100 CM/S E
10 S 180-
140 W 120 W 100 W 80 W
West Longitude

160 ........ ........ ...... ..............
Figure 3. Current velocities at 50 meters 0 10 20 30 40 50 60
determined by ADCP dead reckoning between Time in Minutes
satellite fixes, October - November 1987.
Vectors begin at midpoints between fix Figure 4. various estimates of ship speed
u
minute @MeCn
@2. "lute Mea n.

positions. Note presence of tropical based on GPS information and ADCP bottom
instability waves in zonal sections. tracking.

361

al ADCP'Eastward Velocity ADCP vs Pegasus
0- 3 N October 1987 Least Square Fit 0-200 rn
0

50-
100 DEC 1983 U

100 DEC 1983 V

LUU MAR 1986 V
150

300
200
140 130 120 110
400
b) ADCP Northward Velocity 0 25 50 75 100
0- 3 N October 1987 Standard Error cm/s

Figure 6. Standard error of difference
50- between Pegasus and ADCP shear profiles,
aligned for best fit over the upper 200
100, meters. From the Florida Straits.
0

0
150- During studies of the Equatorial
Undercurrent in the eastern Pacific,
W
comparisons were made between ADCP
7
200 .0( - measurements and moored current meters.
140 130 120 110 The ADCP value was the mean of all'data
within + 7.5 miles along the equator of the
mooring location; the current meter value
was a 24-hour average. The mean (standard
Figure 5. (a) Zonal and (b) meridional deviation) of the velocity differences over
velocity fields at 30 N, October 1987. All 69 comparisons in the upper 160 meters were
profiles were referenced using TRANSIT 3.3 (14.1) cm s'I for eastward and 1.5
satellite data; fifteen minute averages (9.4) cm s-1 for northward velocity. Most
were mapped to a 0.25 degree x 10 meter of the larger differences 1occurred in the
grid for contouring. high shear (up to .03 s ) regions above
the EUC core, where differences from the
nominal mooring depths could change
eastward velocity significantly. The mean
(standard deviation) U difference at 160
ADCP ACCURACY AND CALIBRATION meters was -.6 (13.1) cm s- I ; at 250 meters
it was -18.5. (10.8) cm s- 1. most
.Differences in time and space scales measurements. to date suggest that 200
make comparisons between ADCP data and meters is the maximum depth of accurate,
other current measurements difficult. In measurements for this installation.
OCEANS 1836, agreement was shown between Given a high enough level of back-'
individual Pegasus and ADCP profiles to scattered energy for adequate signal
depths exceeding 250 m. Further studies of processing, the inherent random error in
several hundred such comparisons revealed the doppler current speed estimate can be.
that the two profiling methods agree to reduced to reasonable levels ( < 2 cm
within about 5 cm s-1 when shear profiles s-1 ) by averaging for as little as five
are least-square fit in the upper 200 m minutes. Longer averaging, when
(figure 6), but agreement rapidly decreases. appropriate to the fields being sampled,
below 200 meters. In a comparison of reduces this still further. The primary
absolute velocities at 9 stations across obstacle to accurate current measurements
the Florida Straits (fipure 7), the ADCP is contamination by ship motion. The net
measured about 10 cm s- lower than the effects of nearly periodic oscillations of
Pegasus in the northward velocity the ship's fore/aft, port/starboard, and
component, but the two agreed on the small vertical axes (roll, pitch, yaw) due to
mean deep westward velocities in the wave action are small and can also be
eastern Straits. reduced by averaging. Larger biases are

362

0- introduced into the sy stem by transducer
a misalignment - the transducer coordinate
_5 0
system is rotated relative to that of the
50 Ao ship - or by individual beam misalignments
within the transducer assembly.
.5 0 At a ship speed of 10 knots, a 10
.5 misalignment of the transducer about the
100. 0 yaw axis introduces a cross track velocity
error of 9 cm S The error due to
ISO- pitch-axis misalignment is smaller and
5 takes the form of a constant multiplier.
Fortunately, misalignments are nearly
constant characteristics of an installation
200 and can be estimated and corrected for
80.0 79.8 79.6 79.4 79.2 during data processing.
0- . For this calibration discussion we will
b use the coordinate transforms and equations
.5 (1988)8 and modified
0 derived by Joyce
50' slightly by Pollard and Read (1988)'. The
ship position is determined and the current
0 velocity desired in a true coordinate
100' system (X,Y) with X and Y positive eastward
and northward. The ADCP unit lies in an
(X',Yl) system, rotated @ 0 clockwise. A
.5 factor A scales velocities in the ADCP
150-
0 0 reference frame and includes effects of
-10 pitch misalignment (proportional to
200 (cosE))-' (U1,V1) are east and north
80.0 79.8 79.6 79.4 79.2, velocities in the (X',Yl) frame, found by
vector decomposition of the ADCP velocities
0- using heading from the ship's gyrocompass.
.C
Subscripts (w,d,s) refer to the velocities
of water (current relative to the ground),
50-
180 doppler (velocities obtained from the ADCP,
relative to the ship), and ship (the
absolute velocity of the ship). These
100- 160 100 three quantities are connected by the
120 so equations:
140
150- 60 120 UW = Us + A ( U d Cos + Vd sin
V = V + A (-U sin + V Cos
W d
200
80.0 79.8 79.6 79.4 79.2 The simplest calibration method is t o
0- compare bottom track (U d) to U determined
.d from precise location measur&ments (GPS,,
LORAN C). This eliminates U so that
180 W
50- tan-' <V U U -V > / <U U + V -V >
80 d d S d d
60 140 A -<U -U. + V V@> / <(U ) 2 + (V ) 2> coso
100- d d d d
10 where < > represents time-averaged values,
100
generally over a calibration run at
1550 60 constant course and speed.
Long straight runs in shallow water
00 with accurate position information are
I
200 surprisingly rare for large research
80.0 79.8 79.6 79.4 79.2 vessels. Immediately following the ADCP
installation in the Researcher in 1982,
four 20 - 40 minute runs comparing bottom
track to LORAN C estimated 0 = -2.67* (a
10
0.79), A= 1.0000 (Daubin, 1983)
Figure 7. Mean currents in the Florida This value of @ appeared to overcorrect
Straits, December 1983. (a) Eastward when applied to actual data sets. A simple
velocity, Pegasus (b) Eastward velocity, method of misalignment estimation was
ADCP (c) Northward velocity, Pegasus adopted that utilized existing data and did
(b) Northward velocity, ADCP. Based on 12
repeated sections of continuous ADCP data
and 9 discrete Pegasus locations.

363

not require special calibration runs. in the AOML logging system is that no
Since reference velocities at depth were values of heading are stored in the ADCP
routinely estimated as a residual between data files. Rosro (1985)11 described
satnav fixes and ADCP dead reckoned typical gyrocompass errors; while latitude
position, they were consistently biased due and velocity errors are routinely
to ADCP misalignment. Using the assumption compensated, the acceleration error (or
that over the course of many cruises there Schuler oscillation) is a complex damped
should be no correlation between current oscillation which changes with,each course
velocity at reference depth and ship change and may have an amplitude of a
heading, the estimated reference velocities degree or more. The solid line in figure 8
(usually at 200 m) were decomposed into shows the estimated f, based on the second
along/across ship track components. For' half of a north-south calibration run, as a
ten cruises in 1983/84 this showed a function of time following the 1800 turn to
misalignment (sin-' <cross-track component north. The dashed line is the estimated
of reference velocity> / <ship speed> ) of gyrocompass error due to accelerations
-1.50 + 0.5. This value gave more during the 1801 course change. The
reasonable results and was used until assumptions made for reciprocal run
further calibration runs could be calibrations (constant current velocities)
performed. become less likely as the runs become long
"Water track" runs in 1986 centered on
a value of t = -1.25*. Water track
calibration assumes that currents remain -3.0 -2.0 -1.0 0.0
constant during.the course of a run that
crosses an area several times at different 0
headings; f and A are determined such that
they minimize the current differences among ........................................................
tracks. The equations for 0 and A are the 100- ......
same as for bottom track, with velocities rotation
U and U replaced by the corresponding
constant
d
differences from their ensemble averages: .... ..............
@=tan_'<SV d -SU B-SU d -Vs>/<SU d. &U,+&V d. &V,> 200 ............ ... ............
A=-<6U -SU +&V .&V > <(&U ) 2 +(6V ) 2 >Cost 300
d S d d d 0.98 0.99 1.00 1.01
where 6U = U - <U>.
I Since water track calibration requires
course changes, the method is easily Figure 9. Depth dependence of 20 minute
contaminated by ship heading errors. The mean t and A during a water-track calibra-
ship's gyrocompass reading at the time of tion run in November 1987.
each ping is used to rotate from transducer
reference frame to earth reference frame
for ping ensemble averaging. An oversight 0.

0.5-

0.0-
c 0.4-
calculated 0
measured
-0.5 @2 0.3-
0
fn
0.2

0.1
1.5 . .....
1.00-
A .989 .002
-2.0 -1 r r ...
0 10 20 30 40 50 60 C
0.99
Minutes following turn C
0

0.98
Figure 8. Estimated misalignment angle of 0 30 60 90 120
transducer assembly (t) vs. time after 1800 Time in Minutes
........................
tion
rota st
con ant

me
a

turn in calibration run (solid line), and
predicted gyrocompass error (dashed line). Figure 10. Calibration constants t and A,
Error is the sum of the Schuler damped versus time in minutes, based on a
oscillation and an empirically determined bottom-track GPS calibration run in
linear slope. November 1987.

364

enough to let the oscillations disappear. REFERENCES
In the absence of better ship heading
information (presently available, but 1 . Bitterman, D. and D. Wilson, 1983.
costly) the next version of ADCP Ocean Current Profiling with a
acquisition software will store heading so Shipboard Doppler Acoustic Backscatter
that a first order correction for System. IEEE Proceedings of OCEANS
acceleration error may be made during 183, pp. 27 - 31.
processing. 2. Smith, Orson P. and J. M. Morrison,
A new ADCP transducer was installed for 1987. Shipboard ADCP Data Applied to
one cruise only in October 1987. A deep Transport Studies in the Eastern
water calibration run had to be cut short Caribbean. abstract in EOS, Trans.
and was again contaminated by gyrocompass Amer. Geophys. Union 68(44), p 1305.
errors (figure 9). This run showed a 3. Wilson, D. and A. Leetmaa, 1986.
strong depth dependence in the estimated Acoustic Doppler Current Profiling in
misalignment, especially below 170 meters, the Equatorial Pacific in 1984.
suggesting weak returns below this level. to appear in J. Geophys. Research
A constant speed bottom-track/GPS run was 4. Leetmaa, A. and D. Wilson, 1985. Near
made later in this cruise to estimate the Surface Circulation Patterns in the
misalignment. Values derived from this run Eastern Equatorial Pacific. Progress
(@ = .37 + .1 , A - .989 + .003; see figure in oceanography 14, pp. 339 - 352.
10) were used on the entire data set. 5. Finkel, R., 1980. Acoustic Doppler
Techniques. In Air-Sea Interaction,
SUMMARY AND FUTURE PLANS Instruments and Methods, F. Dobson, L.
Hasse, R. Davis, eds. pp. 171 -200.
In theory and in practice, the 6. Bitterman, D. and D. Wilson, 1983.
ship-mounted ADCP instrument is capable of ibid.
measuring current shear profiles with an 7. Wilson, D. and A. Leetmaa, 1988. ibid.
accuracy acceptable to most oceanographic 8. Joyce, T., 1988. On in-situ
applications. Factors which further "calibration" of shipboard ADCP's.
degrade absolute measurement accuracy are Submitted to Journal of Atmospheric and
dependent primarily on installation oceanic Technology.
characteristics and processing methods, in 9. Pollard, R. and J. Read, 1988. A
particular measurements of the alignment method for calibrating shipmounted
and relative motion of the transducer and Acoustic Doppler Profilers, and the
true reference frames. This paper limitations of gyrocompasses.
describes how we at AOML are routinely Submitted to Journal of Atmospheric and
dealing with these problems. Individual oceanic Technology.
institutions are independently developing 10. Daubin, S., 1983. Effects of Ship
their shipboard ADCP installations, each Notion on the Output of a Tribeam
with different constraints, and there is a Acoustic . Doppler Current Profiling
recognized need for information flow within System. Technical Report DSC TR 02-83,
the ADCP user community. Daubin Systems Corp., Key Biscayne, FL.
one factor emphasized in our approach 45 pp.
has been to record as much high-quality 11. Kosro, P. M., 1985. Shipboard Acoustic
navi.gation data as possible, in an easily Current Profiling During the Coastal
accessible format. By utilizing features Ocean Dynamics Experiment. Scripps
like the ADCP/satnav link, an atomic Institution of oceanography 85-8, 119 p
frequency input for GPS, and high
resolution (up to seven per minute) GPS and
LORAN C position logging, we maximize
navigation input and quality. Navigation
data is extensively edited during
post-processing; this is made more
efficient by compatibility between our
controller/logger and our lab-based
processing computers.
The next generation of shipboard ADCP
acquisition and processing software will
integrate the ADCP/navigation logging
system with a microVAX based shipwide data
acquisit:ion system. This will make
gyrocompass and other attitude sensor
information available for post-processing,
and provide a framework for other updated
sensors. New sensor emphasis would be on a
deeper-reaching ADCP transducer and a more
stable gyrocompass or other ship's heading
indicator.

365

MAPPING THE SLOPES OF EXPANDING CONTINENTAL MARGINS

Richard B. Perry

NOAA National Ocean Service (N/CG224)
6001 Executive Boulevard
Rockville. MD 20852

ABSTRACT covering waters out to 600 meters deep. Plans
for the next year include a Norfolk-based Class
Modern mapping systems, such as GLORIA which III NOAA ship to be equipped with a multi-beam
provides imagery over a wide area, and Sea Beam system capable of operating to 1,000 meter
which provides swaths of detailed soundings, depths. As of July 1988 the NOAA multi-beam
allow us to take a new look at sea floor surveys, which utilize precise navigation
morphology on and near the base of continental systems, have provided total coverage of
and insular slopes. Sediments are transported approximately 45,000 square nautical miles of
down the slopes in canyons, as well as in the sea f loor in the EEZ. These data are
gravity slides, with much of the material gridded at 250-meter intervals and are
winding up in large fans. These sedimentary contoured for 1:100,000-scale maps using Radian
deposits subsequently are uplifted, expanding Corporation CPS-1 routines. With the exception
the continental margins seaward. One set of of small areas around Cordell Bank off
uplift structures with possible economic California and Loihi Seamount off Hawaii, the
potential, which can be defined by Sea Beam in data remain classified at the request of other
great detail, is a large field of diapirs in agencies.
the Gulf of Mexico north of the Sigsbee
Escarpment. Examples of uplift along the The NOAA surveys are coordinated closely with
tectonical ly- active northern California and the Geological Survey of the U.S. Department of
Oregon-Washington coasts also are discussed. the Interior (USGS). A joint project office
It is concluded that the uplift of the has been established by NOAA and the USGS, with
tectonically-active areas results from material staffing by personnel from both agencies. The
pushing up from below, generally associated USGS has concentrated on reconnaissance surveys
with earth expansion. using the GLORIA long-range side-scan system,
and atlases showing the GLORIA images off the
INTRODUCTION West Coast (EEZ-SCAN F34 Scientific Staff, 1986)
and the Gulf of Mexico and Eastern Caribbean
The National Ocean Service (NOS) of the Areas (EEZ-SCAN 85 Scientific Staff, 1987) have
National Oceanic and Atmospheric Administration been published. The atlases contain a
(NOAA) is continuing its program to survey the description of the GLORIA system, which can
Exclusive Economic Zone (EEZ) using multi-beam image a swath of the sea f loor up to 60 km
echo sounding technology. The survey systems wide. They also show seismic reflection
and processing methodology have been reported profiles, which are very useful to marine
at several previous OCEANS conferences (Hill geologists in interpreting the GLORIA imagery
and Lockwood, 1987; Matula, 1986; Perry, 1982; and Sea Beam maps.
Perry, 1985, Pryor, 1985). An NOS plan for
mapping the EEZ (National Ocean Service, 1987) This paper will describe some of the mapping
which describes the survey systems and areas to results, but it will of course suf fer by not
be covered is available by contacting the being able to show the classif ied Sea Beam
writer. data. As a substitute for the Sea Beam maps,
seismic reflection profiles from the USGS and
There are four NOAA ships working at least part others will be used to illustrate some of the
of their field seasons with the EEZ survey major points concerning uplift off the West
program. NOAA has equipped two Seattle-based Coast. By comparing features covered by both
Class I ships (SURVEYOR and DISCOVERER), and Sea Beam and GLORIA in the Gulf of Mexico we
one Norfolk-based Class II ship (MT. MITCHELL) can get a better appreciation of the strengths
with Sea Beam systems. The Sea Beam system can and the weaknesses of the,two systems, as well
cover a swath equal to approximately 75 per as how they complement each other. It should
cent of the water depth with 16 beams. The be realized that both GLORIA and Sea Beam are
Seattle-based Class III NOAA Ship DAVIDSON has used as general reconnaissance mapping tools,
the Bathymetric Swath Survey System for and that more detailed investigations are

366 United States Government work not protected by copyright

100* 80,
31* 9(r 310

Now He .3
I? Co

S H E L F DeSot
4 Mississippi Canyonr anyon T..pa
A, vk Mi ssippi
?0
0 S
/%\ 7
% .11 mi-i
SI Lower 01@ ((1
Mississippi 10 A@
Q Ri Grande Fen Key W.st
Fen 200
000

3000

22. 22-
100 80,
90*

0 200 400 600 KILOMETERS
i II II I I
0 100 2W 300 NAUTICAL MILES

Figure 1. Generalized bathymetry (in meters) and subsea features
of the Gulf of Mexico. From EEZ-SCAN 85 Scientific Staff, 1987.

needed for resource development and scientific sediment cover increases with proximity to the
studies. A major focus of the paper will be Upper Mississippi Fan as evidenced by the
the author's interpretation of the origin of structures having gentler relief on the western
some coastal margins which have been surveyed side of the survey area in the vicinity of the
by Sea Beam. Some of the ideas relative to Fan. The surveys were run in water depths
uplift on an expanding earth differ from ranging from about 800 to 1900 meters.
conventional concepts of underthrusting at
continental margins on an earth of a fixed A check with the GLORIA atlas covering the same
size, and they should be regarded as those of area (Sheet 16, EEZ-SCAN 85 Scientific Staff,
the author rather than as an official NOAA 1987) showed that the GLORIA system had picked
position on the subject. up portions of some, but not all of the
circular structures. The GLORIA side-scan
GULF OF MEXICO images, which cover a wide swath of sea floor
from a single tow fish, depend upon sea bottom
The NOAA Ship MT. MITCHELL began Sea Beam features with good reflectivity. The result is
surveys in the Gulf of Mexico in 1988. The that GLORIA tends to pick up the portions of
first surveys were over the Upper Mississippi the circular structures which are oriented
Fan (see Figure 1), and there were problems in approximately perpendicular to and close to the
getting a good echo return from the gas-charged sound source, but gets a weak return from
Mississippi Fan sediments. Similar problems portions of the circular structures that are
are common on the shelf and upper slope near oriented approximately parallel to and are
the Mississippi River. The vessel was then remote from the sound source. The overlapping
shifted to the east in the area between the Sea Beam swaths, which are generated from a
Upper Mississippi Fan and the Desoto Canyon. ship passing almost directly above the
circular structures, give an excellent
Every once in a while a systematic survey quantitative data set to map the features. The
program will turn up something very interesting trade-off is that Sea Beam may take weeks to
when it is least expected, and the MT. MITCHELL cover an area that GLORIA can image in a single
hit the jackpot. The survey shows many day. The GLORIA imagery also is excellent in
approximately circular structures ranging in showing the areal extent and transport
size from 2.5 to 9 nautical miles in diameter. mechanisms of sediments which have been moving
In a few cases the circular structures abut, down the continental slope into the deeper
sometimes with deformation of each structure, waters of the Gulf.
and a small canyon between them. The largest
circular structure has a scarp of about 300 The circular structures were interpreted in the
meters on its southern side, and it rises a GLORIA atlas as diapirs, similar to ones to the
total of 500 meters from the sea floor on the west of the Upper Mississippi Fan which have
south to its intersection with the continental been confirmed by drilling to be salt domes. A
slope on the north. Many of the circular discussion of the regional geology in the
structures are covered by a ramp of sediments GLORIA atlas indicates that loading of Tertiary
Jeading down from the continental slope onto sediments onto an underlying salt layer of
the top of the structure. The amount of Jurassic age has resulted in diapiric intrusion

367

by the salt. The GLORIA atlas shows a large closer look be taken at two areas where those
area on the continental slope north of the sediments are being uplifted, (1) an area off
Sigsbee Escarpment to be covered with diapirs, northern California just to the south of the
presumably of salt origin. . The atlas also Mendocino fracture zone, and (2) an area of the
points out that seismic profiles across the lower slope off the Oregon-Washington coast.
escarpment by Amery (1969) suggest that a wedge
of salt is overriding sediments that were The Gorda Escarpment represents the face of a
deposited in the deep waters of the Gulf. raised block of sea floor, the Mendocino Ridge,
Sweet (1986) reports that certain on the south side of the Mendocino fracture
undercompacted shales in the central and zone of f Cape Mendocino. The Mendocino Ridge
western parts of the Gulf have deformed in a has extended the continental margin for more
manner similar to the salt, and are responsible than 60 nautical miles off Cape Mendocino. A
for some of the ridges, swells, and diapirs. seismic profiles of the uplifted block is shown
We thus have a general picture of salt and in Figure 2. The block has been shoved upward
shale structures moving upward through the by forces acting from below, producing an
sediments which have come down the continental escarpment on the order of 1,000 meters in
slope, lifting the sediments as they go, and at elevation. There are no indications that the
the same time the edge of the salt expanding Mendocino Ridge block is overriding the sea
outward as it overrides the deeper sediments in f loor to the north. In fact, the Mendocino
the Gulf. Ridge block appears to be permanently welded to
the King Mountain area to the east, with little
While it has been known for many years that the evidence that the San Andreas fault passes
continental slope in the Gulf of Mexico, has between them. There is considerable historic
many salt domes, it was not realized until.now seismic activity along the Gorda Escarpment,
how effective Sea Beam can be in defining supporting the idea that the Gorda Escarpment
uplift structures of that type. If maps may be continuing its uplift.
showing the structures could be released, they
might assist private industry in targeting Note that the sediments on the back side of the
areas for detailed geophysical investigations. block in Figure 2 are being dissected by
Because there is an existing satellite valleys. These are canyons, including the
navigation system with 24-hour coverage, it Delgada and Vizcaino Canyons, which apparently
should be possible for Sea Beam to map the have formed during and since the raising of the
large area of diapirs located to the north of block. The Sea Beam maps show that the canyon
the Sigsbee Escarpment. The petroleum systems bear no apparent relationship to the
discoveries on several Green Canyon blocks river systems an the adjacent land, where the
suggest that the deeper waters of the Gulf of King Mountains rise from a steep cliff near the
Mexico may be very promising. While the shoreline. There are some incipient canyons
present price of oil and the cost of deep-water forming part way down the continental slope,
production may not add up for immediate with no connection to other drainage patterns.
development, long-range prospects look good. Few canyons head all the way in at the
coastline, but those that do, such as Delgada
Good bathymetric maps will be needed not only Canyon, appear to intercept the longshore
as a framework for detailed geophysical surveys sediment transport and then carry the material
and pipeline routings, but also for related seaward down the canyon. The canyons carrying
scientific studies. In addition to their well- more material tend to have well-developed U-
known presence on ocean ridges, there are shapes, with almost flat, gently sloping
reports of chemosynthetic organisms (vent-type floors. Noyo Canyon cuts into the uplifted
communities) in the Gulf of Mexico (Brooks.and block, and then heads along the strike of the
others, 1985; Grassle, 1985). These include San Andreas fault as it passes along the edge
some at a petroleum seep at 700 to 800 meters of the continental shelf off Fort Bragg.
on the Louisiana slope, and some at a depth of
3266 meters at the base of the Florida The point is that in a geologic sense there is
Escarpment. It is possible that'methane which relatively recent, probably currently-active,
is collected into commercial hydrocarbon uplift of the continental margin along the
deposits is derived both from normal biogenic northern California coast. The uplift is
processes in sediments and from non-biogenic extending the continental margin seaward. That
processes deep within the earth. Since uplift uplifted area is being dissected by young
structures provide conduits for fluid migration canyons whose patterns bear no relationship to
upward, as well as structural traps, it is the drainage on shore.
important that we map them.
Shifting northward to the Oregon-Washington
WEST COAST continental margin, there is both a continental
shelf and an extensive lower slope area
The GLORIA imagery (EEZ-SCAN 84 Scientific dominated by a series of ridges which roughly
Staff, 1986) and Sea Beam maps for areas off of parallel the strike of the continental margin.
California, Oregon, and Washington have The ridges generally tend to be higher in
provided good definition of canyons and slumps elevation on the eastern, or landward, side of
which are transporting sediments down the the area. Where the trend of the continental
slopes. It is appropriate in this paper that a margin changes, the trend of the ridges

368

GORDA + +
ESCARPME11T I ;,r
0
Eureka

-2

125T
1 7@ 00
_Z!
'Ir *r-

211 212 23 174/00 01 0'2 5 ATLAS OF THE EXCLUSIVE ECONOMIC ZONE
1
000* WESTERN CONTERMINOUS UNITED STATES
7.5 KTS U.S. GEOLOGICAL SURVEY 1-1792

Figure 2. Seismic profile of the Mendocino Ridge off Cape
Mendocino. From EEZ-SCAN 84 Scientific Staff, 1986.

-0

T_
1 71,1
,114 1
IT
771r -T7ti;, _T7 -TT-TT-TITT-1
.7 rTT T

T-77T-77 ri: f;=_T-F_TT-7r'-
10
!24 0

7 111TTth -TMT-

i wi

001

I I Ill I rl Iq %I f I I I I I I I I I J, I 1 1 1 1 1 1 1 1 1 t I I
Ij I i@,,F
T _TTTTT 777-7,7'
R'
wl!
-2 IFI;7T7FF-TIII7T r-4
T!!

TT-134-53 5

0 10 20 Krn
Trr
V.E. 11:1
4
_!x,
Figure 3. Seismic profiles off the southwest Washington coast
between 46* 501 and 470 201. Seconds are two-way travel times.
Portion of Sheet 26, McClain and others, 1984.

369

changes, with the ends of the ridges fluids, and gases generated deep within the
interfingering much like interfingering earth. These are the same fluids which are
spreading centers on an ocean ridge. The being discharged at spreading centers on ocean
uplifted nature of these ridges is shown in the ridges, where large plumes of discharged
seismic profiles taken off the Washington coast material have been found during the last few
in Figure 3 (From McClain and others, 1984). years. Because similar materials are rising at
Note the incipient uplift beginning in the uplifting continental margins, there should be
offshore ocean basin sediments in the top increasing reports of vent-type chemosynthetic
profile. communities as we study the base of continental
most of the sediment movement across the area margins in greater detail.
of ridges tends to go from basin to basin in The concept of an expanding earth has of course
gaps between ridges. Other sediments move down been around for a long time, and it has been
the major distributary canyons such as Astoria .summarized in books such as those by Carey
Canyon off the Columbia River. Once the (1976). The reasons for the earth expansion
sediments reach Cascadia Basin they are carried are not clear. Reconstructions of the
along the trend of the continental margin by a fragments of the earth's crust into Pangaea can
series of sea channels, and then deposited in reduce the earth to 80 per cent of its present
the basin near the foot of the slope. This size in the late Triassic-early Jurassic (Owen,
trapping of sediments near the foot of the 1976). Others have reconstructed the fragments
continental slope results from the tilt of the on an earth 75 per cent of its present radius
sea floor upward toward the Gorda and Juan de (Kenneth Perry, personal communication).
Fuca Ridges. The GLORIA images show large areas OCEANS '88 is probably not an appropriate forum
of sea floor near the Juan de Fuca Ridge to be for an in depth discussion of such concepts,
almost devoid of sediment cover. but it should be pointed out that there are
certainly other alternatives to sea floor
The boundary of the continental margin spreading on an earth of fixed size.
apparently is advancing seaward (westward) by
uplifting the thicker sediments near the foot The implication of an expanding earth model is
of the slope into a succession of ridges, each that the sea floor basins, which have all been
one forming further to the west. The uplift formed 'since the Triassic, are relatively
probably results from major regional tectonic passive elements which are created at spreading
forces, and not from salt or shale diapirs. centers as the earth increases its
Griggs and Webster (1986) however, do report circumference. Younger sea floor is formed on
some shale "pillows" on seismic profiles in the an earth of greater diameter, so older sea
Coos Bay, Newport, and Astoria areas, as well floor will be found at greater'ocean depths.
as elongate shale diapirs penetrating upward to This is quite the opposite of the notion that
the sea floor in the Willapa area. older sea f loor cools so that its depth is
proportional to the square root of its age.
PROCESSES OF UPLIFT Since the sea f loor closest to the Oregon-
Washington margin is older than that to the
The uplift of the salt and shale diapirs in the west near the Gorda and Juan de Fuca Ridges,
Gulf of Mexico is relatively straight forward. the sea floor slopes upward toward those
It.is a matter of less dense materials rising Ridges. That sea floor is just sitting there
slowly toward the surface while the loading of passively as the earth expands, and it is not
sediments provides hydrostatic pressure. Most being fed under the west coast of North America
of the salt diapirs in the Gulf are roughly on a "conveyor belt". The sea floor near the
circular, suggesting the absence of regional continental margin, where the sediments are
tension or compression that might align the thickest, may be dropped down due to loading
diapirs into linear ridges. from sediments and from the newly-upthrust
The uplift along the Mendocino Ridge, as well ridges moving onto it.
as the Oregon-Washington coast, results from There is an overthrust of sorts as the
broad regional forces. The writer believes continental margin expands by uplift over the
that the dominant horizontal forces acting in adjacent ocean basin. The earthquakes along
the crust and upper mantle near the uplifted the classic Benioff zone (Benioff, 1955)
ridges is tension acting perpendicular to the reflect the movement of the continental block
trend of the ridges. The tension allows upward over the oceanic block. The result is
segments of the earth's crust to push upward compressional seismic first-motions along the
during earth expansion, with the linearity of Benioff zone, much as in the underthrust model.
the features being determined by the stress This means that those who would make a case for
field. The same tensional and uplift forces underthrusting at continental margins and
have formed most of the relief in the western island arc trenches have the sense of motion of
United Stater., pushing up the Rocky Mountains. the blocks correct, but that there is no long
As the blocks are thrust up near the surface of "conveyor belt" pushing material beneath the
the earth, they may overthkust and fold continents or island arcs. Uplift results not
adjacent sediments and rocks. frcm underthrust material, but rather from new
The uplift probably comes from magma, other material moving upward as the earth expands.

370

CONCLUSIONS EEZ-SCAN 85 Scientific Staff, Atlas of the U.S.
Sea Beam maps and GLORIA images are excellent Exclusive Economic Zone, Gulf of Mexico: U.S.
tools for studying the development of Geological Survey Miscellaneous Investigations
continental slopes, particularly when used with Series I-1864-A, 1987, 104 p.
other data such as seismic reflection profiles.
Sea Beam can be used to map large areas of Grassle, J.F., Hydrothermal vent animals-
diapirs in the Gulf of Mexico. The diapirs on Distribution and Biology: Science v. 229, no.
the lower part of the continental slope north 4715, 1985, p. 713-717.
of the Sigsbee Escarpment are of considerable
long-range economic interest because of Griggs, D. and Webster, F., Pacific in United
potential petroleum development. The uplift of States Outer Continental Shelf Basins - Maps
the diapirs in the Gulf is raising up the lower and Descriptions: U.S. Department of the
slope, and at the same time the salt wedge is Interior, Minerals Management Service, OCS
apparently encroaching on the deep ocean basin Report MMS 86-0048, 1986, p. 26.
sediments, thus expanding the continental Hill, G. and Lockwood, M., Seafloor Exploration
margin. and Characterization - Prerequisite to Ocean
The uplift of the Mendocino Ridge has extended Space Utilization: OCEANS 187 Conference
the continental margin more than 60 nautical Proceedings, Marine Technology Society/IEEE,
miles westward near Cape Mendocino., New Halifax, 1987.
canyons are forming on the back side of the Matula, S.P., Using Exclusive Economic Zone
block forming the Mendocino Ridge. Those Digital Swath Data to Select Effective
canyons which head near shore, such as Delgada Bathymetry: OCEANS 186 Conference Proceedings,
Canyon, are able to intercept the longshore
sediment movement and they tend to be wider, Marxne Technology Society/IEEE, Washington,
with U-shaped valleys and gently sloping 1986, p. 136-140.
floors. McClain, K.J., Peper, J.S. and Holmes, M.L.
The uplift of a series of linear ridges on the Single Channel Seismic Reflection Records of
lower slope off the Oregon-Washington coast is Western Washington Margin and Cascadia Basin,
expanding the lower slope westward. . It is in Kulm, L.D. and others (eds.), Western North
postulated that the uplift of those ridges, as, American Continental Margin and Adjacent Sea
Floor off Oregon and Washington: Atlas 1, Ocean
well as the Mendocino Ridge, results from Margin Drilling Program, Regional Atlas Series:
tensional forces acting normal to the major Marine Science International, 1984, Sheet 26.
trend of the features, at the same time that
there is uplift from within the earth. It also National Ocean Service Plan for Mappi ng the
is postulated that the uplift is associated
with earth expansion. Seafloor of the United States Exclusive
Economic Zone, 1988-92: National Oceanic and
REFERENCES Atmospheric Administration, U.S. Department of
Commerce, 1987, 51 p.
Amery, G.B., Structure of the Sigsbee Scarp, Owen, H.G., Continental Displacement and
Gulf of Mexico: American Association of Expansion of the Earth During the Mesozoic and
Petroleum Geologists Bulletin, v. 53, no. 12, Cenozoic: Philosophical Transactions of the
1969, p. 2480-2482. Royal Society of London, v. 281, no. 1303,
Benioff, H., Seismic Evidence for Crustal 1976, p. 223-291.
Structure and Tectonic Activity: Geological Perry, R.B., Scientific and Hydrographic Use of
Society of America Special Paper 62 the Bathymetric Swath Survey System: OCEANS '82
(Poldervaart, A. ed.), 1955, p. 61-74. Conference Proceedings, Marine Technology
Brooks, J.M., Kennicutt, M.C. II, Bidigare,. Society/IEEE, Washington, 1982, p. 396-401.
R.R., and Fay, R.A., Hydrates, Oil Seepage, and
Chemosynthetic Ecosystems on the Gulf of Mexico Perry, R.B., Mapping the Exclusive Economic
Slope: Eos, v. 66, no. 10, 1985, p. 106. Zone: OCEANS 185 Conference Proceedings, Marine
Technology Society/IEEE, Washington, p. 1193-
Carey, S.W., The Expanding Earth: Developments 1197.
in Tectonics 10, Elsevier, Amsterdam, 1976, 488 Pryor, D.E., Overview of NOAA's Exclusive
P. Economic Zone Survey Program: OCEANS '85
EEZ-SCAN 84 Scientific Staff, Atlas of the Conference Proceedings, Marine Technology
Exclusive Economic Zone, Western Conterminous Society/IEEE, Washington, 1985, p. 1186-1189.
United States: U.S. Geological Survey
Miscellaneous Investigations Series 1-1792, Sweet, W., Gulf of Mexico in United States
1986, 152 p. Outer Continental Shelf Basins - Maps and
Descriptions: U.S. Department of the Interior,
EEZ-SCAN 85 Scientific Staff, Atlas of the U.S. Minerals Management Service, OCS Report MMS 86-
Exclusive Economic Zone, Gulf of Mexico: U.S. 0048, 1986, p. 11-18.

371

MODIFICATIONS AND IMPROVEMENTS TO THE SEA BEAM SYSTEM
ON BOARD RIV THOMAS WASHINGTON

C. de Moustier, T. Hylas and J.C. Phillips

Scripps Institution of Oceanography
Shipboard Computer Group, A-023
La Jolla, California 92093

ABSTRACT During this time, the engineers and the technicians of SIO's Ship-
board Computer Group (SCO), who operate and maintain the system,
A Sea Beam multibeam bathymetric survey system has been 1 ,n began to modify some of the components of the system to improve
operation on board the R/V T. Washington of the Scripps Institution reliability and performance as well as to facilitate the maintenance
of Oceanography since December 1981. In response to operational tasks. In the last four years, the improvements to the operation and
requirements, the engineers of the Shipboard Computer Group at maintenance of the Sea Beam system have been tied to an upgrade of
Scripps have implemented a number of modifications to the system's the shipboard computer from an IBM 1800 to a DEC VAX- 11/730
Narrow Beam Echo-Sounder and to its Echo Processor. These system, henceforth referred to as VAX.
include the design and construction of a digital pitch compensator, In the following we discuss the design philosophy adopted for
the ability to use a variety of sensors for vertical reference, the logging the bathymetric data output by the Sea Beam system on the
design and constniction of hardware test equipment, of an interface VAX, and its implications for expansion with new hardware or addi-
to the shipboard DEC VAX-I1n30 computer for data logging and tional interfaces. We then describe the implementation of the new
for automation of start-up procedures as weU as for performance
monitoring. Some of the modifications have prompted the manufac- TERh
EN
to
turer of the Sea Beam system, General Instrument Corporation,
upgrade the system they have delivered in the past 4 years to match R ROLL ECLIPSE
the con-esponding improvements implemented at Scripps and which S/D 0
could prove useful to Sea Beam operations on other ships. CONVERTER CONTOURS U)
(A
---- RT6--V' U.1
0
CONVEFITER 0
CONT M
& CIL.
I INTRODUCTION KEY TIMI ECHO 0
PROCESSOR
RECEIVERS (16)
A Sea Beam bathymetric survey system, manufactured by the KEY
General Instrument Corporation (GIC), was installed on board
SWATH
R/V T. Washington of the Scripps Institution of Oceanography HEADIN T
GYRO F E PLOTTER
E
(SIO) in the fall of 1981. This system is a multibeam echo-sounder
designed to measure depth on 16 discrrte beams, with 2 2/30 angular
resolution, and produce a swath of contours along the ship's track. it DIGITAL VERTICAL, BEAM LINE TEST
PITCH ITCH PO , STBD DRI SIGNAL
consists of two main components: a narrow beam echo-sounder and COMPENSATOR RE VERS (16) GENERATOR
an echo processor (Fig. 1). The echo-sounder consists of two huu- cc
mounted transducer arrays, installed at right angle to each other in a POWER MECHANICA LU
AMPLIFIER POLL 0
T configuration, their associated signal generator, pitch androll com- Z
1 (20) C BEAMFORMING
pensators, timing unit, power amplifiers, preamplifiers, and beam- MATRIX
forming network. The echo-processor utilizes a Data General RO ROLL (4 X 40)
Eclipse S- 130 real-time computer for the digitization of the echo sig- SERVO 0
nals received and detected on each beam, for bottom detection and PREAMPLIFIERS
tracking, for geometric correction (roll and refraction), and for com- (40)
puting depths and horizontal distances to be displayed as an instan- W
taneous depth profile or contoured and output to a swath plotter.
More extensive descriptions of the Sea Beam system and its opera- 0
cc
tion are found inRenard and Allenou [11, Farr [2] and de Moustier
and Kleinrock [3]. Z
Following acceptance tests and sea trials in December of 1981 1>
aboard the R/V T. Washington, the system has been in operation an
average of 180 days per year over the past seven years, with rela- Figure 1. Block diagram of the Sea Beam system with its two main
tively Ettle down time due to system failure. Several failures of the sub-systems: a Narrow Beam Eacho Sounder and an Echo Processor.
pitch compensation unit occun-ed in, the first two years of operation. Elements in the shaded areas are peripherals and external sensors.

CH2585-8/88/0000- 372 $1 @1988 IEEE

components designed by SCG which include a digital pitch compen- VAX
sator, a sound source synchronizer and test equipment for system
calibration. Finally, we discuss the modifications that have
improved the reliability of the system and facilitated the operator's
tasks such as an automated parameter initialization procedure and GPlB
enhanced output capabilities, the ability to use various vertical refer-
ence sensors, and a capability to record the acoustic waveforms
received on each of the 16 preformed beams.
STD
II DATA LOGGING AND PROCESSING BUS

34 SPEED
During each transmission cycle, the Echo Processor in the Sea WORDS SPEED
Beam system computes depths and cross-track distances from the LOG
sealloor echoes received on each of the 16 preformed beams. These
data are displayed on an oscilloscope as an instantaneous depth
profile across track, as well as on a swath plotter as a contour chart. Figure 2. Data flow from Sea Beam to the VAX via the STD-GPIB;
In addition the data are sent to a data logging system. In this section, buses.
we describe the current hardware which conveys the data from the
Echo Processor to the Data Logger aboard the RIV T. Washington. Beam's Echo-Processor sends 34 words (i.e. up to 16 depths and 16
horizontal distances as well as time in minutes within the hour) and
Sea Beam to VAX Interface via STD-GPIB Bus ships heading in a bit parallel, word serial format. That is, a 16 bit
The data logging option sold by GIC only consists of two mag- word is transmitted in parallel and all 34 words are transmitted seri-
netic tape drives. An additional computer is required to merge the ally. Three handshake signals associated with the Data Logger output
bathymetric data output by the Sea Beam system with the ship's board ensure the proper synchronization of the data to the Sea Beam
navigation data in order to produce a contour map in geographic STD Interface board.
coordinates. As the R/V T. Washington was already equipped with Another function of the Sea Beam STD interface is to output
an IBM 1800 real-time computer configured to log and process the the ship's speed to Sea Beam. This is done by sending a digital
ship's navigation and underway geophysical data (i.e. magnetics, representation of the ship's speed to the STD bus from the VAX via
gravity), SIO opted to use this computer to log and process Sea GPIB. This information is then changed into the proper analog vol-
Beam data as well. In February 1984, the IBM 1900 was replaced by tage by a digital-to-analog converter and sent to Sea Beam for use in
a DEC VAX- I In30 computer system [4]. the real time swath chart record. The rest of the hardware on this
The front end of the VAX system has microprocessor con- interface is used for address decoding and timing needed for com-
trolled input/output devices utilizing the IEEE 961 Standard (STD) munication with the Z80 microprocessor.
bus protocol in conjunction with an IEEE 488 General Purpose Inter- III NEW DESIGNS
face Bus (GPIB) bus controller (Fig. 2). The STI) bus is an 8 bit
data bus controlled by a Z80 microprocessor. Data from a variety of
sources are transferred over this bus in an 8-bit parallel format under The Z80 microprocessor architecture described in Section II
control of the microprocessor and stored in local memory until inter- has been used in the design of two new components discussed in this
rogated by a GPIB controller in the VAX. The controller subse- section: a pitch compensator and multiple sound source synchron-
quently sends these data to disk storage through a separate STI) to izer. Retaining the same microprocessor architecture had the advan-
GPIB interface. In this manner the VAX is free to perform higher tage of simplifying maintenance, as spare parts would be inter-
priority tasks before polling the individual STD bus devices for their changeable between the various data acquisition systems and the
data [4], [5]. new components, and of capitalizing on the programming expertise
The choice of the STD bus architecture to route Sea Beam data already acquired for the Z80 microprocessor.
to the VAX was prompted by a larger real-time data acquisition
scheme involving several other underway geophysical measurements A Microprocessor Controlled Pitch Compensator
(e.g. gravity, magnetics, single channel seismics) as well as the The transmitter array of the Sea Beam system consists of 20
ship's navigation (e.g. Loran C, Transit Satellites, dual-axis Doppler projectors mounted along the ship's keel. The array is electronically
speed log) [6]. This architecture has the advantage of providing a steered to compensate for the pitching of the ship. To accomplish
versatile instrument interface and data buffering system independent this, a pitch compensator alters the phase of the input signal fed to
of the main computer [4]. li also allows the addition of new sensors each projector relative to the center of the array, projector #10. This
without affecting the performance of the main computer, and the method ensures that the maximum response axis of the transmitted
choice of sensors is not limited to those which have internal data radiation pattern remains aligned with true vertical.
buffering capability. By comparison, computers installed by the The system originally installed on the R/V T. Washington used
NECOR group in support of Sea Beam operations, and tasked only a mechanical pitch compensator in which a synchro signal represent-
with acquisition and processing of Sea Beam data and the ship's ing the ship's pitch angle was fed to a control transformer driving a
navigation data, rely on sensors providing a serial RS-232C output servo motor in a one speed closed servo loop. The servo motor was
[7). The Sea Beam STD interfa cc uses line receivers and differen- mechanically coupled to a gear train which was connected to 20
tial drivers on its data lines to ensure a high immunity to noise for resolvers. Each resolver was geared so that when a pitch synchro
the data flow. The data flow between Sea Beam and the Data Logger signal was input, the resolver turned accordingly and output the
is illustrated in Figure 2. proper phase adjusted signal to each o 'f the twenty projectors.
According to GIC, very few problems had been reported on similar
STD
BUS

34
S
JORD

Every transmission cycle, the Data Logger output board in Sea units installed in other ships equipped with a Sea Beam system or

373

only with a narrow beam echo-sounder. However, many hours of Each counter is loaded by it's own LOAD command from the
maintenance and down time were incurred because of failure of the software to an initial count value. A 3.112448 MHz clock signal is
pitch compensator on the R/V T. Washington. The gear train con- then gated on by a Modified Sonar Key (MSK) and they commence
nected to each resolver would easily slip and bind causing failure of counting. The Modified Sonar Key is a signal created in software
the compensator or, far worse, subtle inaccuracies of the output sig- which limits the length of the pulse sent to the transmitter to 7ms to
nals. In addition, the device's small physical dimensions (12x3x4 prevent damage to the transmitter power amplifiers. The counters
inches) made it difficult to repair or adjust, particularly in a sea- are wired as divide by eight so that by using the high order bit on
going environment. For this reason SCG decided to design and build each counter, a 12,158 Hz output is obtained. The resulting square
a digital pitch compensator that would perform the same task wave is then gated with MSK and the signals at the output of the
without the problems associated with a mechanical design. counter boards are passed to the filter board. The filter board is
The design had to meet the same input-output criteria as the comprised of 20 switched capacitor filters which were selected for
mechanical compensator did. To accomplish this we use an elec- their stability and accuracy. The filters change the 12,158 Hz square
tronically controlled system based on a Z80 microprocessor and the wave output by the counter boards into a 4.5v p-p sine wave. This is
STD bus. As shown in Figure 3, the resulting design requires six cir- d .one to remain compatible with the original system which requires a
cuit boards: sine wave input to the transmitter shaper amplifiers. These
(1) a Central Processor Unit board which contains the Z80 micropro- amplifiers then output a signal to the transmitter power amplifiers,
cessor and associated hardware used to control the operation of the which send a 7ms burst of 12,158 Hz energy to the projectors.
system. It also contains memory devices for permanent and tem- This digital pitch compensator was first tested on the R/V T.
porary storage of inforniation. Washington in 1984, and became an integral part of the system in
(2) a Syncliro to Digital Converter board which converts the pitch August 1985, greatly improving the Sea Beam system's reliability
signal supplied by the vertical reference (Fig. 1) from synchro to and decreasing system down time. The simple replacement of STD
digital format. This information is then used by the processor to out- bus modules in the compensator has also decreased repair time on
put angle and counter delay values. the system. In parallel with SCG's efforts, GIC also developed a
(3) a Display board which outputs pitch angle obtained from the pro- digital pitch compensator to replace the mechanical unit. All the Sea
cessor to a LED display on the front panel of the compensator Beam systems delivered by GIC since 1985 are equipped with a digi-
drawer. tal pitch compensator.
(4, 5) two counter boards accepting delay values from the processor
as input to 20 counter circuits which output a 12,158 Hz square wave Multiple Sound Source Synchronizer
for each of the 20 projectors. One counter board contains counters During geophysical surveys carried out with the R/V T. Wash-
for channels 1-10 the other contains counters for channels 11-20. In ington, a 3.5 kHz subbottorn profiler, a one or two channel seismic
addition one of the counter boards has the necessary circuitry profiler and the Sea Beam system are often operated simultaneously.
required to synthesize a 12,158 Hz square wave. This creates conflicting transmit-receive sequences and these sys-
And (6), a filter board which changes the square waves output by the tems interfere with each other. The Sea Beam system is affected the
20 counters to sine waves required to drive the transmitter shaper most by such interferences as its bottom tracking algorithm may lock
amplifiers. on an interfering signal for a series of transmit cycles and generate
The system is run by software installed in permanent memory erroneous bathymetry or artifacts in the contours which could be
on the CPU board. This software is. divided into three sections: a mistaken for a bottom feature [3]. Experience has shown that opera-
control program, a counter delay table and an angle display table. tor intervention to avoid interference, by slewing sound sources
The function of the software is to wait for a sonar key from Sea away from each other, usually happens after the fact. Moreover,
Beam's timing logic and, upon receipt, read the pitch angle informa- over rough terTain the task becomes rather demanding when more
tion from the synchro-to-digital (S/D) converter. This angle is then than two sound sources are involved. When the Sea Beam system is
used to index a lookup table containing the appropriate phase delays operated in conjunction with the 3.5 kHz subbottom profiler,
to be loaded into each of the 20 counters on the counter boards. The interferences between the two systems can be avoided by simply gat-
software also uses a lookup table containing angle values to output a ing out the 3.5 kHz transmission during a Sea Beam reception win-
pitch angle to a LED display. 'dow. Such a solution only requires a very simple circuit, however it
PITCH
ANGLE SEA BEAM 612158 Hz
PITCH DISPLAY SONAR SINE WAVE
SYNCHRO KEY TO
INPUT TRANSMITTER
SQUARE WAVE INPI-1
12158 Hz JT SHAPER
AMPS

COUNTER COUNTER FILTER
SYNCHRO Z80 CARD CARD CARD
TO DISPLAY CPU 1 2
DIGITAL Channel Channel Channel
1-10 11-20 1 -20
A&
1
21W z
7 AM
@R C1
JNTE [LFILTER
CARD

STD BUS

Figure 3. Architecture of SIO's digital pitch compensator.

374

is inadequate whenever a third sound source is also operated simul- to facil itate the calibration of the transmitter power amplifiers and to
taneously. measure the response of the Echo Processor receivers.
To deal with this problem, SCG designed and built a multiple To calibrate the transmitter power amplifiers, the most vital
sound source synchronization box, also know as "Synch Box", which piece of test equipment is a digital storage oscilloscope (DSO) with a
schedules the firing rates of the three sound sources according to built-in GPIB interface. A DSO is necessary to be able to verify the
selectable priority assignments and to timing parameters computed shape of the 7 ms CW transmit pulse as well as to measure the phase
by a Fortran program running on the VAX. As this system is difference between the outgoing voltage and the return current in
described in detail by Phillips et al [81, only a brief overview of its each power amplifier. Taking advantage of the DSO's ability to
components is given here. digitize, save a waveform and transfer it to a host computer, SCG
The hardware of the Synch Box uses the same Z80 micropro- developed a simple procedure to perform the calibration. The square
cessor and STD bus system described above, and it consists of three wave output of each power amplifier is run through a 12 kHz
circuit boards (Fig. 4). (1) a CPU board containing the Z80 bandpass filter box built in house and digitized with the DSO. These
microprocessor, EPROM firmware for the application program and data are then sent to the VAX through the GPIB/STD interfaces for
RAM memory for system management functions, as well as an analysis. Currently, the values of the 20 projector channels are
upgraded 4 MHz crystal oscillator with an accuracy of 5 parts per entered into a Fortran program which calculates the output power
million. (2) an input-output board which orchestrates the events corresponding to the measurement, and each channel is determined
associated with each sound source under control of the firmware. to be either in or out of specifications and adjusted accordingly. We
And (3) a display interface board which displays the most current intend to remove the filter box from this operation and use digital
center beam depth measured by the Sea Beam system on a 5-digit filtering programs on the VAX to perforin the same function. This
LED readout as well as transmit events for each of the sound sources procedure proved far more reliable and accurate than one 'based on
on individual LED's. A continuous two-way communication takes visual readings from an analog or non-storage oscilloscope.
places between the Synch Box and the VAX acting as a host. The SCG also built an echo delay generator to perform the neces-
firmware checks the status of the various sound source systems and sary tests on the Echo Processor receivers. This device contains a
relays the information to the scheduling algorithm running on the 12,158 Hz oscillator, matching the transmit frequency of the Sea
VAX. It also controls the triggering of source and graphic recorder Beam system, and presetable counters and amplifiers. After ter-
events by loading counters with the appropriate delay parameters minating the hydrophone inputs with a resistor network, the delays
calculated by the scheduling algorithm and passed by the VAX. and amplitudes specified by the manufacturer are set in the device,
This system has been operational on the R/V T. Washington and the response of each receiver is checked for accuracy following
since the winter of 1987. When all three sound sources are in opera- the procedure recommended by GIC. This device has greatly
tion, the seismic system, having the most rigid firing rate require- simplified the method of insuring accurate response of the Echo Pro-
ments, is given first priority, the Sea Beam system comes next and cessor receivers.
the 3.5 kHz subbottorn profiler is last. Once the initial parameters IV MODHICATIONS
have been loaded and the scheduling program started, the Synch Box
operation requires no further operator intervention. modifications described in this section have been imple-
mented to (1) facilitate the task of the operator and reduce the poten-
tial for error by streamlining the Sea Beam start-up procedure, and
VAX by expanding the output capabilities of the system; (2) improve the
system's reliability by being able to switch one of three vertical
reference sensor online, and by alleviating a problem affecting the
RS-232C accuracy of the ship's roll information received by the Echo-
r--------------- ---------------- Processor, and (3) give access to the acoustic signals at the output of
the beamformer to allow their recorftg for further analysis.
Z80
CPU CENTER BEAM Downloading of Sea Beam Initialization Parameters
DEPTH(M)
STD The operator's console provided by GIC to communicate with
BUS DISPLAY 0 the Eclipse computer in Sea Beam is a DECwriter printing terminal,
serving both as keyboard for parameter entry and as printer for out-
1/0 put of system messages by the Eclipse computer. On the RIV T.
Washington, this console has been replaced by a direct RS-232C
connection to the VAX. Console messages sent by the Eclipse com-
------------- ------- ----------- puter are now saved in a log file on the VAX and echoed on a Gra-
3*5 kHz SEISMICS phOn CRT terminal with an attached Okidata printer for optional
hard-copy output. A switch-box was added to allow this terminal to
SEA BEAM be used with either the Sea Beam Eclipse processor or the VAX host
computer and serve as a laboratory checkpoint to monitor the status
Figure 4. Architecture of SIO's multiple sound source synchronizer. of a variety of equipment.
With this setup and a UNIX* C-shell script, it was relatively
Test Equipment simple to automate the procedure required to initialize the Sea Beam
To ensure the continued optimum performance of the Sea system by downloading a parameter file from the VAX to the Eclipse
Beam system, a regular maintenance schedule must be observed computer. The parameters in this file include the Leg number, the
requiring test procedures and measurement equipment. In general: date, the time of day, the ship's draft and a sound velocity profile
SCO's engineers use the test procedures recommended by GIC, how- consisting-of up to 10 pairs of depth and sound velocity. With the
ever specific test equipment and procedures were developed in-house UMX is a trade mark of AT&T Bell Labs.

375

DECwriter terminal these parameters had to be entered sequentially of failure or malfunction of the data logger, an archive copy is kept
at the keyboard, and typing errors meant having to start the sequence on microfilm. For this reason, a second Complot DP-1 plotter was
from the top. This fact becomes noteworthy when it is realized that connected in parallel to produce a continuous duplicate copy of the
the procedure must be followed not only when the system is first swath output, thereby granting the investigator unlimited shredding
turned on or rebooted after a crash, but every time a new sound velo- privileges on the working copy.
city profile must be entered to account for changing oceanographic As mentioned above, plotting of the contoured Sea Beam data
conditions, as well as every few days to reset the Sea Beam clock in a geographic reference frame requires that these data be merged
which is known to drift about 20 seconds per day. with the ship's navigation. On the R/V T. Washington, this opera-
Once the parameter file in the VAX has been satisfactorily tion is performed by the VAX computer, and the data are plotted on
edited, the automated initialization method requires minimal user one of the two CalCornp 965A belt-bed plotters linked with the com-
input, and the dialogue between the VAX and the Eclipse computers puter. There is no direct connection between the Sea Beam system
is displayed on the terminal and the printer for verification. With the and these plotters. During Sea Beam operations, one of the plotters
terminal switch in the "Sea Beam" position the operator executes the is dedicated to plotting, with a few minutes delay, a swath of seafloor
start-up program for the Echo Processor: 'COMM'. Switching the contours along the ship's track in a geographic reference frame.
tenninal back to "VAX", the operator then executes the UNIX C- Parameters such as chart scale and contour interval are user select-
shell script 'RUN.loadsb' which sends the intialization parameters to able. The other plotter is used for Sea Beam chart making in the
the Eclipse computer. A shorter C-shell script 'RUNsvp.time' is post-processing phase. This involves editing and processing the
used whenever it is only necessary to change the time or the sound ship's navigation and remerging it with the Sea Beam data. Further
velocity profile [9]. information on the real-time and post-processing tasks associated
This automated procedure has been a great improvement by with Sea Beam is found in Moore et al, [61 and Charters [101. Also
reducing the potential for operator error as well as the time required note that a similar plotting arrangement is used in the NECOR Sea
to change parameters during a survey hence minimizing data loss. Beam operation [7).
Output Enhancements Vertical Reference System
The Sea Beam system comes with a limited array of output Because the acoustic arrays of the Sea Beam system are
devices and supporting features. Although it has the required driving mounted in a fixed position on the ship's hull, the system needs an
circuitry to support 2 graphic recorders, I step-driven swath plotter, accurate vertical reference to compensate for the roll and pitch
I analog storage oscilloscope, and 2 LED displays for depth read- motions of the ship. As mentioned in Section III, pitch compensa-
outs; only the storage scope and one LED display are standard equip- tion ensures that each of the 20 projector output pulses are properly
ment. The remaining devices must be purchased as optional equip- phase shifted with respect to the center of the array, to align the max-
ment. On the R/V T. Washington, the output devices added in sup- imurn response axis of the transmitted acoustic radiation pattern with
port of Sea Beam operations include a graphic recorder, two swath vertical. Likewise, the seafloor echoes received at the hydrophone
plotters and two belt-bed plotters. array must be corrected for the ship's roll so that (1) the Narrow
The current graphic recorder is an EPC Labs model 3211 with Beam Echo Sounder portion of the system can output the proper
32 kilobytes of built-in memory. This graphic recorder is used to beams to the port, starboard, and vertical receivers and (2) the Echo
display the acoustic return received either on the vertical beam, or on Processor computer can compute the correct depth and cross track
a selected port or starboard beam, or for one of the combinations of distance for each beam.
all port beams, all starboard beams or all beams port and starboard. The Sea Beam system's specifications call for a vertical refer-
If a second graphic recorder were added it would be limited to the ence which outputs. pitch and roll information in a synchro format
vertical beam display only. with an accuracy of one tenth of one degree. Aboard the R/V T.
The memory feature of the graphic recorder plays an important Washington, one of three sensors can be selected for this task: (1) a
role in the implementation of the Synch Box described in Section III, vertical reference gyroscope (Kearfott-Singer) which was purchased
as it allows keying at non-specific times [8). During Sync Box with the Sea Beam system to serve as its source of vertical reference,
operations the graphic recorder is configured for use with an external (2) the gimbaled table of an Anschutz gravimeter and (3) the vertical
record trigger. When Sea Beam is the only sound source in use, the reference unit of a Bell Aerospace BGM-3 gravimeter. A sensor is
recorder is set to its internal trigger mode and the EPC edge pulse is selected by connecting the appropriate jumpers in a gyro junction
used for Sea Beam keying. Both the 3.5 kHz subbottom proffier and box.
the single channel seismic system use the same model of graphic The Kearfott-Singer gyroscope is the most "sea-worthy" verti-
recorders so that plenty of replacements and spares are available. cal reference because it is not as adversely affected by rapid 180
The Sea Beam system delivered to SIO in 1981 included one course changes or rough sea-states (>4) as the gravimeter references.
Houston Instrument Complot DP- I plotter to display a near real-time It contains all the necessary electronics to provide the proper synchro,
swath of contoured Sea Beam data. This swath display includes the output to the system. It requires 115v-400 Hz for its operation and it
bottom contours, time, course, and contour interval. In addition, the has internal transformers which provide the necessary step down to
rate of advance of the paper is a function of the ship's speed supplied 26v-400 Hz needed by the system for reference. Its only significant
to the Eclipse computer as described in Section II. This display is a drawback is its limited mean-time between repairs: one year of con-
very useful tool during the conduct of a survey as it provides the stant use, after which the gyroscope's rotor bearings must be
investigator with near real-time seafloor based navigation while fol- overhauled at the factory at a cost of $50,000, with delivery times
lowing geological features of interest or positioning the ship prior to exceeding 6 months. For this reason this gyroscope is now only used
instrument and vehicle deployments as well as for coring and dredg- as a back-up unit and in rough sea states. We do not plan to have
ing operations. However, because this display does not contain any this unit repaired the next time it fails.
geographic information, to follow seafloor features often requires As a substitute source of vertical reference, we used the gyrot-
cutting and reorienting the strip of paper to match the current ship's able of an Anschutz gravimeter which was already installed on the
heading. As this swath plot. would be the only data available in case R/V T. Washington. The gyrotable had two synchro devices giving

376

pitch and roll information, so we only had to provide these synchros three high power operational amplifiers, one amplifier for each wind-
with the proper rotor input signals and the Sea Beam system with a ing SI, S2, and S3. The amplifiers are wired as voltage followers so
26v 400 Hz reference from an external transformer. This system that the input voltage is equal to the output voltage and is in phase
worked well for several years but the need for frequent manual fine for each respe ctive winding. Sufficient voltage sources (+,-.24Vdc)
leveling of the gyrotable and constant maintenance due to its 20 are provided to insure linearity over the entire input range, apprDxi-
years of age as well as the gyro's inability to re-erect itself quickly mately 22Vp-p. Output filter capacitors are also included to prevent
after drastic course changes were severe drawbacks. In February of spurious noise from affecting signal accuracy.
1987, the gravimeter system was replaced by a Bell Aerospace This modification has proved very effective in assuring the
BGM-3 unit; however the gyrotable portion of *the system still accuracy of the input roll signal and the ease of switching between
remains on board as an emergency vertical reference source. any of our vertical reference sources.
The new gravimeter currently provides our primary source of Acoustic Data Acquisition
vertical reference for Sea Beam. It is mounted in a small case on
free moving gimbals aligned to true vertical by two independent The ability to make, for every transmission cycles, 16 discrete
gyroscopes, one for roll and one for pitch. The gravimeter was acoustic measurements with an angular resolution of roughly 2 2/3*
delivered with two resolvers mounted on the case to sense gimbal within an angular sector of about 40' centered on the ship's vertical
position. We built an interface to convert the output of these axis is the basis for the distinct advantage of a multibearn echo-
resolvers into the required synchro signals, and to provide the rotor sounder such as Sea Beam over a single point depth sounder,
input signal for the resolvers and a 26v4OO Hz reference signal for whether wide or narrow beam. Such simultaneous acoustic measure-
Sea Beam. At present this scheme is working well but the gyro dev- ments contain much more information than is necessary for bathy-
ices used in the BGM-3 gravimeter also have a mean time between metry. However, there is no internal provision in the Sea Beam sys-
repairs of about 18 months with delivery schedules of 9 months and tem to preserve the acoustic information received, and the signals
cost in excess of $ 10,000. once processed for bathymetry are discarded. In 198 1, when the Sea
In order to meet the vertical reference requirements of the Sea Beam system was installed on the R/V T. Washington, the Marine
Beam system at a substantially lower cost, we are in the process of Physical Laboratory (MPL) at SIO explored ways to preserve this
installing a Datawell PIRO-120 sensor on the R/V T. Washington. acoustic information. A set of buffer amplifiers were built to tap the
The interface between this sensor and the Sea Beam system uses a detected and rectified envelopes of the beamformed echo signals
Z80 microprocessor and technology similar to that described in Sec- received on the 16 beams. These signals were sent differentially to a
tion III for other components of the Sea Beam system designed by special purpose acoustic data acquisition system where they were
SCG. digitized and recorded on magnetic tape for further processing [10].
Acoustic data recorded with this system have proved invalu-
An Input Buffer Amplifier for the Mechanical Roll Compensator able to assess the performance of the Sea Beam system and to
The Narrow Beam Echo Sounder portion of the Sea Beam sys- explain the cause of artifacts that were sometimes found in the con-
tem contains a mechanical roll compensator. This device couples a toured bathymetry output by the system [3]. The addition of buffer
specified beam or combination of beams from the 16 preformed out- amplifiers at the output of the receivers in the Echo Processor also
puts of the beamformer to one or more of the pon, starboard and highlighted a 1 MHz noise endemic to the Echo Processor. This
vertical receivers for output to a graphic recorder and to a digitized problem was successfully alleviated by adding capacitors to the
vertical beam depth display. A roll servo system, consisting of an buffer amplifiers. Since then, GIC has released a Sea Beam applica-
input synchro, a servo amplifier and a servo motor in a closed servo tion note advising the addition of a capacitor on the backplane of the
loop, takes roll synchro data from the vertical reference sensor. It Echo Processor receivers. Duplicates of the buffer amplifier boards
rotates a shaft which moves the roll compensator plates to the correct were also made at MPL and given to the NECOR Sea Beam group
position, thus coupling the proper beams to the receivers. who installed similar acoustic data acquisition capabilities on the
Due to the mechanical nature of the system, wear on the gear R/V Conrad and Atlantis 11.
train and friction of binding of the compensator plates can cause a For seafloor acoustic backscattering investigations, the acoustic
non zero servo error which is coupled back through the control data recorded in this fashion proved only partly adequate as sidelobe
transformer and loads the input signal. This can cause gross errors in interference inherent in the multibearn geometry affect the returns.
the roll angle data as, well as subtle errors which are difficult to This situation is particularly damaging in the near-specular beams
detect. In addition, the roll input control transformer is highly induc- were sidelobe interference and bottom return overlap. As only the
tive and has a low input impedance, resulting in a significant phase envelope of the returns was recorded, there was no way to separate
shift (up to 20') between the reference and output windings depend- sidelobe interference from bottom return; to do so requires that the
ing on the the vertical reference sensor used. full waveform be available [ I 11.
This problem required careful attention because roll data are For this reason, in 1985, the MPL data acquisition system was
also used by the Eclipse computer to compensate for the ship's roll redesigned to allow preservation of the amplitude and the phase of
while calculating depths and horizontal distances. Given that the the echoes received. This is done by tapping the signals at the output
output synchros could not be changed, our first modification was to of the beamformer, where they are still in audio form, and by base-
provide a phase shift (RC) network which matched the reference banding and quadrature sampling to obtain the in-phase (1) and qua-
phase to that of the output. Although functional, this method proved drature (Q) components of these signals. The resulting 32 channels
impractical because of the need to reconfigure the input wiling (16 complex channels) are then digitized at approximately 1 kHz per
whenever a different vertical reference sensor was used, and subtle channel, and recorded on magnetic tape. This complex acoustic data
roll errors could still happen if any loading occurred. acquisition scheme, described in greater detail by de Moustier and
A buffer amplifier circuit, installed between the roll angle input Pavlicek [121, has been used successfully aboard the R/V's T.
and the roll compensator synchro, provided a better fix to the prob- Washington, Atlantis 11 and the French Oceanographic Vessel Jean
lem. This prevents any synchro loading from affecting the input and Charcot. Complex acoustic data derived from this system have also
allows us to use any of our vertical references without the need for confirmed the fact that Adaptive Noise Cancelling technique were
any external phase shift network or rewiring. The circuit consists of well suited to remove the sidelobe interference and give access to

317

seafloor acoustic backscatter measurements with the anticipated V111 REFERENCES
angular resolution of roughly 2 2/3' [131.
[1] Renard, V. and J. P.Allenou, Sea Beam multi-beam echo-sounding
V CONCLUSIONS in "Jean Charcot" Description, evaluation and first results
In spite of the problems encountered during the first two years Intemational Hydrographic Review, LVI(l), pp. 35-67, 1979.
of operation of the Sea Beam system, which led to the modifications [21 Farr, H. K., Muldbeam bathymetric sonar Sea Beam and
described in this paper, the system has proved very reliable. This fact Hydrochart, Marine Geodesy, 4(2), pp. 77-93, 1980.
is a tribute to the designers of the system as most of the Narrow [3] de Moustier, C. and M. C. Kleinrock, "Bathymetric artifacts in
Beam Echo Sounder sub-system delivered to SIO had remained Sea Beam data: how to recognize them, what causes them", J. Geophys.
unchanged since it first came out in the mid 1960's. The Res., Vol 91, No B3, pp. 3407-3424,1986.
modifications made to the system installed on the RIV T. Washing- [4] Abbott J.L., S.M. Smith, J.S. Charters, P.G. Downes, T. Hylas,
ton have reduced the need for maintenance or repairs as well as facil- R.L. Moe, J.M. Moore and D.V. Stuber, Scripps seagoing computer
itated the task of the operator. Although these modifications are centers: real-time data acquisition and processing, IEEE Proc. 4th
specific to the SIO system, the automated parameter intialization, the Working Symposium on Oceanographic Data Systems, pp. 123-129, 1986,
multiple sound source synchronization, the buffering of the roll input [5] Charters J.S., SIO Sea Beam software documentation, Shipboard
and the improvements in calibration procedures can be adapted to Computer Group A-023, Scripps inst. of Oceanog., La Jolla CA 92093,
other Sea Beam installations that include a data logging computer. unpublished document, 1986.
The modifications described in this paper remain relatively [61 Moore, J. M., J. S. Charters and C. de Mounier, Multi-sensor
benign as they were intended to improve the existing system rather realtime data acquisition and preprocessing at sea, Proc. MTS-IEEE
than design a new one. By today's standards [14] the Sea Beam Sys- Oceans'88 Conf., Baltimore Maryland, 1988.
tem will need to undergo more drastic changes to meet the needs of [71 Tyce R.C., S. Ferguson and P. Lemmond, NECOR Sea Beam data
the oceanographic community for wider swath widths and finer collection and processing development, NITS Journ., Vol. 21 No. 2,
angular resolution. This will require modification of the acoustic pp. 80-92,1987.
arrays and the use of digital beamforming techniques. [8] Phillips J.C., J.L. Abbott and C. de Moustier, Multiple sound
VI ACKNOWLEDGEMENTS source synchronizer for seafloor surveying, OTC #5867, Proc.
The development work reported in this paper has been funded Offshore Technology Conference, Houston Texas, 1988.
in part under the Accelerated Research Initiative on multibeam sys- [91 Smith S.M., J.L. Abbott and J.C. Phillips, Scripps Sea Beam start
tems sponsored by the Office of Naval Research (ONR), Contract up procedure and downloading of terminal entry parameters, SIO
N00014-86-G-0142, N00014-85-G-0104 (SCG) and N00014-79-C- Geol. Data Center, Sea Beam Series No. 5. internal
0472 (MPL); and from SIO Institutional funds. The Sea Beam Sys- unpublished repoM 1987.
tem installed on the R/V T. Washington was purchased with an ONR [101 de Moustier, C., Sea Beam Acoustic Data Acquisition System,
grant, Contract N00014-80-C-0440, matched equally with trust funds Technical Memorandum 379, Scripps Institution of Oceanography, San
from SIO. J. L. Abbott directed and was actively involved in the Diego, CA, 1985
development work done by the Shipboard Computer Group. Under [IlldeMoustier,C. "Beyond bathymetry: mapping acoustic
his guidance the main contributors to the hardware development backscattering from the deep seafloor with Sea Beam", J. Acoust.
were P. Downes, T. Hylas and D. Stuber for the STD data acquisition Soc. Am., Vol 79, No 2, pp. 316-331, 1986.
system, P. Downes for the digital pitch compensator, J.C. Phillips, T. [121 de Moustier C. and F.V. Pavlicek, A fully transportable Sea Beam
Hylas and D. Stuber for the Synch Box, and T. Hylas for all the other complex acoustic data acquisition system, OTC #5514, Proc. Offshore
modifications to the Sea Beam system. The Sea Beam acoustic data Technology Conf., Houston Texas, pp. 269-274, 1987.
acquisition work done at MPL was initiated by R.C. Tyce and pur- [131 Alexandrou D. and C. de Moustier, Adaptive noise cancelling applied
sued by C. de Moustier and F.V. Pavlicek who have developed the to Sea Beam sidelobe interference rejection, IEEE J. of Oceanic
necessary hardware and software, with significant software contribu- Eng., Vol. 13 No. 2, pp. 70-76,1988.
tions by R. M. Lawhead. We thank E. Ford for typing and format- [ 14] de Moustier C., State of the art in swath bathymetry survey
ting and J. Griffith for the art work. systems, in Current Practices and New Technology in Ocean
Engineering, G.K. Wolfe and P.Y. Chang Eds, ASME OED-Vol. 13,
pp. 29-38,1988.

378

THEORY AND TEST OF BATHYMETRIC SIDE SCAN SONAR

Donald E. Pryor

Office of Charting and Geodetic Services
National Ocean Service
National Oceanic and Atmospheric Administration
6001 Executive Boulevard
Rockville, Maryland 20852

ABSTRACT corresponding to any point in the side-scan
image.-
Bathymetric, or interferometric, side scan-
sonars offer great improvement over
conventional hydrographic and bathymetric PHASEDIFFERENCE DsInO
techniques because their broad swath makes BETWEEN RECEIVERS 2nA
it possible to survey an area more D = ARRAY SEPARATION
efficiently, their high spatial resolution A= ACOUSTIC WAVELENGTH
makes it less likely to miss an obstruction
or feature, and their image output gives
valuable information about the bottom
composition. Accuracy and data processing
improvements have appeared necessary before
such systems are accepted for routine TRANSDUCER
operational use. A theoretical model of ARRAYS
the phase measurement errors which limit
accuracy has been developed. Tests of a
shallow-water bathymetric side-scan sonar,
the Bathyscan 300, were conducted in
August, 1987, in the Chesapeake Bay. The
results of those tests, as well as the
performance demonstrated by other
bathymetric side-scan sonars, are compared
to the predictions of the model.

1. INTRODUCTION,
Bathymetric Side-Scan Sonar Technique
Figure 1
The National Ocean Service (NOS) is
responsible for mapping and charting of the Systems of this type can be towed near the
waters around the United States. Surveys surface at the same speeds that hull-
conducted to meet this responsibility must mounted systems are operated. They can
be done efficiently, they must meet provide complete coverage of the seafloor
international accuracy standards, and they over swaths of 3.4, and perhaps as much as
must be adequate to withstand legal 10, times the towfish altitude. High
scrutiny in liability cases. Bathymetric, frequency versions can operate in water
or interferometric, side-scan sonar systems depths as shallow as 10 meters or less.
offer substantial potential improvements The spacing between conventional survey
over current techniques in a wide variety lines can be increased and still provide
of conditions. greater confidence that all natural
features and obstructions to navigation
The basic interferometric technique uses a have been detected. In deeper water, low
pair of transducers which are very similar frequency versions can reach to 6,000
to those used for conventional side-scan meters or deeper. The broad swath width
systems. A pulse is transmitted just as in of these systems means that the rate of
a conventional system. In addition to area coverage by a survey ship can be four
monitoring the amplitude of the returning times (or more) greater than with today's
echo, the phase difference between these multibearn systems. In addition to
two transducers is monitored. This phase providing more complete and efficient
difference indicates the direction of coverage, bathymetric side-scan systems
arrival of the echo at any instant. From also provide bottom backscatter imagery
the angle and time, depth can be computed indicating bottom composition.

379 United States Government work not protected by copyright

The disadvantages of this technology, as of systems is now offered by Honeywell. These
recently, have been that the accuracy did systems evolved from equipment developed
not meet international standards and that for manganese nodule exploration and the
the large amounts of data produced could search for the Titanic. International
not be processed efficiently. Progess is Submarine Technology (IST), the Hawaii
being made in both areas. As a Institute of Geophysics (HIG) and Seafloor
contribution to this progess, NOS sponsored Surveys International (SSI) all made
the development of a theoretical model of important contributions to the development
the accuracy of bathymetric side-scan of the bathymetric capability. The newest
systems. In addition, in order to obtain version of SeaMARC is being developed in
direct experience, a field test of a cooperation with Texas A&M University
shallow-water version was conducted in (TAMU). The characteristics of the
August of 1987. Bathyscan 300 and the SeaMARC systems are
shown in table 1.

2. BACKGROUND These systems are not yet available "off-
the-shelf". However, the SeaMARC II
(offered now as Honeywell's series 12) and
The first experiments with the the SeaMARC/S (offered as Honeywell's
interferometric technique for seafloor series 150) have seen extensive operational
mapping were conducted in the 1960s (see use in recent years. They have clearly
Chesterman et al, 1967 and Lowenstein, developed beyond the research phase.
1970). Research continued through the
1970s and 1980s at a number of locations
around the world. The BASS system was 3. THEORETICAL MODEL
built in the United Kingdom (Denbigh,
1979). The SeaMARC II system was built in
the United States (Blackinton and Hussong, The accuracy of a bathymetric side-scan
1983). The IDSS system was built in West sonar system is affected by many of the
Germany (Kolouch, 1984). The TOPO-SSS same factors that limit the accuracy of
system was built in Norway (Klepsvik, multibeam systems. The most important of
1984). The technique has been employed in these common factors are errors related to
Canada (Caulfield, 1984) and the Soviet sound velocity and errors related to the
Union (Aleksandrov, 1983). sonar attitude. The other large factor in
a multibeam, system is the error in
Recently, the first systems of this type estimating the travel time for an echo to
began to be available as commercial return in a given beam. The travel time
products. The Bathyscan 300, offered by error is related to the transmit pulse
Bathymetrics, Ltd., is a direct descendant length, the composite transmit-receive
of the BASS system and other work done in beamwidth, the signal-to-noise ratio, and
the United Kingdom. The SeaMARC family of the shape of the seafloor within the beam.

Table 1
Bathymetric Side-Scan Systems

System SeaMARC SeaKARC SeaMARC Bathyscan 300
Series 12 Series 70 Series 150

Developer IST/HIG IST/TAMU IST/SSI Bathymetrics

Max. Depth 6,000 2,000 300 60
Below Towfish (m)

Frequency (kHz) 11/12 72 150 300

Max. Swath Width
(m) 10,000 2,000 1,000 200
(% towfish altitude) 340 340 340 700

Towfish Length (m) 5.5 2.5 2 2

Towfish Weight (kg) 1750 150 204 180

Max. Tow Speed (kt) 10 10 6 8

Number of Systems 1 under 1 1
Fielded construction

380

In an interferometric system, the error
source analogous to travel time in a
multibeam system is that of estimating the when (p TE which caused the broadest
direction of arrival from the phase distriKution. when Tl=n/2 it was found
difference between the transducers. The that the interference not only broadens the
relationship of this error to the design distribution, producing a standard
parameters of the sonar is not well known. deviation of:
Several investigators have considered this a=1.5/,/(S/I)
problem. The consensus starting point is
the probability density function of the but also shifts the mean of the
phase difference as derived by Ol'shevskii distribution away from the direction of the
(1967): wanted signal by:
1-IRi2 (P)+TEP/2+(j_p) 1/2 ]
f(d(p)= ---------- [Psin A program was written by Science
2Tx(l_p2)3/2 Applications International Corporation
(SAIC, 1987) to incorporate this result
for jd(p-4j<TE into an overall performance prediction for
bathymetric side-scan sonars. The user
and 0=jRjcos(dT-W specifies both the system design parameters
and the operating environment. This model
where dT is the phase difference, Vt is its includes not only the depth errors related
mean value, and R is the crosscorrelation to phase estimation, but also those related
coefficient. This is a peaked distribution to sonar attitude and sound velocity. The
whose width increases as the program also calculates the errors in
crosscorrelation decreases. positioning soundings produced by the
specified system.
The crosscorrelation coefficient is Figure 2 is a representative result of this
difficult to obtain in circumstances of program. For this particular run:
interest. Alexandrou (1987) considered the Sonar Design
cases of homogeneous volume reverberation frequency: 300 kHz
and single surface boundary reverberation transducer width: 0.44 cm
which indicate how the crosscorrelation receiver spacing: 5 cm
depends on receiver seperation and receiver tilt angle: 200 below
orientation. Gapper and Hollis (1985) horizontal
demonstrated the dependence of the attitude errors: 0.10 rms (roll,
crosscorrelation on signal-to-noise ratio pitch and yaw)
and the autocorrelation of the transmit towfish depth error: 10 cm rms
waveform. Klepsvik (1984) showed that the sound velocity error: negligible
crosscorrelation can be factorized to Environment
include the dependence on both the transmit water depth: 50 m
pulse and the reverberation towfish depth: 10 m
characteristics. Additional effects, bottom type: sand
primarily interference produced by back- wind speed: 10 knots
scattering from surfaces other than the
seafloor, are recognized as limiting
factors. Multipath interference involving Depth Error (1.25 WAY)
reflections from the sea surface is the
reason that coverage is limited to 3.4 Coverage (m 2/see) 513
times the water depth in the SeaMARC
systems. Glint, or interference produced
by multiple strong reflectors on the
seafloor, can also have an important effect
on the phase distribution (Blackinton,
1986).

Denbigh (1987), in an effort sponsored by
NOS, developed a relationship between the
probability density function of the phase
difference and the signal-to-interference
ratio, S/I. The interference was assumed
to arrive from a direction different from
the desired signal and which would cause a Dis tance (25 R/div)
phase difference, @pj, between the two
receivers. He showed that the width of the
probability distribution of phase Theoretical Model Performance Prediction
differences, expressed as a standard Figure 2
deviation, could be approximated by:
--Measurements were -a-l-lowed-to -be-aver-aged-

381

over a 10 square meter area as the system was orders of magnitude greater. The final
moved at 5 knots. The figure shows that form of the data was as a 5 meter grid.
the rms errors remain below 1% of the water The rms difference between the gridded data
depth out to a range of 100 meters. To from the two sets of lines was about 0.2
meet international standards, which require meters. Data analysis has not yet been
that the total error not exceed 1% of the completed but the differences with respect
depth or 0.3 meters (whichever is greater) to the conventional echo-sounder data
with a probability of at least 90%, the appear to be no greater than the
range would be restricted to about 75 differences between the two grids.
meters. These parameters are close to the
specifications and typical operating The second portion of the tests, designed
conditions for the Bathyscan 300. The to examine survey suitability, was to cover
model's prediction is in general agreement a larger area over the range of conditions
with the manufacturer's specification. The available within the Chesapeake Bay. The
model indicates only small dependence on area chosen was 1 mile by 5 miles in extent
bottom type, but strong dependence on wind and included water depths from 9 to 50
speed. The model's predictions for other meters. Operations were conducted as a
systems are close to the manufacturer's routine survey as much as possible.
specifications and to the results of Conventional echo-sounder data was gathered
operational experience. simultaneously. Figure 3 is a stacked-
profile perspective plot of the results of
4. FIELD TESTS this portion of the tests. The survey
lines in this area were also spaced by 70
Field tests of the Bathyscan 300 were meters. The work was completed in 3 1/2
conducted from the NOAA ship Rude during a days of routine operations. The towing
two-week period during July and August of depth was maintained at 3 meters. Adequate
1987 in the Chesapeake Bay. The Bathyscan returns were obtained out to ranges of 100
300 was viewed both as an example of a meters except in a small region in the
bathymetric side-scan sonar and as a deepest part of the survey area. A 5 meter
promising tool for surveying of harbors and grid was produced from the data and the
harbor approach areas. The tests were plot drawn from this grid. Random
divided into three phases to test accuracy, variations are clearly not greater than a
suitability for surveying, and capability few tenths of a meter. Comparison with the
for detection of obstacles. conventional echo-sounder data showed
differences of up to 1 meter which extended
The first portion, designed to check over fairly large areas of the grid. These
accuracy, was run over a generally flat appear most likely attributable to errors
bottom in approximately 12 meters of water. in the towfish depth. A combination of
A 3/4 mile square area was surveyed with information from a pressure sensor, an
two sets of perpendicular lines. Data from accelerometer and adjacent swath data is
the standard conventional echo-sounder, a used to establish this depth.
Raytheon DSF-6000N, was collected
simultaneously for intercomparison. The The third portion of the tests was designed
Bathyscan system was towed at a depth of 3 to examine the capability of this type of
meters and a speed of about 5 knots. The system to detect obstacles to navigation.
normal 300 depression angle was reduced to Several objects which had been investigated
200 to get better coverage in this depth of by other means were used for targets. The
water. Ranges consistently exceeded 70 most useful data came from the wreck of a
meters and typically exceeded 100 meters pile driving barge. The dimensions of the
which is the manufacturer's specification. barge were 60 feet by 25 feet. It had been
This is more than 10 times the altitude of investigated using conventional side-scan
the towfish above the bottom. There was no sonar and divers. The bathymetric data
evidence of degradation of the output at from the Bathyscan system forms an image
the range at which surface-bottom which shows the scouring that has taken
interference would be received. An place around the wreck. When the
apparent penalty for this immunity to discontinuity in the bottom profile at the
interference which permits long range edge of the hull is less than a few feet,
operation is a gap in coverage directly the bathymetric data maintains track and
beneath the towfish. To fill this gap the produces an image of the wreck itself. By
line spacing was reduced from the full adjusting the spatial averaging to trade
swath width of 200 meters to 60 and 70 off accuracy for spatial resolution and
meters for the two sets of lines. This also by adjusting the ping rate and tow
also insured substantial overlap between speed, the system can be made to detect
lines of data which is important in the smaller objects.
processing of Bathyscan data. The area
coverage rate was slightly less than that Weather conditions during the tests were
of a conventional system which would consistently good. Within that limited
typically be run at 10 knots and 50 meter range of conditions the system performance
line spacing, but the density of coverage appeared independent of wind speed contrary

382

to the model's prediction. The bottom
material in the test areas ranged from soft 6. REFERENCES
mud to sand. The ranges achieved by the
system were clearly dependent on the type
of bottom. Mid-water targets, probably Aleksandrov, A. A. et al. (1983), "Study of
biological, which were very common in the Sea Bottom Characteristics by Side-Looking
area could also disrupt the system Phase Sonar", Oceanology, 23, 3, 378-381.
performance but controls available to the
operator were generally able to reject Alexandrou, Dimitri (1987), "Computer
these effects. Simulations of Bathymetric Side Scan
Sonar", oceans '87 Proceedings, 1168-1174.
The Bathyscan system proved to be robust
and well-engineered. No major problems Blackinton, J. Grant (1986), "Bathymetric
were encountered in set-up or field Mapping with SeaMARC II: An Elevation-Angle
operations. The experience and results of Measuring Side-Scan Sonar System", Ph.D.
the tests suggested that improvements could dissertation, University of Hawaii.
be made in the towfish depth sensor, the
interface with the positioning system, and Blackinton, J. G. and Hussong, D. M.
the real-time displays. It appears (1983), "First Results from a Combination
possible to design a system to be towed at Side Scan Sonar and Seafloor Mapping System
higher speed and to provide some form of (SeaMARC II)II, Offshore Technology
coverage in the near vertical. Two weeks Conference Proceedings, 4478-4484.
following the tests had been planned for
data processing. The task could not be Caulfield, D. D. et al. (1983), "Stereo
completed during that time however the Side-Scan as a Complement to Echo Sounding
processing software has since been for High Resolution Bathymetric Studies",
considerably improved. The tests provided USN/SEG Three Dimensional Data Symposium.
evidence that it is possible for this type
ofd system to meet international accuracy Chesterman, W. D. et al (1967), "Acoustic
standards. The ranges that were achieved Surveys of the Sea Floor near Hong Kong",
indicate that it should be possible to International'Hydrographic Review, 44, 1,
design a system based on this technique 35-54.
which would be capable of much higher area
coverage rates than conventional systems. Denbigh, P. N. (1979), "A Bathymetric Side
The high spatial resolution data produced Scan Sonar", Ultrasonics International 179
by such a system will prove to be of great Conference Proceedings, 321-326.
value.
Denbigh, P. N. (1987), "Bathymetric
Sidescan Sonar; A Study of its Suitability
5. SUMMARY for the Charting of the Sea Bed",
unpublished report for NOS under contract
#50-DGNC-6-00203.
Bathymetric side-scan sonars have
demonstrated considerable promise for Gapper, G. R. and Hollis, T. (1985), "The
hydrographic and bathymetric surveying. Accuracy of an Interferometric Sidescan
The technique has been shown to be Sonar", Proc. Inst. Acoustics, 7, 3, 126-
applicable to shallow harbor areas as well 131.
as the deep ocean. A theoretical framework
for prediction of their performance has Klepsvik, John 0. (1984), "TOPO-SSS: Real
been established. The development of Time Bathymetric Mapping with a Side Scan
models and operational hardware will Sonar", HYDRO 184 Proceedings, 42-49.
continue. This technology is likely to
play an important role in surveying in the Kolouch, Dieter (1984), "Interferometric
near future. Side Scan Sonar, A Topographic Seafloor
Mapping System'11, International Hydrographic
Review, 61, 2, 35-49.

Lowenstein, Carl D. (1970), "Side Looking
Sonar Navigation", J. Institute of
Navigation, 17, 1, 56-66.

01'shevskii, V. V. (1967), Characteristics
of Sea Reverberation, (Consultants Bureau,
New York).

SAIC (1987). "Theoretical Analysis of
Bathymetric Side Scan Sonar", unpublished
report for NOS under contract #50-DGNC-6-
0203.

383

rs,

-'M
Nil
-Y

4, r"t

41-
4'

Ma
@, @-7 7-
t i- -27---L

wj
L

IXI

All#

P
f If w''YEr

Bathyscan 300 Survey Data
Figure 3

384

PROCESSING AND MANAGEMENT OF UNDERWAY MARINE GEOPHYSICAL DATA AT SCRIPPS

Stuart M. Smith, James S. Charters and J. Michael Moore

Scripps Institution of Oceanography, Code A-023
University of California, San Diego,
La Jolla, CA 92093

ABSTRACT In 1984 the IBM 1800 machines were replaced by Vax
'Underway marine geophysical data (navigation, depth; Sea lln30s for shipboard processing and the shore processing and
Beam, magnetics and gravity) collected on Scripps cruises are data management consolidated onto a compatible Vax 11/750 at
processed and displayed at sea or on shore on Vax computers Scripps. Most computer code was modified or rewritten for the
under the UNIX BSD 4.3 operating system. An on-line Master interactive terminal and disk Me environnient provided by these
Abstracts file allows users to search a data base containing more computers and the UNIX(*) operating system (Berkeley,BSD
thari 1300 cruise legs from Scripps and other sources by geo- 4.3).
graphic region or data type. Sea Beam archive procedures and a In the following sections, we first describe how the pro-
tape archive index are briefly described. cessing and management system has been implemented under
UNIX. We then discuss the processing procedures for data col-
1. INTRODUCTION lected on a single cruise followed by an explanation of multiple
The Scripps Institution of Oceanography (SIO) currently cruise data management After a brief outline of Sea Beam data
operates 4 research vessels to support sea-going research pro- archiving, we end with a description of the options available to
grams in all oceanographic disciplines. In the field of marine plot the specified data.
geophysics, underway data (navigation, depth, magnetics and
gravity) are collected continuously at sampling rates ranging 2. DATA MANAGEMENT UNDER UNIX
from seconds to minutes over the duration of a cruise leg, typi- The UNIX operating system has proved an excellent
cally 30 days. The resulting amount of data becomes rapidly environment for the amount of data processing and archiving
overwhelming and a system has been developed, at SIO to pro- done by CDC, once the early difficulties were past on the learn-
cess and manage underway marine geophysical data collected on ing curve due to the terse, sometimes obscure and "trees but no
SIO ships or obtained from other sources. forest" documentation. Programs are coded in Fortran-77 with
The system had its origins in the early 1960s when H.W. occasional "C" subroutines. UNIX C-shell scripts have been
Menard, one of the pioneers of sea floor topography analysis, written for user interface, for chaining together UNIX utilities
encouraged the senior author to look into computer processing of and programs and for job scheduling. Input and output redirec-
the Institution's bathymetric data. Magnetics were added in tion with pipes and filters add to the flexibility. There are a
1968 in a crash program to digitize these data because of the number of tools provided with UNIX, such as 'grep' to list lines
exploding interest in sea floor spreading. In 1981, a Sea Beam that do (or do not) contain a target pattern; 'awk' for pattern
multibeam echo sounder was installed on SIO's principal geo- scanning and processing; a 'sort' command-, and 'cat' to con-
physical research ship, the R/V Thomas Washington. Procedures, , catenate files. The inverted tree directory structure, few file
for processing and archiving this new influx of data (up to I naming restrictions, the use of wildcards and the ease of moving
Mbyte per day of survey) were included in the system. Gravity, between directories all enhance the processing environment.
previously processed for relatively few cruises, was added in The UNIX definition of a file as a simple stream of bytes is a
1987 when a new gravity meter was installed on the Thomas significant advantage compared to other operating systems which
Washington. require that various kinds of files be defined with different for-
Responsibility for data archiving, as well as the mainte- mats and modes of access.
nance and development of the data management system was Most files in the system are in ASCII character mode (with
assumed by the Geological Data Center (GDQ when it was esta- lines defined by new line characters) to take advantage of the
blished in 1970. Extensive programming support has been environnient, screen editor and tools provided by the operating
provided by the Shipboard Computer Group (SCG). Both system. By local convention, lines with a V" character in the
groups are now part of the Ship Technical Support Division of first column are comments and are ignored by the processing
SIO. programs. This scheme has proved very useful and flexible for
In past years, various parts of the system have resided on headers, file documentation and processing notes that become
different computers: Control Data 1604 and 3600, Burroughs part of the file instead of being lost in a separate file or notebook.
7600 and IBM 1800. The IBM ma@hines were the seagoing Files which are large and require little direct human interaction
computers on the larger Scripps ships from 1967 to 1984. Reli- are in binary integer mode for compactness and computational
able, but out of date and punch card based, they were the first efficiency. Floating point mode is not used for data storage to
installation of general purpose computers in continuous oper-
tion on academic research ships. (*) UNIX is a trade mark of AT&T Bell Laboratories.
385 Unites States Government work not protected by copyright

reduce machine compatibility problems. SIO NAVIGATION PROCESSING
W
For clarity, the data flow charts that illustrate this paper oran
show only the main program names and files. In fact, most pro-
grams are called by UNIX C-shell scripts and may have addi- lora
tional terminal input as well as ancillary list or status files.
........ Z@iv(
mag av.. file. cs.Vdratv
fit :::: SAM f s a@c
3. SINGLE CRUISE LEG PROCESSING 110
The underway (u/w) collected on SIO ships and managed
by GDC include the ship's navigation and geophysical data: gps-interp navrtv msetrtv reformat veccsospd
magnetics, gravity and depth (single point and multibeam bathy- (set) 'awk' script
metry). An overview of the hardware and software used for the
real-time acquisition and logging of these data has been given by transit Loran
Abbott et al., [11 and Moore et al., (2]. In this section, we file: saida file: lordta
expand on these narratives with a description of the post- GPS - 1 m riffle transit Imin avg
processing and management procedures applied to the navigation file: gpsdta file: msatdta file: csespd
and u/W geophysical data from a single cruise.
At sea, files containing data with GMT time from each sen- decimate decimate nevrtv
sor or instrument are dumped to tape approximately every two 'awk' script 'awkl script (csesp
days by SCG personnel using the UNIX 'tar' tape archive utility.
By convention each log file, and other files produced in process-
ing, are named by concatenating the type of logged or processed
data and an end date suffix (e.g. MGpfile.88mai,06 for logged GPS - 0 min. Loran 'significant"
file: goscita. 10min file: cse1spds
magnetic data Me ending on 6 March 1988). All files are for- lordta. lomin file: cspdta
matted in ASCII character mode unless otherwise stated.
Navigation
As seen on the flow chart in Figure 1, the ship's navigation merge
is derived from Transit Satellites, GPS or Loran-C for position, (unix utility)
from a dual-axis Doppler speed log for speed and from a gyro- nav integration
compass for heading. The characteristics of the data derived screen fixes fIxPI
from these sensors and the corresponding processing programs EDIT file: sardta
listed on the flow chart are as follows: P nterfile
TRANSIT SATELLITE: Log file 'satfixes' contains position file nav1st
fixes received on a dual frequency ITT Transit Satellite receiver
at intervals ranging from 20 minutes to six hours. Program Figure I
Inavrtv' converts the data from each fix into a one line record in
Me 'satdta' which is inthe input format for fixes required by the minute average and puts the result into file 'csespd' in the same
navigation integration program 'fixpi'. As a backup, fixes from a format. This file in turn is input to program 'navrtv(cse)' which
single frequency Magnavox receiver are logged in file calculates "significant" course and speed points which are put in
'MSAinpf' and converted to another file 'msatdta' by program Me 'cspdta'. (A "significanf' point occurs if the dead reckoning
Imsatrtv'. (DR) position of a given point falls outside a circle with a radius
of 0.02 nautical mile from the position obtained by extrapolating
GLOBAL POSITIONING SYSTEM (GPS): Log Me 'GPTinpf' the previous course and speed.)
contains fixes received on a Trimble GPS receiver that is set to NAVIGATION INTEGRATION: The 'cspdta' and 'satdta' files
output fixes at 10 second intervals. Program 'gps-interp' inter- are next input to a much modified version of program 'fixpi',
polates between fixes, when necessary, to produce a one line received a number of years ago from the Lamont-Doherty Geo-
record at exact one minute intervals. This line is written to file
'gpsdta' which is further decimated to 10 minute intervals by a logical Observatory of Columbia University. This program
UNIX 'awk' script into file 'gpsdta. I Ontin'. linearly interpolates positions of course and speed changes
between consecutive fix pairs and outputs the integrated naviga-
LORAN: Log file 'loran' contains fixes received on a Trimble tion to a 'navIst' list file (which has proved to be a good naviga-
Loran-C receiver at 7 to 10 second intervals. A UNIX 'awk' tion export format) and to a binary 'navbin' file that is used
script converts the data into the input format for fixes in file internally on the Vax for further processing. In the 'navbin' Me,
'lordta' which is further decimated to 10 minute intervals by time is stored in "UNIX seconds" (i.e. time in seconds since I
another script into file 'lordta.10min'. January 1970) rather than in multiple time units. In order to pro-
FIX MERGING: Fixes from the above files, which share a com- cess historical data collected prior to 1970, we have modified the
mon format, are merged and sorted by time into a combined UNIX system's utilities for time conversion so they can handle
'satdta' file using the UNIX 'sort' utility. An experienced opera- negative seconds. Latitude and longitude are each stored as two
tor must then select the best subset of fix data using the screen 16 bit integers (minutes North of the South pole or minutes East
text editor, a process that may require up to an hour for data col- of the Prime Meridian) and as parts of a minute times 10,000.
lected over a two day period. This method allows a resolution of position to 18 cm on the
COURSES AND. SPEEDS: Log file 'csespd.raw' (binary earth's surface while avoiding floating point round off or
integer) contains 10 second averages of courses and speeds sam- conversion problems.
pled at one second intervals from the ship's gyrocompass and the NAVIGATION EDITING: Once courses and speeds are free of
ccsospd

trarts@
file. sa

2-axis Doppler speed log. Program 'veccsespd' calculates a one errors, bad or questionable fixes are removed until the output

386

shows drifts varying smoothly between consecutive pairs of $10 U1W GEOPHYSICAL DATA PROCESSING
fixes. Producing a satisfactory navigation file may take 3 or 4
M
runs of the integration program and require several hours of
knowledgeable human interaction for two days of navigation ea earn apt
6 sec. total field les
1 sec metelloounts -16 sec. P1
data.
twile 4.
file: G
vpfjje
Geophysical Data
The magnetics, gravity and depth data flow are illustrated X
in Figure 2 in the same format as used for the navigation.
MAGNETICS: Log file 'MGpffle' (binary integer) contains
magnetic total field readings received from an EG&G magrt beilgray sbcrtv
Geometrics magnetometer at six second intervals. Program
'magrtv' calculates one minute averages and puts the results in depth
file 'mag.uwts' (magnetic underway time series). magnetic gravity vertical beam
1 min avg. I min avg. I min avg.
GRAVITY: Log file 'GVpfile' (binary integer) contains one file: ma_q.uwts file: gravuwts file: d@qpxwfs
second interval counts received from a Bell Aerospace BGM-3 EDIT EDIT
gravity meter. Program 'bellgrav' applies a recursive 25 pole
filter and a 148 second time delay to filter the short period ship
accelerations and puts one minute values in file 'grav.uwts'.
DEPTH: For cruises when the Sea Beam multibeam echo navigation UWmerg
Cale mag. grav
sounder is operating, the Sea Beam log file 'SBrtm' contains 34 file: navbin free-air anomalies
data words (up to 16 depths, 16 horizontal distances, heading and
time) per ping cycle (one to 16 seconds, depending on ocean
U Cho
profit
List time
r
depth). The average value of depth measured on the vertical U/W
List time,
beam is extracted over one minute intervals by program 'sbcrtv' merged data
Zia gaps P
file: uwmrg ip ti
rro a
aip ti
into file 'dep.uwts'. Note that only the vertical beam depth rs hi
measured with the Sea Beam system is used here, This is done ....... ..
...... ..............
@i .. .. ... .
profil
to conform to existing geophysical data exchange formats
pot
(MGD77) as described below. Management of the full set of Sea ........ .
Beam data is discussed in Section 5. For legs without Sea Beam, Figurce 2
depths are manually scaled from a depth recorder trace at five
minute intervals and the values typed directly into a 'dep,uwts' merging by subtracting the Gravity Reference Formula (GRS 67)
Me with the help of script 'enterdepths'.
and adding the E6tvos cortection to the measured gravity value.
GEOPHYSICAL DATA EDITING: The 'uwts' files are the ARCHIVING, REPORTS AND MICROFILMING: After edit-
archived form of the data time series and they share a common ing, the processed files, containing approximately two days of
format. Each record contains the date (year, month, day) and six data of each type, are concatenated and the final files named by
pairs of time (hours, minutes, seconds) and data value. Each file file type and cruise ID (e.g. uwmrg.TUGA02WT for the u/w
is passed through two editing procedures. Program 'uwchek' merge file'of Tortuga, leg 2 on R/V Thomas Washington). The
lists times when data errors or gaps exceeding user defined limits final time series, navigation and merged files are copied to tape
occur, whereas program 'tvprof' produces a proffle of data value (2 copies) using the 'tar' utility. A report is prepared containing
vs. ship time on a CRT screen or on a hardcopy dumped to the track charts, a Sample Index and profiles of depth, magnetic and
printer. Errors are then edited with the UNIX 'vi' screen editor. free-air anomaly. The Sample Index is a first level interdisci-
DATA MERGE WITH NAVIGATION: The 'uwts' files for plinary index which gives the begin/end time and position of all
depth, magnetics and gravity are merged with the 'navbin' navi- records, samples and measurements collected on the cruise leg.
gation file by program 'uwmerg' to produce a single 'uwmrg' The original u/w recorder records and the Sea Beam swath plots
merge file. (An ancillary file also lists any time errors that'may are microfilmed onto 35mm continuous flow microfilm to pro-
have gone undetected in earlier processing). Each record in the vide easy access to these data without compromising the original
merge file contains time and position stored in the same fashion records.
as described above for the 'navbin' file, depth (two-way travel
time in seconds times 10,000), magnetic total field and anomaly 4. MULTI-CRUISE DATA MANAGEMENT
(in nano-Tesslas times 10) and measured gravity and free-air The 'uwmrg' files form the principal u/w data archive. In
anomaly (in milligals times 100).
I . order to manage these data, we use a program to search an on-
During the merge process, the magnetic anomaly is calcu- line abstract file and have utilities to assist users in uploading
lated by subtracting the International Geomagnetic Reference files from tape to disk as illustrated in Figure 3. We have also
Field (IGRF) value for the appropriate position and date from the developed an index for locating magnetic tapes in the GDC
measured value. The magnetic field is calculated at the comers archives and methods for data exchange with other institutions.
of a one degree square containing the first position and, as long UNDERWAY MERGED DATA ARCHIVES: The 'uwmrg'file
as the ship remains in that square, the field is calculated by inter- for each cruise leg, typically one month long, is about I Mbyte
polation of the comer values. This method requires considerably when data are stored at intervals of one minute of ship time.
less computation than calculating the field at each point. The Files for older cruises, when data were manually digitized from
maximum difference observed between the direct and interpo- the records at 5 minute intervals, are correspondingly smaller.
lated method is 2.7 nTesslas. The archives hold more than 500 files of Scripps cruises for legs
The gravity free-air anomaly is also calculated during going back to the late 1950s. Although GDC does not attempt to

387

BID U/W DATA IMPORVEXPORT, ABSTRACT AND SEARCH The 'DOsearch' program generates a hit file list of all
cruises satisfying user supplied criteria of geographic location
MGD77 and data type. The 'abstract.header' file is first searched for
Tape tar u merge (Import) bounds overlap of each cruise with the area specified and for
Archives udlity file: uWmrg types of data if also requested. if the header for a cruise passes
the tests, then the 'abstract.data' for that cruise is examined to
header determine if an abstract line segment passes through the search
abstracter uwmrg to Ile: mgdhd area. If depth, magnetic or gravity data are also specified, the
mgd77 .......... data flags for that segment are also examined and must be true
for a hit. The search hit file can, in turn, be input to scripts
MGD77 'DOmake.uwtar.ffles' and 'DOload.uwtar.tapes' to make tem-
entarcrulss xport) porary script files of 'tar' commands and user prompts for tape
adittapold
. . ..... loading. If a large number of cruises are to be plotted, script
'DOmake.infile.lines' produces lines formatted for the 'unlot,
input command file.
DO CRT
absplot
plats RVIPORT/EXPORT: The MGD77 format is the standard for
quick Index
plots
exchange of navigation, vertical beam depth, magnetics and
gravity between academic institutions and for transmission to
DOsearch
11stabscru by area and from the National Geophysical Data Center [4]. MGD77
or data type files received from these and other sources are input to program
'fmgd77' to produce a Me in the 'uwmrg' format (Figure 3).
Usts ot headers search The magnetic anomaly can optionally be recalculated using the
(arid absbr.data) hit file DO make. IntIle.111nes most up to date IGRF reference field. For export, program
DOMake.uwtarmiss Icf 0' 'tmgd77' generates a disk file in MGD77 exchange format using
'--@'-_.Waaalln-d )
coffi,.: the 'uwmrg' file and a 'mgdhd' (MGD77 header) file as input. A
file
script dumps MGD77 files to tape and provides listings of header
temp.flies DOload.uwtar.tapas records, file sizes and other tape statistics.
TAPE ARCHIVE INDEX: In addition to the 'uwmrg' tapes
Figure 3 described above, GDC currently archives over 1000 reels of
magnetic tape of data from Scripps cruises logged by SCG as
duplicate the holdings at the National Geophysical Data Center well as other types of marine geophysical data, such as gridded
(NGDC), throughout the years data from other sources have been bathymetry, heatflow and data sets from the Deep Sea Drilling
acquired in response to requests by Scripps students and staff Project. A relatively simple but effective index has been
bringing the total number of cruise legs to over 1300. These developed to keep track of these tapes. Each tape reel has one
data, totalling nearly 500 Mbyrtes, currently reside off-line on record in a 'tapeindex' data file with fields for reel ID, tape
thirty six 1200-foot reels of 9 track magnetic tape, one copy in name, type of contents, data format (ASCII, EBCDIC, etc), crea-
the computer room and a backup in the archives. tion date, location codes and contents. A user executes the script
ON-LINE ABSTRACTS: Each 'uwmrg' merge file is input to 'DOfind.tape' with a target text such as a Reel ID, cruise name
program 'abstracter' which generates an 'abstr' Me containing a or word in the contents field and the UNIX utility 'grep' win list
header record and times and positions of selected points. The all lines containing the target. A secondary script examines the
header has the latitude/longitude bounds of the cruise and flags location codes in the hit list and provides an expanded storage
indicating the types of data present in the merge file. Points are location description.
selected such that lines between abstract points deviate from the
track by no more than 5 nautical miles. A flag is set for each 5. SEA BEAM DATA PROCESSING AND ARCHIVING
data type (depth, magnetics or gravity) that is present on a given Sea Beam data are processed and archived separately from
segment between abstract points. A typical 30 day cruise leg is the other u/w data. Only the basic data flow (Figure 4) and
represented by several hundred abstract points, a compression to archive procedures are outlined here. Details of the processing,
about one percent of the original merge file. plotting and navigation adjustment programs have been
Each 'abstr' file is, in turn, input to program 'entercruise' described by Cha@rters [3].
which loads it to the Master Abstracts consisting of two files: Program 'sbpfx' merges the Sea Beam data contained in a
'abstract.data' and 'abstract.headers'. The header is appended to 'SBrtrn' log file with the integrated navigation in a 'navbin' file
the 'abstract.header' file with pointers to the location of the data to generate a Sea Beam merged 'SBint' file. This latter file, nor-
that are added to the 'abstract.data' Me. Entries in the mally one to two Mbytes in size for two days of ship time, is the
'abstract,header' can be modified by programs 'edittapeid' to standard archived form of Sea Beam data. The merged file is run
add the tape id on which the 'uwmrg' file is stored and 'editda- through a checking program, 'sbchek', that lists begin/end times,
tacoll' to toggle flags for other types of data collected on the gaps, errors and geographic boundaries of the data and appends
cruise but not contained in the 'uwmrg' file, such as Sea Beam, summary information to a file kept for each cruise leg. The
3.5kHz depth and seismic reflection. merged Mes are dumped to archive tapes with the 'tar' utility
The Master Abstracts can be input to a number of different and can be exported to other institutions with another, more gen-
programs: 'listabscru' for abstract data listings of a known eral, 'dd' utility. At present there is no standard format for
cruise; 'DOabsplot' for quick index plots of the whole leg on a exchanging Sea Beam data but ad hoc exchange of well docu-
CRT screen; the 'uwplot' program described below; and a search mented data in the various formats used by the data producing
program. institutions has worked reasonably well.

388

510 SEA BEAM' DATA ARCHIVING 6. PLOTTING DISPLAYS
Maps and. profiles are the modes of u/w data display most
frequently requested- by data center users. Geologists prefer to
work with large size hardcopies so plots are usually done on
drum or belt-bed pen plotters. Both plot programs described
fx navigation below, as well as the 'sbcontour' program mentioned in the Sea
B with
Inte r led nav. file: na@bin Beam section, use an ASCII text input command file (icf) for
specifying plot commands. Although the 'icf` files differ in
sbchek detail for each program, they share many parameters and an
che k for SB Input
begirdernd times, overall common structure. Command lines contain a keyword
boundaries, Sea Beam wit Command usually followed by one or more parameters in fme-field format
gaps & errors navigation File
file: Silint (e.g. "BOUNDS 12N 5S 121W 119W"). Comment lines can
sbcontauf be included to document parameters or for notes about the plot.
Cruise Leg tar Sea Beam
SB File utility Contour UWPLOT: The 'uwplot' program plots ship tracks and data from
Summary dd Plots one or many cruises on Mercator projection at scales ranging
Archived utility from world-wide coverage to very detailed surveys (Figure 5).
SB files Tracks can be annotated with a choice of time ticks and
on tape Export SB Plot ta!
SB files File Will date/time labels as detailed as one minute or left unannotated.
on tape Depth, magnetic total field and anomaly, and measured gravity
Hardcopy Plots Archived and free air anomaly may each be profiled or printed along track
(I 6"/deg. scale) SB Plots in any of four colors. Profiles values may be projected to a
Achived all GDC on tape specified azimuth as is most often done for interpretation of
magnetic anomaly lineations. To plot data values with program
Figure 4 luwplot', the input files must be in the 'uwmrg' format. If only
tracks are required, the input source may be a mixture of merge
files, binary navigation 'navbin' files, individual 'abstr' abstract
The Sea Beam merged files are used mainly for input to the files or the Master Abstracts. High or low resolution coastlines
Sea Beam contouring program 'sbcontour' which produces may also be specified.
UNIX standard plot files of Mercator projection maps with Sea Other types of data or annotation, such as station numbers,
Beam contours plotted along the ship's track (across track swath pon names, heatflow values, navigation fixes or plot boundaries
width is approximately 314 of the ocean depth). An input com-
mand file (icf) contains parameters for grid annotation and scale, represented by symbols, lines or text may be added to the plot
contour intervals, etc., as well as optional commands to displace from one or more 'token' files. Each record in a 'token' file con-
different segments of the track relative to each other in order to tains a pen up/down flag; symbol type, height and color, plus
adjust for imperfect navigation. The adjustment process can be optional text height, color and content. Although the 'token' for-
complicated and time consuming depending on the complexity mat is bulky for some types of data (a plot boundary requires 4
of a survey and therefore is not routinely done on the standard records), this disadvantage is outweighed by the convenience of
archived data. For archiving purposes, a set of plots at a scale of dealing with only one such format in the plot program and the
16 inches per degree of longitude are generated for each cruise ease with which such files can be generated by scripts from other
leg. One set of hardcopy plots is produced for the chief scientist sources.
and another retained at the data center. The plot files are also UWPROF: The 'uwprof' program profiles u/W data versus dis-
stored on tape for making future hardcopy pjots,. a.procedure. that tance along the ship's track or versus distance projected to a
has proved more convenient and cost effective than regenerafing specified azimuth. Ship time can be annotated along the x axis
the plot files each time they are requested. and multiple consecutive plots, each having a fixed distance, can

SIO IJ/W DATA PLOT DISPLAYS

U/W PLOTS U/W PROFILES

Mhster prof
Abstracts token u/w data profiles uwproficl'
uyy
UWMrg file(s) vs. ship track put =cPrVmica,ds
file(s) ct> uw
UWPI000f distance
(input command
nzbin Uwplot file) NIX Hardcopy Profiles
plots tracks and/or
i e(s) plot
printed or profiled
09> u1w data file(s) CRT
Alternate
Input files UNI Hardcopy plots
plot
file CRT

Figure 5
,Tvbin@

389

be generated for a cruise leg with one execution.

7. CONCLUSIONS
Although the system for processing data from Scripps
cruises and the management of the multi-cruise data base meets
most current institutional requirements, there is always room for
improvement. Navigation needs to be upgraded from one minute
to one second time resolution and modified to work in the full
coverage GPS environment expected to be operational within the
next few years. Methods should to be developed to resolve the
present conflict between precise positioning needed for Sea
Beam bathymetry and smoothly varying velocities required for.
gravity data. User access to the multi-cruise data base would be
greatly improved by storing the merged data on-line on high
capacity optical disk.

ACKNOWLEDGMENTS
University of California curatorial funds supported S.
Smith for system design, programming and implementation.
Financial support for much of the programming done by J. Char-
ters and M. Moore was provided by the Scripps Industrial Asso-
ciates. Many improvements in processing methods have resulted
from the suggestions of GDC staff members U. Albright, G.
Psaropulos and W. Smith. C. de Moustier critically reviewed the
manuscript, E. Ford formatted this paper and J. Griffith drafted
the illustrations.

REFERENCES

[1] Abbott, J. L., S. M. Smith, J. S. Charters, P. G. Downes, T.
Hylas, R. L. Moe, J. M. Moore, and D. V. Stuber, Scripps Seago-
ing Computer Centers: Real-time Data Acquisition and Process-
ing, IEEE Proceedings 4th Working Symposium on Oceano-
graphic Data Systems, pp 123-129, San Diego, CA, Feb. 1986.

[2] Moore, J. M., J. S. Charters and C. de Moustier, Multi-
sensor Real-time Data Acquisition and Preprocessing at Sea,
Proceedings MTS-IEEE Oceans'88 Conf., Baltimore, Maryland,
Nov. 1988.

[3] Charters, J. S., SIO Sea Beam Software Documentation,
Shipboard Comp. Group, Scripps Inst. of Oceanography, La
Jolla CA, unpublished document, 1986.

[41 The Marine Geophysical Data Exchange Format - MGD77,
KGRD NO. 10, National Geophysical Data Center, Boulder, CO,
1917.

390

DUOMORPH SENSING FOR LABORATORY MEASUREMENT OF SHEAR MODULUS

*Samantha K. Breeding
+Dawn Lavoie

*SYNTEK Engineering & Computer Systems, Inc.
+Naval Ocean Research and Development Activity - Code 363
Stennis Space Center, Mississippi 39526

ABSTRACT the dynamic modulus of solid propellants, but
this new technology has potential for use in
The shear modulus of a sediment is directly other applications and materials. The need for
related to shear wave velocity, an essential an accurate laboratory method for determining
geoacoustic parameter not easily measured in the dynamic shear modulus of carbonate
the laboratory. Further, both shear modulus and sediments under varying pressures led to the
shear wave velocity are a function of the application of the Duomorph technique in
effective stress on a sediment. The objective sediments.
of this research is to develop a simple
laboratory test procedure to measure the shear DUOMORPH SENSING AND LABORATORY METHODS
modulus of a sediment under controlled loading
conditions. To do this, a Duomorph has been The Duomorph sensor, app. 2.54 cm diameter,
constructed similar to ones designed by Briar consists of a thin steel plate sandwiched
et al. (1). Compressional wave velocity has between two peizoceramic crystals with a
been measured using the NORDA compressional metallic strain gage adhered to the center of
wave transducers before and after each crystal (Figure 1). The peizoceramic
consolidation to validate the Duomorph results. crystal is a low power electromechanical
Initial results indicate that the concept is transducer capable of converting electrical
feasible and continued testing is planned. energy to mechanical energy and vice versa.
The application of an alternating current
INTRODUCTION across the individual layers of the
peizoceramic crystals causes one layer to
The United States Navy is interested in the expand while the other contracts. The
measurement of shear wave velocity and shear deflection of the crystals is in a dish-shaped
modulus of marine sediments for acoustic and manner.
engineering purposes. Shear wave velocity and
shear modulus are both a function of effective STRAIN GAGE
stress. To produce results close to in situ lzz=@
values, a laboratory procedure that reproduces
the estimated effective stress state of the in WP
situ sediment should be used, such as PIEZOCERAMIC STEEL +
consolidation testing. Currently, the bender CRYSTALS
element shear wave transducers used by the tp V
Navy are only designed to operate under
ambient pressure. Therefore, Duomorph sensors
designed to measure shear modulus have been STRAIN GAGE
fabricated based on ideas originated by Briar
et al. (1). P POINTS IN THE DIRECTION OF POLARITY
The Duomorphs as originally designed by Briar, Fiqure 1. Schematic of a Duomorph wired in
et al. were used as On situ sensors to measure parallel.

391 United States Government work not protected by copyright

point. The Duomorph is then embedded in
zV sediment and placed in a Consolidation
chamber. As -loads are added, the consolidation
of the sediment is noted according to standard
procedures (2). Duomorph strain measurements
are made after completion of consolidation
under each load increment and used in
calculations to determine the shear modulus of
the sediment at each load increment.

scoy T
DESIGN

Two prototypes were built, one with a .020 cm
Figure 2. Photograph from the oscilloscope of thick steel plate and one with a .008 cm thick
the transmitted signal to the Duomorph (top) steel plate. Both prototypes used 350 ohm
and the received signal from the Duomorph resistance metallic strain gages with a .152
(bottom) after being processed. cm gage length, a .254 cm grid width, an
overall length of .381 cm, and an overall width
The deflection of the Duomorph changes the of .254 cm. They are an "EA" type which
resistance of the strain gages which are measures up to 5% strain. The piezocera.mic
connected to a wheatstone bridge in a strain crystals used were a G-1278 type, fired silver
indicator, in a half bridge configuration. The with a thickness of .279 cm and a diameter of
dynamic output from the strain indicator is 2.54 cm.
amplified, filtered and passed to an
oscilloscope to obtain a photograph of the After the strain gages were carefully adhered
dynamic strain (Figure 2). This laboratory to the crystals, the crystals were adhered to
setup is depicted in Ficlure 3. the steel plate forming a sandwich. Thin
To use the Duomorph, the strain gage reading of wires were carefully soldered to the crystals,
the Duomorph is obtained in air as a reference the steel plate, and to the tabs on the strain

AMPLIFIER OSCILLOSCOPE

OUT
IN CH 1 CH2
f
DYNAMIC
CONSOLIDATION 0U-rPU'r RECEIVED SIGNAL
TEST I
STRAIN
INDICATOR
P- S+ S_ (B our
J DLIOMORPH IN FILTER

GROUNDEDON WAVETEK
SHIELDED
CABLE OUTPUT

TRANSMIFIFTED SIGNAL
Fiqure 3. Diagram of the laboratory setup for obtaining the dynamic
2v\i-

strain from the Duomorph.

392

gages. This configuration constitutes the where:
instrument called a Duomorph. The Duomorph
was then coated with a combination of Resin a is the radius of the -Duomorph, and
184 and 186 to protect the 'wires and gages D is the disk flexural rigidity,
from moisture when embedded in sediment.

DATA REDUCTION OF DUOMORPH DATA hm 3 (EM - E,) + Eh3 (5)
D
Reduction of the Duomorph data follows the 12 (1 - v2)
method devi sed by Briar et al. (1). The wave
form displayed on the oscilloscope represents
the amount of dynamic strain detected by the where: v is Poisson's ratio for the Duomorph
strain gages. This strain was used in the (v =.495 for the Duomorphs used in this
following equation to determine the modified research).
moment ratio:
The shear modulus, G, is calculated from (5):

Mc (e/ea) k, (1) E' (6)
G
Mo 1 k 2 (1 + v)

Where: The shear wave velocity, Vs, is calculated
from (4):
e is the strain in sediment under a load, -1/2 (7)
ea is the strain in air, and Vs = [G
k' is a constant dependent on disk design (for
the Duomorph with a .020 cm steel plate, k
-.664)

METHOD USED TO CHECK DUOMORPH RESULTS

1 1 - 3phz(hz + hm) (2 + Emhm/Ezhz) (2) The values of shear moduli calculated from the
Duomorph data are compared to values obtained
k h2 [1 + (Em/Ez - 1) (hm/h)3 from a standard method of determining shear
modulus to check the validity of the Duomorph
results (4).
where:
The standard method used to determine the
hz is the thickness of the PZT crystals, shear modulus of a sediment involves
hm is the thickness of the steel plate, measuring the compressional wave velocity
Ez is the tensile modulus of the PZT crystals, before and after consolidation. The
compressional wave velocity can then be
Em is the tensile modulus of the steel plate, interpolated for intermediate load increments.
h is the thickness of the entire sandwich, and The specific gravity is determined by a
weight/volume technique (3). From the
specific gravity, the dry density can be
(1/2) (1 + hm/hz). (3) calculated, assuming a fully saturated
condition. After a sample has been
The modified moment ratios are used to consolidated (2), the results of the void ratio
determine the modulus, IM, from a nomograph versus log of effective stress graph are used
generated from quasi-static analysis (1). The to calculate av, the coeffiecient of
elastic modulus, E', is calculated by the compressibility, for each load increment
following equation: (Figure 4):
E' = M D a3, (4) 393 av = Ae / Ap, (10)

4.6e+6.

CM
X 4.4e+6 -
z
CALCULATED
-0.9- u) 4.2e+6 -
0 SHEAR MODULUS
0.8
D 4.0e+6
Q 0.7 0
3.8e+6 -
0 06-
>
DI )OMORPH
0.5- W 3.6e+6
X DATA
Cl)
0.4- 1 i
3.4e+6
1 10 100 1000 10000 10 100 1000 10000
VERTICAL EFFECTIVE STRESS (kPa) VERTICAL EFFECTIVE STRESS (kPa)
Figure 4. The void ratio versus log of vertical Fiqure 5. Graph of shear modulus obtained
effective stress curve generated from from Duomorph experiment compared to shear
consolidation data. modulus calculated from a standard method.

where: where:

Ae is the change in the void ratio between two G is the shear modulus,
load increment, and p is the dry density,
Ap is the change in the pressure between two VP is the compressional wave velocity at a
load increment. load, and
K is the bulk modulus.
The bulk modulus, K, is calculated from av and
the initial void ratio, eo (2,4). RESULTS

1C= av + eo) The shear moduli determined from the
measured Duomorph data and the shear moduli
The shear modulus, G, is calculated from (5): calculated from the standard method using p,
V and K are graphed versus the vertical
G = (p V 2 - K 3/4), (12) p
p effective stress in Figure 5. The values from

Table 1. Listing of values calculated from consolidation test and
compressional wave transducer, tests.

P e av 1C Vp G
kN/m2 1 /kN/m2 1 /kN/m2 m/s N/m2

0.0 1.0398 - - 1629 -
17.1 .9808 3.45E-3 1.69E-3 1630 3,730,000
51 .4 .9528 8.15E-4 4.OOE-4 1632 3,740,000
119.9 .9093 6.36E-4 3.12E-4 1637 3,770,000
257.0 .8584 3.71 E-4 1.82E-4 1646 3,810,000
531.2 .7941 2.35E-4 1.15E-4 1'666 3,900,000
1079.0 .7197 1.36E-4 6.65E-5 1704 4,080,000
2175.0 .6264 8.52E-5 4.18E-5 1781 4,460,000
4605.0 .4892 4.89E-5 2.39E-5 1951 5,350,000

394

Table 2. Shear wave velocities calculated using the shear modulus
from the Duomorph data and the standard method.

Standard Method Duomorph Technique Percent Difference
G Vs G VS G Vs
(N/m2) (m/s) (Nlm2) (m/s) % %

3,740,000 44.6 3,630,000 44.0 2.7 1.4
.3,750,000 44.7 3,630,000 44.0 3.0 1.5
3,770,000 44.8 3,590,000 43.8 4.8 2.5
3,820,000 45.1 3,630,000 44.0 4.7 2.4
3,900,000 45.6 3,810,000 45.1 2.3 1.1
4,090,000 46.7 3,790,000 45.0 7.2 3.7
4,460,000 48.8 4,190,000 47.0 7.5 3.8

-the consolidation test and the compressional modulus and shear wave velocities of
wave velocity tests that were used to sediments under effective stresses of interest
calculate the shear moduli using the standard to geotechnical engineers. The advantage of
method are listed in Table 1. The shear wave the Duomorph sensor is that it can be placed
velocities calculated using the shear moduli directly into a consolidation test, whereas, the
from both methods are listed in Table 2. NORDA bender element transducers for
measuring compressional and shear wave
velocity are limited to testing samples before
DISCUSSION and after consolidation testing.

The values of shear modulus and shear wave
velocity determined from the Duomorph ACKNOWLEDGMENTS
experiment are within 2.3% to 7.5% and 1.2% to
3.8%, respectively, of the values calculated This project was supported by Naval Ocean
from the standard method. The Duomorph data Research and Development Activity funding at
has slight fluctuations in the data. Stennis Space Center, Mississippi. The
Fluctuations occur because the strain wave contribution of Messrs David Young who built
from the oscilloscope isn't always a stable the Duomorphs and John Burns who provided
wave; sometimes, it has a tendency to electronics support is gratefully
fluctuate due to background laboratory noise. acknowledged.
The camera speed was set to 1/15 of a second
in attempt to capture a clear wave. In spite of
the fast shutter speed, the signal photographed REFERENCES
was not always at the same point of
fluctuation; and, therefore, the strain (1) Briar, Herman P., Bills, Kenneth W., and
measurements varied slightly. Schapery, Richard A., "Design and Test of the
Operational In-Situ Gage for Solid Propellant
Surveillance", Air Force Rocket Propulsion
CONCLUSIONS Laboratory, 1976.
(2) Bowles, Joseph E., Engineering Properties
The values of shear modulus and shear wave of Soils and Their Measurement, McGraw Hill
velocity obtained from the Duomorph technique Book Company, third edition, 1986, pp.
are within the range of values determined from 107-128.
the compressional and shear wave velocity (3) Lambe, T. William, Soil Testing for
techniques. Therefore, Duomorph sensing Enaineera, John Wiley & Sons, Inc., 1951,
provides a valid method of determining shear p.15-21.

395

(4) Hamilton, Edwin L., "Elastic Properties of (5) . Timoshenko, S., and Goodier, J. N., Theor
Marine Sediments", Journal of Geophysical of Eta McGraw Hill Book Company, Inc.,
Research, Vol. 76, No. 2, January 10, 1971, p.
581. 1951, p. 9.

396

THE USE OF A TOWED, DIRECT-CURRENT, ELECTRICAL RESISTIVITY ARRAY FOR THE
CLASSIFICATION OF MARINE SEDIMENTS
Dawn Lavoie
Edward Mozley
*Robert Corwin
Douglas Lambert
Philip Valent
NORDA, STENNIS SPACE CENTER, MS 39529
*409 Sea View Drive, El Cerrito, CA 94530
ABSTRACT A prototype, direct-current (DC), electrical resistivity
array which can be towed in the marine environment
was fabricated and field tested in the Mississippi
The direct-current (DC) electrical resistivity Sound in the fall of 1987. RASCL acoustic data were
method has been used for a variety of offshore collected concurrently for comparison of the two
geophysical, geotechnical investigations. Model techniques. Ground truth data were acquired using a
studies indicate that the DC technique is feasible for conductivity, temperature, and depth (CTD) profiler
sediment classification and layer structuring. A for water column measurement and a hydroplastic
prototype array was built to test the hypothesis that gravity corer and vibrocorer for sediment sampling.
such a technique can be used while in an underway The purpose of this paper is to describe the DC array
mode in the marine environment. A 60 m, inverted and the results of the field trials in the Mississippi
Schlumberger array was towed both on and off the Sound.
seafloor with electrode spacings appropriate for a
penetration depth of 10 m below the seafloor. Three BACKGROUND
different bottom types, mud, gassy mud, and sand,
were surveyed in the Mississippi Sound using the Electrical resistivity, the reciprocal of conductivity, is a
array. Ground truth was provided with an acoustic measure of the ability of a solution, solid or mixture to
seafloor classification system, CTD casts, and resist the flow of electric current (2). In most
numerous sediment cores. Data were analyzed sediments, the porosity and salinity of the pore fluids
using SUBVERT, a modification of a University of is more important in determining the resistivity than
California, Berkeley, inversion routine adapted for the mineral grains comprising the sediment. The
an IBM-PC AT. Both the layering structure of the seawater pore fluid of marine sediments provides the
upper sediment and lateral variability of sediment ideal medium for the conduction of electric current
types were adequately defined by the DC technique. With the resistance of the formation inversely related
to the porosity and degree of saturation (3).

INTRODUCTION Archie (4) established a quantitative relationship
between the porosity and electrical resistivity of a
The United States Navy is interested in the problem sandstone. He defined formation factor (F) as the
of remotely classifying and characterizing sediments ratio of the resistivity of the saturated sandstone to
in terms of their physical properties from an the resistivity of the pore fluid
underway platform. Two basic techniques are F = ps / pw, (1)
presently under development, acoustic and
electrical. The acoustic technique has been used where ps is the resistivity of the sandstone and pw is
successfully in a number of areas using the NORDA
Remote Acoustic Seafloor Classifier (RASCL) (1). the resistivity of the pore fluid. From this, he
However, acoustic penetration is limited in regions established the relationship between formation factor
where gassy sediments predominate or where highly and porosity
reflective sediments overly "less reflective"
sediments. In addition, the RASCL acoustic system F = 0-m, (2)
requires at least 3 m of water below the transducer; it
is not effective in shallow-water, nearshore where 0 is the effective porosity and m is a constant
environments. These bottom environments, which that represents the slope of the line depicting the
occur often on continental shelves, prove to be relationship between F and 0 when plotted on log-log
amenable to investigation by direct-current, electrical paper (4). Subsequent workers (5, 6, and 2), have
resistivity technique. In other shelf environments with found that the form of Archie's equation
greater than 3 m of water the two systems provide
complementary seafloor information. F = ao -m, (3)

397 United States Government work not protected by copyright

where a and m are constants related to sediment type MISSISSIPPI SOUND EX PERIMENT
with values close to 1 and 2 respectively, results in
predicted porosities close to experimentally measured
porosities for terrigenous, consolidated and relatively
low porosity rocks with low clay content.
RASCL I I M 1 2 1 A B 1'2 3 N
V V X.,
Because marine sediments generally have lower LAYER 1 54 4 1 1 7
LAYER 2'.'*.*
porosities and pore fluids of higher salinity than
terrigenous sediments, formation factors and absolute @@YER I
resistivities are generally lower in marine sediments
than for similar sediments on land. Measured BASEMENT INVERTED SCHLUMBERGER ARRAY
ELECTRODE SPACINGS: (m) GROUND TRUTH:
resistivities for marine clays, silts and sands overlap to M1N1 3.0 COPIES
some degree; thus there is no unique relationship M2N2 4.4 SUBBOTTOM
M3N3 6.4 CLASSIFIER
between sediment resistivity and sediment type (3). M4N4 9.4 Cm
However, given a reference point in a known bottom M5N5 13.7
M6N6 2GO
type, changes in sediment type are identifiable and M7N7 29.4
can be predicted from measured resistivity. AB/2 0.4
In this study, no attempt was made to determine Figure 1. Illustration of the configuration of the
unique values for a and -m to calculate porosity values towed, inverted Schlumberger array used to
for sediments within the Mississippi Sound. Instead, measure sediment resistivity in the Mississippi
initial porosity calculations made from measured Sound.
apparent resistivities are based on the formula
1.2 F -.69 (4) re.lieve the connecting wires of towing stress and
minimize cable stretch.
It was developed by Boyce (6) as a generic form of
Archie's Law and is based on porosities measured on The current transmitter system included an AC power
sediments from the Bering Sea with a composition source, a power supply and control unit and a switch.
similar to much of the sediment in the Mississippi The AC power source was a 6 kW gasoline-powered
Sound. generator that provided 110 VAC output. The power
supply and control unit included a Variac variable
transformer, a voltage multiplier circuit that stepped up
ARRAY DESCRIPTION the input voltage to a maximum of about 800 VAC and
a rectifier circuit that transformed the AC voltage to DC.
The prototype DC system included an electrode cable, Maximum output of the unit was 800 VDC or 4 amps
a current transmitter, and a data acquisition system. DC or 2 kW, whichever was reached first.
Modeling results of an earlier study indicated that the
electrode cable in an inverted Schlumberger The DC output of the control unit was fed into a
configuration could be used to resolve subseafloor normally open, high capacity, manually operated
layers and sediment properties to a depth about 1/3 to double pole, double-throw, switch that was used to
1/4 the array length with a resolution of about 1 m (3). reverse the polarity of the DC current output to the
Since the longest vibrocore that could be recovered current electrodes.
was 3 m, the array length was constructed so that the
maximum electrode spacing was less than 30 m Output current, electrode potential and pressure
(Figure 1). The electrode cable consisted of one pair transducer output were measured with liquid-crystal
of current electrodes for transmitting electrical current display (LCD) digital multimeters (Beckman Industrial
into the sea water and seven pairs of potential model DM800). Resolution on the 200MV range was
electrodes for measuring voltages generated by this 10 microvolts and accuracy was +/- 0.05% of the
current. The electrodes were arranged in an inverted measured value. Current measurements were made
Schlumberger array. The distance AB/2 from the array in the 10A range, with a resolution of 1 mA and
center to the center of each current electrode was 0.4 accuracy of +/- 0.75% of the measured value. Data
m. The corresponding distances, MN/2, for the rate was 2 points per second.
potential electrodes were 1.5, 2.2, 3.2 4.7, 6.85, 10.0
and 14.7 meters from the array center to the outer tip of
each electrode. This roughly logarithmic electrode
spacing resulted in equally spaced data points on the FIELD METHODS
log log plots used for data interpretation. The cable
also included a pressure transducer so that standoff The field platform was the FIN KIT JONES , a 20-m,
distance from the seafloor could be determined. wooden hull, diesel powered vessel. The resistivity
array was towed directly behind the ship with the
The cable jacket was abrasion resistant polyurethane center of the array approximately 65 m behind the
plastic. A kevlar stress member was incorporated to center of the ship. This towing configuration allowed
7YER MlCL'@@

398

the cable to be towed along the seafloor in shallow data, it is relatively easy for an experienced interpreter
water, and within 2 m of the seafloor in water over 6 or to roughly estimate layer parameters by inspection of
7 m deep. the plot of apparent resistivity vs. electrode separation;
however, for seafloor resistivity data, the presence of
In addition to the resistivity system, ancillary equipment the overlying water makes it much more difficult to
included LORAN C navigation (Internav) , a CTD obtain such initial estimates by inspection of the field
system (SeaBird Electronics, Seattle, WA.), RASCL data curve. Because such estimates are required for
subbottom classifier (Honeywell ELAC, Kiel, FRG.) for input to the inversion routine used, an empirical
acoustic measurements and a vibrocorer for bottom stripping procedure was devised for removing the
sampling. Both LORAN systems were checked effect of the water layer and transforming the seafloor
Foritinuously to verify the ship's position. The RASCL apparent resistivity readings to those which would
is a high resolution seismic system which has the have been observed had the water layer not been
capability to measure echo strength both quantitatively present. The basis for this process is the assumption
and qualitatively in ten time windows. For this that the seafloor and the sea water act as resistors in
experiment, the system used a 15 kHz transducer and parallel, so that the measured apparent resistivity is
was adjusted so that each time window corresponded equivalent to the parallel combination of the seafloor
to a sediment depth increment of 0.4 m. The system and sea water apparent resistivities. Since the
applies algorithms based on multilayer acoustic theory apparent resistivity of the water layer can be calculated
to compute acoustic impedance for the ten depth from its known resistivity and thickness, the "stripped"
increments in the seafloor. This impedance profile is seafloor resistivity values are easily calculated. (7). An
then used to predict sediment structure and various inversion routine, SUBVERT, is used to determine the
physical properties including porosity (1). The CTD layered structure corresponding to a given set of
data and cores were used for ground truth both to resistivity values for Schlumberger and Wenner
calibrate the RASCL and compare with the direct expander arrays. The primary limitation of SUBVERT
current interpretations. is that a maximum of four layers can be resolved. This
usually includes the water layer above the array, one
or two subbottom layers, and an infinite half space
DATA REDUCTION AND INTERPRETATION basement.

Several computer programs written at the University of
California, Berkeley, have been used to reduce the FIELD RESULTS
measured field data (7). SUBRED is used to calculate
apparent resistivity values from measured field data for Three test sites in the Mississippi Sound were chosen
each potential electrode pair. For onshore resistivity for their differing bottom types (Figure 2). Results from

05' 890 55' 50' 45'

BILOXI

GULFPOFTr
t
N
DEER ISLAND SITE
20' (SANDY CLAY)

MISSISSIPPI SOUNO

15'
loc.

HORN ISLAND SITE
.2 .8 TIRIANGLE (MED. SAND)
.7 (GASSY MUD)
30c10' i _____j

Figure 2. Location map of the sites surveyed in the Mississippi Sound.

399

STATION 6

% (by weight)
0 10 20 30 40 50

0
U

5
Z
-762

A
4 j",
-P
U)
E

. . ... -1.73

Figure 3. RASCL record of Station 6 with several
subbottom reflectors within 4 mbsf. Note flat echo
strength lines indicating the system had "pegged out" -2.62
due to shallow water depths.

three stations within the Deer island and Horn Island Coarse sand
sites are reported here. Results from the Triangle will Medium sand
be presented in a later paper. In Fine sand
0 v. fine sand
The Deer Island site is a shallow water, sandy-mud 0 silt
bottom environment. The offshore barrier islands, 10 clay
including Horn Island, in the Mississippi Sound are
migrating westward over a mud bottom. The Horn Figure 4. Results from five discrete grain size
Island Site was chosen because bottom sediments analyses, Station 6, Deer Island, Mississippi Sound.
were expected to be sands to an unknown depth, and
it was hoped that the resistivity array would delineate a
less reflective layer below the surface sediments. A comparison of core and array results is presented in
Figure 5 as porosity plotted against depth in meters
Station 6, 3.5 krn south of the western end of Deer below the seafloor mbsf.
Island (Figure. 2), is known from core analyses to be Station 9 is located within the Horn Island site (Figure
fairly representative of the Deer Island stations. Water 2). Water depth (CTD) was 7.3 m. The RASCL
depth is known from CTD casts to be 2.9 m and the analog. record shows good acoustic penetration, about
sea floor is relatively flat, as can been seen from the
RASCL profiles (Figure 3). Because the water depth
was less than the required 3 m below the transducer, TABLE 1: STATION 6
the amplitude of the echo strength returns was --------------------------------- --------------------------------------
clipped." RASCL was unable to compute an accurate layer Pa z F
acoustic impedance and thus not able to predict
sediment type or porosity at this station. Avisual (ohm-m) (m) W
description of the core indicates the sediment to be a
dark grey, organic, sandy mud that is fairly
homogeneous down hole. Grain size analyses done Underway: 1 0.659 0.44 2.56 62.69
at five discrete intervals down the core are presented 2 0.487 4.04 1.89 77.25
in Figure 4 and indicated sandy clay at the surface, 3 2.300 - 8.54 27.32
and clayey sand to fine sand below.
Stationa[y@ 1 0.665 0.52 2.64 61.43
A listing of apparent resistivity values (Pa) and layer 2 0.482 4.08 1.91 76.70
thicknesses (z) derived from SUBVERT, formation 3 2.150 - 8.53 27.43
factors (F), and porosities calculated using Boyce's (6)
formula for resistivity data acquired while underway
and stationary is presented in Table 1. --------------------------------------------------------------------------

400

TABLE'2: STATION 9 STATION 6

------------------------------------------------------------------------
0-
law Pei z E 0
(ohm-m) (M) M -1
Underway: 1 0.212 1.310 core measurements
2 0.739@ 0.642:,. 3.47 50.83 2-
3 1.000 - 4.74 41.03
E
Stationa[y: 1 0.651 0.537 2."89'. 57.73 -3-
2 0.725 5.79 3.22 53.60 underway
3 1.050 - 4.66 41.47 .4-- stationary
-4-

--------------------------------------------------------------------- -5-

-6
4 m, into a fairly "soft" bottom with a( dipping reflector at 20 30 40 50 60 70 80
about 2 meters below sea floor mW (Figure 6). This Porosity (%)
location is assumed to be a fill channel. Echo strength
return indicates the sediments to be silty sand over a Figure 5. Values of porosity measured from Core 6
fine sand. Actual grain size distribution for three and calculated for underway and stationary resistivity
discrete samples is presented in Figure 7. Table 2 lists measurements, plotted against depth in meters
apparent resistivities (Pa in ohm-m), layer depths in rn below the seafloor.
(z), formation factors (F), and calculated porosities (0)
in percent using Boyce's formula (6). A compadson of
RASCL interpreted porosities, array interpreted . . . . . .
porosities, and measured porosities from the Station 6 0
OvtF 0,
core are presented in Figure 8.

Station 10B is located within the Horn Island site but
away from the channel filled area of Station 9. Station
10B is located on a sloping bottom in 9.54 rn of water.
RASCL penetration is 2 m or less and echo strength
returns indicate a "harder" or more reflective bottom
than Station 9 (Figure 9). RASCL bottom
characterization is a silty sand over a finfe sand.
Vibrocore recovery was poor with about 0.6 mi of very
clean, medium sand retrieved. Laboratory measured
porosity, RASCL porosities, and underway and
stationary array calculated porosities are compared in "41 4
Figure 10. The data from which DC porosity was
calculated are presented in Table 3.

TABLE 3: STATION 11013
------------------------------------------------------------------------ Figure 6. RASCL record of Station 9. Penetration is
19= Pa z F 0 about 4 mbsf. Echo strength return indicates a "soft"
(ohm-m) (M) M bottom with a dipping reflector at about 2 mbsf.

Underway: 1 .216 9.10 water
2 .243 0.43 DISCUSSION
3 .832 3.85 47.32
Station 6 was an ideal site to test the DC array
Stationary: 1 .227 10.2 water because the water was shallow, the array stayed on
2 .847 00 3.73 48.37 the seafloor, andAhe water-sediment interface and
subbottom layering structure were fairly horizontal.
------------------------------------------------------------------------- The RASCL record indicates subbottom reflectors at
ore easuremen
m ts

derway
4-- stationary

401

STATION 9 approximately 1, 2, and 3 mbsf. The existence of these
% (by weight) layers is substantiated by core analyses. Porosity
0 20 40 60 80 measurements generally increase to about I mbsf,
decrease from 1 to 2.5 mbsf and then increase to the
bottom of the core at 3 mbsf. Porosity interpreted from
array data suggests an increase in porosity between
the surface and 1 mbsf, then a layer of lower porosity
.127 to about 4 mbsf. Underway and stationary data are
both good at this site. The number of layers that can
be resolved are limited by the inversion software to
four. Because these constraints exist, information
U)
M below 1 rn has been averaged to provide the
E interpretation of a surface layer to 1 mbsf, a deeper
layer of higher porosity than the surface layer, and
CL .94
CD then a basement layer of much lower porosity than
either of the upper layers.

Data from station 9 presented some difficulties in
interpretation of the underway data because the array
was off the bottom. The final inversion of data resolved
1.57 the water into two layers, 7.03 m and 1.31 m, the array
presumably riding 1.3 m off the bottom. Again with the
software limitations, only two sediment layers were
able to be resolved: a surface layer of about 0.6 m of
sediment of 50% porosity and a basement layer of
M coarse sand unknown thickness of 43% porosity. Stationary data
0 medium sand- provided more resolution of the subbottom since the
IM fine sand array was actually lying on the bottom. However,
* very fine sand interpreted values of resistivity and porosity were
* silt significantly higher than core, RASCL, or underway
I M clay interpretations (Figure 8). Stationary data was taken
with the ship rolling at anchor. Generally the signal to
Figure 7. Results of grain size analyses from three noise ratio and data quality was poor under these
discrete samples from core 9. Depth is in meters conditions leading to less reliable interpretations.
below the seafloor. Legend indicates grain size.

STATION 9 0
-n
'2
Underway array 1.3m 4e,
f bollo *aler
m;
resolved Into 2
layers:
CTD
rho -M (U)
.2249 ohm
o 2 2119 ohm-m 7.03M 7.32M 7.3m 4

1.31m V V seaflo,,
:A el,@7 @-,@OA
Depth Co.
(mbst)
RASCLV (U) (S)
11 71111 11, 11

10 W
30 40 60 so 70
_77
Porosity (nol to scale) 77777- 77-

Figure 8. A comparison of RASCL interpreted Figure 9. RASCL record from Station 10B. Note that
porosities, measured core porosities, and underway echo strength return indicates a more reflective
and stationary array interpreted porosities. bottom than at Station 9. Penetration is only about 2
Underway data indicates stand off distance from the mbsf at Station 10B.
bottom.
OEM

7(u)

402

Schlumberger array was able to retrieve resistivity
data of sufficient quality to resolve water depth and
Station 10B gross layering structure within the subbottom. The
resolution and bottom layering structure is limited by
I Sea surface the inversion software used, not by the DC array or
undmy army 1.14 rn configuration.
depth 10.2m off t@ottorn; water
We'*, (S) resolved Inlo 2 layers:
W 9.54M 8.38m hO@21)):.216 Ohm-nn Array data can give resistivity values which can be
;ho .243 ohm-nn converted to formation factors that accurately indicate
different sediment types. It is necessary to make a
judgement call and use formulas relating formation
Wth _DJ SeAfloor factor and porosity for particular sediment types. A
(mbso A2 general, non-unique formula such as the Boyce
C re 4!@L) 1
@.3 (33,9%,@ (35%1S formula will lead to increasingly large errors as
.4 13.4
4.4 % MSCL (54%) sediment types grade into medium sands with little
0.5, clay or silt content.
0.67@ 30 40 50 60 70 It is evident that it is not necessary to take stationary
Porosity (%) (not to wale) data since the underway data are as good or better
than 1he stationary data. This is possibly because the
ship rolled considerably while at anchor and the signal
to noise ratio was higher than while underway.
Figure 10. A comparison of RASCL interpreted
porosity, core measured porosities, and underway The towed array is ideally suited for shallow water use;
and stationary array interpreted porosities. The array with a thinner water layer to resolve, resolution within
Was off the bottom while underway; the water is the subbottom becomes better. The DC resistivity
resolved into 2 layers. technique provides valuable complementary data
when used in conjunction with an acoustic
classification system.
The interpretation of array data from Station 10B was
frought with similar problems because the array was
not only off the bottom, but the bottom was sloping, ACKNOWLEDGMENTS
thus the array was not parallel to the bottom. With the
water resolved into two layers equalling 9.54 m of
water, the same water depth as measured by CTD, the Support for this study was provided by the Office of
underway array data provided us with bottom Naval Technology, Oceanographic Support Block,
formation factors of 3.85, very close to those reported Project RM35685. The authors wish to thank the crew
by other workers for sands (8). Similarly, with the of the R/V Kit Jones for field support and Frank
stationary data, a formation factor of 3.73 was Carnaggio and John Burns of NORDA for excellent
calculated. Using the Boyce formula, porosities of electronics support.
47% and 48% were calculated, considerably different
from measured core porosities. Using a formula
derived by Beyer (8) for quartz sands, F = 0-1.26, REFERENCES
porosities of 34.5% and 35% are calculated for the
Station 10 sediments. It is therefore important to 1 . Lambert, D.M.: "A New Computerized Single
realize that formation factors are indicative of sediment Frequency. Seaf loor Classification System", in: Current
type. Appropriate formulas relating porosity to Practices and New Technology in Ocean Engineering,
formation factor for particular types of sediment must eds. G.K. Wolfe, and P.Y. Chang, ASME, 1988, pp 99-
be used. RASCL-interpreted porosities are 105. .
considerably different from measured porosities. The 2. Erchul, R.A.: "The Use of Electrical Resistivity to
correlation factor used in RASCL algorithms to Determine Porosity of Marine Sediments", Ph.D.
correlate acoustic impedance and porosity is not as Dissertation, University of Rhode Island, Kingstown, Rl,
well defined for clean sands as for sediments 1972, 86 p.
containing significant amounts of silt and clay.
3. Valent, P.J., E.D. Mozley, and R. F. Corwin, "Rapid
Underway Sediment Classification by Electrical
Methods", NORDA Report 211, Stennis Space
CONCLUSIONS Center, MS, 39529, 1987, 72 p.
4. Archie, G.E.: "The Electrical Resistivity Log as an
The use of a towed, DC array to delineate subbottom Aid in Determining some Reservoir Characteristics":
layering and characterize sediment is highly feasible Am. Inst. Mining Metall. Petroleum Engineers Trans.,
in the marine environment. The inverted v. 146,1942, pp 350-366.

403

5. Winsauer, W.O. and McCardell, W.M.: "Ionic 7. Corwin, R. F.: "Model Study and Computer
Double-Layer Conductivity in Reservoir Rock": Am. Programs for Modeling and Interpretation of Seafloor
Inst. Mining Metal[. Petroleum Engineers Trans., v Direct-Current Electrical Resistivity, Measurements",
198, 1953, pp 129-134. prepared under Contract No. N00014-87-C-6012,
NORDA, Stennis Space Center, MS, 39529, 1987,
6. Boyce, R.E.: "Electrical Resistivity of Modern 104 p.
Marine Sediments from the Bering Sea", Ph.D. 5. Beyer, J.H.: Unpublished report retained by R.F.
Dissertation, San Diego State College, San Diego, Corwin, 406 Sea View Drive, El Cerrito, CA, 94530,
CA, 1967,172 p. 1971.

404

SEDIMENT CONTAMINATION BY HEAVY METALS AND HYDROCARBONS

P.F. WAINWRIGHT', B. HUMPHREY2, G. STEWART3

1. Seakem Oceanography Ltd. P.O. Box 2219,2045 Mills Rd., Sidney, B.C., V8L 3S I
2. Seakem Oceanography Ltd., 28 Block C, St. Michael, Barbados
3. Canada Department of Indian and Northern Development, Yellowknife, N.W.T.

monitoring programs may suffer from changes in design
ABSTRACT or inconsistencies in sampling and analytical methods.
Results of the analysis of contaminant data from marine sediments at Complex statistical procedures may be required and
exploration shoTebases collected between 1982 to 1984 by the Beaufort specific hypotheses of concern may not be testable. To
Sea Shorebase Monitoring Program are presented here. Analyses facilitate use by regulators, it is desirable that statistical
determined that the dataset was not suitable for conventional statistical analysis of monitoring results be robust but easily
methods and that inter-laboratory variance was confounded with interpreted. This paper presents the analysis of data from
variance between years. A method was developed based on relations the Beaufort Sea Shorebase Monitoring Program and
between sediment grain size and contaminant concentrations. Stations describes a method which was found to be both robust
where contaminant concentrations exceeded the relation's prediction
interval were identified as outliers and an index of contamination based and easily interpreted.

on the magnitude of the departure from the prediction interval was
used to display their distribution on maps. MONITORING

INTRODUCTION The Beaufort Sea Shorebase Monitoring Program was
initiated in 1982 in response to both public and
The Department of Indian Affairs and Northern government concerns regarding oil and gas related
Development (DIAND), as the lead federal agency, has industrial development along the Tuktoyaktuk Peninsula
the legislated mandate (the DIAND Act) to manage (Figure 1). The program was developed and funded
Canada's northern resources. This involves promoting jointly by government and industry. The goal of the
development of resources while ensuring proper program is to provide a time series sampling of
environmental management, which is particularly concentrations of metals and hydrocarbons in marine
important in the arctic because of the role that renewable sediments near shorebase developments, and to monitor
resource harvesting plays in the life style of northern possible changes in concentrations of these substances.
natives. Monitoring is considered an integral part of the Field sampling was carried out in 1982, 1983, and 1984.
environmental management process. It plays an essential Through the Northern Oil and Gas Action Program
role in providing feedback to the decision-makers about (NOGAP), the samples from all three years have been
the appropriateness of environmental controls and analysed, and data reports documenting the sampling,
mitigative technology. Monitoring programs may also analytical techniques, and analytical results have been
create an information base - in this case, contaminants produced (Thomas et al. 1983, Arctic Laboratories
associated with support bases for offshore oil and gas Limited 1984 and Nuclear Activation Services Limited
exploration and development. Howeveri long term 1986). Summary reports have been prepared for the 1982-

CH2585-8/8810ooo- 405 $1 @1988 IEEE

134" W 130, w

+ lig /120
+ lie
+ 117 /123
++ 122 121

114 70"N
109 190 a
+ 88 B7 I0B,+I + H2 4@10 McKINLEY
+ 89 + BAY
+ + 107 + III
+90 7- TUFT POINT
(,Zr +91 +7 2
82, +83 /3 UHUTCHIS@OON
BI- 4@ 31 BAY CQ3 <2
C> I-- + 31

\ib TOYAKTUK
0 HARBOUR

pprox. 25 km. I

Flpre Ia. Locations of 1982 - 1984 stations, Tuktoyaktuk Peninsula nearshore zone.

133-W 132-30-W 132-W

TU.TOYAKTUX
HAMLET OF 17
14 @
DEW 15 +54
LIN lo"
I

wc +a +is
ONE 21 TUFT POINT Ai +41 +4
CAN.AR ESSO IT? HUTCHISON
GULF4 +35 1 '51 SA@
BE IL + 36
+

+20 +33
+ 29
.pp.. (17 M--.,
Fill- Ili. Locefloom of 1992-1984innions, Tktyakwk Hwbo- RpmLd. Locsti- of 1992 -1984 nations, HutdAmin Bay and Tuft Folm.

131. W
+?I

HEAMICHEL
+70 ISLAND UH
++,`@ 1.2
\Ili\. + 101
+0 - 70-M 96-+ +-R
+ 91 + IOD
+79 +j" \gs
+n +78 495
+50 76'+0 STOK
P
+ POINT
+ III So 62 C;, VLAND+ 75+
74 +75+

64 +73

+72 94
+66 KING + 93
POINT

appros. 5 Ion,

Film le. Locations of 1982-1984 slatimem, McKinley Bay. FIgarete. [Locations of 1992-1994 stations, Yukon Coast (Henchal Basin
new King Poim@

406

83 data (Yunker 1986) and the 1982-83-84 data sediment collected. Therefore, statistical analysis was
(Wainwright and Humphrey 1988). A similar monitoring restricted to observations with concurrent grain size
program was carried out in the nearshore Beaufort Sea determinations, resulting in exclusion of about 40% of the
area in 1984 (Can Test 1984). The four data sets include contaminant observations.
samples from 124 stations.
Relationships between contaminant concentrations and
DATA ANALYSIS grain size were then determined (Figure 2) and the
residual values from the contaminant to grain size
Yunker (1986) determined that the 1982-83 data were regression relations were analysed. The relationships
unsuitable for detailed parametric statistical analysis were better for some contaminants than for others (Table
without applying data transformations and other 1). It was apparent, by inspection, that the residuals were
manipulations. Differences in sampling design, and normally distributed for each laboratory and year
sampling and analytical methods between data sets often analysed, implying that remaining variance was caused by
limit possible approaches to statistical analysis of data. intra- and inter-laboratory analytical variance. This was
Individually and in combination, the data are often subsequently shown using statistical tests. An example is
neither normally distributed nor homogeneous. For this chromium, an element which adheres closely to the grain
study, the entire data set was examined to identify an size relationship. A histogram of the residuals plotted by
approach or methodology that would permit laboratory (Figure 3) indicates that residuals from
simplification of statistical analysis and interpretation. individual laboratories, and therefore years, appear to be
Data may be analysed in relation to a reference data set if normally distributed.
the study data set and the reference data set represent the
same population. The pooled data may then be examined TABLE I
Coeffleleals of determination from
for outliers (anomalous values) relative to some I..t q.am revv"Jon analygis Ron" percentage clay
characteristic and the residuals examined for the effects of C@Mclmt
of
determinn ion
year and location. The approach adopted here was to Contaminant (0) t
assume that the data belonged to a single population. This Barium 603
population was then characterized and examined for Cadmium 60.9
Otornium 78.7
Copper 769
deviating observations or outlieTS.No reasons were found Lead 25'.8
Mercury 74.6
Nickel 73A
to discount the assumption of a single population when S.m Mkanes 31'7
Surn PAHs 32.8
using the Shorebase data and the nearshore data. Zim as

9
Previous workers had observed a relationship between 9
II Y-a75U-2&8
sediment grain size and contaminant concentration f A0787
R 'a0 -356
(Adams et al. 1980, Dossis and Warren 1980), in samples o -R
y
A' E A
A@ E a
from the Canadian Beaufort Sea (Yunker 1986, Hoff and OF J R A
RI R al E A .6%
NN j A u
Thomas 1986, Arctic Laboratories Limited and LGL Ltd. R
A
1987) and in the Alaskan Beaufort Sea (Boehm et al.
1985, 1986). We examined the available data sets for this
an an 100
relationship and found that the contaminants of interest x ew
were determined primarily by the grain size of the

407

Samples with contaminant concentrations in excess of the OUTLIERS BY CONTAMINANT
value predicted by the grain size relations plus a 99%
prediction interval were then identified. Maps of outliers Outliers were defined as observations with a magnitude
were plotted for each contaminant (e.g. Figure 4). either higher or lower than the range of magnitudes which
Inspection of the resulting maps identifies an association statistically characterized the distribution of the data. The
between industrial activity and sites of outliers. Many of range of magnitudes which characterized the data
the outliers identified are associated with loading docks or distribution may be calculated as a confidence interval or
sites of dredging activity. prediction interval.

For each contaminant, the degree of contamination was
calculated as:

Di @ (Co - Cu) Cm
@J: where:
Di = degree of contamination for contaminant i;
Figuml Frequency diaribuden, of mid,ah for ch-niu, (Cr), Co = concentration observed;
Cm= mean concentration of contaminant i from the
+ *+ data used to determine the grain size relation;
p and
+ Cu upper boundary of the prediction interval from
the grain size relation calculated as follows:

+ tcx/2 * s [ I + 1/n + %clay - %claym )2
+ CU CP
+ SS%clay
S2 =(SSC-jo@-SSC%clay)/(n-2)
The environmental significance of anthropogenic inputs and
was assessed by determining an index of contamination,
which was mapped in the same way as the contaminant Po + (A, -%clay)
outliers. This index of contamination is the sum of the CP
deviations from the grain size relations for each where:
contaminant. At the outset of the data analysis it was
envisioned that an environmental risk index would be
determined. However, risk indexes were not determined po and #1 are the regression coefficients;
as a relatively arbitrary "toxicological factor" would have %Clay = percent clay observed;
to be determined for each contaminant. The maps of %Claym = mean percent clay from the data used to
determine the grain size relation;
contaminant deviations and index of contamination were
n = the number of observations on which the
sufficient to identify and interpret areas of risk. The
grain size relation is based;
question of what magnitude of contaminant index should s2 = estimated variance of the random error;
cause concern was not addressed for the same reason.
SS%Clay sum of squares for percent clay;

408

SSC sum of squares for the contaminant; reference areas were produced in which the type and size
SSC%Clay =sum of squares of the cross-product: of symbols represent the magnitude of the index of
contaminant x percent clay; contamination. In these maps a circle presents a value of
t,Z/2 = critical value of t for probability a; and 1>0, where the size of the circle is related to the
CP = concentration of contaminant predicted magnitude of 1, and crosses indicate values of I=O.
by the regression relation. Stations are labeled with the contaminants considered as
outliers.
All values of Di less than zero are considered to be zero,
as a negative value for Di requires that the concentration The maps demonstrated that the majority of the observed
observed is less than that predicted by the grain size outliers were found in Tuktoyaktuk Harbour and
relation. The index of contamination is then calculated by McKinley Bay. There were also some outliers in
summing the values of Di for all contaminants measured Kugmallit Bay (cadmium) and Hutchison Bay (copper
for that sample: and mercury). The pattern of outliers for McKinley Bay
and Tuktoyaktuk Harbour were different.
I Di)
i Simple deviation from expected values does not
The resulting index, 1, will be zero if the concentrations of necessarily imply anthropogenic input. For hydrocarbons,
all contaminants observed are less than, or equal to, the the patterns of specific compounds must be examined. If
prediction interval for probability the CPI is high, it is likely that the source was biogenic,
not anthropogenic. The compositional pattern of the
Two values of a were considered and evaluated, 5% and alkanes and PAHs as determined by gas chromatography
1%, which correspond to 95% and 99% prediction should be examined for all hydrocarbon outliers.
intervals. Figures 5 and 6 present maps of the index of ENVIRONMENTAL SIGNIFICANCE
contamination derived from 95% and 99% prediction
intervals respectively. The 99% prediction interval was The degree of contamination (D) and index of
selected because it gives results which are consistent with contamination (I) values greater than zero indicate that
expectations (that the foci of contamination are the active observed concentrations were in excess of those predicted
shorebase areas), because falsely concluding that a station on the basis of sediment grain size. It cannot be assumed
is contaminated is undesirable, and because departure
from the prediction interval does not necessarily imply that values greater than zero result in negative. effects on
the environment, but rather that values of I greater than
environmental impact. It is preferable to adopt the more
conservative value. zero imply some risk of environmental effects.

The geographic distribution of the index of contamination Thomas et a[. (1986) reviewed information on ecological
(I) (based on the 99% prediction interval) is presented in effects of contaminants and concentrations of concern.
map form (Figures 7 to 10). Mapping was performed Concentrations of concern are generally available for
using ESLMap, a software mapping package developed by water column exposure but are limited for sediments. In
-ESL Environmental Sciences Limited. This software Canada, the only criteria for sediments are from the
allows the data to be displayed interactively on detailed Ocean Dumping Control Act (ODCA) applicable to "bulk
maps of variable scale, with symbol sizes and types wastes" and thus to sediments when disposed pursuant to
determined on the basis of criteria defined. by the the Act. The ODCA specifies limits for only mercury,
operator. Detailed maps of the embayments and cadmium and petroleum-related compounds: cadmium -

409

132'. Ise.. J32- W

0-
0
+

p7, 70-N 70- N
0,+ +1 44+ + 'a .
+ ++ + + Q7 Is
+ + + ++ + ++
+
A 0
Q7 + C?
to A
0 Is
16
o 16

pp-. 50 lan. aps,a,. 50 lort.

Ftipars S. Map of distribution of Index of contannination (1) hued on 95% Flipin 6. Map of distribution of index of contarnination (1) bound on 99%
lnedietion intemal. pdiction 1.1-1.

Is,- w 153-W

L-st + *+ q
zo,
Ph 0-

+ Cu.Cd.ln+
a It 4@# `+ + OrApb
+ Alt
+ + C 'A

+ Go' 30' N
o + 0
+.: .11M. "dic- C1. Alt +
A I a.- Alk
0 DO. Alt a. ho'l- 0: ba'Wra
Cd d.l.. eadral..
C, C,
Car'Call In -.I..
+ C -r + Cwpor
No anarcur
NI fa
PAW =" . PAN: 4wan PAN
PAM
Ph load 'ad
,Ppmx, 5 Um. Z. up-.. 2 kon, sine

Fightell. Fillorre 7. map or distribution of Index of ecrataraluation (1) for Tuktoyakluk
HA,b-,.

132- 301 IN 132-W 134- W

+
+

+ 0 CA +

+ 0 Cd

+ + 0
+ + + + Wind- ix.d'.-Aths + + pradl. W
+ A": Zon -1 :lk: sous I Ch"ah
on sodral..
usurer. 5 Am. Cd ..d.A. n. Act,
cr Cr eltraml,
a cu.
+ c cup;p., GV40-N /C' jCd No . eapps,
No ashr-, ans- 10 AM. ahar@
Air atk., PSI .1
PAN PAN: Mwe" PAN
Ph 0
Zn 4 n
Fliparell. M.frofdistribution of Index ofconlannination (1) for Hutchison Bay. nitsumlit Mpofdistrlbudonorinduofmmffdmdm([) Far KuSchallit Bay.
+ ++0-
+ + 0
+ Ask-

Cd
on
u

tharn PAN
C

land
Ana
PA
40 n

410

0.6 ppm, mercury - 0.75 ppm, petroleum-related does not indicate that environmentally significant changes
compounds - 10 ppm of hexane soluble substances. to the sediments have occurred, but that some detectable
changes were occurring in the two major shorebase
No concentrations of mercury in our data approached the embayments.
ODCA limit. Sixty-one observations of cadmium
exceeded the ODCA limit with a maximum observed -REFERENCES
concentration of 1.01 ppm. The upper boundary of the
95% prediction interval exceeded the ODCA limit of 0.6 Adams, D.D., DA. Darby a nd RJ. Yuong. 1980. Selected analytical
ppm at a value of about 38% clay. It must be expected techniques for characterizing the metal chemistry and geology of
that cadmium concentrations will exceed the ODCA limit fine-grained sediments and interstitial water. In: R.A. Baker
(editor). Contaminants and Sediments, Volume 2. p. 3-28.
where the sediments are of fine texture. This result
suggests either that there may be an environmental risk Arctic Laboratories Limited. 1984. Beaufort Sea coastal sediment
associated with fine sediments in the Beaufort Sea or that reconnaissance survey: A data report on 1983 geochemical
the ODCA limit for cadmium is not applicable (i.e., is too sampling. Report for Environmental Protection Service,
Yellowknife, NWT; (unpublished manuscript) 39 p.
low). With respect to the limit for petroleum-related
compounds, 244 of 249'observations of hexane-extractable Arctic Laboratories Limited and LGL Limited. 1987. Beaufort sea
compounds exceeded the ODCA limit of 10 ppm. ocean dumpsite characterization. Report for Environmental
Protection Service, Yellowknife, NWT; (unpublished manuscript)
100 p. + appendices.

DISCUSSION Boehm, P.D., E. Crecelius, W. Steinhauer, M. Steinhauer, S. Rust and
J. Neff. 1985. Beaufort Sea monitoring program: Analysis of trace
The method developed in this project was easy to use and metals and hydrocarbons for outer continental shelf (OCS)
activities - Year I results. Report by Battelle New England
will be easy to maintain over time. The data are held in a Research Laboratory for Minerals Management Service, Alaska
dBase IH+ data-base, and the calculations are performed OCS Region; (unpublished manuscript) 162 p.
using a spread-sheet program such as Lotus 1-2-3. The
method clearly indicates deviations from expected Boehm, P.D., E. Crecelius, W. Steinhauer, M. Steinhauer and C.
Tuckfield. 1986. Final annual report on Beaufort Sea monitoring
contaminant concentrations in Beaufort Sea nearshore program: Analysis of trace metals and hydrocarbons from outer
sediments. Any future determinations can be added to the continental shelf (OCS) activities (Contract No. 14-12-0001-30163).
data-base and, if they do not represent outliers, can be Report by Battelle New England Research Laboratory for Minerals
included in the baseline set of data. Any new data which Management Service, Alaska OCS Region; (unpublished
ript) 238 p.
deviate significantly will be flagged by a value of the index
of contamination greater than zero. Can Test Ltd. 1985. Chemical analysis of samples collected for
Beaufort Sea nearshore monitoring program, 1984. Report for
Environmental Protection Service, Yellowknife; (unpublished
The method appears to be sufficiently robust to manuscript) 96 p.
accommodate differences in analytical and sampling
methodology and differences between laboratories. Any Dossis, P. and L.J. Warren. 1980. Distribution of heavy metals between
the minerals and organic debris in a contaminated marine
improvement in consistency between methods will be seen sediment. In: R.A. Baker (editor). Contaminants and Sediments,
as increased sensitivity, but the present results appear to Volume 1. p. 119-39.
be sufficiently sensitive to detect the influence of Hoff, J.T. and DJ. Thomas. 1986. A compilation and statistical analysis
anthropogenic activity and to identify areas of of high quality Beaufort Sea sediment data with recommendations
contamination. for future data collections. Report by Arctic Laborat ,ories Limited
for Environ 'mental Protection Service, Yellowknife, NWT;
Ile analysis of the monitoring results from 1982 to 1984 (unpublished manuscript) 118 p.

411

Nuclear Activation Services Limited. 1986. Beaufort sea Shorebase
monitoring program; data report on 1984 geochemical sampling.
Report for Department of Indian Affairs and Northern
Development, Yellowknife, NWT; (unpublished manuscript) 122 p.

Thomas, D.J., P.F. Wainwright, B.D. Arner, and W.H. Coedy. 1983.
Beaufort Sea coastal sediment reconnaissance survey: A data
report on 1982 geochemical and biological sampling. Report by
Arctic Laboratories Limited for Environmental Protection Service
(Yellowknife), Dome Petroleum Limited, Esso Resources Canada
Limited and Gulf Canada Resources Inc., Calgary, Alberta;
(unpublished manuscript) 459 p.

Thomas, DJ., W.S. Duval, and W.E. Cross. 1986. Development of a
monitoring program for contaminants in Tuktoyaktuk Harbour,
NWT A report to the Environmental Protection Service,
Yellowknife, NWT 68 p.

Wainwright, P.F., B. Humphrey. 1988. Analysis of sediment data from
the Beaufort Sea Shorebase Monitoring Program, 1982 to 1984.
Environmental Studies Research Funds Report Number
090.Ottawa, 147 p.

Yunker, M.B. 1986. Final report; 1982-83 Beaufort Sea Shorebase
monitoring program; Statistical analysis and recommendations for
future programs. Report by Dobrocky Seatecb Ltd. for Indian and
Northern Affairs Canada, Yellowknife, NWT; (unpublished
manuscript) 165 p.

412

A HIGH SPEED MULTI-CHANNEL DATA ACQUISITION SYSTEM 1
FOR REMOTE ACOUSTIC SEDIMENT TRANSPORT STUDIES

A.E. Hay 2, L. Huang, E.B. Colbourne, J. Sheng and A.J. Bowen

Department of Physics, Memorial University of Newfoundland,
St. John's, Nfld.Canada AlB 3X7

Department of Oceanography, Dalhousie University, Halifax, N.S. Canada

ABSTRACT systems which exploit the frequency dependence of
the scattered sound and which can be used for both
A multi-frequency, multi-beam acoustic system for backscatter and forward scatter measurement.
tethered near-bottom sediment concentration and
size profile measurement In the the ocean has been We have therefore adopted from the beginning a
developed. EXADAC, the data acquisition and multi-channel approach to system design which
control section of the system, is described. The incorporates at least three acoustic frequencies
system at present operates with 4 acoustic beams and a choice between several multi-beam
and 3 frequencies: 1.0, 2.25, and 5.0 MHz. The geometries. This represents a significant advance
data acquisition system consists of a CAMAC crate, over earlier acoustic devices for suspended
a PC/AT-compatible microcomputer and a 9-track sediment concentration profiling, which have been
streaming tape drive. The CAMAC crate contains restricted to a single frequency In any given part
several data acquisition and control modules, of the water column (1-4). Previous multi-beam
Including: a programmable four-channel 12-bit devices have been developed to measure suspended
transient recorder, a programmable clock, and a sediment concentration (5) and suspended sediment
GPIB interface controller. Typical 4-channel concentration and size (6), but must be physically
operating characteristics are: total range, 1 m; raised or lowered to obtain profiles.
range resolution, 2 cm; transmitted pulse length,
20 ps; pulse-to-pulse interval, 12 ms; acquisition One of the principal technical difficulties in
rate for 4-ping ensemble-averaged backscatter developing multi-channel profiling systems Is the
profiles for four channels, 2 Hz. The system has high data rate. Typically, a range resolution of
been used in two field experiments. Typical data order 1 cm is needed. Resolving the detailed
are presented. temporal structure of the suspended sediment
concentration field in the nearshore zone requires
that ensemble-averaged backscatter profiles be
1. INTRODUCTION obtained 5 to 10 times faster than the highest
frequency wave involved In transporting sediment,
In November 1986 funding was received to Initiate which is typically of order 1 Hz. Assuming 100
a joint project between Memorial University of range bins, which Is equivalent to a 1 m range
Newfoundland and Dalhousie University to develop interval at 1 cm resolution, and 12-bit samples, a
acoustic systems for sediment transport studies in four-channel system would therefore produce
the nearshore zone and on the continental shelf. roughly I Gbyte of data per day. A further
The project acronym is RASTRAN: Remote Acoustic constraint Is that consecutive backscatter
Sediment TRANsport measurement. profiles be averaged to eliminate the purely
statistical variations in signal amplitude which
The principal initial objective of the RASTRAN occur even in the absence of changes in
Project is to obtain accurate profiles of concentration (2). The data acquisition system
suspended sediment concentration and concentration must therefore perform at least the arithmetic
gradients in the bottom 1 m of the ocean at necessary to produce the "ensemble-averaged"
sampling frequencies approaching 10 Hz. In order profiles mentioned above. In addition, there is
to distinguish contributions to the scattered the requirement that the data acquisition system
signal due to variations in suspended sediment be flexible and expandable, to accommodate
concentration from those due to particle size additional channels as the need arises.
changes, it is in our view necessary to develop
1 We are developing two different RASTRAN systems.
RASTRAN Project Contribution No. 1 System 1 Is connected to an above-water data
2 acquisition and control system, and is Intended
On sabbatical leave at DELFT HYDRAULICS, P.O. Box primarily for use In the nearshore zone. System 2
177, 2600 MH Delft, The Netherlands Is designed to be completely submersible and
autonomous. It Is Intended primarily for use on

CH2585-8/88/0000. 413 $1 @1988 iEEE

the continental shelf. 2.1 Mesotech Acoustic Sounder Modules

There have been two field deployments of System 1: The Model 810 acoustic sounders were produced by
at Queensland Beach, Nova Scotia In 1987; and at Mesotech Systems for this project. They are based
Bluewater Beach, Georgian Bay, Ontario in 1988, on the company's existing line of small immersible
Each has resulted In successful continuous echo sounder modules, with Improvements to the
deployment with the system in the water and transceiver to permit: synchronous generation of
operating for periods of 3 to 4 weeks. The the transmitted pulse; jumper selectable values of
Queensland Beach deployment was a test of the output power, transmitted pulse length, and TVG
prototype system using only one acoustic (time-variable gain) schemes; and a choice of
frequency. Three frequencies and four different at least three frequencies including 1, 2.25, anci
acoustic beams were used at Bluewater, making this 5 MHz.
to our knowledge the first time that a
mul t I -frequency, multi-beam acoustic backscatter The acoustic sounder modules consist of a
system has been successfully deployed for sediment transceiver and single element piezoceramic
transport studies In the nearshore zone. transducer mounted in an immersible pressure case
(Figure 2). The modules are powered and triggered
In this paper we describe the data acquisition and externally. The signal received after pulse
control section which solves the data handling transmission is heterodyned down to 455 kHz, which
problem for System 1. A companion paper (7) Is the frequency of the receiver output. More
describes the design considerations for System 2, detailed specifications and performance
which is still under development. characteristics of the modules will be presented
in another article.

2. RASTRAN SYSTEM 1 2.2 EXADAC: The Data Acquisition System

The complete System 1 is sketched in Figure 1. It The data acquisition and trigger generation system
consists of the above-water data acquisition and has been developed as part of the RASTRAN Project.
trigger generation system which is connected via It has been given the name EXADAC, which stands
armoured multi-element (coaxial and single for EXpandable Acoustic Data Acquisition System.
conductor) cables to the Mesotech Model 810
acoustic sounder modules described below. The A photograph of the system Is shown in Figure 3,
coaxial elements are used for the signals and the and a block diagram In Figure 4. It is based on a
trigger pulses. the single elements are used for LeCroy model 1434A CAMAC crate, which provides a
24 VDC power. At present, four Mesotech acoustic rapid transfer data highway between data
sounders can be deployed on two armoured cables at acquisition modules and a GPIB interface to the
distances from the beach up to 300 m. This host computer. The data acquisition modules
limitation is due to the present cable length. consist of a LeCroy model 6810 programmable

455 KHz SIGNAL
ENVELOPE EXADAC
DETECTORS No

FUNCTION
GENERATOR

TRIGGER
PULSE

MESOTECH 810'S
68 h

Figure 1. Block diagram of RASTRAN System I

414

user. These data are stored In the 6810's internal
memory, and are downloaded between triggers via
the CAMAC crate and GPIB interface to the
computer.

The 6810 transient recorder is capable of
digitizing at a maximum rate of 5 MHz in single
channel operation and at 1 MHz in four-channel
mode, with 12-bit resolution. The digitizing rate,
input voltage range, and pre- and post-trigger
windows are among the programmable features of the
system. For our applications, which typically
involve transmitted pulse lengths of 20 ps, the
digitizing rate is usually 200 kHz, and the
post-trigger window length 1 to 2 ms.

The acoustic sounders are triggered at a rate of
80 Hz. This rate Is 80% of the maximum rate quoted
by the manufacturer, and is slow enough that
d"Y
contamination by surface-to-bottom multiple
"j,
reflections has not yet been a problem even in
Pry, 1"r

Figure 2. The Mesotech 810 acoustic sounder,
showing transducer and under water connector.
ff'Ps P
A'
high-speed, 12-bit resolution, 4-channel transient
recorder, and a model 8501 programmable clock also
made by LeCroy. Data Is transferred from the crate
-compatible microcomputer and
to an 8 MHz PUAT
stored after processing on a Thorne/EMI model 9900
9-track streaming tape drive. Data are displayed
using colour graphics either as time series of
backscatter at selected ranges or as acoustic
Images colour-coded with respect to backscatter
amplitude, intensity or concentration.

As presently configured, up to four separate
acoustic sounder modules can toe triggered
simultaneously and the four received signals
digitized simultaneously using EXADAC. The system MS
Is limited in principle to four units.
not
Additional channels can be accommodated by adding
another transient recorder to the CAMAC crate. The
system can also be operated with 1, 2 or 3 as well -ENNOMM
as 4 channels. We describe four-channel operation
below.

The acoustic sounders are triggered under computer
control. The trigger pulse is generated by a
command to the programmable clock in the CAMAC
crate, which Is then used to activate a Wavtek VF
function generator capable of driving the lines to
the sounders (Figure 1).

The output signal from each sounder module, after
being passed through an envelope detector to
eliminate the 455 kHz carrier (Figure 1), is fed
into one of the four input channels on the 6810
transient recorder. The transient recorder is
begins digitizing when it receives the trigger Figure 3. Photograph of the EXADAC data
pulse from the function generator (Figure 4): that acquisition system. The LeCroy Model 1434A CAMAC
is, at the same time the acoustic sounders are crate is at the top. Below it are, in descending
triggered. All four channels are digitized order, the Thorn/EMI streaming tape drive, the
simultaneously during a preset time window (or envelope detectors, the computer chassis, the
I @

ONNIMW

equivalently, a range window) selected by the function generator and the DC power supplies.

415

INPUT 6810
SIGNALS
TRANSIENT
TRIG IN RECORDER

FUNCTION CAMAC
GENERATOR BACK-
TRIG OUT 8501 PLANE
CLOCK

COLOUR PC /AT GPIB
MONITOR CONTROLLER

9-TRACK
STREAMING TAPE
DRIVE

Figure 4. EXADAC: B lock diagram

water as shallow as 2 m. In the 12 ms interval larger directly addressable memory sizes currently
between triggers the digitized, backscatter available In 32-bit 80386-based machines should
profiles for the four channels are downloaded from allow us to increase the acquisition rate by at
the 6810 to the computer and stored in an array. A least the desired factor of three.
user-selectable number of consecutive backscatter
profiles, usually four, are averaged to to yield a
single ensemble-averaged profile. At this point 3. FIELD EXPERIMENTS
further averaging of the data is performed over
each range bin. The width of the range bins are RASTRAN System 1, using EXADAC for data
also selectable by the user. Finally, the acquisition, has been deployed in two field
ensemble- and bin-averaged profile is written experiments, one at Queensland Beach, N.S. in
either to the hard disk or to a RAM disk. October 1987, and one at Bluewater Beach, Georgian
Bay, Ontario In May and June 1988. During both
It has been necessary to develop our own special experiments independent measurements of the
purpose data acquisition and processing software velocity field, pressure and suspended sediment
to perform these operations. We now have a fully concentration were made in parallel as part of a
operational package which permits us to acquire second joint project with B. Greenwood of the
individual acoustic backscatter profiles University of Toronto and one of us (AJB). The
simultaneously from 4 channels over a range of independent suspended sediment concentration
about 1 m with 2 cm range resolution at a minimum measurements were made using Optical Backscatter
Intervals of 12 ms. Four ping ensemble-averages of Sensors (OBS's).
these backscatter profiles are acquired at a
maximum rate of 2 Hz. At Queensland we successfully demonstrated that
the Mesotech modules and the EXADAC system could
Although the 2 Hz ensemble-averaged profile be used to monitor wave-induced resuspension of
acquisition rate for four channels represents a sand in the field. Preliminary analysis of the
respectable achievement in our view, and is we Queensland data shows that the RASTRAN and OBS
believe unmatched by any other system, it is results are comparable. A typical example of these
slower than would be desirable for certain types data is shown in Figure 5. It can be seen that
of study. The limitation on speed is the 8 MHz, sediment is lifted into suspension primarily
16-bit microcomputer. The faster clock speeds and during the passage of wave groups, consistent with

416

previous studies using OBS's alone to measure as possible (about 15 cm maximum horizontal
sediment concentration(B). This effect Is more separation between transducers). The other 2.25
obvious in the RASTRAN data from the Queensland MHz unit was deployed about 2 m farther offshore
experiment, apparently because the sensitivity of from the first. Data was collected during six
the OBS's was set too low. This difference in storm events, and Is being processed now to test
sensitivity accounts for much of the difference in the capability of the three-frequency system to
the overall structure of the two concentration discriminate between concentration and size.
time series in Figure 5. The differences In the
detailed structure are due to the faster sampling
rate for the OBS, and to differences In the 4. LABORATORY EXPERIMENTS
detected volume and the spatial separation of the
two instruments (0.6 m horizontal distance). EXADAC is used in the laboratory as well as In the
Nevertheless, It Is gratifying that the absolute field. A special purpose tank has been constructed
values of the RASTRAN and OBS concentrations are to perform acoustic calibrations of the Mesotech
comparable. acoustic sounders, to calibrate their response as
a function of suspended sand concentration and
We have also been able to obtain time series of size, and to test standard targets. EXADAC is used
bottom position from the RASTRAN data from to acquire and process the calibration data.
Queensland. The bottom position varies with time Calibration experiments can be performed
by �2 cm, consistent with sand ripples (which simultaneously on three Mesotech modules operating
during this experiment were a few cm high at the at different frequencies. The complete System 1,
RASTRAN location) changing form or migrating including Mesotech modules, cables, envelope
through the sound beam. detectors and EXADAC, can also be assembled and
tested in the laboratory facility. This ability to
At Bluewater we deployed 4 Mesotech modules: two use the same system in the lab and in the f ield
2.25 MHz units, one I MHz unit and one 5 MHz unit. has proved to be very valuable during software
As mentioned above, ensemble-averaged backscatter development and system checkout prior to,the field
profiles with 2 cm range resolution were obtained experiments.
simultaneously from these 4 instruments at a rate
of 2 Hz. The 1 and 5 MHz modules, and one of the
2.25 MHz modules, were mounted as close together

A. NAh A A A
WVV7 vqp V@ qf@pjq

co 0

A AA
1.0 2.0 3.0 4.0 5.0 6.0
MINUTES

Figure 5. Suspended sediment concentration approximately 8 cm above bottom and
wave data from Queensland Beach, Nova Scotia In October, 1987. From top to
bottom: on-offshore component of velocity; OBS sediment concentration; and
RASTRAN sediment concentration.

447

5. CONCLUSIONS Profiling System," Proc. Coasta-l Sediments '87,
(New York: ASCE), 236-249.
EXADAC Is a multi-channel data acquisition and 5
triggering system developed for use with Schaafsma, A.S. and W.J.G.J. der Kinderen, 1986.
high-frequency acoustic transceivers. It is "Ultrasonic Instruments for the Continuous
presently capable of acquiring 4-ping Measurement of Suspended Sand Transport," Proc.
ensemble-averaged backscatter profiles over a I m IAHR Symp. 2n Measuring Techniques Ln Hydraulic
range interval with 2 cm range resolution and Research, A. C. E. Wessels, Editor (Balkema,
12-bit data resolution on four channels Rotterdam), 125-136.
simultaneously at a rate of 2 Hz. The system Is 6
readily expandable to accommodate more channels. Crickmore, M.J., I.E. Shepherd, and P.M. Dore,
1986. "A Field Instrument for Measuring the
EXADAC has proved to be a versatile tool. It has Concentration and Size of Fine Sand Suspensions,"
been used In two field experiments as part of the Proc. Intl. Conf. on Measuring Techniaues of
RASTRAN Project for which It was developed. It Is Hydraulics Phenome In Offshore, Coastal and
also used in the laboratory for acoustic Inland Waters ' London, U.K. 9-11 Apr. 1986,
calibration work and data processing, including 425-441.
acoustic image analysis. 7
Hazen, D.G., A.E. Hay and A.J. Bowen, 1988.
While the 2 Hz averaged profile acquisition rate "Design Considerations for RASTRAN System 2,"
is satisfactory for many types of sediment Proc. OCEANS '88, Companion to this paper.
transport study, rates three to five times faster 8
are desirable and are expected to be achievable by Hanes, D.M. and D.A. Huntley, 1986. "Continuous
replacing the present 8 MHz PC/AT-compatible host Measurements of Suspended Sand Concentration In
computer with a 32-bit 20 to 25 MHz machine. a Wave Dominated Nearshore Environment,"
Future developments will include upgrading the Continental Shelf Res. 6, 585-596.
system with such a machine, and increas-ing the
number of data channels.

Acknowledgements

This work was funded by a Natural Sciences and
Engineering Research Council of Canada (NSERC)
Group Strategic Grant to AEH and AJB for the
RASTRAN Project, and an NSERC Strategic Equipment
Grant to AEH for EXADAC. We thank Mesotech Systems
Ltd., Rayonics Scientific Inc., and to LeCroy for
their assistance and cooperation throughout this
project. We thank the technical support staff and
graduate students from Dalhousie University and
the University of Toronto for their assistance in
collecting the data in Figure S. AEH is grateful
to DELFT HYDRAULICS for their support and
hospitality while this paper was being written.

REFERENCES
IYoung, R.A., J.T. Merrill, T.L. Clarke, and J.R.
Proni, 1982. "Acoustic Profiling of Suspended
Sediments In the Marine Bottom Boundary Layer,"
Geophys. Res. Lett. 9, 175-178.
2Hay, A.E.9 1983. "On the Remote Acoustic
Detection of Suspended Sediment at Long
Wavelengths," J_. Geophys. Res. 88, 7525-7542.
3Hess, F.R. and K.W. Bedford, 1985. "Acoustic
Backscatter System (ABSS): The Instrument and
Some Preliminary Results, " Mar. Geol. 66,
357-379.
4Libicki, C., K.W. Bedford, R. Van Era III and
J.F. Lynch, 1987. "A 3 MHz Acoustic Sediment

418

DESIGN CONSIDERATIONS FOR RASTRAN - SYSTEM 2'

D. G. Hazen, A. E. Hay* and A. J. Bowen

Department of Oceanography, Dalhousie University, Halifax, NS B3H 01
*Physics Department, Memorial University of Newfoundland, St. John's, NF All) 3X7

ABSTRACT et al. (2) have been used successfully by several groups; however,
Estimation and prediction of sediment transport on they still have the drawbacks of being intrusive, although they
the continental shelf is an important problem in fields are very small, and sampling only at discrete points above the
such as fisheries, pollution control, pipeline burial, (presumed) location of the bed. Video techniques have been used
and well-head maintenance. This paper discusses to monitor experiments, but have trouble seeing beneath the top
the design of an autonomous, multi-beam, multi- of the sediment cloud and measuring distances above the bed.
frequency acoustic backscatter suspended sediment The acoustic backscatter approach has several advantages over
profiler which builds on knowledge gained in deploy- other methods in that it is a remote sensing technique. The
ing a shore tethered version. This new system will instrument is mounted over the area of interest and transmits
allow direct measurement of sediment concentration high frequency acoustic pulses which are scattered by suspended
profiles in the bottom 1m of the ocean at a rate of 10 particles in the water column.
profiles/sec, in bursts, over a 2-3 week deployment.
Ancillary data such as flow velocity, temperature and The major problems of this approach are:
pressure will also be recorded. 1. Possible contamination by bubbles and biological scatter-
ers.
1. INTRODUCTION 2. Variations in the speed of sound (and thus time to distance
conversions) due to temperature changes or high concentra-
The measurement of suspended sediment near the sea floor is a tions of suspended sediment or bubbles.
problem which has challenged engineers and scientists for sev- 3. The problems of 'inverting' the backscattered signal to re-
eral years. Reliable suspended sediment profiles under various cover concentration profiles. This requires the ability to
wave conditions are of primary interest to coastal engineers and account for signal losses due to scattering, absorption, and
geomorphologists as a tool to help understand the mechanics of spherical spreading. (Libicki et al. (8) provides a review of
shoreline dynamics. These profiles may also be used in other these problems.).
conditions to help predict sediment transport rates on the con- 4. Calibration. Tamura and Hanes (9) provide an illustra-
tinental shelf. Successful predictions of these rates could in turn tion of the difficulties of providing repeatable calibrations
be used to predict pollution spreading rates or nutrient fluxes for of backscatter intensity versus concentration.
fisheries research.
Non-acoustic approaches used to measure sediment in suspension -Several groups have been working with var .iou.s types of acous-
fall into several groups: pumped sampling, sediment traps, opti- !tic backscatter instruments. Young et al. (10) describes a 3
cal transmission, optical backscatter, and video. Several of these MHz acoustic concentration meter (ACM) system, which has
were reviewed by Huntley (7). Pumped samplers suffer from two been used at several locations off the East coast. Hay and Hef-
problems. The first problem is they sample only from discrete fler (4) describe the design considerations for the Acoustic Sus-
heights above the bed, which can result in undersampling the pended Sediment Profiler (ASSP) which was to have been built
structure of a complex sediment cloud, and the second is the lo- by the National Research Council of Canada as part of the Cana-
gistical problems involved in maintaining them over a deployment dian Coastal Sediment Study (C'S') in 1983-1985. Crickmore et
lasting several weeks. Sediment traps are useful for determining al. (1) describes another system which simultaneously measures
net fluxes but have the disadvantage of being inherently unidirec- attenuation, backscatter and forward-scatter from a sediment
tional and ra@se serious questions about flow disturbance. Optical cloud moving between two pairs of heads. Hanes and Vincent
transmission devices (transmissometers) have been used success- (3) briefly describe a low crst 3 MHz system under development
fully to measure turbidity in the water column but are also rela- at the University of East Anglia and Libicki et al. (8) describe
tively large and, again, flow disturbance becomes a problem near another 3 MHz system which they have used in the Great Lakes.
the sea floor. However, they do have the advantage of providing All of these systems have the disadvantage of being only single
an effectively continuous output in a convenient, electronic, form. frequency devices and therefore requiring a priori assumptions
Optical Backscatter devices, such as those described by Downing about sediment size distributions before further analysis can take
place. They are also limited in the speed and amount of data
which can be gathered and must be tethered to a shore station.
1RASTRAN Project Contribution No. 2

CH2585-8/88/0000- 419 $1 @1988 IEEE

Some of the most recent work with acoustic backscatter systems -I ROM
is described in a companion paper by Hay et al. (5) which de- 256K 64K
scribes the Remote Acoustic Sediment Transport (RASTRAN)
System', developed as part of a joint effort between Memorial T
and Dalhousie Universities. Hay describes the system as used in V2, 0
.the lab and at two field locations, where it operated tethered to a MPU _ilp Parallel InterfaceIto Data Logger
sfiore-based computer system. This paper describes the plans for
RASTRAN- System 2: a self-contained version including data sl-
loggers, etc. for deployment on the continental shelf for 2-3 weeks MPU
at a time. UP1452 Detectore
Mux
2. THE RASTRAN SYSTEM A/D Background
ADC816
The RASTRAN system is composed of a number of sensor heads I M11. sensors
10 bits (u, v, P, OBS)
(single frequency acoustic sounders), manufactured by Niesotech mesotech Hea&
Systems Ltd., coupled to a data acquisition and control corn-9uter
and sensors for ancillary data. For System 1, these heads were
connected by a 300m cable through an envelope detector to a
CAMACTIM crate high speed data acquisition system and an AT Figure 1 - RASTRAN - System 2 Block Diagram
compatible personal computer.
however, will allow a radio link or an interface to the TUDATS
The second phase of the project is to design an autonomous ver- hard wired system (6) to be used instead.
sion of the system, capable of being deployed for up to a month
at continental shelf depths. As shown in Figure 1, the system 2.1 Sensors
will consist of a battery power supply, a Digital Data Acquisition
and Control (DDAC) module, and a module to store or transmit The system will be primarily concerned with acquiring data from
the data. The primary focus of this phase of the project is to put up to four Mesotech Model 810 acoustic sounders. These im-
together a system using a data logger. The modular approach, mersible sounders, as shown in Figure 2, are in an aluminum and

0 0

OW 4 AA46T

0
A

arc

@1
,:I.r.u.d
v,
Mesoted,

Figure 2 - Mesotech Model 810 Acoustic Sounder

420

polyvinyl chloride (PVC) enclosure. Each contains the high fre- / < Data
DDAC
quency circuitry required to transmit a pulse of pre-determined
length and amplify the returned (backscattered) signal. The re- Cable
turned signal is output on a 455 kHz IF caxrier for the next stage Temperature to Shore
of the signal conditioning process. The system uses sounders with -OBS
frequencies of 1.0, 2.25 and 5.0 MHz and beam widths of 2.9' to - Pressure < --t rlare
3.6". These sounders were used successfully in System 1. Roto@ u, v
- EM u, @
The remaining sensors will be used to record ancillary informa - mrsot@,h Heads
tion and obtain independent estimates of sediment concentration.
The first of these sensors is a small array of Optical Backscatter
(OBS) probes, manufactured by HR Labs for Downing & Asso-
ciates. These are the latest version of the probes described by
Downing et at. (2) and provide an estimate of suspended sedi- Figure 3 - DDAC Block Diagram
ment concentration over a small 1.3 crn' volume, in front of the
probe. The OBS probes will be mounted near the cone illumi-
nated by the Mesotech heads and will be used as an independent
estimate of concentration. sounders, acquiring ancillary data, conducting data transfers to
the data logger and controlling power to the A/D converters,
Other sensors include pressure and thermistor probes for depth, sensors and data logger.
wave height, and temperature measurement along with a Marsh-
McBirney model 512 electromagnetic current meter, and a slower The final part of the digitizer hardware to be described is the sig-
response, rotor-type current meter for determination of both nal conditioning required to detect the envelope of the received
mean currents and wave orbital velocities. signal. Two alternatives have been designed and tested. The sim-
pler design is for a rectifier detector which follows the envelope
2.2 The DDAC Module of the amplitude modulated (AM) IF signal from the Mesotech
heads. The disadvantage of this design is it loses phase informa-
The Digital Data Acquisition and Control module, as shown in tion which may be of interest in attempting to recover velocities
Figure 3, is responsible for both digitizing data streams from from Doppler shifts in the returned signal. The other, more com-
the various sensors and controlling other devices in the system. plex, design is a synchronous demodulator which preserves phase
The DDAC is built around a number of cards on a STD bus sys- but requires the application of modulus function at some later
tem. The system includes an 8MHz CPU, 256k RAM, Real Time stage in the process for simple backscatter calculations. (The
Clock, 64 k EPROM, and a custom digitizer board. The STD bus amplitude modulated signal is, in effect, "over modulated" with
system is capable of being placed in a low power "sleep" mode the amplitude of the signal occasionally exceeding that of the
to reduce battery drain. The control program will be written in carrier). We have not yet decided which design is best for our
C, compiled and stored in EPROM. application.
The digitizer board is designed around a slave microprocessor The sampling program, as shown in the modified Warnier-Orr
and a very high speed Analog to Digital (A/D) converter module, diagram in Figure 4, after initialization, consists of running re-
capable of 1 Msamples/sec. The actual speed required is 200k peated sampling bursts on an hourly basis. Data is sampled in
samples/sec for each of the four Mesotech heads to preserve the bursts because the system cannot log (or process) the data fast
100 kHz signal bandwidth. Data from the A/D converter is trans- enough to maintain continuous sampling for over 15..minutes.
ferred via Direct Memory Access (DMA) to a buffer in RAM. The In addition, continuous sampling would shorten the deployment!
slave processor is an Intel UP1452 running at 16 MHz, executing length of the system dramatically.
assembly language code stored in EPROM. The slave processor
also acts as an intelligent 1/0 controller, triggering the Mesotech
- Run 10 profiles
(< 10ms each)
- compile avg. profile
Acquire profile - (Apply inversion
- Power up A/D, etc. - Read OBS algorithm)
- Acquire data at - Acquire ancillary Bin average profile
Power up 5-10 Hz for data (every 2nd
initialization - 15 min time through)
Sampling cycle - Store data - Store in RAM
(loop time from buffer to buffer
1 hr) data logger - Wait for next pass
- Go to 'sleep' until
next burst

Figure 4 - DDAC Program Flow

421

The sampling bursts consist of gathering sets of 5 to 10 pings and 4. ACKNOWLEDGEMENTS
ensemble averaging them to produce 10 averaged profiles per sec-
ond. It is hoped that by sharing the tasks between the master The work on the RASTRAN project, Systems I and 2, has been
and slave processors, inverting and bin averaging the data can funded by a Natural Sciences and Engineering Research Coun-
also be done in real time. In addition, data from the OHS probes cil of Canada (NSERC) Strategic Grant. Figure 2 is provided
will be acquired at 10 samples/sec and the other ancillary instru- courtesy of Mesotech Systems Ltd. of Port Coquitlam, BC.
ments at a lower rate of 2.5 or 5 samples/sec. This data is stored
in RAM until the buffer is full. The buffer fills in approximately 5.REFERENCES
15 minutes. At this stage, sampling is stopped and the A/D and 1. Crickmore, M.J., I.E. Shepherd and P.M. Dore, 1986. "A
sensors powered down. The master CPU then performs any re- Field Instrument for Measuring the Concentration and Size
maining processing required and transmits the buffer through the of Fine Sand Suspensions" Proceedings of the International
slave to the data logger. When the data has been transferred, Conference on Measuring Techniques of Hydraulics Phe-
the system goes into a low-power "sleep" mode until the next
sampling burst is required. nomena in Offshore, Coastal & Inland Waters, p. 425-442.
Bedford, U.K.: BHRA, the Fluid Engineering Centre.
2.3 The Data Logger 2. Downing, J.P., R.W. Sternberg and C.R.B. Lister, 1981.
"New Instrumentation for the Investigation of Sediment
The data logger is being designed around a 240 Mbyte Write Processes in the Shallow Maxine Environment", Marine Ge-
Once-Read Many (WORM) optical disk drive, as shown in Fig- ology 42, 19-34.
ure 5. This drive has higher capacity and fewer moving parts 3. Hanes, D.M. and C.E. Vincent, 1987. "Detadled Dynamics
than conventional 1/4 inch tape cartridge systems. The drive is of Nearshore Suspended Sediment", Proc. Coastal Sedi- .
controlled by a PC bus controller card which, in turn, is con- ments '87,285-299. New York: ASCE.
nected to a low power PC compatible system on a passive PC 4. Hay, A.E. and D. Heffier, 1983. Design Consideration for
bus backplane. Details on the design will be reported in a future an Acoustic Sediment Transport Monitor for the Nearshore
paper as the design matures. Zone, Report C2S24, 37 pp. Ottawa: National Research
Council Canada.
5. Hay, A.E., L. Huang, E.B. Colburne, J. Scheng and A.J.
Bowen, 1988. "A High Speed Multi-Channel Data Acqui-
sition System for Shore-based Acoustic Sediment Transport
V401 256k Parallel Studies", Proc. Oceans '88, Companion to this paper.
CPU RAM 1/0 6. Hazen, D.G., D.A. Huntley and A.J. Bowen, 1987.
to DDAC "UDATS: A System for Monitoring Neaxshore Processes",
Proc. Oceans '87, 993-997.
7. Huntley, D.A., 1982. In Situ Sediment Monitoring Tech-
J port C2S2-1, 35 pp. Ottawa: National Research Council
niques: A Survey of the State of the Art in USA, Re-

Disk WORM Drive WORM Canada.
8. Libicki, C., K.W. Bedford, R. Van Era III and J.F. Lynch,
256k Controller Drive 1987. "A 3 MHz Acoustic Sediment Profiling System",
Proc. Coastal Sediments '87, 236-249. New York: ASCE.
9. Tamura, T. and D.M. Hanes, 1986. Laboratory Calibration
Figure 5 - WORM Drive Data Logger of a 3 Megahertz -Acoustic Concentration Meter to Measure
Suspended Sand Concentration, Technical Report 86-004,,
87 pp. Miami: University of Miami, Rosensteil School of
3. CONCLUSIONS Maxineand Atmospheric Science.
10. Young, R.A., J.T. Merrill, T.L. Clarke and J.R. Proni, 1982.
The design work on RASTRAN-Systern 2 is progressing. The "Acoustic Profiling of Suspended Sediments in the Maxine
new system will allow the simultaneous acquisition of profiles Bottom Boundary Layer", Geophysical Research Letters,
from up to four acoustic sounders operating at three different. 9(3), 175-178.
frequencies at a rate of approximately 10 profiles/sec. This data.
will be stored, along with that from ancillary sensors, on a high
capacity data logger over a 3 to 4 week deployment. RASTRAN
- System 2 will be a significant advance over existing systems in
terms of both the speed and storage capacity of the data collec-
tors and the number of acoustic sounders which may be attached
to the system.
The development plan is for bench testing by January 1989 and
deployment as part of the instrument array at the Stanhope '89
experiment in October-November 1989 at Stanhope Lane, P.E.I.
The significant problems remaining are those of interpreting the
data and inverting the backscatter profiles to obtain concentra-
tions.
@
6k

M
A @@to DDAC

4@2

MINI-PROBES: A NEW DIMENSION IN OFFSHORE IN SITU TESTING

Alan G Young, Lowell V. Babb, Ronald L. Boggess

Fugro-McClelland Inc.
Houston, Texas

ABSTRACT HISTORICAL PERSPECTIVE

In an earlier publication on Underwater Soil Sampling,
Acquiring geoscience data about the near-seafloor sediments Professor I. Noorany predicted that "the practice of under-
is fundamental to using this natural resource. Tradition- water sampling and testing is presently in a state of flux
ally, gravity corers have been used to sample the ocean and there is much promise orl%veloping better and more
sediments. Sample recovery is usually very shallow unless efficient systems in the future. " In reviewing how this
special purpose, expensive systems are deployed. Acquiring technology has evolved since he made this prediction, the
in situ data using push-in probes has been developed in efficient systems he predicted are now available.
recent decades. Existing probes are approximately 3.5 cm in
diameter and require deployment of large seafloor mechanisms The evolution of geotechnical sampling methods and in situ
(10,000-20,000 kg) to achieve significant penetrations. testing devices used offshore today is important in under-
Use of such devices from inexpensive vessels is generally standing today's technology and in developing an apprecia-
not practical. tion of how much progress has been made in the area of geo-
Recent development of a series of miniature probesis technical investigations since the mid-70's and early 80's.
Basically, the development of the equipment available today
described. Because of the small probe size, pushing mecha- evolved through the activities of three different groups:
nisms are also scaled down (1,000 kg typical). Handling (1) Oceanographic and Geologic Research Centers, (2) Naval
equipment onboard oceanographic research vessels is suitable Research Centers, and (3) Geotechnical Engineering Service
for deployment of the self-contained mini-probe systems. Companies.
Installation of the systems on remotely-operated-vehicles
also appears feasible. Penetrations to 12.2 m have been The methods of operation used for geoscience sampling and
achieved. in situ testing employed today depend upon whether the
system is to be deployed downhole or used in a seafloor mode
INTRODUCTION of operation. The mode of operation in turn depends upon
the depth of interest for the particular application or
There are a large variety of marine activities planned for need. For deep penetrations (greater than 10 m), downhole
the U.S. Exclusive Economic Zone and other oceanic areas sampling and in situ testing are performed in conjunction
that will require geological and geotechnical data to with drilling equipment. For shallow penetrations (less
describe the properties of the seafloor sediments. The data than 10 m), self-contained units or equipment mounted on a
requirements will vary from only a few meters for small tethered seafloor platform or submersible vehicle is most
shallow foundations and cables to depths to 200 m for deep widely used.
foundations such as the piles supporting massive offshore
platforms. There are a number of different methods depend- Since 1947, when the first geotechnical investigations were
ing upon the depth of interest and water depth to acquire performed from temporary platforms in about 6 m of water for
these data at the site for the proposed activity. the U.S. petroleum industry, improved technology has evolved
that has allowed drilling and sampling to be performed from
Traditionally, geological and geotechnical data were ob- vessels that can accommodate the move into deeper water.
tained by performing laboratory tests on samples acquired by Anchored barges eventually took the place of the temporary
a self-contained gravity sampler or a downhole sampler oper- platforms and were used extensively until 1962 when a port-
ated through the bore of a drill pipe advanced by a rotary able drilling rig was first mounte n an oilfield supply
drilling rig. In the last decade, there has been much vessel, as described by McClelland.M
greater reliance on in situ testing devices that can be
deployed from seafloor platforms, submersibles, or used with The shift to water depths greater than 200 m, the difficult
traditional drilling equipment. The purpose of this paper logistics associated with anchoring, and the increased use
is to describe a new "concept" and system available for of specialized in situ testing equipment brought forth, in
performing in situ testing which overcomes many of the dif- the mid-1970's, the use of dynamically-positioned geotechni-
ficulties of previous systems. A primary advantage of the cal drill ships such as. the M.S. Bucentau . In some cases,
new system is that now a small oceanographic vessel can oilfield drilling vessels have been used for geotechnical
deploy the equipment, allowing more widespread and cost- investigation; however, high operational costs have fre-
effective collection of geoscience data. quently caused this to be a prohibitive option.

CH2585-8/88/0000- 423 $1 @1988 IEEE

The most inexpensive method for obtaining samples of marine BASIC CONCEPT
sediments has been the self-contained tool such as the
gravit.j (or drop) core sam@@jr as described by Hvorslev and Examples of a standard-sized probe and a mini-probe are
Steton 3) and Kullenberg. Most of these samplers are shown in Figure 1. The mini-probe is 1.27-cm-diameter and
limited to penetration depths of 10 m or less. However, a the larger probe is 3.55-cm-diameter. Both probes are
deepsea gravity corer has been constructed in sizes capable called friction cone penetrometers. Each contains two load
of obtaining "Tples up to 45 m in length as described by cells that measure resistance to penetration as the tool is
Hollister et al. pushed into the soil. Two measurements, the tip resistance
and the sleeve friction, are recorded continuously as the
A number of sampling and in situ testing systems that would Orobe is pushed into the seafloor sediments. By analyzing
operate from either tethered platforms or submersibles were the tip and sleeve data as a function of depth, valuable
developed in the 1950's and 1960's. Tethered platforms yg1d information on the sediment behavior can be determined.
for sampling include: (1) the U.S. Navy's TIPOS,
(2) the University of Rhode I@IqRd's DOSP,R?the Dutch The data signals are usually recorded on a PC-compatible
Geological Survey's Woff II. The Lehigh University's computer. Plots of the tip and sleeve data versus penetra-
Underytter Tower, and Woodward-Clyde Consultants' tion can be generated in the field.
MITS(I were tethered platforms that had the capability to
perform in situ vane and cone penetrometer testing. Probes like the above example are connected to a thrusting
rod which is jacked into the soil. Most thrusting rod
In situ testing equipment operated from a deepsea submers- systems are composed of one-meter long threaded connecting
ible that was e-. geloped during the samt,4Y eriod included: rods. These rods are made up as the probe penetrates or a
(1) UL I -Z@, (2) the Deep Quest, and (3) the pre-strung length of rods is supported by a guideline
AlvinvW
Geise and Kolk describe a system that can be used system. The new mini-probe thrusting rod is formed by a
from a manned submersible to perforr@,T situ testing and continuous long rod which is coiled as shown in Figure 2.
sampling to seafloor penetrations of 1 m.

A wide variety of seafloor testing systems were developed in
the 1970's that allowed in situ testing to be performed to
penetrations up to 35 m. In situ vane testing and cone
penetrometer testing were performed with systems as follows:
(1) McClelland's Stingray, (2) Fugro's Seacalf, (3) Fugro's
Seasprite, (4) McClelland's Starfish, (5) Fugro's Seaclam,
and (6) McClelland's Halibut. These systems, developed N
primarily for the geotechnica troleum industry, ve
been docume?61 by Young, 11% Ndberg, a al, _J
Young et al, and McClelland. 11-03 S,

In more recent years, improved equipment for tethered plat-
forms was developed that would allow in situ testing to be
i2_ @,,,r%
performed in deepwater. The XSP-40 was developed and tested
by the Naval Civil Engineering Laboratory capable of per- offl, A
forming tests to 12-m ftq@h. In 1984, the SEALION as
V
described by van der Wal was developed and tested by the
Netherlands Council of Oceanic Research. It could be oper-
ated as an automated piezoeone probe system in water depths
up to6,000m. The system was designed to perform tests to
seafloor penetrations up to 5 m.
-7

7
72-
Ne 4@rl@
9F
eo
FIGURE 2. COILED THRUSTING ROD AND MINI-PROBE

The thrusting rod is inserted into a hydraulic ram jacking
r
-probe
system. By actuating a rod chucking device, the mini
rod is straightened and thrust into the soil. At the con-
AX:
All
clusion of a push, the chucking mechanism is reversed and
& "r, f @0,@Jypk@e jKi-pr,
ego,
the rod is jacked out of the soil and re-coiled. The
@A* --J W,;61 RJ
A,_ -4- Ar thrusting unit is shown in Figure 3.

The major advantage of the mini-probe system is its compact
size. Since the probes and rod are small, the force and
reaction weight required to penetrate the soil are also
small. A typical system weighing 1,000 kg has penetrated as
deep as 12.2 m. In certain soils, penetrations exceeding
18 m are projected. An added benefit of the small size is
the low power requirement which makes battery-powered
FIGURE 1. CONE PENETROMETERS: MINI-PROBE AND syste ms quite feasible.
STANDARD SIZE UNITS

424

, _7
_Vf @J

-PROBE SYSTEM MOUNTED ON 4-WHEEL
FIGURE 4. MINI
DRIVE LIGHTWEIGHT TRUCK

FIGURE 3. MINI-PROBE THRUSTING UNIT WITH
20-FOOT ROD NAUTILUS CONE PENETRATION TEST

MINT RESISTANCE @- 15,
The major disadvantage of the mini-probe system, compared to PENETRA7[ON 200 400 600 Sao RATIO
FE@T UIE FR@MO@ I - - - @SF , ,
standard-sized probes, is limited penetration. This limita- 0 10 15 20 2 4 %6 8
tion is related to the inherent stiffness of the thrusting
rod. While standard systems have penetrated in excess of
90 m, the maximum mini-probe penetration is expected to be
18 m. ------

MINI-PROBE VARIATIONS

The example mini-probe described above is but one variation
n
of available probes. Additional probes have been developed.

The piezocone penetrometer measures the pressure of the pore
fluid in the formation. Combining the pore pressure and tip
resistance data can improve the determination of soil type.
The standard-sized pie@yTnes have been in widespread use
since the late 1970's. Miniature versions are now
available.
7;
25
Mini-probes have also been designed to acquire soil vapor
samples. This device requires a vacuum that is generated at
the ground surface to pull vapor samples from the soil.
This mini-probe is widely used in the investigation of 30
hazardous waste sites. A modification of this probe allows
the collection of a pore fluid sample for chemical analysis.

35
A probe to measure acoustic properties of the soil is
presently under development. Another mini-probe to measure
soil resistivity has been conceived. Many other types are
possible and will allow exploitation of the mini-probe tech- 40
nology.

MINI-PROBE DEPLOYMENT SYSTEMS
A mini-probe system deployed on the front of a small four- FIGURE 5. EXAMPLE DATA PLOT
wheel drive truck is shown in Figure 4. Three vehicles of
this type are in daily operation in the U.S.A. Primary uses
of the truck-mounted mini-probe systems are vapor sampling A mini-probe system has also been mounted on the flotation
and friction cone penetrometer testing. An example of cone vehicle shown in Figure 6. In this application, 180 sepa-
penetrometer data is shown in Figure 5. Data from the tip, rate coff )tests were performed along a chain of barrier
sleeve, and friction ratio are presented as a function of islands. 5 Because of the vehicle's configuration and
depth of penetration. light weight, it could be towed between the islands. A

425

question that often arises concerns the fatigue life of the By incorporating self-contained data recording and hydraulic
coiled rod. For this project, about 170 tests were per- power systems, the Nautilus can be deployed in even deeper
formed with the same rod. water. A single line mechanical cable can be used to deploy
the system from small oceanographic research vessels as
illustrated in Figure 8. Alternately, a deepwater mini-
IF probe can be installed on a remotely-operated-vehicle (ROV).
4P W,I Jp This concept is shown in Figure 9. The benefits of this
Z5
A; S application are use of the ROV's data and power systems and
"'A
If some
y manueverability independent from the mother ship.

@(-r 4
Handling
Hoisting
Frame
System

FIGURE 6. MINI-PROBE SYSTEM MOUNTED ON A
FLOTATION VEHICLE

In water depths of approximately 30 in, a mini-probe system
has been waterproofed and installed on a submersible plat-
form. This device, called Nautilus, is shown in Figure 7. Nautilus
In this configuration, hydraulic power is provided from the A
surface through adualhose. Data signals are transmitted
to the surface on an electrical cable. Ballasting the
system is accomplished by varying the quantity of lead
weights. This device has most often been deployed from a
floating barge. Geotechnical Probe

FIGURE 8. DEPLOYMENT OF A MINI-PROBE SYSTEM FROM
AN OCEANOGRAPHIC RESEARCH VESSEL

VIDEO CAMERA

REMOTELY MANIPULATOR ARM
OPERATED
VEHICLE

MINI-PROBE
MECHANISM

FIGURE 9. MINI-PROBE INSTALLED ON A
REMOTELY-OPERATED VEHICLE

FIGURE 7. SHALLOW WATER MINI-PROBE SYSTEM

426

CONCLUSIONS 10. McClelland, B. and Ehlers, C.J., "Offshore Geotechnical
Site Investigations." Chapter 9. Planning and Design o
The patented mini-probe systems show great promise for Fixed Offshore Platform . B. McClelland and M.D. Reifel,
extending the use of in situ soils data. The high cost of Eds. New York: Van Nostrand Reinhold Co., 1986,
operating dynamically-positioned vessels to support drill- pp. 224-265.
ing, sampling, and testing activities will limit the oppor-
tunities to acquire geoscience data in deepwater areas' 11. Noorany, I. Underwater Soil Sampling and Testin , San
because of the high daily costs. Project cost savings can Diego State College, San Diego, California, 197 1.
be particularly realized in the frontier deepwater environ-
ment through the use of small oceanographic vessels. 12. Ibid, p. 12.
The new system offers many advantages over previous systems 13. Ibid, p. 13.
including: (1) deployment from any type of vessel,
(2) rapid repetition of testing from site to site, and 14. Noorany, I. "Offshore Sampling and In Situ Testing:
(3) rapid dismantling and transportation to any area of the 1981 Update." Updating Surface Soundings. 1985.
world. 15. Preslan, W.L., Lourie, D.E., and Boggess, R.L. "A Cone
Although the depth of penetration is limited compared to Penetrometer for Beach Reconnaissance." Proceedings.
much larger predecessors, widespread use of the tool is Beagh Preservation Technology conference. Gainesville, FL.
expected to cover a majority of the research and technical March 23-25, 1988.
requirements for the activities planned in many marine
areas. Any activity requiring knowledge of the seafloor 16. Richards, A.F., McDonald, V.J., Olson, R.E., and Keller,
sediments to depths approaching 18 m can benefit from this G.H. "Under-water Soil Sampling, Testing, and Construction
new technology. Control." Special Technical Publication 501. American
Society for Testing and Materials, Philadelphia, PA, 1972.

17. van der Wal, J., Bergen Henegouw, C. van, Hoogendoorn,
H.G. and Richards, A.F. "Sealion: A Snellius-II Expedition
Automatic System for In Situ Geotechnical Testing in Water
Depths of 6000 m." Oceanology. London: Graham and
BIBLIOGRAPHY Trotman, 1986, pp. 241-246.
18. Young, A.G "Marine Foundations." Chapter 14. The
1. Geise, J.M. and Kolk, H.J. "The Use of Submersibles for Handbook of Coastal and Ocean Enpaineering. J. Herbich, Ed.
Geotechnical Investigations." The Design and Operation of Houston: Gulf Publishing Co., 1989.
Underwater Vehicles. Undon: Society for Underwater Tech- 19. Young, A.G, McClelland, B., and Quiros, G.W. "In Situ
nology, 1983, P. 13. Vane Shear Testing at Sea." Presented at the ASTM Inter-
2. Hollister, C.D., Silva, A.J., and Driscoll, A. "A Giant national Symposium on Laboratory and Field Vane Shear
Piston-Corer." Journal of Ocean Engineering. Vol. 2, 1973, Strength Testing, Tampa, FL, 1987.
pp. 159-168. 20. Zuidberg, H.M. and Richards, A.F. "Sampling and In Situ
3. Hvorslev, M.J. and Stetson, H.C. "Free-fall Coring Tube; Geotechnical Investigations Offshore." First Shanghai
A New Gravity Bottom Sampler." Bulletin, Geological Symposium on Marine Geotechnology and Nearshore/Offshore
Society of America. Vol. 57, 1946, p. 935. Structures. Claney, R.C. and Farg, H.Y., Eds. Special
Technical Publication, American Society for Testing and
4. Inderbitzen, A.L. and Simpson, F. "A Study of the Materials, 1986.
Strength Characteristics of Mining Sediments Utilizing a 21. Zuidberg, H.M., Schaap, L., and Beringen, F. "A Pene-
Submersible." Special Technical Publication 501. American trometer for Simultaneously Measuring Cone Resistance,
Society for Testing and Materials, Philadelphia, PA, 1972, Sleeve Friction, and Dynamic Pore Pressure." Penetration
pp. 204-215. Testin . Proceedings of the Second European Symposium on
5. Inderbitzen, A.L., Simpson, F., and Gross, G., A Compari- Penetration Testing, ESOPT 11, Amsterdam, May 24-27, 1982,
son of In Situ and Laboratory Vane Shear Measurements. pp. 963-970.
Lockheed Ocean Laboratory Report 681703, Lockheed Missile
and Space Company, 1970.

6. Kullenberg, B. Deep-sea Coring. Swedish Deep-Sea Expe-
dition Report No. 4, 1955,,p. 35.

7. Lambert, D.N., "Submersible Mounted In Situ Geotechnical
Instrumentation." Geo. Mar. Let . Vol. 2, 1982, pp. 209-214.

8. Lewis, L., Nacci, V., and Gallagher, J. "In Situ Inves-
tigations of Ocean Sediments." Proceedings, Civil Engineer-
ing in the Oceans 11, ASCE Conference, Miami Beach, FL,
December 10-12, 1969.

9. McClelland, B. "Techniques Used in Soil Sampling at
Sea." Offshore. Vol. 32, No. 3, March 1972, pp. 51-57.

427

BOUNDED BEAM TRANSMISSION ACROSS A WATERISAND INTERFACE,
EXPERIMENT AND THEORY

K. L. Williams and L. J. Satkowiak

Naval Coastal Systems Center
Physical Acoustics Branch, Code 2120
Panama City, FIL 32407-5000

AB STRACT solid layers.
A series of measurements were made to investigate the The predictive capabilities of the SAFARI program were
effect of a water/sediment interface on a propagating tested as part of an experiment designed to examine
acoustic signal. Under nearly laboratory conditions the effect of a water/sediment interface on the
data were collected mapping out the sound pressure propagation of both linear and parametric beamS.4 In
field in a homogeneous sand sediment for a 20 kHz the following, we first describe the experimental
linear source. The data were taken at grazing angles arrangement and data reduction methods in Section 11.
that ranged from well-above to weil-below the critical in Section 111, we present a brief overview of the
angle for the sediment. Based on data collected, in- theoretical underpinnings of the SAFARI program and
sediment pressure field contours were obtained. These -its implementation in the present work. Then, in
contour plots are compared with theoretical Section IV, we compare the SAFARI and experimental
predictions calculated using the SAFARI model (H. results and in Section V we conclude.
Schmidt and F. B. Jensen, J. Acoust. Soc. Am., 77, 813-
825 (1985)). Agreement between SAFARI and 11. EXPERIMENT
experimental pressure field contours represent an
experimental confirmation of the SAFARI code. 'The measurements were made under nearly laboratory
conclitions at the Admiralty Research Establishment
,Bincleaves Acoustic Range tank facility in the United
Kingdom. A schematic of the facility is given in Fig. 1.
The 10m x 3m x 2m tank consisted of a steel frame with
fiberglass walls. The tank itself was level to within 30
over its entire length. The transducer mounting
allowed complete control over the height of the array
1. INTRODUCTION and the direction it was pointed. The transducer was
mounted on a full tilt-and-pan mechanism that gave
The propagation of acoustic energy across water/ angular accuracy to within 0.30. The tilt-and-pan
sediment interfaces is of interest in any situation where mechanism was attached to the bottom of a remotely
one wants to examine sub-bottom features. The use of controlled telescoping arm (which was accurate to
the plane wave transmission coefficient to characterize within 0.3 cm). The telescoping arm mechanism was
this propagation suffers from one fundamental flaw- attached to a platform that allowed forward,
real transducers are not infinite in extent. ;Th@ backward, left, and right movement. The platform
bounded nature of real transducers implies that there is was able to move 5 meters in a given direction with an
a spectrum of plane wave components incident onto accuracy of about 0.5 cm. The versatility of the
the water/sediment boundary in any experimental positioning system allowed one to "point" the acoustic
arrangement and it is useful to have knowledge of the beam very accurately. it was felt that the location of
transmitted field structure as well as the transmitted the center of the acoustic beam on the surface of the
level. The theoretical methods needed to attack this. sedimentwas known to within 5 cm.
problem have been known for many years.1 However,
the mathematics is rather involved and if one desires an Eighteen in-sediment hydrophones and one in-water
extension to cases where there are sediment layers hydrophone were employed in the experiment. The
analytical analysis quickly becomes infeasible. hydrophones were 2.2 cm ceramic spheres with a
nominal sensitivity of -202 dB're IV/pPa. The beam
Recently, the analytical methodology of Ref. 1 hasbeen patterns from each hydrophone indicate nearly
implemented numerically.2.3 This Seismic and Acoustic symmetric response. The in-sediment hydrophones
Fast-field-program for Range Independent were held in position by a lattice of quarter inch kevlar
environments (SAFARI) is a full wave solution which ropes secured through holes drilled in the sides of the
allows an arbitrary number of layers (the version used tank. The divers were able to adjust the hydrophones
allows up to 24 layers) each of which,may be a vacuum, both laterally and vertically to within 1 cm of their
an attenuating liquid, or a viscoelastic solid. An planned positions. Acoustic ranging was used to verify
amplitude and phase shaded vertical array of the hydrophone positions (Ref. 5). The tank was then
compressional sources may be placed in the liquid and carefully filled with a well sorted sand of 250 micron

428 United States Government work not protected by copyright

diameter and leveled by careful scraping. The surface where
of the sand was visually inspected periodically by
trained divers and leveled as needed (diver 2 2
measurements indicated the water/sand interface was (V +h
within 10of horizontal).
The data were taken utilizing th.e HP 101/102 V + k 2)A =0
transducer5 which was driven linearly utilizing a 4-cycle nn
sinusoid signal from an HP-3314A signal generator 2
amplified by an ENI 1140A power amplifier. The 20 kHz 2 WPn
beam generated had a beam width of about 200 with h n + Pn
sidelobes that were about 35 clB down from the main n
lobe and located at 25o off the main lobe. The sound
pressure level (SPQ of the transducer was measured 2 W2Pn
using the in-water hydrophone located on the k
landward side of the tank. The SPL was measured to be n Pn
171 clB (rel 1pPa) at a range of 4.7 meters, (2)
approximately the transclucer-to-secliment distance
that was maintained during data acquisition. By In these equations, Xn, Pn, and Pn are the LamO
utilizing the two buried hydrophones at the landward constants and density of the nth layer, and w is the
side of the tank the se iment sound velocity was radial frequency. (Note that (Dn and An, being solutions
measured to be 1677 m/s with an attenuation of 9.8 of the homogeneous wave equation, don't include
dB/m. contributions from any point sources. These
contributions are handled separately3 as discussed
The in-sediment sound field data were taken by below).
illuminating the central probe field at a selected angle,
then by shifting forward and backward (maintaining Use of the Hankel transform then allows one to obtain
the same incident angle), the sound pressure field in integral representations of the potentials in terms of
the sediment was "mapped" oUt.4 The signal from a up and down propagating conical waves
given hydrophone was amplified (Tektronix 502
Differential Amplifiers), filtered (Krohnhite (3202)),
averaged and digitized (Nicolet 4096B with a 4562 _C' (8) Z
digitizer plug-in), and stored (HP9826). The in- (r, z) A We n + A + (s)ea"( 8) z sd (rw) ds
sediment sound pressure levels were determined from n 1)
the measured peak-to-peak voltages accounting for
individual transducer sensitivity and amplifier gains.
From the discrete point sound pressure levels, pressure A,(r,z)= B We - On(s) z + B + Weprz(s @'31 .4 (m) ds
contour lines were extracted using WAR routines6 that n 0 (3)
fit the data with splines under tension. The data
presented will be in the form of equal sound pressure where
level contour lines.

111. SAFARI a (S) = VS 2 _ h2
n n
The SAFARI program is a numerical implementation of
the method described by Ewing, et. al.1 This section On(s) = %IS 2 _ k2
gives a brief description of the concepts used in SAFARI. n
The reader is referred to Refs. 1-3 for a more complete and s is the horizontal wavenumber and the
discussion including the numerical implementation coefficients
technique. A A+,B-,B+
The medium to be examined consists of n horizontally (4)
stratified layers which are isotropic, homogeneous, and are functions of s.
viscoelastic. Time harmonic compressional sources
located at the r = 0 position of a cylindrical coordinate A solution for the total potential in any layer requires
system r, 0, z are allowed. The particle displacements in that the field produced by all the sources in that layer
the nth layer can then be written in terms of be included. For any compressional source at r = 0, the
displacement potentials (Pn, An as potential produced in an infinite medium with the
material properties of the nth layer is
I 11@ (r aA
Wn (r, z) az - r ar dr (r,-,) e 1. s J (rs) ds
a a'A 10 a"(s) (5)
- +
ar a raz where zs is the source depth. Inclusion of multiple
sources requires only a sum of such sources.

429

The potentials at any position in any layer are the parametric experiments where the amplifier
determined via Eqs. 3 and 5 if one knows the values of problems were not present (Ref. 4).
the coefficients in Eq. 4 for all s values. These values are
determined in principle by the boundary conditions at Another indication of agreement between experiment
each interface. In practice, efficient and stable and theory can be obtained by examination of the
numerical computation of these coefficients requires dashed and solid straight lines in Fig. 2. Using the data
care.3 After determination of the potentials, the points input into the NCAR routine, the horizontal
particle displacements can be determined via Eq. 1. position of the maximum pressure was established at
Furthermore, the values of the stresstensor (viscoelastic four depths within the sediment. The dashed line in
solid) or pressure (attenuating liquid) can be the experimental plots is the best fit line to these
determined at any position from the potentials. points. The dashed line has been transferred over to
the theoretical results where it can be compared with
To obtain results for comparison with the experiments the solid line which is its theoretical counterpart.
in Section 11, we placed an array of sources in a water
half space such that the incident grazing angle as well V. CONCLUSIONS
as the angular separation and relative strength
between the primary and secondary lobes of the The experiment discussed here is novel in its physical
incident beam conformed to the experimental range for these types of frequencies. This is due to the
arrangement. Preliminary investigation showed that extent and homogeneity of the sand layer as well as the
the theoretical results for in-sediment pressure lack of confining walls in the water column. The data
contours were not significantly different if we used an thus allow an unique test of theoretical predictions.
attenuating liquid model of the sediment instead of a The results indicate the utility of the SAFARI code in
viscoelastic solid model with realistically low values of examining transmission of bounded beams across
shear wave speed.4 Therefore, in the comparisons water/sediment interfaces. Furthermore, they indicate
between experiment and theory to be shown next, the that inclusion of the shear degree of freedom is not
theoretical results use an attenuating liquid sediment essential for prediction of the pressure contours within
model with wave speed and attenuation as given in unconsolidated sand sediments. As a note of caution,
Section 11. however, if the sediment cannot be treated as infinite
the inclusion of the shear degree of freedom can
IV. COMPARISON OF THEORY AND EXPERIMENT become important.4,7

The in-sediment longitudinal wave speed given in VI. REFERENCES
Section 11 along with the measured in-water
longitudinal speed of 1510 m/sec implies a critical 1. W. M. Ewing, W. S. Jardetzky, and F. Press, Elastic
grazing angle of 25.80. With this in mind, the grazing Waves in Layered Media (McGraw-Hill, New 7-ork,
angles chosen in the experiment were 51.9o, 28.8o, 1957), Chap. 4.
25.80, and 20.1o thus including both pre- and post-
critical angles. In a!l the plots shown, the origin has 2. H. Schmidt and F. B. Jensen, "A Full Wave Solution
been placed where the center of the incident beam for Propagation in Multilayered Viscoelastic Media
strikes the water-sediment interface. All contours are with Application to Gaussian Beam Reflection at Fluid-
in relative clBs. Solid Interfaces," J. Acoust. Soc. Am. 77, 813-825 (1985).

Theoretical results using SAFARI are shown on the left 3. H. Schmidt and F. B. Jensen, "Efficient Numerical
hand side of Fig. 2 for incident grazing angles of 51.9o, Solution Technique for Wave Propagation in
28.8o, 25.8o, and 20.10. Only the incident portion of Horizontally Stratified Ocean Environments," Rep. SM-
the total field in the water is shown. The right hand 173, SACLANT ASW Research Centre, La Spezia, Italy
side of Fig. 2 shows the experimental results at the (1984).
corresponding angles. Examination shows "excellent"
agreement in overall contour shape between 4. K. L. Williams, L. J. Satkowiak, and D. R. Bugler,
experiment and theory except in some portions of the "Linear and Parametric Bounded-Beam Transmission
20o data where the signal to noise ratio was low. Across a Water/Sediment interface - Theory,
Proper prediction of location and amplitude of the Experiment, and Observation of Beam Displacement,"
secondary lobe within the sediment relative to primary Submitted to J. Acoust. Soc. Am.
lobe is apparent.
S. D. R. Bugler, UTH Technical Memo, 66/87, A.R.E.
Fig. 2 also shows, as curved dotted lines, the levels of Portland, U.K., 1987.
the pressure contours within the sediment relative to
the maximum pressure of the incident beam at the 6. More information on NCAR routines can be
interface. In this regard, agreement between theory obtained by writing National Center for Atmospheric
and experiment is not quite so impressive. The levels Research in Boulder, Colorado.
seen in the experiment are always lower than predicted
by SAFARI. However, it is very possible that this is due 7. P. J. Vidmar, "Ray Path Analysis of Sediment Shear
to amplifier problems experienced while performing Wave Effects on Bottom Reflection Loss," 1. Acoust. Soc.
the linear source experiments. This may have made the Am. 68, 639-648 (1980).
peak levels in the incident beam lower than what had
been measured when the source characterization
experiments were performed. In support of this
contention, / we found much better agreement in

430

rt,

w
F-

F--J 0 > -u
(D --1 0
m 0
r-
co r-
m m 0
U) c Z
0
z

rt,

Fl- cA
m
FJ, 00
rt,
x
0 m
z z

4
CA
@
m

1.5

0.0

IL
LLJ

. ..... .1 dB
1.5 1 dB

1.5

0.0
CL
LLJ Ps
I @@L r@@,
, ME--
...... -5 dB
5 dB 15 dB
1.5F -i5 d
0.0 5.5 0.0 5.5

0.0

..... .19 dB
1.5 -26 dB -26 dIB I

1 (d)

0.0

LLI

- - - - - - - - - - - - - - - - -j

23 dB . ..... .23 dB
1.5 1 1 1 1 1 1 1 1
0.0 5 0.0 5.5

HORIZONTAL DISTANCE (M) HORIZONTAL DISTANCE (M)

Fig. 2. Theoretical and experimental results for and their experimental counterparts are shown on the
transmission of a linear beam across a water/sand right hand side. The incident grazing angles are (a)
interface. Theoretical pressure contours generated by 51.90 (b) 28.80 (c) 25.8o (d) 20. 1 o.
SAFARI are shown on the left hand side of the figure

432

REMOTE SEA BOTTOM CLASSIFICATION UTILIZING THE
ULVERTECH BOTTOM PROFILER PARAMETRIC SOURCE

L. J. Satkowiak

Naval Coastal Systems Center
Physical Acoustics Branch, Code 2120
Panama City, FIL 32407-5000

ABSTRACT system consists of a control and receiver unit, a power
amplifier, a power supply for the power amplifier, and
Computer algorithms were developed to extract a transducer. The profiler uses a parametric transmitter
certain marine sediment properties utilizing the to enerate a low frequency acoustic beam which has a
Ulvertech Bottom Profiler system. A series of normal wi3e bandwidth and a narrow beamwidth, to give high
incidence bottom reflectivity measurements was made resolution sub-bottom information. The generation of
over six test tracks. The test tracks were precise paths the low frequency acoustic waves in a parametric
that varied in length from 0.1 to 0.5 miles long over transmitter is accomplished by exciting a transducer
sediments that ranged in composition from hard sand with two frequencies separated by the required
to stiff clay to soft mud. Each track was well- difference (low) frequency. The non-linear nature of
characterized by several core samples. The Ulvertech the water will cause the two waves to interact
system uses a parametric source that has its primaries producing a sum and difference frequency. The
at about 200 kHz with the difference frequency difference frequency is of main interest to
selectable, either 5, 10, 15, or 20 kHz. The algorithms investigations requiring sub-bottom penetration. In
developed make use of the acoustic information the current study the Ulvertech profiler output two
contained in the returns from both the primary and primary frequencies, one at 200 kHz and the second at
difference frequency to extract estimates of the 220 kHz thus generating difference frequency of 20
density, grain size, and shear strength of the top 1-2 kHz.
meters of sediment. Some sample data as well as a
description of the algorithms will be presented. The Ulvertech profiler was chosen for this research for
the following reasons:

1. A narrow beamwidth at a low frequency can
be obtained from a relatively small transducer. The
1. INTRODUCTION narrow beamwidth allows only a small section of the
bottom to be sampled on any given ping. The footprint
A subject generating a ?ireat deal of interest is the on the bottom of the 20 kHz signal is only about 60
remote sensing of Sur cial sediment properties, centimeters in 10 meters of water. The small footprint
surficial referring to the first 1 to 2 meters of sediment. minimizesthe effect of multiple scattering paths.
For example, identifying regions of a water requiring
dredging requires a knowledge of the density and 2. The beam has essentially no sidelobes. This
possibly the grain size of the surficial sediments. helps eliminate multiple scattering paths.
Marine construction may require a determination of
the bearing or shear strength of the sediment. Core 3. Both the high and low frequency signals are
sampling can provide these data but it is time available on the receiver side. The high frequency
consuming, area specific (the core may only represent a primary signals are used to give a precise detection of
localized sediment), and relatively expensive. The the first mud/water echo and also to trigger the time
current work was und.ertaken to acoustically identify varying gain forsub-bottom low frequency signal.
sediment properties using an expeditious, inexpensive,
and accurate method. A series of normal incidence The main disadvantage to a parametric source is the
acoustic backscatter measurements was made in the high powers required to obtain acceptable source
Panama City, Florida, area. Core samples provided the levels. However, modern high power FET's have made
necessary ground truth used to extract the the design of high power amplifiers relatively cheap.
relationships between echo return and sediment
properties. 'in order to measure the beamwidth and frequency
distribution of the difference frequency, the transducer
2. EXPERIMENT was calibrated in the Naval Coastal Systems Center
Acoustic Test Facility. The maximum amplitude of the
The Ulvertech Bottom Profiler signal of the primary (200 kHz) signal was measured to
be 229 dB (re 1pPa @ 1m) with a beamwidth of 2.50
The bottom classifier developed uses the Ulvertech' measured at the 3dB points. The 20 kHz difference
Bottom Profiler (UBP) as an acoustic source. The UBP signal was measured to have a source level of 192 clB

433 United States Government work not protected by copyright

(re 1pPa @ im) with a bearnwidth of 3.00. The signal instead of using peak voltage values, the signals were
levels were more than adequate to overcome any summed over time, thus reflecting the total acoustic
signal loss occurring due to spreading loss, absorption energy returning to the transducer. Each channel or
in the water column,ambient noise, and/or bottom loss. time slice was corrected for background (noise) by
The absence of sidelobes helps eliminate the effect of subtracting off a background contribution. The
multiple scattering paths. In order to investigate the ma nitude of the background contribution was found
relationship between sediment shear strength and the by flitting the "non-peak" area with a straight line. We
change in shape of the return pulse it is important to found a very strong correlation between the h h
frequency signal and the density and grain size of i@
eliminate any spurious pulse stretching due to multiple t e
scattering paths. surficial (first 5 to 20 centimeters) sediment. The
following relationships were determined:
The Test Sites
Grain size (phi units)
A series of five test tracks, precisely defined utilizing 9.0785 - 7.5951 * LOG 10 (Summed signal)
readily identifiable shore reference markers as well as
their LORAN-C coordinates, were set up in the Panama Density(g/cm3)
City area bays and in the nearby Gulf of Mexico. The -0.04586 + 0.9258* LOG 10 (Summed signal)
length of these tracks varied from 0.1 to 0.5 miles long.
The sediment types varied from hard sand to stiff clay The low frequency signal (20 kHz) penetrates the
to soft mud. Each track was well characterized by sediment more deeply and allows sampling via changes
several core samples. The ana sis of these cores in the magnitude of the signal, of the density and grain
N
ielded information on the sellment grain size size of the sub-surface sediment as a function of depth.
nsity, water content, porosity, and soun velocity as' By breaking the low frequency return into equal time
a function of depth in the sediment. The core sample bins corresponding to vertical "layers" of the sea floor
analysis techniques used and the results of the analyses one can obtain a density and grain size profile. The
can be found in Reference 1. following relationships were extracted relating the
summed signal in the first of the equal time bins to the
Data Acq u isition grain size and density:
The data were taken utilizing the UBP transducer as Grain size (phi units)
both a transmitter and receiver. The output from the 2.392 - 5.166*LOG10 (Sum-sig-1)
UBP receive side was filtered, rectified, amplified, and
fed into a Nicolet 4094 Digital Oscilloscope with a Density (g/CM3)
Model 4562 dual channel digitizer. The 4562 was set to 0.729 + 0.6641*LOGlo(Sum-sig-1)
digitize at a rate of 500 kHz to avoid any possible
aliasing problems. The Nicolet d 19 itizer was set to where Sum-sig-1 is the summed signal in the first time
average over a fixed number of pings and then bin, corrected for background.
transmit the data to an HP-9020 computer for storage. There is at least one complication; one needs to correct
The effect of any platform instability was reduced by the return signal magnitude for the effects of
using the bottom return to trigger the digitizer. The attenuation in the sediment. A series of empirical
cligitization of the signal would begin approximately relationships relating the grain size of the sediment to
2-4 milliseconds prior to receiving the return signal. the attenuation has been derived by Hamilton
The cligitization of the signal would continue for 6-8 (reference 5). Since the high frequency signal is
milliseconds after the cessation of the return pulse. essentially sampling just the upper few. centimeters of
the sediment, the density and grain size values
The acoustic data were taken on two separate dates determined from this data can be used to correct the
about one month apart. The first set of data was magnitude of the signal in the first time bin of the low
analyzed and algorithms relating the properties of the frequency pulse. The density/grain size determined
return echo with the core sample (ground-truth) from that low frequency time bin can be used to
sediment properties were derived. The second set of calculate the estimated attenuation for correcting the
measurements was made to confirm the algorithms next time (sediment layer) bin. The procedure is
developed. A description of the algorithms and a repeated iteratively until the magnitude of the signal
discussion of the results follow. drops below an arbitrarily set percentage of the
maximum amplitude. In this manner the grain size and
Density and Grain Size Determination density of the sediment can predicted from the acoustic
return as a function of depth in the sediment.
The density and rain size of sediments can be related
to the magnitUe of the normal incidence acoustic Shear Strength Determination
echo. This type of relationship has been demonstrated
by several authors (references 2-4). In the data It was found that the shear strength of the sediment
acquired using the UBP, there is acoustic echo and the magnitude of the signal, or even the summed
information in both the high (primaries) and low signal discussed above, do not correlate well. However,
(difference) frequency channels. The acoustic data the shear strength does correlate well with the shape,
recorded consisted of both the high and low frequency in particular with the decaying edge or tail of the low
signals that were digitized and averaged on separate frequency return signal. The decaying part of the
channels of the Nicolet but with a common trigger. signal, or tail, can be quantified in several ways. In the
current investigation two methods were employed
with nearly equal success. One can fit the averaged

434

return signal (envelope) with a Gaussian shape with a Table 1. A comparison of sediment properties
decaying exponential tail and correlate the change in determined from core sample analysis versus estimated
the decay time constant with the shear strength. from acoustic return.
Alternately, one can determine the width ofthe return
pulse at the points where the pulse exceeds some
fraction (arbitrarily chosen to be 1/e) of the maximum FROM FROM
and correlate the ratio of the return to outgoing pulse CORE SAMPLE ACOUSTIC RETURN
width with the shear strength. Both methods yielded SITE
good correlations with about the same standard error Grain Size Density Grain Size Density
of estimate, however the latter method requires less
computational time and is easier to implement than (phi units) (g/CM3) (phi units) (g/CM3)
the former. The second method yields the following
relationship: G4 2.44 2.00 2.15 2.01
Shear Strength (g/Cm3) S6-11 1 8.50 1.37 8.21 1.28
698.9*exp [-0. 17104* (Width -ratio) I S6-2 8.67 1.18 9.00 1.19
where the Width-ratio is the time width ratio of the WP_1 9.01 1.23 8.30 1.27
return pulse to that of the outgoing pulse. Note that WP-2 8.97 1.25 8.46 1.25
since the total width of the signal is used, the -shear
strength determined in this manner is "averaged" over WP-3 9.02 1.19 8.33 1.27
the depth of penetration of the acoustic signal.
B14 2.42 2-00 2.23 2.01
3. RESULTS AND CONCLUSIONS BB-1 6.22 1.50 6.76 1.43
Two of the test tracks, denoted BU and G4, had 6.70 1.24 5.20 1.66
sediments composed 'of- medium sand with grain
diameters on the order of 180 to 250 microns. Site WP
sediment consisted of a very soft silty-clay with a small
percentage of sand at the start of the test track. Site BB Since the 56 site has such a dynamic range in sediment,
and HB contained sediments that ranged from a clayey- it was chosen for displaying the predictive capability of
sand to a sandy-clay with average densities ranging the algorithms. Figures la, 1b, and ic show the
from 1.3 to 1.7 g/cc and grain sizes ranging from 15 to predicted density, grain size, and shear strength as a
150 microns. However, due to the close proximity of function of distance along the track. The letters
the Hathaway Bridge, the Port of Panama City, and the plotted indicate the data from three separate passes
shipping channel it was difficult to maintain station for along the track. Note the reproducibility of the data
track HB and the data were non-reproducible from one from set to set. Also shown are the results of the core
test run to the next and was not used in the analysis. sample analysis for comparison. The data agree well at
Track S6 proved to be the most interesting site. The the two core sample locations and qualitatively agrees
track extended from near a marina out to a shipping with the historical nature of the track, harder
channel. The sediment near the marina was dominate sediments at the two ends and relatively soft in the
by a sand fill placed prior to construction of the clocks. middle.
Moving away from the marina the sediment became a
soft silty-clay (mud) until one approaches the dredged In summary, a series of normal incidence acoustic
channel where the sediment becomes a stiffer clayey- measurements were made using the Ulvertech
sand or a sandy-clay. Parametric Bottom Profiler as a source. Algorithms
were derived, utilizing the dual frequency nature of
Table I contains a comparison, as function of test site, the source, relating properties of the acoustic return
of the p reclicted (based on acoustic return) grain size with sediment characteristics determined from core
and density versus the values extracted from the core samples. Tests indicate that the UBP source and the
samples. These predicted values were determined from algorithms derived provide a relatively fast and
the first time bin of the low frequency return and accurate method to remotely characterize marine
represent the first 20 to 30 centimeters of sediment. In sediment.
general, the predicted values for the sediment density
and grain size agree quite well with the values
determined from the core sample analysis., Entrapped
gas, noted in the core sample analysis, on the BB track
caused some fluctuations in the data on that track. The
presence of entrapped gas causes a stronger return
from the sediment thus yielding a higher. value for the
density and grain size. However, the changing shape of
the low frequency signal combined with a relatively
unchanging high frequency signal provides an
indication of this phenomenon occurring.

435

4. REFERENCES Site S6 Bottorn Classifier Test
1. C. Ingram et. al., Marine Geological Laboratory 2.0 1 1 1 1 1 1
Report #640, Naval Oceanographic Office, August 1.9 -<-Channel Marina --- >
1986. - 1.8 -
2. K. Winn, G. Becker, and F. Theilen, "The M
Relationship Between Sediment Parameters and the 1.7 -
FIGURE 1A
Acoustic Reflectivity of the Sea-Bed," from Acoustic
mo 1.6 -
and the Sea-Bed edited by N. G. Pace, University Press,
Bath UK, 1983. >1 1.5 -
3. R. Faas, "Analysis of the Relationship Between 1.4 -
Acoustic Reflectivity and Sediment Porosity,"
Geophysics, Vol. 34, No. 4,1969. 1.3-

4. A. Kaya, A. Tsuchiya, and M. Nishimura, "Acoustic 1.2-
Precise Measurement of Physical Properties of Floating J
Soft-Mud Sediment " from Acoustics and the Sea-Bed 1.1 - X-Core Sample Data
I.J.K-ciassirier Test Data
edited by N. G.. Pa@e, Bath University Pres s, Bath UK, 1.0.
1983. 11
5. E. L. Hamilton, "Geoacoustic Modeling of the Sea 10 -<-Channel Marina-->
Floor," J. Acoust. Soc. Am., 68 (5), Nov. 1980.
9-
x
8 -

4 - FIGURE 1B

2- X-Core Sample Data
1,J,K-Ciassifier Test Data

300

N 270 -<-Channel Marina-*

240-

210-

FIGURE IC
MO 180-
0
150-

120-

90-
M
60-

30- 1 X-Core Sample Date
1,J,K-Classifier Test Data
0 1 1 f I
1 2 3 4 5 6 7 8 9 10
Position (Arbitrary Units)

Figures IA, 1B, JC show the predicted density, grain
size, and shear strength, respectively, as a function
of position along test track S6. The letters, I, J,
and K, indicate the results from three separate data
FIGURE IC
GURE IA V
j

X_C.,e Sample

j -C3.r. Sample
X
I.J,K-Cl .. Ifier

acquisition passes along the track. The letter X
indicates the values of the sediment Droperties
determined from core samples (ground-truth).

436

PARTICLE REWORKING IN GREAT LAKES SEDIMENTS: IN-S17U TRACER STUDIES
USING RARE EARTH ELEMENTS

John R. Krezoski

Center for Great Lakes Studies
University of Wisconsin-Milvaukee
Milwaukee, WI 53201

(2,3,12). Mechanisms and rates of reworking have
ABSTRACT further been determined from laboratory microcosm
studies using natural sediments spiked with
Pollen grain distributions and radioactive conservative elements or radioactive tracers (see
fallout horizons observed in sediment cores reviews in 8,9,13). These techniques have provided
collected over the past decade have served as significant insight into sediment reworking
tracers of sediment reworking and burial rates processes but they suffer from the following
for both the marine and lacustrine inherent drawbacks: (1) Data from palynological
environments. Errors associated with the and radionuclide studies, which are
analysis of these cores, attributable to geochronological in nature, are time-avexaged over
blogenic sediment reworking and shipboard tens of years or more, making it difficult to
handling, have necessitated the development of interpret recent (S 30 yr.) events; (2)
more precise techniques to measure sediment Radionuclide and pollen horizons are subject to
transport rates. We utilize a rare earth sediment focusing where net loss or gain of tracer
element tracer technique, which takes can be a function of grain size, resuspension
advantage of the high neutron-capture cross- events, and lateral transport phenomena; and (3)
section of samarium oxide, and neutron Laboratory microcosm studies are conducted under
activation analysis, to trace the burial and highly idealized conditions and are likely to be
lateral transport rates of surficial sediments unrepresentative of in situ conditions. For
in the littoral (SCUBA diver depths) and example, the sediment reworking rates measured in
profundal (manned submersible depths) of the microcosms without sedimentation (e.g. 8) are
Great Lakes. The results of these likely to be quite different from natural reworking
measurements aid in determining the mass rates. Thus, accurate descriptions of in situ
balance of contaminants in the lakes. sediment reworking and transport processes in the
Great Lakes and coastal marine environment are
1. INTRODUCTION still largely unavailable even though such
information is very important in determining the
Sediment reworking and transport processes in the mass balances and final resting places of
profundal Great Lakes and coastal marine contaminants and other anthropogenic components
environments are important features that influence that have been added to marine systems.
post-depositional distributions of pollen grains,
diatom frustules, and atmospherically-derIved An alternative but virtually unused approach for
radionuclides which are used to interpret the determining profundal or pelagic sediment reworking
sedimentary record and determine recent and transport rates Is through In situ tracer
sedimentation rates, regional historical events, experiments, using environmentally safe but easily
and fates of organic contaminants in the respective detectable tracers, of grain size and density
environments (1-4). Considerable attention has identical to the fine-grained sediments being
been-focused on understanding modes and rates of studied. Marine and freshwater investigators have
post-depositional sediment reworking in the Great employed this technique by placing colored
Lakes and marine environments with the result that particles, microspheres or other materials at the
most theoretical models describing particle sediment-water interface and monitoring their
dynamics or reconstructing the history of pollutant movement over time. By microscopically examining
input to the marine systems, based on sediment particle distributions in infaunal fecal material
profiles, include surficial sediment reworking and in sediment cores taken from the study areas,
terms (1,2,5-11). much has been learned about benthic organism
feeding and life habits and sediment redistribution
Present knowledge of sediment redistribution and processes (13-17).
.reworking In the Great Lakes and other aquatic and
marine systems stems largely from comparative A serious drawback to this technique has been the
studies of pollen-grain or radionuclide Inability to accurately follow a tracer once It is
distributions preserved in the sedimentary record diluted by the surrounding sediment (see 13).

CH2585-8/88/0000- 437 $1 @1988 IEEE

After several days, the labeled particles become 8046.581W), at 125m depth (24). This location was
harder to find and estimates of their densities selected because the sediments were known to be
become increasingly difficult and time consuming. fine-grained, the area was representative of other
Thus, long-term studies using this technique have profundal regions of the Great Lakes where blogenic
not been possible and errors in estimates of tracer reworking had been reported, and it was one of the
particle densities in reworked sediments have been few regions in Lake Superior known to support
high. standing stocks of benthos which approximate those
found in profundal regions of the lower Great Lakes
In order to overcome the problem of tracking the (25).
tracer particles, to increase the precision of the
in situ tracer technique, and to automate sample The objectives of the experiment were to: (1)
analysis, Krezoski (18) developed a radiometric carefully place a thin layer of REE tracer at the
technique using selected lanthanide series (rare sediment-vater Interface in an area of undisturbed
earth) elements as sediment tracers. Rare earth fine-grained profundal sediments; (2) leave the
elements (REE's), which axe found naturally in study area and allow the marked sediments to react
soils and sediments in trace quantities, are non- to ambient conditions for a pre-determined length
hazardous, readily detectable analogs of fine- of time; (3) return to the site later to collect
grained sediments. Moreover, because of their high undisturbed sediment samples, and (4) analyze
neutron-captuxe cross-sections, they can be tracer distributions in sediment cores taken from
detected at nanogram quantities when examined by within and around ("I m radius) the marked area to
instrumental neutron activation analysis (INAA: see determine rates of lateral and vertical transport
19). of the tracer compound.

Another difficulty that currently impedes use of in The REE tracer was prepared as a pellet
situ tracer techniques in deep water is the need to encapsulated in a cylindrical cake of ice (see 26).
relocate and return to a precise sediment study This method permitted the tracer to be carried to
site. Most in situ tracer experiments have been the lake bottom without any loss and allowed the
conducted with the aid of divers and thus have been tracer particles to be gently deposited at the
limited to relatively shallow, diver-accessible sediment-water interface as the ice melted (Fig.
depths (@30 m). Use of this tracer technique by 1).
divers in the high-energy littoral waters of
Sturgeon Bay, Lake Michigan (see 18) was severely
compromised by resuspension of the sediments by
wind-generated waves (20,21). Additionally, the
sediments at these locations are primarily coarse-
grained deposits. Thus, very little information is
available on In situ infaunal and physical
reworking of fine-grained deposits at profundal
depths OtI00 m). Application of deep submergence
research vehicle (DSRV) technology, however, makes 15 cm
in situ tracer studies at profundal depths readily
possible. DSRV's are quite useful and are becoming
increasingly available for limnological and marine
research and have been shown to be effective tools
for work in pelagic environments (22). Highly
sophisticated manipulator arms and underwater video 20 Cm
and photographic systems permit sample collection
and underwater observation which was never Figure 1. Schematic diagram of rare earth element
available before (23). Additionally, underwater tracer pellet.
acoustical beacons can permit return to exactly the
same site on the order of years later. This paper To begin the experiment, the frozen pellet was
presents the results of an in situ REE tracer study carried to the lake bottom by the DSRV Johnson-Sea-
conducted in a profundal region of Lake Superior Link II and was placed on the lake floor. An
using a DSRV to determine actual sediment burial underwater beacon (37 kHz acoustical pinger; Helle,
and transport rates In a fine-grained depositional Inc.) was deployed 0.5 m from the pellet. While
basin and evaluates the utility of tracer the pellet was still frozen (00 minutes after
experiments at profundal depths using REE and DSRV deployment), control sediment samples were
methodologies. collected -3 m away from the study site (7.6 cm
diameter, DSRV Alvin style punch cores) and a
2. METHODOLOGY precise fix on the study site was obtained with the
aid of acoustical positioning (Honeywell Hydrostar)
The sediment tracer study was conducted in the Ile and loran C navigation (North Star 7000) systems
Parisienne depositional basin, a major north-south aboard the surface support vessel, the R/V Seward
trough, in southeastern Lake Superior, 40 km Johnson.
northeast of Sault Ste. Marie, Michigan (46 0 43,131H

438

After 23 days, the DSRV was re-deployed at the Ile
Parisienne station where it descended to the bottom
near the acoustical beacon. Using an on-board
sonar system to home in on the pinqer, the DSRV
maneuvered to the tracer pellet location and 7+0.4
collected eleven undisturbed punch cores. On 5.8+0.3
shipboard the cores were sectioned in I cm 16.5+0.8
intervals and frozen for analysis ashore. 1.9+0.1
3.9+0.2
In the laboratory, the core sections were dried, 20+0.1
weighed (�0.1 mg), and evaluated by instrumental
neutron activation analysis (INAA) at the CENTER
University of Wisconsin-Madison Reactor Lab and the 10.7+0.5 60.1+3.0
University of Wisconsin-Milwaukee Center for Great 30.5+1.5
is5es Studies as described by Krezoski (26). The
Sm gamma ray activity (tl/2=46.8 h; u=5820tlOO
Barns) was measured using a Canberra Series 80
multichannel analyzer and a GeLi detector (see 19).
Calibration standards indicated that the gamma
spectrometer had a counting efficiency of 0.3%. Figure 2. Schematic representation of samarium
distribution pattern at REE study site after 23
3. RESULTS AND DISCUSSION days. Center of figure represents location of
tracer pellet. Vectors indicate compass 1
Deployment of the REE tracer pellet and pinger at direction and concentration of Sm (Ag*g ) in
the beginning of the experiment was successful and cores collected I m from the center.
from the control cores collected at the site, the
natural background concentration of Sm in the Advective transport of the tracer material which
sediments was found to be 5.5 gg .g . This result was deposited responded to bottom currents which
falls within the range predicted by Helmke (19) and were not detectable by the DSRV1s current meter
indicates that the 4.3 g of Sm (5 g SM203) added to (located -2 m above the sediment-vater interface;
the study site via the tracer pellet increased the calibrated �0.1 kt) or the observer's eye.
final concegtratioi of the experimental sediments
to 1.7 x 10 pg' 94 - Thus, the tracer sediments Advective transport of the tracer particles can be
contained -3 x 10 greater Sm than the surrounding shown, assuming the flux is one-dimensional, by:
sediments and, as planned, the tracer sediments
could be distinguished from surrounding sediments dC/dt = 2(dC/dz)
by the observation of Sm concentratirs which
exceeded the background of 5.5 pg*g where a-is the velocity of particle movement
(cm'sec ), C is the concentration of excess tracer
At the end of the 23 day experiment the Johnson- (samarium) In the sediment, t is the elapsed time
Sea-Link If landed within 90 m of.the_jinger and of the experiment (23 days) and z is the distance
began moving toward it at -0.1 km hr. . At -5 m from the tracer pellet to the punch core sample
distance the pinger and marker line became clearly location (-1 m). The velocity, 2, is thus
visible and at -2 m distance the DSRV halted and proportional to dC/dt and dC/dz. From linear
the eleven core samples were taken. Two cores were least-squares analysis_ T! tie data, it was found
collected within -3 cm of the polypropylene line that 4C/dt=-0.085 gg*g s and dC/dz=-49.9
marker (to sample the remains of the tracer pellet) gg*g-L'm-1 (see 26). Thus, the average particle
and the remaining nine cores were collected around velocity of the REE particles was found to be 0.17
the perimeter of a "2 m diameter circle cm*s
circumscribing the tracer pellet (Fig. 2). A
twelfth core sample 5) was lost when it's check It is apparent that the ice pellet technique needs
valve malfunctioned. to be more fully developed or an alternative
methodology must be employed (e.g. 27) and that a
A Sm Inventory was calculated to determine the mass sensitive current meter must be deployed close to
of initial tracer pellet disbursed over the study the experiment so that advective transport
area (see 26). Compared with the original mass of processes can be distinguished from eddy-diffusive
Sm in the tracer pellet (4.3 g), it was apparent processes and anomalies in mass balance
that at the end of the 23 day experiment slightly determination can be correlated with physical
less than 1% of the Sm originally placed at the events.
study site was accounted for. Subsequent work with 4. ACKNOWLEDGMENTS
an ROV has shown that uneven ice pellet melt may
have occurred such that the pellet, when mostly I wish to thank D.N. Edgington, J.V. Klump, G.E.
dissolved, becomes detached from the small Glass, L.F. Boyer, R.A. Cooper, W. Cooper, T.M.
counterweight and floats up the I m marker line. Askew, and R.M. Oven for their helpful suggestions.
P.A. Helmke and R.J. Cashvell provided me with
*@� 0 4
19+ 0.1
3 9 � 0 @2

439

critical information on instrumental neutron 10. Robbins, J.A. 1986. A model for particle-
activation analysis techniques. This research was selective transport of tracers in sediments
supported by grants from the U.S. Environmental with conveyor belt deposit feeders. J.
Protection Agency (CR813538-01-0), the University Geophys. Res.- 91(C7):8542-8558.
of Wisconsin Sea Grant Program (R/MW-38-PD), the
NOAA National Undersea Research Program and the 11. Christensen, E.R., and Bhunia, P.K. 1986.
Center for Great Lakes Studies, University of Modeling radiotracers in sediments:
Wisconsin-Milvaukee. Comparison with observations in Lakes Huron
and Michigan. J. Geophys. Res, 91(C7):8559-
5. REFERENCES 8571.

1. Guinasso, N.L., and Schink, D.R., 1975. 12. Davis, R.B. 1974. Stratigraphic effects of
Quantitative estimates of biological mixing tubificids in profundal lake sediments.
rates in abyssal sediments. J. GeoRhys. Res. Limnol. Oceano9r. 19:466-488.
80:3032.
13. Fisher, J.B., Lick, W.J., McCall, P.L., and
2. Robbins, J.A., and Edgington, D.N. 1975. Robbins, J.A. 1980. Vertical mixing of lake
Determination of recent sedimentation rates in sediments by tubificid oligochaetes.
Lake Michigan using Pb-210 and Cs-137. Geophys, Res. 85:3997-4006.
Geochim, Cosmochim. Acta. 39:285-304.
14. Aller, R.C., and Dodge, R.E. 1974. Animal-
3. Robbins, J.A., Edgington, D.N., and Kemp, sediment relations in a tropical lagoon
A.L.W. 1978. Comparative lead-210, cesium-137 Discovery Bay, Jamaica. J. Mar. Res. 32:209-
and pollen geochronologies of recent sediments 232.
of lakes Erie and Ontario. Quat. Res. 10:256-
278. 15. Winston, J.E., and Anderson, F.E. 1971.
Bioturbation of sediments in a northern
4. Swackhamer, D.L., and Armstrong, D.E. 1986. temperate estuary. Mar. Geol. 10:39-49.
Estimation of the atmospheric and
nonatmospheric contributions and losses of 16. Self, R.F., and Jumars, P.A. 1978. New
polychlorinated biphenyls for Lake Michigan on resource axes for deposit feeders. J. Mar.
the basis of sediment records of remote lakes. Res. 36:627-641.
Environ. Scl, Technol. 20:879-883.
17. Jumars, P.A., Self, R.F.L., and Nowell, A.R.M.
5. Schink, D.R., and Guinasso, N.L., Jr., 1978. 1982. Mechanics of particle selection by
Redistribution of dissolved and adsorbed tentaculate deposit feeders. J. Exp. Mar.
materials in abyssal marine sediments Ecol. 64:47-70.
undergoing biological stirring. Am. Jour. Sci.
278:687-702. 18. Krezoski, J.R. 1985. Particle reworking In
Lake Michigan sediments: In situ tracer
6. Robbins, J.A., McCall, P.L., Fisher, J.B., and measurements using a rare-earth-element.
Krezoski, J.R., 1979. Effects of deposit Abstr. 28th Conference on Great Lakes
feeders on migration of 137-Cs in lake Research, Milwaukee, WI.
sediments. Earth Planet. Sci. Lett. 42:277-
287. 19. Helmke, P.A. 1982. Neutron Activation
Analysis. In Methods of Soil analysis, Part
7. Christensen, E.R. 1982. A model for 2. Chemical and Microbiological Properties.
radionuclides in sediments influenced by (2nd edition)-, eds. A.L. Page, R.H. Miller,
mixing and compaction. J. Geophys. Res, and D.R. Keeny. Am. Soc. Agron.-Soil Sci. Soc.
87(Cl):566-572. Am., Agronomy Monograph no. 9.

8. Krezoski, J.R., Robbins, J.A., and White, D.S. 20. Thomas, R.L., Kemp, A.L.W., and Lewis, C.F.M.
1984. Dual radiotracer measurement of 1973. The surficial sediments of Lake Huron.
zoobenthos-mediated solute and particle Can. J. Earth Sci. 10:226-271.
transport in freshwater sediments. J. GeophZl,
Res. 89(B9):7937-7947. 21. Sly, P.G., and Thomas, R.L. 1974. Review of
geological research as it relates to an
9. Krezoski, J.R., and Robbins, J.A. 1985. understanding of Great Lakes limnology. J.
Vertical distribution of35eeding and particle- Fish. Res, Board Can, 31:795-825.
selective transport of Cs in lake sediments
by lumbriculid oligochaetes. J. Geophys. Res, 22. Rechnitzey, A.B. 1986. On the upward trend in
90(C6):11999-12006. manned submersible use. Sea Technol. 27:10-13.

440

23. Cook, R.W. 1986. Underwater Imaging: One
user's perspective. Sea Technol, 27:29.

24. Thomas, R.L., and Dell, C.I. 1978. Sediments
of Lake Superior. J. Great Lakes Res. 4:264-
275.

25. Dermott, R. 1978. Benthic diversity and
substrate-fauna associations In Lake
Superior. J. Great Lakes Res. 4:505-512.

26. Krezoski, J.R., 1989. Sediment reworking and
transport in eastern Lake Superior: In-situ
rare earth element tracer studies. J. Great
Lakes Res. (in press).

27. Tudhope, A.W., and Scoffin, T.P. 1987. A
device to deposit tracer sediment evenly on
the deep sea bed. J. Sed. Petrol. 57:761-762.

441

IN-SITU TRACER STUDIES OF SURFICIAL SEDIMENT TRANSPORT IN THE GREAT LAKES
USING A MANNED SUBMERSIBLE

John R. Krezoski

Center for Great Lakes Studies
University of Wisconsin-Milvaukee
Milwaukee, WI 53201

ABSTRACT deployment. The results demonstrated that not
only were the depths far from q917t (calculated
Rare earth element (REE) tracer "pellets" were bottom currents were -17 cm*sec , but also that
fabricated by mixing -5 g samarium oxide and/or -2 deep submergence research vessel techniques are
g europium oxide with 250 ml natural (wet), fine- continuing to open the doors to exciting and
grained sediment, freezing the mixture (in -1 cm detailed studies of geochemical processes and
x -10 cm diameter cakes) and encapsulating them In animal-sediment interactions at the sea floor In
weighted, 15 ca x 20 cm diameter cylinder of Ice. fine-grained depositional areas. The implications
of this work are that well established theories
In initial studies (1985) using samarium as a regarding seafloor properties and processes,
tracer, a pellet was deployed on the floor of the especially In the Great Lakes, are being
Ile Parlsienne basin (125 m depth) of eastern Lake challenged and that commonly held beliefs
Superior using the Johnson-Sea-Link II (JSL 11). regarding the ultimate fates of those toxic and
As the ice melted, most of the samarium, a stable, hazardous materials which bind to fine-grained
high neutron capture cross section REE, added at a sediments will have to be more closely
concentration approximately lOE4 greater than scrutinized.
found naturally In Great Lakes sediments, settled
to the lake bottom and permitted tracking of in-
situ sediment transport in response to biogenic
and physical disturbances at the profundal depths.
The labeled area was marked by a -1 m strand of
weighted polypropylene line, and a 37 kHz
acoustical beacon was placed approximately 1 m to
the east of this point.

Tventy-three days after tracer deployment the
study site was reoccupied by the JSL II and 11
punch cores (7.6 cm diameter), collected from
within and around the labeled area, were sectioned
in I cm intervals to 10 cm and were analyzed by
neutron activation analysis. High resolution
gamma spectroscopy of the labeled sediments
revealed that little reworking had occurred during
the 3.3 vk period but suggested that longer term
studies were necessary and could be successful in
the quiet depths.

Subsequent studies (1986-1988), in the same region
of the Great Lakes, indicated that, indeed, the
labeled area could be relocated and sampled on the
order of two years later. In these studies, dual
tracers (Sm and Eu) were used (three.replicate
pellets) and a pinger with a three-year battery
pack was deployed. Moreover, careful loran-C and
Satnav coordinate measurements aided greatly in
relocating the study area.

Tventy-four punch cores were collected and
detailed photographic and video Images were taken
in July, 1988, 2 years and 10 days following

CH2585-8/88/0000-442 $1 @1988 IEEE

VIDEO-SEDIMENT-PROFILE CAMERA IMAGERY
IN MARINE AND FRESHWATER BENTHIC ENVIRONMENTS

Larry F. Boyer

Dept. of Geosciences, and Center for Great Lakes Studies
University of Wisconsin-Milwaukee
P.O. Box 413, Milwaukee, Wisconsin 53201

In-situ video-sediment-profile camera provides views of the sediment-water
(VSPC) images will be displayed, "captured" interface (SWI) that no other tool can
using a combination of manned submersible provide, coupled with the precise choice
sampling techniques, and innovative charge- of sampling provided through manned
coupled-display technology. These live submersibles. The VSPC has already been
images provide an intimate perspective of significant in designing sampling strategy
benthic sedimentary structures, processes, in the investigation of biogenic
and dynamics from two very different disturbances (biogenic sedimentary
benthic regimes. Imagery from 350m in the structures, ie. burrows, etc.) and their
Caribou Island Basin of Lake Superior (1986 effects on infaunal community structure in
dives) will be compared to that from 750m the Great Lakes (380m) and on the southern
depths on the upper continental slope off New England upper continental slope (4,5).
southern New England. Details of the These images complement and enhance our
mesotopographic relief (mms-cms) at the traditional sampling methods. Future high
sediment-water interface (SWI), epi- and resolution direct image capture and
infaunal organisms, their densities and manipulation will allow quantitative
spatial distributions may be addressed analysis of these images.
through this technique. Megabenthic burrow
structures were easily identified from the
sub, and the physical and biogenic PURPOSE AND GOALS
sedimentary microstratigraphy were then
examined in detail with the VSPC system The purpose of this presentation is to
down to 15-20cm within the sediments. exhibit live video VSPC imagery from two
profundal environments. As such, this paper
represents an extended abstract, as the
real focus impact is in the viewing of the
INTRODUCTION images. The utility of these VSPC
techniques for studying larger megafaunal
There exists a need for detailed and infaunal.biogenic sedimentary
qualitative and quantitative information, structures in deep marine and freshwater
derived from all possible complimentary environments is demonstrated.
techniques, on the nature of the physical,
biological, and chemical properties of the
sedinent-water interface, and the processes STUDY SITES
that occur at or in close proximity to this
critical benthic interface (1). This These studies were conducted in two
boundary between fluid and sediment is a locations. The first location, sampled in
crucial one, as it is an active region of 1985,186, and 188, is the Caribou Island
transport, resuspension, and deposition of Basin in Lake Superior, approximately 8.5
materials that are processed, mixed, and nautical miles WSW of Southwest Bank (Lat
altered physically and chemically by a 47 12.031N, Long 86 04.801W, Figure 1) in
diverse suite of benthic organisms (2). The 350m depth (6,3). In this area, sediments
video-sediment-profile camera is yet above Precambrian bedrock consist of tills
another tool enhancing our perspective of and glacial gravels, glacial clays, and
this interface (3). post-glacial clays draped with a thin
veneer of recent fine-grained lacustrine
Video-sediment-profile camera (VSPC) sediments and organics (7). This area
imagery provides the only real-time in-situ contains numerous biogenic sedimentary
view of qualitative and quantitative structures created by deep-water sculpin
structures and processes indeep marine and (Myoxocephalus thorpsoni, Figure 2) and the
freshwater environments in concert with burbot (Lota lota, Figure 3) (8). The
manned submersibles. This technique second site is an upper continental slope

CH2585-8/88/0000- 443 $1 @1988 IEEE

LAKE SUPERIOR
in a zone from 653-1290m where the red crab
P Geryon Quinquedens is the dominant
megafaunal invertebrate both in terms of
density and biomass (9). The surface of the
sediments on the upper slope is
characterized by many mounds and shallow
depressions (Figure 5). The most common
WISCONSIN features are produced by Geryon, and may
include depressions and complexes up to a
meter in diameter (10,11,12). This species
is known to actively burrow into the
sediment seeking shelter and food (13) and
can significantly alter the physical and
1 10 20 30 40 KM ONTARIO biological structure of the seafloor.

4. Zp,

@A, @F,
on

Figure 1. General and specific dive
locations for VSPC studies in Lake
2",
Superior, USA.

Figure 3. Burbot (Lota lota) trench-like
burrow system found in the Lake Superior
dive area shown in Figure 1. Burbot in
residence: fish is approximately 50cm in
total length.

CAPE CW

HOLE

Figure 2. Bottom microtopography from the
Lake Superior dive site shown in Figure 1.
Dish-shaped burrows are produced by the
deep-water sculpin (Myoxocephalus
thompsoni) in approximately 350m water
depths. "Burrows" are about 8-15 cm
diameter.
Figure 4. Dive site on the old DOS station
bottom station off southern New England location on the upper continental slope off
(Figure 4), located approximately 100 miles southern New England in approximately 750m.
equidistant from Martha's Vineyard, Block
Island, and the tip of Long Island, at a
depth of 750-800m. This site sits squarely

444

MATERIALS AND METHODS

All of the experimental dive work was
conducted using Harbor Branch Oceanographic
Institution's vessels, the R/V Seward
Johnson and the DSR/V Johnson Sea Link II.

The total VSPC system includes the
prism with an aluminum "skin", handles,
radon lights, a CCD camera, waterproof
housing and cable terminations, cable
_71
-! 1/ -0 system, control panel, a 411 black and white
SONY Watchman, a broadcast quality JVC
-VHS VCR with audio dubbing, remote
mini
control, and 20 minute cassettes (Figure 6)

Figure 5. Red crab (Geryon quinquedens)
burrow complexes typical of those examined
with the VSPC at the dive site shown in VIDEO PROFILE CA@IERR SYSTEM
Figure 4. Burrow complexes may be as much
as one meter across; burrow openings are CONTROL PANEL THROUGH
VICEO
typically 10-15cm in diameter. SEDIMENT-
111@11E
M.RA,SYST:M
d..

Ik-

PREVIOUS VSPC STUDIES

d
A low-resolution, CCD VSPC system
designed specifically for use with manned RECOROE
submersibles was deployed as a prototype on MICROPHONE REMOTE CONTROL soNy
IATCHMRN
NOAA's NURP-UCAP Lake Superior dive
projects in 1985 and 1986. This system
provides instantaneous in-situ information
on the SWI, and aids in the precise Figure 6. Schematic of prototype VSPC
sampling with tube cores, box cores, system, with components connected as in
enabling deployment of in-situ operation aboard the submersible.
manipulative experiments in known sub-
surface sediment conditions. Specifically,
vertical and horizontal heterogeneity in
sediment thickness, general bioturbational Deployment of the VSPC is similar to
features, and major sedimentary structures the sequences involved in taking a sub-
were visible from primary imagery in real operated box core, except that the prism
time, on scale from tens of centimeters to must go in as orthogonally to the SWI as
approximately one centimeter (3,14). possible. The VSPC is lifted clear of its
receptacle and placed carefully above a
In 1987, deployments in a much larger selected area of the seafloor to be
scale bioturbational system allowed sampled. The prism is lowered into the
detailed observation, video observation and sediment gently; with all power, lights,
image enhancement of red crab burrow and VCR system and monitor on, both the
complexes off the southern New England observer and the sub pilot can see the
upper continental slope (11,15). In 1988, prism entering the sediment directly from
through NURP-UCAP support, we have the sphere and also from the view of the
increased the resolution with a SONY XC-77 internal CCD camera. Therefore,
CCD (570h x 475v TV lines), enhanced our immediately apparent if the system works,
image acquisition and storage with a and real-time, in situ observation of the
TRUEVISION MS Targa ICB board and TIPS SWI can be recorded. Information on time,
software, and our ability to quantitatively sample number, locality, and other
analyze the imagery with IMAGEPRO software. parameters can be added to the VSPC
This VSPC work confirms the concept that a original tape on the audio track and this
detailed view of the SWI and local benthic
R

environs requires information from many information can be cross-referenced with
complimentary techniques. the main sub VCR audio tape record (3).

445

RESULTS Center for Great Lakes Studies, University
of Wisconsin-Milwaukee.
During 12 dives with the DSR/V Johnson
Sea Link I to 750 m on the southern New REFERENCES
England upper continental slope, and 7
dives with the JSL II in Lake Superior we (1) Rhoads, D.C., and Boyer, L.F. 1982.
were able to "dissect" the burrow The effects of marine benthos on physical
structures of the red crab (Geryon properties of sediments: A successional
cruinqueden@q) and the V-shaped, flared perspective. p. 3-52. In P.L. McCall, and
burrows of the burbot (Lota lota) by M.J.S. Tevesz [eds.], Animal-sediment
sequentially coring with tube cores and box Relations. Plenum Press.
cores down the axes of these systems, and
inside and outside of the biogenically (2) McCall, P.L., and M.J.S. Tevesz.
influenced areas (4,5,12). At the same (eds.) 1982. Animal-Sediment Relations:
tine, we obtained live, in-situ VSPC Plenum Press.
imagery. The VSPC system was deployed in
the same locations, or the same sequence of (3) Boyer, L. F., and Hedrick, J. A.
locations in similar burrow complexes, as 1989. Submersible-deployed video-sediment-
were the samples from the coring transects. profile camera system for benthic studies
A video excerpt from the dive sequences in (Special Lake Superior submersible
Lake Superior and The New England Slope symposium: Jour. Great Lakes Research).
will illustrate the microtopographic relief
and textural details created by these two (4) Boyer, L.F., Grassle, J.F.,
bioturbators at the SWI as well as the Whitlatch, R.B., and Zajac, R.N. 1988. Red
details of their respective burrow crab bioturbation on the Southern New
complexes. England Slope: The L.A.S.S.E. project. East
Coast Benthic Ecology Meetings, Abstracts
with Programs, April 8-11, Portland, ME.
SUMMARY
(5) Whitlatch, R.B., Zajac, R.N., Boyer,
1) The VSPC images provide a unique and L.F., and Grassle, J.F. 1988. The effects.
complimentary perspective of the profundal of the Deep sea red crab (Geryon)
sedimentary habitat. Disturbance on the southern New England
upper continental slope. East Coast Benthic
2) Video images provide real-time Ecology Meetings, Abstracts with Programs,
information on the spatial heterogeneity of April 8-11, Portland, ME.
the SWI, and thus allow for choice in
selectively sampling the profundal benthic
environment from manned submersibles or (6) Boyer, L.F., and Whitlatch, R.B.
ROVs at critical scales. 1989. Organism-sediment relationships in
the Caribou Island Basin, Lake Superior,
3) The VSPC system, in conjunction with U.S.A. (Lake Superior submersible
the sub, allows examination of the micro- symposium: Jour. Great Lakes Research).
structural details of physical and biogenic
sedimentary features that are sparsely (7) Matheson, D.H. and Munawar, M. 1978.
distributed. Lake Superior Basin and its development.
Jour. Great Lakes Res. 4(3-4):219-263.
4) vsPc images should not replace other (8) Boyer, L.F., Cooper, R.A., Long,
sampling techniquesf but rather be used as D.T., and Askew, T. 1989. Bioerosion and
a unique and very complimentary view of shelter formation by the burbot Lota lota
benthic systems. in the deep basins of Lake Superior. (Lake
Superior submersible symposium:Jour. Great
Lakes Research).
ACKNOWLEDGEMENTS
(9) Haedrich, R.L., G.T. Rowe, and P.T.
Ocean Instruments provided the VSPC Polloni. 1980. The megabenthic fauna of the
prototypes, NOAA's NURP-UCAP the ship and deep-sea south of New England. Deep-Sea
sub crews of the R/V Seward Johnson and Res.
DSR/V1s Johnson Sea Link I and II, the
University of Wisconsin provided (10) Grassle, J.F., H.L. Sanders, R.R.
photographic and cartographic services. Hessler, G.T. Rowe and T. McLellan. 1975.
This work was funded in part by the Patterns and zonation -- a study of the
University of Wisconsin Sea Grant Institute bathyal megafauna using the research
under grants from the National Sea Grant submersible Alvin. Deep-Sea Res. 22: 457-
College Program, NOAA, U.S. Department of 481.
Commerce, and from the State of Wisconsin.
Federal Grant NA8YAA-D-0065, Project R/MW-
38-PD. NURP-UCAP PSA grant 1987,188. This
article is contribution # 325 from the

446

(11) Wigley, R.L., and R.B. Theroux. 1981.
Atlantic continental shelf and slope of the
United States--Macrobenthic invertebrate
fauna of the middle Atlantic Bight region--
Faunal composition and quantitative
distribution. U.S.G.S. Prof Paper 529-N,
198 pp.

(12) Boyer, L.F., and McCall, P.L. 1988.
Video-Sediment-Profile Camera (VSPC)
perspectives of Red Crab burrow systems.
East Coast Benthic Ecology Meetings,
Abstracts with Programs, April 8-11,
Portland, ME.

(13) Valentine, P.C., J.R. Uzmann, and
R.A. Cooper. 1980. Geology and biology of
Oceanographer submarine canyon. Mar. Geol.
38:283-312.

(14) Rhoads, D.C., and Germano, J.D.,
1986, Interpreting long-term changes in
benthic community structure: a new
protocol. Hydrobiologia. 142:291-308.

(15) Boyer, L.F., R.J. Diaz, and J.A.
Hedrick. This issue. Computer image
analysis techniques and video-sediment-
profile camera enhancements provide a
unique and quantitative view of life at or
beneath the sediment-water interface.

447

COMPUTER IMAGE-ANALYSIS TECHNIQUES AND VIDEO-SEDIMENT-PROFILE CAMERA
ENHANCEMENTS PROVIDE A UNIQUE AND QUANTITATIVE VIEW
OF LIFE AT OR BENEATH THE SEDIMENT-WATER INTERFACE

Boyer, L.F.1, Diaz, R.J.2, Hedrick, J.D.3

1. Geol. Dept. & Centr. Grt. Lakes Studies, UWM, Box 413, Mil, WI 53201
2. V.I.M.S., William and Mary, Gloucester Pt., VA 23062
3. Ocean Instruments, 5312 Banks Street, San Diego, CA 92110

A video-sediment-profile camera system for providing in-situ sediment sampling
in-situ imaging of bottom sediments has strategy, however, the resolution of the
evolved in 3 stages: (1) A low-resolution stored imagery was poor (2). In 1986
(280h x 350v) CCD system with 1/211 VHS as deployments, optical enhancements and
the storage/viewing medium; (2) enhanced manual digitization of the images provided
optics and manual digitization of the a much better, yet still low-resolution
stored images; and, (3) a higher picture of the benthic sedimentary
resolution CCD (570h x 475v) system--images structures in Lake Superior (2). In 1987,
are recorded on broadcast-quality 1/211 VSPC deployments were in a much larger-
tape, captured frame-by-frame on a scale bioturbational system, which allowed
TRUEVISIONTM image capture board and detailed observation, video documentation
altered with TIPSTM and IMAGEPROtM and image enhancement of red crab (Geryon
software. The VSPC was used in Lake quinquedens) burrow complexes at 750m off
Superior in 185, 186,and 1188 to probe the southern New England upper continental
burbot (Lota lota) biogenic structures, and slope (1,3,4,5).
in 1987 on the upper continental slope off In our 1988 dive sequences, examining
so. New England to dissect red crab (Geryon burbot (Lota lota) trench systems in 350m
quinguedena) burrows. Two image sequences in the deep basins of Lake Superior, we
are analyzed for contrast and dynamic increased the video resolution with a SONY
range, then contrast enhancement, and XC-77 CCD (570h x 475v) camera module. In
histogram equalization were applied to addition, we have enhanced our image
enhance sedimentary structures. acquisition and storage with a TRUEVISIONTM
M8 Targa image capture board (ICB) and
TIPSTM software, and our ability to analyze
the imagery with IMAGEPROTM software.
INTRODUCTION In this paper, the original video
images will be displayed to illustrate the
This paper very briefly describes the narrow ranges of contrast and low dynamic
developmental stages and deployments made range present in both the Lake Superior and
with a prototype charge-coupled-display New England Slope sediments. After
(CCD) video-sediment-profile camera (VSPC) analyzing grey scale ranges to pinpoint the
system, specifically designed for use with range necessary to maximize image contrast,
manned submersibles (1,2). The major the images will be enhanced through
thrust of this work was to examine the histogram equalization. In the future, RPD
original in-situ video images within area, depth, and the size of physical and
biogenic sedimentary structures from two biogenic sediment structures, etc., can be
selected profundal marine and lacustrine calculated from the images after these
environments (1,2,3,4). The original video enhancements. Video-sediment-profile
was captured frame-by-frame, and then camera (VSPC) imagery provides the only
enhanced and manipulated. real-time in-situ view of qualitative and
The VSPC imaging system has undergone quantitative structures and processes in
three stages of prototype development. A deep marine and freshwater environments in
low-resolution (280h x 350v TV lines) B&W concert with manned submersibles (1,2).
CCD, fitted to a custom housing,
incorporating a solid acrylic prism, was STUDY SITES
successfully deployed from the DSR/
Johnson Sea Link II in 1985 at 350m in the These studies were conducted in two
Caribou Isia_@d Basin of Lake Superior locations; the Caribou Island Basin in
(1,2). This real-time video imagery of Lake Superior, and an upper continental
conditions at and below the sediment-water slope bottom station off southern New
interface (SWI) was extremely useful for England (1,2,6). These areas contain

CH2585-8/88/oooo- 448 $1 @1988 IEEE

numerous biogenic sedimentary structures-- inside the burrow. The burrow and internal
Lake Superior has burrow complexes and sedimentary features are extremely
trenches created by the burbot (Lota lota); difficult to discern due to low contrast
a zone from 653-1290m on the slope contains and dynamic range in the ambient sediments.
depressions and complexes up to a meter in Figure 2c is an histogram of grey scale
diameter, created by the red crab (Geryon dynamic range and spread (ie. max of 0-256
quinquedens). These organisms can levels) in the burrow sediments along a
significantly alter the physical and profile transect. Note the very narrow peak
biological structure of the bottom (low contrast range), yet relatively high
sediments (l,2,3,4,5,6,7). values per index (dynamic range within the
available grey level). To enhance the
MATERIALS AND METHODS internal features (Figure 2d), histogram
equalization multiplies each pixel's grey
All of our experimental dive work was level from the original image by an
conducted using the Harbor Branch algorithm where each pixel's range in the
Oceanographic Institution's DSR/ Johnson image is stretched over as large a segment
Sea Link I and 11 (1,2). The total VSPC of the 256 levels as possible, and the
system is illustrated in Figure 1 (1,2). dynamic range is smoothed (IMAGEPROTM,
Deployment of the VSPC is similar to the Media Cybernetics). Fluidized sediment
sequences involved in taking a sub- fill, infaunal burrow structures, and the
operated box core, except that the prism average mixed depth (as confirmed in X-rays
must go in (and remain precisely) at right and cores) are now visible.
angles to the plane of the SWI (1,2). The next sequence of VSPC images
video images were recorded in real documents a burbot trench at 300m in Lake
time in the front sphere of the sub. The Superior(Figure 3a-g). This sequence
tapes (1985-86) were then played on a illustrates the trench edge morphology from
broadcast quality VHS system, filtered, the surface down to the internal burrow
and the raster images were photographed fill structure at the base of the trench.
with 35mm Ektachrome, and finally Once again, the original image of the
printed.These photo images were digitized internal structure of the trench bottom
by hand at 1/411 by 1/411 grid scales, and sediments is examined (Figure 3a-d). In a)
reproduced at scale usinga Zenith 158 PC The sides of the trench and the fluid
and a Prodesign II CAD program (2). Video reworked fill material are visible looking
images from 1987 and 1988 dives were toward the shallow end. The upper portion
replayed directly into the TRUEVISIONTM of the burbot trench is visible in b),
Targa M8 Image Capture Board. The video where the left side of the photo shows the
was captured frame-by-frame at 1/30 second sharp edge of the excavation--the VSPC has
and analyzed with IMAGEpROTM scientific not penetrated the bottom of the trench
imaging software, specifically for grey even at 18-20 cm deeper than the SWI. The
scale range or overall image contrast, and lower portion of the trench excavation is
dynamic range of the available grey visible in c). The left side shows the
levels. Then, an appropriate histogram further penetration of the VSPC into the
equalization algorithm (IMAGEPROTM Manual, excavated trench, with the bottom sediments
version 1.06.05, Media Cybernetics) was just visible at the lower portion of the
used to change the spread of the image image, and in d), the image of the internal
contrast. These images can then be edited, burrow/trench sediments displays the 'IV"
with text and graphicsMplaced directly on shaped cut of the trench excavation filled
them, with HALOVISIONT software (Media with sediments. This image was then
Cybernetics). analyzed for grey level contrast and
dynamic range (Figure 3e). Note the
RESULTS extremely narrow grey level contrast range,
yet high dynamic range. Then the image was
Figures 2a-d, and 3a-g show two enhanced (Figure 3f) and mapped onto the
examples of the utility of the VSPC system original image (Figure 3g). The enhanced
coupled with image enhancement and image shows fluidized burrow sediment fill,
manipulation techniques. Figure 2a-d reworked and mixed/transported infilling
documents the VSPC imagery through a red sediment, the sharp boundary between
crab burrow and into the internal sediment and the overlying water, and
sediments. Figure 3a-g illustrates a between the infilling sediments and the
sequence of VSPC images through the burbot original excavation in the highly cohesive
trench system down into the underlying post-glacial clays.
post-glacial clays.
Figure 2a illustrates a red crab
burrow opening. The highly textured surface
sediments sculpted by the crab chelae and
legs are visible around the undercut burrow
excavation. Interior burro fill is just
visible below. Figure 2b i( the original
VSPC image from SWI, down 15-20cm deep

449

SUMMARY REFERENCES

The utility of the VSPC and image (1) Boyer, L.F. this issue. Video-Sediment-
analysis techniques opens new avenues for Profile Camera imagery in marine and
studying the structures, organisms, and freshwater benthic environments. OCEANS 88.
processes at and beneath the sediment-water
interface in marine and freshwater (2) Boyer, L.F., and Hedrick, J.A. 1989.
profundal environments. Submersible-deployed video-sediment-
profile camera system for benthic studies
(Special Lake Superior submersible
ACKNOWLEDGEMENTS
symposium: Jour. Great Lakes Research).'
Ocean Instruments provided the VSPC (3) Boyer, L.F., and McCall, P.L. 1988.
prototypes, NOAA's NURP-UCAP the ship and Video-Sediment-Profile Camera (VSPC)
sub crews of the R/V Seward Johnson and perspectives of Red Crab burrow systems.
DSRI Is Johnson Sea Link I and _I1, the East Coast Benthic Ecology Meetings,
University of Wisconsin provided Abstracts with Programs, April 8-11,
photographic and cartographic services. Portland, ME.
This work was funded in part 1py the
University of Wisconsin Sea Grant Institute (4) Boyer, L.F., Grassle, J.F.,
under grants from the National Sea Grant Whitlatch, R.B., and Zajac, R.N. 1988. Red
College Program, NOAA, U.S. Department of crab bioturbation on the Southern New
Commerce, and from the State of Wisconsin. England Slope: The L.A.S.S.E. project. East
Federal Grant NABYAA-D-0065, Project R/MW- Coast Benthic Ecology Meetings, Abstracts
38-PD. NURP-UCAP PSA grant 1987,188. This with Programs, April 8-11, Portland, ME.
article is contribution # 326 from the
Center for Great Lakes Studies, University (5) Whitlatch, R.B., Zajac, R.N., Boyer,
of Wisconsin-Milwaukee. L.F., and Grassle, J.F. 1968. The effects
of the Deep sea red crab (Geryon)
Disturbance on the southern New England
upper continental slope. East Coast Benthic
Ecology Meetings, Abstracts with Programs,
April 8-11, Portland, ME.

(6) Boyer, L.F., and Whitlatch, R.B.
1989. Organism-sediment relationships in
the Caribou Island Basin, Lake Superior,
U.S.A. (Lake Superior submersible
symposium: Jour. Great Lakes Research).

(7) Boyer, L.F., Cooper, R.A., Long,
_0
D.T., and Askew, T. 1989. Bioerosion and
- Ww
ZA4 shelter formation by the burbot Lota lota
in the deep basins of Lake Superior. (Lake
Superior submersible symposium: Jour. Great
Lakes Research).

Figure 1. The VSPC system. The system
includes the prism with an aluminum "skin".
handles, radon lights, a CCD camera,
waterproof housing and cable terminations,
cable system, control panel, a 411 black and
white SONY Watchman, a broadcast quality
JVC mini-VHS VCR with audio dubbing, remote
control, and 20 minute cassettes.

450

P A

A
2,
t

@@p,A
M"

VIZI "I
FA 111 - 0
YZZ
AN

tZ
f
'r e" 01 J @i eo 1
fig, 0 f 4,2,-e
la
7

I a

A, 4"1
@c)
@; @r @ @11 d)

Figure 2a-d. VSPC Sequence through a red crab burrow interior captured frame-by-
frame on a TRUEVISIONTM Targa M8 board. Screen resolution 512 x 480, 35mm
ektachrome/interneg/to B&W print. a) Red crab burrow opening. Average diameters
typically 10-15cm. The highly textured surface sediments are visible around the
undercut burrow excavation. Interior burrow fill just visible below. b) Original VSPC
image from SWI down 15-20cm deep inside the burrow. The burrow and internal
sedimentary features are extremely difficult to discern due to low contrast and
dynamic range in the ambient sediments. c) Histogram of grey scale dynamic range and
spread in the burrow sediments along a profile transect. Note the very narrow peak,
yet relatively high values per index. d) Histogram equalization of the original image.
(IMAGEPROTM, Media Cybernetics). Fluidized sediment fill, infaunal burrow structures,
and the average mixed depth (as confirmed in X-rays and cores) are now visible.

451

71@ 17
A
POW
-F 4
%
"EM
AVOW
ry

r '0,
frWok"v
0 e"A 9 1 @k,
R-

'04

e@
W-

I.- -W-2
4;6@r@x wzza
4

Figure 3a-d. Sequence of VSPC images in a burbot trench. This sequence illustrates the
.trench edge morphology from the surface down to the internal burrow fill structure at
the base of the trench. a)View of the trench at the SWI through the "eyes" of the
VSPC. The sides of the trench and the material are visible. b) The upper portion of
the burbot trench. The left side of the photo shows the sharp edge of the excavation,
the VSPC has not penetrated the bottom of the trench at 18-20 cm. c) The lower portion
of the trench excavation. The left side shows the further penetration of the VSPC into
the excavated trench, with the bottom sediments just visible at the lower portion of
the image. d) The VSPC image of the internal burrow/trench sediments with the 'IV',
shaped cut of the trench excavation visible filled with sediments.

452

Burbot Trench systems
Deep interior structure

water J

fluidiied
muds

fill
e=&41 Una
efte

post-glacial clays
g)

Figure 3e-g. Sequence of VSPC images in a burbot trench. This sequence illustrates the
trench edge morphology from the surface down to the internal burrow fill structure at
the base of the trench. e) The grey level contrast and dynamic range of the internal
sediments of the burbot trench. Note the extremely narrow grey level contrast range,
yet high dynamic range. f) The enhanced image of the internal structure of the burbot
trench. g) The enhanced image with fluidized burrow sediment fill, reworked and
mixed/transported infilling sediment, and the sharp boundary between sediment and the
overlying water, and between the infilling sediments and the original excavation in
the highly cohesive post-glacial clays.

453

APPLICATION OF THE POLARIMETRIC MATCHED IMAGE FILTER (PMIF) TECHNIQUE TO
CLUTTER REMOVAL IN POL-SAR IMAGES OF THE OCEAN ENVIRONMENT

Wolfgang-M. Boerner, Alexander B. Kostinski, Brian D. James, and Matthias Walther

Communications Laboratory (m/c 154), Dept. of Electrical Eng. &
Computer Sci., University of Illinois at Chicago, SEL-4210,
840 W. Taylor St., Chicago, IL/USA 60680-4348

ABSTRACT recently. Such data, taken with the NASA/JPL
CV-990 dual-polarization L-band (1.225 GHZ)
SAR (Synthetic Aperture Radar) system, have
We focus on image contrast optimization been made available to us. Here, we
between two rough surface classes, which is investigate the potential of a strictly
based strictly on polarimetric filtering and, polarimetric image filteri ,ng which takes full
therefore, no digital image processing advantage of the matrix data provided on a
techniques are employed. The approach is pixel by pixel basis, and complements the
tested on a complete polarimetric synthetic existing scalar contrast optimization and
aperture radar image of the San Francisco Bay speckle reduction techniques. We wish to
area (NASA/JPL CV-990 L-band POL-SAR data). stress from the outset that our goal is
Optimal transmitted polarizations are found contrast optimization (with the corresponding
for each image pixel and the results are speckle reduction) without the help of
analyzed statistically via a set of joint 2D incoherent averaging over pixels or "looks",
histograms. This is done for both of the because of the corresponding loss of spatial
rough surface classes. The image response to or temporal resolution. At first glance,
the "optimal" incident polarization is then speckle reduction is impossible without
simulated digitally by adjusting the receiver incoherent averaging but further consideration
polarization according to the modes of the shows that it is so only for scalar data.
histograms. The corresponding images are Indeed, taking "projections" onto the receiver
computed and displayed with s Iignificant image direction in the polarization space decreases
contrast improvement. amplitude fluctuations and an image appears
less speckled. The goal of this paper is to
find such a choice of the polarization
projection which makes a given rough surface
@lass least speckled and, by doing so, to
improve the image contrast between two given
1. INTRODUCTION classes.
This. paper addresses the problem of coherent The paper is structured as follows: a brief
image contrast optimization between two rough @escription of basic polarimetric definitions
surface classes. We focus on such contrast is provided in Section II, while in Section
which is due to the differences in III the image data are described and the
polarimetric scattering from one rough surface precise problem formulation is given. In
to another. The novelty of this problem is Section IV, our three-stage polarimetric
due to the combination of coherent imaging and optimization procedure is outlined and
polarimetric scattering. The former implemented for each image pixel. In Section
introduces speckle reduction as a major issue V, we make the transition from the single
while the latter provides the full scattering pixel result to a combined description. A
matrix (i.e., complete polarization statistical analysis of the results is given,
information) per image pixel. The second and and the images are displayed and discussed.
equally important task of this work is to The operation of the polarimetric matched
develop efficient statistical tools for filter is summarized in Section VI. Section
polarimetric image data analysis and speckle VII contains concluding remarks.
reduction techniques.
II. ESSENTIAL POLARIMETRIC DEFINITIONS
Speckle has long been recognized as the main
problem of coherent imaging (l)- and many Following (2), we define the origin of the
processing techniques have been advanced to coordinate system at the receiving antenna
overcome it. The vast majority of these terminals with the +2-axis directed toward the
techniques, however, are of a.scalar nature target (pixel) as shown in Fig. 1. Note that
simply because vector/matrix imaging data are in SAR applications the receiver and
so sparse and have become available only very transmitter are co-located but may be

CH2585-8/88/0000- 454 $1 @1988 IEEE

different antennas so that the situation is new basis. With these definitions we now
slightly bistatic. The reflected �R and the proceed to describe the polarimetric SAR image
transmitted E waves, together with the data and to formulate the problem more
T precisely.
antenna "height" h (polarization of the
receiving antenna @5hen used as transmitter III. IMAGE DATA DESCRIPTION AND PROBLEM
(2)), can all be written as plane waves FORMULATION
[lal 9T (E 2+ E2 )@[cos ' + sinYT ej@T y^1 The 4096 x 1024 SAR image of the San Francisco
TI T2 YTX Bay area (6,7) is shown in Fig. 2a for
exp Jj(wt - kz + aT)J horizontal transmitter and receiver
polarizations (HH). The brightness of each
pixel is assigned according to the total
[lb) (E2 + 4 )@[cos + siny ej@R Y] received energy in the horizontal channel.
R YRx R The data was taken with a dual-polarized
x y . exp fj(wt + kz + aR)l antenna and a four-channel receiver system so
that a complete scattering matrix was measured
for each image pixel. The radar wavelength
(1c] W + h2) [cosyhx + sinyh ej@h, Y@] was 24.5 cm and the resolution (size of each
X y pixel size) was about 10m x 10m (see (6) for
exp, (j(wt - kz + ah)) more details). The image texture can be
roughly classified into three main categories:
where in all equations y =_ tan-1 (EY/Ex), and man-made structures (ships, the bridge, urban
area, etc.), vegetated area (park), and the
and m are the relative and absolute phases, ocean region. All three classes can be
respectively (2). From here on, we will considered rough at 24.5 cm according to the
operate with the expressions in square Rayleigh criterion (1). This surface
brackets, written as complex normalized 2D roughness leads to a random modulation of
vectors (also known as Jones vectors (3) in phase of the reflected wave which, in turn,
optics and spinors in quantum mechanics (.4)), produces image speckle (1).
e.g.,
Any kind of averaging is likely to smoothen
COSYT the image and reduce speckle. As an example
of averaging in polarization space, the span
ET sinyT ej@ of [S] image (sum of the magnitudes of all
TI EV IT- four scattering matrix elements) is shown on
when usual assumptions about a linear passive Fig. 2b. The image is essentially an
medium are employed, the input-output incoherent superposition of the four separate
polarization ellipse characteristics of a polarization images obtained per pixel and,
target (image pixel) are given by its therefore, a noticable speckle reduction
scattering.matrix defined as (relative to the HH image) is not surprising,
(8). As was mentioned above, this paper
(21 ER = [SIET focuses on contrast optimization without
incoherent averaging of any kind.
where [S] is a 2 x 2 complex matrix, and o;r is Since a complete scattering matrix is
set to zero by the choice of the time origin available for every pixel of each of the three
Finally, the voltage at the receiving categories, one can simulate the response of
antenna terminals as a function of transmitter the area to any transmitted polarization E T by
and receiver polarizations is given by calculating E R via [2). Furthermore, the
(31 V = t T tT[S]FT response of the image can also be simulated
for an arbitrary receiver polarization h via
[3]. Both equations must be implemente-a for
where superscript T denotes the transpose (as each pixel of the entire image. The
brightness is then assigned to each pixel
opposed to hermitian conjugate - see (2, pp. T * T
1471-1473), and (5) for details). For according to P =- V V = (h ER) (h E stands
reference,.we also include the transformation R
properties of [S] and V from one polarization for complex conjugate). Such numerical
basis to another using . a similarity simulations were recently carried out by the
transformation for [S] to [S] as discussed in JPL group (6,7), demonstrating the ability of
(2, pp. 1471-1473), polarimetric adjustment to substantially
improve image contrast. our goal here is to
[4a] [S], = [U1_l ISHUI develop an algorithm for the search of optimal
image contrast via the combination of our
recently developed Three-Stage-Procedure (TSP)
[4b] V = V - h' T(U] T[U]E.1 TE which is described in the following section,
_R and a subsequent statistical analysis of the
set of polarization eigenvectors computed with
where [U] is the unitary change-of-basis the TSP for each pixel. Again, we emphasize
matrix and primes indicate quantities in the that our algorithm must not include any

incoherent..... averaging and/or smoothing [7b] V = h TE
procedures because of the corresponding loss _R 0-
of information, e.g., temporal (phase) or
spatial resolution. In this paper we focus on In terms of imaging applications, one expects
finding such transmitter and receiver a given pixel to look relatively "bright" when
polarizations that allow significant ocean corresponds to the largest eigenvalue
clutter removal for better contrast with the h
urban area and ship/man-made structure (maximal energy density) and h is adjusted
identification. The method and implementation according to [7a], while the aajustment [7b]
of the optimal polarization search for a ensures that the pixel looks "dark",
single pixel are briefly described in the next especially when supplemented with the choice
section, after which we proceed to the of minimal E T,opt. These observations,
statistical analysis of the results. together with the statistical considerations,
constitute the basis of pixel-by-pixel
IV. THE THREE-STAGE PROCEDURE polarimetric image filtering.
The TSP addresses the following problem (see V. STATISTICAL ANALYSIS OF OPTIMAL
(2,5) for more details): For a given pixel POLARIZATIONS AND IMAGING
(i.e., known scattering matrix), find such
transmitting and receiving polarizations, for Even within a single rough surface class
which the received power is maximal (minimal). (e.g., ocean), there is a considerable
In mathematical terms this means: find E T and variability in polarization properties of
h such that P A = ItT [S]E 12 is optimal pixels in any given patch and we, therefore,
T must introduce a statistical description at
for a given [S], subject to the constraints this stage. we assume that the two terrain
JJhJJ = JJfTJJ classes are sufficiently different in their
polarimetric responses so that their
The TSP accomplishes this in three separate statistics do not "overlap" significantly.
stages (2): Then, with proper statistical tools, a
"threshold" can be found such that the TSP can
Stage 1) The energy density in the reflected be used to "darken" not just one pixel but a
wave (before it has reached the receiver) is majority of pixels in a given class.
optimized as a function of transmitted
polarizations via the following eigenvalue In order to gain insight into the polarimetric
problem response of various terrain and ocean
categories, we have performed Stage 1 of the
[5) J[G] - X[IIJE T,opt = 0 TSP for each pixel of two chosen segments of
ocean and urban areas. Let us consider the
+ ocean vs. city contrast enhancement as a
where [G] a (S] (S] is by construction a specific application. In order to minimize
hermitian matrix for any (S] (+ stands for the ocean return or to maximize the city
hermitian conjugate), [I] is the identity return, the minimum energy eigenvector is
matrix, and �T,opt is the eigenvector computed for each pixel of the ocean patch and
corresponding to the largest (smallest) the maximum energy eigenvector is computed for
eigenvalue giving the largest @he city patch. The eigenvectors correspond-
Xmax (\min) ing to , x are computed according to
(smallest) energy density. The eigenvalues 'tin max
are always real (they correspond to the [5) for each pixel and expressed in the form
measured values of energy density in the (2, p. 1400)
reflected wave) and the eigenvectors are
orthogonal because [G] is hermitian (9). [8a) ET,Opt p - EY/EX
Stage 2) At this stage, the polarization state fil -+Po*)
of the reflected wave is computed using the where p is the complex polarization ratio (3).
known ET,opt After the eigenvectors are computed, we
(61 ER [S]E express them in terms of the more convenient
,opt T,opt* ellipticity c and tilt T coordinates which
describe, respectively, the "fatness" and the
Stage 3) Finally, the receiver polarization is inclination of the polarization ellipse. They
adjusted to ensure that all of the power are defined (3, p. 35) as
contained in E R,opt (reflected wave) is either [8b] s = 1/2 arcsinJ21m(p)/(l-pp*))
absorbed or rejected, depending on the
application. The former is accomplished with [8c] T = 1/2 arctan[2Re(p)/(l-pp*)J.
the choice
Thus, an optimal polarization state which
Pal h - E makes a given pixel darkest (brightest) is
R characterized by two numbers, e and T.
Naturally, one would like to choose the
while the latter requires that incident polarization in such a way as to make
most ocean pixels dark if our goal is to

456

contrast urban area against ocean or to image). The ocean area is quite a bit more
enhance visibility of ships at sea. To this speckled than on Fig..5a and the contrast with
end, we present in Figs. 3a-b joint 2D the urban area is lower. on the other hand,
histograms (c and T) for the two surface there is a better contrast between
categories of interest: statistics of minimal park/vegetated area and the urban region.
eigenvectors are presented for the ocean, and Thus, the results of the JPL group (6) as well
maximal eigenvectors for the urban area. as our experiments clearly show that a much
Each patch contains 40,000 (200 X 200) pixels improved contrast can be achieved between
so that the statistics are quite good. Both man-made, vegetated and ocean areas with the
histogram modes are near the linear vertical proper choice of polarization. Of course,
(c = 01, T = 900) polarization. Thus, if the when the rough surface is such that the
transmitter is adjusted to produce vertically scattering is polarimetrically isotropic
polarized waves (relative to the direction of (i.e., there is no spatial polarization
propagation), the majority of the ocean pixels dependence), this technique cannot work (one
will have relatively low scattered energy, such example is a random sea surface).
while the majority of city pixels will reflect Fortunately, such cases are rather unlikely
strongly. Once the optimal transmitted field and all of the data available to us indicate
is chosen and the scattered field is computed, that real terrestrial rough surfaces exibit a
one can use a similar procedure for the very strong polarization dependence. Even an
receiver adjustment. In Figs. 4a-b, we ocean surface is often modulated by
present the s-T histograms for the scattered well-defined internal wave patterns which show
fields of the two regions. These histograms up clearly in POL-SAR images.
were constructed by letting the incident wave
be vertically polarized and by computing the A sequence -of one-dimensional image brightness
scattered field of each pixel via Eq. 6. distributions in Fig.6 illustrates the effect
Again, the two histograms peak around the same of various stens of the above Drocedure on the
vertical polarization state and the ocean ocean and city patches, separately. One
distribution is more pronounced. In fact, the notices a gradual improvement in contrast
ocean incident and scattered field histograms between the two categories as indicated by the
are quite similar which leads one to conclude decreasing overlap area and better peak
that most of the scattering matrices of the separation. This suggests that the
ocean region are "flat plate-like" identity h-adjustment is responsible for most of the
matrices. This behaviour is consistent with clutter removal, as can also be seen on the
Bragg scattering assumed to be the dominant actual images of Figs. 5a and 5b. Note that
physical mechanism of the ocean scattering identical uniform grey scale assignments have
(6). The urban area histograms, on the other been used for all images so that the effects
hand, differ because the scattered field does are entirely polarimetric.
not have a peak at horizontal polarization.
The fact that this peak vanishes seems to VI. SUMMARY OF THE POLARIMETRIC MATCHED
disagree with the assumption of dihedral FILTER STRATEGY
corner reflectors (6) as the basic scattering
elements of the urban area., Indeed, in such a In this section, we summarize and quantify the
case, the scattering matrix (having entries +1 approach outlined in the previous paper in a
along the diagonal and 0 along the series of well-defined steps. Again, consider
off-diagonal) would produce a mild peak at the suppression of ocean clutter for optimal
horizontal polarizations which would not contrast with man-made structures such as
disappear. ships, etc. we perform the first two steps of
the TSP, and display the "typical" statistics
If the receiver is adjusted to horizontal of the ocean and urban area patches in a form
polarization, most of the energy of ocean of joint bivariate histograms of transmitted
pixels will be rejected because the receiver fT and received E R fields as is shown in Figs.
is perpendicular to the sharp histogram mode 3a,b and 4a,b. we then identify the modes of
at vertical polarization. The urban area the two distributions E and E and adjust h
will not be affected as much because of the -T R
much larger spread. when the brightness is so that the "majority" (i.e., histogram peak)
of the "unwanted" patch pixels "darken". For
assigned according to P = V V (V is computed instance, if h is adjusted in such a way that
from Eq. 3), the image in Fig. 5a results. the peak i-n the ocean distribution E
compared with the original HH image, this near R
HV image has better contrast: the average satisfies
brightness ratio between the urban and the 191 V T
ocean areas increases; but, because of the peak -ER,peak @ 01
fact that the two modes are not separated in
the polarization space, the urban area has
lost some structure. Note, however, that it is ensured that the majority of the ocean
most of the ocean speckle has been "filtered pixels will appear "black" on the actual
out" with the proper choice of polarization, image.
yet, without significant effect on the
man-made structures. The following procedure (see Fig. 7), which is
a statistical extension of the TSP,
The image of Fig. 5b was computed for the constitutes the polarimetric matched filter
vertical polarization of the receiver (near VV for coherent imaging:

457

1a) the energy density of each pixel is conjunction with the polarimetric enhancement,
maximized (minimized) as a function of should be directed towards the increase in
the transmitted , polarization. The peak sharpness of the re@evant field
corresponding eigenvectors, !@T , are found distributions (e.g., N X N block averaging,
from Eq. 5 as in Stage 1 of TSP; discretization, and quantization, etc.).
Here, however, we concentrate strictly on
1b) the joint bivariate histograms of ET (e polarimetric enhancement methods.
Furthermore, the polarimetric image contrast
and T) are constructed for all rough improves with the separation of the histograms
surface classes of interest; in the polarization space (two "spikes" with
no overlap would correspond to a "black and
lc) the transmitted field E T is adjusted to white" image with an ideal contrast). In this
either the peak of the minimal eigenvector respect, the TSP was not successfull because
pdf of the unwanted region (e.g., to the scattered field histograms of ocean and
reject ocean clutter) or to the peak of the urban area (Figs. 4a and 4b) are
the maximal eigenvector pdf of the region approximately at the same c and T. Another
of interest (e.g., bridge, urban area, approach would be to choose the transmitting
etc.). The choice depends on the relative field in such a way as to maximize the peak
sharpness of the modes; separation of the scattered field histograms.
This approach is currently under investigation
2a) the scattered field is computed for in our laboratory and preliminary results
R based on Monte-Carlo simulation of structures
each pixel for the E T chosen in Step 1c, in Rayleigh noise indicate better contrast
see Eq. 6 in Stage 2 of TSP; (relative to TSP) I but less efficient speckle
reduction.
2b) as in Step lb, the joint (c and T) We wish to state here that an immediate
histograms of the scattered field E Rare objective of the research is to establish a
constructed. The histogram mode is "tool-kit" of matrix image processing
identified. techniques,, designed specifically for the
handling of polarimetric scattering matrix
3a) the receiver polarization h is adjusted data on a pixel-by-pixel basis. Consequently,
via Eqns. 7a or 7b to either match or we did not emphasize either the modeling or an
mismatch the polarization of the histogram interpretation of polarimetric scattering
mode found in Step 2b; patterns beyond some very basic physical
arguments based on flat plates, corner
3b) P = V V (received power) is computed for reflectors, Bragg scattering, etc.
each pixel as P (JE (hTE )
T[S] * T - -R --R ACKNOWLEDGEMENT
(h ET) (h [S]E T and the resulting This work was supported in part by the US Army
image is displayed. Research office under Contract DAAK
VII. CONCLUDING REMARKS 21-84-C-0101, the US Office of Naval Research
under Contract US ONR N00014-80-C-0773 and
The potential of complete polarimetric methods N00019-85-K-0483, by the US Naval Air
for radar imaging has already been Development Center under Contract,
convincingly demonstrated by the JPL group N62269-85-0383, and by the US Army Fort
(6.7). In this paper, we have attempted to Belvoir Research, Development and Engineering
quantify and organize a search for optimal Center under Contract DAAR-70-87-P-2565.
image contrast into a systematic polarimetric Computations were carried out on the
filtering method. In addition, no incoherent UIC-EECS/CL VAX-11/750 research computer
pixel/look or spatial averaging was allowed. facilities sponsored by equipment grants ONR
We have accomplished this by combining the TSP N00014-80-C-0073E, ARO-DAAG 29-85-002, and by
search for optimal polarizations on a the University of Illinois at Chicago Research
pixel-by-pixel basis with a subsequent Board and Engineering College.
statistical analysis of polarization We wish to thank and acknowledge the interest
eigenvectors (versus surface category), and shown in our polarimetric research by Henry W.
the digital adjustment of the polarimetric Mullaney, Hans Dolezalek, Robert J. Dinger,
variables E T and h. We find the preliminary Wolfgang Keydel, Otto Kessler, James W. Mink,
results (Figs. 3-7) promising. Walter Flood, and Lloyd W. Root.

The effectiveness of our strategy depends We also wish to thank Drs. Charles Elachi,
sensitively on the sharpness of the relevant Diane Evans, Jakob van Zyl of the CAL-TEC Jet
histogram peaks because such a sharpness Propulsion Laboratory, Pasadena, Ca., who made
reflects similarity of the polarimetric available the JPL/CV-990 L-band poL-SAR data
scattering behaviour of all the pixels within sets of the San Francisco Bay area used in
a given class. Therefore, other image this paper.
processing techniques, when used in

458

REFERENCES

(1) J.C. Dainty, (ed.), Laser Speckle and
Related Phenomena, Topics in Applied
Physics, Vol.9, Springer Verlag,
Heidelberg, 1975.

(2) A.B. Kostinski and W-M. Boerner, "On
foundations of radar polarimetry," IEEE
Trans. Antennas Propag. AP-34(12),
1395-1404 (1986); H. Mieras, "Comment,"
IEEE Trans. Antennas Propag. AP-34(12),
1470-1471 (1986); A.B. Kostinski and
W-M. Boerner, "Reply," IEEE Trans.
Antennas Propag. AP-34, (12) 1471-1473,
Dec. 1986

(3) R.M.A. Azzam. and N.M. Bashara,
Ellipsometry and Polarized Light,
North-Holland, Amsterdam, 1977.

(4) E. Merzbacher, Quantum Mechanics, John
Wiley, 1961, Ch. 13. (a)
(5) A.B. Kostinski and W-M. Boerner, "on the
polarimetric contrast optimization,"
IEEE Trans. Antennas Propag. AP-35, No.
8, 988-991, Aug. 1987.

(6) H.A. Zebker, J.J van Zyl, and D.N.
Held, "Imaging r@dar polarimetry -from
wave synthesis," J. Geophys. Res. 9,(Bl),
-7
683-701, 1987.

(7) N. Donovan, D. Evans, and D. Held (eds),
NASA/JPL Aircraft SAR Workshop Proc.
(JPL, Pasadena, CA, Feb. 4-5,1985),
iPL-publ. 85-39, Jan. 15, 1985.
-.7
(8) W-M. Boerner, Polarization microwave
Holography: An Extension of Scalar to
Vector Holography (invited), 1980
International Optics Computing
Conference, SPIE's Techn. Symp. East,
Washington, DC, April 9, 1980, Session
3B, Paper no. 231-23, Proceedings, pp.
188-198, 1980.
(9) G. Strang, Linear Algebra and Its (b)
Applications, Academic Press, New York,
1976. Academic Press, New York, 1982.
Fig. 2. Synthesized Images: (a) HH Element of
(S] and (b) Span of (S].
Speckle in these.coherent images is due to
random phase modulation associated with
surface roughness. The span image (b) has
less speckle than the HH image (a) because
span [S] = IS HH12 + Is HVI2 + ISVHI 2 + IS,12
is an incoherent average of four images.
Sixteen uniformly spaced gray scale levels
have been used to cover the voltage values in
-3.5 -1.5
the range [10 10 1 on a logarithmic
scale.

Fig. 1. Arrangement of Coordinate System for
Polarimetric SAR.

459

17
J

'1@71

(b)

Fig. 3. Histogram of optimal Transmitted Polarizations: (a) ocean Region and (b) Urban Region.
Optimal eigenvectots were computed for each pixel of the 200 x 200 ocean and urban regions (see
Step 1 of the PMF). These eigenvectors were histogrammed in ellipticity c and tilt T
coordinates (see Eq. [8]). Minimum energy eigenvectors were found for ocean pixels and maximum
for urban pixels. The mode (peak of histogram) location indicates that at E T Of c = 0* and -r
900 (vertically polarized), a majority of ocean pixels will respond weakly. Fortunately, the
modes of (a) and (b) are the same and, therefore, the majority of city pixels will respond
strongly to the same polarization.

(a) (b)

Fig. 4. Histogram of Scattered Polarizations: (a) Ocean Region and (b) urban Region.
The scattered field E R was computed for each pixel of the 200 x 200 ocean and urban regions;
the transmitted polarization was chosen in accord with Fig. 3. These scattered Polarizations
were histogrammed in ellipticity c and tilt T coordinates (see Step 2 of the PMF). The ocean
mode at e = 00 and T = 900 indicates that a majority of ocean pixels are mismatched by
adjusting h to e = 00 and t = 00. Because the two modes coincide, an appreciable portion of
the urban region will also be mismatched. However, the large spread of the urban histogram
still leads to significant contrast improvement.

460

'- flo

(a) (b)

Fig. 5. Synthesized Images: (a) ocean Polarization mismatched and (b) ocean Polarization
Matched.
In (a), the ocean was mismatched by adjusting ET to the peak of Fig. 3a (C 01, T = 90*) and
by adjusting h orthogonal to the peak of Fig. 4a (c =0*, T = 00). In (b), the ocean was
matched by adjusting ET as before, while adjusting h to the peak of Fig. 4a (c = 0", T = 901).
Note that the ocean vs. urban contrast is much higher in (a) than in (b). Sixteen uniformly
spaced gray scale levels have been used to cover the voltage values in the range
[10-3-5,10-1.5 ] on a logarithmic scale.

I SELECT REGION I
I (OCEANIURBAN) I
CITY
01
FC-ONSTRUCT POWER MATRIX I

VOLTA IGE BRIG4HTNESSILOGI VOLTAIGE 51IG4HTNES5 -2 (LOG)
(b) I OPT-IMIZE ENERGY DEN,1ITY
(MAX OCEANIMIN V W AN

."". - T,-.
I HIS IDENTIFY MODE

I COMPUTE-SCATYERED FIELD

r-HISTOGR-AM & IDENTIFY"
.6 .4 .1 -6 4 -4 .3
VOLTAGE BRIGHTNESS (LOG) VOLTAGE $RIGHTNESS (LOG)
W (d) ADJUST RECEIVER
[email protected] OCL NIMAICH URBAN)
Fig. 6. Effect of PMF on Image Brightness
Distributions: (a) HH Reference, (b) optimal
Reflected Energy (see Step 1 of the PMF), (c) I-fFFR- 0 D U C EIMAGE
ocean Matched, and (d) ocean Mismatched (see
Step 3 of the PMF).
All voltage values are normalized by the Fig. 7. PMF Flow Chart with Applications to
transmitted energy and, therefore, their Ocean vs. Urban Contrast Enhancement.
logarithms are negative. The decrease in
relative overlap area between (a) and (d)
indicates contrast enhancement. Also, note
that the decrease in variance of the ocean
distribution between (a) and (d) signifies
speckle reduction.
C
L -1

461

REMOTE SENSING OF PHYSICAL AND BIOLOGICAL PROPERTIES OF ESTUARIES

Charles Bostater, Vic Klemas

University of Delaware, College of Marine Studies
Newark, Delaware 19716

ABSTRACT

Estuarine remote sensing instruments In general, we can safely say that
are described with respect to state interestuarine comparisons have and will
of the art capabilities. Remote probably continue to rely upon spatial
sensing studies of estuaries can be and temporal resolution less than that
considered as being of two different required for intraestuarine comparisons.
types: estuarine intercomparisons For, example,, current inter estuary
(between estuaries) and intra- comparisons have relied upon AVHRR
estuarine (within estuary assess- satellite data with a spatial resolution
ments). Current research has of greater, than a kilometer. The
allowed initial efforts of estuarine frequency of satellite overpasses used in
intercomparisons of light atten- such analyses are based upon daily to
uation and suspended sediments. monthly temporal frequency. For example,
Remote sensing can provide inputs to figure 2 shows the results of comparing
math- ematical models used for 14 different coastal estuarine and near
research and management of estuaries coastal water areas after 17 images were
and near coastal waters. The in atmospherically corrected using NOAA,
situ optical profiling approach is NESDIS atmospheric corrections according
transferable to other estuarine to Stumpf (1,2). Approximately 85% of
studies. the varibility of light attenuation
estimated from these images was
statistically associated with physical
characterisitcs of the estuaries (i.e.
water depth, freshwater velocity,
average wind speed, amplitude of the
tidal current, a scale for the tidally
rectified residual circulation, and
1. INTRODUCTION bottom sedimnet type).

Remote sensing of estuaries can be
conceptualized as being applied to two
types of analysis questions, as Intraestuarine comparisons rely
indicated in figure 1. These are upon, and generally require, remote
interestuarine (between estuaries) and sensing data with higher spatial,
intraestuarine (within an estuary) temporal, radiometric and spectal
comparisons. At a national level, resolutions in order to resolve dynamic
government agencies are interested in processes occurring on smaller space and
comparisons between estuaries in order to time �cales. Low flying aircraft sensors
help provide information for making have an important role in both types of
regional or national policy decisions. analysis efforts.
The information requirements for these
types of comparisons is unique from
assessments concerning a particular 2. REMOTE SENSING INSTRUMENTATION
estuary, subestuary drainage area, water
segment, or water mass boundary such as a Existing satellite sensor systems aboard
front between water masses with different SPOT and Landsat mss (multispectral
optical signatures. scanner) or TH (thematic mapper) systems
can provide the spatial resolutions for

CH2585-8/88/0000- 462 $1 @1988 IEEE

Estuarine Intracomparisons
I Liuing Resources/
Land Estuarine Tributary Estuary Water P Biological Production
Use Habitats Outputs Plumes Quality Assessments
Estuarine
Intercomparisons

Figure 1. Conceptualization of estuarine and interestuarine
comparisons.

estuarine intracomparisons, but they lack At the same time, data from remote
the temporal resolution needed for sensing instruments is becoming a
studies that rely upon a temporal valuable asset in filling temporal and
frequency of less than 2 weeks. Even spatial data gaps used in the application
this estimate is optimal since cloud of research and natural resources
cover typical prevents use these management models. Figure 3 indicates
satellite images more than 2-3 times a the different types of mathematical
season. The AVHRR sensor onboard the models and types of inputs that remote
Tiros N series of NOAA satellites sensing instrumentation can currently
provides the one or two times a day provide for estuarine and near coastal
coverage needed for studying the dynamics ocean assessmnets. Past research has
of estuaries. Except for extremely large documented the capability of providing
estuaries and near coastal waters, this land use from remote sensing satellite
instrument lacks the footprint or platforms. The technology exists for
spatial resolution as well as the determination of salinity using microwave
spectral resolution necessary for radiometry as described by Swift (3).
detailed estuarine intracomparisons. At this time, sea surface water
This instrument lacks the spectral temperature is one of the few remote
resolution (only 2 visible channels) in sensing variables routinely processed
the visible portion of the from AVHHR imagery. Recent research by
electromagnetic spectrum that would allow Stumpf (1,2) and estuarine
application of sophisticated multichannel intercomparisons by Bostater et. al. (4)
inverse algorithms. have documented the advances and possible
routine use of AVHRR imagery for
suspended sediment (seston) and light
attenuation estimation in estuarine and
near coastal waters. Hardisky, Klemas
ro 3@ M 0@ M a 0 N W -N O@ -4 and Daiber (5) have shown the ability to
CAPE COD ... estimate wetland biomass and stress
LONG ISLAND SOUND detection which can be of value for
RARITAN > 0 permit activities within state and
DELAWARE federal programs. Remote sensing of
u.CHESAPEAXE SAV's (submerged aquatic vegetation) has
M.CHESAPEAXE also been described by Ackleson and
L.CHESAPEAKE CD Klemas (6). Table 1 indicates the
T.CHESAPEAKE performance- of remote sensing for
estuarine studies. This table summarizes
POTOMAC the results of a conference (7)
RAPPAHANNOCK concerning remote sensing of estuaries
JAMES T and represents a general consensus on the
ALBEMARLE SOUND 01 state of the art with respect to sensors
PAHL ICO SOUND for coastal ocean and estuarine
MEUSE applications.

1985 IMAGES 1.984 IMAGES

3. DELAWARE BAY REMOTE SENSING
Figure 2. Dimensionless light RESEARCH-OPTICAL MODEL INVERSIONS
attenuation (k/k max) for 14 estuarine
and near coastal water bodies for Ongoing research in Delaware Bay is
selected images during spring 1984, 1985 designed to provide the first estuarine
along the Middle Atlantic coast of the and near coastal waters data base for
US. optical characteristics of Case II and

463

aters Hydrod9nmaic ate u I it9
Node I Runoff Models istribution 81 onstituen

Land Use Tidal
Vegatation circulation Living Resources
Cover Local Currents Models
Sail in"ct On Living
Properties Fronts'plumes arine Resources
IEtc. Surface Salinity Suspended Sed.
Surface Wate
Temperature -Chlorophyll Habitat Quality
wind Fields -Pollution Uariables
tc. Plumes Wetlands And t9
Dredging Marsh Pmductiui
Water Temp. H9droacousticStoc]k
Etc. Assessments
SAV's
Fixed Platform
Remote Sensing
From Bridges
Etc.

Figure 3.' Description of estuarine models and Model inputs that
can be derived from remote sensing techniques.

measurements of reflectance or upwelled
radiance from below the water surface.
Case 111 (8) waters and associated water These models are then inverted in
quality indicator variables. order to estimate state variables which
can be used for mapping purposes or for
During the last 2 years a series of inputs for estuarine and near coastal
estuarine, near coastal water and water modeling efforts, as indicated in
continental shelf cruises have been figure 5.
conducted. During these cruises '(Which
are scheduled to coincide 'with optimal Thus, remote sensing provides the
satellite overpass conditions) in situ background for assessments and atlas
high resolution optical profiles have .(maps) information for considering
been conducted along with water chemistry impacts on living marine resources.
and algal primary productivity. The Figure 6 shows the conceptual approach
optical data collection (profiling) used in the current and proposed research
scheme is outlined in figure 4. '' optical for estimating state variables from
spectra (400-1100 nm) are calculated from remote sensing instrumentation.
downwelling (Ed) and upwelling (Eu)
measurements above and 'below.the water
surface. Edo -1, t EUG
This in situ data provides the
necessary optical spectrums for Z=O Air Water Interface
calculation of light attenuation (k), Edz,,,
backscatter (b) and irradiance z=JL
reflectance (R) spectra, which can then 'rEuz
be related to water concentrations such z=2 Ez=Eoe
as chlorophyll, suspended sediment or z=3
tracers'of freshwater (such as gelbstoff) t
into near coastal waters for estuarine Edo
,fronts and plume detection. z=4 t 2k=ku+kd
These measurements allow Ithe z=i t ku=kd
development and testing of new improved B=Rd2k
optical models based upon a variety of B=bw+bjLC IL+B2C2+B iC i
solutions and approximations to the 1k=kw+kjC1+k2C2+kiCi
radiative transfer problem. As indicated
in figure 4, the goal of this basic
research is to provide the algorithms for
relating water properties to aircraft, Figure 4. Delaware Bay in situ light
shipboard, in situ or satellite sensor measurements and analysis framework.

464

Develop Improved Optical Models.
For Organic/Inorganic Constituent
Mapping Using Optical/Biological/Chemical
Shipboard Data

Analysis of AVHRR,MSS,TM
Time Series and Ship Data Use Satellite Data To Map
to Determine Spatial & Organic/Inorganic Const ' ituents
Temporal Variability of By Inverting Algorithms And
Ph9sical/Biological Properties Performing Atmos. Corrections
of Delaware Ba4

Determine State Variables Of Estuaries
By Remote Sensin To Estimate Their
E Yq
Susceptibility To v irommental Degradation
And Changes In Living Resources Production
Delaware,Chesapeake,Potamac,Albemarle,Pamlico
,Hudson/Haritan,Long Island,Mobile, Others

Figure 5. Conceptualization and approach for optical model
development, data collection, and estimation of state variables
for estuarine comparisons.

4. MULTILEVEL REMOTE SENSING
5. SUMMARY
The above discussion has focused on Estuarine .remote sensi.ng has a
satellite sensors and in situ data unique role to play in estuarine and near
collection efforts. Remote sensing coastal water research and management.
instruments that need greater attention This role has been and will continue to
are sensor systems which canbe flown become more useful as new instrumentation
from low altitude aircraft. To date, can be. applied, and tested with
NASA and NAVY research and development concurrent in situ water column studies
funding has been utilized to develop and routine monitoring programs. Dynamic
these type of systems for ocean color. small scale features. and effects upon
living marine resources can also be
These systems have been used to assessed with new instrumentation that is
assess the optimal bands for detection of becoming available to marine scientists.
water column constituents. Recent
advances in instruments such as charged
couple device (CCD) multispectral video
cameras or other similar detectors need 6. REFERENCES
to be given multiuse development
considerations by agencies. These 1. Stumpf, R., Application of AVHRR
instruments can provide the capability to Satellite data to the study of sediment
cover large scale areas in a short time and chlorophyll in turbid coastal water,
period (hours to minutes), as well as for NOAA Technical Memorandum NESDIS AISC 7,
studying small scale processes in 1987, 50 pp.
estuaries.
2. Stumpf, R., Remote sensing of
suspended sediments in estuaries using
In addition, such devices have the atmospheric and'compositional corrections
potential for serving as fixed platform to AVHRR data, Proc. 21st Inter. Symp.
optical ground truth stations for Rem. Sens. of Env., Ann Arbor, M1, 1987,
satellite sensor systems as well as for 18 pp.
calibration purposes. They can also*can
be designed for use aboard ships as well 3. Swift, C., microwave radiometer
as aircraft, enabling remote sensing measurements of Cape Cod canal, Radio
techniques to be conducted under various Science, Vol. 9, No. 7, 1974, pp. 641-
sky and cloud conditions. 653.

465

Veg.& Biomass Coast- Bottom Susp. Susp. Chloro- Curr
Plat- Land & Veg. line Feat. Depth Sed. Sed. phyll Oil Surf. Water Circ: Wave Surf.
Sensor form Use Stress Erosion SAV Profiles Ptrns. Concen. Concen. Slicks Temp. Sal. Ptrns. Spectra Winds

Film Cameras A 3 1 3 3 2 2 1 1 2 0 0 2 2 1
S 2 1 2 2 1 2 1 1 1 0 0 2 2 1

Multispectral A 3 2+ 3 3 2 3 2 2+ 3 0 0 2 2 1
Scanners S 2 2 2 2 2 3 2 2 2 0 0 2 2 1

Thermal IR A 1 1 1 0 0 1 0 0 3 3 1 2 0 1
Scanners S 0 0 0 0 0 1 0 0 1 3 0 2 0 1

Laser A 0 0 1 3 3 1 0 0 1 0 0 0 3 1
Profilers S 0 0 1 1 1 0 0 0 0 0 0 0 2 0

Laser A 11 0 1 0 1 1 2 3 3 1 1 1 0 0
Fluorosensors S 0 0 0 0 0 0 1 1 1 0 0 0 0 0

Microwave A 1 0 1 0 0 1 1 1 3 3 2 2 1 3
Radiometers S 0 0 0 0 0 0 0 0 1 2 1 1 0 2

Imaging Radar A 2 1 3 0 1 1 0 0 3 1 1 2 3 2
(SAR or SLAB) S 1 0 2 0 1 0 0 0 2 0 0 1 2 1

:CODAR (Radar) G 0 0 0 0 0 0 0 0 0 0 1 3 2 2

RADS(Acoustic) Q 0 0 2 3 2 2 1 0 1 0 0 2 1 0

UW Camera G 0 0 2 3 2 2 1 1 1 0 0 1 0 0

Rating Platform
3 - Reliable (Operational) A - Aircraft (Medium or Low Altitude)
2 - Needs Additional Field Testing S - Spacecraft (Satellite)
1 - Limited Value (Future Potential) G - Ground (Boat,or Field)
0 - Not Applicable

Table 1. Estuarine remote sensing instrument capability - state
of the art consenus (7).

4. Bostater, C., et.al., Remote sensing 7. Klemas, V., Thomas, J. P., Zaitzeff,
of suspended sediment and light J. B., eds, Remote Sensing of Estuaries,
attenuation for estuarine and near US GPO, 1987 0-181-100(72341), 1987, 250
coastal waters - CASE III water types, PP.
(submitted for publication), Aug. 1988,
23 pp. S. BUkata, R. P., Jerome, J. H., Bruton,
J. E., Particulate concnetrations in Lake
5. Hardisky, M. A., Klemas, V., and St. Clair as recorded by a shipborne
Daiber, F. C., Remote sensing salt marsh multispectral optical monitoring system,
biomass and stress detection, In: Remote Sens. Env., 1988, pp. 201-229.
Advances in space research, Vol. 2.
COSPAR, Pergamon Press, London, 1983, pp.
219-229.

6. Ackleson, S. G., Klemas, V., Satellite.
remote sensing of submerged aquatic
vegetation in lower Chesapeake Bay, In:
Remote Sensing of Estuaries, US GPO 1987
0-181-100(72451), 1987, pp. 199-218.

466

RADAR DETECTION OF OCEANIC FRONTS

Donald L. Murphy

INTERNATIONAL ICE PATROL
Avery Point, Groton, CT 06340-6096

ABSTRACT second is a description of a 1986 experiment in
which SLAR images were compared with IR data and
In recent experiments, the Ice Patrol compared in situ measurements. Data from a third
radar images of the ocean with satellite infrared experiment, conducted in 1987, are still being
imagery, and hydrographic and drifting buoy data analyzed and the results will be reported later.
collected at the sea surface. The radar, a
side-looking airborne radar (SLAR), is a real BACKGROUND
aperture radar that operates in the X-band. The
SLAR can detect the sharp thermal fronts Since 1983, Ice Patrol has used SLAR as its
associated with warm eddies. However, the results primary instrument for iceberg reconnaissance.
of the experiments show that interpreting SLAR During the iceberg season, typically March through
images in the absence of supporting data is August, IIP flys approximately 80 iceberg
difficult. This instrument is most useful when reconnaissance patrols. Each flight is about
used in conjunction with other data sources, such 3200km in length and covers 65,000 sq. km, a small
as drifting buoys and AXBT's. portion of the Ice Patrol operations area.

The SLAR is a Motorola AN/APS-135, an X-band (3cm
wavelength), real aperture radar that scans the
INTRODUCTION sea surface in a plane normal to the flight path.
The radar image from both sides of the aircraft is
Since its formation in 1914, the International Ice displayed on a CRT that produces a continuous
Patrol (IIP) has tracked icebergs in the North negative image on photographic film. The radar
Atlantic, warning mariners of the danger to data are not recorded digitally. Navigational
navigation. Circulation in IIP's operations area information from the aircraft's inertial
(40*N to 52*N, 39*W to 57*W), which is east and navigation system is printed directly on the
south of Newfoundland, Canada, is dominated by two film. The SLAR system polarization is vertical
major currents. The first 'is the (VV) and is not selectable.
southward-f lowing, cold and relatively fresh (<
2*C and < 34.3 ppt) Labrador Current (LC); the The U. S. Coast Guard has two SLAR systems
second is the northeastward-f lowing, warm and more installed on its long-range patrol aircraft
saline (> 12*C and > 35.5 ppt) North Atlantic (HC-130). During the iceberg season one of the
Current (MAC). HC-1301s is deployed to Newfoundland on alternate
weeks.
A knowledge of the boundaries and areal extent of
the LC and MAC, and their associated eddies is Routine IIP patrols are flown at an altitude of
useful in predicting iceberg movement, a major Ice 8000 ft (2438m), which permits the radar to map a
Patrol task. Satellite infrared (IR) imagery of lOOkm-wide swath of the sea surface, 50km on each
the ocean's surface has been used with great side of the aircraft. This is the same swath
success in many parts of the world's oceans to map width as the operational mode planned for RADARSAT
thermal gradients and, thus, infer circulation. and approximately the same resolution ( 30m).
However, persistent clouds and fog in the IIP RADARSAT's scan mode permits coverage of a 500km
region restrict the use of IR detection of oceanic swath with a 100m resolution.
features to a few images per month, which is of
limited use for operations. Active microwave Airborne and satellite imaging radars have
systems (radars), which also have the ability to demonstrated the ability to detect the surface
map oceanic features (1), can penetrate clouds and manifestations of oceanic fronts associated with
fog, making them an attractive alternative to the eddies and the Gulf Stream, and internal waves.
IR sensors. Several investigations (2, 3, and 4) documented
the ability of the SEASAT synthetic aperture radar
Two examples are presented. The first is a (SAR) to detect such features. LaVoilette (5)
comparison between a SLAR-observed front on 28 compared imagery from the SEASAT SAR, with imagery
April 1985 and IR imagery of 26 April 1985. The from an earlier, less powerful version of the IIP

CH2585-8188/Oooo- 467 $.1 @1988 IEEE

SLAR, and satellite IR imagery. The results of Oceanic f rOnts . The detection of the f ront was
the experiment, which was conducted in the UP unimpeded' by the clouds and fog. However, the
operations area, showed that both radars could SLAR image does not give quantitative information
detect the fronts seen in the IR images. about the thermal gradient ..on the sea surface.
What was lacking was a surface measurement program
Imaging radars map the sea-surface roughness to document the meteorological conditions, water
through Brags scattering (1), which for the 3-cm. mass characteristics, and current information. An
wavelength and incidence angles of the UP SLAR, Ice Patrol SLAR experiment in 1986 addressed these
results in a sensitivity to ocean waves needs.
approximately 2-cm long. As a result, the UP
SLAR imagery of the ocean is essentially a map of April 1986
the distribution of these 2 em-long waves; the
SEASAT SAR was sensitive to 30-cm, wavelengths in April and May 1986 Ice Patrol conducted a study
(6). on both radars, differences in surface east of the Grand Banks of Newfoundland in which
roughness are indicated on the radar image as SLAR images of the sea surface were compared with
tonal changes. Thus, there are light and dark in situ hydrographic data, drifting buoy
areas on the images that correspond to differences, trajectories, and an AVHRR image (8). The study
in the reflected radar energy. focused on a warm-core eddy spawned from and
interacting with the NAC, which is an extension of
Interpretation of the images requires an the Gulf Stream (east of about 50*W).
understanding of how wind stress, current
gradients, temperature, salinity, etc. modulate Four aerial SLAR surveys, at approximately
capillary and short gravity waves on the ocean one-week intervals, provided radar images of the
surface; but, our understanding is poor. Lichy 2jt sea surface in the study area. The first survey
al (7), who tracked a warm core ring using SEASAT (26 April 1986) covered 127,000 sq km and
SAR data, found that within the warm water there identified a site to conduct the hydrographic
was a more intense radar return than from the study. The three subsequent flights (2, 9, and 17
surrounding area. But much additional research is May 1986) focused on the hydrographic study area,
required before radar images can be interpreted each mapping 56,000 sq km with overlapping
with confidence. coverage.

April 1985 on the last day of the experiment the AVHRR on
NO&A 9 provided the only useable T.R. image of the
The first example is a comparison between the area.
SLAR-observed north wall of the Gulf Stream and
satellite infrared (IR) data, as analyzed by A photo mosaic of the 26 April SLAR survey (Figure
NOAA's National Environmental Satellite and Data 3) shows what we interpret as the WAC, appearing
Information System (NESDIS). as a dark region along the southern and eastern
edge of the image. Over most of the image, the
Figure I shows a simplified reproduction of NOAA boundary between light and dark areas is sharply
interpretation of a Advanced Very High Resolution defined. However, its shape is complex and not
Radiometry (AVHRR) image south of Newfoundland, easily explained. The area enclosed by the white
Canada. The analysis, a composit based on images. box was the site of a 68 station bydrographic (26
taken on 25-26 April 1985, shows a warm-core eddy April - 3 May 1986) survey; during which CTD
interacting with the Gulf Stream in the vicinity (conductivity, temperature, and depth) casts to
of 41*H, 50*W. The inset shows the interpreter's about 1000 m were taken along ten north/south
worksheet on which is marked the location of the transects.
Gulf Stream's north wall.
Figure 4 shows the distribution of sea surface
An Ice Patrol SLAR reconnaissance flight mapped a temperature, showing a sharp thermal front that is
portion of the region two days after the AVHRR nearly coincident with the SLAR-observed
image was taken. on the date of this patrol (28 boundary. As in the previous example, the dark
April 1985) the sea surface was obscured by clouds area indicates the presence of warm water. The
and fog, preventing the collection of IR images. hydrographic survey Iidentified the water
The SLAR image (Figure 2) shows a sharp tonal properties as those of NAC origin (T > 12*C,
boundary that is conincident with the IR-observed salinity > 35.5 ppt). The combination of the
north wall at 50*W longitude. Figure 3 is a hy0rographic. data and trajectories of
negative image, thus the dark area is a region of satellite-tracked buoys showed that this feature
high radar return (i.e., radar bright). This area was a meander of the NAC that was developing into
of high radar return marks the warm Gulf Stream a warm-core eddy. Henceforth, the feature is
waters. This result is consistent with the referred to as an eddy, although determining
findings of Lichy et al. (7), who tracked A exactly when the feature developed a closed
warm-core ring using SEASAT SAR data, that within circulation is not possible from the data.
warm water there was a more intense radar return
than from the surrounding area. Following the evolution of the eddy over the
three -week period of observation was difficult for
The results of this simple comparison demonstrated two reasons: first, the inability of the SLAR
that the UP SLAR had some promise as an imagery to provide a closed boundary on all of the
instrument that could be used to detect major flights, and second, the complexitiy of an eddy

468

interacting with the NAC and trapped against the definition.
Grand Bank and the Labrador Current. The location
of the northern and eastern boundaries of the eddy The 1986 study illustrates the importance of
was well def ined on both 26 April and 2 May SLAR research that blends remote sensing with in-situ
surveys; however, in neither survey were the sampling, with the goal of studying oceanic
western and southern boundaries,well defined. processes. Without the SLAR we could not have
located the fronts as easily, nor recognized the
The 9 May SLAR survey provided the most complex spatial and temporal variability of the system.
and ambiguous images of the study. The Without the in-situ sampling, the imagery would
northernmost frontal location remained nearly have been another opportunity for 'unfounded
unchanged. However,. this survey provided the speculation.
first good image of the southern portion of the
feature. REFERENCES

The 17 May SLAR survey, conducted on the last day 1. Robinson, I. S., 1985. Satellite
of the experiment, provided the most remarkable Oceanography: An Introduction for
image (Figure 5). By this date trajectories of Oceanographers and Remote-Sensinx
sate 11 ite-tracked drifters had confirmed that the Scientists. West Sussex, England: Horwood
feature bad the characteristic anticyclonic Limited. 455pp.
.(clockwise) circulation of a warm-core eddy. The
SLAR image shows an eddy with a complex shape 2. teal, R. C., P. S. DeLeonibus and I. Katz,
interacting with the NAC. Along the southern 1981. Spaceborne Synthetic Aperture Radar
boundary of the eddy is a sawtooth-pattern with a for Oceanography. The John Hopkins
peak-to-peak separation of 35km and a height of oceanographic studies, No. 10. The Johns
20km. As on other dates, not all of the Hopkins University Press, Baltimore, MD,
boundaries are clearly defined, particularly the 215pp.
western boundary. As a result, it is difficult to
estimate the size of the eddy based solely on the 3. Fu, L. And B. Holt, 1982. SEASAT Views Oceans
StAR imagery; the best size estimate is 160 by 80 and Sea Ice with Synthetic Aperture Radar.
km. Publication 81-120, Jet Propulsion
Laboratory, California institute of
May 17 was the only cloud-free day during the Technology, Pasadena, California, 200pp.
three-week experiment. An AVRRR image from the
NOAA 9 satellite taken 8 hours before the SLAR 4. Hayes, R. M., 1981. Detection of the Gulf
image, shows an excellent agreement of the frontal Stream, in Spaceborne Synthetic Aperture
boundaries (Figure 6). In addition, the SLAR Radar for Oceanography. The Johns Hopkins
boundaries are as sharp as those seen on the IR oceanographic Studies, ed. by R. C. Beal.
imagery. The Johns Hopkins University Press,
CONCLUSIONS Baltimore, MD, p. 146-160.

SLkR imagery is difficult to interpret but can be 5. LaVoilette, P. E., 1983. The Grand Banks
used with other data to gain a better Experiment: A Satellite/Aircraft/Ship
understanding of oceanic processes. in addition, Experiment to Explore the Ability of
because SLAR and SAR imagery portray similar Specialized Radars to define Ocean Fronts.
features, the more we learn about SLAR now, the Report 49. Naval Ocean Research and
better prepared we will be to interpret satellite Development Activity, NSTL Station, MS
and airborne SAR imagery when it becomes routinely 39529, 12 .6 pp.
available.
6. Vesecky, J. F. and R. H. Stewart, 1982. The
The use of aircraft-borne SLAR, and eventually observations of Ocean Surface Pheonomena
satellite-borne SAR, in determining ocean Using Imagery from the SEASAT Synthetic
circulation near the Grand Banks holds great Aperture Radar: An Assessment. Journal of
promise for improving IIP operations. However, Geophysical Research, 87(C5):3397-3420.
the work in interpreting radar imagery of the
ocean surface has only started. Experiments such 7. Lichy,D. E., M. G. Hattie, and L. J. Mancini,
as that described here mist be repeated several 1981. Tracking of a Warm Core Eddy, in
times with a broad range of oceanic features. Spaceborne Synthetic Aperture Radar for
Ultimately, the combination of active microwave Oceanography. The Johns Hopkins oceano-
imagery, air-deployed drifting buoys, and AXBT's graphic Studies, ed. by R. C. Beal. The
will permit IIP to gather the required near Johns Hopkins University Press, Baltimore,
real-time data. MD, p 171-182.

IIP SLAR data suffer somewhat from the inability 8. Thayer, V. B. and D. L. Murphy, 1988. SLAR
to record digital radar data aboard aircraft. observations of ocean Fronts East of the
This is not important for the major features, such Grand Banks of Newfoundland. Proceedings,
as the obvious tonal signal that marked the Gulf Ilth Annual Canadian Conference on Remote
Stream in the first example and the eddy in the Sensing. Waterloo University, Waterloo,
second. However, for more subtle features, Ontario, Canada (In Press).
digital processing would have provided better

469

70' G5* so, 53' 50' 43,
so*

GS - GL41 Stream
WE - Warm-core
E(d*dy

r
C

GS - - CS

40*

Nr WALL

Figure 1. Reproduced National Earth Satellite Service (NESDIS) product from April 26, 1985.
Inset NESDIS worksheet from 25-26 April 1985.

422ON

M, 490OW
42DON 4950V

A
@GS GL
WM
41 SU
1
arm,=
E

Figure 2., A segment of SLAR film from an Ice Patrol reconnaissance flight of 28 April 1985.
Warm Gulf Stream water appears dark.

470.

7-7@

";t;@3
77@7-1@"'- ttzQ@11-t

z,K

Figure 3. Photomosaic constructed from SLAR survey of 26 April 1986. The area enclosed by the
black box is the area of the hydrographic survey.

46 N

0 0 0 0
45.
0 13'
0 0 0

0 0 0 0
45 N 13*
+ 0 0 0

0 0
0 leo-
44.5 N 120
+ 120

44 N

40 W 47 N 46 W 45 W
@z

Figure 4. Sea surface (0.5-1.0m) temperature distribution based on the first phase (27 April 3
May) hydrography.

471

IGP
War
2W
P_-'?' - --''
451ON-
W-0

r--#, -d- e

)Ef PAV
01

43*40'N-

4930 w 47PS6 W 45*30'W

Figure 5. Photomosaic constructed from SLAR survey of 17 May 1986.

Figure 6. Infrared image from the Advanced High Resolution Radiometer on. board NOAA 9 on 17 May
1986

472

ADAPTING THE NSCAT DATA SYSTEM TO CHANGING REQUIREMENTS

J.R. Benada, D.T. Cuddy, B.H. Jai

Radar Science and Engineering Section
Jet Propulsion Laboratory
California Institute of Technology
Pasadena, California ?110? USA

ABSTRACT: NSCAT is a spaceborne, 8-beam process it to wind speed vectors, deliver
scatterometer which will measure ocean the processed data to the principal inves-
backscatter. The backscatter measurements tigators and archive the data.
are processed on the ground to oceanic
wind vectors. Vector winds will be re- 2 MISSION & INSTRUMENT DESCRIPTION
trieved over 90% of the ice free, global
oceans every two days for a three year 2. 1 Mission
mission. The Data System is a non-real- The nominal NSCAT 'mission will provide
time ground based science data processing wind speed measurements over 90% of the
system. It ingests backscatter telemetry, ice-free, global oceans every two days.
processes it to wind vectors, archives and The wind speed root-mean-square (rms)
distributes the wind vector data and other accuracy is the greater of 2 m/s or 10%
products. Brightness temperatures from an from 3 - 30 m/s. The wind-direction rm5
on-board microwave radiometer are used for accuracy is 20 degrees in the 'same wind
estimating atmospheric water. The pro- speed range. The spatial resolution
ducts are delivered to a Science Team required is 50 km. The instrument is
composed of oceanographers, meteorolo- designed for a circular, polar, sun syn-
gists, climatologists and other scien- chronous orbit with a nominal altitude of
tists. This paper describes the Data E330 km.
System, particularly the changes to the
baseline design in response to changes in The satellite bus is unknown at this
requirements. The new requirements are in writing. Originally, the scatterometer
five areas: 1) telemetry data source, 2) was to be one of the instruments on the
data catalog, 3) hosting Instrument Opera- Navy's NROSS satellite with a launch in
tions, 4) definition of data granularity 1991. However, the NROSS program was
(revolutions) and 5) Science Team data cancelled in 19B7. The NSCAT project is
needs. The Science Team requirements are: in the process of investigating and evalu-
a) need for new or revised products, b) ating other options.
need for different delivery volume and
schedule and c) need for electronic com- 2.2 Instruments
munication among team and with the pro- The instruments of interest to the project
ject. are the scatterometer, of course, and the
Special Sensor Microwave/ Imager, SSM/I,
both described below.

1 INTRODUCTION 2.2.1 Scatterometer
The scatterometer is an 14 GHz active
NSCAT '(NASA Scat terometer) 1,2 is a space- radar which measures the normalized ocean
borne, active radar which measures the radar backscatter cross-section (go).
surface backscatter of the earth oceans. Eight tran5mit/receive fan beam antennas
These measurements are used to calculate illuminate the earth in three different
surface wind vectors with a resolution of directions on each side of the subsatel-
50 kilometers. Wind vectors are retrieved lite track. The antenna illumination
over most of the worlds open oceans every pattern on the earth is shown in Figure 1.
two days. From these winds, the Principal Antennas 2 and 5 are actually dual anten-
Investigators (PI's), including oceano- nas with different polarizations. The
graphers, meteorologists and climatolo- other antennas are single polarization.
gists, will, for examplex study wind In the 350 km inter-swath gap centered on
forcing of ocean currents, detect meteoro- the subsatellite track, a. data cannot be
logical features, and perhaps predict the used for wind retrieval.
occurrence of an El NiPio. This paper
describes the ground-based Data System The 8 antennas are individually fired in
which will ingest instrument telemetry, a set sequence which is completed in about

CH2585-8/88/oooo-473 $1 @1988 IEEE

4 seconds. The return signal is received the satellite ephemeris calculation.
before the next antenna fires. The re-
ceived signal for each antenna is proces- (d) Decommutation and DN-EU conversion -
sed on board using range gating and dop- unraveling of spacecraft telemetry.
pler filtering of the backscatter power to DN-EU conversion is the conversion of
generate a set of 25 measurement areas on instrument Data Numbers into Engi-
the ocean surface called Cr, cells. Each a. neering Units, e.g., from spacecraft
cell has a spatial resolution of roughly counts to the value of received power
25km x 25km. in watts.

As the instrument flies along, it measures (e) Nadir location - calculation of the
the backscatter of a patch of ocean from nadir point from the satellite ephem-
three different directions and two dif- eris.
ferent polarizations. These different
measurements are necessary to estimate the (f) Cell location 'and geometry - calcula-
wind direction and to improve the wind tion of the earth surface location
speed estimate. and shape of each a. cell in an anten-
na beam illumination pattern. The
2.2.2 SSM/I cells are roughly in a line, running
The Special Sensor Microwave/ Imager (SSM/- out from the nadir point.
I) is the same instrument as is currently
flying an DMSP-FB. It is a passive, scan- (g) Flag land and ice - The instrument
ning instrument which measures the inci- cannot be used to measure winds over
dent radiation, on the SSM/I, at 19, 21, or near land or ice. Each cr, cel 1 is
35 GHz from roughly circular footprints flagged based on a fixed land map and
along the scan. Each scan is an arc cent- a temporally varying ice map.
ered on the nadir point and extending
almost to the edges of the scatterometer (h) Flag excess atmospheric absorption
swath. The SSM/I scan includes data in The instrument cannot be used to
the scatterometer inter-swath gap. The measure winds in areas of large at-
instrument also has an 85 GHz channel mospheric absorption. This is usual-
which NSCAT does not use. The surface ly due to large amounts of atmospher-
brightness temperatures, Tb, at those ic liquid water, namely clouds and
frequencies can ,be calculated. The Tb's rain. The SSM/I data is used for
which are co-located with scatterometer setting flags. The SSM/I Tb foot-
sigma-o measurements are used to calculate prints (usually more than 1) must be
the atmospheric absorption due to water. co-located with one a. cell and the
This is used to set an "excess absorption" flag determined.
f lag in the data to indicate that the
scatterometer measurement at that location (i) a. calculation - this is the normal-
is seriously degraded. ized radar backscatter, corrected for
instrument and antenna effects. Rms
3 DATA SYSTEM FUNCTIONS & ORIGINAL DESIGN errors are also calculated.

The Data System has three basic functions (j) a. grouping - the a.'s are grouped
which correspond to a partitioning among into ground-track based Wind Vector
three subsystems: ingestion, processing, Cells (WVC's) in preparation for
and archiving and distribution. vector wind retrieval. The WVC's are
the cells of a 50kmx5Okm grid running
3.1 Ingestion and Processing parallel to the ground track. The
grid extends slightly beyond the far-
3.1.1 Function% thest possible cross-track distance
The following are the basic functions for a a. cell.
necessary to ingest and process the scat-
terometer data: (k) Vector wind "retrieval" - a calcula-
tion based on an algorithm estimating
(a) Telemetry ingestion - read telemetry the wind speed and direction which
tape, check for errors, merge with would produce the observed collection
previous data and check for missing of U0, s in the WVC. This normally
data. produces 2 to 4 solutions with a
likelihood calculated for each.
(b) UTC time convertsion - convert the These solutions generally have very
instrument timetag into UTC (Univer- similar wind speeds but widely dif-
sal Time Coordinated). ferent direction. In fact, frequent-
ly two solutions will have wind di-
(c) Cut into revs - A "rev" is one satel- rections about 180 degrees apart.
lite orbit . The revs are based on a
set of rev start times generated from (1) Vector selection - this step takes

474

the solutions and probabilities for engineers.
the WYC from the wind retrieval and
selects one wind speed-direction (e) Provide information exchange utility
solution. The algorithm looks at the - for communication among the PI's
neighboring cells to help in the and between PI's and the Data System
selection. staff.

An additional processing step is the pro- 3.2.2 Original Design
duction of a two day vector average global The archive and distribution design as-
wind map. The format and contents of this sumed that the users, mainly the PI'sv
map are under study at this time but it would want to select their data interac-
will not affect the rest of the functions tively over dial-up or network connec-
discussed here. tions. Having selected the data, the
PI's would want it delivered electroni-
It is important to notice the two differ- cally or by mail (tape) as soon as pos-
ent organizations of the data. One is sible, normally within one or two days.
based on the antenna beam illumination Basically, these are the services sup-
pattern, or "beam". The other is based on plied by the NASA Ocean Data System,
the Wind Vector Cell or WVC. The data is NODS. The plan was to use NODS software
organized by "beam" up to "cro Grouping". on a dedic'ated NSCAT computer and only
From that point on, the data is organized make slight modifications to accommodate
by WVC. This is shown in Table 1. NSCAT project data.

The SSM/I data is organized by scan. The 4 NEW OR ALTERED REQUIREMENTS
data are cut into revs but otherwise unal-
tered. The Tb footprint location and During 1987, several new or changed re-
geometry are included in the incoming quirements came to light. Some, such as
data. Thus the data is sufficient for co- items 4.1, 2 and 4, arose from design
locating with a. cells and calculating the considerations. One, 4.3, came from the
atmospheric absorption. project. The rest, all of 4.5, came from
the PI's. This was based on their better
3.1.2 Original Design understanding of their own needs and of
The original ingestion and processing the original Data System design.
subsystem design included the following
features which bear particularly on the 4.1 Data Source Uncertainty - Due to
new requirements considered later: cancellation of NROSS, we were not able
to define the format, data time span,
(a) Division of the functions between the delivery frequency or transfer method for
ingestion and processing subsystems telemetry data. However, we could not
as shown in Table 1. hold up ingestion design.

(b) The original data level definitions 4.2 Data Catalog
shown in Tables 1 and 2A. A catalog functions as a bookkeeper for
the data handled by a subsystem. Each
(c) A "rev" was defined as starting at Data System subsystem has its own catalog
the northward going equator crossing. which keeps information about all the
data types it handles. For example, the
(d) The Data System provided no SSM/I ingestion subsystem catalog contains the
data to the PI's. information about external inputs and
level 1.0 data. The relevant parts of
3.2 Archive and Distribution this must be transmitted to the proces-
sing subsystem for it to know what data
3.2.1 Functions is ready to be processed.
The following are the basic functions of
the data archive and distribution: Originally, the communication among these
three subsystems was to be done once per
(a) Archive all project data - all data day covering the last 24 hours. This was
necessary to reproduce the data prod- due to the subsystems' heterogeneous com-
ucts. puter environment and DBMS software.
Thus, the information is up-to-date only
(b) Deliver data to permanent archive - when the information file is transferred.
i.e. NASA Ocean Data System.
4.3 Host Instrument Operations
(c) Receive orders from PI's which data The project identified a need for a JPL
set was needed and for which time based Instrument Operations facility.
interval. This is a non-real-time function which
will monitor instrument performance and
(d) Distribute data to PI's and project trends. It will also generate parameters

475

to be loaded on the on-board processor. from SSM/I data for comparison with
The real-time monitoring and communication NSCAT winds and for estimating wind
with the satellite and instrument will be in the inter-swath gap.
at the project or spacecraft operations
control center. The project wanted this 4.5.2 Different delivery volume and
activity on the Data System as it would be schedule
the only local, NSCAT-dedicated computer
facility operational during the mission. (a) The capability to provide routine
delivery of large data sets pre-
4.4 Data Granularity scribed in advance. Monthly delivery
Overlapping of revs necessary to retrieve would be satisfactory.
winds greatly complicates processing,
reprocessing and cataloging. The revs are (b) Retain the ability to request data
defined based on the nadir location but many months old, but with a frequency
the NSCAT instrument "looks" forward and and volume expected to be small.
backward of the nadir point by hundreds of
kilometers (with respect to the direction (c) No need existed for an on-line re-
of flight). Likewise the SSM/I scans quest or capability to provide selec-
always aft of the nadir point. Thus, with ted subsets of data quickly.
revs beginning at the equator, a rev of
Level 2.0 data (WVC data) contains data 4.5.3 A need remained for an electronic
from three different revs of Level 1.5 communication between P.I.'s and with the
(beam data) and two different SSM/I revs. project for information and mail.

4.5 Science Team Requirements 5 REDESIGN

4.5.1 New or revised data products The following are the changes made in the
The PI's wanted some changes in the data Data System in -response to the require-
type to be delivered to them. These were ments given above.
suggested by advisory committees and were
discussed with the project in terms of 5.1 Ingestion Redesign
practicality. 5.1.1 Isolation of external interface
Because we knew the basic content of the
(a) Some PI's wanted to retrieve winds incoming data, we defined an intermediate
from the calculated u.'s. These ex- level which we could specify completely.
isted in Level 1.5. However, the The: ingestion subsystem could now be
large size of this file (over 500 designed from that point on. Later,
MB/day) made this impractical. Also, after the external interface was speci-
Level 1.5 is organized by antenna fied, a conversion module would be writ-
"beam" and uD's must be grouped into ten to convert from the true input to the
WVC's for wind vector retrieval. intermediate level.
Thus, a smaller volume product ar-
ranged in WVC's was needed. 5.1.2 Redefinition of Level 1.0
Decommutation and DN-EU conversion at JPL
(b) The wind vector product that was to are traditionally part of the data inges-
be delivered, Level 2.5. contained tion process. Those functions were added
only the selected wind direction to the ingestion subsystem. The defini-
based on the ambiguity removal al- tion of Level 1.0 was changed according-
gorithm. The PI's wanted all direc- ly. See Tables I and 2B.
tional ambiguities delivered in addi-
tion to the selected wind vector. 5.2 Data Catalog
With such a product, investigators A distributed DBMS was adopted for the
could re-select wind direction based catalogs. This will keep the catalogs
on another algorithm or ancillary up-to-date in all three subsystems. In-
data (such as in situ). stead of transferring catalog information
once a day, the catalog information about
(c) A requirement was generated to deli- the products will be transferred to a
ver SSM/I data colocated with a a subsystem whenever the products are
data, in order to make atmospheric transferred. In other words, the minimum
corrections instead of the flagging information which is needed by the recei-
done by the Data System. Furthermore ving subsystem will be redundantly stored
it was desired to have an SSM/I pro- in both the generating and receiving
duct compatible with the Level 2.5 subsystems.
wind vector product. This would
allow calculation of wind speeds 5.3 Host Instrument Operations
Inasmuch as the ingestion function was
relatively low in computer loading, it

476

was a suitable place to host Instrument production and archiving. However,
Operations. The ingestion functions can no "on-line" selection and generation
be completed in four to six hours, leaving for users. Users submit a request
ample time for expected Instrument Opera- (by mail) which is filled by the team
tions activities. Even during periods of and sent to the user on tape.
high activity, such as a instrument anoma-
ly, normal ingestion can keep pace on an (c) No on-line requests-- There is no
altered schedule. need for an outside user interface.
Further, there is no need for rapid
5.4 Data Granularity subsetting in order to limit user
The revs were redefined as beginning (and connect time. This greatly reduces
ending) at the southernmost latitude. The archive sophistication. A fairly
ao s in this region cannot be used for simple file cataloging system, where
wind retrieval because of the presence of data is stored by revs, is suffici-
land and ice. Thus, usable data from one ent.. Because the phone or network
rev never overlaps the adjacent revs. line to this subsystem is no longer
This redefinition resulted in no loss of needed, system security is improved.
wind vector data.
5.5.3 Need for electronic communication
5.5 Science Team Requirements The need for communications among PI's
and with the project for information and
5.5.1 New or revised data products mail will be met by off-the-shelf elec-
The revised level definitions are illus- tronic bulletin board and electronic mail
trated in Table I and summarized in Table software. This provides a low cost, low
2B. Details of the changes are presented maintenance system. Inasmuch as there
below. is no electronic connection to the sub-
systems it provides good security for the
(a) Level 1.7 - Writing minimum data out system.
immediately after binning, produces
a low volume (100MB), binned a. data 6 CONCLUSION
product with little additional pro-
cessing. As a result of the design changes, we now
have a Data System which is simpler,
(b) Level 2.0 changed - Level .2.0 was especially in the data archive and dis-
redefined to be all wind vector solu- tribution subsystem. The Principal. In-
tions but with a "check mark" on the vestigators have data types which more
solution selected by the Data System closely match their needs and give them
algorithm. the flexibility they want. The design of
the Data System is now complete and im-
(c) SSM/I Tb and <Tb> - Providing the plementation will begin soon.
original SSM/I Tb data cut into revs
satisfied the need for a product 7 ACKNOWLEDGEMENT
compatible with a. data. We defined
a new product, SSM/I <Tb>-, in which The refined Data System design was the
the Tb's are averaged over a WVC. product of a multidiciplinary design team
This averaging was done for the en- lead by Steve Gunter. Members of the
tire scan including the inter-swath team included Dan Bonbright, Gloria Con-
gap- This was organized in exactly ner, Robert Insley, Chris Leng and the
the same way as the Level 2.0 data. authors.
We would like to thank the other members
5.5.2 Different Delivery Method of the NSCAT Data System Team for their
assistance with the overall design and
(a) Standing Orders method preferred - their work in software development.
Most data will be requested by PI's
with a standing order which is sent The work described in this paper was
in advance and infrequently changed. carried out by the Jet Propulsion Labora-
Immediately after level data produc- tory, California Institute of Technology
tion, the data specified by the order under-contract with the National Aeronau-
is extracted to disk or tape. When tics and Space Administration.
a sufficient volume of data for a PI
is collected or a time limit (one 8 REFERENCES
month) is passed, the data is shipped
on tape. 1. Martin, B D, Freilich, M H, Li F Kv
Callahan, P S, "An overview of the NSCAT-
(b) "Historical" archive - The archive /NROSS Program," Proceedings of a Work-
and distribution subsystem still must shop on ERS-1 Wind and Wave Calibration,
handle historical requests, i.e. Schliersee, FRG, 22-6 June, 1986 (ESA SP-
those that arrive well after data 262, Sept. 1986)

477

TABLE 2A
2. Callahan, P S, Benada, J R, "NASA ORIGINAL DATA LEVELS
Scatterometer Data Processing System -
Features for Validation," ibid. LEVEL ORIGINAL DEFINITION
LO Rev oriented NSCAT telemetry
data (This is raw data, i.e.,
without any conversion.)

L1.0 Rev oriented, decommutated, EU
SUBSATELLITE converted, and nadir point lo-
ANTENNA 6 TRAC K ANTENNA I cated NSCAT data in engineering
ANTENNA 2 unit.
45-- _45' L1.5 Rev oriented data reduced to
-65'
normalized backscatter cross
section, earth located sigma-0
cells with all geometry param-
% eters for each cell, and flags
REFERENCE for the ice and land.
CELLS L2.0 Rev oriented Wind Vector Cells.
(with wind directional ambigu-
5!_ ities). The data is stored in 50
km grid form.
ANTENNA 5 45!._ _45'

L2.5 Rev oriented WVCS with only
selected solution included.
ANTENNA4- LEFTWINDVECTOR RIGHT WIND VECTOR ANTENNA 3
SWATH SWATH
175 175 600 k----@ L3.0 Non-rev oriented wind map.
- 600 k--Km + k@

Figure 1 NSCAT Antenna Pattern TABLE 2B
NEW DATA LEVELS

LEVEL NEW DEFINITION

TABLE I LO Deleted
FUNCTION LEVEL SUBSYSTEM
ORIG NEW ORIG NEW L1.0 The same as the old definition
except that the data is not
LO.0 nadir located.
Telemetry ingest I I
UTC time convert L1.5 No change.
Cut rev's L1.0 I I
Decom and DN-EU LI.V P L1.7 Sigma-0 cells grouped in sub-
Nadir location P track-grid ("binnned") form,
Cell location i.e., the same form as Wind
Flag land, ice Vector Cell.
Flag absorption
a. calculation L1.5 L1.5 L2.0 Combination of original Level
a. grouping L1.7 2.0 and 2.5 data, i.e., a check
Wind retrieval L2.0 mark will be made on the selec-
Vector selection L2.5 L2.0 ted ambiguity while the rest
ambiguities will be kept in for
I = Ingestion Subsystem reference.
P = Processing Subsystem
L2.5 Deleted

L3.0 No change
-4
5 @o

41 t-

To
@U@FT'WIN.VEC R RIGHT NWIN._V@CMR
1H
.................. __ . . ......... ........

478

THE INFLUENCE OF PACING TECHNOLOGIES ON ENVIRONMENTAL
APPLICATION OF SPACE-BASED SYNTHETIC APERTURE RADAR

Samuel W. McCandless, Jr. Dr. John Curlander

User Systems Incorporated Jet Propulsion Laboratory
4608 Willet Drive 4800 Oak Grove Drive
Annandale, Virginia 2,2003 Pasadena, California 91109

ABSTRACT TABLE 1
SPACE-BASED SAR
space-based synthetic aperture radar (SAR) PACING TECHNOLOGY ASSESSMENT
sensing of ocean and polar regions began
with the NASA SEASAT system in 1978. ELEMENT STATUS
Since then, ocean applications have
continued with space shuttle flights and Exciter Supports applications
free-flyer satellites. Each of these SAR requirements.
systems have been applications limited due
to technology driven design constraints Transmitter Available T/R modules &
such as on-board recording (SEASAT, SIR- tubes over a wide range
A), and data collection rate limits (SIR- of frequencies.
A, SIR-B). Recent breakthroughs in
relevant technology areas permit re- Antenna Area & tolerance
evaluation of SAR system architectures available to meet user
and design approaches. Innovations needs.
include optical/digital on-board image
processors, advanced magnetic tape Receiver Bandwidth adequate for
recorders and data compression applications.
algorithms. This paper explores SAR calibration needs
pacing technology areas in terms of improvement.
projected capabilities that will produce
the most benefits for ocean users. A/D Converter, Data rate limitations
Data Recorder, can reduce swathwidth,
Downlink. coverage, dynamic range
SAR SYSTEM ELEMENTS and range resolution.

oceanic application of space-based SAR is Image Processing Does not support,real
currently limited in the areas of data time application.
rate conversion, storage, downlink and Processing choice and
image processing, as shown in Table 1. location (space or
This barrier can affect important ground) can limit
performance values such as swathwidth, dynamic range,
dynamic range, range spatial resolution, calibration
sensitivity and real time application. sensitivity, range
resolution, etc.
Several of the SAR subsystems can be
eliminated from consideration as pacing
technologies. For most historic, planned These subsystems can be arranged in a
and projected designs the portions of the variety of ways. Regardless of how the
SAR from the radar transmitter through the steps are performed, each can impose
radar receiver can be eliminated as pacing significant limits on SAR application.
technologies. Image signal processing can be performed
on-board (preceding the recording and
The basic pacing technologies for space- downlink steps) or on the ground. Both
based SAR are: options will be discussed.

- Analog to digital (A/D) conversion A/D CONVERSION
- On-board image processing
- On-board data recording If the data is to be processed digitally,
- Data transmission to the ground as is the current U.S. preference, it must
be sampled and converted from analog to
digital form, as shown inFigure 1.

CH2585-8/88/0000- 479 $1 @1988 IEEE

Following the analog-to-digital HIGH RATE RECORDER
conversion of the video signal produced by
the SAR receiver, the digital data stream The high rate recorder system must be
is typically time expansion buffered to capable of recording at data rates of 50
reduce the instantaneous data rate to a Mbps and above with a recording capacity
lower sustained rate which is continuous of at least 8 Gbytes (about 20 minutes of
over the interpulse period. operation) to handle even the simplest SAR
system (e.g., single channel, 50 km
For most SAR applications that require swath). For the multiple channel, wide
high precision estimation of the echo swath, multi -polarization SAR required to
power, an 8 bit ADC is needed to prevent meet future observation requirements, an
significant saturation or quantization order of magnitude better performance is
noise in the sampled data. A 300 MHz\ 8 needed. To date, the most advanced
bit ADC manufactured by Tetronix is recorder flown in space, manufactured by
currently operationally used in airborne Odetics Corporation, is capable of a
radar systems. Techniques have been maximum record rate of 60 Mbps with a
developed for adaptive selection of the maximum capacity of 8 Gbytes and a BER of
four most significant bits from the 8 bit 10-7. This system has been flown on SPOT,
ADC output to reduce the effective data on the Shuttle with SIR-B 'and is planned
rate. A threshold level (or exponent) is on upcoming SAR satellites such as JERS-1.
calculated to accompany a block of data
based on the average power in that block. The , recording technique used in the
This technique, referred to as Block Odetics and most ground recorders (e.g.,
Adaptive Quantization (BAQ), can reduce Honeywell HD-96) is called linear
the data rate by a factor of 2. The BAQ recording, where as many as 42 parallel
technique is being used by the NASA-JPL longitudinal tracks have been packed onto
Magellan Venus Radar Mapper and the a one inch wide tape. The helical
Shuttle Imaging Radar SIR@C, both recording format can realize 600 tracks
scheduled for early 90s operation. per inch producing a significant increase
in packing density. The helical format
A large amount of work has been devoted to Schlumberger Industries model PV6410 (MIL
another data reduction technique STP 2179) is capable of 240 Mbps a
identified as pre-image data compression. capacity of 50 Gbytes at a BER of <1;_10
However, this area is much less promising. (Reference 1). This machine is being
Using noiseless (lossless) coding utilized in military combat aircraft,
techniques, such as a Huffman code to helicopters, and ships. It is currently
eliminate redundancy in the data, will undergoing testing by NASA for use with
yield no more than a 20 to 30% reduction SIR-C.
in data volume. This limitation is
primarily due to SAR "speckle noise" Although high rate recording systems are
effects. Lossy coding techniques for still a factor of 2 less than what is
reducing the data volume/rate of the raw required for projected SAR systems, the
signal data, include a prototype system compact size (1.3 ft.3), low power (0.3
developed by Unisys. This , system KW) and light weight (66 lbs) of high rate
implements a vector quantization fixed recorders permit multiple systems to be
code-book algorithm implementation. A flown on a single platform. It appears
prototype is capable of handling data that over the next decade these systems
rates of 100 Mbps and above. Preliminary will increase in performance even above
testing demonstrates that compression current specifications and that the
ratios of 4 to I can be achieved with recording technology will not be a
little degradation in the resulting image significant limitation in future
quality, however, this was done only for spaceborne SAR systems.
airborne X-band SAR data. More work
remains to prove that this is a viable OPTICAL IMAGE SIGNAL PROCESSING
technique for spaceborne applications.
An analysis of the current and future
Following the buffering/coding stage to technology as it relates to the signal
reduce the peak data rate (typically by a processing of SAR data must consider both
factor of 4) the data stream passes into a optical and digital computing systems. In
data steering network. This network addition to the historic limitations of
routes the signal to any of three processors, such as weight,
destinations: 1) On-board high rate power and environmental limitations
recorder system; 2) SAR signal processor (radiation shielding, outgassing),
for image formation; or 3) Downlink consideration must also be given to system
transmitter system for relay to a ground control, image calibration, processor
station. flexibility and reliability (graceful
degradation) issues.

480

The early versions of ground-based SAR dynamic range of only 30dB can be achieved
processors, some of which are still in which is too small for high precision,
operation today, were analog (optical) calibrated SAR image processing.
systems utilizing a laser light source
with a series of lenses to perform the Although tremendous strides have been made
two-dimensional convolution processing. in electro-optical computing over the last
Typically, film is used for both the input. decade, major technological advances are
and output media. optical processing still required before these systems can
systems feature high throughput (real- compete with digital computers in terms of
time) relative to a digital processor, but image quality and performance. However,
are constrained in terms of the dynamic the potential advantages of optical
range of the film, limitations of the systems for on-board SAR processing are
lenses for high resolution imaging and tremendous. Specifically, the high speed
swath width limitations. ADC and buffers shown in Figure 1 can be
eliminated from the SAR system and the
Recent technology advances in the,field of signal can be routed directly to the
electro-optics have resulted in optical processor which in turn generates
improvements in state-of-the-art optical the image nearly instantaneously. This
computing. A functional block diagram of image is captured on a digital CCD and
an optical SAR processor is shown in routed to the downlink processor for
Figure 2. Specific advances include; transmission to a ground receiver before
available semiconductor light sources the satellite has. passed from reception
(LEDs, laser diodes) that are more range.
reliable than previous light sources,
acousto-optic devices (AOD) that have Considering the size of the engineering
improved for input spatial. , light community working in optical computing
modulation use and semiconductor detector today, it is reasonable to assume that a
arrays (charge coupled devices -.. CCDs) high performance spaceborne SAR processor
that replace the film (Reference 2). With will become technically feasible within
todays technology it is feasible to the next decade.
construct a real-time, compact, power
efficient optical computer to perform on- DIGITAL IMAGE SIGNAL PROCESSING
board processing of SAR imagery. However,
improvements must still be made in the The tradeoff in digital versus optical
area of performance, flexibility and processing is typically performance for
reliability. throughput. Digital signal processing is
not theoretically limited in terms of the
The performance in terms of the image dynamic range, swathwidth, or resolution
pixel resolution will be limited by that the processor can achieve but rather
factors such as aberrations in the optics, by the quality of input signal data and
mechanical/electronic stability and light the ancillary data such as platform
source coherence. Resolutions on the ephemeris,,attitude and sensor calibration
order of SEASAT SAR (25m) are achievable data. It is fair to say that the
with todays technology, but an order of processing algorithms/techniques are
magnitude better resolution is not sufficiently mature today such that they
currently feasible. Additional problems contribute little or no degradation to the
exist with the light source. Since the resultant image quality. The major issue
duration of the light source pulse must be in implementing an on-board digital
shorter than the inverse BW df.the signal processor ,is in- achieving the necessary
to avoid range smearing, extremely short computational power within the size
pulses (in the 10 nanosecond range) must weight and power limitations of the
be used. This presents a problem with platform.
coherence and gain transients that
effectively degrade the resolution. This Recent gains in semiconductor technology
problem could be overcome if a pulsed gas using VHSIC Gallium Arsenide (GaAs)
laser were used, but this is only circuitry and advanced Complementary Metal
currently feasible for a ground processing Oxide Silicone (CMOS) devices produced
system. The other major area of with Silicon On-Insulator (AOI)
technology improvement is in the CCD array architecture, bode well for major
technology. Currently CCD chips are performance gains in the near future.
limited in width to 1000 elements. This Looming on the horizon is the potential of
limitation can be overcome by interfacing superconductivity revolutionizing the
a number of chips to realize a wide range microelectronics industry. Superconductor
swath, however, the time bandwidth product research is on-going to greatly increase
of the AOD may then become a constraint on the cycle - time through improved
the swath width. Another severe junctions.
limitation of existing CCD arrays is the
dynamic range. Without special cooling to
reduce the dark current, a reliable

481

The state-of-the-art in terms.of currently the requirements on the digital downlink
available or near operational technology and the ground data handling systems.
primarily utilizes CMOs circuitry which
exhibits excellent speed/power DIGITAL DOWNLINK
characteristics (Reference 3). For
spaceborne systems, a 64K radiation With recent technology advancements in
hardened RAM and a 16 bit radiation digital telecommunication systems, wide
hardened microprocessor (8OC86RH) are to bandwidth digital links have become
be used in the Mars Observer satellite. commonplace. The NASA Tracking and Data
One of the most advanced processor Relay Satellite (TDRS) has two 150 Mbps
candidates for space use is the IBM Common communication channels for a total
Signal Processor' (CSP) developed for the capacity of 300 Mbps. To increase the
Advanced Tactical Fighter (ATF) program link capacity or equivalently the system
and used in conjunction with the bandwidth requires an increase in power to
Westinghouse Ultra Reliable Radar (URR) offset the commensurate increase in noise.
(Reference 4). The architecture consists
of a number of processing elements each Considering the potential of on-board data
rated at 125 MFLOPS contained in a module processing and compression the next
with 32 Mbytes memory. The system can be generation of direct downlink systems and
configured with a number of modules each data relay satellites should be able to
interfacing into a common data network handle data rates on the order of 0. 5 to
achieving a maximum p@erf ormance of 1, 8 1.0 Gbits/sec, which even after error
GFLOPS. The CSP uses CMOs technology, correction coding is equivalent to
does not require special cooling and is instantaneous rates of 8 Gbits/sec coming
currently constructed with 6 x 9 in. from the SAR ADC (Reference 6). As
boards (2 ft.3 volumey. Considering that previously reviewed, the recorder
this system is within a factor of 4 of technology to capture such a data stream
what is required for SEASAT real-time is currently available. A number of
processing, it is highly probable that the recorders, each with a 240 Mbps capacity
technology will soon exist for spaceborne could be interfaced in parallel to achieve
high resolution real-time SAR processing. the desired capacity.

An,alternative approach for the SAR signal GROUND DATA SYSTEM
processor architecture is to. develop a
custom chip taking advantage of the highly It is reasonable to conclude that existing
repetitive nature of the SAR pr ocessing IC technology and data networks (using
algorithm. The range processing could be fiber optics) make the question of
performed with an.analog device (such as a feasibility of real-time ground processing
SAW filter) before digitization as shown mute. The pertinent question is not the
in Figure 3. Following the ADC, a azimuth feasibility of such a system, but rather
correlator chip could be designed that what is the optimal architecture and what
does not require a corner turn of the data systems exist or are under development? A
from a range file to an azimuth file. number of organizations have already
The most efficient algorithm for this type built custom signal processors with GFLOP
of implementation is the time-domain computation capabilities. Specifically,
approach where the azimuth correlation is the Advanced Digital SAR Processor (ADSP)
a convolution,operation (Reference 5). A designed and built by NASA/JPL to support
custom chip is required that performs the Magellan Venus Radar Mapper and the
resampling for. the range migration Shuttle Imaging Radar (SIR-C) has been
correction fo-llowed by a complex demonstrated to be capable of 6 GFLOPS.
multiplier for the reference function
weighting and an accumulator (complex An architecture with dedicated pipeline
adder with memory for one range line) . processing modules, custom designed to
This chip would be replicated for each perform a specific function will reduce
element in the synthetic aperture system reliability, since a failure will
followed by a multiplexer to recombine the typically halt all processing. This
data. This approach does not appear to architecture is used in the ADSP. It is
represent any significant technology capable of extremely high computational
drivers, although this chip has never been rates when the pipeline is full, but
fabricated. relies on every module to be functioning
for the system to be operational (i.e., no
The signal processor should also be graceful degradation).
followed by a compression subsystem which
performs spatial compression of the SAR The IBM CSP represents an alternative
imagery. Studies have shown that approach using multiple identical boards
compression ratios of 20:1 are achievable for each type of processing (e.g., FFT,
with little degradation in the image memory, complex interpolators) and routes
quality. This would significantly reduce the data using a high speed switch to each

482

board as required in the processing A more achievable specification for the
algorithm. One drawback to this calibration within the next ten years is
architecture is the extremely high data 0.2 to 0.3 dB. This will require internal
rates at the data transfer node. However, calibration signals to characterize the
this system does feature graceful amplitude and phase characteristics of the
degradation at the computational board system (including the antenna) over the
level. entire signal bandwidth. This
information must then be input to the
A third potential architecture is the processor to modify the appropriate
concurrent processing system such as the processing parameters in near-real-time.
Massively Parallel Processor (MPP)
developed by Goodyear for the NASA Goddard High precision GPS and attitude
Space Flight Center (GSFC) or the determination are also required for
Hypercube developed by Caltech/JPL. These estimation of the Doppler parameters and
systems consist of a large number of the antenna boresight. All of these
identical microprocessors each with local elements are technically feasible but have
memory. The microprocessors are yet to be incorporated into an
interconnected to each other with operational system.
different configurations for each
application. For example, the MPP is a REFERENCES
planar array with each microprocessor
communicating with its nearest neighbor. 1. "Digital Helical Scan Cassette
The Hypercube permits multiple Recorder", Schlumberger Fairchild
interconnection schemes such that any Weston Systems, Inc., Recorder
processor can communicate with any other. Division, Sarasota, Florida, Product
As the microprocessor technology improves, Announcement, 3-88.
this type of architecture becomes more
desirable due to its high redundancy, 2. Psaltis, D., K. Wagner, "Real-Time
system reliability and flexibility. optical Synthetic Aperture Radar (SAR)
Processor", optical Engineering,
SU14MARY 9/10-82.

The current state-of-the-art and expected 3. Bursky, D., "Tackle Real-Time DSP with
future developments in data handling and CMOS Chip Set", Electronic Design,
on-board signal processing were presented 3-31-88.
at a functional module level. The
analysis indicates that the primary 4. "IBM Will Deliver Initial common
technology driver is in the spaceborne Signal Processor to Air Force",
image signal processor. optical, digital Aviation Week and Space Technology,
or combination hybrid designs can be 7-4-88.
considered for space-based
implementations. 5. Assal, H., J. Vesecky, "Spaceborne SAR
Azimuth Processor VLSI Implementation
If an on-board processor is implemented, Assessment", Stanford University,
this system could be followed by a data Scientific Report D909-86-1, 12-86.
compression system that would reduce both
the data rate and volume by a factor of at 6. Pritchard, W.', "Satellite
least 20. This would alleviate any issues Communications - An overview of the
regarding the capacity of available high Problems and,Progress", Proc. of the
rate recorders and downlink systems. IEEE, Vol 65, no. 3, pp. 294-307.

The only area of lingering concern is the
control of the processor and the data
calibration. The requirement that the SAR
must be capable of discerning small
changes in the radar cross section is an
extremely difficult specification to meet
in the laboratory, not to mention in
space. A rule of thumb is that the
calibration accuracy is inversely
proportional to the resolution. Thus, the
higher the resolution, the poorer the
4formance.
calibration per Typical
specifications for planned spaceborne
systems (e.g., SIR-C) are on the order of
0.5 to 1.0 dB for the relative calibration
(stability) of the system. It is
difficult to make a measurement to better
than 0.1 dB.

483

F)GURE I FUNCTIONAL BLOCK DIAGRAM OF SAR
MULTI-CHANNEL SPACEBORNE RADAR IlF
SYSTEM WITH ONBOARD PROCESSOR OUTPUT

SAW X LPF AOC
FILTER v
0
A AC AC AC
=BUFFRl N
ER/ T
RCVR Coo FE A A 1 2
r
CODING AZ REF FUNC
s @GENERATOR
T ----

RCVR BUFFER/ E
CODING I RANGE
N. MIGRA71ON INTERPOLAT ON MULTI-PLEXER
N
RCVR BUFFER/
ADC T
CODING W OUTPUT IMAGE
0 com
R REF FUNC MUL IPLY
K B

+
A ----------- ON-BOARD
T
A -4 ---------0PROCESSOR
A
C;
T c

OAT
COMPR
OUTPUT
N
Z FIG 3 ARCHITECTURE FOR:
T
w GH RATE A) HYBRID ANALOG/ DIGITAL TIME DOMAIN SAR CORRELATOR
0
R B) AZIMUTH CORRELATOR AC CHIP
K

DOWNLINK TRANSMITTER

SAR VIDEO SIGNAL SHIFTED TO CENTER OF ADD BAND

S (t)

LIGHT COLLIMATING FOCUSING ACOUSTO
SOURCE P. LENSE -------- 0 LENSE OPTICAL@
DEVICE

MATCHED CCD OUTPUT
FILTERING MASK OUTPUT IMAGE
(3 LENSES)
@
'TPUT PL,
SAW
FILTER

FIG 2 FUNCTIONAL BLOCK DIAGRAM OF OPTICAL SAR PROCESSOR

484

A COMPUTER PACKAGE FOR THE PARAMETER OPTIMIZATION OF GROUNDWAVE RADAR

D.S. Bryant, A.M. Ponsford and S.K. Srivastava

NORDCO Limited

ABSTRACT their differing Doppler shifts.

The increasing popularity and diverse
applications of groundwave radars operating A typical example of a groundwave radar is
in the HF band has necessitated the shown in Figure 1. An omni-directional, or
wide beam, antenna is used on transmission,
development of a generalized computer
simulation package. to "floodlight" the desired sector.
Directional information of the reflected
This paper describes such a software package target signal is then obtained, on reception,
and shows how it can be used at the design using a wide aperture, phased array, antenna.
stage to optimize the many radar parameters Effective target detection, in the presence
for a particular application. These of sea clutter returns, involves the coherent,
applications may vary from the remote sensing integration of the echo over a period of time
of the sea state to an early warning system which is dependent upon the type of target
for the detection of low flying aircrafts. being interrogated. For a ship target, with
a Doppler frequency within the Doppler
spectrum of the sea, a coherent integration
period in the order of hundreds of seconds is
1. REVIEW OF GROUNDWAVE RADAR required. For the detection of a high speed,
low flying, aircraft, whose Doppler frequency
Over-the-horizon groundwave radars utilize lies well outside the Doppler spectrum of the
the low attenuation rates of vertically sea, the coherent integration time can. be
polarized HF electromagnetic waves when reduced to the order of tens of seconds. The
propagated over the conductive sea surface. remote sensing of the sea state however,
These waves will be scattered from objects requires long Dwell times in the order of
having a substantial vertical dimension and tens of minutes.
the resulting radar return will have a
precise Doppler frequency shift that is Typically the radar will be required to cover
proportional to the radial velocity of the a specified angular sector in a given period
target. In addition, the sea surface acts as of time. A single narrow beam scanned in
a spatially distributed collection of azimuth, as in microwave radar, is therefore
scatters for which the radar return will non-ideal. This is because of the need to
typically be greater than that of the targe t. maintain a sufficient search rate to ensure
Discrimination between the target echoes and the early detection of new targets and the
the sea echo is achieved on the basis of accurate tracking of known targets whilat

CH2585-8/88/0000- 485 $1 @1988 IEEE

still allowing a sufficient dwell time to these Bragg returns that information
eliminate the clutter return. The HF radar concerning the ocean currents and surface
must therefore use multiple beam synthesis wind direction can be obtained. The
techniques to cover the desired surveillance backscatter from a target will also have a
area. Doppler frequency shift which is proportional

to the r adial component of the target
operation in the spectrally congested HF band velocity. Doppler frequency shift for a
dictates that the radar is compatible with variety of target radial velocities, as a
the narrow band communication user. The function of radar carrier frequency, are
radar must be designed to minimize plotted in Figure 2.
interference both t o, and from, other users.
To control spectral spread the envelope of The ability of a groundwave radar to track
the transmitted pulse is shaped. since the ships using multiple beam synthesis
bandwidth of transmission is regulated at HF, techniques, has been demonstrat -ed by Ponsford
the range resolution will be in the Iorder of et al; 1987.
several km's.

To maximize the operational range of the 2. GENERAL OVERVIEW OF MODEL
radar it is necessary that the receiver has
atmospheric noise level sensitivity. To
avoid losses associated with beamforming at The necessity to optimize the design of an HF
the radar carrier frequency and to maximise groundwave radar has led to the development
the sensitivity of the radar, each array of a generalized computer simulation package.
element must be allocated its own receiver. Barrick (1972) has developed models for the
The adoption of this approach has the first and second orders of the backscattered
additional advantage of increased flexibility radar cross section of the ocean surface.
offered by post-detection beam synthesis. Using the concept of generalized functions,
This in turn requires that the receiving Walsh and Srivastava (Walsh; 1980,
channels must now be matched in both gain and Srivastava; 1984, Walsh et al; 1986)
phase characteristics to a degree of accuracy developed a similar model for the ocean
consistent with the specified b eam patterns clutter. However, their second order model
(Ponsford et al; 1988). consists of three parts. The first part of
second order represents the case where both
The radar distinguishes between targets and first order and second order scatterings
clutter returns by spectral analysis of the occur on the angular section (defined by the
backscattered signal. The clutter return receiving system beamwidth) of the circular
from the sea surface has a characteristic annulus (defined by the range resolution).
Doppler spectrum. The dominant contribution The second part of second order represents
is produced by scatter from surface water the case where the first scattering occurs at
waves with a wavelength exactly one half of the transmitting point and the second
the radar signal wavelength and which are scattering falls on the angular section. The
moving either directly towards, or directly third part of second order represents the
away from, the observing radar. This case where the two scatterings occur off the
condition is referred to as Bragg resonant angular section. A geometrical illustration
scattering by analogy with X-ray of the three second order parts for
crystallography. It is from the analysis of omni-directional transmission and reception

4.86

is given in Figure 3. antenna coupling losses, etc.)
- pulse length
A cross section model for the target has also - pulse repetition frequency
been developed by Walsh (Walsh et al; 1986), - antenna array height above sea
which makes extensive use of Fourier level
techniques in order to solve Maxwell's - system noise level
equations for different physical targets.
ii) Environmental Parameters
Models have also been implemented to
describe: the radio noise level, the - orientation of sea with respect
propagation loss 2 and the wave height to radar
spectrum. The propagation loss model - surface current, speed and
employed has been adapted from Walsh et al; direction
1986, and Rotherham; 1981. - swell magnitude, period and
direction

3. COMPUTER MODEL

The software structure consists of a series iii) Target Parameters

of subroutine modules linked together by a
main controlling program. These subroutines - target speed and direction
are called where needed via a menu driven - target cross section

screen.

b) System Characteristics: This module
The software package is broken into the implements the antenna gain in the
following subroutine packages: appropriate form.

a) input Parameters c) First Order and Second Order Sea Clutter:
b) System Characteristics This module utilizes a model designed to.
c) First and Second Order Sea Clutter calculate. first and second order sea
d) CCIR Radio Noise (atmospheric and man clutter on a Doppler basis.
made) Level Calculation
e) Propagation Loss Calculation d) Calculation of CCIR Radio Noise Level
f) Generation of Wave Height Spectr um This. module generates external noise
g) Output Options levels (ie., man7made, atmospheric and
galactic) utilizing standard CCIR
A brief description of each of the above algorithms and data (CCIR, 1982).

follows:
e) Propagation Loss: A standard spherical
a) Input Parameters earth propagation model is used to

account for diffraction around the
i) System Parameters earth' s surface. A modified surface
- radar carrier frequency impedance is used to account for the
- peak transmitter power effects of ocean waves on the transmission
- antenna (array) gain loss.
- horizontal beamwidth (3 db)
- system losses (cable losses,

487

REFERENCES
f) Generation of Wave Height Spectrum: A wave
height spectrum is generated based on the Barrick, D.E., (1972) REMOTE SENSING OF THE
sea state characteristics provided. The SEA STATE BY RADAR, in Remote sensing of
actual model utilized is one developed by the Troposphere edited by V.E. Deit, US
Pierson and Moskowitz (1964), with Government Printing Office.
cardoid directional distribution. C.C.I.R. (International Radio Consultative
Committee) (1982), WORLD DISTRIBUTION AND
g) output Options- The outputs provided are CHARACTERISTICS OF ATMOSPHERIC RADIO
listed below and are illustrated in
NOISE, Report # 322, Geneva, 78 pp.
Figure 4. C.C.I.R. (International Radio Consultative
i) A Doppler spectra of the Committee) (1982), MAN-MADE RADIO NOISE,
received signal Report # 258-4, Geneva, 177-183 pp.
ii) A Doppler spectra of the Pierson, W. J. and Moskowitz, L. (1964). A
PROPOSED SPECTRAL FORM FOR FULLY
normalized cross section of
DEVELOPED WIND SEAS BASED ON THE
each component of the sea SIMILARITY. THEORY OF S.A. KITAIGORODSKII,
clutter and the total sea J. Geophys. Res., 69, 5181-5190 pp.
clutter, and Ponsford, A.M., and Bagwell, D.J.,, (1988),
iii) A target visibility profile RECEIVER DESIGN FOR HF GROUNDWAVE RADAR,
as a function of range. IEE Int. Conf. 'HF Radio Systems and
Techniques", Conf. Pub. 284, 254-258 pp.
Ponsford, A.M., Bagwell, D.T., Money, D.G.,
and Gledhill, M.H., (1987), PROGRESS IN
The computer simulation package has been
extensively tested against published data SHIP TRACKING BY HIP GROUNDWAVE RADAR, IEE
(Wyatt et al., 1985). The two .separate Int. Conf. 'Radar 87', Conf. Pub.
Rotherham, S., (1981), GROUNDWAVE PROPAGATION
approaches by Barrick, and by Walsh and Part 2, IEE Proc., Vol. 128, No. 5 Oct.
Srivastava give consistent agreement with the
Srivastava, S.K. (1984), SCATTERING OF
measured data.
HIGH-FREOUENCY ELECTROMAGNETIC WAVES FROM

AN OCEAN SURFACE: AN ALTERNATIVE
NORDCO is in the process of developing its APPROACH INCORPORATING A DIPOLE SOURCE,
own groundwave radar facility. This facility
Ph.D. Thesis, Memorial University of
will be used to further validate the Newfoundland, Canada, available from
simulation package in order that optimum Canadian Theses on Microform Service,
operational performance can be achieved for National Library of Canada, 395
given target and environmental conditions. Wellington, Street, Ottawa, Ontario KIA
ON4, Canada, 305 pp.
ACKNOWLEDGEMENTS Srivastava, S.K. and J. Wal sh, (1985), AN
ANALYSIS OF THE SECOND ORDER DOPPLER
The authors would like to thank..NORDCO Ltd., RETURN FROM THE OCEAN SURFACE, IEEE J.
for their permission to publish this paper. Oceanic Eng., OE-10,.443-445 pp.
The work has been supported by DND Canada and Walsh, J., (1980), ON THE THEORY OF
the Atlantic Accord Development Fund. ELECTROMAGENTIC PROPAGATION ACROSS A
ROUGH SURFACE AND CALCULATIONS IN THE VHF

488

REGION, Technical Report prepared for the
Defence Reseach Establishment Atlantic
(DREA), Department of Supply and
Services, Government of Canada, available
from ocean Engineering information Centre
(OEIC), Report # N00232, Memorial
University of Newfoundland, St. John's,
Newfoundland, Canada, AlB 3X5, 195 pp. 9
Walsh, J., B.J. Dawe and S.K. Srivastava
(1986), ANALYTIC MODE DEVELOPMENT FOR TR 0
01PIE. WF' '.W -IW
STUDY OF GROUND WAVE RADAR AS REMOTE

SENSOR IN AN OCEAN ENVIRONMENT, Prepared
for the Department of National Defence, ........
6 0 11 1. a 1. 20 22 24 2. 2. 30
Ottawa, Ontario, Canada. RADAR CARRIER FREQUENCY IMHA
Wyatt, L.R., Burrows, G.D., and Moorhead, FIGURE 2 Target Doppler as a function of radar carrier freciusncy
M.D., (1985), AN ASSESSMENT OF A FMICW
GROUND WAVE RADAR FOR OCEAN WAVE STUDIES.
Int. J. Remote Sensing, 1985, Vol. 5 No.
1, 275-282 pp.

T@MT first and second
RQW CAN scatter

first
scatter

Fb
mpatition 2P
P x
0
'second first
scatter
sc itter
transmitting and
receiving point 260
P.I. idlh

second
scatter
*A"C A.,

(PO-A P P @s (PO + AP)

Fig. 3 Three parts of second-order backscatter
from a circular annulus and off the
annulus for omidirectional transmission
CEIIANJ and reception. (Po= distance of
DMIN AWAY annulus;.2 &9@= radial width of annulus;
F's groundwave attenuation functions
Figurel.l. GROUNDMVE PULSE DOPPLER RADAR with modified surface impedances.
-*' "'ru "",
scatter

,489

-0617

RECEIVED POWER SPECTRUM CdIB/M.> NORMAL37ED SPECTRAL CROSS SEC71ON (dB/M.@

RECEIVED POWER cdowi NORMALIZED SPECTRAL CROSS SECTION (dB/H.)

NORMALIZED SPECTRAL CROSS SECTION (dBIM.)

m I EA
-4

I r
-4 A 'G !' 0@ I
rq p 1-4
rn
ta

m
M

RESULTS OF EXACT INVESTIGATIONS ABOUT THE CHARACTERISTICS OF THE EXTREMELY FAST AND
ACCURATELY, MEASURING KIEL MULTISONDE AND REPRESENTATIONS ABOUT ITS NEWEST PERFORMANCE

Werner Kroebel

Institute for Applied Physics, University of D-2300 Kiel, FRG

A B S T R A C T totally new generation-of C-T-D sondes.
Prearrangings to further developments in the The author already came to the same conclusion a
ocean research are actual and challenge the de- couple of years ago, but other than the considera-
velopment of a total new generation of C-T-D tions of the "WOCE" committee, that generally for
sondes. Therefore, many efforts have been made example temperature measurements in situ in the
during the last 2 years. Anyway, this problem is Sea need probe devices with definite smaller time
still of high actuality. Now, the author has constants for all C-T-D sensors than those used
found a new and special solution for this problem. before. Since the commonly used temperature sensors
It concerns new devices of the sensors as well as with their relatively big time constant caused, in
of the needed electronic circuits. At present terms of suddenly appearing changings in the tempe-
this new C-T-D sondes have been exactly tested. rature profiles, faults by great delay of time as
Theiresults are: All C, T and D sensors have a well as suppressions of the fine structures in the
very short time constant. For example T .5 msec, graphic representations. These faults cannot be
the accuracy .3 mK, the resolution .12 mK, compensated by calculation. Furthermore in these
the electric circuits allow to repeat all cases, the measured temperature data range does
measurings in periods of 2 msec. The spatial not correspond,to the measureddata taken at the
resolution for C is .5 cm, the precision .001, same time of the electric conductivity and the
the resolution .0005. At present, the pressure pressure to the a 'ctual same depths. Besides, the
sensor can reach a precision of about �1,5 m calculation of deviative units of measurands as
between 0-600 bar, a resolution of �30 cm. of salt content and density leads to the well
known spikes in the graphic depiction of profiles.
Moreover, real existing changings in the kind of
jumps in thetemperature profiles notify important
I N T R 0 D U C T 1 0 N marks for bounderies between superimposed different
layers of water. Thus they are re.levant and give
About 4 to 5 years ago it seemed to be about time characteristic informations for the interpretations
to undertake an universal determination of hydro- ofprofile courses under the different parameters.
graphic data of the Sea. This lead to developing Since, on the other hand, such fine structures of
and planning an international program of measuring, those transition regions from one to the other
which became popular about 1984 under the name of layers can have extensions between cm, dm and m.
"WOCE". Based on intense thoughts about the extent In order to find out this transitions with its
of this program, it was deducible that it was fine structures all measurable parameters must
realistic and possible only, if beforehand essen- have time constants ranging between msec. and less
tial progress, on the used measuring instrument referring to the high speed of lowerings of the
was achieved. This concerned especially the C-T-D C-T-D sonde. A further reason for the necessity of
sondes. Therefore, they had to be developed in fast measurings is the requirement to quasi synop-
order to meet the necessary standards. tic measurements.

This requirement referred to the-feasibility of BRIEF REPORT ABOUT FEASIBILITIES OF HIGH SPEED
fast measurements in terms of times of only a few C-T-D MEASURINGS IN SITU
milliseconds, on increasing the accuracy in
measurement up to 16 bit per measuring range for Based on these considerations the author took [1-1
the temperature and the electric conductivity efforts in the seventies to develop first of all
parameter plus a new pressure sensor for the depth a high speed measuring temperature probe. 1978 he
destination with an accuracy of � I meter for a succeeded in realizing such a thermic sensor with
range of 0 up to 600 bar as well as essentially time constants in the range of .5 msec. Some of
increased speed of lowering of the C-T-D sondes those high speed measuring temperature sondes were
through the hydrosphere. Since all available first applied in a very thorough hydrographical
C-T-D sondes did not meet these standards, all surveying of the Lake Constanz in July and
producers of these sondes definitely knew that the November of 1979 [2]. They came up with results
"WOCE" experiment required the development of a which depicted a relevant meaning for understanding

CH2585-8/8810000-, 491 $1 @1988 IEEE

the dynamic of the lake even for temperature values dissolving depth of .5 cm for measurements of the
within the range of mK. Later in 1982 with this electric conductivity in situ. The diameter of the
type of probe it was possible to depict by means cell tube was 16 mm, thus the seawater could flow
of an expedition with the German research vessel very quickly through the cell. Besides, the cell
Poseidon that in the northern part of the Atlantic was performed mechanically very stable.
Sea changings in temperature occur in a number of
ranges in depth within depth of more than 1 900 m In respect of measuring the depth there were indeed
in which the temperature only run up to about .1 K. great problems confronting the new generation of
(see fig. 1) DI the C-T-D sonde by the available multiple types
POSEIDON 86 DATE. 9.4.82 STATION. 329 PROFILE, 72 21.S.32 of pressure sensors on the market, because they
13 did not allow high measurings speeds and simul-
'4S as 11 taneous sufficient accuracy in measurement.
74
Besides, most of them were also not sufficient free
of drift and the best of these were too expensive.
Measuring of the drift of pressure sensors has
been made by Wearn. R. B.., L@rson.. N. G. The
result by these authors .'is to:,see-in Deep Sea Re-
search Vol* 29, NOIA 82 [3]. This situation gave
the author the suggestion to seek for a new and
T4 To genuine solution for measuring of a good fitted
pressure sensor concerning marine measurements
in situ. He found a new way about the pressure
t dependence of the electric resistance of a mercury
filament. About a first performance of such a probe
he gave a presentation on the "Ocean 8711 [5a].
Parallel to his efforts, independent of him, also
Neil Brown had himself turned towards to seek a
new solution about this problem. And also Neil
Brown presented his new interesting ideas and re-
flections on this problem with his first results
7------- 7-- on the "Ocean 87" [5b].Since that time the new
solution for a pressure gauge by the effect of
changings of an electric mercury filament re-
sistance has been carefully investigated under
the authors guidance by his student B. Steffens in
his thesis as well as forward developed by the
author himself. The result is now a pressure
Figure 1. Temperature profiles at a depth of sensor device, measuring high speed and with an
1492 m to 1562 m in the atlantik near the azores. accuracy better than required by the "WOCE"
T is a Rosemount thermometer which takes a long committee. Thus a new generation of a Kiel C-T-D
time to reach the final resolution value with a sonde, respectively multisonde by the author with
sampling rate of 10/sec. T T T are thermo- new genuine sensors and new performances are
meters designed by Kroebel'@hiN ak characterized existing. To this are added advanced as well as
by time constants of approximately .5-1 msec with decisive improved electronic circuits. To demon-
a sampling rate of 50/sec. The differences are real strate the properties of this C-T-D sonde it was
and are caused frcm horizontal distances of about necessary to investigate and to test the complete
15 cm between the thermometers T4 and T7 and T 91 instrument very carefully.
in us,ing a combination of many high speed measuring THE ARRANGEMENT FOR TESTING THE NEW GENERATION OF
temperature probes by this expedition it was THE KIEL C-T-D SONDE WITH AN ACCURACY ACCORDING TO
possible to prove that in the transition layer of THE STANDARDS AT PRESENT
such a jump always water turbulences of minimal
spatial extension in the range of some cm occur. For this work especially the author's co-;worker
K.-H. Mahrt has prepared..' a special arrangement.
The sampling rate for these temperature sensors, To this belonged a big pressure vessel for the
depending on the number of simultaneously measuring pressure range of 0 - 1000 bar. It has a volume
parameters, increased up to 50 and 100 per second. of about 80 litre and is sufficiently big enough
Based on this, it was possible to reach a dis- to enclose the complete C-T-D sonde. This pressure
solving of depth for the temperature up to about vessel is combined with a pressure balance. It was
10 cm in using the lowering velocity of the Kiel shortly calibrated by our bureau of standards PTB
C-T-D respectively multisonde device by the author (Physik techn. Bundesanstalt) on an accuracy-of
of that time (see fig. 1). 5@1e-10 Besides, we have for the temperature
calibrations and tests water triple point cells and
In the course of realization of high speed a gallium cell for the temperature fix.point of
measured parameters, the author succeeded in de- 29,77140C. In addition to that we have platin
veloping in the years 1980-1982, a new and rela- thermometers with an accuracy of 2 mK even shortly
tively long conductivity cell of about 10 cm using chequed by our bureau of standards. For chequing
measuring speeds of about 1 msec for a spatial the electric resistances we use an ASL Bridge and
a
T11 I
sf%

492

an inductive resistance box. And for measurings feasibilities of a linearisation of the bridge
of atmosperic pressure and comparisons we own a big circuits. He found several ways of which it could
tank of sea water of 3,6 ml. be done. For the new C-T-D sonde, the author has
been chosen an electric circuit arrangement which
For the calibration of the salinity respectively was first represented on the "Ocean 85" and de-
electric conductivity we use the autosal of guild- scribed in the proceedings.[71
line. For using the autosal the pressure vessel is
arranged in such a way that it is possible to The linearisation of the above explained bridge
take out water samplings under pressure, which were reached by splitting the measuring into three
can be measured with the autosal at atmospheric steps. With the first step the sine signal voltage
pressure. is led to the input on an A-D converter of about
12 bit 181. It gives a rough value of the signal
BRIEF INFORMATIONS ABOUT THE.DEVELOPED ELECTRIC, peak voltage in numeric numbers. This happens in
ELECfRONIC CIRCUITS FOR MEASURING C, T AND D a very short time interval near the signal peak
voltage in any half wave period of the sine. With
Since all three measurands C, T and D by electric the next step the numeric defined voltage is led
measurements cause changings of the values of to the input of a D-A converter at the present
electric resistances, the author has developed a time of 18 bit. On this way it gets an exact known
special A.C.resistance bridge. This bridge has value for the sine servo voltage on the output of
been first deviced by the author 1972 161. This the D-A converter. This voltage appears temporary
bridge works with three branches. (s. fig. 2) 171. directly after the time point if the numeric num-
Two of these 1, 2 are well known and general used ber is produced. This takes still place in the
in resistance bridges. Through this two branches first half wave period of the sine. The exactness
flows a constant sine current. This current pro- of the sine servo voltage of the D-A converter
duces a sine signal depending on the measured corresponds to 18 bit. Now the exact known value
value between the two branches..This sine voltage of the servo voltage in general is very less dif-
can be understood as an algebraic vector. The ferent against the signal voltage from the bridge.
third branch is connected with a constant cosine Therefore to know its exact value the d ' ifference
current. This current produces at a fixed re- of these two sine voltages is produced in an adder
sistance a constant cosine voltage. The addition and than after a suitable amplification directly
of the sine signal voltage and the cosine fixed measured respectively reproduced as a numeric
voltage results a new and with the signal voltage number about the above explained two phase bridge
in its phase changeable vector. By this addition in a range where the bridge should viorir linear til a re-
a phaseshift is produced. From this changeable solution of � I Isb for a range of 18 bit.
phaseshift can be deduced a changeable time inter- 7
val which is counted out with a very high puls In the present time this bridge works with an A.C.
frequency. By this procedure the described bridge voltage of 500 cycle. Therefore the value of the
transfers the measurands directly into a digit measured parameter will be received every two
number. Therefore this bridge is also an A-D milliseconds. However the measuring of the dif-
converter. ference voltage respectively its corresponding
i._ konslant und time interval needs only a measuring time of
gegeben durch
i. sin d about 100-150 psec.

C THE RESULTS OF TESTS OF THE LINEARISED TWO PHASE
1 2 3 A-C BRIDGE
R3 imm . !L,.Co s Wt Since the described and explained bridge circuits
R211 + OP, are an essential part for the quality of the new
instrument. Therefore these circuits must be in-
R, R4 vestigated and tested as well as the single sensors
themselves. This test has been performed with great
r,.R6
Rx under his guidance by working on his thesis. The
carefulness by the authors student R. Duch5teau
Rr Rx + imm
4 + @, -
R
1_@' tests were started by connecting a very stable
.[ R. electric resistance to the input of the bridge.
Figure 2. This alternating current bridge has By this arrangement at the output the value of the
three branches 1, 2 and 3. The branches 1 and measurand in kind of numeric number appears. There-
2 have the same source and a constant sine
current.flowsthrough them. R is the changeable fore this numeric number corresponds to the size
and R,gthe reference resistaAce. U is the out- of the input resistors. The values were written
put s nal. Through 3 flows a constafit cosine current. down by using a magnetic tape from which later
They.are vectorial added in OP U is than a com- the values were carried out. On this way the mag-
plex output signal 3' 1 netic tape contains all elementary measurings in
time distances of 2 msec from which the value can
be transferred to a graphic projection.
However the dependence of the output signal from
the input is non linear. The consequence is that To,test the circuits we made many thousands elemen-
the resolution is not the same for the whole range. tary measurings with a fixed inductive resistance
Therefore the author has thought about
op,
R,

6
@,R

493

of high quality of the input of the circuit. With T
this arrangement the deviations of all 2 msec Calibration of the linearised two phase AC - Bridge with
measured values were noted on a graphic in de- Rosemount Pt-oceanographic thermometer with Rosemount
pendence of the time in sec.in fig. 3 on the Pt - standard thermometer.
ordinate. Hereby the chosen range was.devided in
18 bit. As the fig. shows there were no deviations
til one's of 2 400 data greater than �1 bit of
18 bit. The structure of the distribution of the Deviations from quadratic regression
deviations depending on the timedet expect, that
the measurings were,still something disturbed from
outside. resolution precision
Stability of the linearised two phase AC-bridge 0.12 mK 0.3 mK
lsb A/mK I presently
of 18 0.4- X X claimed
0.2- accuracy
0 .10`X-X so 2 mK
Y 3@ t I -C
0.2 - 1@ 1; \iO X
-0.4-.
Figure 5. traceable to PTB

11A In fig. 5 is graphically represented the result of
Elementary measurements every 2ms of a constant data, simulated with a high
accuracy ratio transformer at room temperature ( 2400dato I a calibration of the linearised two phase bridge
with a Rosemount Fit oceanographic thermometer
Figure 3. Elementary measurementc,every 2 ms of against a Rosemount standard thermometer. The
a constant data, stmulated with a high accuracy measurings cover the whole range between 0' and
ratio transformer at room temperature-(2 400 data) 300C. This graphic represents only the deviations
versus a quadratic regression curve because in such
In fig. 4 a, b, c are drawn the deviations from a way the result is better to demonstrate. These
the linearities of the used and linearisized two measuring has been done in our big seawater con-
phase A-C bridge of the author. These measurings tainer with a content of about 3.6 m". As the
are referred to an arrangement by which the graphic shows the result is �.3 mK for the preci-
sensor resistance at the input of the bridge has sion and .12 mK for the resolution.
been substituted by variable resistances of our
inductive resistance box. The measurings in Differences 6weeks after repetition of calibration
fig. 4 a, b, c are done in a temperature con-
trolled container for fig. a at 100C, for b at
200C and for c at 300C. The result is that all
measurings stay in an interval of �2 bit of 18 bit. T
However it can be concluded that the values lie AymK_
within a band with deviations of �1 bit of 18 bit. 1mK
Besides, as measurings of the D-A converter have
shown this'band is the following on non linearities 7.5.88
by this converter.
N-fl- f- time-ty f the I i me a,,&, t-1 ph... AC-b,idg, 2,0
a' diff-It t"Pelpl-1, 10.11.d t, .2 -C
11@b
23 b
30*c I- r
I V
-2 - - - - - - - - - - - - - - - - - - - - - - - - - - - -
1,0

2- - - - - - - - - - - - - - - - - -
20*C .1 b
I 21.S.88
-2 1 T
-3-a02 0,4 6 0.8 10 0 - 5 10 f5 2' Z5 3,l) T
It.b .18 Calibration A. C
21 r
lopc@ a

-2- . . . . . . . . . . . . . . . . . .
-3-00,2 0,4 0,6 ps 10 Fig. 6a, b measured in a shrred 3,6m3tank
Figure 4. This figure shows a g raphic represen- In fig. 6 a, b two calibration curves are repre
tation of a calibration of the tested circuits sented with the same arrangement as for fig. 5 be
in a at 100C, in b,at 200C and in c,at 300C Hereby fig. 6 b is referred to, dates which have
been measured six weeks laterthanthe calibration
room temperature for the complete linearised dates in curve of fig. 6 a. The comparison of both
AC-bridge.
figures show a drift between the temperature

494

values of about 2 mK. However hereby is to attend However this cell has been further developed under
that shifts in the temperature notification of Pt the author guidance in connection with the thesis
sensors could happen also by a push against it. of the student E. Falkenhagen. The cell has eight
Therefore it must not be a real drift. annular electrodes 1-8. The constant alternating
current for measurings of the conductivity is led
A measuring in the same tank by a fixed temperature in the electrode 4 and flows,to the electrode 1. A
produced the graphic in fig. 7. It shows separate- flow of this current to the electrode 8 is blocked
ly the results of elementary measurings over a by a controled servo current which is flowing in
time interval of .4 sec with a sampling rate of the electrode from 7 to 8. On this kind the path
500/sec. In this graphic the measured values over which the conductivity is measured only lies
fluctuate to about �.5 mK. However this fluctua- between the electrodes 2 and 3. Its distance is
tion is caused very probably by the natural and about 5 mm and in general the diameter of the tube
real noise in such a big tank. 9 about 16 mm. This cell is combined to a compact
The linearised two-phase AC-bridge with a Rosemount sensor with thermic'sondes. This cell has been
Pt oceanographic thermometer calibrated with the arrangement as explained above.
That means the calibration is made with the
26-_1mK linearised two phase bridge. For the cell Wormley
water is used in a comparison with a autosal in-
21,1120- strument. The result is displayed in fig. 9 by a
temperature variation from DOG til 250C of the
1mK
Wormley water. The reached precision was �.001 in
a A salinity, this means in deviations of a square
21,1100-T regression line. For the resolution. could be found
@s .0005 in A S.
Elementary measurements every 2ms in a stirred 3,6m3tank Calibration of the linearised two phase AC-Bridge with Wormley water
and outosal salinometer
Figure 7. Deviations from quadratic regression
b s resolu hon precision
INVESTIGATIONS AND TESTS I PF THE CONDUCTIVITY CELL 0,004- 0,0005 -0,001 presently
claimed
accuracy
As above mentioned for measurings of the conducti- 0,002- tO,003
vity it was used a special cell which has been de- - 42.902 Z
0-
viced by the author and is described in several 0!6/ 0,7 0
publications 191. Therefore it should not be -0,002- X-9 -.0
necessary to repeat a description of this sensor
in details. From this reason for this paper it -0,W4-
may be sufficient to show the construction in
fig. 8 a, b. 9 Figure 9.
A MERCURY PRESSURE SENSOR FOR MARINE RESEARCH

1982 Wearn R. B and Larson
N. G. 141 have made an
investigation of the sensitivities and drift of
pressure sensors which at present are commonly used
W
in the marine research. For all scientist's who
need pressure sensors, the work of Wearn and Larson
is very instructive. By this work it can be con-
cluded that these are many pressure sensors in the
20 marine field, however it exist's no really suitable
20 sensor for purposes by which precision measurings
are reached as required. Therefore the author has,
tried to find out another fundamental idea to
-710 realize such a sensor. From this basic he thought
mm
MM 3 3, to use the pressure dependence of a liquid elec@-_
tric resistance. Then a liquid electric resistance
0 2, sensor cannot show any drift phenomenons. Therefore
in respect of the well known electric properties of
mercury, such a sensor could be a mercury filament.
In a first approach the dependence of a mercury
electric resistance on pressure R@jo by a variable
P and a constant temperature T i given by
0
R R (1+a - AP)
PoJ 0 PoJo P,To
Figure 8 a, b Conductivity cell with 8 annular respectively ARPJ
electrodes 1-8 in a quartz tube 9. Fig. 8 a demon- o I A P, OL
ttrdtes in a cross-section drawing and b-in per- R P, T wher3 a
specti ve depiction PoJo Pj0 0

495

is the pressure coefficient at the temerature T o' But our measurings prove that up to a sufficient
A diagram for this dependence is to be seen in approach these deviations can be corrected by a
fig. 10a curve (1) for an interval of 0 til 600 bar. square function. By this reason it is possible to
ARp R reach on the basic of the described mercury
R 2 6 OC
P. @4@ i RTO resistance pressure sensor a precision at least til
0:2 0,4 0,6 *C �1.5 m in the range of 0-600 bar. Because this work
could not be finished til this time a detailed
publication will follow very soon.
2
60- -60 Tegrature edibrahm of i Ng4nortz dt&k mktance
At
40- Abe mK
40 1621. "1 Dwidiam of fto Unftr mVmionlirw .80
P,-2,063SO10 -70
20- 20 -60
__3
0- - 0 so
_20- 160 2@O 3100 4'0 0 56 '600 P/b., width '0
-30

.10
-60- Regmsionstine _0
-3- 10
_80-

-30

-40
120] 0 . . . . S' 10 . . . . i's ia@ i5 30 r'C
Figure 10 a, b. Diagram for the dependences Fiftre 11. REFERENCES
AR P AR
__R__ 10-4 of P in curve (1) and for R_ T .' 10-4 (1) a)Kroebel, W. German Pat. Nr. 29.10957 (1979)
P0 T WEEE Journ. OC. Engeneering, Vol. OE-6, No.4,
0 118..124 (1981)
for T = 6'C in curve (2).and for T 0,60C in c)Kroebel, W. and v. Bosse, N. OCEANS V, IEEE
cu Irve (3). Publ. No. 82-CH 1827-5, Washington D.C., USA,
272..276 (1982)
However,@ a mercury filament shows also a dependence (2) Kroebel, W. Hydrologische Untersuchungen mit der
of its electric resistance with the temperature., Kieler Multisonde im Bodensee. Publ. by
This dependence is represented in ajirst approach Landesanstalt fOr Umweltschutz Baden-WOrttemberg,
in fig. 10b curve 2 and 3 for a temperature inter- Karlsruhe (1980)
val from DOC til 61C in curve 2, respectively (3) Kroebel, W. and v. Bosse, N. OCEANS '82, IEEE
0-0,6*C in curve 3. In the fig. 10a is recorded in Publ. No. 82-CH,1827-5, Figures 3-6, 273..274
the abszis the pressure P as well as in fig. iOb (4).Wearn, R.B., Larson, N.G. Measurements of sensi-
the temperature T versus the ordinate ARP 4 tivities and.drift of Digiquartz pressure sensor,
R 10 Deep Sea Research Vol. 29, No IA, 111-134 (1982)
,IRT P0 (5) Kroebel, W. OCEANS '87, IEEE Publ. No. 87-CH2498-4
respectively -4 Halifax, N.S., Canada, Vol.1, 331..334 (1987)/
R 10 Brown, N.L., OCEANS '87, IEEE Publ. No.
T 0 87-CH 2498-4, Halifax, N.S. Vol.1, 335..340 (1987)
The magnitude of the ordinates is the same for both (6)'a) Kroebel, W. German Pat. Nr P220059 89 (1972)
abszisses. Therefore the diagrams allow to compare b) Kroebel, W. A frequency independent Two-Phases
the influence of the temperature on a pressure AC-bridge for precision in situ Measurements
measuring. This comparison shows that it is impor- of Physical Parameters in Sea Water,
tant to try to suppress the temperature influence Interocean 1973, KongreBberichtswerk,
by a suitable pressure sensor device. A first de- DOsseldorf,817..822 (1973)
vice which fulfills this conditions leads to the (7) Kroebel, W. OCEANS '85, IEEE Publ. No. 85-CH-2250-9,
result of a deviation of about �1,5 m. For this re- San Diego, USA, 610..615 (1985)
Isult has been arranged in a rugged block in one (8) Kroebel, W.A.,new Generation of CTD-Probe for
borhole a pressure protected mercury filament and in situ Measurements with high speed, Accuracy and
in another borhole of the same block near by a sensitivity, Part 2, Marine Technology Vol. 18,
pressure unprotected similar mercury filament. The 149..154 (1987)
changeable temperature interval by a measuring (9) a)Kroebel, W. German Patent No. 3238956 (1982)
between 0-600 bar has been about 2 K. b)see (8)

However, the deviations concerning the coefficient
T,P0 from the linearity are shown in fig. 11.

496

FIELD PROVEN HIGH SPEED MICRO OPTICAL DENSITY PROFILER SAMPLING
1000 TIMES PER SECOND WITH 10-6 PRECISION

Karl-Heinz Mahrt and Christoph Waldmann

Institute of Applied Physics, University of Kiel
Olshausenstrasse 40, 2300 Kiel 1
Fed.Rep.of Germany

ABSTRACT vinced that the addition of high recision refrac-
tive,index determinations.to simuTtaneously mea-
We developed a miniature fibre optical point sea- sured "first class" CTD- data should give new in-
refractometer with industrial prototype status. It sights into the dynamics of structures and distri-
was designed to be used in high speed profiling butions of water bodies. We really were led to
applications, e.g. for quick survey of the density think about a whole refractive index concept rather
field of the ocean by directly measuring the index than building just another new instrument./ I
of refraction with extremely high precision of 10-6 / 2 /./ 12 /,/ 13 /.
as it is required for fine description of the den- At the M.I.T.- workshop on "Profilers of the Futu-
sity distribution. re" (WOCE- meeting) in Boston,August 1984,/ 3 / we
reported under the topic "Is there anything new
Thanks to the kind invitation by Albert Bradley under the sun" on our-experience with refractive
from Woods Hole Oceanographic Institution and,to index measurements. It was there where we learnt
his substantial and invaluable help,we were able about details of A.Bradley's research work,on his
to mount our sensor on his newly developed most high performance deep sea instrument carrier "Fly-
progressive high speed free falling deep sea in- ing Fish" / 4 /, following an inspiring and very
strument carrier "Flying Fish" who at the same, courageous concept of high speed surveying of the
time was furnished with the most recent Neil Brown oceans as a large step forward to tackle the prob-
Mark V CTD electronics. lem of what Walter Munk called the "terrible under-
sampling of the oceans" at the occasion of his talk
Field data evaluation studies showed at the very in the opening plenary session of last year's
first glance: Index of refraction profiles are an "Oceans'87" conference in Halifax,.N.S.,Canada./ 5
extremely good representation of the actual densi- We found it to be of mutual interest .and benefit to
ty distribution. Evaluation of all measuring data come together,' -and so we did some "merge in"- re-
turns out that there are significant correlation search work with "Refractive Index",and "Flying
anomalies between CTD- derived density data (using Fish" that culm,inated in 1987/88. The crucial joint
the equation of state EOS-80),and densities as.de-
ternined from index of refraction measurements.
These differences may well serve as a new parameter LSD PMF
for a much more detailed description of water
bodies. TST

1. INTRODUCTION

6PD
On June 26, 1984,from board of the research catama- rb
ran M/S"Haithabu" our sea going experimental labor- db
atory, instrument delivered in situ in the Baltic
Sea the first continuously measured absolute index
of refraction profile, clearly demonstrating the
good feasability of "first class" optical density E
measurements with unparalleled small measuring vol-
ume, very high measuring speed, and very high pre- PR
6
cision of 10 . This most important experience to-
gether with the results of our intensive recherche
rb
and study of all the old and recent related litera- 0 11
ture we could get, made us conclude that it would IRB
- Arb
be a regrettable, limitting misunderstanding to lb
consider refractive index measurements just to be lb
another substitute for e.g. conductivity measure-
ments to finally arrive at "conventional practical" Fig.1: Measuring principle of the extrinsic fibre
salinity and density data. Instead, we became con- optical point sea- refractometer

CH258S-8/88/0000- 497 $1 @1988 IEEE

field experiment took place on WHOI R/V "Oceanus" 4. INDEX REFERENCE STANDARD
195 cruise, Oct./Nov. 1987 / 6 /, when the refrac-
tive index sensor proved to be a high performance SCHOTT Optics Division, R&D Physical Optics,Mainz,
instrument. West- Germany supplied us with special K5- type
reference material with best chemical resistance
2. OPTICAL PRINCIPLE properties with respect to the aggressive sea
water. From that raw material we had the index
To begin with, have a look at figure 1 showing the reference body shaped according to figure 2.
optical foundations of the instrument. a
750 nm laser light from the single mode laser diode 3 mm
LSD is launched into the polarization maintaining
fibre PMF which guides the light through a bulk
head into the sea water in such a way that it hits
as incident beam ib at nearly grazing angle the
index reference glass body IRB at it's glass / sea
water interface. The refracted beam rb re-enters
the pressure protected inner part of the vacuum
optical bench system and is lead on through the
magnifying projecting optics PRO and focused on
the silicon crystal surface of the posicon, serv-
ing as beam position detector BPD. The angular dis- 0.8mm
placement of the deflected beam db is a function
of the refractive index n of the sea water. Fig.2: Favourable shape of the sensor sting for
best hydrodynamical behaviour
In order to fulfil accuracy requirements the laser
diode and the posicon chip are kept on constant This standard material shows fairly low temperature
temperature by means of the common thermostat TST. and pressure coefficients:
3. RESOLUTION An = - 2 jo-7 K
Center part of the point sea- refractometer is the AT
index reference body IRB, a standard material of An 10-8
best available quality with refractive index N. As --ip = + 4 dbar
we have vacuum at the inner side of the optical
bench system, the following relations hold: These numbers have to be used to simply correct for
siny = N sin 0 = N sin ('cL - E temperature and pressure influences.
y = arc sin ( n cos F, N2 - n2 sin c 5. BEAM POSITION DETEKTOR
as sinCE = n / N. The lateral-effect photo detector used for sensing
the position of the refracted and deflected beam
The sensitivity of the influence of n on follows has a sensitive-area of 1 x 3 mm2 and is a special
from differentiation: silicon chip made according to our requirements.As
relatively often @mall but not negligible crystal
in F disturbances occur, all incomming chips had to un-
Cos E + s dergoe a thorough check for crystal defects. For
dn Cos Y NF(-N/n)'- 1 that purpose we developed a high precision inter-
ferometric length measuring machine in a well con-
of . the trolled environmental chamber to measure the beam
As we choose Y = 0 for the mean value no position output signal as a function of the light
total range of refractive index variation in sea spot position with the ultimate possible accuracy,
water, we arrive at: reaching the limits of noise. Spot size, light in-
i tensity, and some other parameters could be varied.
dn Cos & __ - The light spot position signal is continously mea-
n=n @1 - (n/N)2 sured while the spot moves along the mean axis of
Y=00 the crystal surface. Figure 3 shows the character-
istic of a passed crystal (solid line) in compari-
To get a numerical idea about the sensitivity the son with a rejected one due to obvious defects
following figures are given : For our standard ma- (dotted line).
terial of the index reference body we have
E = 62.060, and with the * imum requested re- The lower diagram in figure 3 shows the simultane-
so m19
lvability of dn = 5 - 10- we get dy = 0.22" ously measured intensity signal UL to be almost
what corresponds with a slope of about 10 cm to independent on the beam position.
100 km ! 6. EVALUATING ELECTRONICS
This should be strong argument that the stability
of the whole optical bench system in the full In order to improve the signal to noise ratio an
?ceanographic range of pressures and temperatures AC- modulation technique has been used. The laser
is of paramount importance.

498

60 7. SPECIFICATIONS
A Center scan
d ZEISS Lurninar f=25mm After the brief description of the main features
Spot: 0.360.. of the instrument's measuring principle, the
30 main specifications are to follow now:

--- extremely small measuring volume: 0.1 pI
-------- --- high measuring speed: 1000 samples per second
--- measuring wavelength: 750 nm
B
--- measuring range: 1.323 !!: n t 1.369
-30 --- precision: 10-6
--- accuracy: +10-5 ( estimated with distilled
water and -Felieved to be achievable for the
whole range of measurement as soon as re-
-6 liable sea water data for 750 nm will become
-4060' -2000 0 2000 1 40100 available )
dX/8 --- over all length: 48 cm
- ---max.diameter: 8 cm
4- 1 --- weight in air: 5.5 kg
u,./v Ointensity of --- max.working depth: 6000 m
2- light 660n. ---materials exposed to sea water: titanium,
glass, neoprene

The industrial prototype instrument is depicted
-4000 -2000 0 2000 d AIS 4000 in figure 5, and a close up photo in figure 6
Fig.3: Calibration functions of posicons shows details of the rugged refractive index
A good silicon crystal sensor sting.
B crystal defect
I light intensity belonging to curve A 8. PRELIMINARY CALIBRATION
The curves illustrate the deviation from At present time, refractive index data for a com-
linearity versus light spot position in plete calibration of sufficient quality are not
units Of d ),/8 . ( d X= 632.8 nm, He-Ne- available from the literature. Until we will have
laser wavelength of the length measuring in- finished our own laboratory interferometer for
terferometer 750 nm index of refraction measurements of sea

light is modulated with kHz sinusoidal
r
signal. The pre- amplified signals of PA1 and A B.P S&H X:
PA2 in figure 4, which shows the basic con- Filter INTENSITY
cept of the evaluating analog electronics,. PA1 PA COMP. 11"S I
are led through band pass filters and then 3s L I(L-5) U,
are peak- rectified by use of sample,and 6S.1
hold modules. The proper control signals for O.S. POSITION
the sample and hold circuitry are,generated BPD
by comparators as zero crossing detectors and B B. p
monostable multivibrators ( one shots O.S. ) Fitter S&H
with suitably adjusted time constants T T db
Then the well established standard techAl?qug* COMP US
for analog computation of intensity UL and '172 EVE: EvALuATiNG
position U+_is applied as e.g. described in ELECTRONICS
7 a S. (ANALOG
The over all electronics noise level is kept
below a 30 nm position signal equivalent,
thus being below 5 - 10-7 refractive index PHOTO CURRENT 3, = )S+ J(L-S)
equivalent. s
and 3 = lo( 1 -7) hotds
or )(L-Sf 30( LLS respect,
INTENSITY INDEPENDENT POSITION SIGNAL:
Fig.4: Basic concept of the AC- measuring photo- IS -)(L-S) =I - 2 S
electronics and beam position detection )S +](L-S) L
3(L
S-1

499

Fig.6: Refractive index sensor sting

--------------
C,T-sensors!

-------- - -- - -
--- -------

48*

C,T- sensors

1.64.

0.65.
2. TS_
EF12,L Industrial prototype of fibre optical
point sea- refractometer -C
TF
water, we have to put up with the 400 ... 700 nm data
compiled by Austin and Haliks / 8 / and extrapolate
them for 750 nm. As these data remain very doubtful --- J__
we use them only as fixed check points against which
to observe drifts, etc. n-sensor
( Note: It is not important to know the accurate
values in the context of this paper

.em
9. FIRST HIGH SPEED PROFILING TEST Tf
In order to best demonstrate the capabilities of C
our refractive index sensor we were looking for the
best test facilities available today, and these Fig.7: "Flying Fish" for surface to bottom high
were found at Woods Hole Oceanographic Institution: speed density profiling
A.Bradley's "Flying Fish" / 4 / well equiped with
most recent Neil Brown Mark V CTD- electronics / 9 The-upper right hand picture represents an axial
and Mark III CTD- sensors. Figure 7 shows the rear view and illustrates the relative angular po-
"Flying Fish" together with the arrangement of sen- sition of n- and C,T- sensors.At the middle right
sors. Is a drawing of the Mark III CTD- sensor cluster
with Ts Slow Rosemount Pt-thermometer,
In order to properly accomodate the optical experi- TF Fast thermistor bead thermometer,
ment, a special mid section with two "short wing :C Ceramic conductivity cell.
tanks" had to be built. The left tank holds the
refractive index sensor, while the other tank con-
tains units of power supply and solid state memory At the lower right: Location of thermistor bead
@48-'

for refractive index sensor operation and data relative to the conductivity cell.
logging.

500

How the optical sensor fits into it's short wing _@o An/,Cj. 6P 120
tank can be seen from the cross section drawing in
figure 8. 0

300--
An
EVE

MOD
LS TST
500-
BPD-_ PA2

PA1__ SLB

700
VO8
10cm

PRO
%dbar
PM IRS 900-

10 i/.C 20
Fig.8: Compact arrangement of the components of Fig.9: Optical density profile. 190 000 refractive
the refractive index sensor index data taken in less than 200 seconds.
The abbreviations stand for: "Flying Fish" drop #4, Nov.11,1987
X= 55023.7'W, (P= 37049.8'N
EVE: Evaluating electronics
MOD: 10 kHz modulator Now, if we subtract the very low frequency part of
LSD: Laser diode the An- profile in figure 9 from the an- profile
TST: Thermostat itself and express the refractive index differences
BPD: Beam position detector in equivalent density differences APn, we get the
PA1: Pre- amplifier channel I spiking- free profile in the third column of figure
PA2: Pre- amplifier channel 2 10. Taking the CTD- data, we calculate the PEOS 80
SLB: Stream lined body density profile / 10 / and subtract in the same way
VOB: Vacuum optical bank it's very low frequency part to get the I&PEOS 80
PRO: Projecting optics profile in the first column of figure 10, referring
PMF: Polarization maintaining fibre to the use of the slow Rosemount Pt- thermometer
IRB: Index reference body data. In case we were using the fast thermistor
bead data, we got the curve in column 2. The last
The first high speed field test took place on WHOI column in,figure 10 depicts the result of the same
R/V "Oceanus" cruise 195, Oct.27 - Nov.15, 1987 procedure applied to PSS 78 salinity calculation
/ 6 /. The refractive index sensor was put in for / 11 / and expression of salinity differences in'
the very first time at "Flying Fish" drop #4 on terms of equivalent density differences.
Nov.11,1987 and produced the first refractive index
profile together with CTD- data as shown in figure Comparing these results, we learn that the refrac-
9 . On the extremely smooth ride down with 4.6.m/s tive index measurements perfectly resolve the
all data were taken within roughly 200 seconds, density structures. This is also.obvious from fi-
about 1000 refractive index values and 40 CTD data gure 11 illustrating the directly measured profiles
sets every second. of a thin layer in the pycnocline. Here every
measuring point of refractive index and CTD data is
The graph in figure 9 demonstrates at first glance printed out. From looking at the curves one might
the refractive index profile to be an extremely be tempted to deviate towards the discussion of
good estimator for density as suggested by the sophisticated anti- spiking algorithms and proce-
Lorenz-Lorentz equation. Today, there is no other dures, instead of this one should rather realize
in situ measuring single sensor with such a close and think about the advantages rising from the
approximation of the searched density profile. fact, that the optical density profile is a direct

501

APn /Ib to-' NEW 80 10-5
A A -4 0 4 -8 0 4
*,r, 0-
EM-3465 JL
CM.4W CAM-3-11" EgM-340'

100-1

5-
300 L
.7 -120

-2 P
/dbar
to-
T TF -125
9

3
700- 1

Fig.12: Deviations-from the mean density profile, observed
in the pycnocline with refractive index measure-
%bar' ments on the one hand and with CTD- measurements
on the other hand.
900 "Flying Fish" drop #4

high speed single sensor result as compared to the
Fig.10: Registrated "density roughness".Third indirect multi sensor PE?S 8? - profile.
column depicts spiking- free optical density Figure 12 is another examp e o the excellent reso-
measurement. lution of the density field in the pycnocline by
means of refractive index measurements. To traverse
TS calculated profile basing on slow the 14 m layer the "Flying Fish" needed some 3 sec-
Pt- thermometer data onds. The region between the dotted lines is the
TF dto.,basing on fast thermistor data same thin layer as regarded in figure 11. However,
this time the curves are difference plots of the
"Flying Fish" drop #4 same kind as in figure 10 in order to get a distinct
resolution for the comparison of the original data.
P/.JL
0 1 2 3 0' 4 1.02585 1.02690

120

0.2 ............ . .

-121
.OA An T@F C
Fig.11: Optical sensing Of 2-
the fine density
structure in a P/dbar
thin layer of the -122
pycnocline AZ/,'
"Flying Fish" drop #4 -0.6
21.9 22.0 T/9C 22.1
3
na 61.9 %S sio
CM

502

Ap/-VO-5
-32 -16 0 16 32 -32 -16 0 16 32
100- 400-

300- Soo---

DO
Soo- Soo-

700- 1000- f

P/dbar
Pldbar-

900-- 1200-

1 1.5 2
-1 0 AO/10-3 I An/10-3
Fig.13: A An curve Fig.14: Same as figure 13, but refering to another
B CTD- derived P[Oj 80 curve, adjusted water body, measured on location:
in such a way that o h curves have common "Flying Fish" drop #5, Nov.14,1987
profile end points. 6702.7'W, @D= 4007.5'N
C : Differences A-B, expressed in units of
density equivalents. 10. ANOMALIES
"Flying Fish" drop #4 If one compares the density plots derived from op-
tical refractive index data with those calculated
from electrical CTD- measurements, as it is done in
figures 13 and 14, one discovers characteristical
differences due to the variability of the composit-
ion of sea water. ( Note: An- plots in figures 13,
14 refer to the exclusive use of the linear part of
the refractive index calibration function. The true
density differences are somewhat smaller and will
500
be published elsewhere ). These differences are
small and normally concern only the 5th and 6th de-
cimal place, but they may well serve for more de-
5
j__ tailed studies of water bodies, as has been sug-
400 gested by by Bein,Hirsekorn,and MbIler / 14
t 0
GO *4 E.g. they determined difference contours as repro-
Se duced from / 14 / in figure 16, illustrating a sect-
ion through the Danmark Strait between Iceland and
k
30* Greenland. The West to East variation of the com-
position of sea water is as well seen as the pro-
70o 60' so. nounced indention of the Irminger and East-Green-
land currents.
Fig.15: Stations locations of drops #4, #5. 11. CONCLUSIONS
MarKed on surface currents map taken
from.Dietrich and Ulrich 15 The advent of new optical techniques and components
diode
like laser s, fibres, silicon chips, etc. make
us believe, that the concept of refractive index

503

measurement of sea water is a "hundred years old 5/Munk,W.: "Remote sensing/ contact oceanography"
good idea whoe's time has come". This is partcular- talk in plenary session:"The Ocean frontier
ly true with respect to high precision and high in 1990"
speed measurements."First class" n,C,T,D- instru- OCEANS187 conference,Halifax N.S.,Canada
ments on high performance vehicles are a powerful Sept.28- Oct.i, (1987)
new oceans research tool.
0. 69 70 71 68 67 5 3 62 /6 /Bradley,A.M.: "Fast hydrographic profiler"
in cruise report R/V "Oceanus" 195
WHOI, Oct.27- Nov.15, (1987)

/7 /Kelly,B.O.,and Nemhauser,R.I.:"Techniques for
using the position sensitivity of silicon
photodetectors toprovide remote machine
100@
Profil control"
21 st Annual IEEE Machine Tool Conference,
15Wr Profil Ya -1500 Hartford, Conn.,USA, (oct.1973)
8/Austin,R.W., and Halikas,G.: "The index of re-
12M fraction of sea water",
-1,SCRIPPS Inst.of
5 Tech.Rep. SIO Ref.No. 76
Oceanography, La Jolla, CA 92093,(Jan.1976)
2500L____
31- J21
Abb. 24. 9 /Brown.N.L.: "New generation CTD sensor system",
OCEANS'87,IEEE Publ.No.87 CH 2498-4,01.1,
Fig.16: Contours of differences between chlorinity Halifax,N.S.,Canada, 280 ... 286,(1987)
determined density P17 5(CI) and refrac-
tive index determined d6nsity P17.50), /10 /Fofonoff,N.P. and Millard,R.C.:"Algorithms for
reproduced from / 14/.(numbers are in 10-6) computation of fundamental properties of
12. ACKNOWLEDGEMENTS sea water"
UNESCO techn.papers in marine science
The authors wish to express their deep appreciation No.44, (1983)
for Dr.Albert Bradley's,WHOI, most friendly invi-
tation to use his "Flying Fish" and facilities,and /11 /UNESCO: "Background papers and supporting data
for his substantial help in preparing for the field on the Practical Salinity scale 1978"
work. UNESCO techn.papers in marine science
We greatfully acknowledge Dr.N.Neuroth and W&Jochs No.37, (1981)
from SCHOTT-Glass,Mainz, for donating some high
quality glass samples to this project. /12 Mahrt,K.-H.,and Kroebel,W.: "Quantitative per-
We would like to thank the German Research Society formance data of a new automatic optical
(DFG) for financial support. bench salinometer/ densitometer"
OCEANS185, IEEE Publ.No.85 CH 2250-9,
13. REFERENCES San Diego,CA,USA,622 ... 627,(1985)
Mahrt,K.-H.,Waldmann,H.C.and Kroebel,W.: 13 Mahrt,K.-H.: "Neue Methoden zur Pr6zisions-
"A newly developed in situ-measuring ocean- messung des optischen Brechungsindex von
ographical probe sensing the optical index Seewasser in situ im Ozean und im Labor
of refraction of sea water with new aspects dargestellt an erhaltenen MeBresultaten"
of salinity and density determinations" Verhandlungen d.Deutschen Physikal.Ges.,
OCEANS'82,IEEE Publ.No.82 CH 1827-5, 51.Physikertagung, Vol.5,ME-4,Berlin(1987)
Washington,DC,USA,266, ... 271,(1982) 14/ Bein,W.,Hirsekorn,H.G.,M61let,L.: "Konstan-
2 /Mahrt,K.-H.,and Kroebel,W.:"Optical interfero- tenbestimmung des Meerwassers und Ergeb-
metric bench salinometer of high precision nisse Uber Wasserk6rper"
with electronical read out" Ver6ff.d.Inst.f.Meereskunde,Univers.Berlin
OCEANS'84,IEEE Publ.No.84 CH 2066-9, Neue Folge A,Heft 28, 240 S.,(Nov..1935)
Washington,DC,USA,219 ... Z23,(1984) /15 Dietrich,G.und Ulrich,J.: "Atlas zur Ozeano-
3 /Heinmiller,R.H.: Report:"Profilers of the fu- graphie".
ture" Meeting (WOCE),Earth Sciences Bldg., Bibliogr.Institut Mannheim,S.45, (1968)
Massachusetts Institute of Technology,
Cambridge,MA,USA,Aug.8-9,(1984)

4 /Bradley,A.M.: "High performance free-fall pro-
filing vehicles"
Marine Tech.Soc.Journ.21,33 ... 41,(1987)

504

MULTI-SAMPLE PARTICLE FLUX COLLECTOR

STAFF*

Oregon State University
College of Oceanography
Oceanography Admin. Bldg. 104
Corvallis, OR 97331-5503

ABSTRACT 4) Sufficient 0.50 square meter collecting area to provide
- gram-sized samples for multiple analyses
A non-metallic, conical fiberglass sediment trap has been 5) In situ sample poisoning and preservation
designed by personnel at Oregon State University to collect 6) Fifteen modular sample cups which can be opened and
uncontaminated, preconcentrated samples compatible with closed in situ
trace element analytical requirements. The sediment trap has a 7) Isolation of each individually sealed sample from contam-
0.5-m2 collecting area and a 15-multiple sample cup mechan- ination or loss during deployment and recovery, and
ism to allow the sediment trap to collect gram-size monthly 8) Removal of sealed sample from the sediment trap without
samples during a typical year-long deployment and to isolate exposure to atmospheric contamination.
them until recovery. MULTI-SAMPLE PARTICLE
The sample acquisition timing is controlled by a programmable COLLECTING SEDIMENT TRAP
CMOS microprocessor for variable sampling intervals from 32
seconds to tens of months. In addition, the microprocessor The design concept of the sediment trap is based upon facilita-
stores a permanent event record of trap performance. ting clean sample transferral and trap deployment. Sealed sam-
ple cups can be removed and archived individually without ex-
Particular emphasis has been directed at keeping the trap ternal shipboard contamination. The modularity of the sedi-
modular for ease of sample handling and trap manufacturing.
Components of the trap are compatible for moorings to 6000
meters depth.
0 im
INTRODUCTION nylon braid
mooring lines
In recent years (1,2,3), particle collectors have been used in- llei@ 'A.
creasingly in the marine environment to investigate the com-
A
plex cycles of marine sedimentation from surface production `.11
and mid-water transformation, to eventual burial, diagenesis,
@ F., hexcel baffle
"@upw@
and resuspension at the ocean floor. In order to fully under-
stand the complex, rapid changes in oceanic productivity and
sedimentation, it will be necessary to collect samples on a
K-111
monthly or bi-weekly basis over the course of several years. A,
PkJll,!WV@,l fiberglass
E, ", E
In addition, increasing interest in trace element chemistry of the
ti collector funnel
oceans makes it extremely important that samples remain un- @k4
contaminated during typical year-long deployments.

The following description is for a particle collecting sediment
trap designed to address these rigorous sampling criteria. The
trap incorporates the following features: funnel outlet
and coupler
-metallic construction materials compatible with trace
1) Non
element analytical requirements
2) Conical shape for preconcentration of sample and ease of
multi-sample
processing
carousel assembly
3) An accurate and reproducible trapping efficiency under
normal oceanic current environments

Contributions by:
Dr. Roderick Mesecar, Head
Technical Planning & Development Group
Chris Moser, Research Assistant
Marine Geology Figure 1. Multi-sample sediment trap assembly

CH2585-8/88/0000- 505 $1 @1988 IEEE

ment trap design also allows removal of the sample changing
mechanism for servicing in a clean, trace element laboratory. anodized 6061 T6
aluminum
The multiple sediment trap, shown in Figure 1, consists of pressure case
three subsystems: (1) a conical fiberglass collector with feeder funnel
honeycomb baffle; (2) a multiple 15 sample collection mechan-
ism controlled by a programmable timer; and (3) a non-metallic
nylon rope harness for flotation and mooring attachment. compression fit
FIBERGLASS CONE o-ring seal

The fiberglass sediment trap cone has a 0.50 square meter col-
lecting area and was modelled after a design from Soutar et al.
(4). The diameter of the trap mouth is 80 cm., that of the low-
er opening is 7.5 cm., and the height is 160 cm., giving the
cone a 2:1 height/diameter aspect ratio for maximum trapping
efficiency. The trap mouth is covered by a polyester resin,
honeycomb baffle with 1-cm. diameter cells that are 5 cm.
thick which creates an entrapped water boundary at the mouth
of the sediment trap to reduce internal eddies. A horizontal lip
extends 5 cm. beyond the edge of the baffle to help deflect cur- eccentric
rent eddies downward and reduce the turbulence in current motor drive
flow over the trap mouth. The walls of the cone slope only crankshaft and
14* from vertical and are coated with an extremely smooth, sliding actuator
clear acrylic resin to allow trapped material to fall rapidly to the mechanism
multiple sample collector. vertically compensating
MULTIPLE SAMPLE COLLECT10N MECHANISM sleeve with Teflon o-ring
[email protected]
A mating flange on the fiberglass cone bolts to a lower multi- valve dy wt
sliding gate
sample collector assembly which can be removed and serviced

fiberglass carousel
unit support frame Lexan polycarbonate
sample tube
microprocessor/timer
shaft and torque stop mountm*g cage
for torsional spring

removeable end cap

Figure 3. Schematic drawing of the Collector assembly
"@Ojl
components.

independently in the laboratory (Fig. 2). This assembly con-
sists of a feeder funnel and rotating sample carousel plate with
fifteen independently sealable sample tubes; a pressure housing
that contains a microprocessor based timer unit, high-torque
motor and batteries; and an actuator alignment mechanism that
opens and closes each sample tube as it sequentially rotates
@V into precise alignment beneath the feeder funnel aperture.

The essential collector assembly components are shown sche-
matically in Figure 3. The collector assembly mates to the
fiberglass cone through a feeder funnel with a vertically com-
ON pensating, telescoping tube faced with a teflon coated silicone
rubber O-ring which seals tightly against each of the fifteen,
wedge-shaped sample tubes. Each sample tube is capped by a
unique sliding gate valve sealed with a teflon-coated silicone
rubber O-ring. This configuration allows every sample tube to
carousel sample plate for be loaded and removed from the sediment trap without being
one of three carousel plate fifteen sample tubes opened. The collected sample is only open in situ to prevent
roller bearing supports any contamination from the ambient seawater or the atmos-
._r
p
in
es
a
s
@su@u
emc a, @e

phere until analysis in the laboratory. All feeder funnel, valve,
carousel torsional and sample tube components are constructed of LEXAN poly-
spring assembly carbonate plastic to prevent leachable trace metal contamina-
tion. Polycarbonate also withstands extremely rough handling
Figure 2. Sample tube carousel assembly at cold temperatures without breakage.

506

Oregon State University personnel laboratory experiments BASIC to open each sample tube for variable intervals from 32
have shown that upwards of 30% of the total dissolved sample seconds to many months or more, if needed. The micropro-
fraction can be lost from even a 20 cm. long sample tube over cessor also keeps an independent history of motor jog events
a two week period. This soluble fraction can amount to a sig- utilizing a backup lithium battery in the event of main battery
nificant portion of the total flux of some sample components. failure.
Therefore, each individually valved sample tube is 3.8 cm.
I.D. and 44 cm. long and is filled with 500 ml of a dense anti- ne trap is deployed with a sample valve open under the feeder
bacterial sodium azide brine. This long, narrow sample tube funnel which constitutes position #1 in the sampling sequence
design minimizes any loss of soluble sample fractions from the as shown in Figure 5. Under command from the microproces-
stable brine while the tube is open and collecting. sor, the motor begins turning and the crankshaft pushes the
valve pin forward to begin closing the valve. This also moves
The 15 sample tubes are bolted to a carousel plate that rotates the pin against a sloping ramp channel that breaks any sticktion
on three non-metallic bearing supports and relies on a linear, and allows the sample valve and sample carousel to rotate out
torsional spring to directly rotate the sample tubes into posi- from under the feeder funnel to position #2. Here the sample
tion. This high-tension spring is an integral part of the carou- valve pin escapes the actuator mechanism and the carousel ro-
sel mechanism (Fig. 4) and is located on the carousel axis in a tates freely powered by the torsional spring until the next valve
sealed, pressure compensating chamber filled with water-solu- pin enters the actuator at position #3. This second valve is
ble lubricant to protect it from corrosion. The spring applies a closed when it enters, and, as the crankshaft continues to
continuous torque which can be adjusted simply by how many rotate, the actuator pulls the valve pin back and the valve opens
turns the spring is loaded before deployment. The motion of through another sloping ramp channel which guides the new
the spring-loaded carousel sample plate is regulated by a sam- sample tube into registration under the feeder funnel back to
ple actuator mechanism and microprocessor timer that deter- position #1 where the motor rotation stops. The second sam-
mine how long each sample tube is open and collecting mater- ple tube is now open and ready to receive sample material for
ial. its preset period.

carousel shaft
sample tube under
sample tube valve body feeder funnel
with sliding gate
sample carousel plate 0 0 0
Z"I'MIMM131M,

0 0 0 0

corresponding
motor shaft represents path
positions of valve pin

Figure 5. Sample valve actuator mechanism
operating sequence

NYLON NON-METALLIC ROPE HARNESS
carousel torsional sample tube
spring housing During a typical oceanic deployment, each sediment trap is
location of spring suspended within a completely, non-metallic, nylon rope
assembly pressure harness (Fig. 1) for easy coupling into any mooring design.
compensation diaphram Since each trap is usually deployed in a mooring below glass
Figure 4. Carousel sample holder assembly flotation on galvanized chain, the harness is designed to keep
any source of metal contamination at least 50 meters above the
trap mouth. To further reduce potential metal contamination,
MICROPROCESSOR TIMER AND MOTOR the traps have also been deployed with PVC plastic radio
buoys, nylon flotation harnesses, and type 316 stainless steel
The motor, microprocessor timer, and batteries are enclosed in shackles to eliminate corrosion and galvanized metal compon-
an anodized 6061 - T6 aluminum pressure case tested to ents. The harness and trap have been successfully tested to
10,000 psi. The high torque, gear-reduction motor engages an 3600 Kg working load with failure occurring at 6300 Kg
eccentric drive and crankshaft to power the sample valve actua- tension. The trap weighs 54 Kg in air (22.5 Kg in water) and
00

tor and sample changing mechanism. The low power-drain, can easily be deployed and recovered with typical shipboard
CMOS microprocessor can be programmed in user-friendly facilities.

50.7

SUMMARY

In order to monitor the, increasingly complex changes in ocean-
ic productivity and sedimentation, we have designed a new
multi-sample particle flux trap which can collect and preserve
15 individually seated samples during a typical year-long
deployment. The non-metallic fiberglass and LEXAN poly-
carbonate plastic construction materials and type 316 stainless
steel hardware are compatible with trace element analytical
requirements. Individual samples can be removed from the
trap without exposure to atmospheric contamination. A micro-
processor controlled sample changer can be programmed for
an unlimited variety of sampling rates while storing a perman-
ent historical event record of trap performance. These sedi-
ment traps can be deployed to 2700 Kg meters depth and under
tension loads up to 3600 Kg.

Three of these sediment traps are currently moored at 1500
meters depth on three moorings off the Oregon coast and are
due to be recovered in late September 1988.

REFERENCES

1.Spencer, D.W. The sediment trap intercomparison experi-
ment. Some preliminary data. In: R.F. Anderson and M.
P. Bacon, eds., Sediment Trap Intercomparison Experi-
ment, Woods Role Oceanographic Institution Technical
Memorandum WH0I-1-181,1981, p. 57-104.

2. Asper, V.L. A review of sediment trap technique. Marine
Technology Society Journal 21, 1987, 18-25.

3. Honjo, S. and K.W. Doherty. Large aperture time-series
sediment traps: design objectives, construction and appli-
cation, Deep-Sea Research 35, 1988, 133-149.

4. Soutar, A., S.A. Kling, P.A. Crill, E. Duffrin and K.W.
Bruland. Monitoring the marine environment through
sedimentation. Nature 266, 1977, 136-139.

508

MULTI-SENSOR REAL-TIME DATA ACQUISITION AND PREPROCESSING AT SEA

J. M. Moore, J. S. Charters and C. de Moustier

Scripps Institution of Oceanography
University of California, San Diego
La Jolla, California, 92093

ABSTRACT second, or "offline" system, is available for data processing, software
This paper describes the real-time underway data acquisition development, general use (word processing, correspondence, etc.)
system developed by the Shipboard Computer Group of the Scripps and as a backup for the "online" computer. Both machines run under
Institution of Oceanography in support of seagoing research. Since the Berkeley UNIX (BSD 4.3) operating system, supporting applica-
February 1984, this system has been implemented on a DEC VAX- tion software written in C, FORTRAN-77 and UNIX shell script pro-
11/730 computer which runs the UNIX operating system and com- gramming languages.
municates with various sensors through an IEEE488 General Pur- The following peripheral devices are supported by each sys-
pose Interface Bus and slave microprocessors. The sensors inter- tem:
faced in this fashion acquire navigation data (i.e. ship's heading, - Two Fujitsu 160 megabyte Winchester disk drives, interfaced
speed and position) and geophysical data (gravity, magnetics, Sea through an Emulex Unibus controller. On the R/V T Washington
Beam bathymetry and expendable bathythermograph profiles). After each system also includes a Fujitsu 300 megabyte Winchester disk
a brief overview of the hardware involved, the paper discusses near drive.
real-time processing of the underway navigation and geophysical - One Kennedy 9300 tape drive, 800/1600 bits per inch, 125 inches
data made available to the investigator during the course of a survey. per second, connected to the VAX via an Emulex Unibus
L INTRODUCTION controller/formatter.
For 17 years, reaf-@time acquisition and processing of ship navi- - A National Instruments Unibus IEEE488 GPIB controller to pro-
gation and marine geophysical data on the R/V T. Washington vide the communication link between the VAX and its slave
operated by the Scripps Institution of Oceanography (SIO) were han- microprocessors.
dled by an IBM 1800 computer system permanently installed aboard - One Calcomp 965 four color belt bed plotter (R/V T. Washington
ship. From the start, the Shipboard Computer Group (SCG) at SIO only) with an effective plotting area restricted to approximately 34
has been responsible for operation and maintenance of this system by 36 inches. Note that the plotteron the "online" system is devoted
and of nearly identical installations on R/V's Melville and Argo. In to plotting in near real-time the ship's track in geographic coordi-
February 1984, these shipboard computer systems were replaced by nates and bottom contours when Sea Beam is used. The "offline"
DEC VAX-1 1/730 systems on the two ships still in operation: R/V's plotter is used both as a post processing output device and as a
T. Washington and Melville [ 1 ]. To use these new computer systems backup to the "online" plotter. These plotters communicate with the
for real-time data acquisition, SCG built a communication interface VAX through a 9600 baud RS-232C serial line. The RIV Melville
based on an IEEE 488 General Purpose Interface Bus (GPIB) and an has only one Calcomp 502 flatbed plotter connected to a VAX-
IEEE 961 Standard bus (STD) to acquire, buffer and transmit data to 111730 through one of the STD/GPIB interfaces described below.
the VAX from various navigation and geophysical sensors. - Up to eight Graphon-140 (DEC VT-100 compatible) terminals pro-
In the following we first give a brief overview of the VAX- vide user input/output and data display functions. All have graphic
11[730 computer hardware and peripherals installed aboard the ships capabilities emulating the Tektronix 4010; some have attached
and of the various sensors interfaced to the computer through the Okidata-92 printers with graphic enhancements, providing parallel
STD/GPIB data acquisition system. In Section III, we discuss the hard copy output.
real-time data acquisition system including descriptions of the STD Hardware interfaces to real-time data input/output devices are
and VAX-l lf730 software components, along with how these com- built into eight industrial instrumentation systems using the IEEE
ponents interact over the GPIB. We describe the near real-time 961 STD bus protocol. Each STD system contains a single Z80
VAX-lln30 software used for data processing, storage, and to drive microprocessor supporting from four to sixteen kilobytes of memory;
the displays for the various devices in section IV. And we conclude direct memory access (DMA); a GPIB for communicating with the
with comments relating to both the strong and weak aspects of the "host" computer, together with a variety of digital, analog and spe-
data acquisition and processing system, together with ideas for future cialized data acquisition interface cards. Each STD system also has
improvements. its own software clock which is used to time stamp all data passing
from one of the acquisition devices through it to the VAX. A
H. SYSTEM OVERVIEW hardware clock which displays day-of-the-year, hours, minutes and
On board ship, two Digital Equipment Corporation VAX- seconds can be set and interrogated through the STD/GPIB VAX
lln30 mini computers provide the "main frame" computing link.
envirortment. Each machine is equipped with a floating point As shown in Figure 1, the navigational sensors interfaced to the
accelerator and four megabytes of memory. One is designated as VAX through the Z80/STD/GPIB link include: a gyrocompass for
"online"; its primary responsibility is to interface with the various the ship's heading, a dual-axis Doppler current profiler for the ship's
shipboard instruments for real time data acquisition and display. The speed, a Transit satellite receiver and a Loran-C receiver for the

CH2585-8/88/0000- 509 $1 @1988 1EEE

NAVIGATION DATA DEVICES REA IME INITIALIZE CLOCK INTERR
CLOCK Reset clock K41MR u@
GYYFO DOP Clear poll and transfer tables
Set-up GPIB/DMA protocol rR`.-d-4o-.k_.7.7
Idata buffer _j

MAIN POLLING Loop
7 Read clock and update address Into transf
timeout counters table and rot., rn
Z801STD BUS MICROPROCESSOR SYSTEM
Data Collection, Data Buffering and Software Clock dela Decode Identity and
yes read in branch to service
available routine
from buffer
ROM CODES
A
AK - acknowledge
recelved,clear
UNDERWAY DATA HARDWARE DEVICES
data appropriate slot
no valid yes in transfer table
01. download buffer
to specified
RAM address.
MAGNE-Tncs IEEE-488 GPIB LINK XQ_ fill in poll
ISP@
Y
fDISPLAY TO VAX 111730 __7 Check transfer table and perform address to turn
7
any pending data buffer, device on or off.
acknowledge, or retry transfers
Figure 1. Architecture of the real-time data acquisition system. to VAX 11/730 RAM CODES
Ind! cate which
device service
ship's position. Note that the receiver for the Global Positioning routine is to be
System (GPS) transmits its data directly to the VAX by a serial (RS- evok .
232C) link. Return from
device Call device service routine
Underway scientific data collected on the RN T. Washington servi yes
'Ice
may vary from leg to leg, depending upon the particular scientific no address 1. service
in poll utine
requirements for that leg. The "standard" devices installed on board table
ship which can be serviced in any desired combination include: a
gravimeter, a magnetometer, a Sea Beam mutlibeam echo-sounder, Increment
an expendable bathythennograph and a one- or two-channel seismic poll table
system. A subset of these devices, no gravimeter and no Sea Beam Iaddress
system, is also available aboard the R/V Melville. Special "one Figure 2. Flow chart of the Z80 ROM software common to all STD
time" devices have also been accommodated at the request of many units.
chief scientists, but will not be discussed here.
111. STANDARD BUS TO VAX-11/730 INTERFACE The bi-directional communication link provides the
input/output path between the STD/Z80 and the VAX for data
The UNIX BSD 4.3 operating system has no "real-time" acquisition, device control and the downloading of RAM device
pretensions in the data acquisition and control sense; however it does specific software. These routines accept packets sent from the VAX,
provide flexible capabilities that would be difficult to reproduce on a inform the VAX of a full data buffer ready to be sent, transmit pack-
computer required to respond in microseconds to external events. ets to the VAX, construct check characters for each transmission and
By incorporating the STD/Z80s to handle the input/output details initiate retransmissions as needed.
and bringing buffers of data into the mainframe computer only as . Great care was taken in the design of the polling loop to insure
fast as is convenient for navigational and human response times, the each device an equal "time slice" of the Z80's attention. No "loop"
VAX is allowed to concentrate on the interactive human-to-data or "wait" states are allowed; if a device is not ready to accept or to
interface where UNIX performs so well. send data, the next device in line is interrogated. By maintaining
Our STD/Z80 design philosophy was to allow a single tight control over both the ROM and RAM software routines, a dev-
microprocessor to control a maximum of three external devices. ice can be serviced less than a millisecond after its "ready" flag is
This limitation keeps the Z80s lightly loaded for increased reliability activated. This method eliminates the complexity and overhead of
and simplicity of programming. Configuration and control informa- an interrupt driven scheme without compromising data integrity.
tion data are maintained in random access memory (RAM) tables to Some of the device interfaces requiring high data rates incorporate
allow dynamic activation/deactivation of any device without disturb- first-in-first-out (FIFO) buffers to enable the STD/Z80 to sample and
ing the operation of other devices sharing the same STD slot. Data store many bytes within a single polling cycle.
buffer sizes vary from a few dozen bytes up to sixteen kilobytes, The clock in each STD unit consists of a sixteen bit hardware
with larger buffers possible. Most of the buffers are made small so register, driven by a rubidiurn vapor controlled pulse train with one
the VAX will always have sufficiently recent data and so that any millisecond resolution, together with a sixteen bit high-order counter
problems with the instrument or transmission path will be detected containing the time-of-day in tens of seconds. A "global" interrupt
early. can be generated by the VAX to force all STD clocks to report their
Each Z80 read only memory (ROM) contains approximately current values. Clock synchronization checks are performed
7

Yes

Yes
,F.tZy"

one kilobyte of code to provide the following elements: routines for automatically every five minutes. The VAX support software allows
communicating via the GPIB with the VAX; a polling loop to allow users to interrogate the clocks at any time. Every data buffer
frequent testing for device activity; a clock synchronization interrupt includes at least two time readings, the time of the first and the time
routine; clock reading and setting routines; and a power-up and reset of the last data samples contained in the buffer; each sample is timed
initialization routine (see Figure 2). in some data streams.

510

Device specific software resides in RAM and operates under INITIALIZE
the control of its associated ROM polling and transfer routines. Set-up Interrupt servicing Set flag to
Although the detailed functions of the RAM code vary according to Clear GPIB interface read
the instrument it services, the overall operations are quite similar. Read device configuration file configuration
Data input routines are responsible for testing device ready status, file
reading in the data in either serial or parallel format, decimating and
or reformatting the data as necessary, attaching a time stamp to the MAIN PROGRAM LOOP Reconfig.
data buffer and notifying the ROM when a buffer is ready to be Dormant until one of
transferred. Output routines normally act as buffers between the following occurs: SRO Service
VAX and the device receiving its data; additional services such as
flow control and data reformating are supplied in some cases. Interrupt signaled Wakeup
All Z80 software is written in assembly language using utilities 1 second loop active Timeout
developed on the VAX by SCG. These utilities include a cross- I
assembler, an object code downloader and a monitor which runs in Send STD clock read Return from
an unused STD slot. The monitor communicates with the VAX command If timetic file Interrupt
downloader to check and bum EPROMs and to provide code exists; delete file routine(s)
verification and debugging capabilities. - +
The software for the real-time data acquisition system in the Retry fan, unacknowledged Identify STD
VAX is controlled by a "core program" called: GPCON (GPib CON- bfe's sent to STD a issued SRO
troll . This program is run as a high priority UNIX server (i.e. a max U, of two times and urn If queue
'm
p
F 'Int message if failure ern t
background process which lies dormant until activated by an external
event) to maintain communication with the STD/Z80s. A Initiate GPIB transfers to
configuration file is read by GPCON at start-up or upon command STD for all data Read in GPIB buffer,
from a "privileged" user. This file contains an STD slot number, buffer files found send VAX acknowledge
download and execute addresses, and an identity tag for each device. and decode identity
This information insures the instrument support software and com-
munication table entries are, in place to accept and/or transfer data. sad In configuration If
As mentioned above, the "dynamic" configuration allows instru- F Rfile If flag is set Sm yes
ments to be "turned on or off' without affecting other devices in the acknowidg.
Decrement one second reeved delete Its
same STD slot. _@@cycle count file
After initialization, GPCON puts itself into a ten minute sleep _F
cycle until one of the following events occurs (Fig. 3): no
1. A VAX program requests a data transfer to the STD by creating a Write the buffer to Its file
file whose name is unique to a particular STD slot containing the
data together with header information describing the device to be Figure 3. Flow chart of the GPIB control program GPCON.
serviced. The program then sends a "wakeup" signal, forcing
GPCON out of its sleep cycle to begin transferring data from any 3. A special signal can be sent to GPCON to force it to wake up and
STD files it finds. Upon awakening, GPC0JV cycles for 6 one read its configuration file. The polling loop is again reactivated after
second intervals to insure all new or pending retry transfers are ini- the new configuration information has been obtained.
tiated before resuming its ten minute sleep cycle. 4. The ten minute "watchdog" timeout should never occur under
Clock interrogation programs create a "timetic" file which will normal operating conditions. Even with no instruments active, the
cause GPCON to send an "interface clear" interrupt to all STD sys- STD clocks automatically report their readings every five minutes;
tems. This interrupt signals each STD to read its clock and send the therefore this wakeup condition indicates a problem in the GPIB sys-
results to the VAX. tem.
2. One of the STD systems initiates a service request (SRQ) inter- The GPIB VAX support software include both mainline and
rupt when it has a data buffer ready for transfer to the VAX. The subroutine versions of utilities to: (1) Provide "wakeups" to signal
header section of the buffer contains a device identifier, time stamp, GPCON to initiate a data transfer or reread its configuration file. (2)
data byte count and any additional information specific to that partic- Download and activate Z80 software for an entire STD system
ular instrument. Control is then transferred to an interrupt routine configuration or for selected device(s) on the system. (3) Deactivate
which polls the GPIB to determine which STD system requires atten- selected device(s) on the STD. (4) Set, interrogate and report the
tion and initiates a transfer to the VAX. A checksum is performed STD clocks.
and, if all is well, the device identifier is decoded and its data are
written (or appended) to a file associated with that identifier and a All the VAX GPIB system and-support software is written in
"received acknowledge" transfer is initiated over the GPIB to that the C programming language. The ease of defining device structures
STD. and implementing the "signal-interrupt" capability made C the logi-
A data buffer may contain an "acknowledge identifier" which cal choice over other high level languages available. Some addi-
informs GPCON that a previous VAX to STD transfer has been tional user interface utilities have been developed using the UNIX
received and verified by the STD. Upon receipt of this identifier, any shell script feature. These are mainly "one-time" programs to pr6-
pending retransmissions for that device are cancelled. vide some non-standard information or control options; useful scripts
are recoded in C for increased efficiency.
The interrupt routine exits back into a now active polling loop
to service any VAX to STD transfer requests that were made while IV. NEAR REAL-TIME PROCESSING
the STD data buffers were being received. All of the VAX data acquisition, process control and display
Poll to
slot that
and ret
Is

Y
-s
.

511

programs receive and or transmit their data buffers under virtually
identical procedures (Fig. 3). Input routing is controlled through the
use of unique device identity files; output to a selected STD is deter- IEEE-488 GPIB LINK SERIAL DATA LINK
mined by the device identifier contained in the header portion of the IFFIDIVISMSYMSTEM FROM DEVICE,
data buffer.
VAX input is obtained from files created either by GPC01V or GPCCN GPS Navigation
by a UNIX "cat" command to a serial port. These files continue to VAX GPIB Interface
grow until the servicing program is activated to perform a specific Controller
processing function which always includes the removal of its input op
.*IMIM
file so no data overlap can occur between processing cycles.
.4 DISK 1/0 DATA
AND
di AM
A program initiates output to the STD by creating a file and HANDSHAKING FILES
.M
sending a "wakeup" signal to GPCON. This file has the name logelock
S I
Hard/ o t Clocks
readcloks
GPoutN where N corresponds to the STD device slot number (0
-Doppler Log dmav, s
priate
through 7). The data buffer is sent over the GPIB to the appro
..........
STD slot; the ROM allocated to that slot uses the header identifier to
0 MV
Gyro
determine which RAM servicing routine to activate.
The three following subsections describe program operation,
fiffiR,
together with the data and communication files required to process,
display and archive the real-time information presented to the VAX a 000 Transit Sat
. . . . . . ......
. . . . . . . . . .
(Fig. 4). These subsections am: STD CLOCKS, NAVIGATION and
lorxfr
UNDERWAY DATA. Loran
a
IV (a) STD CLOCKS Gravity
Data integrity requires that all information received over the cypmag
GPIB be synchronized by time. As mentioned before, our procedure
is to include at least the times of the first and last samples in each
....... . .........
s@draw Sea Beam
buffer. Time differences between STD slots are monitored by the
logclocks and readcloks command scripts. Both of these'scripts . . . . . . . . . .. . . . . . . .
. . . . . . . . ..
elsmic
seisS[1-21
S
interrogate the clocks, wait until the clocks have reported their time
.....................
02.
. . ... M.77-- 77 ."O.M.7-5
count registers (by noting the existence of the numbered clock iden-
tity files), and then execute the program clkmn to convert the counts bt XBT
to time in hours, minutes, seconds and milliseconds. Every five
minutes logclocks appends the times to the stdtimes file; readcloks is Figure 4. Block diagram of programs and data Mes associated with
invoked from a terminal to print the current STD times on that termi- each sensor. Programs appear in the boxes within the shaded area,
nal and to append the values to the stdtimes file. The output format with data files names listed directly above the box.
is:
Tue Oct 27 18:45:43 GMT 1987 entire water column within beam range) are sampled every 600 mil-
STD2time= 18:45:39.589 offset= 482 liseconds, packed into 16 kilobyte buffers and sent to the VAX every
STD3time= 18:45:39.589 offset= 190 twelve seconds.
STD4time= 18:45:39.589 offset= 2751 A background* process (drnav) performs the functions involv-
STD5time= 18:45:39.589 offset= -87 ing measurement of ship's speed and heading, inference of ship's
STD6fime= 18:45:39,589 offset= 412 true velocity over the ground and plotting of DR and fix positions.
HARDtime= 1-8:45:39.589 date= 1987/10/27 (300 of year) This process "wakes up" at one minute intervals to read values of
The first line is the VAX time when the clocks were recorded speed and heading buffered in the VAX. Speed is determined by
and is typically three to five seconds after the interrogation time. applying drift history to the two dimensional Doppler speed vector
Offset values represent millisecond adjustments needed to synchron- course is calculated by combining the gyro reading with the Doppler
ize the individual software clocks with the hardware (HARDtime), speed vector. The output includes 10 second averages for speed and
clock. One of the tasks of program clknwn is to compute these clock course, together with updated one minute ship positions.
offsets and to send the corresponding adjustment to the appropriate Inputs to drnav are the the raw Doppler D)oinpf and gyro
STD units. GYinpf files created by GPCON (Fig. 4), together with a file called
IV(b) NAVIGATION saved containing all the information that may be shared by con-
current processes (i.e. current position, last Transit satellite fix posi-
Dead-Reckon (DR) tion, course, speed, drift, as well as times associated with these data).
Analog heading information is obtained from the ship's Sperry A simple software locking scheme was implemented to prevent the
Mark-37 gymscope, passed through a 10 bit synchro-to-digital con- programs that both read from and write to this Me from obtaining
verter and sampled at one second intervals by the STD system. simultaneous access to the data and possibly corrupting them.
When 20 readings have been accumulated, the data buffer is File DPinpf is also used by the background speed display pro-
transferred to the VAX via the GPIB. gram spddsply. The Arrietek-Straza Doppler Current Proffier has no
A dual-axis Ametek-Straza Doppler Current Proffier provides direct speed indication capabilities; spddsply filters the raw values to
ship's speed and, if desired, ocean current profile data. Selected bins
(depth windows) are collected at 1.8 second intervals, buffered in the *All background processes discussed in this paper, with the exception of GPCON
STD and passed along to the VAX every six seconds. If ocean which is activated by "signals", use the UNIX "sleep" mechanism to control their
LINK
D
1
@
CON
VAX GPIB 1.
C..troll

current profile data are to be collected, aU bmis (representing the cycle periods.

512'

B
0&1200- - + 0 A@
SEA BEAM OFF
+ J%..
I _%diRk
O&JI28- MAGNETIC CENTER
ANOMALY BEAM
06, + PROFILE, DEPTH->,o,%
1100@@ ftv,900
ftft'14

04,
Irooo $000 q%00
TRACK
ADJUSTMENT
FROM
1 10 km, 01 0418 FIX
0--
449
LEGEND

+ 15 MINUTE DR
POSITION

x TRANSIT SATELLITE
POSITION
SOLID LINE RE "0410. 10 km
PRESENTS 1100 _.J
SHIP'S SMOOTHED TRACK
0610, BETWEEEN LAST TWO
1100 SATELLITE POSITIONS.

Figure 5. Examples of real-time plots: a) real-time DR plot of the ship's track with sattelite fixes information,
b) real-time Sea Beam contours and profile of magnetics data along the ship's track.
provide a smoothed data stream for driving serial LED speed (3) a permanent archive file (csespd) which contains 10 second aver-
displays located on the bridge and in the scientific watchstander's ages for time, course and speed. This file is used primarily as input
area. to the navigation post-processing program, but is also available to
In addition to updating position, course and speed values in various utility programs that use Us information.
saved, drnav outputs include: (4) plots of the ship's position in Mercator projection (Figure 5.a)
(1) archive files DPpfile and GYpfile which contain the raw Doppler are also generated by drnav. Spool files containing plot moves are
and gyro data in a compacted format. These files are not normally created for user defined repetition rates (usually 5, 10, 15 or 30
retained for more than a few days; transfer to permanent magnetic minute intervals) to output the most recent DR position to the Cal-
tape is done only if Doppler current profiler information is desired. comp 965 plotter. As discussed in Section W(c), near real-time Sea
(2) a circular file (drfib which is large enough to hold approximately Beam generated bottom contours can also be included on this plot
the last 12 hours of one minute entries for the ship's time, course, (Fig 5.b). Transit satellite fix positions are automatically plotted
speed and position, This file is used by the Transit satellite fix calcu- along with the DR; or, if desired, fixes may be plotted without the
lation and smooth plot program discussed below, and by utility pro- DR.
grams which provide position data for survey sites, XBT stations, . Unless otherwise indicated, all archive files are periodically
etc. transferred to permanent magnetic tape storage. Two tape copies are

513

made; one is returned to SIO at the end of each cruise leg and the the magnetics data to be plotted, at user defined scales and offsets,
other remains onboard to serve as a backup to the transported tape, along the ship's track (see Figure 5.b).
Transit Satellite Global Positioning System (GPS)
Automatic updating of ship's position is accomplished through Depending upon the ship's location relative to the present GPS
fixes obtained from our Transit satellite (TRANSAT) system. Both satellite constellation, absolute positions to within a few tens of
GPS and Loran are used to supplement TRANSAT, but their current meters can be obtained for periods of four to twelve hours each day.
availability cannot be counted upon for 24 hour coverage in all the Serial (ASCII character format) position fix data are obtained at user
oceanic research areas of interest. selected intervals, usually 10 seconds, from a Trimble-4000AX
In the TRANSAT system, a dual channel (150 and 400 receiver and transferred directly to the VAX via a 1200 baud RS-
megahertz) ITT 5001 receiver performs an analog-to-digital conver- 232C serial line. Acquisition and logging of GPS position fixes are
sion of the incoming 4.6 second Doppler count interval data. At the controlled by a UNIX C-shell script (GPT.Tr-set.up) which
end of a satellite pass, it sends these data to the VAX, through its configures the RS-232C port and starts a UNIX background process
STD/GPIB interface, for position fix calculation. Position fixes may to read the port and save the position fix information in file GPTinpf.
be obtained at intervals varying from tens of minutes to six hours, The drift adjustment program is the only near real-time process that
depending upon the ship's latitude and the number of healthy satel- uses all the GPS fixes made available by the user selected transfer
lites in orbit. A series of fix position computation, drift determina- interval. Terminal display and plotting processes use one minute fix
tion and smoothed track plot programs are executed under control of intervals.
the satloop script. As its name implies, this script runs in a closed When GPS becomes an active world-wide 24 hour system, our
loop invoking the processes in the order required. current DR scheme will give way to GPS navigation. Until then, the
The first program in the loop, satacq.out, checks the current only contribution of GPS to real-time navigation is through drift
condition of the satellite data input file SAinpf. Three possible con- adjustments and North and East components of drift applied to the
ditions and the program's corresponding actions are: ship's DR positions are calculated from Transit satellite fix differ-
1. The file does not exist; satacq.out sleeps for twenty seconds, ences.
wakes up and tries again. A script (DOclosedloop) operates on a five minute cycle to
2. The file exists and its status (time of last update) has changed. update the drift values. This script activates the process
Again, satacq.out sleeps for twenty seconds before performing triDRF.closedloop which reads the last few thousand bytes from
another status check. GPTinpf and averages a number of carefully screened fixes. This
3. The file exists and its status has not changed since the last time it average is fitted to speed lines in the North and East directions for
was tested. This condition indicates that all the data have been gath- standard deviation calculation. If all is well, the new drift adjust-
ered from the satellite's pass and the processing phase can now be ment is written to saved. If no new inputs are available or if the fixes
started. appear to be in error, the values in saved are allowed to slowly revert
The file is then renamed SAinpflunp to clear the way for the back to a smoothed version of recent drifts. A history of drift
next acquisition cycle and its raw data compacted and appended to changes produced are logged into the file TRIlogCL.
the permanent archive file SApfile. Data validity checks are per- Various utility programs have been developed to display and
formed and if any problems am detected, an appropriate error mes- plot GPS fixes from the data in GPTinpf. GPS positions may be
sage is written to the satfixes file. added to an existing DR Mercator plot by gpsplt which can be exe-
The next program in line, satOr, collects, reformats and edits cuted either under operator control or automatically on a preselected
the information from the files SAinpf ttnp and drfil. These data am time cycle. Because the status of the GPS navigation system is still
written to file file.0r; again, any error condition is reported by writing undergoing changes, the corresponding processing software is regu-
a diagnostic message to sa#ixes. File file.Or is used as input to pro- larly updated as new GPS developments are implemented.
gram sat1r which does the preliminary short count calculations. The Loran-C
results are written out to file file.Ir with error reports again directed Due to the lack of world-wide coverage and occasional unrelia-
to satfixes. bility of Loran-C, positions obtained from this source are not
The last program in the satellite fix computation loop is integrated into our real-time navigation. These fixes are used as a
sat2r.out which calculates the ship's position and fix statistics using supplement to the information provided by the Transit satellite and
the information obtained from file.1r. This information is written to GPS systems.
file last-fix and appended to satfixes. Users can obtain the most Loran-C lines-of-position are collected and converted to char-
recent fix from the former or a longer term history of fixes from the acter navigation inform ation by a Trimble-200 receiver, and supplied
latter. The saved file is then updated by this program. to a 1200 baud serial STD interface at approximately ten second
Two other processes are included in the satloop operational intervals. The STD extracts time, latitude, longitude, and transmitter
cycle. The first program (drift) calculates the North and East com- time differences from each fix and, when three fixes have been accu-
ponents of a drift vector by comparing the TRANSAT position with mulated, it passes these parameters to the VAX which stores them in
the DR position at the time the fix was obtained. A smoothing func- file LOinpf. Under stable operating conditions, data are appended to
tion factors in past drift values to minimize errors due to either a bad the file at approximately thirty second intervals. However, if less
TRANSAT fix or bad DR. than three fixes have been accumulated and 45 seconds have elapsed
The second process invoked by sattoop after a successful since the last fix, the available parameters are passed in order to
TRANSAT pass is the real-time smooth plot program sphnn. From empty the buffer.
the fix position information in saved and the course-speed data in A background program torxfr loops waiting for any new entries
drfil, spbnn computes a smoothed ship's track between the current to appear in Winpf. lorxfr reformats the data and writes a one line
and previous fixes. A plot spool file is created and the Calcomp 965 summary to the file loran for each fix position encountered.
then plots this smooth track (Fig. 5.a). User options allow the navi- The loran file serves both as a permanent archive and as a
gation data to be merged with magnetics (file magfil) and profiles of source for utility programs to display and plot Loran-C fix positions.

514

Program lorplt operates either under operator control or automati- This file is used by the smooth plot process for profiling magnetics
cally on a preselected time cycle to add Loran-C positions to an (total field or magnetic field anomaly) along the ship's track (Fig.
existing DR Mercator plot. This feature, together with the GPS plot 5.b).
program gpspit, allow all navigational fixes to be plotted and com- Several utility programs exist to manipulate and display mag-
pared in near real-time. netic data in near real-time. Both terminal graphics and hard copy
IV(c) UNDERWAY DATA printouts of magnetics profile versus time are available on demand.
Gravity Other features permit the scientific watchstander to call up on the ter-
One second pulse counts (1 pulse count is roughly 5 mGal) are minal the latest reading logged, or a series of the logged readings.
supplied by a Bell Aerospace Textron BGM-3 gravity meter to its Sea Beam
STD interface. The STD checks the validity of each 16-bit count, A General Instrument Sea Beam bathymetric survey system
appends the current time to each accepted value and, when ten values installed on the R/V T Washington is also interfaced to the VAX
have been received, transfers the buffer to the VAX Program grvcpy through an STD/GPIB link. This multibeam echo-sounder forms 16
is executed on a five minute cycle to process the one second gravity beams with 2 2/3* angular resolution over an athwartships sector of
"pulses" accumulated during the last time interval in file GVinpf. about 400 to provide cross-track coverage of roughly 3/4 of the water
The input Me is renamed by adding a amp extension to prevent depth. It contains a Data General Eclipse S-130 computer which
conflicts between the logging and processing functions. After controls timing and echo processing within the system and generates
compressing and archiving the pulse counts in permanent file anear-real-time swath of seafloor contours plotted in the direction of
GVpfile, the temporary Me is removed. ship's travel. For every ping cycle (I to 16 sec. intervals depending
Scripts used to monitor the gravity data stream include two upon ocean depth), it also sends up to 16 pairs of depth and cross-
utilities to convert the pulse count data to milligals and print the, track distance together with Eclipse time and ship's heading at
results. Script avGVin prints a series of the latest one second read- transmit time, for a total of 34 16-bit words, to be logged on the
ings taken from GVinpf and script avGVpf prints the average of the VAX via the Z80/STD/GPIB link.
values stored in the permanent file GYpfile over the latest one minute The software in the Z80 microprocessor serves two functions.
interval. These scripts are normally setup to write their outputs to Its primary function is to acquire the data sent by the Eclipse com-
data files which the scientific watchstander can routinely query to puter and, when 3 pings have been received, to transfer them to the
check the integrity of the data collected. VAX. SCO's standard GPIB transfer header information is added to
Other data monitoring procedures include two near real-time this group of three pings. The second function is to receive the ship's
display programs: graphGV and grvchart. Script graphGV produces speed from the VAX and to send it to the Eclipse.
a time series plot of the last five minutes of one second gravity - On the VAX, the Sea Beam data acquisition software requires
counts from GVpfile. Maximum, minimum and arithmetic mean about 10 megabytes of disk storage for the data files and 2 mega-
values are also displayed. gravchart is a more elaborate C program, bytes of storage for the programs. Program GPCON appends each
contributed by scientists from the Lamont-Doherty Geological set of three pings transferred by the GPIB to the Sea Beam data file
Observatory, which computes the free-air gravity anomaly. The pro- SB!npf. These data are then handled by three background processes:
gram reads one minute averages of DR position, course and speed SBstdrw, SBmerge and SBplot.
from the common data file saved, and one second pulse counts from . Program SBstdtrw (Sea Beam standard bus to raw file program)
GVpfile. An estimate of the observed gravity is then calculate 'd, the is executed on a 40 second cycle to process the data accumulated in
F,6tvos correction is applied, and an estimate of the normal gravity SBinpf during the last time interval. The first processing step is to
and of the free-air anomaly are computed. The observed gravity, its rename the file to SBinpftmp and append its contents to file SBpfile.
F,dtvos corrected value and the free-air anomaly obtained in this This file is a raw backup in case of problems ftirther on in the real-
fashion are displayed symbolically on an Okidata printer. Such a time processing. The next operations include adding the current
near real-time display is useful during gravity surveys at sea, as the UNIX time to the data in SBinpfttnp and writing these data into both
free-air anomaly displayed can be compared directly with the free-air the SBpfile file and a circular file called raw.seabeam. The
anomaly values shown on existing gravity maps for the oceans. raw.seabeam file holds the data until navigation for the same time
Magnetics period (one second resolution) is available.
Binary Coded Decimal (BCD) magnetic field readings are At the start of every 40 second cycle of processing new data,
obtained at six second intervals from a EG&G GeoMetrics G-801/3 SBstdtrw also finds the center beam and the closest return. (first
magnetometer. The STD software drives a six segment Light Emit- arrival) and writes these two values together with UNIX time to a file
ting Diode (LED) device to display the most recent reading and, called depfil which can accomodate 299 such records. UNIX time is
when ten readings have been collected, transfers the buffered data to stored in a 32 bit integer word and isthe number of seconds elapsed
the VAX. The processing and display of magnetic data is a routine since Jan 1, 1970. These data are saved to be used by other programs
procedure carried out on all cruise legs when the EG&G GeoMetrics such as the sound source synchronizer which schedules the firing rate
magnetometer is in operation. A real-time display, driven by the of the Sea Beam system, the seismic and the 3.5 kHz subbottom
STD software, shows the latest six second total field reading received profilers [2].
to allow the computer technician to monitor the data stream to the The second background real-time job is then started with pro-
VAX. gram SBrItnerge which sleeps several minutes between passes of the
Data compression and archiving functions are performed by data files. It merges the Sea Beam data found in the raw.seabeam
program cpymag which is also executed on a five minute time cycle. file with the dead reckoning navigation found in file drfil, written by
This process follows our standard procedure of renaming the input program drnav. The merged Sea Beam data is written into file
Me by adding a tmp extension to it; in this case MGinpf becomes rtin.seabewn in the SIO standard merge format of 50 16-bit words.
MG1npf.t?np. The BCD total field readings are then compressed and This file is a 5 Mbyte circular file and its pointers are kept in file
archived in the permanent file MGpfile and the temporary Me is seabewn.rtcommon.
removed. Program cpymag also maintains an additional circular Me, To output the ship's speed to the Eclipse, Ortmerge creates a
magfil, containing magnetic readings and their associated times. file GPouO which contains the ship's,speed converted to a binary

515

integer in the range of 0-1023 hex corresponding to speeds of 0 to 15 Program oneshot (or twoshot for channel 2) can be executed to
knots. Note that due to a problem in the Eclipse software the speed display the last recorded event in near real-time. The file Sldspl is
sent by the VAX is limited to a lower bound of 1 knot. When used to produce a time series display of the recorded signal. By
GPCON finds this file it transfers it to the Z80 processor which then monitoring these plots, improper hardware setup or equipment mal-
sends the 10 bits of speed data to the digital-to-analog converter functions can be detected and quickly corrected.
hardware which provide ship's speed to the Eclipse computer. XBT (eXpendable BathyThermograph)
The last of the three background real-time jobs is handled by Temperature measurements from a Sippican MIK-9 XBT sys-
program SBplot. This program reads the Sea Beam data from tem are transferred through Sippican's GPIB interface to the
rtin.seabeam, contours it, and sends the output to the real-time Cal- appropriate STD where the data are buffered and passed along to the
comp 965 plotter (Fig. 5.b). SBplot gets the file pointers, contour VAX.
interval, and labeling information from the file seabeam. rt. common. The da ta once acquired are stored on disk (about 4000 bytes
The contents of seabeam. rt. common can be monitored with program per T-4 drop) and displayed in near real-time on the operator's termi-
Cbeam. This allows the operator to use a CRT terminal to set the nal, with an optional hard copy printout available. The time needed
contour interval for the flatbed plotter display, monitor the amount of for an XBT drop depends on the type of probe used, a typical T-4
space left in the holding file and turn contour plotting on and off. probe requires about 80 seconds to acquire the data. The actual cpu
In order to archive the Sea Beam data it must be extracted from usage is only several seconds during this 80 second interval.
the rim.seabeam file approximately every other day. This rate To do an XBT drop, program xbt is started on the VAX. This
depends on the ping rate, but them is enough room to last about 24 program communicates with program xbt.src which is running in the
hours at a 2 second ping rate. Program sbcopy copies the data from Z80. xbtsrc establishes that the Sippican MK-9 interface is opera-
the circular,file rtm.seabeam and writes them into a file to be tional and relays this information back to xbt. The user informs the
archived in the SIO merge format. MK-9 of the type of probe being used through the the VAX to
Seismics STDAZ80 link. Once the probe type has been established no more
The seismic system is used for acquiring seismic reflection or computer operator intervention is needed. The MK-9 interface waits
refraction signals received by hydrophones or sonobuoys. An inter- for data to become available and then transfers them to the VAX by
face designed at Scripps controls cycle time, delay time and acquisi- way of the STD. As data are collected, they are displayed on a
tion time for each event to be sampled. This interface applies the TEK-4010 compatible terminal in a graphic form. This is just a
appropriate filters to the incoming signal, digitizes it and passes it to rough look at the data (uncorrected for thermistor non-linearity and
the VAX via its STD/GPIB link. Parallel systems are available for AID drift). At the end of the data acquisition phase, program xbt
simultaneous acquisition and processing of two input channels. Sam- reformats the raw data, adding the ship's position, ship's name, time
ple and delay times are controlled by switch settings on the and date. It then writes these data to a disk file in the XBTDATA
operator's control panel with 250 microseconds as the shortest sam- directory.
ple interval available. The STD data buffers accept a maximum of A shell script (DOxbr) assists the operator during the XBT data
sixteen kilobytes for each shot or event; therefore, to prevent data acquisition process. It runs all the commands needed to acquire, pro-
overflow, the operator must control the balance between sample rate cess, plot and archive the data. It starts by creating a new filename,
and event duration. The following paragraphs describe the process- unique for this leg. It then runs xbt,.xbt1soth, xbtplot, bathymessage,
ing applied to channel 1; except for program and file names ("J." and xbt-Smpl[ndx. It also produces a hard copy output on a printer
replaced by "2") the processing for channel 2 is identical. attached to the terminal running the script.
A l6kbyte buffer containing the sixteen bit data words is Two other scripts are used to help maintain order in the data
passed to the VAX at the conclusion of each event. The buffer directory. The script DObegincrulse which initializes the sequential
header contains a shot sequence count and a configuration word with number and prompts for cruise name, and the script DOendcruise
sample rate, delay time and sample time. Other words with urnique which nms program xbt-nodc converting all xbt drops to NODC for-
bit patterns are used to flag beginning and end of shot conditions. mat and which reminds the..operator to copy the data to tape.
Data words contain twelve bits of digital information and a four bit
scale value representing the gain applied to the value in the other V. CONCLUSIONS
twelve bits.
A background process (seisSl) is activated every ten seconds to The system described in this paper has been in continuous
check for the existence of file Slinpf and, if found, rename it operation since February 1984. Hardware reliability has been excel-
lent with only one significant failure on the R/V T. Washington's
Slinpf.imp to ready the system for the next event. The 16kbyte "offline" computer. The system was back on-line within 48 hours
buffer represents a maximum shot sample size of 8000 words (e.g. after a faulty crystal oscillator had been detected and replaced; none
an eight second sample window at a one kHz sampling rate). As the of the data acquisition capabilities were affected by this interruption.
actual number of samples may not fill the full 16kbyte buffer passed, Had the failure been in the "ordine" system, the other computer
a header word indicates the number of samples present in the buffer would have been switched in by changing the STD/GPIB cable (no
to process. power down/up required) and starting up the appropriate data
A "shot record" of appropriate length is constructed by unpack- acquisition processes.
ing each 16 bit sample (four gain plus 12 data) and converting it to
its equivalent 32 bit integer SEG-Y* format. Each shot record has a Once the system is booted and the data acquisition processes
240 byte binary header containing sequence count, shot are started, only routine system monitoring (primarily archiving of
configuration, date and time prepended to it. This buffer is used to data to magnetic tape, purging the archived files and checking input
overwrite the near real-time plot data input Me Sldspl and is also devices) is required. This job, together with program development
appended to the permanent archive file SEGY]file. and hardware maintenance, is normally handled at sea by a techni-
cian from the Shipboard Computer Group. Collection of navigation
*SEG-Y is the data exchange magnetic tape format proposed by the Society of Ex- along with a subset of underway data has been successfully done on
ploration Geophysicists Technical Standards Comn-tittee. the R/V Melville without a trained computer operator in attendance.

516

Software command scripts handle the necessary details to provide a 4. Integration of gravity and Sea Beam center beam depth profiles
designated operator (usually the ship's Resident Technician) with the into the smooth plot system. These features should be fairly
tools for activating and monitoring the data acquisition processes. straight-forward and easy to implement but, as mentioned above,
Although little concern was given, during system development, will require ship time for final testing and debugging.
to the eventual transport of this software to other computing Improving the computational and inter-communication capabil-
machines; use of the Fortran.77 and C programming languages ities of our existing mainframe machines is another area of high
should simplify this task if a switch to some other UNIX based sys- priority. Personal compute 'rs (PCs) have been routinely used at sea
tem is desired. The only current requirement of the target machine is for some time with access to the VAX- 11/730 real-time data bases
that it provide hardware and software support for the GPIB data through primitive RS-232C file transfer procedures.
communication protocol. With appropriate acquisition device inter- On a cruise of the R/V T. Washington in the spring of 1987, an
face modifications, the STD/GPIB link could be replaced by RS- Alliant FX/l cornputer* was attached to the VAX-l 1/730's through
232C communication on a mainframe computer able to support the an Ethernet link. Prior to this installation, resource sharing between
overhead of the serial transmission of these data. the two VAX-l 11730's had been accomplished by use of magnetic
Most of the current VAX- 11/730 - BSD 4.3 real-time data tape, the UNIX utility "uucp" and, a serial data network scheme
acquisition software was adapted from the IBM 1800 system "slipnet." The purchase of Ethernet hardware and a simple
developed during the past 20 years. The increased power and flexi- reconfiguration of the UNIX system provided true high speed net-
bility of these new computers have allowed us to continuously working between the three mainframe computers.
improve and fine tune our operations in cautious and slow incre- This successful mating demonstrated the versatility of a
ments. Caution is required, as in any software modification project, networked system where each machine can be dedicated to a portion
to insure the new changes do not affect any of the existing features. of the tasks required to be performed. In our case, the "online"
The sometimes painfully slow rate of system refining is due to the VAX-11/730 handled only the data acquisition and near real-time
small portion of time at sea that can be dedicated to hardware and processing; the "offline" machine served as backup and for software
software testing. Ship time is expensive and the scientists whose development; while the FX/l provided a powerful time shared com-
grants pay for that time do not want to donate it to non-scientific pur- puting environment.
poses. The old adage "if it's not broke, don't fix it" prevails on most
cruises. ACKNOWLEDGEMENTS
Hardware and software development and testing has, until Most of the software development work reported here has been
recently, been carried out on our shore based VAX-1 1/750, This has funded by SIO institutional funds. Sea Beam related work was
not been a very satisfactory configuration because; 1) the system funded in part through the Accelerated Research Initiative on multi-
must remain "up" for use by Scripps' radio station WAID, the beam systems sponsored by the Office of Naval Research, Contract #
scientific and administrative communities and; 2) subtle differences N00014-85-G-0104. J.L. Abbott directed and was actively involved
exist between VAX-11/730 and VAX-11/750 architectures which in the development work performed by the Shipboard Computer
have affected the operation of some software packages. Recently, Group. Under his guidance, the main contributors to software
the availability of used computer equipment at low cost has made it developments reported here were G. Bouchard, J.S. Charters, R.L.
possible to put together a shore based VAX-l 1/730 and STD/GPIB Moe, and J.M. Moore. P. Wessel and W.H.F. Smith at Lamont-
system to better emulate the shipboard environment. Doherty Geological Observatory contributed the grvchart program
Software modification and enhancement projects currently mentioned in Section IV(c) while at sea on the R/V T. Washington.
under consideration include: C. de Moustier brought this paper to publishable form. E. Ford for-
1. Automate the STD and VAX-11/730 clock corrections using matted the text and J. Griffith did the art work.
either the available high precision Transit or GPS satellite times.
The ability to detect and correct clock offsets greater than some REFERENCES
defined threshold should be added.
2. Smoothing out the interaction between GPS and the dead-reckon (1] Abbott, J. L., S. M. Smith, J. S. Charters, P. G. Downes, T.
program will be a high priority item as the GPS system approaches Hylas, R. L. Moe, J. M. Moore and D. V. Stuber, Scripps Seagoing
full 24 hour coverage. More effort needs to be spent to determine Computer Centers: Real-Time Data Acquisition and Processing,
the best combinations of data input editing and processing. IEEE Proceedings of the Fourth Working Symposium on Oceano-
3. Providing our own time stamp of Loran data so as not to rely on graphic Data Systems, pp. 123-129, February 1986.
the information provided by the receiver. Invalid times, due to tem-
porary receiver malfunctions or improper initialization, sometimes [2] Phillips, J. C., J. L. Abbott and C. de Moustier Multiple sound
make it difficult to correlate Loran positions with other navigational source synchronizer for seafloor surveying Proc. Offshore Technol-
and underwav data. ogy Conference, OTC #5867, Houston, Texas, May 2-5, 1988.

*On loan from Alliant Computer Systems Corporation. Acton Massachusetts.

517

A COMMON XBT/PERSONAL COMPUTER INTERFACE

STAFF*

Oregon State University
College of Oceanography
Oceanography Admin. Bldg. 104
Corvallis, OR 97331-5503

ABSTRACT The Apple II computer was one of the first to approach a
solution to the interface problem. It was provided with a
Ile use of expendable probes, such as expandable bathy- number of 1/0 slots" for peripheral devices such as floppy
thermographs (XBTs) as temperature profilers, in the disks. One of the early peripherals was an RS-232 interface
oceanographic research and operational data gathering (intended for modem access). A unique characteristic was a
community is well established. Thousands of XBT probes are memory chip in the interface which provided the necessary
cast annually to contribute to a broad suite of national information to the resident software (frequently BASIC) to
oceanographic research programs and the National Weather provide easy peripheral access.
Service.
Now, most small computers, including those of the "IBM PC-
This paper will report on the improved system design AT-XT and Clone" group provide an RS-232 port as a
flexibility for a digital XBT electronic interface created by the standard feature complete with high-level language access.
simplified serial communications and hardware features of RS- Most of the popular languages, including BASIC, C,
232 parts on a variety of personal computers. The application PASCAL and the Eke have either resident commands or readily
of this design philosophy is also applicable to other marine available subroutines to access the RS-232 port.
instrumentation needs.
But why the interest in RS-232? Surely there are other
1. VIRTUAL INSTRUMENTS AND RS-232 schemes for moving data between an interface and a computer.
The following is a review of some of the data link options:
Since the beginnings of small computers, the advantages of
applying the power of these computers to instrumentation have 1. Processor bus - probably the highest data rate of all choices
been apparent. Gradually, the concept of using a small is available when going directly onto the data bus. This
computer together with a specialized interface and software to method usually requires machine language code to support it.
construct an instrument has grown and acquired a name: It is the method used by "peripheral cards" such as analog-to-
"virtual instrument" (1). This name applies particularly to digital (A/D) converters. On the negative side, this method is
equipment where software has replaced major functions processor-specific and a device built for one family of
previously accomplished by hardware. This concept has computers will usually not operate with a different one; also,
particular import for oceanography because the numerical many computers (including "lap tops") do not provide access
quantities of equipment tends to be small in comparison to to the processor bus. Very little physical separation is allowed
many other fields of endeavour and the demands on the between a computer and a device on its processor bus (perhaps
equipment tends to be rather varied. The virtual instrument can less than half a meter).
frequently satisfy these specialized requirements better and/or
more inexpensively than conventional hardware-only 2. IEEE-488/GHIB/HPIB - IEEE-488 and GPIB (General
equipmem Purpose Interface Bus) are synonyms for an 8-bit parallel
communications bus. HPIB (Hewlett-Packard Interface Bus)
One of the frequent frustrations of designing virtual was the precursor to IEEE-488 and differs in a few respects
instruments with early small computers was the problem of from IEEE-488. In many applications, the two are
connecting the specialized interface to the computer. A interchangeable. Data transfer rates up to one-half megabyte
Commodore "PET" computer had entirely different per second are possible. The interface has a reputation for
requirements from a Radio Shack "IRS-80". Frequently, the being difficult to implement for both hardware and software.
processor bus was the only connection available and machine Hardware design has been simplified to some degree with
code was almost always required to access devices on that bus. specialized integrated circuits. Relatively few computers
Different processors made the machine code and hardware provide a "native" IEEE-488 interface. The simplest way to
different for each computer. Even connectors were different. implement the computer-side of the interface is to purchase a
'Mere was no such thing as a "standard interface". commercial product; this Emits implementation to computers
with "expansion slots". Separation between computer and
devices on the IEEE-488 bus is limited to 20 meters without a
Contributions by: repeater.
Dr. James Wagner
Dr. Roderick Mesecar, Head 3. HPIL - (Hewlett-Packard Instrumentation Loop) is a serial
Technical Planning and Development Group loop initially intended to provide a simple interface to pocket

CH2585-8/88/0000. 518 $1 @1988 IEEE

calculators. Outside of a few Hewlett-Packard calculators and
computers, this system does not appear to be widely used.
4. LAN - (Local Area Network) comes in a variety of styles Sigral
including Ethernet, STARLan, and others. Data rates can Serisor(s) Conditiordrg AID U Host
exceed 1 megabittsec over distances up to 500 meters (2). If Cornput!r
simple communication between two only points is needed, the
hardware overhead can be rather high. Software, especially at
the computer side of the connection, is commercially available.
Access is generally limited to computers with "expansion ................... ControUer
slots" for the LAN interface cards.

5. RS-232 - this is a duplex serial system with control lines.
It is best suited from point-to-point communications between a
single pair of devices. Usually, each end of an RS-232 line
requires receivers and transmitters which convert between 5
volt logic levels and RS-232 levels. Each end also requires
UART (Universal Asyncronous Receiver-Transmitter). These Figure 1. Virtual Instrument Block Diagram
devices are readily available and are frequently included in
many small computers. Data rates are commonly limited to
19.2 kilobaud (about 2000 8-bit characters per second). 'Me Note that there is no requirement that the specialized interface
maximum recommended length of an RS-232 link is 15 meters be connected a computer at all times. A case in point is the EG
with longer distances permitted at lower data rates or reduced & G Inc. Vector Averaging Current Meter. These instruments
load capacitance. When an RS-232 port is included in a are used by sending sampling parameters to the meter via an
computer, software to access the port is usually included in the RS-232 link, then disconnecting the meter and placing it into
operating system. use. During measurement, data is stored in a local memory.
At some later time, the meter is reconnected to the computer
6. Current loop - here is a system very much like RS-232. It and data is transferred from the meter's memory to the
differs only in the electrical characteristics of the link. It is a computer. Processing and display then occurs within the
simplex serial system using current rather than voltage for computer. This is again a virtual instrument but one in which
signaling. No control lines are used. Maximum data rates are the data collection is carried out disconnected from the
lower than RS-232 and are frequently in the 1.2 kilobaud.area. computer.
Noise immunity is good.
11. AN XBTE*4TERFACE
7. RS-422 - this system is also very much like RS-232 and
differs only in the electrical characteristics of the link. It is a An example of a virtual instrument for oceanographic
duplex serial system. It utilizes differential signal lines (for instrumentation which will be considered in more detail is an
noise immunity) which can default to RS-232 when connected XBT (eXpendable Bathy-Thermograph) interface designed at
in a certain way. Protocol is the same as RS-232. Interface Oregon State University by the Technical Planning and
cards for some computers are available. Development Group of the College of Oceanography.
With all of these possibilities, how is a choice to be made? If This interface is made, to operate with the standard XBT probe
there is only one to be produced,,then the choice will probably and launcher produced by Sippican Corporation. It provides
be. what-ever fits equipment on hand while meeting other sensor excitation and analog signal conditioning appropriate to
requirements such as data rate. But if there are to be many the sensor. A digitizer converts analog voltage signals into
built for wide distribution, the answer is probably RS-232. digital numbers. Sample rate and other characteristics are
The reason is fairly simple: implementation cost is low and it is programmed into the interface by the host computer. Data is
supported in a wide variety small computers. While computer stored in the interface and transferred to the host after the
programs to support a specialized interface on an RS-232 port measurements have been completed. An 8-bit microprocessor-
may not be transportable at the compiled-code level, programs based controller containing an EPROM interprets commands
at the higher-level source codes usually are transportable with from the host and schedules operation of the interface. The
most of the required changes occurring in instructions which interface is fairly simple, fabrication costs are low and the
generate displays. design was carried out quite quickly.

The impact of these factors on oceanographic instrumentation For maximum portability to ease placement of XBT equipment
is fairly significant. It leads to a virtual instrument which uses aboard ships-of-opportunity, the interface is tailored to operate'
a computer and an RS-232 link to connect to a specialized with some of the newer "lap-top" computers of the "PC-
interface. The interface usually contains some analog sensor compatible" type (Toshiba, Zenith, and others). It occupies an
excitation and signal conditioning, an analog-to-digital enclosure 12" square and 3" thick which is close to the size of
converter, a serial data port, and some type of controller. The most lap-tops. Figure 2 shows the XBT interface with a lap-
controller may be implemented as a simple state-machine if the top computer. Operation is possible with any computer having
required tasks are very simple. More likely, however, one of an RS-232 port, however. Communication between the host
the simpler 8-bit microprocessors will be used; for simplicity and the interface is via an RS-232 link for the reasons
and ease of implementation, the choice is often a previously outlined. Virtual instrument software is written in
microprocessor which contains its own ultra-violet erasable BASIC which is fully adequate in execution speed with the
read-only memory (EPROM). Other choices might be made if possible exception of "bathy message" processing. The bulk
low power is special concern. The almost universal block- of the program was written in about five days with some fine
diagram of such a virtual instrument is shown in Figure 1. tuning requiring additional time later.
qeD:
.......... L

519

I `I-IMOM!'n
There are other controlling characters which may be sent
between the host computer and the interface. For example,
the host computer can send an "M" (memory status) to the
interface. If the interface has unread data, it responds with "P
(full); if there is no unread data, the response is "E" (empty).
Similarly, the host computer may send characters to set the
sample rate, change the serial link baud rate, or reset the
interface controller. The interface also provides a software-
M" N
controlled data switch allowing one host computer serial port
to drive either the XBT interface or serial satellite transmitter
J
J."N",2,
H F.J@ port on the interface.

ME In a virtual instrument, there are two sets of specifications.
One includes the performance of the specialized interface.
These specifications establish boundaries to the system
performance which programming wizardry is not likely to
circumvent. The second set outlines the system performance
utilizing provided software. Table I gives interface
specifications for the XBT interface while Table H gives the
U specifications of the complete XBT systen-L
Figure 2. XBT Interface with lap-top computer Table 1 - XBT Interface Specification
Design of a virtual instrument requires careful consideration of * Sample Rate: 1, 2, or 4 per second programmable
how tasks are to be partitioned between the host computer and * Sample capacity: 1009 samples
the specialized interface. The design of the XBT interface was * Resolution: 0.05' Centigrade
no exception. An example is the test which the interface does e Accuracy: 0.10* Centigrade
prior to each probe drop. The numbers obtained help gauge
the operation of the interface and are used in the conversion of o Serial host interface: RS-232
the A/D output to temperature. A decision had to be made * Serial baud rate: 300 to 9600 baud, programmable
concerning which unit, host computer or interface, was to
determine acceptability of test results. That responsibility was o Serial data control: XONIXOFF
assigned to the host computer so that the test criteria could be & Satellite port: switchable by host
more flexible than if contained in the interface controller's 9 Satellite baud rate: 300 to 9600 baud, programmable
EPROM.
9 Interface testing: built in
Another aspect of virtual instrument design is the protocol * Launch detection: built in
chosen for exchanges between the host computer and interface.
Brevity and lack of ambiguity ease decoding at both ends.
Completeness insures that all probably situations can be
handled. The interchange protocol between the host computer Table H - XBT System Specifications
and the interface is best described by following the sequence of
operations during a probe launch. Normally, the empty probe
canister is kept in the launcher until just before the next drop. e Host: Zenith or Toshiba "lap-top"
The interface senses when the canister is removed and sends e Operating Sys tem: MS-DOS 3.0
an ASCII "C" to the host computer. The interface
automatically begins the self-test procedure; when the test is * Language: BASIC
finished, the interface sends the A/D output values from each * Floppy disk drives: dual 3.5" required
test step to the host computer. The AID is a 12-bit converter * Data storage: all samples stored as temperature and
and each output value is sent as three hexidecimal values in depth
ASCII. The host computer verifies the test values and if they
are acceptable, sends an ASCII "G" (go) to the interface. At 0 Data display: temperature vs depth graph (operator
the same time, an operator message indicates readiness for a optional)
new probe canister and the next launch. When the probe is e Additional processing: significant profile points
launched, the interface sends an ASCII "L" (launched) to the (operator optional)
host computer and begins sampling at the rate specified earlier e Bathy message: screen display and disk storage
by the host computer. Sampling continues until the interface (operator optional)
either reached the end of available memory space or receives
from the host computer an ASCII "E" (end). When the host * Satellite message: Suited to user requirements.
computer is ready for the data, it sends an ASCII "D" (data) to
the interface. 'Me interface responds by returning the data
using three hexidecimal characters (sent in ASCII) per sample.
K necessary, the data flow can be interrupted by the host
computer by sending "XOFF" (ASCII ctl-S or DC3) and Host computer software handles operator displays, movement
resumed by sending "XON" (ASCH cd-Q or DC1). The host of control information to the interface, movement of data from
computer processes the data into the desired form, then waits the interface, conversion of data from arbitrary A/D units into
for the signal from the interface that the one empty canister has temperature units and from saffi ple time into depth, storage, of
been removed for the beginning of the next launch. depth and temperature data, graphic display of temperature vs

520

depth, compression of a temperature-depth data set into
standard World Meteorological Organization (WMO) bathy-
message, formatting the bathy-message (if desired) for satellite
transmission, and, finally, (if desired) movement of
compressed data to a satellite transmitter. Since the host
computer software is accessible to any user who knows
BASIC, it is easy to change, for example, operator interface
displays, processing algorithms or data formatting rules.
Additional functions may be readily added. These might
include temperature limit alarms, mixed layer depth
determination, or identification of profiles leading to certain
sound transmission characteristics.

Note that these functions are independent of the host computer,
operating system, or programming language. While the
provided implementation runs under MS-DOS 3.0 on Zenith or
Toshiba lap-tops, many of the functions have also been
implemented on an HP-85.

While few details of the interface have been provided, it should
be sufficient to say that the design for sensor excitation and
signal processing is taken from one of our existing designs (3)
and includes about 8 analog integrated circuits (I.C.s) and
some discrete parts. The digitizer is a single I.C. and includes
some discrete parts. Two digital I.C.s are used to control the
analog circuitry. The interface controller is an 8748
microprocessor with an additional 2K of memory and a few
logic I.C.s. The RS-232 port includes a UART and a receiver-
transmitter.

Four of the systems described have been constructed. Several
are undergoing initial sea-tests at the time of preparation of this
paper. Initial results appear to indicate that the specifications
of Table I have been met. Two of the interfaces have been
built with battery-backup so that operation is possible while
separated from the host computer and line power.

III. SUMMARY

The idea of a virtual instrument has been explored. The
problem of how to move data between a host computer and the
specialized interface was discussa If use with the greatest
variety of host computers is a goal, then RS-232 would seem
to be the interconnection method of choice. An XBT interface
is described as an example of how these ideas can be applied to
the unique requirements of oceanographic instrumentation.

REFERENCES

1. Fitzgerald, K. Instrumentation, IEEE Spectrum 25 (1),
Jan 1988, 47-49.
2. Stix, G. Telephone Wiring: a conduit for networking
standards, IEEE Spectrum 25 (6), June 1988, 38-41.
3. Mesecar, R. and J. Wagner. An XBT Digital Recording
and Display System. OCEANS 1979 Conference,
September 1979, 598-601.

521

HYDROBALL - A NEW EXPENDABLE:
USES AND ISSUES

H. Tremblay

NewTech Instruments Limited
P.O. Box 13635, Station A
St. John's, NF, Canada
A1B4G1

ABSTRACT i - The Eulerian method - measures simultaneously the
direction and speed at fixed points, and;
Early in 1988 the HYDROBALL Ocean Current Profiling System,
under development since 1985, was introduced to the ii - The Lagrangian method - measures the water move-
marketplace by NewTech Instruments Ltd. The HYDROBALL ments by tracing a path of a water particle over a longer
System is a transportable, easy to operate, low-cost expendable time interval.
current and temperature profiling system which operates from
stationary platforms and moving vessels in water depths up to The commercial products existing at the time used one of
500m. two technologies to derive current speed and direction
The HYDROBALL System uses an expendable Lagrangian throughout the water column: cloppler measurements of
drifter, spherical in shape, which houses an acoustic transducer, acoustic waves reflected from particles suspended in the
pressure sensor, thermistor and timing electronics. The un- column (RD Instruments Doppler Current Profiling Sys-
tethered, free-falling drifter transmits data hydroacoustically as it tem), and measurements of the electrical current generated
descends through the water column. The changing range and bythe motion of seawaterthrough the earth's magneticfield
depth of the drifter, aswell as its azimuth with respectto the ship's (Sippican Ocean Systems Inc. Current Profiling System).
heading, are detected by an ultra-short-baseline phased array
hydrophone, and converted to digital data by a signal analyzer. After researching these and other potential technologies, it
Adedicated microcomputer calculates flow velocity components, was decided to use a direct, physical method of current
as well as temperatures, at regular depths throughout the drop. profiling - monitoring the movement of a free falling Lagran-
gian drifter in three dimensions, and using the resultant data
to extrapolate current speed and direction along the path
1. INTRODUCTION of descent.

The potential benefit of absolute real time current profiles It was determined early in the development cycle that
used in support of offshore operational activities has only hydroacoustics would be a reliable and cost-effective
recently been identified as a long term growth market for method of tracking the path of the probe. The use of
manufacturers of oceanographic equipment [1]. The sur- hydroacoustics would also allow the transmission of sen-
vey identified several sectors which could derive tangible sor data from the descending probe.
benefit from absolute current profiles obtained prior to un-
dertaking sub-surface activities. The applications identified The completed system consists of an expendable probe
included uses in manned diving and remotely operated (the HYDROBALL), a hydrophone, and shipboard
vehicle operations, drilling unit positioning, dredging and electronics. The data acquisition process is shown in
dump site management, and tactical ice management Figure 1.
programs in support of offshore drilling.
2. PRINCIPLE OF OPERATION
In 1985 National Petroleum and Marine Consultants Ltd.,
of St. John's, Newfoundland, Canada, undertook the The HYDROBALL System operates on the principle of cal-
development of a current profiling product which would culating a current velocity profile from the dynamic
meet the requirements identified. The objective was to response of a freely falling object [3,4,5]. The development
develop a cost-effective, easy to use technology that could of the product began with the construction and verification
reliably measure significant currents in near real time. of a numerical model to determine the dynamic response
of a spherical probe of known physical characteristics (Fig.
In general there are two methods of measuring water move- 2) [6]. Both theoretical and field research demonstrated a
ments (currents) [2]: predictable dynamic response of the HYDROBALL probe.

CH2585-8/88/0000- 522 $1 @1988 IEEE

rate of 1.024 Hz. Because the HYDROBALL probe is
Analog to calibrated to fall at a constant vertical velocity of ap-
Digital proximately 0.36 m/sec., the effective resolution of the sys-
Circuit tem is approximately three measurements per metre.

Navigational information, including ship's heading and
+ speed (for moving vessels), is input manually by the
operator. Corrections for changes in heading during the
descent of the probe are input through a direct electronic
Probe connection to the vessel's gyrocompass. In practice,
4 moving vessel profiles are undertaken at a speed of ap-
proximately two knots, with vessel heading and speed
maintained throughout the profiling operation. An addition-
FIGURE 1. The HYDROBALL Current Profiling System hydroacoustic al verification of the profile data is established by continu-
data acquisition system. ing to collect data from the probe for several minutes after
it has reached the seabed.
The results of this theoretical study were used to guide the During the profiling operation the incoming data is dis-
development of Inverse Dynamics Analysis (IDA) algo- played to the operator in numerical form in real-time. The
rithms. The IDA algorithms are used to express the incoming data is processed by the IDA algorithms to
hydrodynamic relationship between the response of the
sphere to flow velocities in a given horizontal field, and the produce current and temperature profiles in near-real-time.
actual current speed. The data is stored to disk for later analysis, printing or plot-
ting.
Tracking the expendable sphere requires the resolution of
its position in 3D space at regular time intervals as it des- The HYDROBALL System is designed to operate in water
cends through the water column, The technique involves depths of up to 500m., with a maximum range between the
the determination of probe range, depth and azimuth, rela-. probe and the hydrophone of 1500m.
tive to the hydrophone. The completed system uses three
techniques to determine positional information. 3. TESTING

The range of the-probe is determined by measurement of Throughout the development the HYDROBALL System
the time delay of incoming probe signals due to the speed was tank tested at the Acoustics Laboratory facilities at
of sound in water. A precision electronic. clock in the ship- Memorial University of Newfoundland and at the New-
board electronics package is synchronized with a similar foundland and Labrador Institute of Fisheries and Marine
clock in the probe prior to deployment. Astherangeofthe
probe from the hydrophone increases, the time delay of in-
coming signals versus the synchronized clock pulses is
measured and the range calculated.

The depth of the probe is measured by a pressure
transducer housed in the expendable. Data from the
transducer is converted to digital form, and transmitted
using pulse position modulation of the hydroacoustic sig-
nal.

The measurement of the azimuth of the probe relative to
the hydrophone is made possible by ultra-short-baseline
phased array hydrophone. The measurement of small
changes in the phase of incoming signals as they pass over
the three-element array allows an accurate determination A,
of azimuth.

In addition to establishing probe position, the system also
recognizes and decodes data from a thermistor built into
-V
the probe, and uses the incoming data to produce a simul-
taneous temperature profile.

The probe emits range, depth and temperature signals at
three distinct frequencies (27, 26, 25 kHz.) at a repetition FIGURE 2. The HYDROBALL expendable probe.

523

HYDROBALL SIPPICAN DOPPLER
TWO VELOCITY COMPONENT ANALYSIS
SHIP: COLUMBUS ISELIN FUZ NAME: MIAM13
LAN: 25.46.5.N LONG: 80,03.4.W 12-AUG-88 14:58:23

200-- LEGEND

180-- HYDROBALL N-S

160-- SIPPIOM N-S

....... DOPPLER N-S
140--

HYDRORA-1 W-E

120--
.......... SIPPICAN W-E

rn 100--
--- DOPPLER W-E

so--

60--

40--

20 . . .........

.............. ...........
0.

-20 i i i i i
10 165
DEPTH
M

FIGURE3. HYDROBALL Sippican /Doppler field data. was felt that the most practical method of demonstrating
the accuracy and function of the system would be to ar-
Electronics. These facilities were used to determine the range comparative, simultaneous trials with two other near-
hydroacoustic properties of a variety of transducer real-time profiling systems.
materials, constructions and assemblies for both the
hydrophone and the HYDROBALL probe. On August 12, 1988, NewTech Instruments Ltd., the
manufacturers of the HYDROBALL System, arranged for
In the absence of a suitably large testing facility that could the charter of the research vessel ORV Columbus Iselin
simulate current shear and representative water column from the University of Miami, to obtain comparative current
depths, open-water testing was conducted. A prototype measurements in the Gulf Stream. The vessel was out-
hydrophone and a supply of probes was manufactured and fitted with an acoustic Doppler Current Profiling System
the system was installed on an inshore fishing vessel for manufactured by RD Instruments Ltd., San Diego, CA, an
field trials. The field program also included the use of MK-10 current profiling system manufactured by Sippican
moored current meters and a three-point shore-based Ocean Systems Inc., Marion, MA, and the HYDROBALL
navigation system as a means of data verification. In sub- System. Third-party personnel were responsible for the
sequent trials the system was installed on an offshore operation of the doppler system (University of Miami per-
anchor handling supply vessel, and tested on the Grand sonnel) and the XCP system (Shell Development Com-
Banks offshore Eastern Canada. I pany).

Although these tests provided some comparative data, al- At the time of preparation of this paper the data from the
lowed the construction of mathematically generated data three systems was still undergoing detailed numerical
sets for software testing, and ensured that the measure- analysis; however, a preliminary examination of the results
X

ment methods and equipment were functioning properly, it (Fig. 3) suggests that the three instruments were produc-

524

ing data sets with a high degree of correlation. A tempera- extrapolate current speed and direction from the motions
ture profile from the same HYDROBALL probe is shown in of the drifter.
Figure 4.
The HYDROBALL System has been demonstrated as a
A full report on the testing results will be available from the practical alternative to profilers using other technologies.
author at time of publication.
5. ACKNOWLEDGEMENTS

Wer TeNperdture ('--- C.) The development of the HYDROBALL System was funded
in part by the Government of Canada's Department of
Regional Industrial Expansion Ocean Industries Develop-
ment Program, and through the Industrial Research Assis-
20 tance Program of the National Research Council of
Canada. Additional technical support was provided by
Memorial University of Newfoundland and by the New-
40 foundland and Labrador Institute of Fisheries and Marine
Technology.

The author also wishes to thank Mr.George Z. Forristall of
the Offshore Engineering Research Department, Shell
Development Company, the University of Miami and the
80 Husky Oil East Coast Project for their participation in the
testing program. 6. REFERENCES
W
1.Omnifacts Research Ltd. (1985), "Market Survey for Cur-
rent Profiling Systems", National Petroleum and Marine
120 Consultants Ltd. Internal Document, St. John's New-
foundland.

140 2.Neuman, G. and Pierson, W.J. Jr. (1966), "Principles of
Physical Oceanography," Prentice-Hall Inc.

160 3.Rossby, H.T. (1969), "A vertical profile of currents near
Plantaganet Bank", Deep-Sea Research, Volume 16, pp.
337-385.
180 4.Pochapsky, T.E. and Malone, F.D. (1972), "A vertical
profile of deep horizontal current nearCape Lookout, North
Carolina,"Journal of Marine Research, Volume30, pp. 163-
167.
6.0 12.0 18.0 24.0 5.Luyten, J.R. and Swallow, J.C. (1976), "Equatorial Under-
currents", Deep-Sea Research, Volume 23, pp. 999-1001.
FIGURE 4. HYDROBALL temperature profile from the field test il-
lustrated in Figure 3. 6..Npmc i;ii,. (1985), "Numerical analysis of the dynamic
response of the HYDROBALL due to wave condition and
current velocity profile", National Petroleum and Marine
4. CONCLUSIONS Consultants Ltd. Internal Document, St. John's New-
foundland.
The HYDROBALL Current Profiling System has
demonstrated in laboratory and field trials the validity of
using hydroacoustic tracking of a Lagrangian drifter to
measure current speed and direction in water depths up to
500m@

The IDA algortihms used in the system software have, in
comparative testing, demonstrated the ability to properly

525

SHIPBOARD TACTICAL COMPUTER

The Coast Guard's Combat Information Center Modernization

Lcdr Rex A Buddenberg USCG and Lt Andrew Givens USCG
Commandant (G-TES-1) commanding Officer (cs)
US Coast Guard USCG Electronics Engineering Center
Washington, DC Wildwood, NJ

II. REQUIREMENTS
ABSTRACT
Assembling definitive requirements
We chronicle the motivations, for a modernized CIC has turned out to be
architecture and development of a a nearly insoluble problem. The first
modernization of information handling in aspect is that articulation of existing
Combat Information Centers aboard Coast requirements for a high intensity law
Guard Cutters. The mission requirements enforcement/low intensity conflict
are specialized for Coast Guard application is not an easy task.
applications, but the solution is being Secondly, the rapid mission evolution
fabricated from general purpose industrial experienced by the Coast Guard in the past
and defense components. decade means that whatever requirements
stated now are destined to be obsolete by
the time a system can be fielded to meet
them. This problem of finding a solution
I. THE PROBLEM to a problem that isn't well stated has
bedeviled us from the beginning.
Two classes of Coast Guard cutters Nonetheless, we found it quite
have wholly manual CICs -- the WMEC 210 practical to perform some requirements
and WHEC 378 classes. Indeed, a high analysis once we understood requirement
endurance cutter CIC at general quarters #1: provide a flexible, growth-oriented
resembles a college telephone booth structure that is highly adaptable to the
stuffing contest. Despite the existing varied and ever-changing Coast Guard
situation, the information management mission profile.
tasks placed on these CICs is steadily We first examined what we perceived
growing and now dwarfs the capability. to be the 'growth' missions, particularly
Additionally, the WMEC 270 class cutter in their demands on our CH structure.
has a considerable degree of automation, These include air interdiction and surface
but the close coupling of its architecture law enforcement in peacetime and
presents major problems in its evolution continental shelf ASW and mine warfare in
over the life of the hull. the wartime environment. In all these
As technology has placed a personal cases, it is evident that cutters are
computer on everyone's desktop, it has components in a larger milieu and we could
become obvious that the existing Combat not confine our architectural thinking to
Information Centers on our cutters can within the hull. In particular, data is
handle data in a much more effective imported from external sensors and
manner than they do now. Equally databases in all of these missions. This
obviously, the Coast Guard does not have data export aspect has a seminal impact on
the budgetary or organizational means to the communications subsystem requirements.
reinvent solutions from the ground up.
Unfortunately our vendor survey revealed
that there does not exist a canned CIC III. THE ORGANIZATIONAL ENVIRONMENT
that meets both the wartime and peacetime
Coast Guard needs. This problem, and impetus to do
Consequently, we built the Shipboard something about it, has appeared in a time
Tactical Computer project around a clear of budget and personnel shrinkage. This
understanding of our architectural presented us with the constraint of having
requirements, and then a low risk to articulate the larger architecture,
adaptation of existing technology to meet then trim the problem to practical size
them.

CH2585-8/88/0000- 526 $1 @1988 IEEE

without sacrificing the growth potential. Similarly, we are unable to solve the
This meant that notions of a 'do all' external communications problems in the
automation system that fully integrated ship-shore-air network environment within
ship control, tactical planning and combat our present scope, but we are able to
direction were impractical. The present provide a well specified interface in the
scope of the project has brought it within cutter's communications center once this
the capabilities of the three dedicated area is addressed. The key is to
project personnel. understand the internetworking
In our pruning of the project scope, relationship between an intra-ship network
we set into the future any requirements and a ship-shore-air network.
that fall under the aegis of the conning Finally, by defining the immediate
officer and weapons liaison officer in problem as transporting data from sensors
favor of automating the tactical decision to a decision support system and providing
support process. This confines the the evaluator with an integrated displayr
technical risk, provides the most payoff we can capitalize on well understood
in terms of crew workload and becomes a technology.
practical objective.

Intel

On board Display
Sensors Identify
Acquire & Assess
targets
Off board
Sensors

ssess
Threat &
Target file
Recommend
Nell",II %I % % %'1,%,%,V,
Engagement
% % % %

....... ....
% % % % % % %% % % Control
Combat &
% %
% % % % % %
I I I I I I I I . III . I I I . I I I
, -@,"X'l I I II%I%I%I
APTOM01 Z, I "% , , % % Weapons
%
% I I I I . I . I I 1 11
systems

Nav
e/
Navigat llpp le I
Senso
Maneuver &
r c e p t

I . I . I I I Analyze
%I %I & Display
I I I I I I I I I I II

IN1,111I%
Track file

Automation Domain

527

IV. C31 STRUCTURE But first, how much survivabilty is
enough? We have set, as a working
The structure for automating certain hypothesis, the requirement that the
CIC functions does not change the data network must be able to isolate any
flow within CIC. Rather we concentrate on damaged portions of the network so the
replacing.manual communications (sound remainder can continue to operate in the
powered phones) and manual plotting (dead event of a single combat hit (or a single
reckoning tracer) with faster, less error- ramming by a desperate drug smuggler or
prone and less costly mechanization. The single incidence of damage internally by
Sense --> Decide --> Act cycle remains fire or externally by ice). In such
fully intact and will be easily events, the damage control system in the
recognizable to the operators using the ship is likely to be effective and the
system. And the decision making still communications.network must be able to
rests with the human operators, not the support it. The second combat hit in a
machinery; our modernization provides Coast Gaurd cutter.renders the network
computer assistance, not complete survivability problem moot -- there won't
automation. be a ship left for the network.to operate
in.

Erin

Automation Domain Boundaries

The Command & Control Cycle

A. COMMUNICATIONS SUBSYSTEM An additional requirement was that
the network should be able to accomodate a
The intra-ship communications system very wide variety,of communications
that relies on sound powered phones and applications aboard ship, not just the
liberal application of Norwegian steam is tactical data ones. Examples include
obviously inadequate. But this time- logistic, administrative, and engineering
honored practice has the attributes of control data communications applications.
being quite survivable and flexible. How This desire drives the solution strongly
do you make the necessary upgrades in in the direction of a LAN.
capacity without reducing the existing A final implicit requirement is that
features? We knew we wanted a Local Area such a network must be affordable -- we
Network (LAN) because it offered the would have to meet the organizational
promise of capacity, flexibility and budgetary considerations mentioned
abili0y to internetwork to other earlier. And this constraint applies not
communications nets, but existing just to the initial configuration of the
standards for networks lack the network, but would continue over the life
survivability needed aboard ships. cycle.

528

increased, Unix has essentially become
B. DISTRIBUTED COMPUTING real time as its sensitivity to data
requirements approaches that of a true
Once the survivability problem in the real time executive. This shortcoming is
communications network is solved, the far outweighed by the standardized,
survivability aspect of the decision general purpose nature of Unix and its
support system becomes much easier to deal ready availability.
with. The network can supply track data'
to two or more computers as easily as it
can to one. This allows us to use the D. APPLICATION SOFTWARE
Coast Guard's proven-tactic of relaxing
the individual component hardening Decision support software is probably
requirements (with considerable cost the most difficult area in which to obtain
savings) and installing more than one articulate requirements from users or
(much more economical than it looks at resource sponsors. Additionally, it is
first -- the logistics pipeline doesn't obvious that we-will not have the
need to be duplicated). We've done this resources to specify, assemble or maintain
successfully throughout the service with a significant software product within the
navigation receivers, radars, and radios; Coast Guard. Therefore, we have mounted a
we can do it with tactical computers as concentrated, continuing effort to locate
well -- within the budgetary constraint. software products already in development
We can today purchase and send to sea or use in other services. Since the
tede a tactical computer that exceeds our taxpayers have already paid for the
ability to tax it. Indeed it is generally software, we see no reason why we should
economical to provide about twice the duplicate the expense.
horsepower that the forecasted This intelligence effort is very
requirements indicate because of the low interesting, and usually unorthodox.; the
cost of hardware compared to high cost of military services do not answer requests
squeezing ten pounds of software into,,a for information in the Commerce Business
five pound bag. Daily. No central clearing house exists
leads,have come from thelold-boyinet,
symposia, contractors (yes, most of them
C. SOFTWARE FLEXIBILITY are quite ethical) and general 'big ear'
habits of listening. And like any
We wanted very much to avoid intelligence work, the raw data that comes
constraining ourselves to a single in is random, and a large part of it
software package. Similarly, our ever unusable. But we have identified a few
present budgetary constraint argues.very usable and suitable packages.
strongly toward use of an existing package
already owned and supported.by the
government rather than creation and.
support of a custom package -- assuming we V. COMMITMENT TO OPEN STANDARDS
can find one. So the software
requirements drove the computer hardware Permeating the entire structure of
choices to open standards. A color the project, including out-year.
graphics workstation.running the Unix expansions, is our committment to open
operating system and supporting open system standards and layered, flexible
graphics (Programmers Heirarchical architectures. This is our hedge against
Interactive Graphics - PHIGS or Graphics obsolescence, our.guarantee of
Kernel - GKS) and database standards interoperability, and our means to
(Structure Query - SQL) is our choice. integration without the close coupling
Many software packages either are ' or soon that stifles growth.
will be, available off-the-shelf and we More specifically in communications,
can preserve.our ability to choose an this means employment of the International
appropriate package or packages and change Standards Organization layered
them as the mission requirements change. architecture model, and use of the
Unix provides the best tradeoff Government'Open Systems.Interconnect
between availability and standardization, Profile standards in its execution.
cost, flexibility, and real-time This overriding desire to preserve
performance. Unix is not 'real time' in flexibility is embodied in use of the Unix
the sense of being event driven. But its operating system for the decision support
scheduled, time slice task managementy workstations. -Since we are obtaining
particularly in a high performance turnkey decision support software, the
environment characteristic of todays language it is written in is not directly
workstations comes close enough to a true relevant. liowever, anything not written
real time operating system that we are in C, Ada or another high level, machine
within the performance envelope that we independent language was nearly
need. As task switching speeds have automatically ruled out.

529

And in data formats, we are planning VI. ENUMERATING THE SOLUTIONS
to use those prescribed in JCS Publication
25, the JINTACCS message formats for Once we had charted the requirements
anything new we must generate. More on carefully, the search for solutions
this subject in the system integration and yielded to the concentrated search effort
the Presentation Layer Pre-Processor we Coasties are trained for. our search
function below. patterns didn't exactly conform to those
Cartographic standards also hold a in the SAR Manual, but they turned out to
promise of improved chart presentation in be effective enough.
the future. Most current workstation- The Navy's Survivable Adaptable Fiber
based decision support systems generally optic Embedded Network (SAFENET)
use World Database points (from Defense standardization effort will meet all the
Mapping Agency) to create on-screen map intra-ship communications requirements and
representations. A few simply use a more. SAFENET's emphasis on multi-vendor
digitized representation of a paper chart. interoperability promises a long.term
These representations are deficient for a growth path. We have been participants in
variety of applications due to the lack of this standardization effort for about a
detail. Our decision support software year and a half at this writing and are
criteria include a desire to replace the starting to glimpse the promise of
current cartographic database with the deliverable hardware.
Electronic Chart Display and Information A suitable graphics workstation was
System being standardized by RTCM for located in the Navy's Desktop Tactical-
digital chart displays as that standard support Computer-II which is in the
matures and products become available. contracting process at this writing. Our
Navy colleagues were happy to include the
Coast Guard as optional buyers,off their
open contract.

Architecture

Implementation

n't (D T

Implementation of the Decision Cycle

530

We have lined up at least one VII. DEPLOYMENT AND CONCLUSIONS
software package at this writing that has The past year has been spent largely
been in use in the Navy for a few years in surveying the marketplace,
and is currently undergoing a major troubleshooting our architecture,
rebuild to remove exactly the hardware eliciting user feedback and observing and
dependencies that worry us and to meet the aiding the maturation of several products,
general purpose requirements needed for and lining up contractural and logistic
growth and maturation. vehicles. This writing marks the occasion
of a turning point in our project.
The coming year's plans consist
A. INTEGRATION primarily of assembling a shore based
mockup using the above described
The final identifiable component is components and existing sensors in common
glue. Most of todays critical sensors in Coast Guard use. With a bit of fortune,
the Coast Guard's inventory, such as the the project will see a workable,
surface search radar, navigation receivers deployable, and supportable structure in
and visual sighting reports have some sort about a year. This Block One will replace
of native digital output, but it isn't in the existing hardware in the surface
a format understood by the decision warfare/surface law enforcement/search &
support software. Similarly, imported rescue module in our cutters starting in
track reports come in a variety of formats about a year.
most of which need translation. Expansion to the Anti-Air Warfare and
In some cases, particularly over the Anti-Submarine Warfare modules in CIC and
long run, the problem can be solved by re- incorporation of modernization of conning
educating sensors to provide data in a functions on the bridge and communication
correct format. In the nearer term, a functions in radio central are all future
track report translater function will have growth areas fully accounted for in the
to.be provided. This function, which we architecture.
have termed a Presentation Layer Pre-
Processor, can run.as a separate process
on the multi-tasking decision support VIII. THE PAYOFF
computer, or on a separate machine on the
LAN, or as a distributed function at each At the moment, the half dozen crewmen
of the LAN interface points. Specific manning the surface warfare module in CIC
implementation decisions in this area will during general quarters become swamped
remain one of the major work items in the with more than a dozen active contacts to
project in the coming year. Fortunately, account for. This expansion should
this is the only real development work provide a track capacity about two orders
that we must do -- all other parts of our of magnitude larger. Conversely, the
system have been imported from other efficiency payoff is that the job will be
sources without the burden of reinvention performed about half as many crewmen.
or customization. Similar manpower efficiencies will be
realized as we expand into the other CIC
modules.

531

NETWORKING AND SHIP-TO-SHORE SIUP-TO-SHIP COMMUNICATION

Ronald L Moe

Shipboard Computer Group, A-023
Scripps Institution of Oceanography
La Jolla, California 92093

ABSTRACT fleet and the shore based station; and how we have
This paper discusses shipboard communication as it applies attempted to utilize them in networking schemes.
to computer networking between ships and the shore facili- SFHP-TO-SHORE
ties at the Scripps Institution of Oceanography. Different Since the early days, the primary means of ship-to-shore
communication links are discussed including satellites and communication has been radio-teletype (RTTY) or high
VHF (high frequency radio). Further detail is given relat- frequency voice (VHF). Radio-teletype is sometimes the
ing to the networking options aboard ships and at SIO only communication path available because of distance,
while pointing out some of the advantages, disadvantages, atmospheric conditions, etc. Since it is slower and less
and limitations of the different systems. flexible than voice communication, voice is the preferred
INTRODUCTION method to communicate.
Communication between the shore facilities at the Scripps Approximately ten years ago, satellite communication
Institution of Oceanography (SIO) and the ships in our fleet became available to the oceanographic community. NASA
has always been a necessity. The need to communicate had two transponder relay satellites that were -originally
.may be for ship's personnel to have contact with the VHF communication experiments. NASA wanted to deter-
Marine Facilities or for scientists to be able to pass mes- mine the reliability of VHF communication with the satel-
sages to SIO or to their own institutions if they are not lites. The results from the experiment showed that VHF
affiliated with SIO. Members of the scientific party are communication was not very reliable. The satellites had
usually made up from many diversified groups, such as served their intended purpose and were no longer being
scientists and technicians from SIO or other institutions used. Access to the satellites by the ocean 'ographic com-
both in the U. S. and abroad, as well as observers from munity was allowed by NASA for communication pur-
foreign countries if survey work is done in their territorial poses. The satellites are referred to as ATS-1 and ATS-Ill.
waters. The larger of the ships can be anywhere in the Both satellites are in geosynchronous orbits and appear to
world with the associated problems of long distance com- remain stationary in the sky if the observer remains ' station-
munication. To meet the communication needs, SIO has ary on the ground. As a ship station moves on the surface
always operated and maintained its own licensed radio sta- of the earth, the apparent position of the satellite in the sky
tion, call letters WWD. will change. This change is very slow and changes in the
In 1966, it was decided to install permanent computer ship's heading will cause tricking problems.
centers on our larger ships. The computers were to be used The satellites were to be maintained stationary over a
for real-time data acquisition and control. The post- specific longitude by NASA ground control. ATS-I was
processing of data and general purpose computing were maintained over 149 degrees west while ATS-III was main-
also to be supported. The Shipboard Computer Group was tained over 105 degrees west. The satellites oscillated
created and charged with the responsibility for this task. about the equator and ATS-1 would move about 11 degrees
Along with the shipboard installations, an identical facility above and below the equator. ATS-Ill would move about 9
was established on shore to support the ships and to pro- degrees above and below the equator. The period of the
vide general computing services at SIO. Soon it was oscillation is twice per day. One problem that has
recognized that there was a real need for computer to com- developed is that the fuel for the satellites positioning
puter communication. Initially what was envisioned was motors has been used up so no control can be exerted over
the ability to transfer messages between individuals, small them. ATS-IH is parked in a gravity anomaly and should
programs or documentation, and maybe even small remain there-forever. ATS-I has drifted into and back out
amounts of data. of ATS-111's territory and both have a wobble in their
Ship to shore communication has evolved from CW, Morse orbits. these satellites have long outlived their projected
code sent with a telegraph key keying a transmitter, to lives but are still in useful condition.
currently using satellites that provide near telephone qual- The ATS satellites are translating devices rather that
ity service. We will briefly discuss how some of the com- repeaters. They re-transmit exactly what they receive and
munication facilities function between the ships in the SIO the amount of power that they provide during transmission

CH2585-8/88/'0000- 532 $1 @1988 IEEE

Table 1 ATS Channel Definitions

Channel Principal Usage

SDR (Ship Data Receive) This is the "low data channel" and the ship stations receive
data on this channel. Ile land stations transmit data on
this channel.

CH2 (Channel Two) ATS-III voice channel.

CH3 (Channel Three) This is the "center channel" and is reserved for ATS-1
operation only.

CH4 (Channel Four) ATS-HI voice operations.

SDX (Ship Data Transmit) This is the "high data channel" and the ship stations
transmit data on this channel. The shore stations receive
data on this channel.

Note that SDR and SDX combined give full duplex operation. These five channels are defined for the convenience
I of the satellite's users. At the satellite, they appear as a single wide channel.

is dependent on the amount of power received from the vice with simplex data transmission rates of 56 Kbytes per
earth station. The stronger die received signal, the stronger second. Inmarsat's three Marisat geostationary satellites
will be the re-transmitted signal. ATS-I is a very basic were launched in 1976 and are positioned to cover the
translator and can support only a single user at any given Atlantic, Pacific, and Indian Ocean regions. The complete
time. This means that voice must be operated simplex and Inmarsat space segment is made up of eight satellites which
there are no data channels on this satellite. It is quite easy provides near global coverage with the exception of the
to transmit data on ATS-I but the data must be transmitted polar regions and a band in the middle of North America.
one way at a time. ATS-111 is a more advanced translator Four new dedicated satellites are currently under construc-
and can support several users simultaneously. There are tion and will be launched beginning in 1989.
five defined channels for the ATS satellite communication Besides voice and data communication over the satellite
system (Table 1) and ATS-Ill can support all of them. The systems, it is possible to send graphics data. The ATS does
only real limitation on multiple user operations on ATS-HI this throught Qwip which we have used to send bathymetry
is the total power available from the satellite. maps, administrative forms, and program listings. The
The current schedule calls for three one hour windows for Inmarsat has FAX capability. Besides being monochrome,
oceanographic use for voice traffic. The times are 140OZ our shipboard FAX machines have gray scale processing
to 150OZ, 180OZ to 190OZ, and 230OZ to OOOOZ. The data and have been used to send high resolution pictures to
channels on ATS-IH can be used at any time although by shore facilities.
only one user at a time. The data channels are operated on
simulated low power. The simulation of low power results EARLY COMPUTERIZED ATTEMPTS
from the location of the data channels on the satellite's
base band. The data channels are near the edge of the In the early 1970's, our first attempts at computer to com-
satellites response so it takes a lot of power to operate over puter message passing used a computer to send or receive
them and they normally provide relatively weak signals. 16 bits of binary data to a black box built in-house. The
This allows the ship systems to use maximum available box provided two frequencies that represented O's and Vs.
power and still avoid interference with voice channels. These were then sent to key the appropriate equipment
While the ship use of the data channels must never inter- which were then transmitted over the RTTY. This attempt
fere with a voice channel, it is often the case that a voice did work, but because it involved three people on shore in
channel will interfere with the use of a data channel. This three different locations and two people aboard ship to set
is one of the limitations on the use of the data channels. up the communication and make it work, and because of
One of the more recent communication systems installed the lack of constantly reliable transmissions, this project
on SIO ships is the Inmarsat, International Maritime Satel- was never used.
lite Organization, satellite link. It uses the Mazisat satellite A later attempt was made utilizing the ATS data channels
system as a marine communication system that provides to provide a path for the computers aboard two ships. This
Telex, telephone, and data services on a worldwide basis. demonstrated to us that it was indeed feasible to communi-
Data transmission rates of up to 2400 baud are provided. cate this way, although with varying degrees of bit error
More recently however, AT&T International using Inmar- rates. Operating from two different moving platforms try-
sat facitilies has made available a data communication ser- ing to keep their antennas tuned to the satellite does not

533

always produce the best of conditions. Today almost all To implement the Telex connection required still another
computer traffic is from ships to our shore facility and set of rules. Access to Telex was made over telephone
vice-versa and ship-to-ship communication via computer is fines, and SIO subscribed to several companies to provide
rarely done. Telex services. Communication protocols for the different
services were never quite the same. In order to connect to
CURRENT IMPLEMENTATION a phone line, we purchased a NU-DATA Series 106 con-
The Shipboard Computer Group of the SIO uses a DEC troller which contains an originate/answer modem and an
VAX-lln50 in its shore facility and two HMO's aboard auto-dialer. This box took care of the basic protocol
the R/V Melville and Thomas Washington. The computers requirements to communicate over Telex. Output from the
run under the Berkeley UNIX* 4.3 operating system. The box was RS-232. The Telex services we subscribe to also
computers aboard ship are networked together via Ethernet use ASCII data code so no data conversion was necessary.
with provisions to connect other computers brought on Four more lines to the computer and the Telex hardware
board if they have the appropriate connecting hardware. requirements were met.
The computers on shore are networked via Ethernet to the The radio station required that all traffic transmitted, or
University of California at San Diego's campus wide Eth- received, be logged on hard copy. Therefore, all equip-
ernet network and also to UCSD's local area network ment also had a printer in parallel. For equipment operat-
(LAN). In addition, smaller computer systems around SIO ing in particularly noisy environments, like the ATS, a
are linked to these facilities by our computer. device built in-house was also provided to filter out non-
In early 1985, it was decided at SIO to "computerize" the ASCII characters to the printers. This prevents the printers
radio station, and to a lesser extent, the larger of the ships from taking unexpected form feeds, going into strange
equipped with computers operated by the Shipboard Com- modes, etc. One other requirement by the radio station was
puter Group. The goals for the radio station were to be to cover times when the computer was not available due to
able to send and receive electronic mail (e-mail), between hardware problems, preventive maintenance, or some other
the radio station and the people using its services. The reason. For this, all equipment also have a terminal that
messages transmitted between the ships and the shore radio can be switche d in place of the computer. Data can be sent
installation were then to be sent under computer control to directly as it is being typed in by the operator or up to four
the various facilities. The communication facilities of screens of data, 24 lines by 80 or 132 columns per screen,
interest were the ATS, RTTY, Telex, and two message can be saved in the terminal.'s memory and transferred after
repeaters ashore where all messages are printed, one in the. being entered.
administration offices and one in the marine operatiorls Up to eight RS-232 lines were needed to support the radio
offices. stations requirements. However, after following the vari-
To implement e-mail for the radio station was an easy task. ous tunnels, conduits, etc., the length of the wires far
It was only necessary to run the appropriate hardwire lines exceeded the RS-232 standard. Therefore it was decided to
between the station and our facility, and to fumish them string only two twisted pair lines, four wires, and have a
with a terminal and an account on the computer. multiplexor on each end that would provide the eight lines
The ATS was the next easiest implementation. Because needed. Traffic through the multiplexor was not expected
data channels in the radio station were already connected to to be very heavy and well within its limits. Only the termi-
a hardcopy terminal via RS-232 lines, it was only necessary nal ports runs at 9600 baud. Ile ATS port is set at 300
to install a switch to select the current terminal or a coin- baud and the other ports at 110 baud.
puter connection, and to string more hardwire lines from The original implementation of the repeater loops for mes-
the switch box to the computer. In this case, the lines were sages to the administration and Marine Facilities was one
connected directly to one of the computer's RS-232 ports. loop on a dedicated teletype line. The hardware required
Data had been routinely transmitted in ASCII code so no was identical to the RTTY. Recently the teletype line has
data conversion was necessary. been removed and the two printing functions separated.
The RTTY connection was more difficult to implement. The administration's printer is located close enough to the
The output from the transmitter was usually connected to a computer that wires were strung directly to the printer from
teletype. This meant current loop connections and BAU- the computer and a print spooler implemented to output all
DOT code. A Black Box system was purchased to convert messages. The link to the Marine Facilities spanned
the current loop to RS-232 signals and back; and another approximately 15 miles and was more involved as it was
box was purchased to convert the BAUDOT code to ASCII necessary to go over telephone lines. In this case the com-
code or vice-versa in real time. The data conversion could puter port was connected to a modem with autodial capa-
have been done easily by the computer but since an in'ex- bilities and the printer was connected to a dedicated phone
pensive box could be purchased to do the code conversion line with another modem. A print spooler was imple-
and to manage the RS-232 1/0 requirements, it was decided mented to deal with potential printer error conditions (e.g.
to do it in hardware. Once again four more wires would off line, out of paper, paper jammed, and phone line noise).
need to be strung to connect the transmitter to a computer The ibove descriptions cover most of the hardware require-
port and the hardwire requirements would be satisfied. ments needed to route traffic at the radio station by the

Trademark of AT&T

534

computer. Aboard the ships, only a subset of the hardware including personal computers such as IBM PCs and
described above was needed, and it was identical to that Apples, this capability allows for a convenient and reliable
used in the shore installation. means of transferring files.
In order to make all of this hardware work, a certain Another UNIX utility that could be used, although we have
amount of software was required. UNIX provides most of not, is the UNDC to UNIX COPY, (UUCP), program. It
the utilities needed although in some cases maybe not quite could be used to advantage for sending electronic mail
as well as the user might like. The main problem encoun- between ship and shore. The electronic mail utility in
tered in the automating of the radio station has been accep- UNIX uses UUCP and Ethernet to transfer messages. The
tance by the radio operators. Most of them had no com- UUCP program, like Kermit, will keep retransmitting pack-
puter backgrounds and were reluctant to learn the neces- ets until a successful transfer takes place. With UUCP,
sary computer operations. Fortunately, UNIX provides a mail is held until a connection is established. Once the
facility for writing scripts containing lines of commands connection is established and verified, all the mail that is
that are executed by the operating system. To make it being held by both computers is sent without operator
easier for the operators to use the computer system, this intervention.
facility was used to construct a menu to walk an operator Aboard ship or on shore, people can connect their comput-
through whatever function he wanted to perform. Func- ers to our VAX systems via RS-232 lines or the Ethernet
tions were included to read or compose e-mail, start and links. These computers can then go further. Once they
stop receiving traffic from any of their communication have logged into the remote host computer aboard ship or
facilities, starting editors to compose messages, list con- ashore, they can log into any desired computer that has
tents of directories, remove files, or any other function that been connected to the remote Ethernet network. Ashore,
they would need. The disadvantage to this way of operat- since UCSD has access to most of the major nets, a user
ing is that scripts are slow since they are interpretive and aboard ship can transmit mail or attach themselves to virtu-
are submitting commands to the operating systems. Also ally any net or computer anywhere via the UCSD network-
there is the necessity of doing extensive error checking on ing facilities.
operator input. The system does work well and has As already stated, our system is used on a daily basis in the
achieved the desired objective with half the number of more passive mode, with the computer listening at the port
operators required at the station. or sending data out the port to the transmitter. This mode
THE NEXT STEP requires very little operator interaction. The operator need
only turn it on or off by software. To perform real com-
Our usual way to establish a connection between computers puter communication requires the radio operators at both
is to listen or send on a port with no interaction between ends to establish the communication path. On shore,
computers. However, once the various transmitters have because the radio facilities and the computer facilities are a
been connected to the computer, the link for computer to distance apart, the radio operator must then call the com-
computer hookup has been established. The radio link is puter user over the telephone to notify him that communi-
really nothing more than a copper wire link with probably a cation has been established. The user can then proceed
somewhat higher bit error rate. To go a step further is to with his computer link. After he is through, he must then
use one computer to connect to the other computer. The call the radio operator back so he can break the radio con-
only, change for the computer is that on one end, the port nection. Setting up the "normal" computer mode must then
corresponding to the 'communication device must be be done. Radio operators would prefer not to go through
configured so that the computer can be "logged into". Con- this routine, so it is only done in cases where large file
necting computers together can be done through RTTY or transfers need to be completed in an error-free environ-
voice channels but with the current hardware we only do it ment. To alleviate this problem, the necessary ATS equip-
through the ATS. Hardware does exist for the RTTY and ment should be purchased to allow the Shipboard Com-
voice channels that provides more reliable communication puter Group to perform all the necessary functions required
and in a full duplex environment but so far there has been to set up the ship-to-shore link themselves. This would
no interest in purchasing it. expedite matters tremendously but funds have not been
Once the computers are talking to one another, it is possi- available to do this.
ble to do remote computing or other computer functions. It One other means of connecting computers has been
is possible to make data transfers under computer control achieved by way of Inmarsat. Using a triple speed modem
making the transfers much more reliable. One often used (300, 1200, and 2400 baud) with error correction capabili-
file transfer facility is Kermit. It does not have error ties, a user aboard ship can use a voice channel to call any
correction abilities like cyclic redundancy check (CRC), compatible modem. Our experience has been that using
checking to enable data stream correction as the bits go by. voice channels can result in a noisy environment. The
It works by using checksums to determine if a packet has error correction option within the modems has never been
been sent error free. If the checksums do not match, the tried. Our use of this system has been restricted by the cost
receiving computer requests the sending computer to of Inmarsat voice channel communication. The current
retransmit the packet again. We have used Kermit to cost is $30 for the first three minutes and,$1 for every fol-
transfer files between computers with 100% reliability. lowing tenth of a minute. In most of the cases where the
Because Kermit has been ported to so many computers data collected aboard ships would be usefully sent to shore

535

for analysis, the quantity of data is too large to be cost ing on the ship, different people may operate the ATS. On
effective. Some of the typical large data streams collected the smaller SIO ships that do not have radio operators, the
at sea include seismics profiles and Sea Beam bathymetry, captain or a member of the scientific party may operate it.
both of which accumulate many millions of bytes of data On the larger ships, it may be the radio operator
on a daily basis. exclusively or a member of the scientific party. There still
Some experimentation has been done with intelligent con- is apprehension by captains and radio operators for anyone
trollers sending out packets with error correction and/or using the ATS or any other radio without their direcr con-
packet retransmission over the different transmitters. This trol. In a computer-to-computer hook-up, this control is
makes connecting computers together by RTTY or VHF obviously lost.
more practical. These controllers have internal micropro-
cessors to take care of all decoding, signal processing, and CONCLUSIONS
protocol requirements, and their output can be made com- For us, ship-to-shore computer-to-computer data communi-
patible with VHF, RTTY, and RS-232 requirements. The cation, even under the best of circumstances, has the prob-
internal modems can transmit packets at rates up to 1200 lem of various degrees of bit error rates. Satellites provide
baud with the option of using an external modem for higher the best and easiest transmission mediums. The limitations
baud rates. One controller tested also supported many on ATS are its slow transmission rates, limited availability,
parallel, dot matrix graphic printers with a direct connec- manual antenna tracking, and lack of global coverage. Its
tion to print BF monitored FAX signals. main advantage over Inmarsat is that after its initial cost,
Another concern to consider is that the captain and the its use is free except for maintenance on the equipment.
radio officer are responsible for all communication to and The limitation on Inmarsat is its expense, $30 for the first
from the ship. The ATS in particular is technically a tele- three minutes and $1 for each additional tenth of a minute
phone and not under FCC regulation. It therefore does not to North America. Its advantages are transmission speeds
need a licensed radio operator. Also, on SIO ships it is up to 2400 baud over voice channels or 56 Kbytes over
considered as a piece of scientific equipment and depend- data channels, it is always available and with near global
coverage.

536

THE DEFENSE MAPPING AGENCY'S
NAVIGATION INFORMATION NETWORK

Steven C. Hall

Defense Mapping Agency
Washington, D.C. 20305-3000

Abstract The weekly Notice to Mariners is a joint publica-
tion of the Defense Mapping Agency (DMA) , the
More than a decade ago, the Defense Mapping National ocean service (NOS) and the U.S. Coast
Agency made a ccmmitment to improve the Guard (USCG) . Its use is intended primarily for
means of processing, managing, and producing the offshore mariner; that is, U.S. Navy units
navigation safety publications and information operating worldwide, ccmmercial U.S. vessels oper-
using automation to the fullest extent possible. As ating in international trade or coastal trades and
the present Automated Notice to Mariners System lastly, the offshore sailor who goes beyond the
developed and matured, it became apparent that the limits of the USCG local Notice to Mariners. It is
future of dissemination of these data lay in a time-dated publication essential to the preserva-
teleccrmnun icat ions, thus the creation of the tion of the integrity of the nautical products
Navigation Information Network (NAVINFONET). This which are necessary for the safety of all ship-
paper will review the history, design, and use of ping, large or small. It is critical to all
the ANYS and then discuss the present and future concerned that the information contained in the
utility of the NAVINFONET. As the age of the Elec, Notice be accurate, current and distributed in a
tronic Chart Display and information System (ECDIS) timely manner.
approaches reality, the potential of the NAVINFONET
as the only functional existing system to support When viewed as a battle against time, three
corrections to ECDIS at sea may well prove its separate aspects of the Notice to Mariners process
greatest value. In the interim, its worth is proven were defined; namely, compilation, composition and
daily by the myriad of users who seek up-to-date distribution. Within the canpilation phase of a
marine safety information to correct their charts weekly Notice, the first target was chart correc-
and publications far in advance of receipt of the tions as the Notice is the legally designated
printed word through the mails. vehicle to correct any or all of the 5,000 odd
charts produced by NOS and DmA. The first thing
that had to be done was to break away f ran the
The Defense Mapping Agency began the design of former loosely defined format of a chart correc-
the Automated Notice to Mariners System (ANMS) in tion and create specific guidelines to simplify
1975. The intent of this project was to maximize and standardize information presentation. Suffice
the use of automation to improve the composition, it to say that DMA consciously f lew in the face of
management and distribution of several products international tradition in designing this new
which prcmote navigational safety, primarily the computer compatible format which:
weekly Notice to Mariners. In view of the ambitious
scope of this project, and keeping in mind the 1. corrects each chart individually
blinding pace at which ADP systems were reaching 2. tracks edition number, date, correction
the market in the last decade, a phased development authorty and previous correction
approach was selected. This allowed DMA to add the 3. uses a three (3) colLnn format for text
latest off-the-shelf hardware and certain pieces of that tells action (in f ive (5) verbs --
commercial sof tware to the ANMS as the system ADD, ]DELETE, SUBSTITUTE, FELOCATE, aiANGE)
matured. The initial design concept of this devel- --subject (specific standards, minimum of
opment was to allow ccmmunications access to the words) - position (level of accuracy tail-
extensive files to be developed not in the tra- ored to specific scale and required action),
ditional meaning of a data base but actually as By changing to columns, reducing to only the mini-
data files. The success of this phase develop- mum text necessary and custom designing coordinate
ment approach has already been proven even though presentations, data entry for a computer based
its full potential has yet to be achieved. Inter- system became simpler, canpiling (and training
estingly enough, the communications portion of the compilers) became faster and easier, and accounting
system design has been so popular with the user for various horizontal datuns in a common area
ccmmunity and so beneficial to the Defense Mapping ceased to be a problem.
Agency and the U.S. Coast Guard among others, that
this portion of the workload has now been given its Use of the new guidelines established as the
own computer and its own name -- The Navigation system design proceeded has allowed the Marine
Information Network or NAVINFONET. Information Specialist (MIS) compiling a notice to

CH2585-8/88/0000- 537 $1 @1988 IEEE

move through his tasks more easily and more accu- via remote communications and to improved effi-
rately than in the past. Data entry to the weekly ciency of Notice to Mariners data processing.
work file is facilitated by the use of intelligent
graphics terminals which will soon be available on Here it is appropriate to mention that periodi-
each desk. The quality and integrity of chart cally throughout the year special recaps of infor-
corrections and, in fact, all sections of the mation appear in the weekly Notice. Automation of
Notice have been maintained throughout the develop@- weekly compilation of the Notice allowed automatic
ment process. updating of numerous summary files for later pro-
duction; thus, time to compile these listings was
Automation was used to expedite the compilation eliminated completely. The files referred to here
of the catalog correction portion of Section I of include the, monthly mobile oil drill rig update,
the weekly Notice to Mariners as well. The DMA the quarterly drill rig status list and the quar-
indexes its hydrographic products in ten (10) vol- terly list of charts affected by notices, to name
umes of catalogs. These dynamic publications are but a few.
corrected weekly. Naturally, if mariners are
expected to have the latest available chart on In each of the aforementioned cases, the informa-
board, there must be a rapid means of notifying tion is processed by an MIS, checked, entered to
them when a chart is ready for purchase. A side, the work file by an Editorial Assistant working at
benefit of this development was a file of chart a terminal, proofed from an output device, checked
titles, scales, coordi nates and other attributes and revised as.necessary and loaded to a weekly
for future use. Ebr now, suffice it to say that work file on the ANMS computer. Although this input
this change has trimmed compilation time for cata- procedure represents several new steps in the com-
log corrections in half. pilation proceSSr it remains far more efficient
than the past manual methods used at DMA. The
Computer assisted compilation of the List of composition process begins at this point.
Lights (produced by DMA) and Light Lists (produced
by U.S. Coast Guard) has resulted in a measurable At the conclusion of the compilation of the
decrease in compilation time and a corresponding weekly work file, several computer routines are
improvement in product accuracy and dependability. executed by the ANMS. These procedures simultane-
In this case, automation was a necessity as well as ously update a number of history files from which
a desirable improvement because the former method various products are composed or queried. In
used by DMA and USCG was using totally archaic addition, a tape is generated which drives an
equipment and technology. Adding the List of Lights automated photo typesetter wbich produces page
and Light Lists to the ANMS provides far more negatives of chart corrections, list of light
up-to-date information than ever before due to corrections, light list corrections, radio aids to
simplified data entry and storage procedures. The navigation corrections, catalog corrections,
compilation data files for all seven (7) volumes broadcast warning message texts and special cumula-
.of the DMA List of Lights and six (6) of the seven tive listings -- each ready for the press. The
(7) USCG Light Lists are now mature and strictly reader should contrast the speed of this automated
maintained on a weekly basis. These latter volumes, operation with the old hot lead type operations of
although produced by the US Coast Guard, are main- the Government Printing Office or even our own
tained through the weekly Notice to Mariners and vari-Type list of lights process where each line of
compiled and produced usingthe computerized work type was manually typed on to one_3@'_x -- 10 " card
files of the AkmS. unique to these operations is that required the operator to change type elements
the fact that all the USCG light files (weekly, as many as four times per line/card. Computer
summary, and annual) are accessed and processed driven typesetters at DMA can set up to 4,000
remotely by Coast Guard personnel, thus eliminating characters per minute, a speed which makes even
their need to be on site at DMA. the most proficient manual operator obsolete.

Corrective information for certain radio aids to Because this phase of production is controlled by
navigation are now compiled on the ANMS either as software, it has enabled DMA to upgrade or replace
part of the appropriate List of Lights or as part@ its former typesetting equipment without interfer-
of Pub. 117, Radio Navigational Aids. Further ing with the critical schedules of the Notice to
automated compilation of this latter volume is Mariners. ANMS Software modifications are neces-
planned for the near future. sary, of course, but that is far more desirable
than complete system redesign. Once again the
The next compilation task to be addressed as part phased development procedure has proven advanta-
of ANMS was the Broadcast Warnings portion of the geous.
weekly Notice. These warnings are the texts of
messages issued by DMA as part of its duties under At the same time that the various summary and
the worldwide Navigational Warning Service and history files were created and updated by the
include NAvAREA Iv and NAvAREA xII warnings; weekly work filesr an additional compilation/compo-
HYDROLANT and HYDROPAC warnings; and Special warn- sition benefit was realized. As the master data
ings. The compilation advantage here was not re- files reached maturity, it became possible to
lated as directly to the weekly Notice as it was to produce with vastly improved efficiency the five
the production of the Daily memorandum and the (5) voluiies of Summary of Chart Corrections twice
quarterly summaries of messages in force. NDnethe- yearly, seven (7) volumes of List of Lights and six
less, this software package did contribute to the (6) volumes of Light Lists on demand or on a sched-
overall usefulness of the ANMS data to mariners ule simply by generating a print tape to produce

538

page negatives frcm the same auto photo typesetters also was selected because all ships equipped with
previously mentioned. To give an idea of the extent an INMARSAT ship earth station have TELEX and the
of this improvement, the total summary of correc- ship driver is our main target. Up until early
tions file contains every chart correction to every 1987, NAV-INFONET was also accessible by TWX;
DMA and NOS chart published since mid 1975 or since however, it proved to be of little use so the line
the latest edition date of the individual chart. In was converted to a TELEX rotary circuit.
other words, it contains everything the user needs
to bring his charts up-to-@-date to the point where As of today, NAVINFONET is accessible over
maintenance can be easily assumed by the weekly voice grade telephones (including INMARSAT voice
Notice to Mariners. This represents 40 megabytes of circuits) on: the U.S. telephone systems by five
data. The lights data f iles on the other hand (5) 300 baud lines, f ive (5) 1200 baud lines, and
contain the current status of over 82,000 aids to two (2) lines at 1200/2400 baud; the European
navigation worldwide. These files hold 107 standard CCITT V.21 by two (2) lines at 300 baud;
megabytes of data. To put this improvement in and one dual TELEX line compatible with any Inter-
perspective, composition, negative production and national Record Carrier. This equates to 16 lines
printing of a List of Lights formerly took on the plus several hard-wired access ports.
order of.90 days prior to automation. Now it can
be accomplished in two weeks, although we normally At the time this paper was prepared, there
allow four. On the other hand, the constant changes were over 1,600 individual user identification
representing new data for chart correction codes on issue that permit access to the NAVINFONET
summaries were never an easy update process for system. of this number, about half have used the
manual compilation whereas now the summaries are NAVINFONET and make an average of 1,000 queries to
produced practically at the touch of a button. the system each month. These figures do not reflect
special query options used by DMA, the Federal
The final time oriented design element that was Republic of Germany, Lykes Brothers Steamship
chosen was information distribution. This was a Ocxnpany, or selected. other customers with whom DMA
tough problem because broadcast warnings are trans- collaborates on special projects. Current usage is
mitted by Morse telegraphy radio, the Notice to at about 16% of capacity which leaves much room for
Mariners is sent by first class mail, and the expanding the customer base as well as improving
remainder utilize various rapid automatic distribu- our available services.
tion methods to reach the user. So --where did we
go from here? After all, already we had cut compi- A pioneer in managing the development of the
lation and composition time down by a factor of ANMS and NAVINFONET, Mr. Glenn R. DeYoung, a former
50%, that is, from 42 days to the user under former head of the Notice to Mariners, was adamant about
methods to 21 days for the printed paper copy with the fact that the system must be user friendly. His
the computer assisting with compilation and compo- position was (and ours still is) that the system is
sition. Enter the Navigation Information Network for the end user, the navigator and cartographer;
(NAVINFONET) which for our primary target of chart therefore, it must operate in plain language, not
and light corrections, reduced another eight (8) in code. With this guideline in mind, nine (9)
days off the user access time. separate modules were developed within NwiNFONET,
each with appropriate sub-elements. As with most
NAVINFONET, like the ANMS, is centered around a things, an advantage can become a disadvantage at
second Prime 2550 super minicomputer. In point of times. While working in plain language was advan-
fact, they are very close to being totally redun- tageous overall, it had the effect-of slowing the-
dant systems.. This is-- necessary --to ensure user's response time by requiring numerous prompts
dependability, data integrity and uninterrupted and replies. This became just another on-going
service to the user. Each time the ANMS data files challenge for us to design an optimum compromise
are updated, a similar update is performed on the position and we're still striving for that plateau.
master disk of the NAVINFONET. Thus, its data files
are up@to-date making the latest information No matter whether the user selects "good ole" 50
available for remote query. The NAVINFONET computer baud TELEX or the more efficient 2400 baud
is a multipurpose computer which simultaneously voice-grade circuit, a conversational series of
supports new system development, broadcast warning -prompts will le 'ad the user step by steR to the
processing, manages the query system and stands, needed data. Preformatting a request can be very
ready to take over from the ANMS if necessary. helpful to the user unless he is just experimenting
with the system. it must be emphasized again that
NAVINFONET's primary function is to bring the although the data are free, the connect cost is
user closer to the data needed while at the same paid by the user. Tb receive a user ID, one need
time providing selective query options to minimize only convey the request to:
the connect time and associated costs involved with
interactive data exchange over commercial communi- Director
cations circuits. As the potential user community DMA Hydrographic/Topographic Center
was examined, it was obvious that the versatility ATTN: MCN
of the worldwide voice grade telephone system Washington, D.C. 20315-0030
presented the most efficient and accessible cir-
cuitry readily available to NAVINFONET. Not so and an identification code will be assigned and a
simple to access, or as versatile, but universally user manual supplied by return mail. Nearly any
accepted, was TELEX. Even though it is slow and data terminal with an internal MODEM or acoustic
somewhat unfriendly in a, conversational mode, it MODEM or a micro computer with a communications

539

software package may be used. NAVINFONET is is called "Query by Port." By entering your choice
configured to operate full duplex, parity off, 8 of all notices or notices back to a certain point
data bits, 1 stop bit. The 300 baud modems are following the two digit figure wbich corresponds to
compatible with the Bell 103 standard while the one of 92 ports or regions, the user avoids typing
1200 baud equipment correspondsto the Bell 212 or a long list of chart numbers. A word of caution is
Racal-Vadic 3400 standards. The 2400 baud circuits necessary here. The charts selected for each port
are compatible with Bell 201 modulation standards may not be suitable for each individual user. It is
and, of course, there are two 300 baud modems up to the user to make that decision and supplement
con-forming to CCITT V.21 standards for our the data received as he feels necessary.
overseas user community.
Lastly in this section is program 22 which we
Module 1 is the ANMS Mailbox/utilities group. informally call the autanated chart card. By enter-
There are only two sub-elements to this section at ing a chart number, the user obtains a printout of
the present time. Program 10 is the ANMS Mailbox. the Notice week numbers in which that chart was
This allows the user to send a message to DMA from corrected beginning with its announcement as
his location and to get a copy of exactly what he available from a distributor. This program is used
sends to us. This is a supplementary, one way extensively by U.S. Navy units and others who track
canmunications program. That is, it is not in- the number of Notices charged against a chart.
tended to compete with commercial electronic mail
systems or bulletin boards, but merely provides a Module 3 is concerned with the text of various
narrow set of users with an alternative method of radio broadcast warnings. This particular group is
canmunicating information to DMA for immediate valuable to retrieve previously transmitted radio
consideration for action. In point of fact, an messages which were missed for whatever reason.
incaning MAILBOX message prints out at the World- This is the only program on the system that gives
wide Navigational Warning Service watch desk which information - af ter - it-was pranulgated-by another
is operated 365 days per year. most of our oil, means; however, it provides a highly valuable means
drill rig movement data canes in over our MAILBOX of restoring radio warning continuity after a break
!as do many requests for information fran companies in reception. Selections are based on all warnings
under contract to DMA. occasionally, DMA will or warnings since a specific date. Further deci-
promulgate outgoing information on the MAILBOX; sions are made using the DMA subregion system to
however, in order to receive these messages, the narrow the affected area or else NAVAFdW
user must call into the system. As can be seen, HYDRCLANT-HYDROPAC designations can be used. Spe-
that is not particularly dependable or efficient in cial Warnings may be queried by date or by number
the outgoing mode. as can MARAD Advisories. The intent of this module
was to provide the user with sufficient parameters
The second option under this module is the NA- to narrow his search for these vital data as much
VINFONET User's Manual. As the Manual is canposed as possible. Last, but not least, with this in
on the ANMS, this option was added to allow changes mind, option 32 prints a list by number only of all
to the system to be made available to users without effective warnings in a chosen warning series so
continually reprinting the instruction booklet. the user can check his inventory and request only
This option is preceded by some very interesting the missing items.
numbers concerning the down load times to transfer
the Manual from its file to the user. Quite simply, Modules 4 and 8 are practically identical. The
it states "This selection prints out the 1987 major difference is that 4 deals with the DMA Lists
Blition of the NAVINFONET User's Manual. This takes of Lights for foreign waters and 8 deals with the
4 minutes at 2400 BAUD, 8 minutes at 1200 BAUD and USCG Light Lists for U.S. waters. The user first
32 minutes at 300 BAUD. TELEX users should not selects the volune number he wishes to query.
exercise this selection." These statements illus- Thereafter, he may select individual lights by
trate quite graphically the advantages of voice number or groups of lights consecutively listed or
grade use over TELEX. Although voice grade may cost even all lights between two selected light num-
more per minute, ultimately its efficiency and bers. This last option will allow a user to print
speed override the lower per minute TELEX rates. an entire volume if he so chooses. All these option
selections print the light data including any cor-
The second module, subsystems 20, 21 and 22, deals rections processed against the selected lights
with chart corrections for NOS and DmA nautical since the latest printing of that particular vol-
charts. Program 20 is the most active query pro- ume. On the other side of the coin, an option is
gram on the system. It gives the user access to all offered that allows the selection of just the
corrections to selected charts or corrections from lights that have changed since the last edition
the current notice back to a user selected notice date of the publication. These programs are de-
number for selected charts. The most common and signed to facilitate updating small portions of a
efficient use of this program is as a last minute List of Lights or Light List until the hard copy
update before making port or as a supplement to the Notice is received through the mails. This will
printed Summary of Corrections and weekly Notice allow the corrected volume to be used safely as a
when a change of orders is received to divert to a supplement to the appropriate up@to-date nautical
port or area where the chart portfolio may not have charts.
been maintained previously.
The last unique user selection in the lights
Option 21 was created to save the user from program allows the navigator to select only those
having to enter a long series of chart numbers. It lights exceeding a certain range of visibility.

540

Ibis option may be useful when coasting in order to Module 7 was designed to give the user convenient
see at a glance the specifications of all lights access to DMA Hydrographic Products catalog correc-
visible at 12 miles or more. other uses are as a tions. Sales'agents, ship chandlers, and navigators
voyage planning tool for the navigator and for the reviewing portfolios for chart replacement can make
cartographer. it simplifies lighted aid selection use of this group of subroutines to stay on top of
for certain smaller scale charts which show only what's new in the world of charts and publications.
the more powerful navigation lights. Selections are made first by catalog volume number.
The numbers one through nine correspond to the 9
Module 5 doesn't really "correct" anything, but DMA catalog regions. Selections may be further
it contains information that could help save a ship defined by subregion using two digits. The excep-
and crew. It is called the Anti-Shipping Activity tion to this rule is volume 10, miscellaneous
Message (ASAM) subsystem. This file was developed Charts and Publications, which can be searched only
at the request of the U.S. Interagency Working by volume number. In any case, the user may make
Group on piracy and Maritime Terrorism. it contains all corrections to the volume or subregion or all
random reports of various forms of aggression corrections since a certain notice. Further,
against shipping around the world. Events are within the subregion query group is the option to
categorized by date and by geographic area based on query a specific notice number alone.
the DMA subregion system. ASAM reports can be
f iled with DMA using option 50 which is the f irst Programs 97, 98, and 99 are operational in nature
subroutine of the module. Step by step prompts in that the first two supply a short or long menu
help the user enter full particulars of the inci- for reference purposes and the last selection
dent to be reported and then automatically transmit actually terminates your system access.
the message to DMA over the ANMS mailbox subsystem.
Upon receipt at DMA, the text is reviewed, evalu-
ated for further action or disseminated, edited, By way of recap, you have here, readily available
and f iled in the ASAM data file for use by all at your fingertips, a full file of chart correc-
system users when needed. This is an immature tions tied to individual charts by edition date and
f ile as of this writing. It is expected to gain number, broadcast warnings with worldwide applica-
favor as a voyage planning tool in the future by tion, List of Lights, and Light List data for the'
providing cautionary ihformation to owners and world, anti-shipping intelligence information,
masters concerning security conditions in and near mobile drill rig/ship location data, and chart
ports and narrow channels around the world. Exam- catalog updates. Each program was designed to be a
ples of data in this file include the ACHILLE LAURD functional tool for the user--the mariner at sea or
incident, robberies of ships transiting the Malacca the planner in port. of course the cartographers
Straits, attacks on fishing boats and merchant within DMA and NOS received great consideration
ships coasting off Western Sahara, and certain when our system was designed, expanded and modi-
events occurring in and around the Persian Gulf fied. In fact, DMA remains the highest volume
over the last few years. remote user of the system. Several instances have
occurred wherein a user requested a specific pro-
As the world need for petroleum increases, more gram and we were able to accommodate the request,
and more offshore oil exploration has become neces- such as the Automated chart Card and the Query By
sary. Despite the present slackening of interest, a Port file. It is in our interest to continue re-
point had been reached a few years ago when there sponding to such requests whenever possible, not
were so many mobile oil drill rigs and drill ships only to create new options, but also to discontinue
plying the seas that it became extremely awkward to unused selections now that sufficient user statis-
trace their whereabouts. Thus, module 6 "Oil Drill tics are available to make competent decisions.
Rig Locations" (ODR) was created. This subsystem is
used to track nearly 1,000 of these very fluid What does the future hold for ANMS/NAVINFONET?
exploration vessels. The ODR file makes an ideal Technology is moving ahead so rapidly it is almost
supplement to the weekly radio broadcast warning futile to look too far into the future, but plan
messages which update the movement status of the one must in order to be a part of the future rather
rigs. The user may use this file in several ways. than an afterthought of history. The primary goal
If he knows the name of a particular unit, of autanating the Notice to mariners and associated
entering the correct spelling will produce the information was to be able to produce that weekly
current location, assuming the information has been periodical, cover-to-cover, by digital means. ith
accurately reported to DMA, of course. Secondly the is about 60% complete. Still to be completed are
user may query an area by subregion, which corre- all of the free text sections, chartlets (or wall
spond s to the f irst two digits of a DMA or Nos paste-on sections of charts) , depth tabulations,
chart nunber. Thirdly, he may query by broadcast weekly sailing directions corrections and miscella-
warning area; that is, HYDRoLANT, HYDRoPAC, NPVAREA neous graphics. Automation of full sailing direc-
IV or xiI. The last subroutine in this file allows tions will be accomplished in the near future using
the user to enter a rectangular area of operation the DMA Consolidated Navigation System and the
by coordinates and the NAVINFONET will search for Navigation Graphics Workstation. The major benefit
and print the names and locations of all oil explo- of these improvements in relation to NAVINFONET
ration units within the designated area. This file will be the remote access availability of the
is updated daily from a variety of sources, many weekly Sailing Directions and Coast Pilot correc-
reports of which come to bMA over the mailbox tions and ultimately, remote receipt of chartlets,
subsystem. depth tabulations and graphics.

541

A project of major concern at this time is, of cost may not sufficiently offset higher costs
.course, the development of the Electronic Chart associated with duplex operation of voice (up to
Display and Information System (ECDIS). It would 2400 baud) or data speeds. This will bear further
seem only a halfway job if a functional ECDIS were investigation, especially since someone will have
created to meet international standards without a to 'absorb the cost for the data.
means of providing corrections to the ECDIS data
while at sea. Should the United States endorse the I've given sane background on the development of
ECDIS concept and even produce an ECDIS, it will, the Automated Notice to Mariners System, discussed
no doubt, be incuibent upon the Defense Mapping the content of the various files accessible using
Agency to provide the means to keep it corrected. the Navigation Information Network, and presented a
The ANMB/NAv-iNFONET is available for that purpose few ideas and potential avenues to pursue in the
when the time comes. future. our existing systems consist of dual PRIME
2550 super mini computers with a full range of
once more, it should be emphasized that the data peripheral devices. Each is equipped with one 600
is already in the files of the ANMS, although I'm megabyte disk drive and two 80 megabyte disk
sure the format will need some revision. The oommu- drives. We have designed room for expansion and can
nications system is there in the form of NAVINFONET support future uses as well as new users of our
waiting for a heavier workload. systems.

The Federal republic of Germany has attempted to The NAviNFONET was designed for flexible exter-
study chart correction sizeing under ECDIS condi- @nal access to these vital data. Tbe mission of the
tions and has generally arrived at a f igure of 60 Navigational Aids Division which controls the
kByte to correct a 100 chart portfolio. A use of ANMS/NAVINFONET is to promote safety of life at sea
the NAviNFoNET not previously discussed is to through up-to-date, accurate and inexpensive nauti-
electronically transfer chart corrections contained cal charts and publications. Today we accomplish
in each weekly Notice to Mariners digitally from most of the in-house processing paper to paper or
Washington, D. C. to New Orleans, Louisiana. This digits to paper. Tomorrow, we'll do it digits to
process is done over a 1200 baud voice grade line digits. Satellite communications using Standard A,
using an error checking protocol and takes about 20 Standard B, or Standard C, or land based radio
to 30 minutes to complete an error free transfer. using vHF or high frequency narrow band direct
Coincidentally, the size of this transfer, wiien printing, or conventional distribution of unconven-
converted to the 10 bit German byte is 57.7 kBytes! tional products such as mailing floppy disks or
One item that remains to be resolved is how to larger printouts are potential avenues worth inves-
transfer these data to the user. INKARSAT Standard tigating. The list for tomorrow is endless. DMA
C may be a viable candidate system, but at a speed will be part of tomorrow with the Navigation Infor-
of 600 bits per second in simplex mode, its low mation Network.

542

COLD WEATHER EFFECTS UPON MARINE OPERATIONS

CAPT John D. Crowley, USCG (Retired)

Marine Consultant*
606 Montauk Avenue
New London, Connecticut 06320

ABSTRACT Ice continues to accumulate, destroying the stabil-
ity of the ship until, under the force of sea and
Cold weather causes unique and unexpected problems wind, the vessel capsizes. Annually, perhaps ten
in marine operations. These problems affect both or twelve ships are lost in cold, northern seas be-
ships and offshore structures. The paper reviews cause of this threat. Well known examples include
a few of these problems. Several incidents are the British fishing vessels LORELLA and RODERIGO,
discussed to provide examples of the severity or lost northwest of Iceland in 1955 (Reference 1),
the unexpectedness of the events. Design features the British ROSS CLEVELAND and KINGSTON PERIDOT,
and operational procedures are noted and some needs also lost in Icelandic waters, and the U.S. crabber
are discussed. GEMINI, lost in Alaskan waters in 19,80.

The most serious cases of ship icing involve freez-
1. INTRODUCTION ing spray. Spray can be the wind-driven spray from
wave tops and breaking seas, or the spray generated
Vessels and structures are brought into cold weath- by the sea striking the ship's hull. Significant
er operations in a variety of ways. In some cases, spray also arises from the interaction of the wind
icebreakers and offshore structures designed for and sea with the ship generated wave train. Of
Beaufort Sea or northern Norwegian Shelf opera- these sea spray sources, spray generated by the sea
tions, attention is paid in design to meet the spe- striking the ship's hull is by far the most
cial problems involved. Similarly, Great Lakes important in topside icing.
ore carriers and most northern fishermen antici-
pate cold weather operations; special design fea- Of lesser importance is the freezing water from
tures may be installed and cold weather operational atmospheric sources: freezing rain, freezing fog-
precautions are developed. Otfiervessels or struc- Dr snow. However, all contribute to icing and
tures have been designed to operate in warmer re- worsen the situation if sea spray icing occurs con-
gions. When the demands of commerce shift their currently. From the small data set available, it
operation to colder areas, careful winterization appears that snow in the presence of sea spray ic-
should be accomplished. Finally, there is the case ing occurs in most cases of extremely rapid icing.
of the ship caught in a one-time situation: the Whether this is true because of the interaction of
freighter for years trading between the Gulf Coast snow and freezing spray, or because the.particular
and South America, sent to pick up a cargo of news- meteorological conditions that produce extreme sea
print in the St. Lawrence in January, or the ship spray icing also generate heavy snow flurries has
caught in Boston or Portland, Maine on that odd not been established. (See POLAR LOWS below.)
week when the temperature never rises above 10 0F.
In such cases, the crew can rely only upon its own The three key factors are wind velocity air temp-
0
resources, knowledge and experience. This paper erature and sea temperature. If below about 4 C,
describes a few cold weather problems, some design sea temperature has a weaker effect on icing rate
or operational procedures to relieve the problems, than the other two. Nomograms have been developed
and the need for continuing data, development and to indicate the relation of these factors to the
communications in this broad area. rate of icing. These are usually based upon sta-
tistical analyses of actual icing cases, and the
2. SUPERSTRUCTURE ICING trends have been supported by laboratory studies
and by mixed empirical-theoretical studies.
Perhaps the most spectacular and certainly the most (References 2-6)
dangerous of all cold weather effects for small
vessels is that of superstructure icing. With high 2.a. Examples of Superstructure Icing
winds, low air temperature and low sea temperature,
ice accumulation is very rapid. For small vessels Large ships with high freeboard experience severe
caught in an extreme icing situation, severe con- icing less frequently than smaller vessels. Yet,
ditions bar on deck ice removal efforts by the crew. they are not exempt from heavy icing. Figure I is
a photograph of the bow of a Finnish Containership
*Formerly of Bath Iron Works Corp., Bath, Maine

CH2585-8/88/0000- 543 $1 @1988 IEEE

tV,

6iA

Figure 1. Bow of a containership in Helsinki Figure 2. The semisubmersible OCEAN BOUNTY in
(Photo courtesy of Lasse Makkonen) Cook Inlet (Photo courtesy of L. David Minsk)

upon its arrival in Helsinki. The mass of ice on A very severe icing event on the East Coast of
the rigging is typical of severe ice formation and North America involved the semisubmersible 0
the ten to twelve inches of ice on the deck and on SEDNETH II, drilling on the Scotian Shelf (43 30'N,
the winches reflect a severe encounter. 62034'W) on the evening of February 25, 1970.
At one time during this incident, the rig's draft
Another example of heavy icing on a large vessel is was increasing at.a rate of one foot per hour, and
described by Sachse (Reference 7): the icing of the crew was preparing to jettison several hundred
the German freighter ELSFLETH on a voyage between tons of mud, barite and drill pipe. In the actual
Bremerhaven and Baltimore. The icing began south event, the temperature rose and these extreme
of Cape Farewell, Greenland, continuing until the measures were not needed. The icing of this event
ship passed Nantucket. The growth of ice was was concentrated below the main deck of the rig.
steady and enduring. The total ice accumulated was (Reference 9)
estimated at 250-300 tons forward and an additional
100 tons amidship. The addition of such weight 2.b. Problems Created by Superstructure Icing
above the main deck not only reduces stability, it
grossly impedes any work on deck, may damage ma- The problems associated with superstructure icing
chinery, and creates both the hazard of personnel are many. The issues of stability and buoyancy
falling and the hazard of falling ice. alluded to above are familiar naval architecture
problems associated with adding high weight. In
Figure 2 shows the ice accumulation on the semisub- addition to this, icing increases the wind resist-
mersible drilling rig OCEAN BOUNTY during one of a ance of the structure, adding to the heeling load
series of severe storms in lower Cook Inlet, Alaska while decreasing stability. Design rules of the
in the winter of 1979-80. Extremely high winds, regulatory administrations and the classification
very low temperatures and the very steep seas societies incorporate criteria to meet the expected
associated with short wind fetches resulted in ice ice loading and wind. In operation, taking on sea
accretions of 2 to 10 inches per day. The spray water ballast may be an effective measure, assuming
icing in this case extended to a height of over 30 buoyancy is not an issue. If reserve buoyancy is
meters above the sea surface. During the most se- critical, ice removal or jettisoning high weights
vere icing event, total ice loads were estimated at may be,the only solutions. Clearly, the basic re-
500 tons, and it was deemed necessary to dump quirements to ensure safety must be incorporated
drilling mud to ensure adequate stability. into the design.
(Reference 8)

544

Even moderate icing creates problems. Bridge win- Most severe icing incidents take place with the
dows, even fitted with deicing features, can become wind coming off a continental shore or ice shelf.
covered by ice; safety of personnel working on deck Moving further offshore, to provide a longer fetch
is drastically impaired; and the operability-of to allow the sea to warm the overlying air, or
machinery on deck or in the rig is impaired. moving into warmer waters (out of the Labrador
Access to vital equipments: winches, windlasses, Current toward the Gulf Stream) frequently prove
ship's boats, lifesaving apparatus, deck firefight- effective methods to terminate a severe icing in-
ing equipment and valves may be attained only after cident. If the windward shore or ice edge is close
slow, laborious deicing. And the danger of damage by, moving into the lee of the shore will result
to equipment during icing exists. in a reduced.sea. However, that reduction that
reduces spray is only very close to shore, and the
Light icing or freezing snow on a radome, unless air temperature will be coldest closing the coast.
eliminated by adequate anti-icing, will cause a Each event is special, and requires the best data
blind arc decreasing the safety of navigation. For available on the current weather, oceanographic
some radio transmitting antennas, salt water ice conditions and knowledge of local geography.
on an insulator can result in an arcing path to
ground, causing both a marked attenuation and broad From the viewpoint of design, a clean deck and
band interference with other communications. superstructure, uncluttered by machinery, rigging
and other "ice catchers" is optimum. The purpose
2.c. Alleviating Icing of the unit often mandates against such total sim-
plicity. A number of steps can be taken to provide
There currently exist no magic cures for icing. a cleaner structure. On offshore structures, this
Efforts are presently underway to develop long term includes placing bottom deck stiffeners inside the
effective coatings to which ice will not adhere structure, using fewer legs for a structure when
tightly. As yet, it appears that economical coat- possible,and, as much as possible, moving cross
ings are still over the horizon. Heating for gen- bracing out of the icing zone.
eral deicing or anti-icing is tremendously costly
(first cost, maintenance costs and energy cost) and 3. POLAR LOWS
is usually fitted only for small, critical areas
and vital equipment. High pressure water jets are While addressing the topic of sea spray icing, men-
being investigated for this use, but successful tion must be made of Polar Lows. Polar Lows, or
application at sea is not yet confirmed. "Arctic Hurricanes" are small scale storms of in-
tense fury and relatively short duration that are
I have seen pictures taken a century ago of whalers usually born near the edge of the ice cap, often in
de-icing the rigging of a ship, wielding the heavy an area of high horizontal sea water temperature
cooper's mallets used to barrel whale oil, as well gradients. Like the tropical cyclone, they have
as axes and other tools at hand. In the mid-1980s an eye. While they are mesoscale or subsynoptic
the tools available for most ships are wooden phenomena, it has been found possible to identify
baseball bats, fire axes, hammers and crowbars. them using satellite data. Considerable recent
Not much progress has been made. However, studies scientific attention has focused on these lows, as
on anti-icing and deicing procedures continue. reflected by a mini-symposium in Copenhagen (1984),
(Reference 10) a Workshop on Arctic Lows at the National Center
for Atmospheric Research (May, 1985)(References 11,
Operationally, using the customary impact tools 12) and a complete issue of TELLUS, the Swedish
described above, it is important to start ice re- Geophysical Society publication (Reference 13). The
moval efforts early. The ice as initially formed Fifth International Workshop of Polar/Arctic Lows
encapsulates considerable brine and is weak. The was held last March in conjunction with the
brine encapsulations coalesce into larger brine American Meteorological Society's Second Conference
pockets, leaving a low salinity ice which quickly on Polar Meteorology and Oceanography. (Proceed-
hardens, especially if the temperature is falling. ings to be published)
In many cases, an early started and continuing ice
removal effort may control build up. If the wind For the marine operator, the nature of the Polar
and temperature allow, washdown with a solid stream Low can be appreciated by the situation encountered
of sea water from a full stream nozzle, undercut- by the Coadt Guard Icebreaker POLAR SEA on 13-15
ting the ice, is an effective method. Some ships October 1985, near Barrow, Alaska (Reference 12):
are fitted with boilers to heat this firemain
water. If the wash water itself begins to freeze, "The ship had just completed helicopter recovery
or if the deck drains are frozen, the use of sea- operations that require good weather conditions.
water in this fashion only adds to the total ice. Less than 5 hours later, the ship measured winds
If adequate steam is available (and reserve feed at 75 knots, seas estimated fifteen feet, icing
water), this can be a most effective,ally in ice of superstructure, low visibility in snow.and
removal, but does pose some hazard to personnel. the ship sustained damage from this unpredicted
For a ship, running before the sea and wind reduces storm. Excerpts from the POLAR SEA situation
both the spray coming on board and the relative report message from 13 October 1985 read as
wind that convectively cools and freezes the spray follows:
into ice. Thus, downwind courses are preferable
in severe icing situations.

545

"P 140224 OCT 85 4. COLD

FM USCGC POLAR SEA TO AIG EIGHT NINE FIVE SIX Cold, of itself, is a severe problem. Freezing of
INFO REQIFUA/COGD THIRTEEN SEATTLE WA//O// any exposed fresh or saltwater piping should be ex-
pected. Consequently, all topside firemain must be
SUBJ: AWS-85 DAILY SITREP: 13 OCT 85 secured and the hydrants and risers properly drain-
1. 14020OZ OCT 85 POSIT: 70-53N 160-42W ed. Proper attention must be paid to any topside
2. WX: TEMP; 26F SST: 32F foam systems and monitors. It is important that
BARO: 29.24R SIG: OVC/SNOW firemain shore connections be properly drained, as
WINDS: 240/67KTS VIS: .3NM the ship may encounter its coldest weather in port.
ICE: OPEN WATER SEAS: 240/7FT
CONFUSED,INCREASING. Sanitary systems and overboard discharges are
3. SUMMARY OF ACTIVITIES: subject to freezing. Steam condensate lines, if
A. ARRIVED VICINITY OF PT.BARROW LATE not properly drained, will freeze. Recognize that
P.M. 12 OCT. MAJORITY OF TRANSIT TO PT.BARROW insulation does not protect against freezing if
WAS THROUGH FY THIN AND YOUNG ICE WHICH DIS- there is no flow in the line. Thus, below deck
ALLOWED FOR ANY FURTHER GLOBAL LOAD DATA COL- piping in unheated holds or compartments, or ex-
LECTING. posed via an unheated ventilation system to outside
B. 13090OU: CONDUCTED FLT QTRS FOR PAX/ air must also be isolated and drained. Compressed
CARGO TRANSFER ----- air lines and piping, unless extremely dehydrated,
C. 131227U: RECOVERED HELOS ------ will condense water vapor and freeze. How does
D. 13170OU: WINDS SUDDENLY AND UNEXPECT- this affect design and operation? Questions of
EDLY INCREASED TO 60KTS AND CONTINUED TO IN- this sort require careful, integrated review of
CREASE TO 70KTS WITH GUSTS TO 75KTS; WINDS OF the detail design and the interplay among systems.
THIS MAGNITUDE NOT FORECASTED. OMEGA ANTENNA Further questions: Must this piping system be re-
BLOWN DOWN. BOTH ANTENNAS HAVE BEEN RECOVERED. routed or heat traced to prevent freezing? ...
ADVISED BARROW RADIO VIA H.F. OF PRESENT WINDS What spaces are affected by unheated vent systems?
/WX.BARO. Unheated vent systems can carry cold air deep into
4. INTENTIONS: a ship, causing freezing in unanticipated places.
A. PROCEED TO KODIAK. SOA MAY HAVE TO BE Even if the designer envisages that a vent cover
REDUCED IF CONDITIONS CONTINUE DETERIORATE." would be in place in bitter cold, that does not en-
sure that it will be .... even the vent fan may be
In a supplementary situation report, POLAR SEA not- running!
ed that winds in excess of 50 kts. had continued
throughout the evening, maximum seas of about 15 ft In addition to freezing, cold can cause other prob-
with freezing spray, snow and fog. Sea and air lems. Cold lubricating oil, hydraulic oil or fuel
temperature remained the same. from tanks near the shell can increase in viscosity
so as to decrease the net positive suction head at
Clearly, for a small vessel without the extreme the pump rotor low enough to result in severe
stability of an icebreaker, an unexpected storm of cavitation and rapid pump wear. If cold fuel oils
this ferocity and duration could lead to disaster. are not heated well above the wax point, and fuel
Any vessel could expect some damage to topside filters closely follow the heater, rapid fouling of
equipment and structure from the icing and the wind the filters by residual waxes is a possibility.
loads. Norwegian studies cited by Twitchell report
a maximum frequency of these storms of one or two 5. CHEMICAL EFFECTS
a week in the months of January and March. Much
study has been conducted because of the development These effects are effects of cold, but are listed
of offshore resources on the Norwegian Shelf, but separately because they all concern the effect of
similar storms occur in the Bering Sea and the Gulf low temperature on chemical processes. First, and
of Alaska. Typically, the Polar Lows are small, probably most familiar is the effect of cold upon
with a size of 300 km or less. Wind speeds in@ the lead-acid storage battery. The same low volt-
crease to over storm force in a brief time, and the age, high resistance characteristic that immobil-
storm is accompanied by very heavy snow and low izes your car on a cold morning will evidence
visibility. With the winds developing the seas itself in the cold marine environment.
over a short fetch, and with rapidly changing wind
direction, the seas would be very steep and con- Not only the lead storage battery, but dry cells
fused, a situation contributing strongly to ship or are affected. Recalling their use for emergency
structure icing. While the fury of the storm may lights, transmitters, emergency beacons, etc.
be no greater than that of many North Atlantic or raises the question, "What batteries or dry cells
North Pacific winter cyclones, the sudden, unpre- should be used for these emergency and vital needs
dicted onset and the resulting steep sea makes in cold weather?
them a particular danger.
Finally, the now common chemical light, used for
lifefloat and lifejacket markers, glows but dimly
in bitter cold weather. Special chemical lights
for cold weather use burn brightly, but with a
short life at warm temperatures. Should two lights
be used, one for cold weather, one for warm?

546

6 . TWO EXAMPLES On both the design and operation side, there is
need for more up-to-date knowledge and better
We will cite two examples of:cold weather problems. communication with the environmental science com-
Each is rather simple, unexpected, almost an oddity. munity. Designers and operators should know the
However, each caused an expensive loss. data and tools available to them and use them.
Considerable progress has been made in the last
6.a. Cargo Contamination few decades through such conferences as those on
ice and icing of the International Association for
A ship had encountered severely cold weather for Hydraulic Research, the POLARTECH Conferences,-..abd
several days prior to reaching its discharge port. the meetings sponsored by the ASME's Ocean Mechan-
Much of the last day was in broken ice. Nearing ics and Arctic Engineering Division. Effort is
the port, steam was turned on to the cargo tank needed by operators and designers to keep abreast
heating coils. It was then discovered that the of this developing technology.
coils had not been totally drained, and that con-
densate had frozen in and ruptured several coils. There is a need for a breakthrough on anti-icing
This was not discovered until after the cargo had and deicing technology. The high cost of systems
been significantly contaminated by water. Costs for occasional use results in the fitting of
included both the damage to the cargo and the re- systems that are capable of handling moderate
pair costs for the damaged heating coils. icing situations. As long as operations are con-
ducted in regions where the icing incident can be
Note that fresh water freezes several degrees above expected to be short, and to be followed by a re-
sea water, and that cold air acting on the ship's prieve to allow deicing, this is satisfactory.
hull can chill cargo spaces and tankage to a level When operations are in regions where the ice of
below that of the warmer seawater. two weeks ago is still on the structure, and is
rock hard, effective deicing will be a necessity.
6.b. Transferring Fuel Between Tanks
Finally, as always, there is need for excellent
A ship bound for Montreal had encountered some sea communications between the designer and the user
spray icing, but nothing abnormal for the area and of a ship or platform. This is particularly true
the season. The engineer on watch decided to because many cold weather problems are infrequent
transfer fuel, bringing one of the double bottom occurrences, almost oddities. This infrequency of
tanks close up to capacity, normally a routine op- occurrence is in itself a serious problem. It can
eration. Some time after the transfer had begun, result in lack of maintenance of systems designed
it was found that the tank top of the receiving to protect from cold, and a lack of understanding
tank had been set up 1@-2 inches. Investigation by operators of both the procedures and the equip-
revealed that the topside icing, usually of little ments fitted to cope with cold. Further, it tends
concern to the watch engineer, had effectively to a collective loss of memory--from design all the
sealed shut the tank vents. As oil was being way to operation--of the problems of cold. Too
pumped into the tank, the unreleased air was being often we find that we have created or discovered
compressed, finally producing a.structural loading an old problem anew. To paraphrase the late George
sufficient to buckle the tank top. Vedeler of det Norske Veritas, "Must each of us
always learn from bitter experience?"
Like cold, superstructure icing can have effects
that are apparent not only immediately at the cold 8. ACKNOWLEDGEMENTS
envelope of the structure, but manifest themselves
deep within the ship or structure. it is this sort Numerous far-flung individuals have contributed to
of complex interplay of systems that makes design the author's interest and knowledge in this area.
review of the integrated detail design essential to Several generations of Coast Guard officers with
identify potential hazards and to develop proper extensive experience in cold weather issues have
operating guidance and instructions. been generous in assistance. Particular mention
must be made of the U.S. Army's Cold Regions
7. CONCLUDING REMARKS Research and Engineering Laboratory for continuing
cooperation.
This paper has been a sampler. Its purpose has
been to provide a reminder of cold weather prob- Parts of this overview result from data developed
lems and to provoke thought on these issues. Note performing related studies for Bath Iron Works
how climatology, meteorology and oceanography have Corporation, and for the U.S. Navy by Bath Iron
a complex interplay with design and operational Works under Contract N00024-82-C-2011. The views
engineering for cold weather operations. On the expressed are those of the author, and do not re-
environmental side, there are many needs. Some flect official positions of the U.S. Coast Guard,
key phenomena are subsynoptic. As a result, their the U.S. Navy or Bath Iron Works Corp.
documentation has been poor and our present under-
standing and predictive techniques are inadequate.
In many important areas, weather data are scant,
and do not support the detail needed for operations
or prediction.

547

9. REFERENCES

1. Hay, R.F.M., "Ice Accumulation upon Trawlers in
Northern Waters", Meteorological Magazine,
Vol. 85, No.1010, 1956, p.225.

2. Lundqvist, J.E. and I. Udin, "Ice Accretion on
Ships with Special Emphasis on Baltic
Conditions", Swedish-Finnish Winter Nav-
igation Board Report No. 23, 1977

3. Makkonen, L., Atmospheric Icing on Sea Struc-
tures, CRREL Monograph 84-2, April 1984

4. Minsk, L.D., "Ice Accumulation on Ocean Struc-
tures", CRREL Report 77-17, August 1977

5. Overland, J.E., C.H.Pease and R.W. Preisen-
dorfer, "Prediction of Vessel Icing",
Jnl.Climate and Applied Meteorology, Vol.
25, No.12, Dec. 1986, pp.1793-1806

6. Stallabrass, J.R., "Icing of Fishing Vessels,
An Analysis of Reports from Canadian East
Coast Waters", National Research Council,
Canada, Division of Mech. Engrg.,
Lab. Report LTR-LT-98, June 1979

7. Sachse, K., "Ein Sonderfall von Schiffsvereis-
ung von 8stlich Neufundland his vor
Nantucket im Februar 1968", Der Seewart,
Band 30, Heft 1, 1969, pp.1-9

8. Naumon, J.W., "Superstructure Icing Observa-
tions on the Semisubmersible OCEAN
BOUNTY in Lower Cook Inlet, Alaska,
Proceedings of the Second International
Workshop on Atmospheric Icing of Struc-
tures, Trondheim, Norway, June 19-21,
1984.

9. Mycyk, 0. (Rapporteur), "Session 2: Icing Case
Studies", Proceedings International
Workshop on Offshore Winds and Icing,
Halifax, Nova Scotia, Oct. 7-11, 1985
pp. 388-9

10. LOset, S., "Investigation on Anti-Icing and
De-Icing Devices for Marine Application",
Proceedings International Workshop on
Offshore Winds and Icing, Halifax, N.S.,
Oct. 7-11, 1985, pp. 95-101

11. Kellogg, W.W. and P.F. Twitchell, "Summary of
the Workshop on Arctic Lows, 9-10 May
1985, Boulder, Colorado", Bulletin Am.
Meteorological Soc., Vol 67, No.2,
Feb. 1986, pp. 186-193

12. Twitchell, Paul F., "Arctic Lows", Naval
Research Reviews, Vol. XXXIX, No. 4,
19879 pp. 15-22

13. TELLUS, Vol. 39A, No. 4, August 1987
(Polar Low Special Issue II)

548

ARTICULATED LIGHTS IN ICE

Shawn M. Smith & Dennis Strahl

U. S. Coast Guard Headquarters
Commandant (G-ECV)
2100 Second Street, S. W.
Washington, D.C. 20593-0001

ABSTRACT joint type hinge above the sinker.) Safety
chains are attached to the light assembly
An experimental and theoretical study on and the sinker in case of failure at the
the behavior of articulated lights in mov- shackle connection. A typical articulated
ing ice covers was conducted. A scaled light is shown in figure 1.
model of a United States Coast Guard proto-
type design was tested under various icing The articulated light has the potential to
conditions. Parameters varied during test- replace numerous conventional buoys
ing included ice strength, thickness and throughout the country. The U.S. Coast
velocity. Data collected included in-line Guard's program of replacing standard
and tranverse inclination, acceleration and channel buoys with these lights is now over
axial loading at the hinge. A clear depen- five years old, with over 20 articulated
dence on ice velocity and ice strength was buoys having been deployed.
demonstrated in describing the model's
behavior. A computer program was concur- The purpose of the study of articulated
rently developed to predict in-line motion lights in ice was to model and predict the
and axial loading at the hinge. structure's response to moving ice. The
study preceeded the placing of the articu-
The United States Coast Guard has since lated lights in ice-covered waters.
deployed articulated lights modified for
ice conditions. Coast Guard experiences THE ICE DESIGN
with these structures is reviewed.
Since ice loading is expected only near or
above the waterline, the only modification
INTRODUCTION made to the articulated light pictured in
figure I was to replace the top assembly.
Articulated lights are aids to navigation On the modelf this was accomplished by
with a narrow watch circle, used to mark merely capping off the top of the steel
channels with greater precision than cylinder. On the actual light, the plat-
conventional buoys. They are compliant form subassembly was removed and replaced
structures, designed to withstand minor with a Polycarbonate Dome Ice Top Sub-
environmental loading and to yield to se- Assembly. This ice top includes ..sacrifi-
vere loading conditions. The basic con- cial" foam dayboards. A similar dome was
struction consists of a narrow vertical placed on the ice buoy deployed in the
pipe with a flotation collar providing Mackinac Straits with good success [ 5
vertical stability. It oscillates around a Figure 2 depicts the ice design.
universal coupling connected to a sinker.
The entire structure is designed to be At certain ice thickness, strengths and
nearly neutrally buoyant. velocities, the ice cover will submerge the
articulated light. This is particularly
The Coast Guard design features a series of likely in the Great Lakes where, at the
flanged steel 12" pipe sections, open to height of winter conditions, ice ranges in
flooding, a flexible foam collar to provide thicknesses up to five feet in the channels
buoyancy, a cast iron counterweight to these lights may mark. As configured for
reduce tension on the sinker, a maintenance icing conditions, the articulated light
platform with dayboards, and the light. will employ the "push-down, pop-up" con-
They are designed to mark waters 25 to 60 cept. As thin or weak ice impacts the
feet deep. structure, it will incline until the
restoring (buoyancy) force exceeds the ice
The light assembly is anchored to the sin- failing strength, at which time it will
ker with a mooring bail connected to the rebound toward the vertical position. In
bottom cylinder. A steel shackle is led stronger, thicker ice, the structure will
through the mooring bail and pinned to a incline until submerging. When the ice
steel plate cast into the concrete. (A cover passes over, or when the ice weakens,
recent design modification uses a universal the structure will "pop back up".

549 United States Government work not protected by copyright

ICE PROPERTY REFRESHER The higher values are characteristic of
hard winter ice, the lower values for
The ice property most influential in ice- spring ice.
structure interaction is ice strength. in
general, ice strength increases with TABLE 1 (Ice Strengths)
decreasing temperature, decreasing brine
content, increasing ice age, and increasing EatLgmg ipgll ggg-igni2211
thickness. Compression 50_1BOO 7S-400
Flexure 44-414 65-200
Ice is much stronger in compression than in Shear 40-132 60-120
tension and its strength is generally
characterized by its failure`mechanism, The variability of ice properties makes the
principally bending (flexure), crushing, task of modeling ice strength and failure
buckling or shearing. The vast majority of mode complex. One usually has to choose an
reports concerning ice-structure interac- average, representative property, or else
tion list crushing or bending as the domi- select a worst case scenario.
nant ice failure modes. Crushing strengths
of ice are generally two to four times that THEORY OF RESPONSE
of bending. Many engineering designs have
taken advantage of the lower bending Motion of an articulated light in wind,
strength of.ice by building marine struc- waves and current can be predicted using
tures with inclined components at the ice- techniques similar to those employed by
structure interaction point [5,13,16]. Bose and Rao [1] and Elledge [3]. In pre-
dicting the structure's response to ice
Crushing occurs when a moving ice mass loading, these techniques were utilized,
strikes a vertical or, nearly vertical with the added assumptions that the ice
structure. Michel and Toussaint (11] impacting the structure is an isotropic,
demonstrated three zones of crushing constant thickness, constant strength
behavior: ductile, transition and brittle. cover; and that waves will not occur cnn-
Although ice failure occurs in all thre ,e currently with ice. It is realized that
zones, crushing strength generally such assumptions reflect only a portion of
increases with increasing strain rates the conditions in nature; however they were
through the ductile zone, peaks in the made to simplify the calculations.
transition zone, decreases and then levels
off in the brittle zone. Under conditions The articulated light is first divided into
of quick loading, ice behaves like a discrete increments; drag forces due to
brittle material. wind and current calculated; and moment
arms calculated about the hinge (shackle).
Ice failure by bending on inclined struc- Added mass and the structure's mass moment
tures is dictated by a decrease in the of inertia are calculated, together with
horizontal component of ice pressure and the buoyancy force and gravitational
the appearance of a vertical component forces. Finally, the angular acceleration,
proportional to the angle of structural change in angular velocity and the resul-
inclination 191. The failure of an ice tant change in the light's angle of incli-
sheet in bending is generally preceeded by nation are determined. These calculations
the formation of radial cracks propagating are repeated at predetermined time inter-
outward from the loading point. Circum- vals, and the incremental change in incli-
ferential cracks then form as the ice sheet nation applied to the previous angle to
is bent, until the flexural strength at one arrive at the instantaneous inclination.
of these circumferential cracks is exceeded This method was proven to be successful in
by the load an the ice. The distance of predicting a model's motion by Elledge [3].
this failure is apparently related to the
characteristic length of the ice, a proper- In predicting the li,ght's response to ice
ty related to its elastic modulus. loading, the computer program written by
Elledge was adapted to include an ice
Tracks behind isolated structures in ice loading moment. The impacting ice sheet
fields can often give a clue as to the mode heels the structure over until the ice
of failure which occured [6,8]. A path fails, or the structure is submerged. it
roughly equal to the structure diameter, was assumed that the ice would fail in one
with small rubble piles on either side, is of two modes, bending or crushing. The ice
generally the result of ice failure by failing forces and resultant moments for
crushing or shearing. Bending failure is these modes were calculated. Then, the
characterized by a series of circumferen- other environmental and structural moments
tial shaped cracks on either side of the are summed. If these moments exceed the
track. ice failing moment, then the ice is assumed
to fail. If not, then the light
Table 1 lists extreme ranges of ice inclination increases as determined by the
strengths measured by various authors as velocity of the ice and the selected time
well as recommended design values [10]. increment. The forces and moments are

550

recalculated, until ice failure occurs. The model was tested under a variety of ice
conditions at the U.S. Army Corps of Engi-
The calculation of which angle of structure neers Cold Regions Research and Engineering
inclination the ice failing mode ..shifts" Laboratory (CRREL) in Hanover, N.H. The
from crushing to bending posed a problem CRREL test basin measures 30 feet wide, a
since there is no well defined threshold. feet deep and 120 feet long. Ice is grown
Danyaa and Bercha [2) found that ice fails to specification by variation of room tem-
by crushing at an angle of up to 15 degrees perature and time to attain desired ice
from the vertical. From 15-30 degrees a thickness and strength. Ice was seeded
crushing made could be assumed, yielding using a urea doped water solution of
conservative forces. Between 30-45 degrees approximately one percent concentration, as
a bending failure could be assumed. Saeki described in Hirayama [7]. Ice sheets are
et al (13) found 0-10 degrees to be pushed along the tank by means of a D.C.
crushing; 10-25 degrees transitional; and motor driven carriage which spans the width
above 25 degrees, bending. of the tank.
Obviously, the type of failure which takes.
place depends on which kind of ice Just prior to testing, the ice was sampled
strength, bending or crushing, is surpassed to determine its thickness, flexural
first. For the purposes of this study, an strength and elastic modulus. The flexural
angle-dependent formula was derived which strength was measured using a cantilever
assumed failure to be somewhere between beam test; the elastic modulus was deter-
crushing and bending. mined by measuring the characteristic
length as described in Kato and Sodhi 181.
FORCES TRANSMITTED TO THE HINGE The crushing strength of ice was then cal-
culated, -again following a relationship
Since the articulated light is designed to reported in Kato and Sodhi, which uses the
quasi-neutrally buoyant, the hinge will downward flexural strength to arrive at the
normally see only a small load due to net compressive strength of ice.
buoyancy (total buoyancy minus total
weight). This force should remain somewhat In addition to instrumented measurement of
constant as the structure inclines since the structuve's motion, an inclination
the hinge is moment free. If ice impacts panel was mounted behind the structure from
the structure however, an axial force a foot bridge which spans the test basin.
component will arise due to friction A video tape record of the tests war. then
between the ice and structure. made.
Additionally, if the structure becomes
frozen-in, potentially high forces would be TEST RESULTS
generated in breaking the adfreeze bond
(4,12]. Friction forces were calculated Tests were conducted on five ice sheets,
using the standard friction formula, with three velocities per sheet, except the
modified by the instantaneous angle of fifth (2 velocities), yielding fourteen
inclination of the structure. The separate tests. Table 2 presents the model
resultant axial load can then be resolved ice characteristics and velocities used:
into vertical and horizontal components to
ascertain forces imposed on the hinge and TABLE 2 (Ice Sheet Characteristics)
sinker system.
sheet Thlglinggg gtrgugth YgIggitz
MODEL TESTS 1 1.5 in 4.2/S.5 psi .16,.32,.49
11 2.0 in 3.8/7.0 psi .08,.32,.49
A model of the prototype articulated light 111 1.3 in 4.8/10.4 psi .08,.32,.66
was constructed based on Froude scaling and IV 1.2 in 7.8/20.9 psi .16,.32,.66
a water depth of 30 feet. The same V 2.0 in 11.9/21.0 psi OBf.32
materials were utilized for the model as
a r d- found on those lights currently Note: Strength values are upward/downward
deployed: steel pipe, cast iron counter- flexural strength. Velocities are ft/sec.
weight, surlyn foam and the paint system. Strength is pounds/square inch. All values
A battery compartment and the light were rounded off.
not modeled, allowing room instead for
instrumentation. A 1000 lb capacity load On the first four ice sheets, the model was
cell measured axial loading at the hinge; cleared from all ice prior to testing. On
inclinometers were placed orthogonal to the fifth ice sheet, the model was left
each other to measure motion in the in-line frozen-in to simulate a similar situation
and transverse planes of motion; and torque in the environment. A summary of test
balance accelerometers were installed to results is given in Table 3. Angles of
measure structural vibration. A ball joint inclination given are the maximum angles,
was used to attach the structure to the in degrees, from the vertical achieved.
tank bottom. The failure mode is the dominant mode
observed during the tests.

551

TABLE 3 (Model Test Results) length. Figure 6 shows a test run pre-
Sheet Velocity Ang19 f@IjVrg Mgjg dominated by bending failure of the ice.
1 .16 fps 41.5 Bending Figure 7 demonstrates crushing failure.
1 .32 fps 44.0 Bending
1 .49 fps 45.0 Crushing THEORETICAL FINDINGS
11 .08 fps Submerged
11 .32 fps Submerged The computer program written by Elledge [3]
11 .49 fps Submerged was adapted to predict response to ice
111 .08 fps 28.9 Bending loading. Emperical results from the model
111 .32 fps 33.5 Sending tests were added to refine the program,
111 .66 fps 45.0 Crushing specifically the maximum ice failing angle
IV .08 fps 43.9 Bending and the angle of rebound. The computer
IV .32 fps 39.3 Bending model predicts reasonably well the struc-
IV .66 fps 45.0 Crushing ture's response, and in every case,
V .08 fps Submerged predicted the structure's submergence.
V .32 fps Submerged Figure 4 provides a sample comparison of
the computer model results (solid, straight
The typical test proceeded in the following line) with visual records (dashed line) and
manner: The ice was set into motion and model test results (solid, jagged line).
impacted the structure with no visible
structural rebound. The structure inclined FIELD EXPERIENCE
at a steady rate. Approximately 5-10
degrees before failing, the ice surrounding Several articulated lights have been
the structure rode up the cylinder positioned as aids to navigation in ice
,slightly, forming a circumferential crack. environments, including one in the
The ice then failed about this crack and Chesapeake Bay and two in the Great Lakes.
split to either side as the structure Only one of these, the Lake St. Clair
rebounded. The rebound did not stop at the structure, has survived the winters
site of the initial circumferential crack, relatively unscathed and it has been in
instead the ice was crushed/sheared for place for two years.
some additional distance before stopping
and repeating the process. To prepare for winter, the top assembly,
including the light, dayboards, lantern
A typical in-line inclination plot is shown stand and platform, is replaced by the
in Figure 4 (jagged line; the other lines Lexan ice dome assembly shown in figure 3.
will be explained later). The structure In the spring, the process is reversed.
inclines as the ice pushes against it, The Chesapeake Bay Bridge articulated
until the ice failing force is exceeded. light, on one occasion, became unshackled
The structure then rebounds, perhaps from the sinker. On another occasion, the
failing the ice several times (noted in the foam collar was found to have absorbed much
visual record) before the process is more water than usual. Ice conditions the
repeated. Little transverse motion was past two years in the vicinity of this
observed, as might be expected in the light had been on average one to four inch
uniform ice sheets produced in the lab. pack ice, generally with about seventy
percent coverage (although ice floes coming
Typical load cell readings are depicted in off the Chop Tank River were observed to be
Figure 5. Generally, the load decreases up to 12" thick). Upon one inspection,
with increasing angles of the structure, the structure was observed leaning 20-30
after an initial peak at about 12 to 20 degrees in 3" running slush ice, with newly
degrees. When the structure rebounded, the installed "sacrificial" dayboards missing.
load increased. This is most likely due to
the decreasing axial component of buoyan- The Munuccong Channel Junction light was
cy. There is one major ice failure, either discovered off location and a larger sinker
bending or crushing, followed by a series was retrofitted to provide a larger bearing
of smaller failures. surface. The light had apparently been
dragged along the muddy bottom. On
The average angle of inclination before ice another occasion, the sinker bail was
failure increased with velocity, as did the observed to have worn excessively, probably
frequency of these failures. There ap- due to loss float floatation causing in-
peared to be a shift in the type of failure creased stresses at the hinge. The light
as well, with bending predominating at the has been removed from station. Average
lower velocities and crushing at the ice thickness in the channel during the
higher. This agreed& with the findings of height of winter is five feet.
Haynes et al [6). Bending failure occurred
at higher angles, crushing occured at the On the first occasion to inspect the Lake
nearly vertical angles. As might be St. Clair Light after installation, the
expected, the failure angle increased with Lexan dome was discovered smashed and ice
the elastic modulus and the characteristic was packed around the top of the light.

552

Ice conditions at the time of inspection REFERENCES
were 6-8 inch of refrozen pancake ice, 9/10
coverage. On another occasion, the Light's [1] Bose, K. and Rao, E., "Development of a
Ice Dome assembly broke off at the flange Theory for Analysis of Articulated Beacons
above the float. Ice conditions were for use as Offshore Light Structures",
severe - up to six feet of moving packed Bulletin de L'A.IS.M., P. 13-19, 1979.
brash ice, full coverage. The floatation
collar was also found to have been torn (2] Danyas, J. and Bercha, F., "Investiga-
slightly and the dayboards were destroyed. tions of Ice Forces on a Conical Offshore
After three months on station, the light Structure", Ocean Engineering, Vol.3,p.
had been dragged about 100 yards. 299-310, 1976.
Of the original 23 deployed lights, only 10 [3] Elledge,L.,"Dynamic Study of Articula-
still remain an station. The other, non- ted Beacon",Thesis Dissertation, URI, 1983.
ice lights, suffered from much the same [4] Gaythwaite, J., The Marine Environment
problems as the ice structure, principally
failure at the hinge/sinker joint. and Structural Design, 1981.

MODIFICATIONS [5] Glahe, P., Recent Developments in a
Navigational Buoy for use in Ice Condi-
Based on these experiences, several tions, U.S.C.G. Report, 1980.
modifications are planned, with some
already implemented. First, the dayboards (6) Haynes et al, Ice Forces on Model
have been changed to Surlyn foam Bridge Piers, CRREL Report 83-19, 1983
"breakaway" dayboards. Second, the foam
float collar will be made longer and thin- [7] Hirayama,K., Properties of Urea-Doped
ner to facilitate buoy tender deck handling Ice in the CRREL Test Basin, CRREL Report
operations. It will also be made of a dif- 83-8, 1983.
ferent material. Shrinkage of the original
foam floats had occurred, decreasing the [8] Kato,H. and Sodhi, D., Ice Action on
volume of the float and thereby reducing Pairs of Cylindrical and Conical Struc-
the net upward buoyancy. These floats also tures, CRREL Report 83-25, 1983.
were subject to water intrusion due to
inadequate "close cell" properties, further [9] Michel, B., Ice Pressure on Engineering
reducing buoyancy. Structures, CRREL Monograph III-Blb, 1970.
More flexible (neoprene) and reliable uni- [10] Michel, B., "Advances in Ice Mecha-
versal bearings will be used at the sinker- nics", P.O.A.C. 1981, p. 189-203.
light connection. Problems with the sinker [11] Michel B. and Touissant N.,"Mechanisms
bail are partially a secondary effect and Theory of Indentation of Ice Plates",
caused by the loss of float buoyancy. Re- Symposium on Applied Glaciology, 1976.
duced buoyancy results in the loss of pro-
per tension on the hinge, leaving the en- [12] Sackinger, W. and Sackinger, P."Shear
tire weight of the structure "bottoming Strength of the Adfreeze Bond of Sea Ice to
out" against the sinker bail. Ultimately Structures", P.O.A.C. Vol. II, p.607-614,
the sinker bail was chipped out of the 1977.
concrete and failed at the welded joint.
[13] Saeki, H. et al, "Experimental Study on
A thicker lens will be used for the Lexan Ice Forces on a Cone Shaped and an Inclined
Dome and longer top sections will be Pile Structures", P.O.A.C. Vo1 II, p. 695-
mounted to give the dayboards a greater 706, 1979.
height of eye. Finally, thicker flanges
will be used to prevent shearing off of the (14] Smith, S., "Response of Articulated
top section due to the severe bending Beacons in Ice", Thesis Dissertation,
moments caused by collisions and ice URI.,1984.
loading.
[15] Strahl, D., "Articulated Lights", USCG
It is hoped that these modifications will ATON Bulletin, Vol 17. No.2, p.10-11, 1988.
solve the major problems encountered with
the articulated lights. Five of the new [16] Tryde, P., "Ice Forces Acting on In-
designs are scheduled for field testing, clined Wedges and Cones", Ocean Engineer-
although none of them will be placed in icy ing, Vol 3, 1976.
locations until the designs are proven in
the more benign environment. Ideally, the [17] Walker, S., "Articulated Beacons Could
articulated light will eventually be placed Replace Buoys", Commandant's Bulletin, 14-
as a year round navigational aid in criti- 82, U.S. Coast Guard, p. 30-31, 1983.
cal navigation locations.

553

U@
FM&PL/M -T

wwo

:rT

Figure 1 C.G. Articulated Figure 2 Articulated Light
Light Design. Ice Top Design.

Figure 3 -,Ice Dome
Assembly.

COAST GUARD CG-L412 C AST GUARD CG- I all
;4 1.0 Y. Be.,. ca V.I. 98.00.1.
Stop-Glli WiSOH-s. Stop-88 I DOI 0M.G.

45.0 [Be

Ic
I S , 05;__:

10.8 4E

C 9.0 26,

0
75
-9.0 SECONDS -2 SECONDS
c-'e.8 -4

27.0 -6
136.0 -8

-45.0

Figure 4 In-line Inclination Plot. Jagged Figure 5 Plot of Axial Loading at Hinge.
line is inclinometer record; Ice Sheet III.
dashed line is visual; solid line
is theoretical. Ice Sheet IV.

1@4," y4f,

1,41i
4 44 4 fla A

AA
Of
6T

W @0

Figure 6 Typical Ice Failure by Bending. Figure 7 Typical Ice Failure by Crushing.

654

ARCTIC ICE ISLAND CORING FACILITY

Michael Gorveatt
Mark Chin Yee

Bedford Institute of Oceanography
P.O. Box 1006
Dartmouth, Nova Scotia
Canada B2Y 4A2

ABSTRACT delivered to the PCSP in Resolute Bay, but for
logistical reasons, the 1600 metre runway was not
In 1982 a three-by-seven kilometre iceberg calved completed and the planned delivery by Canadian
from the Ward Hunt ice shelf on northern Ellesmere Forces Hercules aircraft had to be delayed. The
Island and began moving west into the Arctic same feat was attempted in 1986 with the same
Ocean. The Canadian Department of Energy, Mines result. PCSP then negotiated with the Canadian
and Resources used this as an opportunity to take Forces to put the winch and a bulldozer on the
sediment samples and piston cores from this area island via a Low Altitude Parachute Escape System
of the Arctic Ocean and other areas as dictated by (LAPES) drop. This was successfully carried out
the track of this Ice Island. With the assistance in the early spring of 1987.
of the Engineering and Technical Services Division
of the Department of Fisheries and Oceans, a BUILDINGS
substantial sediment sampling system was de-
veloped, built, and installed on the island. There are many different buildings on the Ice
Island, each with special functions. Buildings
This sediment sampling system consists of a winch, that house essential camp functions include the
gantry, and ice melting system capable of taking generator hut , the radio /of f ice hut , the
piston cores to a water depth of 4000 metres. navigation hut, the sleeping huts, and the all
Sampling is carried out through a 1.5 metre important dining hut. These buildings are the
diameter hole melted through ice 44 metres thick responsibility of the PCSP. Other buildings
by a 249 kW melting system. This paper describes designed specifically for piston coring and
the evolution, the logistics, and capabilities of sampling were constructed by the Atlantic
the sampling system. Geoscience Centre (AGC). These AGC buildings
have grown into quite a complex. The main
building is the coring hut which houses the winch,
gantry, hydro-hole, office, and core working lab.
INTRODUCTION Attached to this is the pump room which contains
the hydraulic power pack to run the coring winch.
The indication that another ice island had calved on the other side of the coring hut is the hut
from the large Ward Hunt ice shelf was discovered which houses the oil-fired hot-water system used
by the geographer and , ice specialist Martin to melt the hydro-hole in the ice and to keep it
Jefferies during his 1982 summer research on open for the summer sampling program. This
northern Ellesmere Island. George Hobson, building also contains the diesel generator which
director of Canada's Department of Energy, Mines provides the power for the coring facility. Part
and Resources Polar Continental Shelf Project of the complex is a 3 by 4 metre Weatherport tent,
(PCSP) identified this ice island on August 11, which is used for storage. Each of the buildings
1983. At this time it was offshore of Ayles Fiord on Ice Island is surrounded by ablation shields,
and moving to the southwest. A telemetering buoy which are sheets of plywood, painted white to
was placed on the island so that overflying reflect the heat from the sun to retard melting
aircraft could plot its location until the PCSP and erosion around the footings during the summer.
were ready to build and man a station. (Fig. 1)

The Ice Island, [named Hobson's Choice after its HANDLING GEAR
founder] , measures approximately 3 by 7
kilometers, is 44 meters thick, and presents an The semi-permanent nature of an installation on
excellent semi-permanent floating station from an ice island of this type permits the use of
which to study the relatively unknown Arctic handling equipment with somewhat greater sampling
Ocean. In 1984, plans were made and funds were capability than might be found on most arctic ice
found to set up and man a small camp on the island camps. The primary concern is that erection is
to operate a gravity corer and a 3.5 KHz sub- possible using a limited variety of hand tools
bottom profiler. Piston coring was. attempted in with little or no machinery available for
1985 without success. In 1985, a new winch was carrying, lifting, or placing of components.

CH2585-8/88/0000- 555 $1 @1988 IEEE

Most operations involve lowering sampling devices to ease handling of the core barrels and
through the hydro-hole using the winch located at heat probe.
one end of the coring hut. The cable is led e. The supports should be detachable at ice
almost horizontally from the winch, through an level in case the coring building had to be
intermediate (heel) block, up over the main sheave moved on its skids because of excessive
block hanging from a gantry, then down the hydro- melting or deterioration of the hydro-hole.
hole (see Fig.2). The use of the heel block
minimizes the side loading on the gantry, and SUPPORTING STRUCTURE
reduces the overturning moment on the winch.
The steel structure is assembled from pipe
Construction of the coring facility took place sections with bolted flange joints. Each section
during the spring of 1985, amidst all the other weighs a maximum of 70 kg so that they could be
preparations for the sampling program on the handled by twopeople. The winch, heel block, and
Canadian Ice Island. A smaller, more portable gantry are mounted on columns frozen into the ice
winch which had been used on earlier ice camps, to a depth of 2.0 metres. The columns are fitted
was installed until the new winch could be with flanges below ice level in order to increase
delivered. The temporary winch was attached to the the pull-out strength by increasing the shear area
wooden beams supporting the floor of the building, of the supporting ice. Where the columns protrude
with most of the load taken by a restraining cable through the floor of the coring hut another
shackled to the back of the winch and anchored to flange connection is provided at ice level to
a pipe frozen into the ice. The heel block was allow removal should the building have to be
held in place by a heavy wire strap wrapped around moved.
a structural steel member spanning two of the
floor joists. Positioned over the hydro-hole is a The gantry was raised by three meters, on columns
standard portable workshop gantry, which straddles which are cross-braced by wires and turnbuckles
the coring hut. The gantry is rated at 4500 kg for lateral stability, to give approximately 6.8
and was mounted on a one-metre high wooderi meters headroom over the hydro-hole (Fig. 3). A
trestle frozen into the ice. The gantry also had jig was constructed from available timber on site
guy wires attached to the base of the building at to locate and support the columns during freezing
the opposite end from the winch in order to resist in. The one utilization of mechanical lifting
any side loading on the structure. assistance was the use of a helicopter in removing
and re-positioning the gantry atop the steel
A limited sampling operation was carried out structure (fig. 4). Finally, the top of the
during the late spring and early summer of 1985 gantry was guyed as added security against side
which included bottom photography, plankton tows: loading.
grabs, dredges, and gravity coring. Finally in
August, with the necessary equipment assembled to The heel block is shackled to a pivot held by two
do piston coring using an 820 kg core head, a structural members frozen two metres into the ice.
trial run was conducted using a 3 meter barrel. Based on our previous sobering experience, much
Everything appeared to go smoothly until pullout debate took place about the way that the heel
was attempted. The resulting line tensions pulled block support was to be installed. It was decided
the heel block up through the floor of the coring to angle the bottom of the columns away from the
building, which in turn altered the rigging hydro-hole to avoid the weakened ice around the
arrangement. The cable angle now produced an hole, and also the possible melting that may occur
upward force on the winch, unresisted by the in that area. An additional flange was installed
horizontally acting restraining cable. Despite on each support to resist the potentially high
valiant attempts to pull out the core barrel stuck pull-out forces.
in the ocean bottom, the wire cable had to be cut
in order to save the winch, which by now was The winch is bolted to steel beams running under
steadily creeping toward the hydro-hole. It did both sides of the base and spanning the columns
not take much imagination to see that piston located near each corner of the winch. The
coring was going to require the coring winch that original intention was to place the supporting
was stranded in Resolute Bay, and a completely columns through holes in the existing floor, but
redesigned support for the winch, heel block, and site conditions made it simpler to mount the winch
gantry. outside, then to extend the building at the winch
end. This also added some much needed room to the
A number of options were considered for the working area inside the hut.
supporting structure with certain constraints
placed on the design: CORING WINCH
a. The forces must all be transferred to the
ice, not to the building. The hydraulic winch is a robustly constructed ma-
b. All components and structural members must chine capable of bare-drum line pulls up to 28.5
be light enough to be hand carried and kN (6400 lbs) at 55m/min (180ft/min), when powered
assembled. by the dedicated 43 kW (58hp) diesel /hydraulic
c. Steel is the preferred material since some pump unit (see Fig. 5). The primary hydraulic
basic steel welding is available on the Ice system utilizes two pumps in parallel, delivering
Island should modifications become oil to an open loop circuit. Whenever the
necessary. horsepower limit of the prime mover (the diesel
d. The height of the gantry should be increased engine) is approached, one of the pumps is

556

automatically unloaded, which effectively drops ships there is mechanical assistance available to
the winch speed by 50%, but doubles the line pull help with the moving and rigging of the corer.
capability by allowing the hydraulic pump to The second is there are no real constraints on
attain higher pressures without stalling the the size of the piston corer assembly used from a
diesel engine. The winch has many of the basic ship since there is usually enough space to rig
features expected of a hydraulic winch: the corer and collect the sample. The third i s
a. infinitely variable speed control, with good most ships carry a deck crew to assist the scien@-
low speed high torque characteristics tist in handling the corer and operating the
necessary for this type of operation; mechanical handling equipment. None of these
b. excellent resistance to wet environment; things exist on Ice Island. Chain blocks and
c. high power to size ratio; come-alongs" are used to compensate for the lack
d. self-contained and simple in design. of mechanized and manual help.

The drum can hold up to 4000 meters of 9.5 mm (3/8 The problem with the trip arm arrangement for the
in.) diameter cable, and is supported on one end corer was not as easy to overcome. Due to the
by a hollow stub shaft suitable for slip ring limited size of the hydro-hole, it was a very
mounting if used to handle electromechanical tedious and nerve-racking experience to lower a
cable. On one side of the winch is an independ- piston corer with the trip arm only inches from
ently driven warping head (capstan) which is the side of the hole. Contact with the side could
convenient for handling short lifts around the pre maturely trip the corer with disastrous re-
hydro-hole. sults to equipment and personnel. On average, the
sides of the hydro-hole would grow inward at the
The winch and pump units are welded steel fabri- rate of about 5 cm per week, so reaming the hole
cations with frames that completely surround the had to be done often to ensure the corer would not
drum and drive components. This design provides catch on the ice walls. After many long discuss-
strength, rigidity, and protection for vulnerable ions of this problem we decided to streamline the
parts, which proved critical when the winch was corer. This would allow us to cut and maintain a
air dropped onto the ice in 1987. The undersides smaller hydro-hole. Our second consideration was
of the frame are enclosed skid or toboggan shaped safety, which meant eliminating the trip arm
.to facilitate dragging the units across the snow. assembly. This was done by using an acoustic re-
lease mounted just above the core head in the
MELTER same manner as the trip arm assembly. Using a
separate pinger (EDO 462;12KHz) attached a known
The melter system (Fig., 6) used to cut the 1.5 distance from the end of the corer, we could
meter diameter hole through 44 metres of ice is a watch the corer's descent and with the aid of a
large and powerful system. This melter uses re- graphic recorder, control the distance between
heated melt water as the melting agent. Since the corer and the ocean floor. When the corer was
this water will become sea water, which would the prescribed distance off bottom (5 m) we could
eventually corrode and contaminate the boilers, a talk acoustically to the release and trigger the
system of transferring heat from a glycol closed corer. This system worked very satisfactorily.
loop to the seawater melting medium was devised.
This allowed a mixture of glycol and water to be There were still problems de-rigging the corer.
heated in the two boilers and circulated through a Limited room in the core hut made it impossible to
stainless steel heat exchanger by a 5 HP steel lay the corer down horizontally to unscrew the
circulator pump (Fig. 7). A second circuit ear- barrels. These had to be separated in the vert-
ploying a 1.5 HP submersible vertical turbine pump ical mode, resulting in the loss of 6 to 10 cms of
in a stainless steel housing draws the surface core before we could cap off the upper barrel.
water from the hydro-hole and sends it to the heat Since the core system was designed to be used at
exchanger via a 4 cm diameter low pressure wash- sea, the barrels were screwed together for greater
down hose. The salt water collects its heat from bending strength. Due to the relatively stable
the heat exchanger and returns to the melter conditions when coring from an ice island, it was
head, the descent of which is controlled by a decided that this lateral strength was not
small winch in the coring hut. A descent rate of required. Several systems of connecting the bar-
25 cm every 10 minutes will produce a hole no rels were explored and a system of movable coupl-
less than 102 cm diameter. The boilers are each ings attached by set screws was used. Whe n t he
capable of producing 249 KW (1.2 millon BTU/hr) barrel coupling arrived at the disassembly point
of heat. just above the hydro-hole, the lower barrels were
clamped and lowered to a support rail. The set
CORER screws were then loosened and the whole sleeve
was allowed to slide past the pipe joint. A core
A modified Benthos piston coring system is now cap was then slid under the end of the upper bar-
being used on the island (Fig. 8). The f irst rel and secured to the pipe without the loss of
attempts at coring employed a Benthos 2171 gravity any sediment. The single barrel was then placed
corer with a 1.5 meter core barrel. The first on its side and the liner removed for cataloguing
piston core attempted was with a 6 meter Benthos and packing.
2450 system with a Benthos 2171 gravity corer as
the trip weight. There are some significant
differences between piston coring from a ship and
coring from an ice island. The first is on most

557

CONCLUSION

Due to the unique problems involved with coring in
the Arctic, many modifications have been made to

the standard shipborne coring system to improve
its safety and performance. This system can now
be operated by two persons and up to three cores a
day are possible. Besides piston coring, the
winch system is capable of handling heat flow
equipment, sediment grab samplers, camera
equipment, and gravity corers. Future plans
include side scan sonar and sediment drilling.

1100 1000 900
820) -820
C5

1985

1986

1987
10,
800- 800
.0 AXEL
HEIBERG
ISLAND

5/06/88

1010
0 '@?o
1)>
PRINCE
'00 GUSTAF
ADOLF
SEA -1 . ......

.1X

780.-.
-780
'Z
63

1100 00 0 900

Fig. #1 Ice Island Track
From 1985 - 1988

558

PLAN VIEW

HEEL BLOC
K HYDRO-HOLE

STON CORER

L
2@ GENERALPURPOSE LCORINGHUT '---GANTRY
HYDRAULIC WINCH

ELEVATION
GANTRY

.GENERAL PURPOSE CORING HUT
HYDRAULIC WINCH

HEEL 'PIMOORER

DYNAMOMETER

.......... .

....... .

........... ",
TO-H

Fig. #2 Coring Facility

'Yo, 001-50@

t M

1
,A
1A
A
il"MA I

Fig. #3 Gantry

Fig. #4 Placing Gantry

559

MEN

Ll
Fig. #5 coring Winch
Fig. #6 Melter System

CABLE
CLAMP
CABLE

100 gal.
EXPANSION ACOUSTIC
TANK RELEASE

HEAT
EXCHANGER
'0

@INL@ r IN@ INLU

1.2 M.B.T.V. 1.2 M.B.T.V.
BOILER BOILER

5 H.P.
CIRC.
PUMP PINGER

r - - - - - - - - - - - -
CORE HEAD
T'
QuTt
BLE
-50%GLYOOL/
1----SALTWATER

CORE
Fi 9 , '77 Schemati -c. of Melter BARREL
Now

System
Fig. #8 Piston Coring System
HEAT
L EXCHALNGER
_[ __ _0
[_r N;-
N_ N__
INL- @r IN. _,N_

560

OPERATIONAL ICEBERG FORECASTING CONCERNS

Walter E. Hanson

INTERNATIONAL ICE PATROL
Avery Point, Groton, CT 06340-6096

ABSTRACT As evidenced by the many years of safe passage by
trans-Atlantic shipping, the IIP seems to have
The International lee Patrol uses parametric some skill in determining the extent of the ice
deterioration models to forecast iceberg melt danger. To quantify this skill has been hard
rates in the vicinity of the Grand Banks of because of the need for accurate data to represent
Newfoundland. These predictions, sometimes the initial iceberg, and interim iceberg and
extending as long as two weeks, are used to aid in environmental conditions. Between 1983 and 1985,
resighting or deleting icebergs. The daily the IIP studied the drift and deterioration of
forecasts ' are used along with reconnaissance four icebergs. Although no firm conclusions could
efforts to establish the limit of the ice danger be drawn from such a small data set, which
to trans-Atlantic shipping. in June 1987, the represented an average drift of 4.5 days, the
Patrol conducted a week-long study of iceberg prediction models did a fair job hindcasting the
deterioration in the Labrador Current on the drift and deterioration of the icebergs when
northeast Newfoundland Shelf (50-45N, 53-30W). A actual observations were used as inputs (3). A
cluster of six medium-sized icebergs was studied. similar study was performed, using U. S. Coast
Iceberg histories were compared to operational Guard iceberg data, for the Atmospheric
melt predictions. Environmental data from the Environment Service of Canada (4). Again the
U.S. Navy's Fleet Numerical oceanography Center, results were mixed. Thus in June 1987, the IIP
the Canadian Forces Meteorological and conducted another cruise to collect similar data
oceanographic Center and on scene observations for a longer period on a cluster of icebergs.
were compared and used to hindcast iceberg melt.
The findings are used in a discussion of the The objectives of this study were to compare
Patrol's system input errors and efforts the iceberg deterioration predictions derived from
Patrol is pursuing to reduce them. inputs collected in situ to inputs available from
operational data centers. These latter inputs
were divided between global and regional scale
products.
1. INTRODUCTION
2. BACKGROUND
Since 1971, the International Ice Patrol (IIP) has
used computer-based drift prediction models to The iceberg deterioration model used by the IIP
assist in evaluating the extent of the iceberg provides the watch officer with an estimate of the
danger in the vicinity of the Grand Banks of "melt" rate of each iceberg entered in the drift
Newfoundland to North Atlantic shipping. A model. The computer-based application (2), vihieh
dynamic model began operational use in 1979 (1). computes the melt rate, is derived from the Coast
This model, along with a parametric iceberg Guard Research and Development Center Report No.
deterioration model which began operational use in CG-D-62-80, "Theoretical Estimate of The Various
1983 (2), has grown in importance as limited Coast Mechanisms Involved In Iceberg Deterioration In
Guard resources restrict iceberg reconnaissance to The Open Ocean Enviroment" (5). The model sums
an every other week schedule. During the peak of the effects of:
the iceberg season, April through June, the limits
of the iceberg danger expand rapidly, requiring . solar radiation;
reconnaissance missions to concentrate on . buoyant heat convection;
patrolling the limits. The majority of the . heat convection caused by iceberg movement
iceberg population is reconnoitered as time relative to the water mass (forced beat
permits. often icebergs go one to two weeks convection); and
before being resighted. Resighting icebergs o waterline wave erosion, followed by calving of
depends heavily on these models effectively the resultant ice overhang.
predicting drift and deterioration rates. These
predictions are also routinely used to set the Based on the 1980 report, warm air heat convection
limit of all known ice twice daily, as reported in is considered insignificant and not calculated.
the international lee Patrol bulletins. The 1480 report also identified other iceberg

561 United States Government work not protected by copyright

deterioration processes; however, they were only the drift and deterioration of six icebergs which
partially addressed and difficult to quantify. were within a circle of approximately 55km radius
Consequently those processes are not modelled. made both compilation of environmental factors
affecting each iceberg and verification of iceberg
The model calculates melt in terms of length identity late in the experiment difficult. The
versus mass. This measure of melt accommodates distance TAMAROA was from each iceberg was
IIP's operational constraints, in which nearly all paramount in determining the applicability of
iceberg dimensions are only reported qualitatively environmental observations. The average distance
by general size categories. These categories are wave data was collected from each iceberg was
based primarily on the maximum observed length of 48km; wind and weather data, 61km; and sea
the iceberg. temperature data, 7km. These distances were
computed form the interpolated positions of each
3. 1987 DATA COLLECTION EFFORT iceberg for OOOOZ and 1200i as derived from a
cubic spline.
This study collected data on six medium to small
icebergs for a period ranging from 2.1 to 6.3 The spatial separation of the wind and wave
days. The icebergs were studied as they drifted observations is much smaller than the 250km data
south with the Labrador Current on the Northeast grid-spacing used by the U. S. Navy Fleet
Newfoundland Shelf (centered around position Numerical oceanography Center to provide input
50-45*N, 53-30*W). The study was conducted data for the IIP's drift and deterioration
between 15 and 21 June using the USCGC TAMAROA, a models. Because the study area was at least 1051cm
68m (205 f t) U. S. Coast Guard law enforcement offshore, the wind and wave field were assumed to
cutter. be spatially uniform over the study area.

Iceberg areal dimensions were taken during in mapping the sea surface temperature, the
daylight using a camera and reticulated laser icebergs were in a tongue-like feature of cold
rangefinder. Iceberg shape and size were water which protruded southeastward. The
calculated from photographic images scaled southeastward protrusion of the tongue measured
according to rangef inder measurements. This about l8km across. This complicated observations
required a 360 degree look at each iceberg; since the sea surface temperature field could not
measuring/photographing all prominent faces. be assumed uniform. As a compromise, only
Measurements were accurate to +-8% of the observed observations within 9km of an iceberg's position
dimensions. No underwater iceberg dimensions were were accepted. Because of this restriction and
measured. having only one observation platform, the data
sets for some icebergs were incomplete. The OOOOZ
The rangefinder-derived distances were used with temperature values necessary to model
visual bearing to fix the icebergs' positions deterioration were linearly interpolated from
during daylight. At night, radar bearings/ranges these data sets.
were used to fix their positions. The ship used
LORAN-C and SATNAV to fix its position. 4. ICEBERGS STUDIED
Positional accuracy for iceberg positions was
estimated (by summing system errors) to be Six non-tabular icebergs were studied. Five were
750m. classified medium in size; one was small 0620).
The majority of the icebergs did not deteriorate
Hourly environmental observations included: air enough to change size. The numbers ref er to the
and sea surface temperature, cloud cover, and wind sequential numbering system that IIP uses to track
(at 22.3m). Sea surface temperature was taken by individual icebergs during the course of the ice
bucket thermometer (error was + O.I*C); wind was season. These are the same numbers used in
measured by the ship's anemometer. Wave height, archiving IIP iceberg data at the World Data
period and direction were visually estimated every Center for Glaciology, Boulder, Colorado.
six hours. Visual wave observations were
estimated to have an error of + 0.5m for wave Although the icebergs did not change size during
height; and + 2 seconds for period. Surface the course of the study, all were in a rapid stage
currents were inferred from the drift of two of deterioration. Because of the recurring
sate 11 ite-tracked drifters, which had window-shade presence of growlers and bergy bits in the
(2m X 10M) drogues. Both drifters were deployed vicinity of all icebergs, except #620 and #744,
near the center of the cluster. Both drifters calving was assumed a major factor in the clusters
were deployed at the same time; one drogued near deterioration. The study could not document all
the surface (center of the drogue 8m deep), while calving for any one iceberg since no iceberg was
the other was drogued at 58m, at the core of the observed around-the-clock. However, two events
Labrador Current. Temperature vs. depth profiles were documented: iceberg #784 on 19 June; and
were taken in the vicinity of each berg and 1#747 on 21 June. Because of the warm water
transects. were made at the beginning and end of (greater than 3*C), the brash was expected to
the study to determine iceberg drift in relation fully melt between observations. Bergy bits and
to the Labrador Current. The measurements were growlers which did not fully melt between
made to a depth of about 300m using T-4 eXpendable observations were tracked (in one case up to 18
BathyThermographs (XBT). hours), to keep the calving statistics for the
cluster from being inflated.
Having only one observation platform to monitor

562

Only icebergs #785 and #787 appeared stable below 40m depth were calculated and included in
throughout the study period. Stability in this the buoyant and forced heat convection
context meant that the icebergs length and height contributions. For all other icebergs,, convection
constantly decreased. Most of the icebergs contributions below 40m depth were ignored.
changed shape during the study, many probably from
rolling. iceberg #784 rolled while TAMAROA was The rest of this paper evaluates modelled iceberg
close aboard on 19 June. In this case, the deterioration by examining environmental
rolling caused height to double although length parameters. The environmental assumptions
increased insignificantly (5%). regarding each melt process are reviewed. using
White's research (5), various observed values are
5. THE WATER COLUMN evaluated as melt parameters. Using Anderson's
operational computer model (2), melt estimates are
The cluster of icebergs was in a tongue of the calculated from operational data center inputs.
Labrador Current as evidenced from both sea The implications of error estimates for various
surface temperatures in the area and the XBT model inputs on IIP operations are then discussed.
profiles. The tongue of Labrador water had a cold
(-l*C) core at 60m, below a shallow thermocline at 7. WARM AIR CONVECTION
40m. Surface temperatures ranged between 3.4* and
7.6*C. Temperatures of -I*C or colder, that would Melt from warm air convection is ignored in the
preclude melt (5), existed primarily from 40m to model. For March through mid-May, no melt is
between 90m in the eastern portion of the study estimated for air convection. For July through
area, to 160m in the western portion. XBT casts September the average melt is estimated at
taken in the vicinity (within 28km) and within 6 8cm/day, assuming an average daily air temperature
hours of the OOOOZ interpolated icebergs' of 10*C and average wind of 37km/hr.
positions were used to estimate the average heat
available in the water column to melt the The daily average air temperature warmed during
iceberg. The water temperature, relative to -I*C, the study period from 6*C for 15 June to 8*C for
was averaged in 10m increments over the estimated 21 June. The average wind speed for the study
draft of the iceberg. Iceberg draft was estimated period was 33km/hr. Warm air heat convection
as 3.95 times the average sail height observed was estimated to be approximately 4cm/day.
during the study (6).
Climatological average air temperatures for the
When there was no XBT cast near a particular lip region could be used to make monthly melt
iceberg within six hours of OOOOZ, the temperature estimates. Likewise, daily global-scale air
information was calculated by linearly temperature values could be requested from an
interpolating in time. operational data center; however, the IIP level
of effort for this major revision seems
From analyzing temperature profiles taken about impractical for operational forecasting purposes.
four days apart, this tongue of the Labrador
Current had advected south 74km. The advection of 8. SOLAR RADIATION
the cold core at 60m agrees well with the
deep-drogued drifter. Its drift indicated a The modelled melt due to solar radiation is fixed
predominantly southerly flow (186*T at 21cm/s for at 2cm/day, which represents the minimum melt rate
4 days (from 15 June/OOOOZ through 19 June/OOOOZ), for the period March through August (2). The
then an easterly flow (112*T at 12cm/sec) for the model assumes cloudy conditions.
last 1.5 days of drift. The westward displacement
of the thermal field above the thermocline agreed The daylight (080OZ to 240OZ) cloud cover/obscured
well with the shallow-drogued drifter. From 15 skies averaged 100% every day of the study except
June/OOOOZ to 17 June/160OZ the drift was 193*T at for the afternoon of 17 June and morning of 19
27cm/s. From 17 June/160OZ until recovered on 20 June. For those half day periods the skies were
June/1243Z, the drifter showed a constant partly (averaged 50%) cloudy. Assuming a 35%
deceleration, averaging 206*T at 9cm/s. All of albedo for an iceberg, the average melt rate for
the drifters recorded sea temperature at 1m depth the June study period was 4cm/day (5).
between 3* and 5*C, which agreed well with the
bucket thermometer measurements. The model could be adapted to the monthly melt
estimates derived by White; however the benefit
6. MODEL EVALUATION BASED ON MELT FACTORS would be minimal. Likewise, global-scale
radiation estimations could be requested from
The deterioration processes were evaluated based operational data centers (7); however the level of
on computations for the four medium, non-tabular effort to identify those periods of clear skies
icebergs which had estimated drafts from 98 to would only provide an additional melt of 2cm/day.
146m. These icebergs were #747, #784, #785, and
#787. This cluster was studied for an average of The IIP level of effort to make this major
5 days. revision, and the data center's level of effort to
extract a parameter, which is an interim
Based on in situ temperature, all of the icebergs, computation to predict other atmospheric
except #784 and #785, had insignificant melt from conditions, seems impractical for operational
convective processes below 40m depth. For forecasting purposes.
icebergs #784 and #785, the convective processes

563

9. WATER TEMPERATURE product to potentially reduce melt rates appears
The model uses sea surface temperature to estimate inadequate for operational forecasting purposes.
both buoyant and forced convection contributions 10. RELATIVE VELOCITY
to iceberg melt. The melts due to buoyant and
forced convection were computed: as a function of Forced convection is also a function of relative
sea surface temperature (To); and as a function velocity between the iceberg and the surrounding
of the temperature (fTO) of the water column water column. The model equates relative velocity
integrated over the estimated draft of the iceberg. to the difference between iceberg drift and the
IIP historical current in the iceberg's vicinity.
a. BUOYANT VERTICAL CONVECTION The wind-induced component to the sea current is
ignored.
Buoyant convection is considered solely dependent
upon the relative temperature between a near Melt rates for forced convection using relative
vertical wall of ice and the water column. The velocities derived from different current
cluster's average daily melt due to buoyant parameters were compared. The shallow-drogued
convection using f To was estimated at 2cm/day drifter was assumed to represent the velocity of
with average values for individual icebergs the water mass between the surface and the 40m
ranging from 1cm/day 0787) to 3cm/day (#785). thermocline, that portion of the total water
The melt rate as a function of To averaged column which contributed most to iceberg
7cm/day greater; with daily differences ranging deterioration. The relative velocity between each
from +3cm/day (9787/21 June) to +11em/day (#747/20 iceberg and the following were calculated as
June). These differences are associated with inputs to the -model: shallow- and deep- drogued
surface temperatures which are approximately I.S*C drifters, and the lip's ..master" historical
warmer than the averaged temperature for the first current field velocity, which for the entire study
10m of the water column. area is 160*T at 23cm/s. The deep-drogued drifter
represents real-time current data which when
b. FORCED HEAT CONVECTION available is used to "modify" the historical
current. The "modified" current is forced by a
Forced convection is primarily dependent upon the two dimensional spline to fit the observed
relative temperature between ice and the water current. Sea surface temperature was used to
flowing past the iceberg. The cluster's average compute the relative temperature term.
daily melt due to forced convection using fTo
was approximately 15cm/day with average values for The average daily melt due to forced convection,
individual icebergs ranging from 12cm/day 0787) using iceberg drift relative to the
to 21cm/day (#785). The melt rate as a function shallow-drogued drifter, was estimated at 59cm/day
of To averaged 47cm/day greater; with daily with average values for individual icebergs
differences ranging from +17cm/day (#784/20 June) ranging from 55cm/day (#785) to 62cm/day (#747).
to +69cm/day (both #747 and#784 on 18 June). The melt rate as a function of iceberg drift
These differences are associated with surface relative to the "modified" historical current
temperatures which are about 0.9*C warmer than the ranged from 25cm/day slower (#747/19 June) to
averaged temperature for the first 10m of the 38cm/day faster 0784/20 June). Using the
water column. "master" historical current, the melt rate ranged
from 53cm/day slower (#747/19 June) to 63cm/day
c. COMBINED EFFECT IN USING SEA faster (#784/20 June). These differences in melt
SURFACE TEMPERATURES equated to velocity differences between the
shallow-drogued drifter and the "modified"
By using sea surface values to compute the current, and between the shallow-drogued drifter
relative temperature tent , the waterline loss and the "master" current of +/-9cm/see and
could be over estimated by 20 to 80cm/day. This +/-16em/sec respectively.
error represents summer conditions (surface
warming). Errors for the period March through In comparing the melts due to forced convection
mid-May should be significantly smaller. Although between the "master" and "modified" currents, the
the error associated with summer sea surface real-time input serves to dampen the daily
temperatures appears significant, it is an order differences for each iceberg by about half. The
of magnitude less than the sum of all modelled meteorological conditions during the study also
deterioration processes. was a limiting factor. Rapid changes in the
weather prevented wind direction from remaining
Subsurface temperature values can be requested constant (within a 60* are of the compass) for
from operational data centers; however, the periods longer than 27 hours; wind shifts averaged
quality of the analyses are highly dependent upon every 12 hours. Consequently, the sum of the
the availability of observational data. More differences for each iceberg never exceeded
daily observations occur for sea surface +/-70cm for the study period, or an averaged error
temperature than for temperature over depth in the of +/-21cm/day.
IIP region. Additionally, data center thermal
depictions often have an accuracy no better than Assuming that these are the least errors
the temperature differences enumerated above. anticipated for the IIP region, where 5- to 7-day
Consequently the justification for an operational wind events occur, an)effort to control the growth
data center to provide an additional temperature of these errors may be appropriate. Wind-induced

564

components of the sea current could be extracted TABLE 1: AVG DIFFERENCE BETWEEN DATA CENTER PRODUCTS AM OBSERVATIONS
f rom the dynamic iceberg drif t model and OBSERVED - PNOC OBSERVED - MITOC
substituted for the existing input. This is DATE or SET WAVE HT WAVE PD SET WAVE HT
perceived to be a moderate level of effort for the 000OZ ICEBERGS 00 (m) (see) (*C) (m)
IIP. is am 2 +1. 1 -1.2 -8 (2) +0.5 -0.9
16 JUN 1 +0.7 -0.9 -5 (5) +1.0 -0.6
17 JUN 5 +0.3 -0.1 -5 (5) +0.5 -0.6
11 . WAVE EROSION: is JUN 6 +1.0 -0.9 -4 +0.2 -1.2
GLOBAL VS REGIONAL SCALE PRODUCTS 19 JUN .5 +1.5 MA -0.9 -2 -0.1 -0.3
20 JUN 4 +1.8 (2) -1.2 -4 (1) +0.2 -0.9
Wave erosion, which is induced by heat convection 21 JUN 3 1+2.5 (2) -0.6 -2 +0.7 -1.2
from the turbulent maximum orbital velocity caused *Vote %Numbers In Farentheass indicate number of times a value "a
auto ids the following error bounds:
by the wave field surrounding the iceberg, is SET +/- 1.6-C Wave Height +/- 1.8 m. Wave Period +/- 4 see
computed by the model. This convection is FNOC and METOC products differed primarily in
proportional to wave height (H) times relative their representations of sea surface temperature.
temperature (T), and inversely proportional to The FffOC-produced sea surface temperatures
wave period (P). The model assumes the effects of averaged 1.3*C colder than the surface-measured
the wave field are non-directional, implying that values for the cluster; the METOC-produced
the iceberg is melted uniformly from all temperatures were 0.6*C colder. Averaged
directions (5). This assumption generates a melt differences between observations and FNOC products
factor which may over estimate the wave-generated for individual icebergs ranged from 1.9*C 0784)
convection. Wave erosion is the most important to 0.6*C 0785) colder. The differences listed in
process in the iceberg melt model. It is up to Table I which are greater than the reported system
ten times greater than melt by forced convection, error are due to the presence of sub-scalar
and routinely 100 times greater than buoyant thermal features. in this case, iceberg #747 and
convection. The applicability and accuracy of the #784 had crossed the surface thermal front between
environmental parameters used to model wave the colder Labrador water and the warmer
erosion thus greatly affect the daily melt rate Newfoundland Shelf water.
estimated by the model.
Little difference existed between the observed
The model computes T from sea surface wave height and those produced by YNOC and METOC.
temperature. Significant wave height and a wave The FNOC-produced wave heights averaged 0.9m
period, which is that period associated with peak higher than the observed height for the cluster;
energy observed in the wave spectrum, are the METOC-produced height was O.8m higher. Daily
currently used by the model f or H and P differences between observed and predicted wave
respectively. Sea surface temperature is assumed heights for individual icebergs ranged: from 0.3m
to be the best parameter from which the relative (all/17 June) to 1.5m (#785/20 June) for FNOC
temperature term for wave erosion is calculated, products; and from 0.2m (all/19 June) to 1.2m
and it is readily available from data centers. (all/18 June). The wave period differences in
Data center products to represent H and P are Table I which exceeded system error may be based
significant, sea, and swell height and peak partly on the limitations of visual observations
periods for the total energy spectrum, and for the and the quality assurance employed by each
sea and swell spectrums. When the model was observer.
implemented in 1983, the wave parameters currently
used were the only ones available. The cluster's average daily melt due to wave
erosion using observed values was estimated at 379
Table I shows the differences between observations cm/day with average values for individual icebergs
and those values produced by operational data ranging from 330 cm/day 0785) to 408 cm/day
centers. The global scale (250km grid-spacing) (#747). The cluster's average daily melt using
parameters were produced by U. S. Fleet Numerical FNOC (global scale) products averaged 152 cm/day
oceanography Center (FNOC) using its computerized faster. Individual iceberg's average melt ranged
Expanded Ocean Thermal Structure (EOTS) analysis from 144 cm/day (#784) to 200 cm/day (#785)
and Global Spectral Ocean Wave Model (GSOWK) (7). faster. The cluster's average daily melt using
The regional scale (estimated from 50 to 100 km METOC (regional scale) products averaged 195
grid-spacing) parameters were produced by the cm/day faster. Individual iceberg's average melt
Canadian Forces Meteorological and Oceanographic ranged form 149 cm/day (#785) to 218 cm/day (#787)
Center (METOC) Halifax, Nova Scotia. METOC faster.
depends on human interpretation of surface thermal
observations and uses a parametric ocean-wave The higher melt estimated using both FNOC and
model (8) which is qualitatively blended with ship METOC parameters was a function of their 'higher
observations. All values are for OOOOZ, except wave height predictions. The melt estimate using
for the METOC H parameter, which was analyzed at FNOC products appeared better than METOC because
180OZ. The . METOC 180OZ sea state analysis the colder FNOC temperatures helped offset the
normally contains more ship observations than the error in predicting wave height. The model's
METOC OOOOZ analysis; improving the quality of the sensitivity to wave height makes that input
180OZ analysis. For all data center inputs, the parameter highly important to IIP. Thus it is
values, which were compared to observations, were highly prudent for IIP to investigate new wave
interpolated to each iceberg's OOOOZ position. data inputs for the model, and thoroughly evaluate
these inputs before substituting them for the FNOC
significant wave height and peak period inputs.

565

12. SUMMARY The iceberg sizing methodology and study
time-constraints made comparisons of model
Table 2 summarizes the deterioration processes estimations to observed lengths inconclusive.
examined by this paper. For this ensemble of Either a better methodology must be used in future
medium-sized, non-tabular shaped, icebergs, a studies or studies must be extended over much
daily melt rate was estimated by the model to be longer periods (14-21 days). Given the potential
4.0 m/day. Using FNOC products, the melt rate was for errors associated with operational
over estimated by 1.7 m/day. METOC products over reconnaissance, which depends heavily on
estimated the melt by 2.2 m/day. These melt qualitative descriptions, and the inability to
estimates are of the same magnitude as the iceberg address all deterioration processes, the IIP
sizing error. Because of the short duration of policy to require icebergs to deteriorate 175% of
the study, no firm conclusions could be drawn from their original estimated length is prudent.
the observed iceberg measurements. Icebergs #785
and #787, which seem to have remained stable 13. REFERENCES
throughout the study period, appeared to have
melted faster than the prediction based on the 1. MOUNTAIN, D. G., 1980. On Predicting Iceberg
optimum-observed environmental parameters; Drift. cold Regions Science and Technology,
however, both predictions based on data center Vol 1 (3/4), p ?73-282.
products were within measurement error bounds.
TABLE 2: AVERAGE MELT RATE FOR VARIOUS INPUTS (cm/day) 2. Anderson, 1., 1983. Iceberg Deterioration
Model. Report of the International Ice
OPERATIONAL MODEL Patrol in the North Atlantic Ocean, Season
Relative Temperature T/D SST SST FNOC IMOC
Relative Velocity Waft) V(sfc) V(deep) v(sfe) V(ofe) of 1983, (CG-168-38)9 p 67-73.
Wave Height Primary Primary Primary FROC METOC:
Wave Period Peak Peak Peak Peak Peak 3. Anderson, 1., 1985. Oceanographic Conditions
Warm Air Convection 4 4 4 2 2 On The Grand Banks During The 1985 IIP
Solar Radiation 4 4 4 0 0 Season. Report of the international Ice
Buoyant Convection 2 9 9 6 a
Forced Convection 15 62 6- 122 33 41 Patrol in the North Atlantic Ocean,
Wave Erosion 379 379 379 531 574 Season of 1985, (CG-188-40), p 56-67.

TOTAL AVERAGE MELT 404 458 402 - 520 572 625
4. El-Tahan, M., S. Venkatesh, and H. El-Tahan,
T/D - Temperature contribution as function of depth 1987. Validation and Quantitative Assesment
V(sfc) - Differential velocity between iceberg and surface-drogue4
drift r Of The Deterioratoin Mechanisms of Arctic
V(desp) - differential velocity between iceberg and deep-drogued Icebergs. Journal of offshore mechanics and
drifter Arctic Engineering, February 1987, Vol 109,
Sea surface temperature appears to be a suitable p 102-108.
parameter from which the relative temperature term
is calculated in the model. In this study the use S. White, F. M., M. L. Spaulding, and L. Gominho,
of sea surface temperature to solely represent the 1980. Theoretical Estimates Of The Various
relative temperature term vice using temperature Mechanisms Involved In Iceberg Deterioration
versus depth to represent the term for buoyant and In The Open Ocean Environment. Report
forced convection, caused the melt rate to be over CG-D-62-80, U. S. Coast Guard Research and
estimated by 12%. Global-scale thermal products Development Center, Groton, CT 06340-6096,
cannot adequately represent the Labrador Current 126 pp.
as it flows below 48*N latitude. Regional scale
temperature products currently available can 6. Robe, R. Q., 1975. Height To Drift Ratios Of
improve the resolution of the temperature data. Icebergs. Proceedings of the Third
international Conference on Port and Ocean
over simplification of methods used by IIP to Engineering Under Arctic Conditions, 11-15
derive the relative velocity between icebergs and August 1975, Vol 1, p 407-415.
the surrounding water contributed to an error of
up to 16% of total melt. Because the wind-induced 7. Numerical Environmental Products Manual, Vol
component of the ocean surface layer (between the II, August 1986. Prepared under authority
surface and 50m depth) is computed in the IIP of Commander, Naval Oceanography Command,
iceberg drift model, this velocity component could NSTL,MS 39525, 200 pp.
be added to the known or historical surface
current. The resultant current value could then 8. Macdonald, K. A., and S. Clodman, 1987. The
be used to compute relative velocity to reduce the AES Parametric Ocean-Wave Forecast System.
magnitude of this error. Proceedings of the International Workshop on
Wave Hindeasting Forecasting. Report 065.
Wave height over estimation causes daily melt to Envirorunental Studies Revolving Funds,
be over estimated. This significant (about 38% of Ottawa, pp 119-132.
total melt) error in determining the wave erosion
factor may help offset the model's inability to
represent other deterioration processes that the
model ignores. New wave products that have
recently become available should improve the
representation of wave erosion.

566

AN AUTONOMOUS ATMOSPHERIC PRESSURE RECORDER
FOR
ESTABLISHING POLAR SEA SURFACE HEIGHT

GEORGE STEEVES STEPHEN GRANT

Head, Systems Engineering Regional Tidal Officer
Engineering & Technical Services Canadian Hydrographic Service

Bedford Institute of Oceanography
Dept. of Fisheries & Oceans
Scotia-Fundy Region
P.O. Box 1006, Dartmouth, N.S.
Canada 82Y 4A2

ABSTRACT

An accurate long-term recorded
measurement of atmospheric pressure in
Arctic regions is essential for
correcting absolute-pre5sure measuring
submerged tide gauges. A device capable
of unattended operation will make
possible an accurate measurement of polar
sea surface height. This data is
required for hydrographic surveys,
physical oceanography, and satellite
image analysis. This paper will
describe an autonomous low-power device
for the precise measurement (+/-0.7 mbar)
of atmospheric pressure versus date/time
in Arctic conditions (-55 deg C, 120
km/hr wind). Data is stored using
"burned in silicon" technology.
Measurements are made every hour on the
hour over a period of 12 months. Static
pressure is measured using a special part
designed to be insensitive to the affects
of wind. Endurance problems associated
with a "one-year" instrument operating in
extremely harsh conditions are discussed.

1. INTRODUCTION FIGURE 1. SYSTEM COMPONENTS

The tides in the vast areas of the Canadian Arctic achieved in the past , using bottom mounted
continue to be one of the principle concerns Of submersible tide gauges that measure the pressure
hydrographer5 and oceanographers. In order to meet changes as the height of water over the gauge
the needs of marine shipping, construction changes with the tide. Except for ice berg
industries. environmental impact studies , and scouring, the gauges are relatively safe beneath
resource development the Canadian Hydrographic the surface of the ice. The pressure sensed on the
Service has recognized the need for a planned ocean bottom is the sum of that exerted by the
approach to the tide survey of the Canadian Arctic water mass and that exerted by the atmosphere. The
Archipelago. However, sea surface height as a pressure of the atmosphere varies by the equivalent
function of time is a difficult parameter to of approximately 12S cm of water, and if
measure, in the Arctic, because of the destructive unaccounted for would result in a significant
forces of ice and the almost total lack of error. The atmospheric component must be
facilities (power, communications, transportation, subtracted from the total pressure to arrive at the
people). A limited amount of success has been sea water pressure and thus the sea surface height.

CH25 85-8/68/0000. 567 $1 @1988 IEEE

In near shore areas not infested with sea ice, the The third circuit board, designated DS-64, is a C-
measurement of sea surface height is a relatively 44 bus compatibledata store composed of 64k bytes
5 i pip I e procedure. Adifferential pressure of CMOs EPROM. The AIR digital barometer is
transducer is installed on the sea floor with one interfaced directly to the CPU board using the
Port exposed to the sea water and the other port parallel port and connected to the atmosphere
vented to the atmosphere. There have been many through a wind dampened pressure port. Energy is
attempts to construct this kind of installation in supplied by a lithium battery pack.
the Arctic. The failure rate is high because in The instrument is contained in a section of 8 inch
most cases ice movement has destroyed the vent to diameter aluminum schedule 40 pipe, with 1/2 inch
the atmosphere; even armored vent tubes in5talled thick aluminum end caps and internal bulkhead. The
in protected trenches have not survived. The battery and instrument compartments . are
obvious alternative is to measure the total and individually sealed.
atmospheric pressures independently versus time and
subtract them later to determine the pressure
variations due solely to the vertical water
movements. The actual wa t er level changes are 3. SENSOR & CLOCK
calculated from the pressures using the specific
gravity of sea water. A search was initiated for a pressure sensor with
To address this problem, the Canadian Hydrographic lowI power, consumption and a 12 month accuracy
Service commissioned the design of a device which equivalent to +/- I cm. of water while operating
would record atmospheric pressure versus time. under Arctic conditions. An almost ideal device,
The accuracy of the pressure measurement and t he t called "INTELLISOR", is manufactured by
of time to be at least as good as that of the Atmospheric Instrumention Research Inc. This
submerged tide gauge. The pressure sensors in the pressure transducer combines a stable capacitance
tide gauges now in use have an accuracy of +/- 0.3 type sensor with microprocessor control. it
mbar, but the uncertainty in the value of the Measures pressure 10 times per second and makes it
specific gravity of the sea water above it, make : available in digital format to a computing device.
it Relevant specifications include:
likely that the calculated sea water height has an
accuracy of +/- I cm. The system required an Pressure accuracy (12 mos.) +/-0.7 mbar
endurance of one year and survivability under Pressure resolution 0.01 mbar
Arctic weather conditions. The design and Pressure range 800 to 1200 mbar
development was initiated in 1987, with a test Temperature range -5S to +85 deg C
deployment later that year. Power consumption 99 mW (oper)
150 uW (stdby)

Accurate time keeping for a 12 month period is a
problem, when the temperature range is expected to
be 60 deg 0. Power restrictions rule out oven
h controlled oscillators. A temperature compensated
crystal oscillator (TXCO) provides an adequately
accurate time base. Accounting for temperature and
A ageing stability, the oscillator chosen would
worst case error in time of +95 to -65
produce a
seconds per year.

P "le
4. OPERATIONAL SEQUENCE

m The sequence of states the system moves through
af ter start up reveals the design strategy and
Of
describes It's operational details.
xTn-
FIGURE 2. VIEW of ELECTRONICS
1. Empty EPROMs are installed on the data store
2. SYSTEM OVERVIEW board. The (year,month,day.hour,minute) initial
date/time is set into the micro thumb wheel array.
The System is based on the C-44 microcomputer bus The system is powered up manually on the minute
architecture for battery powered Systems. The selected.
system is managed by an Onset CMOs single board 2. The power sequencing circuit delays the
computer (CPU-8085A). Asecond custom designed application of power to the digital barometer until
circuit board, contains the real-time the CPU conditions the control lines. This
calender/clock circuit, an array of 10 micro thumb prevents the barometer's microcomputer fr .om locking
wheel switches (to initialize the date/time), and a
circuit to synchronize the application of power. up.

568

PRESSURE
PORT

AIR POWE
DATE/TIME SET DIGITAL S QUENCER
BAROMETER
(MICRO-774UMBMIEEL ARRAY) Ho I M, MO] I k A SHU TEMPERATURE
1@ 1y01M11M0jDjjD6jH1 00,19 DOWN COMPENSATED
CRYSTALL
2 OSCILLATOR

CMOS EPROM CMOS SINGLE BOARD WAKE REAL TIME
DATA STORE COMPUTER UP CALIENDER
DS-64 CPU-8085A CLOCK

F C-44 BUS

FIGURE 3. SYSTEM STRUCTURE

3. The CPU then reads the micro thumb wheel array, daily record. The high voltage supply (used in
the programming made) on the data store board is
sets the initial date/time into the real time turned an and the data written into the non-
clock chip and sets the alarm for the next integral volatile EPROMs. This event happens each midnight
hour. up to and including day 368, when the 64k byte data
4. The power down command is then issued to the store is considered full.
barometer and the CPU places the system in low- The microcomputer code executed by the CPU has two
power HYBERNATE mode. This mode not only holds up special modes to facilitate system testing and
instruction execution, but also reduces the system verification. FAST mode speeds up the real time
supply voltage. However, a TXCD (temperature clock, such that I hour of real time is replaced by
compensated crystal oscillator) time base continues 10 sec. of computer time. FAST/N0 HYBERNATE mode
to drive the real time clock chip. speeds up the clock and keeps the system awake.

5. On the hour, the clock alarm sends a WAKE-UP
signal to the CPU. This signal initiates the S. ENERGY SOURCE
sequence which brings the supply voltage up to it's
operating value. The CPU then begins execution of The energy source for an autonomous instrument
the interrupt service routine. operating in a remote area under harsh conditions
is a system component of crucial importance. For
6. The barometer is turned on and commanded to make unattended year long Arctic , deployments,
10 Measurements and average them. The CPU then reliability and low temperature operation are of
initiates a sequence of hand-shaking signals to particular concern. Operating temperatures as low
transfer the measurement data bytes to CMOS memory. as -SS deg C preclude the use of almost all common
battery Lechnologies; this factor alone would point
7. The system places itself in HYBERNATE mode once to one of the lithium types. Design considerations
more and awaits the next hourly alarm. The day's that must be accounted for include:
data is accumulated and stored in CMOS RAM.
1. Load Voltage versus Capacity (ampere hours)
8. When the alarm coincides with midnight, the 24 2. Capacity (ampere hours) versus Discharge Current
hourly measurements of atmospheric pressure are 3. Capacity (ampere hours) versus Temperature
POWEI
@SEQ@UEN
T MP;=PjAj

concatenated with the date to form a 1?8 byte 4. Mid-point Voltage versus Temperature

569

Al .. 11 . Al 2 .1
J1 8 1211411011812M01POM2124 14 15 J 7 19 J 2 18 J 11113115 17 19 2f 23 2 .0

I T
I I I F IIT
. I .. I III T
1 2 3 4 5 6 7 8 12 1 @1 17 1! @1 @1@2 31 33 3 V
DN)DE 2- 2-
7 7
1. 12 1 2 21 3 3 1 V 74- w- Ixo-
& 7 C =1
ED aG) 1-1 .1 VW
ml LID 2 2 5 2 1 1.

IN4= DO=
C,
-->DH> .1
.1 N2

.2 21.1
'4 3 3 4 4,
2. k'r2
.2

MC14MIS
1 2 3
74 HC3D
.IPj A--
7 11 1 11 1 7
Nora. m 'MINNIGM
2
3

R IS I T I U I V JW I X 118 120 JH

FIGURE 4. SCHEMATIC

The system components which are significant cells in the double D size. To achieve the voltage
consumers of electrical current are: and capacity required, 4 parallel groups of 5 cells
in series were employed, each group isolated by
HYBERNATE EXECUTING blocking diodes to prevent charging. This
configuration produces an open circuit voltage of
CPU Board I mA 8 I'l A 19.Sv, a mid-point voltage of 15v, and a minimum
Data Store 10 uA 55 mA (programmingi capacity of 60 ampere-hours at -40 deg C. This
Barometer 10 UA 6 mA specification is more than adequate for d location
Oscillator 5 MA 5 mA such as Eureka NWT, in the high Arctic, where the
mean annual temperature is -18 deg C and the mean
The system spends less than 2 sec. per hour making daily minima for the coldest month is -46 deg C.
measurements and less than 0.5 sec. per day burning
the data in EPROM. Therefore the power consumed
in execution made is insignificant compared to that
consumed in the hybernate made. The major energy 6. DATA STORAGE & RETRIEVAL
usage results from maintaining an accurate time
base in the presence of fluctuating temperature, The digital barometer, produces it'5 ASCII coded
and riot writing data into EPROMs, as might be readings in the form xxxx.xx in units of
expected. Calculating for a one year deployment, a millibars. The daily record of 24 hourly
battery pack capacity of S2.5 Ah is required; the measurements of atmospheric pressure are combined
momentary peak current is 74 mA. with the date and delimiting characters to produce
a 178 byte ASCII string. The "64k" (6S536) byte
The switching regulators on the microcomputer capacity data store can thus record for 368 days.
boards and more importantly the TCX oscillator can Since the information is stored "on silicon" in
be operated over a wide range of input voltages, ASCII code, the EPROMs themselves form a data
but they are significantly more energy efficient at archive.
,the high end of the voltage range (maximum 18v).
A small read-out device has been constructed into
Lithium batteries using oxyhalide chemistry which the complete data store board may be plugged.
provide the highest performance in low to moderate When initiated the device will output all the
discharge rate applications. Abattery pack was records to either a serial (RS232C) device or a
M
1331

2 ly,

designed using an array of Electrachem BCX primary parallel (Centronicsi' device.

570

In August 1987, one of the two completed units was
deployed on the coast of Labrador for scheduled
recovery in August 1988. The second unit was
tested on the roof of the Bedford Institute during
FUSE January 1988. The resulting measurements were
compared to those obtained using a mercury
+ + + + barometer at the nearby Shearwater Naval Air
Station (readings corrected Using the Smithsonian
Meteorological Tables, 6th edition). The agreement
between the readings was well within the +/-0.7
mbar accuracy of the digital barometer. This test
+ + + + verified the immunity of the pressure port assembly
to the affects of wind, ice, rain, and snow.
ELECTROCHEM
5 x 4 ARRAY
+ + + + BCX LITHIUM
OXYHALIDE 8. FUTURE DEVELOPMENTS
PRIMARY CELLS The independent collection of atmospheric and total
+ + + + sea bottom pressure is a logical first step for
establishing accurate long term records of sea
surface height in remote regions, However, there
are instances where the information is required
+ + + + more frequently than once per year. In these
cases, it would be useful to monitor, both the
barometer and the tide gauge remotely. The
availability of reliable and inexpensive satellite
up-links make this a straight forward engineering
task were it not for the problem of communication
0 with the submerged tide gauge. Just as the
differential pressure transducer could not be
FIGURE S. BATTERY PACK vented t a the atmosphere, due to the destructive
ice surface, the tide gauge cannot be electrically
connected to the shore. Underwater telemetry in
near shore situations is plagued with problems of
it's own, not the least of which is the ice to
RS-232C water interface. It is known that hard rock miners
COMPATIBLE COMPATIB E CMOs EPROM
DEVICE DATA STORE use HF radio to communicate with the surface from
DEVICE thousands of meters underground using the rock as
the transmission medium. If such a radio could be
coupled to the rock under the tide gauge, it may be
SERIAL CMOs possible t 0 transmit the data t o a receiver
COMMUNICATIONS SINGLE BOARD ashore, thus avoiding the water-ice-air boundary
BOARD COMPUTER altogether. This topic is the subject,of some
WART CPU-8085A interest for the Arctic sea surface height
application and will be pursued in the coming
T months.
E C-44 BUS

FIGURE G. DATA STORE READER

9. ACKNOWLEDGEMENTS

The authors wish to thank Steven Witiker of
Atmospheric Instrumentation Research Inc. and
Dr.R.Siegler of Electrochem Industrie,@ for their
7. TEST PROGRAMME technical assistance.
Members of the technical staff of the Bedford
An extensive series of tests were conducted with Institute of Oceanography who contributed to the
the system placed in FAST mode and when cooled to design include R.Vine, G.Awalt and G.Cooke.
-55 deg C. These tests verified the operational
sequence of the system by simulating an entire year
of operation n 25 hours of real time.
rFUSE

Measurements were made of electrical current
consumption and time keeping accuracy which
confirmed the correctness of the design
calculations.

571

10. REFERENCES

Tait,B.J. et al (1986). "Canadian Arctic Tide
Measurement Techniques and Results", International
Hydrographic Review, Monaco, LXIII(2)

Stephenson,F.(1977): "An Assessment of the
Permanent Water Level Stations in the Canadian
Arctic", Lighthouse, Ed.16

O'Leary,T. (1986): "Report on the Stability and
Accuracy of Two AIR-DB Series Digital Barometers",
Schwien Engineering Inc., Pomona CA

Krehl,P.W. et al (1987). "Safety-Enhanced. High-
Rate, Non-Magnetic Lithium D-Size Cells for
Oceanographic and Naval Use", Oceans 87
Proceedings, Volume One pp. 271-279

572

ESTIMATED ICE-GOUGE RATES ON A MANMADE SHOAL
IN THE BEAUFORT SEA

Thomas K. Newbury and Allen J. Adams

Minerals Management Service, Alaska OCS Office
Leasing and Environment Office
Anchorage, Alaska 99508-4302

ABSTRACT gouging, has not yet been analyzed and is the
subject of this paper.
In 1987, slope-protection fabric was removed from
the top of an artificial gravel island in Harrison Past experience with abandonment of Alaskan
Bay, southwestern Beaufort Sea. As a result, Beaufort Sea artificial islands is mainly with
currents began to erode the island's top. After islands such as BF-37 and Duck II, which were
it is eroded to a subsurface shoal, ice keels will constructed on the shoreward side of natural
slowly gouge and disrupt the remaining shoal. This islands where ice-gouge rates arelow. These
paper analyzes the ice-gouge process and estimates naturally protected islands have been gouged very
the rate of sediment disruption by gouges. The infrequently by ice keels during the years since
rate is derived from existing information on abandonment in the early 1980's.
gouges in the surrounding seafloor in Harrison Bay
and for a time when the shoal is at half of the Abandoned artificial islands that have been
water depth (-7.5 m). It is estimated that ice studied in the Canadian Beaufort Sea are located
keels will typically disrupt I vertical meter of in an extensive landfast ice zone in which there
the whole surface of the shoal within a period of are infrequent ice gouges. For example, a study
about 2 decades. of seafloor gouges in 15 m of water near artifi-
cial islands off Tuktoyaktuk concluded that there
1. INTRODUCTION is an average of one new gouge per year per
kilometer (gouges/year/km) of survey trackline
Mukluk Island is an artificial island constructed (4). The maximum gouge rate in the Canadian
in 1983 in 15 m of water in outer Harrison Bay, Beaufort Sea is about four new gouges/year/km (5).
southwestern Beaufort Sea. The island is made of In contrast, an average of eight to nine new
sandy gravel fill material from an onshore borrow gouges/year/km was observed in Harrison Bay where
site and contains negligible fine-grain material. Mukluk Island is located (6).
The material has been protected from wave erosion
by a slope-protection system. The system includes The purpose of this paper is to use existing
gravel- filled bags that are made of heavy woven information on ice-gouge rates to analyze the
polypropylene cloth and an underlying polypropy- process and estimate the rates of sediment disrup-
lene mat (1). In 1987, the owner decided to tion on Mukluk Island's base. The actual rates
abandon the island and, therefore, removed the eventually can be determined through bathymetric
slope-protection fabric from the top of the island monitoring.
to a depth of -4.6 m (1).
2. ICE GOUGES IN HARRISON BAY
This unprotected island in outer Harrison Bay will
be subject to erosion by waves to form a submerged Harrison Bay is gouged at "high" to "very high"
shoal in a relatively brief time period. The time rates between water depths of 10 and 30 m (7).
will depend mainly on the frequency and severity Side-scan-sonar records and bathymetric profiles
of open-water storms. Storms in both 1986 and from this area show remnant traces 'of 50 to 100
1987 that were comparable to 100-year events gouges per square kilometer (7, 8). The benthic
eroded about 20 percent of the island's top. With survey for Mukluk Island (9) also noted the many
the removal of the slope-protection materials in ice gouges in the area. Another study (10)
1987, the complete erosion of the unprotected top describes the seafloor seaward of the 12- to
to -4.6 m might occur after several more open-wa- 15-meter isobath in Harrison Bay as "saturated"
ter seasons (2, 3). The subsequent fate of the with ice gouges.
island's base, which is in an area of severe ice

CH2885-8/88/0000- 573 $1 @1988 IEEE

The maximum incision depths of new gouges in Three sets of studies provide relevant data that
Harrison Bay are usually about 10 percent of the show that the variation in ice-gouge-disruptlon
water depth (6). The most extreme gouges may be rates due to sediment composition is minor.
from ice keels driven by forces amassed from an
encompassing solid icepack (7). Some of the One study observed the effects of ice keels on
widest gouges (gouge "multiplets") are caused by Mukluk Island fill material (1). Just after
ice-pressure-ridge keels. There are traces across Mukluk Island was constructed in 1983, summer
the central Beaufort Sea shelf, including the drift ice caused extensive damage to the slope-pr-
waters 15 m deep in Harrison Bay, of about two otection system. The slope-protection polypropy-
wide multiplets per kilometer of survey trackline lene fabric was ripped extensively and the gravel
(7). One of the gouge multiplets they observed in fill in the bags was dispersed by wave-induced
outer Harrison Bay had a width of 78 m, maximum currents (1). This study shows that Mukluk Island
gouge depth of 4.0 m (excluding the adjacent fill material is gouged and disrupted by the
ridges), and a cross-sectional area of incision forces of isolated pieces of summer ice.
estimated at 234 m2.
A second set of studies concerns ice-push events
The gouges near Mukluk Island are made primarily near Point Barrow (13). In this area grounded ice
during the winter by ice-pressure-ridge keels ridges and floes are frequently incorporated into
(11). The pressure ridges form when the offshore the ice canopy during freezeup. During winter
pack ice drives the landfast ice toward the ice-push events, the grounded ice moves easily
southwest. Stringer observed that massive ridges shoreward across the beach fronts, gouging into
are created in December shoreward of Mukluk many sediment types. The sediment types include
Island, around the 10-meter isobath. The ridges consolidated sandy gravel (13) and consolidated
gradually build seaward to the region around clays (14). This set of studies shows that
Mukluk Island. During the late winter, very deep consolidated, coarse-grained beach fronts and
ice keels gouge into and ground on the offshore consolidated clays are gouged similarly by winter
shoals in water 20 m deep and more than 20 km ice forces.
offshore of Mukluk Island. The late-winter ridges
that ground offshore protect the shoreward ice A third set of studies concerns gouges in the
around Mukluk Island from pack ice forces, so widespread occurrence of "concrete like" over-
there is little additional ridging and gouging in consolidated silty clay in Harrison Bay (15).
waters 15 m deep (11). The seasonal change in This study notes that some of the clays are so
formation of the ridges indicates that most of the hard that vibracorers failed to penetrate them.
gouging around Mukluk Island probably occurs The overconsolidated clays in Harrison Bay are
during the middle winter. located in the area in which many ice gouges have
been observed during a five-year study (6). This
Ice gouges in Harrison Bay are mainly oriented in set of studies indicates that overconsolidated
two directions, northwest-southeast and north- sediments, perhaps similar to some reworked
east-southwest. This dual orientation was inter- materials in Mukluk Island, are gouged by ice.
preted (10) as evidence that there may be a second
forcing mechanism on gouging ice in Harrison Bay, The last set of studies indicate a further point
aside from the force of the offshore pack ice as about gouges. If the overconsolidated silty clays
described above. The second forcing mechanism was in Harrison Bay influenced gouge-disruption rates
identified as wind and currents pushing isolated over a geologically long period, a flattening in
pieces of ice toward the southeast during summer the depth profile or relief at the outcrops would
(10). be expected. However, this is not observed;
bathymetric charts show that the bay's topography
The intensity of ice gouging that occurs in the is flat in spite of the different sediment types.
summer is probably low compared to that of winter The flat bathymetry indicates that "concrete like"
gouges. During summer a few deep depressions may overconsolidated sediments are disrupted at the
be formed by grounded ice through a process called same rate as other sediments.
dynamic ice wallow (12). This process is caused
by the open-water wave-induced motion of grounded Harrison Bay shows no change in slope along the
ice. This motion creates depressions by pounding survey testlines (6) where the overconsolidated
the seabottom. Ice-wallow depressions formed near silty clays crop out adjacent to or on the
Reindeer Island in up to 10 m of water were 0.5 m testlines. Rather, bottom depth increases at a
(nearshore) to 3 m deep (12). Ice wallow easily uniform rate with distance from shore. The
reduces the cohesion of sediments and enhances bathymetric features in and studies of Harrison
currents around the grounded ice such that even Bay (6, 15) indicate the following: ice-gouge
coarse-grained sediments are transported. variability that is related to sediment charac-
teristics is minor.
3. EFFECT OF ICE GOUGES ON MUKLUK ISLAND
To summarize, ice keels have been observed to
The above studies describe ice gouge character- gouge many types of sediments similarly. The
istics in natural shelf sediments, as opposed to types include consolidated sediments and un-
the sandy gravel material in Mukluk Island. The consolidated coarse-grained sediments similar to
applicability of gouge data for natural sediments Mukluk Island fill material.
to the Mukluk Island shoal is discussed below.

574

4. SEDIMENT DISRUPTION RATE IN HARRISON BAY crests due to positive relief of shoals." Other
investigators (17) state that they also were
Harrison Bay is the site of a study by Rearic (6) "impressed by the intensity of (ice) gouging on
of Beaufort Sea ice-gouge-recurrence rates in shoal crests in the Beaufort Sea compared to that
addition to the previously mentioned studies of on the low-lying surrounding terrain."
ice-gouge characteristics. Rearic (6) studied the
rate of seafloor disruption by ice gouges in The magnitude of the difference in the disruption
Harrison Bay along testlines that he resurveyed rate on the top of shoals can be estimated. The
yearly for 5 years. One of the testlines (number observed gouge rates on the crests of the three
1) extends from shore to the 14-meter isobath and shoals in Harrison Bay shoals (6) will be compared
passes within 7 km of the Mukluk Island site. to the rates on the surrounding seafloor. Rearic
Rearic's data (6) for Testline 1 show that the observed generally lower disruption rates both on
annual volume of sediment disrupted increases with and off the shoals in eastern Harrison Bay than
water depth in almost a linear relationship the rates in the central part of Harrison Bay.
(redrawn in Fig. 1). Rearic's calculations (6) The low level of the rates possibly is because of
result in an average minimum disruption rate. His the presence of offshore shoals that protect from
relationship indicates that for the water, around intensive gouges the eastern part of the bay.
Mukluk Island (which is 15 m deep), a minimum of Also, the three shoals which are only several
7,000 m3 of sediment is disrupted annually within kilometers apart may influence the intensity of
an area of 1 km2. gouging on one another. Regardless of general
level of the rates, the data for the three shoals
5. SEDIMENT DISRUPTION RATES ON SHOALS indicate the comparative rates on and off of
shoals.
The sediment -dis rup t ion rate on the top and sides
of the Mukluk Island shoal may be substantially The rates on and off the shoals will be related to
different than the sediment-disruption rate on the Rearic's (6) data on. the height of the shoals
surrounding seafloor. The unusually high dis- above the seafloor. The shoreward and leeward
ruption rate on the top of shoals is illustrated sides of a shoal may be protected from gouges by
in Figure 1 which shows the rates on the top of the shoal. Therefore, comparisons will be made
three shoals in eastern Harrison Bay. Rearic (6) with the seafloor rate on only the seaward side of
describes the "high (ice) impact rate on shoal shoals, which is the primary direction of ice
encroachment at Testline 2 (6).

- One shoal extends 3.5 m above the surrounding
seafloor (16, Figure 7). The disruption rate on
this shoal crest is 6 times greater than the rate
6- on the seaward side of the shoal (16, Table III) .
Two of the shoals extend 2 m above the surrounding
seafloor. The disruption rates on the shoal
TESTLINE 1. crests are 1.5 and 4 times greater than the rates
on the seaward sides of the shoals.

The above data indicate that the disruption rate
A 4
0 C4 Shoal on the top of the Mukluk Island shoal, when it
-E Location extends just 2 m above the seafloor, will be at
_X along least 1.5 times greater than the seafloor rate.
E TESTLINE 2. When the Mukluk Island shoal is still 3.5 m above
E the seafloor, the disruption rate may be 6 times
V) 02- greater than the rate on the seaf loor. When the
0 shoal crest is 7.5 m above the bottom (and is at
-7.5 m water depth), the disruption rate will
probably be greater than 6 times the seafloor
rate. The factor of 6 is probably a minimum
estimate for the Mukluk Island shoal at -7.5 m
because of the shoal's relatively small size (less
0 5 10 15 than 250 m diameter).
Water Depth (m) For the following estimations of the disruption
rate on Mukluk Island shoal, the rate is assumed
Figure 1. Mean volume of sediment disruption by to be at least 6 times the seafloor rate at Mukluk
ice gouges along two testlines in Harrison Bay, Island when the shoal is at -7.5 m water depth.
redrawn from Rearic (16, Fig. 14). Testline I Rearic (6) calculated that the seafloor disruption
extends from shore to the 14-meter isobath and rate near Mukluk Island equals 7,000 m3/km2/year,
passes within 7 km of the Mukluk Island site in so we estimate the rate on a small -7.5 m shoal
central Harrison Bay; Testline 2 passes over three crest to equal at least 42,000 m3/km2/year.
shoals in eastern Harrison Bay. Rearic explains
that the gouges are measured for a 250-m-wide
corridor along 1-m-water depth intervals and that
the disruption estimate is a minimum.

575

6. SEDIMENT-DISRUPTION RATE ON 8. Barnes, P.W. 1981. Ice Gouging. In:
MUKLUK ISLAND SHOAL Beaufort Sea Synthesis-Sale 71, D.W. Norton ;-nd
W.M. Sackinger, eds. Juneau, AK: U.S. Dept. of
The effect of the above disruption rate on Mukluk Commerce (USDOC), National Oceanic and Atmospheric
Island shoal will be estimated for just one depth. Administration (NOAA), Outer Continental Shelf
The depth of -7.5 m will be used because it is Environmental Assessment Program (OCSEAP), pp.
half of.the water depth. Further, the -7.5 m depth 100-108.
is several meters below the -4.6 m. water depth to 9. Dames and Moore. 1983. Final Report, Sohio
which currents are expected to erode the island Mukluk Well No. 1, Harrison Bay Alaska, Biolog-
relatively quickly (2). At a depth of -7.5 m, the ical Survey, February 16- 22, 1983. Unpublished
Mukluk Island shoal has a radius of about 94 m report for Sohio Alaska Petroleum Company,
(1). The area on top of the shoal at -7.5 m. will Anchorage, AK, 14 pp.
equal about 28,000 m2. At the ice-gouge-disrup-
tion rate that is estimated above (42,000 m3/km2/- 10. Craig, J.D. and G.P. Thrasher. 1982.
year), the sediment disrupted annually might equal Environmental Geology of Harrison Bay, Northern
about 1,200 m3 for Mukluk Island shoal when it is Alaska. Open-File Report 82-35. USDOI, USGS, 11
-7.5 m. deep. At this disruption rate, it should PP.
take an average of 23 years to disrupt I vertical
meter of material on the entire surface of the 11. Stringer, W.J. 1981. Seasonal Ice Morphol-
shoal (about 28,000 m2). ogy Maps. In: Beaufort Sea Synthesis-Sale 71, D.W.
Norton anT W.M. Sackinger, eds. Juneau, AK:
In conclusion, ice gouge rates indicate that it USDOC, NOAA, OCSEAP, Appendix C, 13 p.
should take an average of about 2 decades to
disrupt 1 vertical meter of material on the entire 12. Reimnitz, E. and E. Kempema. 1982. Dynamic
surface of the shoal. Ice-Wallow Relief Of Northern Alaska's Nearshore.
Journal Of Sedimentary Petrology 52(2):451-461.
7. REFERENCES
13. Shapiro, L.H., R.C. Metzner, A. Hanson, and
1. Anderson, L.M. and C.B. Leidersdorf. 1988. J.B. Johnson. 1984. Fast Ice Deformation During
Arctic Island Abandonment: Planning and Implemen- Ice-Push and Shore Ice Ride-Up. In: The Alaskan
tation for Mukluk Island. Proceedings of the 20th Beaufort Sea: Ecosystems and Environments, P.W.
Annual Offshore Technology Conference, Houston, Barnes, D.M. Schell, and E. Reimnitz, eds. New
TX, pp. 49-60. York: Academic Press, Inc., pp. 137-158.

2. Standard Alaska Production Company. 1987. 14. Phillips, L. and T. Reiss. 1983. Nearshore
Mukluk Island Reclamation and Abandonment Plan. Marine Geologic Investigations, Pt. Barrow To
Anchorage, AK. Skull Cliff, Northeast Chukchi Sea. In: USDOC,
NOAA, OCSEAP, Final Reports Of Principal Inves-
3. Reimnitz, E. 1987. Personal communication tigators 34:157-181.
to T. Newbury, Oceanographer, U.S. Dept. of
Interior (USDOI), Minerals Management Service, 15. Reimnitz, E., E. Kempema, R. Ross, and P.
Anchorage, AK, in June 1987 from Erk Reimnitz, Minkler. 1980. Overconsolidated Surficial
Marine Geologist, USDOI, U.S. Geological Survey Deposits On The Beaufort Sea Shelf. Open-File
(USGS), Menlo Park, CA. Report 80-2010. USDOI, USGS, 37 pp.

4. Gulf Oil Canada Limited. 1979. Analysis of 16. Rearic, D.M. 1985. Temporal and Spatial
Side-Scan Sonar Sea Bed @magery from Repeated Character of Newly Formed Ice Gouges in Eastern
Surveys off Pullen Island -Beaufort Sea. Un- Harrison Bay, Alaska, 1978 -1982. In: USDOC,
published report prepared in May 1979 for Gulf Oil NOAA, OCSEAP, Final Reports Of Prin@_ipal Inves-
Canada Limited, Ottawa, Ontario, Canada, 28 pp. tigators 52:605-700.

5. Shearer, J., B. Laroche, and G. Fortin. 17. Reimnitz, E. and E. Kempema. 1985. Pack Ice
1986. Canadian Beaufort Sea 1984 Repetitive Interaction with Stamukbi Shoal, Beaufort Sea,
Mapping of Ice Scour. Environmental Studies Alaska. In: USDOC, NOAA, OCSEAP, Final Reports Of
Revolving Funds Report No. 032. Ottawa, Canada, Principal Investigators, 34:251-282.
43 p.
18. We thank Peter Johnson and James Craig for
6. Rearic, D.M. 1986. Temporal and Spatial their expert technical reviews, Elinore Anker for
Character of Newly Formed Ice Gouges in Eastern editorial help, and our supervisor, Tom Boyd, for
Harrison Bay, Alaska, 1977-1982. Open-File Report general encouragement.
86-391. USDOI, USGS, 18 pp.

7. Barnes, P.W., D.M. Rearic, and E. Reimnitz.
1984. Ice Gouging Characteristics and Processes.
In: The Alaskan Beaufort Sea: Ecosystems and
Fn-vironments, P.W. Barnes, D.M. Schell, and E.
Reimnitz, eds. New York: Academic Press, Inc.,
pp. 185-212.

576

DESTRUCTION OF OFFSHORE PLATFORMS BY ACCELERATED GALVANIC CORROSION

Risque L. Benedict

Corrosion Consulting Service Corporation
1530 Tioga Trail
Fallbrook, California 92028
ABSTRACT were 630 platforms which will need to be
removed in the next 12 years from the Gulf
There are two principal drawbacks to existing of Mexico. The cost for this platform removal
platform removal techniques. The most common is estimated at $7.5 billion by the U.S.
method is using explosives to sever the National Research Council.3 In 1987 it
piles below the mud line, which kills fish was reported that there were 139 offshore
and sea turtles. In addition, many deep structures in the North Sea of which 40
water structures are too heavy to remove are deep water structures. The estimate
with the existing derrick barge capacity for platform removal for the North Sea is
after toppling. Accelerated galvanic approximately $8.9 billion.4 BHP estimates
corrosion is proposed as a new concept to total costs for clearing 13 platforms in
destroy offshore platforms. In one scheme the Bass Straights offshore Australia would
envisioned, the use of accelerated galvanic run around $1 billion.5
corrosion would reduce the quantity of
explosives required for platform toppling, Alternatives to platform removal have been
thereby minimizing ecological damage to considered. One possible use is recycling
@ea life. The accelerated galvanic corrosion the platforms and reutilizing them. Still
is achieved by applying a special copper another program has been toppling the
clamp-on, device to the platform structural platforms in place to create artificial
member that is to be destroyed. Galvanic reefs for marine life habitats, such as
corrosion rates of eight inches of steel Louisiana's Artificial Reef Program (LARP).
per year occur at a slit placed in the copper However, the U.S. Navy has raised objections
film. Steel members one inch thick are over artificial reefs, claiming foreign
severed by corrosion in about two months. submarines could use them for cover.6 Another
innovative solution proposed for the North
Sea by the British Ministry of Energy is
the Fairy Lights concept. In this concept
BRIEF HISTORY OF PLATFORM REMOVAL the platforms would be illuminated and left
in situ, becoming beacons in the ocean as
A 1958, Geneva treaty requires that offshore a warning to marine traffic.7
platforms be removed within a year after
they stop producing. Branches of the Interior EXISTING TECHNOLOGY FOR PLATFORM REMOVAL
and Commerce departments would like some
platforms left in place or used as building Virtually all of the platforms that have
blocks for artificial reefs. The United been recovered offshore to date have been
States, along with Britain and Norway, is removed by explosives. This usually consists
urging treaty changes to allow this. However, of placing a charge within the platform
the U.S. Navy has called the platforms a legs below the mudline and remotely detonating
threat to national security. Several coastal the. explosive. The platform jacket is then
states have enacted "Save the Ri gs" refloated and picked up by a derrick barge.
legislation. Louisiana has designated eight Finally the jacket is placed on a dumb barge
areas where obsolete rigs can be scuttled. and brought ashore.
It has also established conditions for
transferring liability for the old rigs The problem with explosives is the destruction
from oil companies to the state.1 of fish and sea turtles, which congregate
in large numbers around offshore structures.
Over the past 40 years, 4,187 platforms An additional problem is the lift capacity
have been installed on federal Gulf of Mexico of existing crane barges. Until recently
leases. Of that total, 3,080 fixed platform crane barge capacity was limited to 2,000
drilling structures remain in place. it tons. There is now a single crane barge
is estimated that between January 1, 1947 with a pickup capacity of 16,000 tons.
and August 31, 1987 a total of 659 structures Even this capacity is far short of the largest
have been removed from the Gulf of Mexico.2 steel jacket placed so far, platform
In June 1988 it was reported that there Bullwinkle, which weighs over 50,000 tons.

CIH2565-6/88/00007 577 $1 @1988 IEEE

Proposed Platform Removal Techniques Equations 1 and 2 below.
There are a number of innovative techniques8 Anodic Reaction
that have been proposed for the removal (oxidation)
of offshore platforms, among which are cutting
and sawing devices, either manned or operated Fe -@ Fe++ + 2e- (1)
by remote operated vehicles (ROV's). These
proposed solutions have obvious drawbacks. Cathodic Reaction
For example, the cutting and sawing operation (reduction at the copper cathode)
by either man or ROV's is potentially
dangerous. At some point during the cutting 02 + H20 + 4e- -4 40H- (2)
or sawing operation when the structure
collapses either the man or the multimillion Our experiments show that reaction 2 is
dollar ROV will be imperiled. In the Fairy the controlling mechanism. Therefore, the
Lights concept the oil companies face the more oxygen that reacts at the cathode (the
continued liability of injury to trespassers copper surface of the ACA), the greater
on the platforms. In addition, the structures the steel corrosion rate at the anode.
pose an obstacle to both military and Because of limited oxygen solubility in
'commercial marine traffic. sea water, about 6 to 10 mg/kg (milligrams
per kilogram), the reaction is limited by
The author believes that the use of the cathode (copper) area. The result as
accelerated corrosion as a method of platform shown in Table 1 is that the larger the
removal has merit when compared to existing copper to steel ratio, the higher the
concepts. Destruction of platforms by corrosion rate of steel.
accelerated corrosion can be utilized to
remove existing structures. DESCRIPTION OF THE
ACCELERATED CORROSION ASSEMBLY
HOW THE CONCEPT WORKS
Figure 1 shows the details of the accelerated
Accelerated corrosion and ultimate collapse corrosion assembly (ACA). A variety of
of the platform member is achieved through ACA sizes will be required to fit the many
galvanic corrosion. Accelerated galvanic tubular diameters encountered in offshore
corrosion is achieved through the well-known pl atform construction. A custom fit is
technique of using a large cathode and a not required, however. For example, a single
small anode. Corrosion rate is proportional ACA is expected to fit tubulars four feet
to the cathode/anode ratio. This is achieved to six feet in diameter. As shown in Figure
by utilizing a large copper cathode with 1, the essentials of ACA are the critical
a thin circumferential slit at the center. gap of 0.005 inches in the slit at the center
This slit is centered over the platform of the assembly. This slit is located at
member at the point where the tubular member the exact point where the platform tubular
is to be severed by corrosion. This geometry is to be severed. The ACA is first secured
produces cathode to anode ratios exceeding at the location on a vertical or horizontal
10,000:1. Such a cathode to anode ratio tubular where destruction by corrosion is
will produce galvanic steel 'corrosion rates desired. Securing is accomplished easily
in sea water of approximately eight inches by cinching straps as shown in Figure 2.
per year. Table 1 shows corrosion rates Thermite charges are then battery detonated,
measured in our laboratory for varying cathode which welds the stainless steel blocks of
to anode ratios. the ACA to the tubular members as shown
in Figure 3. The lightweight copper frame
Corrosion Mechanism of ACA is securely fastened to each stainless
steel block by stainless steel rivets.
The accelerated galvanic corrosion cell The in situ ACA is shown in Figure 3. The
has four components. These are a large resultant geometric configuration will give
copper cathode; a small steel anode (the the large cathode to anode ratio necessary
.005 inch slit in the accelerated corrosion to achieve the high rate of corrosion required
assembly); an electrolyte (sea water); and for rapid destruction of platform tubulars.
a metallic conductor between the anode and
cathode. In the accelerated corrosion DISMEMBERMENT SEQUENCE
assembly (ACA), this conductor is the thermite
weld which metallically fuses together the Platform dismemberment by accelerated
ACA and the platform tubular. The corrosion requires the placement of the
electrochemical reactions are shown in accelerated corrosion assembly (ACA) over

578

the tubular member to be severed. The protrudes nearly 400 feet above the ocean
assembly is placed around the tubular member floor. Proper placement of ACA's could
by diver or ROV. In some cases the ACA reduce this 400 foot protrusion to between
can be utilized for platform toppling. 30 feet and 50 feet after about six months
In other cases it will be used for platform of accelerated corrosion. This would create
dismemberment. When used for platform a barrier reef marine habitat in situ and
dismemberment, placement of the ACA can avoid costly refloatation and the leasing
occur either prior to or subsequent to of a heavy duty derrick barge.
platform toppling depending on the economics.
Water depth will usually determine whether There are a number of platform removal
the accelerated corrosion assembly (ACA), variations where the use of ACA would be
when used for platform dismemberment, is advantageous. ACA would be more economical
placed before or after platform toppling. than platform removal for in situ destruction
As a hypothetical example, for a jacket of the platform and creation of marine
in 1,000 feet of water, let us make the habitats. As a natural process, corrosion
assumption that it is acceptable to leave not only does not ham the environment,
the lower 150 foot section of the jacket but it also eliminates or reduces the use
secured by piles in place. The balance of explosives and their attendant detrimental
of the jacket is to be cut by ACA into pieces, effect on marine life.
none of which weigh more than the capacity
of a medium sized crane barge, e.g., 2,000 COMPARISON OF ACCELERATED CORROSION
tons. The accelerated corrosion assemblies CONCEPT TO ALTERNATIVES
are placed at the appropriate points. if
all the ACA's have the same cathode to anode Table 2 shows the ACA in comparison with
ratio, the uppermost platform segments, other forms of platform removal. Two things
since they have lighter wall tubulars, would stand out. ACA reduces the risk to both
be expected to topple first. However, the man and marine life by a substantial margin.
time to failure sequence can be controlled ACA is estimated to cost about 10% more
by varying the cathode to anode ratios on than platform removal using explosives and
the ACA's. The time frame for platform medium capacity crane barges plus floatation.
dismemberment ordinarily would vary from
two to six months. The crane barge recovers CONCLUSIONS
the more manageable 2,000 ton subsections
and places them on dumb barges. These From the above analysis it can be concluded
subsections can be either brought ashore that since economics usually dictates choices,
or placed in marine barrier reefs. Removal the existing techniques of explosives, medium
of the 150 foot platform base will require capacity lift barges, and floatation equipment
explosives or other means. The pile guides will continue to be used over the alternatives
prevent attachment of the ACA's directly analyzed unless society is willing to pay
to the piles. However, even if explosives a premium for reduced risk to man and nature.
are now used to remove the 150 foot base, If reduced environmental damage warrants
the great majority of marine life that was a premium, then this needs to be spelled
clustered around the upper 850 feet of the out in our national policy. The accelerated
jacket will be spared. In this concepti, corrosion assembly concept could become
two important goals are realized. Damage a viable alternative to existing removal
to marine life is minimized. Large savings techniques within a year once national
are achieved by sectioning the platform priorities are established.
into smaller segments that can be removed
by a 2,000 ton crane barge instead of REFERENCES
floatation tanks or a costly 16,000 ton
crane barge. 1. Sterba, James P., "Save the Oil Rigs?
Yes, Some Say, They Are Habitat-Forming",
In another scenario, again using the 1 ' 000 The Wall Street Journal, April 29,
foot jacket as an example, the jacket is 1988, page 1.
severed into two sections of approximately
the same weight. Assume that the sections 2. "3,080 Platforms Still in Place in
@elected will require cutting the platform U.S. Gulf", Ocean Oil Weekly Report,
into two parts at approximately -650 feet Volume 22, Number 4, October 5, 1987,
MLLW. The ACA will be appropriately placed page 1.
around platform legs or braces at the -650
foot level by divers or ROV's several months 3. "Platform Removal Costs Could Run in
in advance of platform toppling. Platform Billions; Reuse on Rise", Ocean Oil
toppling is accomplished in the conventional Weekly Report, Volume 22, Number 36,
way by explosives severing the piles below May 16, 1988, page 2.
the mud line. The base of the platform

579

4. Editorial Comment, "Could Corrosion TABLE 1
Assist with Offshore Platform Removal?",
Corrosion Prevention & Control, Volume COPPER/STEEL RATIO VERSUS CORROSION RATE
34, Number 2, April 1987, page 31.

5. "Ease Platform Removal Guides, Operators
Ask Australian Government", Ocean oil
Weekly Report, Volume 22, Number 41, Cu/Fe Ratio Steel Corrosion Rate
June 20, 1988, page 2. inches per year (1py)

6. "Platform Removal Costs Could Run in 1:1 0.152
Billions; Reuse on Rise", Ocean Oil
Weekly Report, Volume 22, Number 36, 1:10 0.638
May 16, 1988, page 2.
1:100 2.59
7. Editorial Comment, "Could Corrosion
Assist with Offshore Platform Removal?", 1:1,000 4.86
Corrosion Prevention & Control, Volume
34, Number 2, April 1987, page 32. 1:10,000 8.11

8. W.L. Alexander Jr., T.G. Jackson, and 1:20,000 10.5
D.J. Hardin, "Engineering the Cost
Out of Platform Removals and Salvage",
Presented at the 20th Annual Offshore
Technology Conference in Houston, Texas,
May 2-5, 1988.

TABLE 2

COMPARISON OF REMOVAL TECHNIQUES FOR A 1,000 FOOT PLATFORM

Explosives, Underwater
ACA and Medium Crane Cutting
Medium Explosives Barge, and Devices
Crane and Heavy Buoyancy and Medium
Barge Crane Barge Tanks Crane Barge Comments

1. Total Time for Platform 18 4 8 12 Explosives plus heavy
Removal (months) crane barge are most
rapid for removal.

2. Total Time for Divers or ROV 4 1 3 8 Explosives plus heavy
on Station (months) crane barge minimize
diver or ROV time on
station.

3. Comparative Time for ROV or 4 1 2 20 Explosives plus heavy
Divers Underwater crane barge minimize
ROV or diver time
underwater.

4. Danger to Man and Marine Life 1 10 20 5 Accelerated corrosion
assembly reduces risk
to man and marine life
by a large margin.

5. Relative Cost 1.1 1.5 1.0 2.5 Existing removal
techniques with
explosives, medium
crane barges, and
buoyancy tanks is
least expensive.

580

.FIGURE 3

ACA ATTACHED TO TUBULAR

FIGURE I

ACCELERATED CORROSION ASSEMBLY

WELD CONTAINER

TUBULAR

STAINLESS STEEL
ATTACHMENT DEVICE FOR
15-- ELEC./MECH . CONNECTION

A
10 KL. COPPER (TYP.)
A 5 NIL. GAP WNCHING STRAP
OVERLAP TO 30 ML. STEEL (TYP.)
PERMIT VARYING
DIAMETER TUBULAR

STAINLESS STEEL THERMITE
A-9-- WELD CONTAINER

WELD METAL FUSED TO SURROUNDING METAL

STAINLESS STEEL RIVET -
ATTACHING STAINLESS STEEL
WELD CONTAINER TO COPPER
CLAD STEEL

5 ML. GAP (TYP.)

- 10 ML. COPPER (TYP.) COPPER TUBULAR
14- CLAD -z@
SECTION DETAIL AA ju 14L. STEEL (TYP.) STEEL SECTION DETAIL BB

ELECTRICAL/MECHANICAL ATTACHMENT

FIGURE 2

CINCHING STRAP

PLASTIC TENSION
HANDLE

PLASTIC
RATCHETING
HEAD

PLASTIC
STRAP
CA
'0

OV RLAP To
PE
RMIT VARY ING
D
I
TER TUBULAR

TAINLESS S
@&--@WELD CONTAI
WELD METAL
?IBI

581

CORROSIVE-WEAR OF BUOY CHAIN

Craig A. Kohler

U.S. Coast Guard Research and Development Center
Avery Point
Groton, Connecticut 06340-6096

ABSTRACT

Laboratory studies were conducted to
determine the causes of buoy chain
degradation and how the corrosive-wear buoy
resistance of steel could be improved. wave wave
Undisturbed corrosion and accelerated h ight crest wind
corrosive-wear tests were conducted on
plain carbon and alloy steels. Analysis
of the tests included the effects of
composition, hardness, and microstructure
on wear resistance. T
wave Chain Mooring
Chain moorings were evaluated for pitting water trough (90 ft) ---------
corrosion resistance, reduction in link depth
diameter, overall weight loss, and ob- (35 ft)
served appearance after a two year f ield current
test. Results suggest which steels may sinker -------
provide increased corrosive-wear resis- (20001 .bs)
tance and longer chain service life.

"i4e, r
1. INTRODUCTION

One of the primary missions of the U.S. FIGURE 1. Buoy and Mooring
Coast Guard is maintaining the aids to
navigation within the waterways of the
United States and its territories. Per- Although the mooring chain acts as a
formance of this mission requires periodic damping mechanism, the chain itself is in
inspection and repair of a large number of nearly continual motion. It is the
fixed and floating structures. Servicing motion of the chain which produces wear
of aids to navigation accounts for a of the links. Interlink wear occurs be-
significant portion of the Coast Guard's tween the contact surfaces of adjacent
budget. In an effort to reduce servicing links. The section of the chain which is
costs, the Coast Guard has established a in contact with the bottom, referred to
goal of a six year life cycle for its as the chafe section, is subjected to
fixed and floating aids to navigation. abrasion. Therefore, the chafe section is
exposed to additional wear on the outer
The typical floating aid to navigation, surfaces on the links, referred to as
illustrated in Figure 1, consists of a barrel wear.
buoy, steel chain, and concrete sinker.
Each component is designed to withstand The corrosion of plain carbon steel in
the maximum anticipated wind, current, and seawater is well established. Uniform
wave f orces. Resistance to these forces corrosion rates of 10 to 15 millimeters
are provided by the horizontal component per year (mpy) in the f irst year, then
of the tension in the chain created when slowing to less than 5 mpy after 2 to 3
the chain is picked of f the bot 'tom and, years for undisturbed specimens have been
ultimately, the horizontal holding force documented [1]. However, the ferrous
of the sinker. Additionally, the weight oxide surface layer which reduces
e
_7+ 7__@
,noring _ r

of the chain suspended in the mooring acts corrosion is loosely adherent. When the
as a shock absorber against dynamic oxide layers are removed, the initial
loading. corrosion rates are re-established.

582
United States Government work not Protected by copyright

Corrosive-wear is the synergistic effect The first set of laboratory experiments,
of the corrosion and wear processes which conducted in 1984 by D. May, will be
results in the rapid degradation of steel referred to in the subsequent text as
chain in seawater. The protective oxide Experiment A. The second set of laboratory
layers are continually removed by the wear tests, conducted in 1985 by C. Kohler,
process exposing unprotected metal. The will be referred to as Experiment B. Both
corrosive-wear mechanism is most destruc- studies were funded by the U.S. Coast
tive to the chafe section of the chain Guard [2].
mooring. In Figure 2, the chafe section
in the right of the picture was worn to
the point where failure of the mooring
?ccurred. Chafe section wear such as this
is most common in areas of high currents 2. LABORATORY TESTING:
and wave action. EXPERIMENT A

The research in Experiment A included
three objectives. The first objective was
to identify which mechanisms were
responsible for the majority of chain
material loss. The second was to design
and construct a laboratory testing
'apparatus which would simulate buoy chain
degradation. Finally, by testing several
other steels and comparing the results to
"A the baseline 1022 steel it could be
determined which material properties were
most important in reducing corrosive-wear.
V
The apparatus used to conduct the testing
4
is shown in Figure 3. A plastic tank
contained the distilled water and seawater
for the tests. Mounted on the bottom of.
h,
the tank was a sanding disk used to
provide an abrasive surface for the
corrosive-wear tests. The carriage and
s
v, pecimen holder maintained contact between
the steel specimen and the abrasive disk.
The loading and motion of the apparatus,
were designed to simulate those seen by
FIGURE 2. Abraded Buoy Chain buoy chain links.

Presented in this paper are the results The steels selected for study in
from two laboratory studies and a two year Experiment A, in addition to the baseline
field test on the effects of corrosive- ASTM 1022 steel, were ASTM 1045 and 4140
wear on plain carbon and alloy steel buoy steels. Selection of the 1045 steel wag
chains. The compositions of the steels to determine the effects of increasing the
that were tested are listed in Table 1. carbon content from the baseline 1022. The

TABLE 1
STEEL COMPOSITION

Hardness
steeL 9 tin !LO (Rc)

1022 0.18-0.23 0.7-1.0 0.04 0.05

1022HT 0.18-0.23 0.7-1.0 0.04 0.05 41

4140 0.38-0.43 0.75-1.0 0.035 0.04 0.15-0.35 0.8-1.0 0.15-0.25 28

4140HT 0.38-0.43 0.75-1.0 0.035 0.04 0.15-0.35 0.8-1.0 0.15-0.25 20

4340 0.38-0.43 0.65-0.85 0.035 0.04 0.15-0.35 1.65-2.0 0.7-0.9 0.2-0.3 27

8740 0.38-0.43 0.75-1.0 0.035 0.04 0.15-0.35 0.4-0.7 0.4-0.6 0.2-0.3 25

8620 0.18-0.23 0.70-0.9 0.035 0.04 0.15-0.35 0.4-0.7 0.4-0.6 0.15-0.25 30

583

The weight loss coefficients were plotted
as a function of time. of particular
interest was the shape of the curves for
each material. For the f irst few hours,
there appeared to be an incubation period
Motor which was characterized by a rapid
increase in the material loss coefficient.
This phenomena is characteristic of the
F- Drive Shaft @@DM ion=> wear process due to the abrasive cutting
grooves into the material. Following the
Calomel incubation stage was a steady state
Electrode region. During this phase the surface of
the specimen experiences consistent wear
due to the reduction of asperities on the
Ll Specimen abrasive surface. corrosive-wear of buoy
Weight Abrasive chains occurs almost entirely in the
Disk y state region. Therefore, it is the
stead
region which
slope in the steady state
I provides the best indication of corrosive-
wear rates. A summary of the data
obtained from all the graphs is given in
Table 2.
FIGURE 3. Tank and Motor Set-up Apparatus:
Experiment A
TABLE 2
SUMMARY OF MATERIAL LOSS COEFFICIENT RATES

MATERIAL LOSS COE -FJICIENT
STEEL TYPE OF TEST RATES (x1O )
influence of alloying additions on 2_
corrosive-wear would be investigated in 1022 UNDISTURBED SEAWATER 1.84 mg/cm hr
the 4140 steel. The effects of hardening
by changing the microstructure was 1045 UNDISTURBED SEAWATER 1.93
achieved by heat treating the 1022 steel
at 1650*F for thirty minutes followed by a 4140 UNDISTURBED SEAWATER 1.79
water quench. The heat treatment produced
a martensitic surface layer. 1022HT UNDISTURBED SEAWATER 2.07

Undisturbed corrosion tests were con- 1022 CORROSIVE-WEAR 161
ducted first to provide a baseline for (DISTILLED)
analysis of the corrosion resistance of
the steels and its effect on the 1622 CORROSIVE-WEAR 240
corrosive-wear process. Identical spec- (SEAWATER)
imens of each steel were immersed in
distilled water and seawater for 120 and 1045 CORROSIVE-WEAR 289
180 hours. Weight losses were calculated
for all specimens. 4140 CORROSIVE-WEAR 52

Following completion of the undisturbed 1022HT CORROSIVE-WEAR 73
corrosion tests, corrosive-wear tests were
performed using the 1022 steel in both
distilled water and seawater. The data One of the objectives of Experiment A was
would provide insight into the effects of to determine the effects of the corrosive
the corrosive medium on wear. Corrosive- medium on the corrosive-wear process. By
wear tests in seawater were performed on comparing the results of the undisturbed
all the steels. As with the corrosion tests with the corrosive-wear tests,
tests, weight loss measurements were differences of one to two orders of
determined. magnitude occurs in the material loss
rates. Wear was the dominant material
The corrosive-wear testing apparatus loss mechanism. The results between the
presented a problem with the measurement 1022 steel corrosive-wear tests in
of material weight loss. The steel distilled water and seawater demonstrate
specimens had unequal abraded surfaces. that the wear increases by 50% in a
Therefore, the applied pressures on the corrosivemedium.
contact surfaces were inconsistent. A
material loss coefficient was calculated wear rates are greatly effected by the
by dividing the specimen material loss by material properties of the steels. The
the abraded area to account for the 1045 steel had the highest material loss
difference in area. rates as shown in Table 2. This steel is

584

identical in composition to the 1022 steel The steels tested. in this experiment
with the exception of a higher carbon included the 1022, 4140, heat treated
content. Abrasion resistance improves with 4140, 4340, and 8740 steels. With the
increased carbon content and hardness. exception of the common buoy chain steel,
However, with sufficient stresses, wear 1022, all the other steels had similar
may occur through delamination of the carbon contents. Heat treatment of the
carbide plates below the surface (3]. 4140 steel consisted of austenitizing at
Large amounts of material are removed by 1600*F for one hour followed by slow
this process. The martensitic micro- cooling over a 24-hour period, which
structure produced by heat treatment of produced a more ductile metal. The 8740
the 1022 steel resulted in improved wear and 4340 were selected due to their
resistance. The greater hardness produced increased nickel content.
by this microstructure resists abrasion
while preventing delamination from The corrosive-wear testing procedure was
occurring. Alloy additions in the 4140 similar to that used in Experiment A.
steel provided the lowest material loss Weight measurements were taken prior to
rate. The solid solution hardening of testing and following completion of the
this alloy is most effective in reducing testing. The material loss measurements
abrasion. were graphed as a function of the test
durations. The graphs showed initial
rapid material loss followed by reduced
3. LABORATORY TESTING: steady state wear. The slope of the
EXPERIMENT B curves in the steady state region, given
in Table 3, were obtained to compare the
The results of Experiment A indicated that corrosive-wear resistance of the steels.
the microstructure of the steel was the
most important factor in controlling
material loss in a corrosive-wear TABLE 3
environment. Several different micro- MATERIAL LOSS RATES: EXPERIMENT B
structures were examined in Experiment B
to identify a better wear resistant
microstructure. STEEL WEAR RATE

A new corrosive-wear testing apparatus was 1022 0.46 mg/hr
designed for Experiment B due to problems
in the first experiment with uneven 4140 0.19
abraded surfaces. The apparatus, shown in
Figure 4, was designed to provide abrasion 4140HT 1.42
across the entire surface of the steel
specimen. Although the magnitude of the 4340 0.10
load was less than that used in Experiment
A, the frequency of the loading was 8740 0.21
identical.
The material loss rates indicate that the
4340 steel had the best corrosive-wear
resistance, followed by the 4140, then the
an 8740 steel. The 1022 steel was relatively
Specimen Holder Shaft poor in comparison to the alloy steels,
F- Abrasive Disk Linear Bearing except for the heat treated 4140 steel.
Shaft Guide - The controlling factor for the corrosive-
wear rate of steels with ferrite-pearlite
structures appeared to be the ferrite to
pearlite ratio. Material loss decreased
with a decreasing ferrite to pearlite
ratio. The control of this ratio by
suitable alloy selection, heat treatment,
and processing techniques would therefore
j seem to hold promise in improving buoy
chain wear resistance.

Due to the differences in the loading of
the specimens in Experiments A and B,
quantitative comparison of the material
FIGURE 4. Laboratory Testing Apparatus: loss rates are not possible. However, the
Experiment B relative corrosive-wear resistance were
consistent for the same steels tested in
both experiments. Of greater importance
was whether the results could be verified
by field testing.

585

4. FIELD TESTING bottom characteristics in the vicinity of
the sinker are provided from the diver
A two year buoy chain field test was reports. Percentage weight losses of the
conducted using several of the alloy complete chains are listed in the second
steels examined in the laboratory tests column. In the third column the maximum
[4]. The objective of the field testing reduction in link diameter are listed as a
was to determine whether the results of percentage of the original measurement.
the accelerated laboratory tests were In all chains this occurred in the curved
valid for long term exposures. section of the links. The relative
Particularly of concern was pitting pitting corrosion resistance compared with
corrosion of the alloy steels. Pitting the other chains are shown in the fourth
corrosion was determined to be a major column.
cause of failure of moorings in tests
conducted by the Canadian Coast Guard [5]. By observation of the weight loss data it
appears that it is largely influenced by
The buoys were moored at a location the ocean f loor rather than the type of
approximately one mile south of chain. Both steels exposed to a sandy
Charlestown Breachway on the Rhode Island floor lost the least weight while a rocky
coast. The location of the buoys was floor resulted in maximum weight loss.
selected for several reasons. The most Although general weight loss is not
important of these is that the buoys could critical to the strength of the mooring,
be monitored at frequent intervals. lower weight results in greater movement
Secondly, the environment was semi-exposed of the chafe section due to reduced
with an average wave height of 2.4 feet damping in the riser section.
and current of one knot. Finally, the
location provided a water depth of 35 feet The link diameter measurements were the
with a rough bottom consisting of rocks most critical part of the data. Re-
and sand. duction in link cross section will
eventually result in failure of the
The five different steels selected for the mooring. These measurements also have the
mooring chains were the conventionally highest potential for error as evidenced
used 1022 plain carbon steel, and alloy by the 8740 steel chain. The reduction in
steels 4140, 4340, 8740, and 8620. The link diameter was much less than all the
compositions of the steels are given in other chains; however, the chain weight
Table 1 along with their respective loss was one of the highest. A reasonable
hardness values. The variations in explanation for this occurrence was
composition were primarily chromium, discovered in the diver inspection
nickel, and molybdenum. The 4340 steel reports. The 8740 chain was found to be
contained the highest total of these tangled in rocks in both June and October
alloying additions with the 8740, 8620, 1986. Estimating the chain to have been
then the. 4140 steels decreasing in alloy tangled for at least six months, the
additions. Following the manufacture of reduction in wear would be accounted for
the alloy steel chains, tempering to a by the lack of movement in the chafe
hardness of less than Rc30 was conducted section. The remainder of the chain would
to avoid premature failure by stress still experience wear on the hard rocky
corrosion cracking in the marine bottom which would explain the high
environment. overall weight loss. The 8740 chain link
diameter data was therefore considered to
Prior to commencement of the testing in be invalid.
August 1985, the chains were weighed and
link diameter measurements were recorded A more practical method for evaluating the
at two locations on each of ten links. performance of the chains is to compare
The moorings were inspected quarterly the smallest link diameter measurements.
during the testing to insure they were in Using the weakest link data, the 4340
good condition and that none were wrapped steel performed better than all the other
around their sinkers. The buoys were steels even though it was located on a
subjected to several severe storms mostly rocky bottom. The typical buoy
including Hurricane Gloria in September chain steel, 1022, exhibited the worst
1985. The testing was concluded in July performance.
1987. The chains were reweighed and dia-
meter measurements were taken on each link A comparison of the 4140 steel with the
in the chafe section at the same positions 1022 steel chain showed some unexpected
on the links as had been previously results. Both chains were located in.sand
recorded. Visual examinations of the and from the laboratory testing the 4140
chains were conducted. Evidence of steel should have greatly out-performed
pitting or unusual wear was recorded. the 1022 steel. However, there was little
weight loss and diameter reduction
The results of the buoy chain tests are difference between the two steels. A
shown in Table 4. In the first column the likely possible explanation for this re-

586

TABLE 4
SUMMARY OF BUOY CHAIN FIELD TESTING DATA

Weight Reduction in Resistance
Steel Floor Type Loss Diameter to Pitting

1022 Sand 9.5 41.1 Poor

4140 Sand 9.0 39.3 Poor

4340 Sand+Rock 11.2 32.6 Excellent

8740 Rock 12.5 Invalid data Fair

8620 Sand 13.1 39.6 Good

sult is that interlink wear was the the majority of the , material loss.
primary mechanism for material loss. Link Additionally, the 4340 steel showed
shape reflects this theory as the curved superior performance in the comparison of
section having considerably more material maximum link reductions. Improved service
loss on all chains than the other sections life of chain moorings can be achieved
of the link. However, the 4340 steel through the use of 4340 steel. The steel
showed only minor differences in wear will be of maximum benefit in areas where
along the curved section in comparison to the conventional chain has been worn in
the rest of the links. Consequently, relatively short periods of time.
barrel wear was responsible for the
majority of wear observed for this chain.
Therefore, the results demonstrate that 6. REFERENCES
interlink wear is a much more damaging
process than barrel wear. This conclusion (1] Dexter, S.C., "Handbook of
is not only supported by the data but Oceanographic Engineering Materials,!' John
observations as well. Wiley & Sons, Inc., 1979.

5. CONCLUSIONS [2] Kohler, C.A. and May, D.A.,
"Corrosive-Wear of Buoy Chain," Interim
From the data collected in this study, the Report, U.S. Coast Guard Report No. CG-D-
best material to replace the 1022 steel 21-86, July 1986.
for buoy chain moorings would be 4340
steel chain tempered to a hardness of (3] Suh, N.P., Jahanmir, S. and
Rc28. The accelerated laboratory testing Abrahamson, E.P., I'Delamination Theory of
indicated a four to five fold decrease in Wear," Progress Report to DARPA, Materials
material loss. The two year buoy f ield Processing Laboratory, Department of
test proved the 4340 steel to have the Mechanical Engineering, MIT, Cambridge,
best pitting resistance due to its high MA, September 1974.
nickel content. From a weight loss
perspective, other chains performed (4] Brown R. and Kohler, C.A.,
better. However, responsibility for the "Corrosive-wear of Buoy Chain," Final
majority of the weight loss was due to the Report, U.S. Coast Guard, in publication.
rocky ocean floor. The link diameter
measurements were supporting evidence of [5] Laing, A.K., Buhr, R.K., and
the benefits of the steel. The 4340 Gertsman, S.L., "Navigational Buoy Mooring
exhibited the lowest average percentage Chains," Technical Report by the Canadian
reduction in diameter in the chafe section Department of Mines and Technical Surveys,
where interlink wear was responsible for 1965.

587

COLLISION TOLERANT PILE STRUCTURES: DESIGN ANALYSIS SOFTWARE

Marc Briere, Kenneth C. Baldwin, and M..Robinson Swift

Mechanical Engineering Department and Ocean Engineering Program
University of New Hampshire, Durham, NH, 03824, U.S.A.

ABSTRACT
A Collision Tolerant Pile Structure
(CTPS) is a compliant, hinged structure
designed to carry Aids to Navigation
(ATON's) in shallow navigable shipping
channels subject to heavy barge traffic. ATON
The project goal is to develop a software
package to serve as a tool for full scale
CTPS design based on site specific
conditions. The software runs on MS-DOS,
is interactive and menu driven, and
utilizes the LOTUS 123 spreadsheet.
The software includes recent design MIDDLE BUOYANT SECTION
improvements involving an improved spring
configuration and the use of buoyancy and
simulates CTPS dynamics during collision,
storm and normal operating conditions.
Typical model inputs include wave height,
current velocity, water depth, and barge
draft, speed, freeboard, and bow angle.
Using site specific values in a generic
model, based on Coast Guard
specifications, parametric studies are LOWER SPRING SECTION
.readily performed.

UPPER BASE

LOWER BASE
I.INTRODUCTION

Background

The rigid pile structures currently in
use for deploying aids to navigation
channels and
(ATON ' s) in coastal WINGED FOUNDAT,
intercoastal waterways are frequently
destroyed by towed barges straying out of
the main channel. Replacement of these Figure 1. Central Stay CTPS system and
structures is costly and their absence component configuration. Foam cladding
from the waterway poses both an can also be attached to the lower spring
environmental and safety hazard. A section.
possible solution to this problem is the
replacement of existing rigid pile
support by a compliant Collision Tolerant
Pile Structure (CTPS) - This compliant section has a buoyant core and outer foam
structure would maintain verticality cladding while the lower section contains
during normal operating conditions and a spring assembly. Upon impact the
yet be able to "give way" to a colliding structure rotates about the bell / base
vessel. pivot point allowing the barge to pass
The CTPS configuration consists of an without seriously damaging the structure.
ATON mounted upon a series of two pile After the barge passes the spring /
sections and a bell / base combination buoyancy recovery mechanism acts to
which is inserted into a winged, restore the pile to an upright position.
foundation (see Fig 1). The middle pile The objective of this present work is
0, Tr@

CH2585-8/88/0000- 588 $1 @1988 IEEE

to develop software to serve as a tool in the central stay system is a clear
analyzing and designing a full scale CTPS advantage over that of the peripheral
for site specific conditions. The stay.
software allows the designer to perform The success of the 1/15 scale, central
parametric studies to determine the stay physical model prompted efforts to
effect of the design parameters on the design, build and test a 1/4 scale (10
structure's performance. The program ft.). central stay physical model. The
requests environmental inputs and encouraging results of the investigation
provides the user with information on how are described in Swift et al. (1986).
CTPS dynamics are modified due to design Baldwin et al. (1987) summarized all the
parameter changes. Utilizing a menu 1/4 scale results and confirmed that the
driven structure the user is also central stay hinge concept is indeed a
provided the option of viewing viable design alternative to the
intermediate calculations (ie. hinge, peripheral stay configuration. The next
gravitational, and buoyancy moments vs. step, therefore, is to design, build and
pile inclination). test a full-scale structure utilizing the
central stay concept.

Review of previous work

The initial development of the CTPS II. PRESENT FULL SCALE DESIGN
concept was performed by Swift and
Baldwin (1985). Their efforts were Hinge Concept
directed toward developing analytical
tools for evaluating design concepts, The present full scale hinge concept
choosing the best hinge concept for is based on the 1/4 scale, central stay
further study, and constructing and physical model. However, the increased
testing a 1/15 scale (2 1/2 ft.) physical inertial and gravitational forces
model. Using a mathematical model and resulting from scaling considerations
design criteria specified by the place. higher demands on the spring
Coast Guard, S w i f t and Baldwin mechanism. The existing 1/4 scale spring
(1985,1986) identified two potentially design is incapable of sustaining the
successful hinge concepts - a central increased loading requirement. The
universal joint, peripheral stay investigation of a new spring design is
arrangement and a central stay system. described in Ward et al. (1988). The
The peripheral stay concept received results of this investigation indicate
initial emphasis and its development has that a compressive rather than a tensile.
been reported by Swift and Baldwin spring is best suited for the
(1985,1986), Cloutier et al. (1985), application.
Durkee (1986) and Mielke (1986). The new spring design consists
Subsequently, however, the central stay primarily of three compression coil
arrangement was developed as a design springs connected in series by couplings
alternative. and a termination plate on top (see Fig.
1). A chain running through the centers
The major components of the central of the springs is connected to the
stay system, as described in Swift et al. foundation as well as the termination
(1986), include the base, the bell, the plate. In addition to this change in
central stay attached to a pre-stressed spring design, buoyancy is added via a
tensile spring, and the hollow pile. As foam core and foam cladding to increase
the pile tips, it pivots about the the restoring moment aiding in the
bell/base contact point, while the stay recovery process. Thus, as the pile tips
and buoyancy forces provide the restoring the spring mechanism is enabled and the
moment. buoyant section submerged creating a
The development of the central stay restoring moment about the bell/base
system began during the 1985-1986 pivot.
academic year when a senior student There are a number of advantages in
design team embarked on the design, u s i n g t h i s n e w compression
construction and testing of a 1/15 scale spring/buoyancy recovery mechanism. To
physical model. Their objective was to begin, the pile no longer acts as a
determine the bell/base geometry which compression member thereby significantly
would minimize the possibility of jamming reducing the possibility of buckling.
while maintaining the stay moment arm Secondly, the configuration of this new
with respect to the contact point. Their spring mechanism places its center of
results showed the central stay concept gravity closer to the hinge which reduces
to be successful and thus a viable the gravitational moment produced by the
alternative to the peripheral stay spring weight. Since the springs are in
arrangement. In fact, the removal of compression the possibility of over
friction during the recovery process of extension failure is eliminated. In.

589

addition, the buoyancy reduces the
requirements on the spring by acting to Rb
offset the gravitational moment.

III. MATHEMATICAL MODELING

The modeling approach used is similar
to that found in Swift and Baldwin
(1986). The CTPS system is considered a
rigid pile - flexible hinge system where Ly@ Fs
the hinge is omnidirectional, possessing
a restoring moment stiffness via the A
spring mechanism described above. In HC
keeping with the worst case approximation
the directions of the wind, wave, current
and barge motion are assumed collinear. Hb
In addition, all loadings and resulting
motions are restricted to the vertical RC
plane.
The governing system dynamic equation
is the time rate of change of angular
momentum equation applied at the "fixed" Figure 2. Central stay hinge concept
bell-base pivot point, showing pertinent parameters for
IH 6 = I MO (1) determination of hinge moment (MH).

where IH = moment of inertia about the
pivot point 0, e = angle of pile with
respect to the vertical, ( .. ) indicates
two time derivatives with respect to time
t, and ?10 refers to moments applied about
the pivot 0. Depending upon the
application the applied moments may
include the effects due to the hinge ye
stiffness, buoyancy, gravity, wind,
relative fluid motion and barge contact. 61 G D
The hinge moment, MH, created by the
central spring has the form
MH = -FS(RCcos0L+ HCsin-w.)sgne (2) WT

where FS = spring force, and RC and HC
are the cone radius and cone height,
respectively, as shown in Fig. 2. The
(-) sign and sgnG are in accordance with R91
the sign convention used to indicate that r
the hinge moment always acts to restore @b/ R 82
the pile to vertical.
The upsetting gravitational moment,
MG, due to the "dry" weight of the CTPS
is given as
Figure 3. Central stay CTPS dynamic
MG (yGsinesgne - RBcOs@)WTsgnO (3) modeling nomenclature.
where YG and RB are shown in Fig. 3 and
WT is the total CTPS weight. The moment, MW, acting on the pile as
The buoyancy moment, MB, resulting a result of (steady) wind is given in the
.from fluid displacement has the form form

MB = -62.4(RBlVSI + RB2VS2)sgne (4) MW = FW @D + 0.5 (LT-LS)COSOI (5)
where RB 1 and RB2 are shown in Fig. 3 and where FW = the applied wind force and the
VC1 and VS2 represent the submerged term in brackets represents the moment
v@fumes of the lower and middle pile arm (see Fig. 4). The wind force, FW, is
sections, respectively. The volume of approximated using a drag force approach
the ATON is small and it's buoyancy and can be written as
effect was neglected.
LY,
@H,

590

FW = 0.5 TaCaUa 2 f DO(LT-LS)cosG (6) collision sequences the barge speed is
unaltered and assumed constant. Assuming
where C and U represent the that the barge-pile contact is maintained
3 at at a throughout the collision process allows
density of air, drag coefficient of pile the pile motion to be analyzed
in air and air velocity, respectively, independently of the forces involved.
and the bracketed term f ... I indicates
the projected CTPS area exposed to the Thus, the solution to Eqn. (1) is found
wind. via the barge-pile kinematics. Eqn. (1)
is solved numerically, however, in those
cases where the kinematics do not fully
define the motion. In these cases a
FV Fourth Order Runge-Kutta numerical
(Ly AS casS integration technique I s implemented.
Both of these techniques are fully
developed in Swift and Baldwin (1986).

III. DESIGN SOFTWARE

GIONMT AMC
0+ 1/2(L-r-4)c*s8 Purpose
e D The purpose of the software is to
provide the user with a structured and
easy to use design tool for developing a
site specific CTPS. The goal is to allow
various CTPS configurations to be easily
generated and their performance studied
and compared, giving the user the ability
to conveniently investigate a number of
01 design alternatives and examine their
limits.

Description
Figure 4. Diagram showing the location
of the resultant wind drag force for the A number of conditions are placed on
CTPS. the type of software used in developing
The moment load induced 'by the the design program. It is required that
relative water movementr MCr is evaluated the software used have the ability to
using a form of Morrison's equation easily access tables of pertinent
information, be menu driven, have
MC @ MD + MI' (7) graphics capabilities, and perform
iterative calculations. To do this
where MD and MI' represent the drag and LOTUS 123, a spreadsheet software package
inertial moment loads, respectively. with these capabilities, is employed.
These moments are evaluated using linear The design program (worksheet) is
wave theory and a relative velocity and completely menu driven containing several
relative acceleration normal to the pile. menu levels (see Fig. 5). Each of the
Note that MI I is that part of the total entries in a menu, called menu items, are
inertial forcing moment remaining after used to describe a type of operation. To
the "added inertia effect" has been select an item the user can either type
removed. The added mass itself is the first character of the item desired
incorporated into the mass moment of or, using the arrow keys, position the
inertia about the hinge pivot. A more highlight over the item and press RETURN.
detailed description and full development The main menu categorizes the program
of this approach can be found in operations into three basic types:
Medzorian et al. (1988). Documentation (description), Analysis
(dynamic and static), and output (screen
The applied barge moment, MA, viewing and printing) The lower level
resulting from the barge contact force
during a collision is given in the form menus are Invoked by a higher menu
selection and are more specific than the
higher levels. They provide the user
MA = FALCsin9C (8) with the flexibility to "pick and choose"
where FA = applied barge contact force, the precise operation desired. For
LC = distance from hinge pivot to point example, if the user is interested in the
of contact and OC = angle between pile
direction and direction of barge force.
The approach to solving Eqn. (1) also
varies with the application. During the

591

impulse loadings occurring during the performed by inputting parameter ranges
upper and lower bow impacts he/she need instead 'of single values. These studies
not evaluate the entire collision. By can be done on the system dynamics
selecting Dynamics (D) Collision (C) / (collision, recovery, storm) or the
Impact (I) (see Fig. 5) the user obtains static moments. This item, (S), is still
the evaluation of the impact dynamics under development.
only. The output is easily obtained using
the menu items View (V) and Print (P) .
f-INTRODWT100 DYNAMICS MOMENTS STUDY view PROT QUITI View (V) allows the user to inspect any
- I output resulting from a design session
Ma. Print while Print (P) allows formatted hard
ISterm CoLlWan Recovezy-ExTtl copies of the output. The output formats
I I i include graphs, the corresponding
11mact All --Bxl-ti Warta Data Graph EX10 numerical data, input tables, output
tables, and the CTPS configuration. All
Iftige gravity nuoyawy Restaging of the output, except graphs, is sent
Hil directly to a printer using Print (P). A
hardcopy of the graphical data is
Istorm C01118ken Rocovety ramnts Impact exit obtained without leaving the worksheet by
first viewing the graph and using the
Storm CoIlLsion R*-cmerv Koments lim "Print Screen" keys on the keyboard.
[XnPuts StQ= C01irs-lon- ftcovecy into Parts Exiti Note, however, that the graphical output
produced during a design session is also
saved and is available through the LOTUS
Figure 5. Diagram showing the menu graphics printing program (PrintGraph).
structure. The menu item "Exit" returns Application
the user to the previous menu. The
"Study" item is still in the development Two scenarios are presented to show
stage. the usefulness of the software in
Once inside the worksheet the user is performing parametric studies. The first
allowed to go up and down the menu considers the effect of spring stiffness
structure by selecting the proper menu degradation while the second considers
items. To go from a lower level to the the effects of buoyancy loss.
next highest level the user selects Exit Spring stiffness degradation over time
(E). Only if the user selects Quit (Q) is expected due to cyclic loading and
when in the main menu will the program corrosion. This directly effects the
terminate. Upon completion of a selected hinge moment as well as the overall
operation the program returns to the menu restoring moment. The effects of this
where the operation resides. I The user reduction in the central spring stiffness
then has the choice of performing that on the hinge moment, overall restoring
operation again or going to another moment and dynamic response during a
operation somewhere else in the menu recovery process are easily evaluated by
structure. varying the parameter K (spring
A general description of the program constant) from 100% to 7 @% of its
and its use is contained in the worksheet original value. Figures 6, 7, and 8 show
and can be viewed or printed. To view the output obtained and indicate that the
this information on the screen the user hinge and restoring moments are sensitive
selects Intro (I)/ View (V) . A hard to the parameter change and the pile
copy is obtained by selecting Intro (1) recovery response is not.
Print (P). The outer foam cladding is subjected
The analysis performed by the program to harsh conditions and is considered
is divided into three categories: sacrificial. This effects the buoyancy
Dynamics, Moments, and (parametric) moment and the overall restoring moment,
Studies, and are represented as Dynamics but, again the effect on the recovery
(D), Moments (M), and Study (S), dynamics is marginal. The results are
respectively, in the main menu. The shown in Figures 9, 10, and 11. The
Dynamics (D) item allows the dynamics of static moments are directly affected by
the Collision / Recovery sequences and the loss of foam but the recovery
CTPS under Storm conditions to be response varied little.
investi gated. The Moments (M) item The previous results indicate that the
allows any of the static moments about recovery dynamics are little effected by
the hinge or the spring force to be the individual changes in both spring
evaluated as a function of pile stiffness and buoyancy. A recovery
inclination. The@selection Moments (M) simulation using both a reduced stiffness
/ All (A) evaluates all of the moments (70%) and no outer foam indicates that
and combines the results for each the combined effect results in a recovery
individual moment on one graph. The item time increase of about 4 seconds, or
Study.(S) allows parametric studies to be increase of about 25% (see Fig. 12).

592

HINGE MOMENT VS PILE INCLINATION BUOYANCY MOMENT VS PILE INCLINATION

am -

?a-
'M
do-

do-
so

30-

to-
30-

0a a 20 00
a 20

10-100S + KS - 9= 0 Ks WX ks - MX ?ht - 3 W. + Tht 1.5 W. 0 Tht 0.0 1..

Figure 6. Curves showing the effects of
spring stiffness degradation on the hinge Figure 9. Curves showing the effects of
moment (MH). exterior foam loss on the buoyancy moment
(MB)-

RESTORING MOMENT VS PILE INCLINATION
RESTORING MOMENT VS PILE INCLINATION

23- 40

W
30-
j30 4@ zo
KS - 10019 0") 0
KS - 9= 9 KS - aDX A KS 71119 W so
W - 3 W + Thr 1.3 K Tht 0.0 itt.
Figure 7. Curves showing the effects of
spring stiffness degradation, on the Figure 10. Curves showing the effects
restoring moment (MR). of exterior foam loss on the restoring
PILE INCLINATION VS. TIME moment (MR).
PILE INCLINATION VS. TIME
O@2
70- am @

W - 70-

ao - Go.

30-

20-

a 4 a 12
TIM (MC) _20
KS - 100X t Ks gwg 0 Its - a0X A KS It= Is 2@
Tht - 3 bo. (SM)
5 W 0 Tht 0.0 W
Figure 8. Curves showing the effects of
spring stiffness degradation on pile Figure 11. Curves showing the effects
recovery behavior.. of exterior foam loss on pile recovery
behavior.

593

PILE INCLINATION VS. TIME
00 Swift, M.R. and K.C. BAldwin (1986) "The
Development of a Collision Tolerant
70- Pile Structure Concept", Ocean -
0. Engineering, Vol. 13, No. 2, pp. 131-
156.
Swift, M.R., A. Dudka, R. Villeneuve, J.
Ward and K.C. Baldwin (1986)
30 "Development of a Central Stay
20 Collision Tolerant Pile Structure". a
10 UNH CTPS group technical memorandum
a submitted to the U.S. Coast Guard,
2100 2nd St. SW, Washington, D.C.
02593.
Baldwin, K.C., M.R. Swift and D. Mielke,
-30 (1987), "Quarter Scale Collision
Tal(WI Tolerant Pile Concepts: Peripheral and
KS 70X M W - 0 1. Central Stay", OCEANS 87
Swift,M.R. and K.C. Baldwin (1988).
Figure 12. Curve showing the pile "Collision Tolerant Pile Structure
recovery response with a 30% decrease in Hinge Concepts". ASCE Technical Note,
spring stiffness and total loss of in press.
exterior foam. Ward, J.S., K.C. Baldwin and M.R. Swift
(1988), "The Design and Analysis of a
Full Scale Prototype Collision
Summary Conclusion Tolerant Pile Structure", Submitted to
ASME Ocean Engineering Division, 4th
The CTPS design software presented Annual Symposium on Current Practices
here is a multi- functional package which and New Technology in Ocean
provides the user with a flexible design Engineering.
tool. All of the modeling and analysis Medzorian,J., M.R. Swift, J.S. Ward and
is self-contained in a single worksheet K.C. Baldwin (1988), "Microcomputer
eliminating the need to exit the Modeling of Collision Tolerant Pile
worksheet to perform multiple design Structures Dynamics", Sea Grant report
tasks. All pertinent system parameters number UNH-MP-T/DR-SG-88-3 , -
are allowed to vary enabling the user to University of New Hampshire, Durham,
alter the configuration during the design NH, 03824.
session. The user is also able to view
or obtain hard copies of the output
without terminating the program.

References

Swift, M. R. and Baldwin, K. C. (1985).
"The design and model testing of a
collision tolerant pile structure",
Final report submitted to the U.S.
Coast Guard R&D Center, Avery Point,
Groton, CT.
Cloutier, R., T. Bier, C. Byrne, T.
Sears, and C. Tuttle (1985) "The
Design, Development, and Testing of a
Quarter Scale Collision Tolerant Pile
Structure", Tech 697 Sea Grant Ocean
Projects Course, University of New
Hampshire Mechanical Engineering
Department, Durham, NH.
Durkee, R. A. (1986) "A Mathematical
Model for the Three-Dimensional
Dynamics of a Collision Tolerant Pile
Structure", M.S. Thesis, Mechanical
Engineering, UNH, Durham, NH 03824.

594

UNITED STATES NAVAL EXPERIENCE WITH ANTIFOULING PAINTS

Theodore Dowd

Naval Sea Systems Command
Materials and Assurance Engineering Office - Code 05M1
Washington, DC 20362-5101

ABSTRACT from June 1981 to September 1987. In both cases
the ships did not require any underwater brushin;
Performance of several antifouling tributyltin in the 6 years that the organotin was on the hull.
(TBT) paints on a variety of Naval ship types is Other examples of outstanding long service, where
presented. The current Naval assessment of TBT organotin is still performing, are a submarine
paints is presented in terms of economic, environ- painted in 1982, and an aircraft carrier painted
mental, and health and safety issues. The in 1984. Organotin paints were applied on assorted
alternate ablative paints tested have given mixed combatants and auxiliary ships up to the end of
results, but acceptable economic and environmental 1985, just before the ban went into effect, and
results. Future antifouling paint needs and these ships are continuing to perform well.
directions are discussed.
A complete list of the ships painted with organo-
tin and their length of service is contained in
Table 1.
UNITED STATES' NAVAL EXPERIENCE WITH ADVANCED
ANTIFOULING PAINTS Table 1.
Extended drydocking intervals and foul-free ser- COPOLYMER ORGANOTIN TEST SHIPS
vice of up to 7 years became a reality for the DATE LENGTH OF
marine industry with the introduction of ablative SHIP TYPE PAINTED SERVICE
organotin antifouling paints. These coatings
resulted in reduced propulsion fuel costs up to Coast Guard Cutter Feb 1979 7 Years
16 percent, completely eliminated the expense of Destroyer Nov 1980 6 Years
underwater scrubbing, and halted the ensuing Destroyer Jun 1981 6 Years*
damage suffered by the paint from the mechanical Submarine Jan 1982 Still Active
process which further increased the savings. This Aircraft Carrier Apr 1984 Still Active
is an outstanding achievement because,.at 201knots, Frigate Jul 1984 Still Active
a fouled ship may require 19 percent more horse- Frigate Jan 1985 Still Active
power to maintain its speed. Aircraft Carrier Feb 1985 Still Active
The ablative organotin copolymer paints replaced Supply Ship Mar 1985 Still Active
the old conventional cuprous oxide paints, which Tug Sep 1985 Still Active
are good for only 18 to 24 months, and are no Destroyer Dec 1985 Still Active
longer able to meet the increasing demands of the Frigate Feb 1987 Still Active
commercial marine industry and combat naval forces. Frigate Sep 1987 Still Active
Frigate Sep 1987 Still Active
Ablative copolymer-type organotin paints were first *This ship has an ablative organotin, while all
introduced in Europe in 1974 on commercial ships. the others have the copolymer.
By 1978, the news of its superior performance had
become well-publicized in the United States, and However, this superior antifouling coating has come
led to the application, in February 1979, on a under fire from environmentalists and labor unions
U. S. Coast Guard cutter. This ship performed as being dangerous to man and marine life. In
foul-free for the next 7 years. At the interim 1986, the U. S. Congress imposed a ban on the use
dockings in 1980, 1983, the ship was water washed of organotin for Navy ships only. This is unfor-
to remove the slimej mechanically damaged areas tunate because there are no alternatives that match
were touched up, and the ship was returned to the the performance of organotin paints. Organotin
water. The coating continued to perform well until paints are legally used worldwide on 70 percent to
1986. go percent of the major commercial ocean-going
In November 1980, a U. S. Navy destroyer was ships-including the naval forces of many countries.
painted with organotin and that ship performed
foul-free until June 1986; another destroyer went

595 United States Government work not protected by copyright

The Navy Environmental Assessment report claimed TABLE 2.
that organotin can be safely used without danger
to marine life if the water quality standard did ABLATIVE COPPER TEST SHIPS
not exceed 50 parts per trillion of tributyltin. SOURCE #1
However, the Environmental Protection Agency (EPA)
recently recommended a 10 parts per trillion con- DATE LENGTH OF
centration in the water. This is going to be very SHIP TYPE PAINTED SERVICE RESULTS
difficult, if not impossible, to meet, due to the
existing background levels in busy harbors or re- Frigate Aug 1983 48 Months 15% Fouling
pair areas, and because drydock discharges could Ordnance Apr 1984 36 Months 70% Fouling
be more than one order of magnitude higher. Supply Apr 1984 35 Months 1% Fouling
Carrier May 1984 46 Months 0% Fouling
The other impediment to organotin implementation Frigate May 1984 40 Months 5% Fouling
is the cost of application in naval facilities supply Jan 1985 Still Active 0% Fouling
'Where increased concerns have escalated recent Carrier Mar 1985 12 Months 100% Delami-
costs. Costs at navy facilities can run over nation
$100 per square foot to remove and reapply organo- Jul 1986, paint removed/reapplied
tin, while a commercial shipyard may charge $15 per Carrier Mar 1985 24 Months 0% Fouling
square foot for the same job. The main cost dif- May 1987, docked and repainted
ference could be the result of different interpre- Auxiliary May 1985 Still Active
tations of environmental and health requirements. Destroyer Jul 1985 Still Active
Assault Aug 1985 Still Active
As early as 1982, the U. S. Navy had reservat ions Cruiser Sep 1985 Still Active
on the ultimate use of organotin paints, based on Destroyer Oct 1985 Still Active
the lack of environmental regulatory inputs. This Supply Oct 1985 Still Active
apprehension led to Navy testing of ablative Carrier May 1986 Still Active
copper paints as an alternative, with the first Carrier Jul 1986 (Repair of 3/85 failure,
ship being painted in 1983. This was the first Still Active)
ship test of this paint in the world. Destroyer Sep 1986 Still Active
Oiler Nov 1986 Still Active
In 1987, the Navy resorted to the fleetwide imple-
mentation of the ablative copper antifouling ABLATIVE COPPER TEST SHIPS
,system because organotin paints were no longer a SOURCE #2
viable option. This paint is similar to Formula
121, but employs self-cleaning properties to renew Assault Mar 1986 Still Active
its surface. The system was originally advertised Oiler Apr 1986 Still Active
by the manufacturer to give a minimum of 5 years Carrier Sep 1986 Still Active
service and be the alternative to organotin Carrier Nov 1986 Still Active
copolymer paints.

The history of that experience includes applica- It is too early to report any long-range results
tions on submarines, and combatant and noncombat- of test ships painted in 1985 to 1986, but in one
ant surface ships from 1983 to the present. In case, the paint became powdery after only 13 months
August 1983, ablative copper was first applied to service. This paint had to be removed and
a frigate. That ship was inspected 48 months replaced, and had it not been discovered, a mas-
later, in September 1987, and found to have about sive paint failure and heavy fouling was
15.percent fouling. An ammunition ship painted in anticipated.
April 1984 was found to have 70 percent heavy
fouling after only 36 months service. During At best, the present version of ablative copper is
1984, ablative copper was also applied to an improvement over F121, but no match for the
2 frigates, I aircraft carrier, and an auxiliary superior 7-year performance of organotins. One of
supply ship; they are performing well, with little the Navy suppliers has reformulated his ablative
or no fouling, after 4 years. This illustrates copper product line in an attempt to approach the
that the ablative copper system is unreliabile and performance of organotin. This new material has
not the equal alternative that the U. S. Navy is been applied to a combatant surface ship, and its
seeking in place of organotin paints. performance will be monitored and compared with
that of the existing systems serving as controls
The complete list of ships coated with ablative on the same ship.
copper, from 1983 until 1986, is shown in Table 2.
In addition, the quest for radically new and
superior antifouling paints continues. There is
a program to develop and test fouling-resistant
paints which do not use any toxicants, but rely on
the structure of the polymer to repel the attach-
ment of marine organisms. Along this line, a new
rubber-based material that has no toxicant, but
nevertheless is able to repel the attachment of
marine fouling is being.tested on ships following

596

2 years of laboratory testing. The material can be
used on both steel and aluminum hulls, as well as
rubber surfaces. It also has the potential for use
in areas that require resistance to heat and ultra-
violet rays.

This is all a part of the continual investigation
for new and improved antifouling systems that can
best serve the maritime industry without endanger-
ing the marine environment or mankind. The U. S.
Navy places its reliance on the private sector to
develop a long-lived antifouling paint system.

597

TAPERED INTERFACE IN HARSH ENVIROWENT CONNECTORS

ALAN BERTAUX

JUPITER ELECTRONICS INC.

- Reduced visibility;
The mission of standard commercial, or - Temperature extremes;
even military electronic connectors is - Radiation.
typically to join 2 or more electrical
conductors with a minimum of Yet, a connector operating in such
interference to current flow, within environments should, in addition to
the minimum amount of space and cost, meeting above constraints:
maximum reliability and ease of - Present as little signal loss as
operation for the task at hand. possible;
- Be capable of handling power,
Connectors made to operate in high signal and coax circuits;
fluid pressures or corrosive environ-,
ments face additional requirements - Allow for fast and zero-force
which translate into "trade offs" mating and unmating (since the environ-
(price vs. reliability, size vs. ease ment is often dangerous to human
of operation,etc.). beings):
Have unmistakable mating, in spite
A conical interface represents a of poor visibility or uncomfortable
significant breakthrough in terms of situations;
performance, cost, and ease of - Have positive and secure connec-
operation. tion, to resist shocks, vibrations and
rough handling;
- Offer a wide range of shell and
insert combinations, including high
contact densities;
Harsh environments, such as found in - Be field installable, without
the oil, nuclear, mining, chemical or requiring cumbersome neoprene molding
naval industries dictate demanding con- equipment.
nector specifications not found in the - Have a knurled and circular shape
office environment. Traditionally, to enable easy handling even with
designers have used military cylindri- gloves;
cal connectors to meet the demand of - Be cost effective.
these industries. Unfortunately, such
connectors have inherent design and All above design parameters can be
performance limitations, particularly satisfied,, in part, by the choice of
in the area of mating Interface. The connector raw materials. Housings can
Purpose of this paper is to demonstrate be made of stainless steel, marine
the advantages of conical (or tapered) bronze or nickel plated brass; inserts
interfaces as applied to harsh environ- can be made of neoprene, Nylatron,
ment applications. Tefzel, or vespel; and contacts can be
gold plated.
A harsh environment typically includes
one or more of the following con- In addition to offering various raw
straints: material or plating options, military
- High humidity; connectors have long offered keying,
- Corrosive liquids or atmosphere, polarization and knurled cylindrical
including salt water; bodies to enable easy and fool-proof
- Dirt and dust; mating. Moreover, some of the newer
- Fluid or gas pressure; military connector developments also
- Severe shocks or vibrations, Provide for good resistance to shock
including human abuse; and vibrations (example: MIL-C-38999
Series III).

CH2585-8/88/0000-598 $1 @1988 IEEE

However, even with military-class In addition, even the best 0 ring mate-
circular connectors, there is a limit rials can dry out or harden because of
as to how Much humidity or water age, chemicals, or cold temperature.
pressure can be tolerated. in fact, by They then become brittle and fail.
design, there are some inherent
mechanical conflicts related to water- Another compromise is the use of a
tightness in standard military circular molded-on "rubber boot" wherever high
connectors. Tightening the tolerances fluid pressures are present. The
that exist between two mating cylindri- cables are prepared and terminated in a
cal connector surfaces results in shop, and a rubber-like material (often
better waterproofinq. However, the neoprene) is molded over the connec-
tighter these tolerances are, the more tion.
difficult it is to mate and unmate the
connector halves. Both the metal Unfortunately, there are several draw-
friction and the air pocket in the backs to this approach besides high
interface area must be overcome cost. Field repair is generally impos-
(Figure 1). sible. Also, if the connector is
damaged, it has to be cut from the
rip. I cable. If not enough cable remains to
join the connector halves that have
been substituted for the severed
connector, an entire harness has to be
replaced. Furthermore, a molded-on
boot only guarantees hermeticity if the
plull joined cables remain straight. Past a
given cable curvature, fluids may be
allowed to seep through the open sDace
created between the cable and the
molded boot. In short, standard
Hence, the tighter the hermeticity, the circular military connectors were
harder it is for the air to be squeezed designed for specific purposes that may
out. Conversely, the unmating, of not meet some harsh commercial
tightly mated cylindrical connectors environments. A connector with conical
requires overcoming the vacuu-n left in mating surfaces may answer such design
the interface area. Hence, the first parameters. The conical mating
conflict inherent in military cylindri- surfaces allow quick air motion during
cal connectors is hermeticity vs. ease mating and unmating.
and speed of operation. ho. 2

Another conflict exists with high fluid
pressures. Cylindrical mating surfaces
necessarily require tolerances, even if
only 0.0011, or less. A zero tolerance
in the contact area would ensure abso-
lute watertightness but would render
mating impossible. Hence, the second
conflict is herimeticity vs. basic rules
of mechanical design.
When two 'cones engage, air pressure or
vacuum does not exist to hamper the
Closer tolerances also mean higher connection. in addition, two conical
manufacturing costs, fragility to mis- surfaces can be allowed to "bottom,"
handling, and susceptibility to the metal-to-metal, thus ensuring complete
presence of foreign particles in the watertightness while avoiding problems
mating area. These conditions create a related to close tolerances of cylin-
third conflict, hermeticity vs. cost drical mating surfaces. To avoid the
and practicality. cost of high precision conical sur-
faces, two 0 rings can be used:
As a compromise, 11011 rings can be F19.3
placed in the mating area even though
they are not used in standard circular
military connectors. 0 rings, however,
do not offer the mechanical resistance
of metal. Under high fluid pressures,
0 rings can actually get squeezed
within the tolerance gap existing
between the mated cylinders and be
flushed out of the mating area.

599

However, these 0 rings no longer serve Furthermore, a connector with conical
as the ultimate line of defense against mating surfaces does not depend on
fluid pressure, chemicals, dust, or backshell accessories, like a rubber
temperature extremes. in fact, with boot, for hermeticity. The connector
this design, the greater the outside can be assembled from standard,
pressure, the tighter the conical modular components and becomes fully
surfaces are jammed against each other, field repairable without specialized
thus relieving the 0 rings from their molding equipment. Repairs can be
hermetic role and making it impossible performed without the disposal of the
for them to be "flushed" out under entire connector or harness.
pressure. Such connectors have been
tested under pressure of up to 22,000 The contact size, density and plating
psi when mated, and 1,500 psi unmated. can be chosen to min.1mize bulk and
voltage drop, while guaranteeing pro-
The unmated hermeticity is enhanced by longed usage in the most severe envi-
the contact design, which also takes ronments. A connector can also be
advantage of conical interfaces: provided with a double internal seal
rig. 4 Directl for extreme fluid pressures and to
ofposalble allow sheath capture for perfect
grounding and electromagnetic inter-
ferencelradio frequency interference
(EM11RF1) shielding:

Potting can be applied in the
transition area existing inside the
connector at the level of contact-to-
wire junction (See floure 5).
Fq,
7@777777777

Military cylindrical connectors serve
military purposes well. However, a
00U.'al civil engineer faced with severe envi-
.1 wh romental constraints should select
comerpial connectors that meet his
However, the potting serves more as a specific design requirements rather
strain relief in this case, than as a than adopting existing military pro-
hermeticity barrier. The potting ducts. A connector with conical or
allows for mechanical tension to be ap- tapered interface is more like ,ly to
plied to the cable without risk of wire answer those needs, than a standard,
rupture inside the connector. cylindrical connector.

Figure 7 below shows the whole concept.

Concept (Fig. 7)

high precision flarnele
DU81 0 ring ffmdtilned receptacle
PrIotection Insert male
plog Grounding Blip boot
clip Ino over-molding required)

Conical Optional
Intertbee cornpou; seal feature
lCone slope exaggerated for Conical (tapered) interfacing is a Jupiter patent.
concept visualization)

600

BIOGRAPHY

Alan Bertaux has been president of
Jupiter Electronics, Inc., the U.S.
subsidiary of C.E. Jupiter S.A. of
France, since its creation in the
Chicago area in 1982. Prior to that
assignment, he was director of Latin
American operations for Molex Inc.
where he worked for 5 years. He also
worked for AMP Inc. where he spent 6
years.

Mr. Bertaux holds a degree in
mechanical engineering from the
University of Paris, and a B.A. and
M.A. from the University of Maryland.
Mr. Bertaux also taught at Goucher
College and Tufts University, and spent
3 years in the U.S. Army. In addition
to ECSG, he is a member of Phi Kappa
Phi, SME, SMTA, ASQC, IEPS and ISM.

601

NADIA: WIRELINE RE-ENTRY IN DEEP SEA BOREHOLES

Jacques LEGRAND, Andr6 ECHARDOUR, Luc FLOURY, Henri FLOCH,
Jacques KERDONCUFF, Tanguy LE MOIGN, G6rard LOAEC, Yves RAER

IFREMER, Centre de BREST - Direction de I'Ingenierie et de la Technoloy*e
D6partement Instrumentation, Capteurs et Acoustique - BP 70 - 29263 PLOU ANE
FRANCE

ABSTRACT The first trials at sea were completed successfully in
NADIA is a system clevelopped to complete re-entry ope- 2300 m of water in the Mediterranean Sea. Operations in
rations in deep sea boreholes, drilled for twenty years by DSDP Hole 396 B will be conducted in July and August
the drilling vessels Glomar Challenger and JOIDES Resolu- 1.988.
tion.
NADIA is operated by a manned submersible at water
depth of up to 6000 m, with a capability of lowering PHASE 2 - HOVERING ABOVE THE CONE
instruments in holes of 1000 m of penetration. & I
This paper describes the system and re-entry operations, RELEASE OF THE SECONDARY
with some technical details about specific equipments J. FLOTATION ASSEMBLY
developped within the project NADIA.
The work to be cqmpleted during the upcoming cruise
FARE is described. Then, a programme plan of develop-
ment, following FARE is briefly exposed.

INTRODUCTION
Deep sea boreholes have been drilled bythe D/VGlomar
Challenger of the Deep Sea Drilling Project and then by
the D/V Joides Resolution operated by Texas A & M Uni-
versity for the Ocean Drilling Programme, since 1968, in
allmost every parts of the world Oceans. Figure I - NADIA Configuration in the docking phase
To day, nearly 800 boreholes exist and about 40 of them
are so called re-entry holes, i.e. fitted with a cone at the EQUIPMENTS DESCRIPTION
seafloor and casings in the unconsolidated section of the NADIA is a non-propelled, free falling device. Descent to
geologic formation. the sea floor and return to the surface are achieved by
The only mean to re-occupy these boreholes, either to gravity and buoyancy. The horizontal displacement bet-
replace the drill bit during drilling operations or to come ween the landing point and the cone location is achieved
back to deepen a hole previously drilled, has been the drill by the submersible.
string, hanged below the drilling ship.
The concept of wireline re-entry, i.e. re-occupation of a NADIA is composed of four sub-systems:
drill site in order to lower instrumentation in the borehole - A main frame, built of welded aluminium alloy tubes,
using a conventionnal reseach vessel, was born within the fitted with a winch and its hydraulic control system, the
DSDP community about 8 years ago, but was never com- logging tool, the electro-hydraulic connector, a 10 m
pleted despite several attempts. umbilical link to the submersible, and ancilliary equip-
This paper describes an original solution, developed in ments: (mechanical releases, dead weights, cable cutter,
FRANCE by IFREMER. etc.)
The approach is the use of a special frame NADIA - A main flotation assembly supporting the weight of
(NAvette cle DIAgraphie - logging schuttle) fitted with a the main frame. It is fitted with an accoustic navigation
logging winch and docked into the re-entry cone by a beacon for tracking NADIA during the vertical trips and
manned submersible. rendez-vous with the submersible.

CH2585-8/88/0000- 602 si @1988 IEEE

- A secondary flotation assembly released to ballast 4 Sensors
the main frame and to dock it down in the re-entry cone. Lowering and rising of the instruments in the borehole are
- A descent dead weight, released by the submersible controled by two parameters: cable length and cable
before the horizontal displacement at the sea floor. tension. The cable length is measured at the top pulley,
fitted with 3 Reed switches. A permanent magnet, fixed to
A more detailed technical description is given of the fol- the sheave, actuates successively the switches, giving
lowing sub-systems: pulses directly related to the length of wire payed in or
I -the winch, out. The sign of the wire motion is detected bythe chrono-
2 - the electro-hydraulic connector, logy of the 3 pulses.
3 - the NADIA/submersible umbilical security release, With 1000 N of tension on the cable, no slipping of the
4 - the sensors, wire with respect to the pulley has been measured.
5 - the control and data acquisition system. This simple device gives a precision of � 0.33 m.
I Winch The cable tension is measured with a strain gages tensio-
meter monted between NADIA's frame and the top pul-
In the present stage of development, NADIA is fitted with ley. This tensiometer is pressure compensated up to
1000m of 4.6 mm steel wire. Spooling on the drum is 6000 m of water.
completed in 12 layers, controlled by a diamond screw
level wind. Precision is 0.1 % in the range of 0 to 50 000 N.
A slow POCLAIN hydraulic motor, modified to work in
pressure equilibrium, drives the drum under control of a C"a TENSION
closed-loop hydraulic circuit. Primary energy is provided SEM"
MEMO- C"Lf LMGTH
bythe submersible hydraulic purripworking ata maximum WD"UUC MEA"Ih* PULLEY
CONNECTOR WCKWCAL
pressure of 14 MPa and a maximum discharge flow of AXTUATOR
6.2 Ilmn. "BILICAL CABLE CWTER
Two different rotating speeds are obtained by changing WMC+l LOGMING TOOL
the cubic capacity of the motor, resulting in linear speeds Om WrIcAiT
of the wire from 225 to 450 m/H.
2 Electro-hydraulic connector
Hydraulic and electric links are needed between NADIA
and the submersible. This function is achieved, at the sea
floor, by an underwater matable-unmatable connector.
This device uses an hydraulic locking piston to plug the Figure 2 - NADIA General arrangement
male part of the connector on NADIA to the female part
moved into the correct position by the arm of the submer-
sible. The disaccoupling is obtained by springs when the
piston pressure is released. 5, Control and data acquisition
Three hydraulic links and sixteen electrical conductors can The different functions (electro-hydraulic connectorwinch
be connected under 6000 m of water. operations and data acquisition) are controled by the
The hydraulic connection is realised by self-obturating operator in the submersible.
front water-tight coupling devices while the electrical A micro-computer Epson PX8 is used to perform these
continuity is realised by standard Electra Oceanic connec- tasks. The commands are:
tors. - connect, disconnect,
- speed selection,
3 Security release - pay the cable in or out.
In order to allow the physical separation between NADIA Data from the cablemeter and the tensiometer are dis-
and the submersible in case of problems with the electro- played on the screen for real time control of the operation
hydraulic connector, the interface between the submersi- and stored in the RAM for playback with data coming from
ble and the umbilical link with NADIA isa security release. the logging instruments and plot.
This release is actuated by an electrical pulse causing the A joystick and a separate display are available to ease.the
ignition of explosive bolts. The,forces needed for the dis- control operations inside the exiguous sphere of the
connection are given by compressed springs. submersible.

603

OPERATIONS
A logging operation with NADIA proceeds in the follo-
wing steps:
1 - Launching of the frame and the flotation assembly 7 - Raising the logging tool back in NADIA'sframe. This
(system buoyant 50 to 100 N). Lauching of the descent operation is similar to phase 6. If the tool get blocked in
weight when ship is in a position that gives the closest the hole, and if it cannot be worked free by operating the
impact of the system from the cone (system negatively winch back and forth, two levels of security are available:
buoyant -900 N). a - a shear pin rated 6000 N provides a weak point at the
2 - Landing at the sea bottom. Length between descent cable head. A fiching neck at the tool upper end will
weight and frame is chosen to prevent NADIA from over- permit the use of the fiching overshot in use on board the
shooting or hitting thelsea floor. drilling vessel Joides Resolution,
3 - The submersible dives. Rendez-vous with NADIA with b - an hydrostatic cable cutter is f itted in NADIA's frame.
the help of the accoustic navigation system and a flash In case the shear pin cannot be used, the cable can be cut
light. The submersible holds NADIA with one arm, adjusts to set the system free from the cone.
her ballast to become neutrally buoyant with NADIA and 8 - At the end of operations, the logging tool is raised
releases the descent weight. The submersible moves back in theframe. The submersible disconnectsthe electro-
NADIA towards the cone using its propulsion system. hydraulic link and replaces it in the basket.
4 - Hovering of NADIA above the cone - the secondary 9 - The submersible holds NADIA, lifts it up out of the
flotation assembly is released and NADIA isset in the cone cone and moves a few meters off. The ascent weight is
with -600 N negative buoyancy. released. NADIA is set free and pops up to the surface.
5 - The submersible connects with the electro-hydraulic 10 - The system is recovered on board the ship.
connector to NADIA.

3114,j MP@@

Q 1T,

J,
@e
W
Ig
NO' M,

A
"J,
W

T,-!
'W@
01

'4"

;2-
ii@TT,
Figure 3 - NADIA on the fan toil ready for launching

6 - Lowering the logging tool. Winch control and data
acquisition in the submersible sphere. If the tool is blo-
cked, the wire tension decreases and the operator stops Figure 4 - NAUTILE: 3 men, 6000 m copobity submersible
the winch.

604

FARE: TESTS IN HOLE DSDP 396 B Appendix
The first attempt to re-enter a deep sea borehole will take NADIA - GENERAL CHARACTERISTICS
place in July and August 1988 on board N/O NADIR with
the deep submersible NAUTILE. FRAME
The target hole is Hole 396 B, drilled by D/V Glomar overall height ............................................. 3,5 M
Challenger in February 1976. It's position is 22059'N and overall diameter ......................................... 4,0 m
43031'W, about 90 nautical miles east of the Mid-Atlantic weight in air ............................................. 1300 kg
Ridge. weight in water (descent) .......................... 9 000 N
This hole iscased onthefirst 163 m, on the whole sediment maximum operating depth ........................ 6 000 m
section. Open hole extends to 405 m below the sea floor.
About two thirds of the open hole were drilled in pillow FLOTATION ASSEMBLY:
basalts, the bottom third was mainly basaltic gravel and 2 assemblies of syntactic foa@n packs
sand. height ....................................................... 1.6 m
Although the cruise FARE (FAiseability RE-entry) is a tech- diameter ................................................... 1.3 m
nical test, the opportunity will be taken to collect scientific weight in air ............................................ 2 000 kg
valuable data. After relocation and inspection of the re- buoyancy ................................................ 7 700 N
entry cone, operations to be conducted in the hole are the
following: WINCH
I - Sampling of fluids as deep as possible, maximum cable length .............................. 1 500 m
2 - Temperature continuous logging, cable diameter ........................................... 4.6 mm
3 - Taking pictures from the walls, pulling capacity
outerlayer ..................................... 8 300 N
4 - Test with a 200 mm dummy tool. on drum ...................................... 16 600 N
cable speed
PROJECT - PHASE 2 outer layer ................................ 0. 15/0.30 m/s
The cruise FARE representes the milestone marking the 540/1080 m/h
end of the first phase of the project NADIA. If these tests on drum ................................... 0.07/0.15 m/s
are successfull, a phase 2 will be initiated with the main 270/540 m/h
objective of transforming NADIA in a real logging unit.
The evolution of the system will involve the following LOGGING TOOLS
transformations: diameter ...................................... 100/ 150/200 m m
I - Replacement of the cable by a 7 conductors logging length .......................................................... 2 m
cable; this implicates the replacement of the winch drum weight in water ....................................... 1 000 N
to support a greater volume and to provide for the electri-
cal continuity with the logging tools; the cable length TENSION SENSOR
measurement device will also be replaced by a sensor range ..... .0 - 50 000 N
similar to those used in standard logging operations with
a special packaging for high pressure environment. sensitivity ................................................... 100 N
2 - An autonomous data acquisition system, as compatible
as possible-to the similar units used in oil well logging will CABLE LENGTH SENSOR
be added on NADIA'sframe; an electrical power unitwill increment .................................................. 0,1 m
also be added.
These transformations will give NADIA the capability of ANCILLIARY EQUIPMENTS
recording clownhole data for periods of up to a few hydrostatic cable cutter (minimum operating pres-
month; the re-entry operations and logging tools lowe- sure 100 bars),
ring will still be completed, in this phase, by a manned mechanical actuators for:
submersible. cable cutter operation,
release descent and ascent dead weights,
A final phase 3, in the future, will be to perform the entire electro hydraulic NADIA/NAUTILE connector,
suite of operations using a remote operated vehicle. hydraulic winch control.

605

IMPEDANCE MEASUREMENTS OF BIOFOULING IN SEAWATER CONDENSERS: AN UPDATE

Patrick K. Sullivan 1,2 and Bruce E. Liebert3

10ceanit Laboratories, Inc., Honolulu, Hawaii
2Department of Ocean Engineering, University of Hawaii
3Department of Mechanical Engineering, University of Hawaii

Electrode Impedance Spectroscopy Hawaiian Electric Company (HECO),
(EIS) is a new technology presently the host utility, is interested in
being investigated for its EIS technology to improve power
usefulness in measuring biofouling plant efficiency. To assist in the
in seawater cooled power plants. first industrial testing of EIS
EIS is based on the hypothesis that technology, HECO has provided space
biofouling can be characterized by within their Honolulu power plant
observing its steady state response facility as well as technical
to a small amplitude sinusoidal support.
perturbing potential imposed on the
biofilm material. From the change our research is leading to the
in phase and amplitude of the development of a diagnostic
response signal a real-time engineering tool that would enhance
characterization of the biofilm can the cost effective operation of
be determined. power plants; enhance understanding
of biofilm accumulation and factors
EIS was developed at the University that effect its rate; and allow
of Hawaii in 1985 for use in Ocean plant operators to assess the
Thermal Energy Conversion (OTEC) effectiveness of cleaning
materials testing research. techniques. Additionally, it would
However, it has wide reaching improve methods for designing
applications in the power industry condenser systems; testing and
for measurements of corrosion as selection of materials; and
well as biofouling. The Electric evaluating complex surface
Power Research Institute (EPRI), in geometries. Results from the
conjunction with the Hawaiian ongoing research at the HECO plant
Electric Company (HECO) sponsored will be presented and discussed.
Oceanit Laboratories, Inc. (OLI) to
conduct research to determine the
usefulness of EIS as an insitu tool
for monitoring biofouling in power
plant condensers. Results from
this research will largely
determine its use as a tool for the
power industry.

WORLD'S FIRST RIGID FREE-STANDING PRODUCTION RISER

E. A. FISHER
H. P. HACKETT

CAMERON OFFSHORE ENGINEERING, INC.
580 WESTLAKE PARK BLVD., SUITE 1650
HOUSTON, TEXAS 77079

ABSTRACT FLOATING PRODUCTION SYSTEM

The first deep water (15501) floating In reviewing characteristics of the
production system in the Gulf of Mexico Green Canyon Block 29 discovery, and in
has been installed in Green Canyon Block 29 reviewing the potential technology avail-
for the Placid Oil Company. The key compo- able for the production system, a set of
nent of this system, and the first of its criteria evolved for the economic develop-
kind in the world, is the rigid free-stand- ment of this field.
ing production riser which serves as the 0 System to accommodate simulta-
conduit for the well fluids up to the neous drilling/workover and
floating production facility and for the production;
processed oil and gas back down to the 0 System to be capable of accommo-
pipelines on the seafloor. dating 24 wells;
This paper reviews the overall float- 0 System to accommodate surface
ing production system and focuses on the well flow control devices,
riser and its components: the bottom (chokes) and have individual
connector, a one-piece tapered titanium tubing and annulus lines for each
stress joint, the 601 long steel riser well;
joints wrapped in syntactic foam, the upper 0 System to be capable of gas
manifold, and the flexible hoses and lifting any well by means of
tension system. surface controls;
0 System to be able to accommodate
production from or injection to
any satellite well;
INTRODUCTION 0 System to have the flexibility to
work in the water depth range of
The Placid Oil Green Canyon 29 Project 1,5001 to 3,5001;
is a pioneering effort in many ways: 0 The subsea control system compo-
� it is the first floating produc- nents must be accessible for
tion system in the Gulf of Mexico maintenance;
� it is the first floating produc- 0 The floating platform will be
tion system designed and certi- equipped with "bare bones" pro-
fied for a 20 year life with the duction facilities --- a shallow
ability to withstand a 100 year water production platform site
storm will be used for final processing
� it has the deepest oil well of oil and gas;
completion in the world at 22431 0 Pipelines will be installed to
of water depth, exceeding the transport the field's oil and gas
previous record of 16131 production to a sales point.
� it has the deepest opertional A variety of sytems and designs were
pipelines in the Gulf of Mexico reviewed, analyzed, and costed out; and the
� it has the world's first rigid optimal design was selected (Figure 1).
free-standing production riser
� it will be profitable even at
today's reduced oil prices.

CH2585-8/88/0000- 607 si @1988 IEEE

FIGURE 1. PLACID OIL GREEN CANYON BLOCK 29 FLOATING PRODUCTION SYSTEM

The floating production system
consists of:
0 a floating semi-submersible spin resistant spiral strand
.converted to allow simultaneous jacketed 4-1/211 wire, a submerged
drilling and production. The buoy with 130 kips net buoyancy,
particular vessel selected was approximately 1500, of spin
the Penrod 72. Rig modifications resistant spiral strand jacketed
included the addition of sponsons 4-1/2" wire, another submerged
to "house" the new mooring equip- buoy with 50 kips net buoyancy,
ment and to increase the payload, approximately 12001 of spin
the relocation of the helideck to resistant multi-strand 5 wire,
provide a clear area for the and the bending shoe and linear
processing equipment, the winch system in conjunction with
"opening up" of various areas to large capacity spooling winches.
facilitate inspection, the addi- o A 24 well slot template located
tion of vent booms, subsea on the ocean floor directly
controls, and ROV support, the beneath the rig. The template is
modification from a standard constructed from standard struc-
drive to a top drive drilling tural steel shapes to accommodate
system, and the addition of the the design depth requirement, and
process equipment itself. The has four rows of six slots. The
major purpose of the process two outer rows of slots can be
equipment is to split the well used for connections of remote
fluids into the liquid phase and satellite wells or for wells
gas phase, where the liquids will drilled directly through the
flow through one pipeline and the template. The template measures
gas through another to a shallow out at 1651 long, 821 wide, and
water platform for further pro- 201 tall. It weighs approximate-
cessing. The process-equipment ly 1250 tons and is supported by
is designed to handle 40,000 eight 40" piles.
barrels per day of oil and 120 0 Up to 12 remote satellite wells.
million cubic feet per day of The wells would be drilled by a
gas. different semi-submersible and
0 An eight-point mooring system for would typically be 1coated 5 to
the Penrod 72. Each mooring line 10 miles from the template. The
consists of a 77-kip Bruce Flat flow from these wells would be
Fluke Twin Shank anchor and a 7- routed to the template through
kip chain depressor, approxi- pipeline on the ocean floor.
mately 25001 of 4-1/8" ORQ (oil 0 A shallow water processing plat-
rig quality) plus 20% (20% higher form. Although the well fluids
break strength than ORQ) chain, are separated into the liquid and
10001 of 4-1/4" ORQ plus 20% gas phases on the rig, additional
chain, approximately 1250' of processing is required. This is

608

performed on a shallow water
platform approximately 60 miles
from the rig/template, with
separate pipelines on 'the ocean
floor routing the liquid and gas
from the template to the plat-
form. The processed oil and gas
are then piped to shore. TENSIONE9
0 The rigid free-standing produc-
tion riser. The basic function
of the riser is to route the well CA:LE SUPPLY
fluids up from the template to POOL
the floating production system
and the separate liquid and gas .-A
back down from the rig to the TURNDOWN
template and ocean floor based SHEAVES
pipelines.
WATER LINE
RIGID RISER

Cameron offshore Engineering, Inc.
(Houston, Texas), after working in conjunc- F@'
tion with Placid in developing the overall CONNECTION
f loating production system concept, PORCH PONTOON
designed and supplied the rigid riser
system. The rigid free-standing production CENTRALIZER
riser described by Placid as the most
innovative component of the overall system FLEXIBLE TENSIONER
and a key to its economics, is fully PIPE CABLE
installed and free-standing beneath the
Penrod 72 (Figure 2). TENSIOW SHEAVE
The riser is the critical link between PACKAGE
the floating rig on the surface and the
template on the ocean floor, and holds the CUPPER RISER
ONNECTOR PACKAGE
production and annulus lines from each
well, the oil and gas sales lines, and the
control line bundles. From the bottom up,
the riser has: a structural riser base
welded into the template, to transfer the UPPER RISER FLOTATION
riser loads through the template into the AIR TANKS
piles and soil; a collet connector, to
provide a strong reliable attachment to the
riser base and to provide a connect point
for running the riser and a disconnect
point for retrieving the riser; a one piece
titanium stress joint, to provide the
necessary flexibility; the riser joints
with internal air chambers and external
foam modules, to provide the necessary lift
to be self-buoyant and to act as support
for the production, annulus, and sales
lines; the upper_,,,air tanks, to provide
additional lift; the upper riser connector
package, to provide a single point where
the entire array of flow line connections STRESS JOINT ASSEMBLY
could be separated to quickly release the LOWER CONNECTOR
riser, or where each flowline could be
connected/disconnected separately for RISER BASE
maintenance; the flexible hoses, to allow 1EMPUIE
the flow up to the rig and back down to the OCEAN FLOOR
riser while simultaneously serving to
decouple the relative motions of the rig
and riser; and the specially designed
tensioner system, where the tensioner,is
not used to hold the riser up, but rather FIGURE 2. RIGID RISER
,is used to maintain the proper relative
position between the rig and riser.
'.IN' AS -MBLY

CONCOR

IS G
. LREWR.TA ES

609

- Lower Riser Package
The lower riser package serves several
S
into the template, to connect the riser and
flowlines to the template, and to provide
the necessary flexure between the fixed
functions - to transfer the riser load

template and the relatively mobile riser.
The riser base is welded into the
template, and serves to transfer the riser L@er riser connector
loads into the template, and to act as a
fixed connection point for the riser and
flowlines.
The lower c6llet connector (Figure 3) Hydraulic cylinder (8)
is flanged to the bottom of the stress
joint and lands on the connector hub which
is welded into the riser base. This
connector hydraulically locks the riser to
the template, thereby, transmitting the
riser loads into the riser base. Special Collet segrnents
high strength steel with excellent fracture
toughness and fatigue properties was Locked Unlocked
selected for these components to ensure the
20-year system design life was achieved.
The stress joint assembly, which is -Riser base hub
located Immediately beneath the lowest
riser joint, provides the flexibility
needed for the relatively abrupt bend in
the riser near its foundation. The stress
joint (Figure 4) provides a smooth dis- Connector body stab
placement curve for the production riser,
producing manageable static loads and a
lengthened fatigue life for both the stress
joint and its included production, annulus,
and export lines. Titanium was selected
over steel for the stress joint due to its
substantially superior properties, namely
twice the flexibility, half the weight,
higher strength and toughness, and excep- %.,- ENGINEERWO, INC
tional corrosion resistance and fatigue
life in a seawater environment. The FIGURE 3. LOWER COLLET CONNECTOR
titanium stress joint allows the riser to
stay connected to the rig even in a 100-
year storm.

The flowlines are run after the riser
has been connected.
The production and
annulus tubing anchors are stabbed into the
mating receptacle located in the riser
base, and are
locked In through weight
setting of an anchoring latch containing
elastomeric seals like those used to set
tubing in downhole packers. The tubing can
be retrieved using either right-hand rota-
tion or overpull. The export tubes are
locked Into their riser base receptacles
using tieback connectors containing elasto-
Z"
meric lip-type seals. The tieback connec-
tors are similar to mechanical connectors
used to tieback casing strings in subsea
wellheads, and are locked and unlocked
using a setting tool which is run on drill
pipe through the export tube.

- Intermediate Riser Package
The riser consist of 26 50-foot joints
.of flanged and cathodically protected high
strength steel pipe. Each joint has an
open bottom internal air tank with a cas-
cade tube welded inside to air the joint L
above it. These joints are encased in FIGURE 4. TITANIUM STRESS JOINT

610

syntactic foam modules that provide both riser and the tension lines. The tension
passive buoyancy and path ways molded in the lines remains essentially constant.
within them to guide and support the pro- The pontoon porches added to the rig
duction and annulus lines along with the are made of structural steel members suit-
oil sales line. An external air valve Is able for transferring the vertical moment
also provided on each joint for an ROV to loads from the dynamic motion of the flex-
increase or decrease the amount of air In ible hoses into the pontoon structural
each joint as desired. frames. The flexible pipe itself runs from
Vortex shedding strakes are attached the porches to the top of the riser located
to the outside of the foam modules. The 1501 below the water surface, and serves as
riser is instrumented to monitor its per- a conduit for the flow and to decouple the
formance and stress level. An environmen- relative motion of the rig and riser.
tal instrumentation package has also been
provided to correlate the riser stress
levels with environmental conditions.

- Upper Riser Package
The upper riser package serves several
functions - to connect the flowlines to the
rig while simultaneously decoupling the
relative motion of the rig and riser, to
maintain the riser In the proper relative
position with respect to the rig, and to
1W
allow the rig to quickly disconnect from
the riser If desired or to individually
disconnect each flowline for maintenance.

The air tanks at the top of the riser
(Figure 5) provide additional lift and
"righting" force to offset the forces from
the current. These tanks are modularized
and the top tanks are tapered to avoid
interference with the hose bundles.
The upper riser connector package
(URCP) provides the link between the rigid
tubing which protrudes from the top of the
riser and the flexible pipe which Is sus-
pended from the rig. In doing so, it
F
provides the capability to quickly discon-
from the rig if desired or
nect the riser
to individually disconnect any given flow- @y
.@A
line for maintenance. Individual hydraulic
mini-connectors are used in conjunction
with rigid pipe goosenecks to connect the
tubing strings to the flexible pipe. A
large hydraulic connector is located at the
center of the URCP. This connector trans-
mits forces from the riser tensioners to
the riser, as well as providing a connec- FIGURE 5. UPPER AIR TANKS, AS RISER IS RUN
tion for the 1211 gas sales line. It also
provides the one point quick disconnect
feature.
The production riser tensioner system
acts more as a riser centralizer than as a
typical riser tensioner. The production
riser is free standing and the tensioner is
not required for structural support. By
applying a restraining force to the top of
the rigid production riser and limiting Its
motion relative to the vessel, the flexible
flowlines between the vessel and the rigid
riser can be kept to a more easily managed
length and can thus remain connected during
more severe environmental conditions. The
tensloners supply such a restraining force
and the farther the production riser moves
from Its centrallzed position the greater
the restraining force. This is due to the
changing angle between the top of the rigid

611

A SINGLE BOARD COMPUTER BASED SAIL CONTROLLER

Timothy F. Pfeiffer

University of Delaware

85 we were able to develop a fairly clear
set of specifications for the new
ABSTRACT controller. Foremost, the data were to be
recorded on standard MS DOS disks in a
Several companies are now offering flat ASCII format that could be read into
IBM PC compatible single board computers. spreadsheets, statistical, or plotting
In their simplest form they do not have programs with no further processing.
keyboards or display adapters, but do Secondly the data collection portion of
include the program should be capable of starting
controllers for standard MS DOS disk itself, and restarting after a power
drives. Thus, while they cannot run much failure, with no operator intervention.
off the shelf software, programs can be Our final specification was that the
written on a standard PC and then executed controller have enough power and memory to
on the singleboard computer. In addition perform some real time processing of the
to MS DOS software compatiblity, some data such as computing salinity.
models also include a PC bus expansion slot
which provides hardware compatiblity for As we evaluated hardware to fill these
standardvideo and interface boards. We requirements we realized that a standard
have implemented a SAIL controller using a MS DOS computer actually had more
single board computer from Ampro Computers, capability than the job required. Ampro
Inc. The software does include a few Computers Inc. offers a single board
provisions to allow it to operate without computer, known as the Little Board 186,
an operator, but essentially it is ordinary which runs under MS DOS. The Little Board
PC software. These boards provide an 186 is driven by an Intel 80186 processor,
inexpensive and easily accessible method of has 512 kilobytes of memory, two serial
adding substantial intelligence to ports, one parallel printer port, and,a
instrumentation systems. disk drive controller. All of this is
contained on a single circuit board which
is the same size as a 5.25 inch disk drive
(5.75 inches x 7.75 inches). The Little
Board 186 does not include any display
The Serial Ascii Instrumentation Loop driver or keyboard interface. With the
is a hardware and software protocol addition of a power supply and standard
developed under the leadership of Dr. 360 kilobyte 5.25 inch floppy disk drive
Roderick Mesecar, Oregon State University, this one circuit board becomes a
which is used for collecting data from a functional MS DOS computer. The only
variety of sources aboard ship. cables required are one power cable to the
Essentially a SAIL system is composed of floppy drive, a second identical power
data modules which read information from cable for the Little Board 186 and a
various instruments and devices aboard the single ribbon cable between the disk drive
ship and a controller. The controller and the Little Board 186. In this minimal
periodically interrogates the data configuration the computer will start from
modules, processes and records their a standard MS DOS system disk and execute
responses. The original controller for our any commands in the autoexec.bat file just
system was an HP 85 computer which as any other MS DOS computer would. Of
provided a display, printer, and tape course with no display or keyboard
cartridge data storage in one integrated included in the system the program cannot
package. Recently, as both we and the interact with an operator. However all
scientific users of our system have forms of file manipulation and internal
standardized on MS DOS based computers for processing will be carried out properly
many data acquisition and analysis tasks and the serial and parallel ports are
it became apparent that we needed an MS available. Just these three components
DOS based SAIL controller. could, for example, if connected to a
printer, act as an off-line print spooler.
All that would be required would be to
Based on our experience with the HP include an autoexec.bat file with the

CH2585-8/88/0000- 612 $1 @1988 IEEE

appropriate print commands on the disk module. The data from unknown modules is
containing the files to be printed. The simply stored as received with no
disk files are completely MS DOS formatting, scaling, or other calculation
compatible. One of the Little Board 186's applied to it. For example we have a flow
two serial ports is configured to act as through Turner Designs fluorometer which
the system console. If a standard RS 232 has two analog outputs, one for range and
ASCII terminal is attached to this port the other for the dial. We have assigned
the Little Board 186 responds to the the fluorometer to addresses 30 for range
terminal's keyboard and displays program and 31 for dial. Thus during the loop
output on the terminal's screen. Programs poll the software will expect to find
which are strictly character based operate analog modules at those two addresses and
exactly as one would expect. This will assume that analog modules at those
configuration is very different from the addresses are indeed connected to the
hardware found in a standard PC, so the fluorometer. If one of those analog
program must be restricted to simple MS modules is to be used with an instrument
DOS and ROM BIOS character based input and other than the fluorometer, the module
output. The software cannot write directly must be opened and set to a different
to the display hardware and there is address. While this approach does reduce
essentially no graphics capability. This the flexibility of the system, I believe
means that a great deal of commercial MS that it increases the integrity of the
DOS application software, such as data by reducing the possibility of
spreadsheets and full featured word confusion over which module was recording
processors, will not run on these boards. which signal. If the data from a module
However this is hardly a drawback for a is not required, that module is simply
machine which will be dedicated to disconnected from the loop.
controlling a SAIL loop through its serial
interface. Once the software has established the
population of modules which are present on
The hardware for our SAIL system is the loop it begins collecting data by
composed of OSU data modules, an OSU addressing each module in turn. The
coupler module, and the Little Board 186 function associated with each known module
with its power supply and disk drive. The formats the data into engineering units,
coupler module provides both power for the applies calibration coefficients, or
loop and an RS 232 interface to the loop. otherwise processes the data. The data
The Little Board 186's serial interface is from unknown modules is simply stored as
connected to the coupler module. There received. After data has been gathered
normally is a terminal attached to the from all the modules, the functions are
Little Board 186's second serial port to called which utilize the data from two or
provide a display. more modules to compute a value. For
example, when the salinity function is
called it first verifies that both
The software which drives the Little temperature and conductivity modules are
Board 186 has been designed to run with no present, then it retrieves the stored data
intervention on the part of an operator. for conductivity and temperature, computes
In fact, the system is normally deployed salinity, and stores the result.
with no keyboard attached to it so there
is no possibility of a human interfering At the conclusion of each frame, the
with the execution of the program. The data is written to disk. The disk access
first task performed by the program is to is buffered by first writing the data to a
poll the loop, by issuing the information 300 kilobyte ram disk which is created
address for each of the possible loop with standard MS DOS software when the
addresses. According to the SAIL standard system is first powered up, Every five
when a module receives its information minutes the contents of the ram disk is
address it is to return a description of copied to the physical floppy disk. The
itself. We have established standard software traps the DOS critical error
addresses for those modules which we handler so that if there is a problem with
normally use, based on the UNOLS Standard the disk, such as no disk in the drive or
Ship Parameters, extended where necessary. the disk being full, the program is able
The software then compares the responses to continue instead of stopping at the DOS
received from polling the loop with the "Abort, retry, ignore?" prompt. If the
expected messages which are stored in the floppy disk is not accessible for any
program. If any other type of module reason the program continues to add data
answers, it will be flagged as an unknown to the file in the ram disk and check the
module. Any module which responds from an floppy disk every five minutes. When the
address which has not been assigned within floppy does become available the entire
the program is considered an unknown ram disk file is copied to it. If the

613

floppy is not available and the ram disk return at the end of the line. At the
fills, the program will continue to run, present time we have no modules which
but the data is not being stored. return text messages but if one were added
to the system its formatting function
After the data is written to disk it would enclose the text in quotes to ensure
is displayed. The display function that it would be read properly. The data
includes formatting and labelling files contain data only, with no titles or
information for all the known modules and other non-data items to interfere with
labels data from any unknown modules with reading the data. Each time the program
the address of the module. The display opens a new data file it also writes a
screen for the Little Board 186 is simply header file. The header file is simply
a terminal driven by a serial interface. two lines of text which identify the data
There is no handshaking required between and their units in the associated data
the Little Board 186 and the terminal so file. This approach does present a slight
if the terminal fails, or if no display is possibility that the wrong header might be
desired, the terminal can be turned off applied to a data file, but I feel that
without effecting the execution of the the risk is small compared to the benefit
software. of the ease of access to the data. if
desired the two files can be concatenated
The software continuously monitors so that both headers and numeric data are
the loop for continuity. If the loop is in a single file.
interrupted at any time, as by unplugging
a module, data acquisition halts and the MS DOS based single board computers
program attempts to copy all data in similar to the Little Board 186 which I
memory to the floppy disk and then the used for this project are available from
data file is closed. Since an several manufacturers and are a simple and
interruption of the loop may well indicate relatively inexpensive route to creating
a change. in the loop configuration, the data loggers and intelligent
software simply restarts, itself from the instrumentation. Since the software.can
beginning when continuity is restored. be developed using any MS DOS tools there
Thus after every break in the loop the is.no need to purchase cross compilers,
modules are polled, and new header and prom programmers, or development systems.
data files are opened. Since it is not The cost of creating a complete system,
possible to add a module to the loop assuming that a desktop computer and
without creating a break condition which language are already available, is simply
the software is able to recognize, the the cost of the Little Board 186, some
data from all modules will always be hardware to mount it in an appropriate
recorded. enclosure, and whatever time it takes to
write the software.
The printer interface has been
handled with the same philosophy. Once The project described in this article
every 10 data samples the software checks was supported in part by the National
the status of the printer port. If the Science Foundation.
printer is not available the program
simply continues. If the printer is
present and on line, the current data
frame is printed. Headers are printed at
the top of each page. If the length of a
data frame is greater than 132 characters
additional lines are used and a blank line
is inserted before the next frame. I have
made no provision for further formatting
of the data or selectively printing the
data from certain modules as I believe
this is best done off line with a
spreadsheet or statistical data base
program.

In order to allow the data files to
be read directly into spreadsheets or other
programs with no further processing the
format has , been kept as simple as
possible. The files are standard DOS
ASCII text with the data from each sample
interval written on a single line,
separated by spaces, and with a carriage

614

ON THE-KNOWLEDGE-BASED EXPERT SYSTEM FOR MARINE INSTRUMENTATION

M.R.Nayak

National Institute of Oceanography,
Dona Paula, Goa-403 004. INDIA.

ABSTARCT Artificial Intelligence (AI) originally
a branch of computer science has entered
in commercial technology enabling
The combination of intelligence and computers to perform intelligent actions
marine instrumentation has pushed the normally attributed to human beings. One
frontiers of oceanography to great of the most promising areas of applied
heights. One of the frontiers is the AI is the 'expert system' .(ES) used to
development of knowledge-based expert solve problems requiring the knowledge
systems (KbES). Marine instrumentation and skill 'of human experise Z-1-7. The
demands a comprehensive set . of last few years have 'seen an increasing
requirements, integrating diversified interest in the development of knowledge
technologies into the intelligent system based expert systems (KbES) Z2_7.
(that can be programmed, controlled and
made easy to operate). Extensive search
of knowledge-base is, in general, 2. METHODOLOGY
involved particularly while choosing the Knowledge repre .sentation is basically
appropriate digital IC logic family ' the. link that binds much of AI together.
Knowledge development process for an
In contrast to conventional database
:expert system is discussed in the paper.
systems, AI-systems require knowledge-
base with diverse kinds of knowledge.
These include knowledge about - objects,
time, goals, motivation, actions, etc;
in brief, classified; as objective
(quantitative) and subjective
(qualitative). A . general Ynowledge
representation is provided by Brachman
and Smith Z-- 3 _7, while summaries are
1. INTRODUCTION given by Barr and Fiegenbaum f 4-7. Most
of the knowledge-systems incorporate
In-situ data collection in the ocean knowledge through rules and data
is expensive and unpredictable even when. structures.
the programme is well planned and: I
executed. Inspite of the inherent @Marine instrument designer has to make
difficulties, little guidance is found a large search for choosing the
in the literature to assist in the appropriate components, particularly
planning ' and organizational stages; digital Integrated Circuits (ICs) logic
Marine instrument design,programme can family from a knowledge-base.. The
.be conceptualised into the following knowledge development process -for an
steps: expert system leading to a proper
- identification of the objectives, decision on the selection of ICs is
- identification of the data to be being discussed.
obtained,
- detailed electronic design A long standing need is felt to have a
(including selection of sensors with large knowledge-base expert system
proper resolution, accuracy and particularly to provide advice on the
,logistics) circuit design for marine/oceanographic
- implementation of the design (inco- instruments. The rules are a combination
rporating any modifications whenever of a set of procedures which are user
found necessary). modifiable. The knowledge-base consists

CH2585-8/88/0000- 615 $1 @1988 IEEE

of a large number of criteria, Once the design engineer , has
conditions, specifications on the identified the required specifications:
various types of digital logic families. evironmental and functional parameters
The criteria for decision making is: can. be .expressed in terms of 0knits'
Whether sufficient information is out of) Ch to C."% . The number of
available, functions thus identified for a
whether conditions for selection particular@application, now represents
of proper components for a parti- the availability of complex circuits.
cular design,are satisfied. Comparison can now be made based on the
AND-OR rules within the category of.ICs
Generally,a marine instrument engineer Z-8 _7.
requires' an exhaustive search of
component particularly. constrained by The basic concept underlying this
two factors Ir 5.7: design,. is the use of minimum but
- lack of electrical power from the sufficient amount of knowledge-base(Kb)
shore (low power devices); to design an optimum solution in the
- harsh marine environment (high form of an expert system (ES). The
relative humidity). standard minimization techniques such as
Hence, out of the electronic (digital) Karnaugh-map or Quine-McCluskey methods
logic family of RTL, DTL, TTL, LSTTL, could be used to arrive at the minimised
,nMOS, pMOS, CMOS, PLEs the choice falls design in a marine instrument design
on*. programme with particular emphasis on
* LSTTL : Low Power Schottky TTL, oceanographic applications.( Fig.1)
* nMOS : n-channel MOS,
* PMOS : P-channel MOS,
* CMOS : Complementary Symmetry MOS,
CMOS PLEB: CMOS Programmable Logic
Elements.

The basic characteristics of these Logic Family
devices ake described in RL Morris and
JR Miller C 6 -7, and PLE Handbook C 7 -7. FLF
The selection of devices. can be made
based on the following characteristics:
- fanout (n),
- noise immunity (dc and ac), F1 1C, 1C, C, C, Ic
In
- power dissipation per gate (m.W), LSTTL nMOS pMOS C110S PLE
- ambient temperature (* C),
- operating RH ( %k, fan-out
- clock rate (MHz)f.
- propagation delay per gate (nS), :noise imm
- switching speed,
input characteristics, 6 power diss: * 0
output characteristics, 'and
power supply voltage. a amb. temp: 0 9
The knowledge bank consists of ' all the 0 opg. RH
,above mentioned categories of
information (values). In addition, the * dk rate
.above categories can further be
sub-divided into a large number of sub- 9 prop. delay:
categoiies. As an example, 'fan-out.' < 5
'or 5<f<10; etc. These bits of knowledge
@can bs coded. and stored as knowledge
.information bits ("knits*, in short) in
@the knowledge-bank, after sub-dividing
the main categories into a large number
of thinkable categories. These
!categories and sub-units can thus be!
arranged in a matrix organisation as:

C% C111 C11 C41 CS1 C& C%1
C %a. Cau C32. C43L C52 C&I Ch,. e switchg spd:

C C C C C C
Fig. 1 Logic Families. classification

616

All the knowledge representation 5. REFERENCES
characteristics can be correlated during
the selection of knowledge 1. Kaisler S. EXPERT SYSTEMS:AN OVERVIEW
representation formats to reduce IEEE Trans.Ocean Engg,VolOE-11,No.4,
repetition. The variety of knowledge is pp442-448, October 1986.
stored using a data based on a widely
divergent internal storage structure. 2. Helen C. Shen & GFP Signarowski, A
The decisions for the selection of KNOWLEDGE REPRESENTATON FOR ROVING
appropriate components by accessing the ROBOTS. Proc.II Conf.on AI applicatn
@knowledge representation formats, and (CAIA).IEEE Comp.Soc.1985 pp621-628.
the system control structure is an
important stage of the design. Although 3. Brachman RJ & BC Smith, SIGART
the sequence and extent of operations NEWSLETTER, No.90 (Feb.1980).
execution may change, the manner of
monitoring, tracking and extracting the 4. Barr a & Feigenbaum E, THE HANDBOOK
relevant inference decisions does not. OF ARTIFICIAL INTELLIGENCE, Voll,
Harris Tech.Press,Stanford,USA)1981

3. CONCLUSIONS 5. Desa E, MR Nayak, RGP Desai.MICRO-
PROCESSOR DESIGN ASPECTS FOR COASTAL
In general, marine instrumentation can OCEANOGRAPHIC INSTRUMENTS. Jour.of
be conceptualised into several steps, Coastal Res.(USA). 4(3) 1988.
vi z;.:
- identification of objectives, 6. Morris RL & JR Miller. DESIGNING WITH
- identification of the data from TTL ICs. 1982.; McGraw-Hill Kogakusha
the knowledge-bank, Ltd; Japan.
- detailed design of the programme
including selection of the compo- 7. PLE HANDBOOK. Monolithic Memories Inc
nents and monitoring network, 1987 USA.
data processing, and.
data analysis. 8. Mazlak LJ. USING SUFFICIENCY THRESHOD
RULES IN A KNOWLEDGE BASED SYSTEM.
As an example of the application of Proc.I Int'l Workshop on Expert ,
knowledge engineering in the early Database System.October 1984,p730-734
stages of the design, the role of
knowledge-base in the understanding of
the specifications is considered here. ---------
When the product is to be designed, the
desires of the user must be represented
in the form of the specificitions
according to which the design process is
carried out. The designer then addresses
the knowledge-base for decision-making
to select the appropriate logic family
for the specific marine application.
Emphasis is given particularly to the
role of knowledge-base in expert systems
of the future. This being a new field,
it is expected to have a wide range of
applications particularly in the design
of marine instruments for long-term,
remote, unattended operations. The
continuation of effort is anticipated in
the future development in this area.

4.ACKNOWLEDGEMENTS

The author wishes to thank Dr.B.N.Desai
Director, Dr.E.Desa, Head, MICD and
.S/Shri J.S.Sarupria and G.V.Reddy for
@@useful discussions in, the preparation
.and presentation of this paper.

617

OPDIN ONE WAY THE OCEAN COMMUNITY INFORMS

Ronald J. Smith

Ocean Pollution Data and Information Network
National Oceanographic Data Center
National Oceanic and Atmospheric Administration
Washington, DC 20235

ABSTRACT identified as potential field data col-
lectors. Summaries by year, project
In 1978 the National Ocean Pollution region, and pollutant type were gener-
Planning Act established NOAA's respon- ated prior to a more detailed project
sibility for the timely dissemination of review.
ocean pollution information produced by
Federally funded or sponsored programs Further review of the flagged projects
to persons having an interest in ocean involved tracking them across success-
pollution research, development and mon- ive annual Catalogs to remove redundancy
itoring. To this end, NOAA's National and to identify any changes in primary
Oceanographic Data Center (NODC) created investigators. A report was then
the ocean Pollution Data and Information generated that lists the projects for
Network (OPDIN). Since 1981 the Network each fiscal year by Agency, Department,
has been providing the marine pollution and Program. For each project the
community with an extensive list of pro- report provides project title, region,
ducts and services. One of the Network's data type, point of contact, institu-
recent efforts is to identify Federally tion, and Catalog project number.
funded projects listed in the FY 1978 -
FY 1983 editions of the Federal Marine This second stage of project review
Pollution Project Catalog that have col- resulted in a list of 393 projects as
lected marine pollution field data. A potential sources of marine pollution
final report is targeted for the end of data. The types of data which these
1988. It will list data-producing pro- projects may have produced have been
jects by agency, program, institution, corre
and primary investigator. lated with the pollutant categories
(Table 1) used in the National Marine
Pollution Plan (NMPPO 1985b). Each
INTRODUCTION project is also identified by its region
of interest (Table 2). The number of
Section 8 of the National ocean Pollu- projects potentially producing data in
tion Planning Act of 1978 (Public Law each of the 'data type' categories is
95-273, supplemented by PL 99-272 to summarized by marine pollution program
include specifically the Great Lakes and in Table 3. The number of projects pro-
estuaries of National importance) man- ducing data for each region is summar-
dates that the results of Federally ized by program in Table 4.
sponsored marine pollution programs be
disseminated to all interested persons.
In an effort to fulfill NOAA's PROJECT WORKSHEET AND
responsibility under Section 8 of the ASSOCIATED DATABASE
Act, the Central coordination and A worksheet was created to facilitate
Referral Office (CCRO) of the Ocean individual project characterization and
Pollution Data and Information Network future database key-entry. The work-
(OPDIN) has identified those Federally- sheet contained project information
funded marine pollution related projects gathered from the Catalogs. It also in-
described in the annual Catalogs of Fed- cluded the following attribute fields to
eral Projects published for Fiscal Years be completed by the most recent project
1978 through 1983 (COPRDM 1978, 1979, contact or primary investigator:
1981, 1982, 1983, 1984) that may have
collected pollution related measurements * Did this project collect field data?
in the marine environment. * Region of study
* Zone of interest
Approximately 20% of the over 5,300 * Data types ( pollutant or parameter)
separate Catalog projects were initially * Survey periods

618 United States Government work not protected by copyright

� Number of surveys projects did collect marine pollution
� Approximate quantity of data field data. Most project data is avail-
� Present storage media of results able to interested parties in the form
� Location of stored data of journal papers or agency publica-
� Data availability to secondary user tions. Other projects have generated
� Cost of making data available digital data which is available in vary-
� Further comments ing degrees to secondary users.

copies of individual project worksheets
were forwarded to the most recent re- FUTURE EFFORTS
spective primary investigator as listed
in the Catalogs. The project worksheet responses have
been key-entered into the database and
DATAEASE, a commercial DBMS package, was a report will be generated in the latter
used to design a database to hold the part of 1988 identifying data and infor-
information gathered from the worksheet mation resources potentially available
responses. to secondary users. For easy access, the
report will be indexed by data type and
region.
PROJECT WORKSHEET RESPONSE
Depending on the success of the current
A total of 393 project worksheets were effort, a follow-up evaluation of the FY
forwarded to the respective project 1984 and FY 1985 Catalogs (NMPPO 1985a,
primary investigators or contacts. To NMPPO, 1985c) may also be produced. The
date, approximately 150 completed marine pollution data and information
worksheets have been returned to the user will, as in the past, help
Network. 'Initial review of these determine the utility and future Network
returns indicate that 90% of these resources expended on this task.

TABLE I

Occurrences Pollutant Category
165 Metali-and inorganic chemicals (F)
140 synthetic organic chemicals (B)
92 Petroleum and petroleum products (A)
91 Nutrients and other biostimulants (E)
63 Habitat modification and sediment deposition (H)
44 - Miscellaneous (I)
40 - Pathogens and microorganisms (D)
30 - Radionuclides (G)
3 - Halogens and halogenation products (C)
---- - ------------------------------------------
668 Total Number of Occurrences

TABLE 2

Region Count Region
92 - NE - New England and Mid-Atlantic States
80 - GL - Great Lakes States including portions of New
York and Pennsylvania
67 - us - All regions or possible multiple undefined areas
43 - SE - North and South Carolina, Georgia, the east
coast of Florida, Puerto Rico and the Virgin
Islands
39 - NW - Washington and Oregon
38 - GM - West coast of Florida and all states bordering
on the Gulf of Mexico
35 - SW - California, Hawaii, Guam, American Samoa and
the Trust Territories
19 - AK - Alaska

619

TABLE 3

Data Type Distributio of Potential Data Activities

Department/Agency-Program Total Data Type Count
Proj. A B C D E F G H
------------------------------------------------------------------------

USDA-Effects of Agricultural Practices 5 - 2 - - 3 1 - 3 -

NOAA-Coastal and Estuarine Assess. Program 15 7 10 - 1 110 - 1 2
Ocean Use Impact Program 4 - - - 2 2 2 - - -
Deep seabed mining Environmental Res. 3 - - - - - 1 - 1 2
National Marine Sanctuaries Program 1 - - - - I - - 1 -
ERL Ocean Pollution Studies Program 3 - 2 - 1 2 2 - I -
ERL Great Lakes Pollution Studies 4 2 3 - - 3 - I - -
Sea Grant Ocean Pollution Program 43 6 14 1 8 611 1 1 2
National Fishery Ecology Program 23 8 12 - 7 517 1 5 2
Microconstituents Program 5 3 4 - - - 2 - - -
Financial Assistance Program 6 1 3 - - 1 1 - 1
Long Range Effects Research Program 5 - 3 - - - 2 - -
Ocean Pollution Monitoring Program 2 2 2 - 1 - I - 1 1
Ocean Dumping Program 11 - 6 - - - 7 - -
Marine Ecosystems Analysis Program 9 2 7 - 4 4 3 - -
Hudson-Raritan Estuary Project 2 - 1 - - - 2 - - -
Puget Sound Project 2 1 - - - - 2 - - -

DOD/ACOE-Environmental Quality Program 3 - 1 - - 1 1 - 3 -
Development and Monitoring Program 9 - - - - - - - 9 -

DOD/Navy-Environment Protection Technology I - - - - - 1

DOE- Ocean Thermal Energy Conversion Prog. 4 - - 1 - - -
Regional Marine Program 23 2 2 - 1 5 7 5 - 7
Radioecology Program 9 - - - - - - 9 - -
Physiological Ecology Program 3 2 - - - 1 3 - - -
Strategic Petroleum Reserve Program 2 - - - - 2 1 - 1 2
Subseabed Disposal Program 1 - - - - - - I - I
Estuarine Program 6 1 - - - 1 4 2 2
Stable and Radioactive Elements Prog. 3 - - - - - - 3 -
oil and Gas Program 2 2 - - - - - - - -
Cooling Systems Program 1 - - - - - - - - 1

HHS/NIEHS-Extramural Grant Program 3 1 2 - - - 3 - - -

HHS/FDA-Shellfish Sanitation Program 2 1 1 - 2 1 1 - 1 -

DOI/MMS-Washington, D.C. Office Studies 4 4 - - - - 2 - - -
Atlantic OCS Regional Studies 6 5 - - - I - - 1 1
Gulf of Mexico OCS Regional Studies 2 2 - - - - - - - 2
Pacific OCS Regional Studies 3 2 - - 1 1 - - - I
Alaska OCS Regional Studies 12 12 1 - 2 - - - - -

DOI/FWS- Fishery Resources Program 5 - 4 - - 3 3 1
Habitat Resources Program 2 - - - - I 1 1
Env. Contaminants Eval. Program 6 1 6 - - - 5 - - -

DOI/USGS- Water Resources Division Program 15 - 5 - 1 8 8 - 5 4
Geologic Division Program 6 3 - - - - 4 - 2 1

DOT/USCG-Port and Env. Safety Program 1 1 - - - - - - - -
Marine Env. Response Program 5 5 2 - - - I -

EPA- Marine Waste Disposal Program 11 1 2 - 3 - 3 - & 4
Energy Related Research Program 4 3 - - - - 1 - - 1
Water Quality Research Program 3 - 2 - - I I - 1 1
Great Lakes Research Program 41 - 24 - 6 19 25 1 3 -
Chesapeake Bay Program 15 - 7 - - 7 9 - 4 -
Exploratory Research Program 6 3 3 - - 1 3 3 1 -
Marine Ecology Studies Program 7 2 6 - - 2 4 - 2 -
Energy-Environment Interagency Prog. 5 5 - - - - I - - -
Evaluation of Radioactive Waste Prog. 1 - - - - - - 1 - -
Ocean Dump Site Evaluation Program 1 - 1 - - 1 1 - - -

NSF- Division of Ocean Science 7 1 1 - - 1 4 - - 1
Chemical oceanography Program 3 1 1 - - 2 2 1 2 -
Biological oceanography Program 3 - - - - 3 - - 3 -

NRC- Environmental Impact Assessment Res. 3 - - 1 - - 2 - - 1

NPS- National Park Service Program 1 - - - - I - - 1 -

620

TABLE 4

Regional Distribution of Potential Data Activities

Department/Agency-Program Total Regional Count
Proj. NE SE GM SW NW AK GL US
------------------------------------------------------------------------

USDA-Effects of Agricultural Practices Program 5 1 3 - - - - -

NOAA-Coastal and Estuarine Assessment Program 15 5 - 2 2 4 1 1
Ocean Use Impact Program 4 2 1 1 - - - 1
Deep Seabed Mining Environmental Research 3 - - - 3 - - - -
National Marine Sanctuaries Program 1 - I - - - - - -
ERL Ocean Pollution Studies Program 3 - - 1 - 1,- - 1
ERL Great Lakes Pollution Studies Program 4 - - - - - - 4 -
Sea Grant Ocean Pollution Program 43 8 4 7 4 1 1 17 2
National Fishery Ecology Program 23 9 3 7 1 4 3 - -
Microconstituents Program 5 1 2 1 - 1 - - 1
Financial Assistance Program 6 1 1 1 1 - - 2 -
Long Range Effect Research Program 5 2 - 1 2 - - - -
Ocean Pollution Monitoring Program 2 2 - - - - - - -
Ocean Dumping Program 11 6 5 - - I - - -
Marine Ecosystem Analysis Program 9 9 - - - - - - -
Hudson-Raritan Estuary Project 2 2 - - - - - - -
Puget Sound Project 2 - - - - 2 - - -

DOD/ACOE-Environmental Quality Program 3 - - - - - - - 3
Development and Monitoring Program 9 1 1 1 2 1.- 2 1

DOD/Navy-Environmental Protection Technology 1 - - - - - - - 1

DOE- Ocean Thermal Energy Conversion Program 4 - 2 1 1 - - - 1
Regional Marine Program 23 5 5 - 2 6 - 2 3
Radioecology Program 9 2 1 1 2 2 - - 1
Physiological Ecology Program 3 - - - - 2 - - 1
Strategic Petroleum Reserve Program 2 - - 2 - - - - -
Subseabed Disposal Program 1 - - - - - - - 1
Estuarine Program 6 2 2 - - - 1 - 1
Stable and Radioactive Elements Program 3 - - - 1 - 2
Oil and Gas Program 2 1 - - 1 - - - -
Cooling Systems Program 1 1 - - - - - - -

HHS/NIEHS-Extramural Grant Program 3 - - - - 1 2

HHS/FDA-Shellfish Sanitation Program 2 - - 1

DOI/MMS-Washington, D.C. Office Studies 4 1 1 1 - - - - 3
Atlantic OCS Regional Studies Program 6 4 3 - - - - - -
Gulf of Mexico OCS Regional Studies 2 - - 2 - - - - -
Pacific OCS Regional Studies Program 3 - - - 3 - - - -
Alaska OCS Regional Studies Program 12 - - - - - 12 - -

DOI/FWS- Fishery Resources Program 5 - - - - - - 5 -
Habitat Resources Program 2 - - - - - - 2 -
Env. Contaminants Eval. Program 6 2 1 - - - - 2 2

DOI/USGS- Water Resources Division Program 15 4 3 1 3 - - 1 3
Geologic Division Program 6 1 1 2 1 1 1 - 2

DOT/USCG-Port and Environmental Safety Program I - - - - - - - 1
Marine Environmental Response Program 5 - - - - - - - 5

EPA- Marine Waste Disposal Program 11 4 2 - 2 2 1
Energy Related Research Program 4 - - - 1 1 2
Water Quality Research Program 3 - - I - 1 1
Great Lakes Research Program 41 - - - - - - 41 -
Chesapeake Bay Program 15 15 - - - - - - -
Exploratory Research Program 6 - - - - 1 - - 5
Marine Ecology Studies Program 7 - - 1 - 5 - - I
Energy-Environment Interagency Program 5 - - 1 1 1 - - 2
Evaluation of Radioactive Waste Program I I - 1 1 - - - -
Ocean Dump Site Evaluation Program 1 - - - - - - - I

NSF- Division of Ocean science 7 - - - 1 1 6
chemical oceanography Program 3 - - - - - - - 3
Biological Oceanography Program 3 - - - - - - 1 2

NRC- Environmental Impact Assessment Research 3 - - - - - - - 3

NPS- National Park Service Program 1 - 1 - - - - - -

621

REFERENCES COPRDM. 1983. Catalog of Federal
Projects, IY 1982 Updat pendix 2 to
the Federal Plan for Ocean Pollution
COPRDM. 1978. Catalog of Federal Re- Research, Development, and Monitoring
search, Development and Monitoring Interagency Committee on Ocean Pollution
Programs for Fiscal Years 1978-80. Research, Development, and Monitoring,
Interagency Committee on Ocean Pollution Federal Coordinating Council for
Research, Development, and Monitoring, science, Engineering and Technology).
Federal Coordinating Council for
Science, Engineering and Technology). COPRDM. 1984. Catalog of Federal
Projects, EY 1983 Update, A-p-pendix 2 to
COPRDM. 1979. Catalog of Ocean Pollution the Federal Plan for Ocean Pollution
Research, Development and Monitoring Resea-rch, Development, and Monitoring.
Programs for Fiscal Years 1978-80. Interagency Committee on Ocean Pollution
Fqorking Paper I for tHe fe-deral @;_Ian for Research, Development, and Monitoring,
Ocean Pollution Research, Development, Federal Coordinating council for
and Monitoring. Interagency Committee on Science, Engineering and Technology).
Ocean Pollution Research, Development,
and Monitoring, Federal Coordinating NMPPO. 1985a. Catalog of Federal
Council for Science, Engineering and Projects, FY 1984 Update, Appendix 2 to
Technology). the Federal Plan for Ocean Pollution
Research, Development, and Monitoring.
COPRDM. 1981. Catalog of Federal National Marine Pollution Program
Projects, LY-1980 Update, Appendix 2 to Office, NOAA.
the Federal Plan for Ocean Pollution
Research, Development, and Monitorin . NMPPO. 1985b. National Marine Pollution
Interagency Committee on Ocean Pollution Program: Federal Plan for Ocean
Research, Development, and Monitoring, Pollution Resea-rch, Development, &
Federal Coordinating Council for Monitoring -- Fiscal Years 1985-1989.
Science, Engineering and Technology). Prepared E-y the National Marine
Pollution Program Office for the
COPRDM. 1982. Catalog of Federal Interagency Committee for Ocean
Projects, fY 1981 Update, Appendix 2 to Pollution Research, Development, and
the Federal Plan for Ocean Pollution monitoring.
Research, Development, and Monitorin
Interagency Committee on Ocean Pollution NMPPO. 1985c. Summary I of Federal
Research, Development, and Monitoring, Programs and Projects, F 985 Update,
Federal Coordinating Council for National Marine PoliTt-i-on Program
Science, Engineering and Technology). Office, NOAA.

622

THE USE OF WORM OPTICAL DISKS
IN OCEAN SYSTEMS

Dennis Stamulis and Michael P. Shevenell

Marine Systems Engineering Laboratory
University of New Hampshire
Durham, New Hampshire 03824

ABSTRACT many sensors must be packaged in small
cylindrical pressure housings, size must
The need for large amounts of non-volatile be kept to a minimum. The need f or a
storage has led the Marine Systems robust removal media is critical because
Engineering Laboratory to consider the use it is often impossible or impractical to;
of WORM (Write once Read Many) optical transport the entire instrument. Instead,'
disks for data and program storage. only the medium is removed and transported
Although optical disks have the capability to the lab for data analysis.
to store huge amounts of data (e.g. 800
Mbytes), being a write once medium The capabilities of present WORM optical
presents some interesting file management disks meet all the criterion expressed for
problems. high density, non-volatile removal
oceanographic storage. There are myriads
This paper presents t Ihe system design of uses for current optical disk
which includes a file, system, hardware technology; these include: bathemetric
interface, and performance measures. A' surveys, current profiling, water column
unique feature of the file management measurements, side scan sonar and digital
system allows multiple asynchronous files image storage. Although a single digital
to be written. In addition single files image may contain a large amount of data
are handled efficiently by utilizing the (256k bytes for a 512 x 512.pixel image)
large buffer and high transfer rate of the over 1500 images can be stored on one side
disk controller. The disk is interfaced of an optical disk.
to the development system running Unix as
well as the run time system running PSOS, The application which has driven the
a commercial real-time operating system. Marine Systems Engineering Laboratory
This scheme which is used on EAVE, an (MSEL) at the University of New Hampshire
underwater autonomous vehicle and can be is the EAVE AUV research and development
applied to data collection by the ocean program. The development of intelligent
community. .,autonomous vehicles has seen a great
increase in the complexity and size of
code to perform useful tasks. As the code
size increased, field downloading became
I. INTRODUCTION impractical with previous methods. As an
example, a LISP-based processing
As more advanced systems and sensors are environment requires over 2 Mbytes of data
used in oceanographic data collection, the storage. In addition, as more complex
demand for high density data storage missions are performed the need for large
increases. An example of this data data storage increases.
explosion is shown in the digital storing
of side scan sonar data. Assuming a side 2. HARDWARE
scan sonar with two channels each having a
bandwidth of lOkHz produces 40k bytes per As mentioned, the integration of a WORM
second when the samples are digitized to 8 optical disk comes in the context of the
bits, a single hour of digital data EAVE autonomous vehicle program.
collection would require 144M bytes of specifically, the disk has become part of
storage. the MSEL Knowledge-Based Control
Architecture based on the VME bus (see
Figure 1). The disk interfaces to the VME
in addition to the high data capacity bus System Controller board which contains
required in present oceanographic data an industry standard Small Computer
systems, small size is desirable. since Systems Interface (SCSI) port. In

CH2585-8/88/0000- 623 $1 @1988 IEEE

addition to the SCSI port, the System traditional method of updating lists could
Controller board has special hardware and not be implemented.
a DMA controller to perform fast data
transfers without requiring a CPU to This file system is essentially a one
handle the data. As shown in Table 1, the level hierarchical system that contains.
maximum data transfer rate of the disk four types of blocks: a super inode (SIN),
drive is 1.0 Mbytes per second; this data a file inode (FIN) , a directory inode
rate is easily handled by the DMA (DIN),, and a data block (DB) (see Figure
hardware. 2). Special information such as a time
stamp, and date/time of creation
There are presently several manufacturers identifies the particular optical disk
of high capacity WORM optical disk drives. being used and is stored in the SIN.
At the start of this project Maxtor was There is only one SIN assigned to each
selected as the vendor because it offered side of an optical disk and is always
the highest capacity in the 5 .1/4 inch block number 191951, the highest numbered
form factor and had an embedded SCSI block on the optical disk. Information
controller. The model chosen is the about a file, its name, time of creation,
Maxtor RXT-800S which has 400 Mbytes of a pointer to its f irst block of data,
formatted storage per side of the media. etc., is stored in the FIN. The size of
the file in bytes may or may not be stored
TABLE 1 depending on which write mode was used
Selected Maxtor RXT-800S Specifications (this will be explained later) . The DIN
is unique in that it stores all the
Capacity 800 Mbytes information held in the previous twenty
Average Seek Time 108 milliseconds FIDs. Finally, the DB contains a byte
Peak Power 5v 1.3A count, a pointer to the next DB used, and
12v 3.5A the actual file data being stored.
Included in this scheme is a special file
SCSI transfer rate 1.0 Mbytes/sec (OSTAT) stored on the host winchester disk
Disk transfer rate 2.5 Mbits/sec that contains information pertaining to
Operating Shock 2G @11 milliseconds the state of the optical disk currently
being used. Each optical disk used must
Data Reliability <1 error in 10-12 first be initialized. This initialization
Data Life >5 years process writes a time stamp and other
Dimensions Width 5.75 in optical disk identifying information to
Depth 8.00 in the SIN and to the f ile OSTAT. Besides
Height 3.25 in containing the same optical disk
identifying information found in the SIN,
3. THE FILE SYSTEM five pieces of state information are
needed to be stored in OSTAT for each
During the design of the file system optical disk in use. They are:
structure three requirements had to be
addressed: compatibility, speed, and 9 The block number of the last written
simplicity. A file system would not only DIN.
need to run under UNIX, but be easily
ported to run on the EAVE vehicle under. 9 The block number of the next free FIN.
its own operating system PSOS. Since the
optical disk would be used to download 9 The block number of the next free DB.
large amounts of code to the vehicle in
the field, optimizing the read/write e The number of DINs written.
commands to the disk was essential.
Lastly, keeping the structure simple e The number of FINs written since the
enhances the speed, allows for easy last written DIN. This implies that
maintenance, and assures the size of the each optical disk initialized is a new
executable will be reasonable. The size of entry in the host file OSTAT.
the executable is . an important
consideration since the vehicle has a 5. READING AND WRITING
finitememory.
As mentioned above, the optical disk is
4. THE FILE SYSTEM ORGANIZATION composed of contiguous blocks numbered 0
to 191951 with the largest numbered block
Each side of the optical disk- is a designated the SIN. In this scheme, FINs
contiguous set of indivisible blocks, 2048 and DINs are allocated from block number
bytes in length, and numbered from 0 to 191950 decrementing towards lower numbered
191951. Blocks may be written only once blocks ' and DBs are allocated starting
but read as many times as needed. It would from block number 0 incrementing towards
be an error to attempt to read an higher numbered blocks as needed. Anytime
unwritten block or attempt to write to a an operation is to be performed on an
used one. Because of this, the optical disk, its SIN is read and the

624

OSTAT file is searched in order to find a FINs that have not yet been wri tten to a
matching entry. This procedure verifies DIN - nineteen at most. If the named file
that the disk has indeed been initialized has not been found, only the remaining
and its state information has been entered DINs need to be searched since only they
in OSTAT. contain information about twenty files
previously written. Searches are always
There are two specific write modes that. performed in such a manner to insure only
can be applied when writing to the an the most current version of duplicate
optical disk, the linked list (LL) mode named files will be located first. Using a.
and the sequential (SQ) mode. When a new file's FIN only requires one optical disk
file is to be written, all pertinent file access to find where the file resides on
information is entered into the next free that disk. This option was needed to
SIN, including a pointer to the next free provide for reading any version of a
DB. If the file to be written is duplicate named files. It should be noted
redirected from standard input, the write that when reading, a file would be
routine cannot determine in advance the accessed in the same mode it was written.
actual size of the file, and therefore If it was written using the LL mode it
cannot include the f ile size as part of must be read one block at a time.
the FIN. In this case, a zero file size Otherwise, it may also take advantage of
is entered in FIN to record that the LL the size of optical disk controller
write mode was used. This LL mode causes transfer buffer and read up to its maximum
one disk block to be written at a time size for each disk access.
with each new DB written to the optical
disk pointing to the next free DB. If the
new f ile to be written is given on the 6. UNIX VS. PSOS
command line as an argument to the write
program, the size of the file can be This file system and its associated
predetermined before the FIN is written to routines are compatible so they run under
the optical disk and is included as part both UNIX and PSOS, the EAVE vehicle's
of the file information stored there. In operating system. It is an ideal, method
this SQ mode the number of DBs needed can to download large amounts of code from our
be determined in advance enabling each development system to the EAVE vehicle.
write to the disk to be up to the maximum First, code is written to an optical'disk.
size of the optical disk controller Then, after completion of data transfer
transfer buffer (i.e. 62 blocks of data). the disk is removed and then installed in
After a file has been completely written the optical disk drive of the EAVE
to the optical disk, if the number of vehicle. The vehicle is booted and the
FINs written is equivalent to twenty, a code on the optical disk is loaded into
DIN containing all information recorded in the vehicle's RAM and executed. Both the
the previous twenty FINs is written to the above writing and reading are done in SQ
next free FIN location. At this time, if mode to insure that writing and reading
a previous FIN has a 0 byte count, it must will be done as fast as possible.
have been written in LL mode and its links However, when the vehicle is actually
would be traced, byte count calculated, performing a mission and storing data, all
and then recorded in its corresponding writing is done via the LL mode. All data
entry in the DIN. A special code is recorded is a continuous stream and
entered in the DIN for each file that has emanating f rom, many sources on the EAVE
been written in SQ mode so that any vehicle.
reading of files can also take advantage
of the size of the optical disk controller 7. THE BOOT SEQUENCE
transfer buffer.
Activating the prommed boot code on the
Each access to the optical disk must be EAVE vehicle is a simple matter of
preceded by the four basic SCSI phases: entering 11BU11 on the active terminal after
select, command, data transfer (if any), a reset has been initiated. The program
and status. It can be seen that the LL will then prompt for a SIN number and,
mode is inefficient since the four SCSI once given, will read the proper code into
phases must be executed for each block RAM and execute it. At this time a Master
written. However, it does provide a means Console (MASCON) program will be up and
of writing multiple open files. running. This program is similar to the
Conversely the SQ mode needs to execute UNIX shell in that it allows the user to
the four basic SCSI phases only once per execute certain useful commands.. one of
each 62 blocks transferred. these commands brings up other -modules
necessary for vehicle operation. This
A read operation can also be performed in command loads each modules' image from the
one of two ways, searching for the file by optical disk and then executes it.
its name, or accessing it by using its FIN
number. When searching for a file by
name, it is only necessary to scan those

625

8. SYSTEM PERFORMANCE

Preliminary performance measurements have
been made using a VME bus based 68020 CPU
board handling the data transfer in a
programmed 1/0 mode. The DMA mode of data
transfer has not yet been implemented.
The results of these tests show the
transfer rate for a large sequential block
file to be about 75 kbytes per second.
Although the DMA controlled will handle
data at speeds approaching 1 Mbyte/second,
the programmed 1/0 transfer rate may be
sufficient for many applications.

9. CONCLUSION

The experience at the Marine Systems
Engineering Laboratory has shown the WORM
optical disk to be a vital piece of
technology to be include in oceanographic
data systems as well as advanced
autonomous vehicles. The unit provides
high density, non-volatile removal storage
at low cost (a typical unit costs about
$2000 with media costing $100 for 800
Mbytes).

This file system has proven to be
portable, simple, and flexible, running
under UNIX host as well as PSOS; the EAVE
vehicle operating system. It can be used
to download code from one system to
another in a fast and efficient manner.
For use in the f ield, it is a greatly
improved method for data collection since
the data collection medium can be
transferred to another system where that
data can be easily analyzed.

10. REFERENCES

1. Gait, Jason, "The optical File Cabinet:
A Random-Access File System for Write-
once Optical Disks", IEEE Computer,
June 1988, pp. 11-22.

2. Shevenell, Michael P. "Hardware and
Software Architectures for Realizing a
Knowledge Based System on EAVE11, Fifth
International Symposium on Unmanned
Untethered Submersible Technology,
Marine Systems Engineering Laboratory,
UNH, Durham, NH, pp. 220-237.
3. "A cooperative Research Project on
Intelligent Control for Multiple
Autonomous Undersea Vehicles", MSEL
Report #88-03, Marine Systems
Engineering Laboratory, UNH, Durham,
NH.

4. "Maxtor RXT-800S Product Specification
and OEM Manual", Maxtor Corp. San Jose,
CA. "NCR 5380-53CSO SCSI Interface Chip
Design Manual", NCR Microelectronics
Division, Colorado springs, CO.
"Ironics IV-3273 VME bus System
Controller Users Manual", Ironics Inc.,
Ithaca, NY.

626

VME BUS

-1 F
SIMTW 68020 68010 68020 4M RAM 68020
m 256K RAM
CNTRL P 8 SERIAL 4M RAM 4M RAM

DATA GUIDANCE SITUATION SUPERVISOR
ASSESSNM ASSESSMENT PLANNER

IWSEL BUS 68000 COMPUTERS

TOP LEVEL ARCHITECTURE

FIGURE 1.
@
6
25 6

DAT
SESS
@AS

627

191951 SIN

191950 FIN

191949 FIN

191931 FIN

191930 DIN

191929 FIN

4 Byte DB LINK
Count
Byte LINK
3 Count DB

Byte
2 Count DB LINK

Byte
1 Count DB LINK
0 Byte D LINK
Count B

OPTICAL DISK FILE SYSTEM FORMAT

FIGURE 2.

628

A METHOD FOR OPTIMIZING
ENVIRONMENTAL OBSERVING NETWORKS

W. Brett Wilson

National Data Buoy Center (NDBQ
Stennis Space Center, MS 39529-6000

ABSTRA CT understood, can lead to improper conclusions. Among these are: 1) the
need to properly develop the input data and interpret the results for ap-
Linear programming techniques, based on calculus of variations, have plicability and reasonableness, and 2) the need to understand the restric-
Proven useful in a variety of disciplines for optimizing multi-attribute tions inherent in the techique, such as the linearity assumptions in linear
objective functions subject to constraints. Thispaper documents the ap- programming.
plication of linear programming to the problem of designing an en-
vironmental observing network consisting of multiple sensor and plat- In 1988, an independent project to create such a numerical model was
form systems in light of budget and other limitations. initiated.

An environmental observing network optimization numerical model,
developed by the author at NDBC, is described, and its strengths and 2. OBSERVING NETWORK OPTIMIZATION
limitations discussed. An example numerical model case study is described NUMERICAL MODEL
and evaluated. The value to the planner of the model and its output is
also assessed. The numerical model OPTNET I was developed by the author, in the spring
of 1988 using the BASIC language on an IBM PC-AT. The model is bas-
ed on the linear programming technique derived from calculus of varia-
tions. The technique, which is described in [2] as well as other sources,
1. INTRODUCTION optimizes an objective function subject to constraints. In OPTNETI, the
objective function is the maximization of network "utility" or value for
One of the difficulties facing meteorologists, oceanographers, and other a network composed of multiple system types. The constraints may in-
geophysical scientists is the acquisition of field observations within the clude limits on the quantities of individual systems, initial and recurring
limits of available technology, the constraints imposed by the infrastruc- budget limits for individual systems, and intitial, recurring, and total (both
ture of logistics, maintenance, communications, and data quality present value and nondiscounted) expenditure limits for the network over
assurance, and the restrictions of budget. The issue of the optimal alloca- its life. Figure I depicts the components of the utility function that must
tion of resources is important for operational environmental observing be input separately for each system, as well as the input for system cost
programs as well as experiments, for both short- and long-term observa- data and system and network constraints.
tional projects, and for programs of local concern in addition to those
of global extent. The types of observing systems to, he operated, their The utility of an observing system may be evaluated from the perspec-
particular sensor capabilities, and the quantities of 'systems, etc., are tive of a numerical modeler (involved in operational numerical weather
several of the "questions" that must be resolved. Where observing systems or oceanographic predictions, typically on the synoptic and mesoscales);
of varying capability are already in place, such considerations may in- a forecaster (involved in issuing forecasts on the local or regional scales);
clude the deactivation of existing systems, the modification and enhance- and a climate researcher (involved in investigating climatic anomalies
ment of selected existing systems, and the installation of entirely new through modeling or statistical studies, primarily on the global scale).
equipment. The program allows the input of a weighting factor for each of the above
three user categories to reflect the objective of the observing network (e.g.,
Typically, the composition of an observing network is determined by in the example case that follows, primarily for forecasters and secondarily
negotiation between the various principals involved-the scientists, plan- for climate researchers). OPTNETI also requests for each user-system
ners, engineers, technicians, equipment suppliers, accountants, and, in combination the following:
some cases, politicians and lawyers. This process is complex and not in- I . Ideal coverage factors and corresponding weighting factors for meso,
frequently time-consuming, but does generally result in networks that synoptic, and global scales
achieve an acceptable level of the principals' objectives. In many situa- 2. The utilities and weighting factors for the following characteristics
tions, the negotiation of requirements, solutions, and budgets, despite eParameters sensed
its lack of analyticity, is essential to the success of the program. However, *Reporting frequency
under certain conditions, it is conceivable that analytical techniques could *Timeliness of reporting (the time interval from data acquisition to
be effectively employed to define the composition of optimum observing receipt by the user)
networks as an aid to the network planning process. These techniques, *Data quality (the adequacy of the data quality assurance methods)
which are derived from the field of operations research, can be very useful *Data continuity (the continuity of the observation from a given
to the planner, scientist, or engineer, who must either formulate and ex- location over the network life)
ecute the observing network program plan or recommend it to some ap- 9System performance (end-to-end reliability and data through-put).
proving authority. In general, optimization techniques are intended to The weighting factors enable each user to place different priorities on
supplement the planning and negotiation process, and not replace it the characteristics of each system.
altogether. Since optimization methods, such as linear programming, are
intended for computer application, they offer the network planner a For the linear programming method to work, the premise of linearity in
number of advantages, among which are: 1) the quantification of benefits an operations research sense must hold true. This means that economies
(as well as costs); 2) the ability to investigate a large number of candidate of scale either do not exist or are negligible with respect to both the costs
network configurations and assess the corresponding effects of changes and the utility of the systems in the network. In other words, the change
in requirements or budget (sensitivity analysis); 3) the capability to in network utility and costs must be invariant of the quantity of systems
analytically determine the optimum solutions in terms of benefits versus or show decreasing returns to scale. Similarly, the utility assessments must
costs; and 4) the ability to do 1, 2, and 3 above quickly and efficiently. remain functionally independent in order for the multi-attribute utility
Numerical optimization models also have limitations that, if not function to be developed. For example, the utility of the TIMELINESS

629 United States Government work not protected by copyright

SYSTEM NO budgetary and other constraints. The output includes the quantities of
the systems in the optimum network, a listing of the slack on the con-
straints (non-zero values where a constraint is not binding, i.e., where
that particular limit has not been reached), and the shadow prices on the
constraints (the amount the objective function would be improved if the
binding constraint or constraints were to be relaxed one unit).

3. EXAMPLE NETWORK ANALYSIS

An example analysis using OPTNETI was performed for an imaginary
data-void marine region in which surface observations were required. Four
??USER WEIGHT types of observing systems were assumed to be available: moored buoys,
FACTOR?? ??UWF?? F?? EUWF= 100 Coastal-Marine Automated Network (C-MAN) stations, drifting buoys,
and simplified, meteorological-parameters-only Shipboard Environmental
Data Acquisition Systems (SEAS). Costs, utility considerations, and con-
straints were assumed as shown in Table 1 (note the constraints on the
COVERAGE number of systems, which can be considered as either hardware availability
limits or other restrictions on the availability of suitable installation sites).
The input values for costs and utilities were estimated by the author and
??MESOSCALE ??SYNOPTIC @'@'GT"O'RA?.',
FACTOR?? FACTOR?? 1-100 EACH Table 1. Input Cons Iiderationsfor Example Marine Observing Network
Optimization Study
??WEIGHT?? EWEIGHTS=100
USERS AND WEIGHTING FACTORS
1. FORECASTER - 75
2. CLIMATE RESEARCHER - 25
SYSTEM CHARACTERISTICS 3. MODELER - 0
USER FORECASTER CLIMATE
RESEARCHER
??UTILITIES (1-100 EACH) FOR: ??CORRESPONDING COVERAGE WEIGHTING FACTOR
� SENSED PARAMETERS WEIGHTS E 100??
� REPORTING FREQUENCY MESO HIGH LOW
� TIMELINESS OF REPORTING SYNOPTIC MODERATE MODERATE
� DATA QUALITY GLOBAL VERY LOW HIGH
� DATA CONTINUITY SYSTEM CHARACTERISTICS
� SYSTEM PERFORMANCE?? WEIGHTING FACTORS
PARAMETERS HIGH HIGHEST
REPORTING FREQUENCY HIGH LOW
TIMELINESS HIGH LOWEST
DATA QUALITY MODERATE MODERATE
TI-TS9 DATA CONTINUITY LOWEST MODERATE
??e FIRST-YEAR COSTS PER UNIT I PERFORMANCE LOW I LOW I
RECURRING COSTS PER UNIT?

SYSTEM CONST AINTS SYSTEM TYPE MOORED DRIFTING SEAS* C-MAN
BUOY BUOY
??* MAXIMUM QUANTITY OF SYSTEMS COVERAGE
9 TOTAL INITIAL COST FOR SYSTEMS OF TYPE MESO HIGH MODERATE LOW MODERATE
e TOTAL RECURRING COSTS FOR SYSTEMS OF TYPE?? SYNOPTIC HIGH MODERATE MODERATE MODERATE
GLOBAL GOOD MODERATE MODERATE I MODERATE
- - - - - - - - - - - - UTILITIES
ETWORK CONSTRAINTS PARAMETERS HIGH MODERATE VERY LOW HIGH
REPORTING FREQUENCY HIGH LOW MODERATE HIGH
TIMELINESS HIGH MODERATE GOOD HIGH
??* NETWORK LIFE DATA QUALITY HIGH MODERATE VERY LOW HIGH
� COST OF MONEY DATA CONTINUITY HIGH LOW VERY LOW HIGH
� NETWORK INITIAL COST CONSTRAINT PERFORMANCE GOOD MODERATE, GOOD HIGH
� NETWORK RECURRING COST CONSTRAINT SYSTEM UTILITY
� NETWORK TOTAL LIFETIME EXPENDITURE CONSTRAINT 79.5 37.7 20.1 46.8
� NETWORK TOTAL PRESENT VALUE CONSTRAINT?? (NONDIMENSIONAL)
COSTS,$K
NOTE: INPUT INDICATED BY INITIAL 50 15 20 35
RECURRING 25 15 5 12.5
Figure.l. OPTNET Input CONSTRAINT ON to 40 25 1-7
@,GUMTITY-OF-SYSTEMS
of an observation is assumed to be unaffected by the utility of any other
system characteristic, such as REPORTING FREQUENCY or DATA
QUALITY. NETWORK LIFE - 5 YEARS
NETWORK TOTAL EXPENDITURE CONSTRAINT - $5,000,000
Based on the utility, cost, and constraint input, OPTNET 1 solves the *HYPOTHETICAL SEAS EQUIPMENT, NOT NECESSARILY SIMILAR IN
linear programming problem of maximizing network utility subject to CAPABILITY OR COST TO NATIONAL OCEAN SERVICE SEAS SYSTEMS.

630

do not necessarily represent expert judgment-rathe'r, they are believed 4. CONCLUSION
to be generally representative values intended for an example case study. An analysis of an example marine surface observing n6twork using a linear
The results are shown in Figure 2, where the quantities of systems in the programming optimization numerical model has been conducted. The
optimum network are plotted against the initial year budget constraint analysis demonstrates the capability and effectiveness of the linear pro-
(for a fixed 5-year total expenditure of $5,000,000 and equal recurring gramming technique for assessing alternate network configurations and
costs). All systems are assumed to be installed in the first year of net- determining optimum network composition. The actual run time for the
work operation. Drifting buoys are incorporated into the network at the numerical model was on the order of one to two days, including prepara-
lowest levels of network initial cost constraint. In effect, they have the tion of all input data. The complexity of the problem, even with only
highest ratio of utility to initial cost, and the numerical model selects their four system types and several constraints, is such that solution without
implementation where network initial costs are severely constrained. a numerical model would be very inefficient.
Moored buoys are installed next as the network initial cost constraint is
raised, followed by C-MAN stations and SEAS. (Note that at an initial The use of a numerical model such as OPTNETI cannot take the place
cost = $1,000,000, the network initial and recurring costs are equal; it of thorough planning and negotiation of requirements. Nevertheless, the
is unlikely that a network would be implemented at an initial cost level value of this technique, if appropriately and judiciously employed, can
less than this value.) Below an initial cost constraint level of just over -be of significant benefit. While requiring care in the formulation of the
$1,300,000, the total expenditure constraint of $5,000,000 (first year plus problem and judgment in the interpretation of results, the linear program-
four recurring years), is not reached. Between an initial cost of $1,300,000 ming numerical model approach enables a network planner to quickly
and $1,626,000, both the inital cost and total expenditure constraints are and efficiently investigate a broad spectrum of network possibilities and
binding. SEAS units, with their low recurring costs, are introduced at to quantify their relative merits.
the expense of drifting buoys for this range of initial cost constraint. The
network overall utility, having reached a maximum at an initial cost con-
straint of $1,300,000, remains constant as the initial cost constraint is 5.REFERENCES
raised. Above a network initial cost constraint of $1,626,000, all quan-
tities of systems are frozen, since the network total expenditure constraint I ."Evaluation and Upgrading Strategies for the Marine Monitoring
has been reached, but all the initial budget cannot be spent. System," Ocean Data Systems, Inc./Global Weather Dynamics, Inc.,
January 1986.
As a last item of note, the optimum of optimum network configurations,
given that all the budget can be spent, occurs at an initial cost constraint 2. deNeufville, Richard and Joseph H. Stafford, Systems Analysis for
equal to or greater than slightly above $1,300,000. This point is where Engineers and Managers, McGraw-Hill Book Company, New York,
the maximum network overall utility,is first obtained. 1971.

DRIFTING BUOYS
40-- 4.0

30-- --3.0

QUANTITY SLACK ON
OF SYSTEMS BUDGET
CONSTRAINTS,
$M
20 --2.0

NETWORK
UTILITY x 10-2

10-
1.0

C-MAN

W, \\N\
r"i
0 0
0 0.5 1.0 1.5 2.0

NETWORK INITIAL COST CONSTRAINT, $M
Figure 2. Optimum Marine Observing Networks for 5- Year Life and Total Expense $5M

631

AN USER-FRIENDLY MULTI-FUNCTIONAL CTD SOFTWARE PACKAGE

Woody C. Sutherland

Duke/UNC Oceanographic Consortium
Duke University Marine Lab
Beaufort, NC 28516-9721

ABSTRACT The SMART CTD has been upgraded by Neil Brown
Instrument Systems to allow interfacing of
An in-house software package was developed on additional sensors. A Sea Tech fluorometer and a
a Hewlett Packard 9816 microcomputer for CTD Sea Tech transmissometer are currently used in CTD
operations aboard the R/V CAPE HATTERAS. The operations aboard the R/V CAPE HATTERAS. The
package includes routines for data collection, analog signals from these instruments are
data storage, and tabular and graphical display. digitized inside the SMART and the data
Pre-cast setup of the software permits the multiplexed with the conductivity, temperature and
scientist to select from a number of parameters to pressure values. The data string transmitted up
customize the operation to meet his/her individual the sea cable then contains five measurements.
requirements. Post-cast routines read stored data
from discs, perform tabular print-outs and hard The Hewlett Packard 9816 is based on a 8 MHz
copy graphical display, and transfer of data to Motorola MC 68000 16-bit microprocessor with 32-
IBM-compatible mass storage media. bit internal architecture. The system on which
the CTD software was developed and runs has a 400
Presently used with a Neil Brown SMART CTD x 300 resolution monochrome monitor, 3/4 megabyte
having a sampling rate of approximately 4 Hz, the RAM, a 15 megabyte hard disk which is used for
software/computer combination is able to collect permanent storage of the CTD software and
all available data. Use of faster sampling CTD's, temporary storage of CTD data, and a 720 K byte 3
e.g. NBIS Mark III and SeaBird, would probably 1/2" double sided floppy disk for the final
require a subsampling of available data. Written storage of CTD data. Output devices are an HP
in Hewlett Packard Pascal and utilizing a modular ThinkJet printer and an HP 7475A six-pen graphics
design, the software can be easily ported to plotter.
other, possibly faster, computers. This will soon
be done for the IBM AT class of microcomputers. The software was written in HP Pascal version
3.0, a superset of American National Standards
Institute (ANSI) Pascal. Due to the structured
nature of the Pascal programming language and
INTRODUCTION extensive use of procedures, functions and modules
(HP terminology) the CTD software package can be
An integrated software package was developed ported easily to other computer systems having a
on a Hewlett Packard 9816 microcomputer to Pascal compiler. This is being done for the IBM
collect, display and store oceanographic data from AT class of microcomputers at the request of a
a Neil Brown Instrument Systems SMART CTD. The number of regular ship users.
SMART CTD, operating in raw data continuous mode,
outputs conductivity, temperature and pressure INITIALIZATION
measurements at a rate of approximately 4 samples
per second. This data string is transmitted via Upon running the main program the user is
frequency shift keying (FSK) communications greeted with a CRT screen listing the version of
through a two-conductor sea cable to a shipboard the software, presently 3.0, and requesting a
mod ul ator/demodulator unit. The modulator/ choice of CTD instruments. Duke / University of
demodulator unit translates the FSK signal to a North Carolina Oceanographic Consortium (DUNCOC),
RS-232 signal which is input into the which operates the Research Vessel CAPE HATTERAS,
microcomputer. Data communications are at 1200 owns two Neil Brown Instrument Systems SMART
baud. The raw data signal from the FSK unit is CTD's, designated "A" and "B". The SMART CTD is
recorded on audio cassette as a backup in the designed to have the calibration coefficients for
event of microcomputer failure. the conductivity, temperature and pressure
channels stored on EPROM inside the instrument.

CH2585-8188/0000- 632 $1 @1988 IEEE

As DUNCOC performs its own calibrations it is more most frequently used. The data are averaged over
convenient to have the coefficients stored in a one-half meter depth intervals and are catalogued
disk file to be read by the main program. Thus, as the deeper depth. Thus, data collected between
when the user selects one of the CTD's, a 1.01 m and 1.50 m depth are averaged and stored on
specified file on the hard disk is opened and the disk as 1.5 m depth data.
following variables are read into the main
program: The next screen prompts for station
a) Letter designation of CTD statistics. These are: a) Cruise number, b)
b) Serial Number of CTD Station name, c) Position, i.e. latitude,
c) Last date of calibration longitude, and time differences if Loran C used as
d) Quadratic calibration coefficients for positioning device, d) Log of cassette tape, i.e.
conductivity, temperature and pressure Tape number and side, and Meter wheel reading for
e) Calibration coefficients for beginning of data recording. Values from the
fluorescence and light transmission previous station are listed and only those
f) Disk volume where to temporarily different need be changed. The position
store CTD data statistics can be entered via the keyboard or an
g) Disk volume where to permanently interface with the SAIL (Serial ASCII
store CTD data Instrumentation Loop) can supply the digital data
The user is then presented with a screen directly from the ship's Loran or Global
displaying this information so that he/she can Positioning System unit. The user is prompted to
Verify the selection of the correct CTD in use and verify the correctness of the entries before being
examine the date of last calibration. allowed to continue.

The next screen prompts for verification of The file name for the CTD data is constructed
use of additional sensors on the CTD underwater from the station name and cruise number. Cruises
package. The SMART is mounted on a frame, usually on board the R/V CAPE HATTERAS are numbered from
with a General Oceanics rosette assembly for water the beginning of the calender year. For example,
bottles. Frequently a pinger is also attached to CH0588 signifies the fifth cruise of 1988, the
accurately measure altitude above the sea floor. beginning two letters standing for the name of the
When desired, the Sea Tech fluorometer and ship. The file name is the station name followed
transmissometer are mounted to the underwater by an underline, the last digit of the year and
package. The use of these two additional sensors the two digits of the number of cruise for that
is queried because it affects the length of the year. Thus, CTD data for a station named "PRIMO"
data string and the need for calibration of their on cruise number "CH0288" would be stored in a
measurements. Also, the fluorometer has an file named "PRIMO 802". When the user answers YES
internal switch which changes the range of its to the correctness of the station statistics, the
readings, and thus the calibration coefficients. disk volume where the final copy of the CTD data
The setting of this switch must be entered into will be stored is checked to see that there is not
the CTD software so that the proper coefficients an existing file with the same name. If one is
are used. present then the station statistics screen is
again displayed with a message that a file with
Seven columns of data can be displayed on the that name already exists. The user is prompted to
CRT during the lowering of the CTD. The first enter a different station name and then must again
column is reserved for depth in meters. The user verify the values. This loop continues until an
selects the next five column designations from unique file name is formed. This precaution
conductivity, temperature, salinity, sigma-t, guards against the overwriting of a file and loss
fluorescence and light transmission. The values of CTD data.
for salinity and sigma-t are calculated from
conductivity, temperature and pressure A sea-surface measurement is taken to zero
measurements according to Fofonoff and Miller the CTD pressure data, thereby eliminating the
(1983). Either time of day (hour: minute: second) need to physically zero the pressure transducer
or lowering rate of the underwater package which can be affected by fluctuations in
(meters/minute) is displayed in the final column, atmospheric pressure. This value is entered into
by selection of the user. the CTD software as a pressure offset and is
subsequently subtracted from all further pressure
A real-time graphical display of the data can readings for the lowering to yield a more accurate
be performed in addition to the tabular Output. depth calculation. This offset can be entered by
If the graphical display is selected, the user the user via keyboard or directly from the CTD.
must choose a property to display, e.g. salinity, The preferred method is to lower the CTD over the
temperature, etc., and enter parameters with which side of the ship and hang it at the water's
to scale the graph, i.e. maximum depth of CTD surface. Pressure readings from the CTD are
lowering and minimum and maximum expected values displayed on the computer terminal. The user
of the property. presses a key to accept a value for the offset.
Hanging the dTD underwater package at the surface
All of the valid data can be stored on disk is not practical in rough seas as it may be
(raw mode) or the user can select to store damaged by striking the ship. Thus the option of
averaged data. Since raw data are available from entering a pressure offset via the keyboard is
the cassette recording and disk storage space given.
usually needs to be conserved, the average mode is

633

CTD LOWERING data was selected then values for these properties
which passed the three levels of filtering are
With input of the pressure offset the CTD written to the disk file. In average mode,
underwater package is ready for lowering. A brief filtered values for these six properties are
reminder to start the cassette for recording of averaged over one-half meter depth intervals and
the raw data signal is flashed on the screen, a the mean values written to disk.
number of program variables are initialized, files
are opened on the appropriate disk volumes, The collection, calibration, filtering,
routines are called to prepare the real-time display and storage of CTD data continues until a
graphical display and the bottom line of the CRT designated key is struck by the user. At that
is labeled with the names of the properties time the data file is closed and saved, the
selected to be displayed. The winch operator is graphic routines are terminated and a reminder to
requested to begin lowering the CTD. stop the cassette recorder is flashed on the
screen. A menu is displayed giving the user the
Each string of multiplexed data sent to the option to: 1) begin raising the CTD package, 2)
HP 9816 from the modulator/demodulator'utiit passes plot CTD data on the CRT, 3) plot CTD data on the
through three levels of filtering routines before HP pen plotter, 4) print CTD data, 5) transfer
being accepted as valid data. First, the string data to disk volume for permanent storage or 6)
is checked for the proper number of characters and terminate the program.
the proper sequence of numeric characters and Output of data to devices other than the CRT
commas. Values for the individual properties are time consuming and these options are given
conductivity, temperature, 'pressure, and if again once the CTD is back aboard ship. Thus
appropriate fluorescence and transmission, are their operations will be discussed later. They
extracted from the string. These values are are practical only if a hard copy of the data is
entered into their respective calibration needed before raising the CTD or if the CTD needs
equations. The pressure offset is subtracted from to hang at depth, e.g. to allow reversing
the calibrated pressure value. The second level thermometers to come to equilibrium. Graphs can
of filtering checks these calibrated values be quickly repeated on the CRT, although with only
against the published minimum and maximum for each one property at a time, to aid the user in making
instrument. Each property is finally compared decisions on where to trip water bottles on the
with its last valid value to confirm that return to the surface. The first graph on the CRT
empirically predetermined differences are not for each property selected extends from the
exceeded. surface to the maximum depth and is scaled from
the minimum to the maximum value collected.
Derived properties, e.g. depth, salinity, Following this first graph the user is queried if
sigma-t, are calculated from the accepted values he/she desires to change the depth limits in order
of conductivity, temperature and pressure. to zoom in on a portion of the graph. If the user
Minimum and maximum values for each property, so chooses, he/she is prompted to enter the
which are used in post-lowering routines to scale minimum and maximum depths for the new graph. The
graphical presentations, are updated with the graph is redrawn with the new depth scale and a
current values. Values of depth and the six new property scale with the minimum and maximum
selected properties are formatted and displayed on values collected within this depth interval.
the CRT. The lowering rate is calculated every
five seconds ; it was discovered that ship motion Striking the key to begin raising of the CTD
from surface sea state had too much effect on the brings up the tabular display screen. The bottom
calculated lowering rate using shorter time line is labeled with the names of the selected
intervals. properties and input from the CTD is accepted.
Data collected during the raising of the CTD is
If the real-time graph was chosen then the not stored; values displayed pass through only the
CRT screen can be switched from tabular to first filtering process. A key is activated so
graphical display by striking a key. The that when pressed a routine for closing a water
corresponding depth of the selected property must bottle is begun.
be greater than that of the previous value before
it is graphed, thus yielding a smoother curve. The CTD is raised through the water column
Graphing of multiple properties was attempted in until a depth where the user wishes to collect
previous versions of the software, but the only water is reached. CTD ascent is halted, the
method of distinguishing between properties was by designated key is pressed on the computer
different line types and this proved visually keyboard, and a button is pressed on the General
confusing on the HP 9816 monochrome monitor. if Oceanics rosette shipboard unit. Program
the depth of the CTD exceeds that originally execution is halted while the rosette unit
entered then an additional 100 m is added and the provides a high voltage signal to the underwater
graph rescaled, thereby facilitating a search for unit activating the closure of one of the water
an interface which may be deeper than expected. bottles. The delay in the program is preset to
The addition of 100 m increments will continue as match the time necessary to trigger the rosette
needed until stopped by the user. underwater unit. Following the delay the computer
inputs CTD data for one minute, averages this
Data is stored in a disk file as a series of data, displays this information on the CRT
real numbers: depth, conductivity, temperature, indicating that a water bottle has been closed,
pressure, fluorescence and transmission. If raw and writes the data to a file which will include

634

data on all bottles closed during the raising. fluorescence, and light transmission may be
The program continues to collect CTD data and selected for vertical profiles on each graph.
display it on the CRT until the water bottle key Temperature vs salinity graphs are also an option.
is struck again or a key assigned to the end of Graphs are labeled with the cruise number and
the raising is pressed. station name, and scaled with the minimum and
maximum collected for each property.
When the CTD reaches the sea surface, or the
user is no longer interested in watching the data The user is returned to the final menu
collected, a key is pressed to terminate the following the completion of each output.
raising CTD routines. The file containing the Therefore, he/she can produce as many print-outs
water bottle data is closed and saved, and a copy or graphs as desired, in any order, and then
is sent to the printer. The user is given the choose to begin another station with the same
opportunity to enter a comment for the station and parameters.
is prompted to enter the cassette meter wheel
reading for the end of the raw data recording. SOFTWARE FOR STORED CTD DATA
The file for permanent storage of the CTD data is
opened as an ASCII text file. The station An independent portion of the CTD software
information, i.e. cruise number, station name, package is designed to operate on the permanently
date and time, position, cassette log and comment, stored files. A separate HP 9816 microcomputer
is written to the file. CTD data from the system is available on the R/V CAPE HATTERAS to
lowering are translated into strings of depth, permit scientists to reexamine data from previous
conductivity, temperature, pressure, fluorescence CTD 'stations, or the primary system can be used
and light transmission and written to the storage while not actively collecting data. The opening
file. The permanent disk volume is most often a menu gives options for: 1) tabular or graphical
floppy disk. Should the disk fill up during the output from single -CTD stations, 2) graphical
writing of the data the file is closed and the output of one property from multiple stations, 3)
user is prompted to insert another formatted disk assimilation of statistics from a number of
to continue. A file is opened on the new floppy, stations, and 4) output ASCII data file via RS-232
the station information is written to the interface.
beginning of the file, and transfer of the CTD
data begins from where it was discontinued in the Selection of the first option prompts for
previous file. Data for the water bottles is keyboard input of a CTD data file name. The data
written to a . permanent file on the same disk file is opened, the station statistics read and
volume following transfer of the CTD lowering displayed on the CRT. The user confirms that this
data. is the desired station by virtue of the display
and, if so, the remainder of the data file is read
The final screen gives the options of into memory. Choices of printed tabular output
graphical display on the CRT or pen plotter, and graphical output to CRT or pen plotter are
printed tabular output, initiation of another CTD presented. These routines perform as discussed
station with the same parameters, or termination previously. Following each output the user may
of the program. If the user chooses to begin choose to return to the main menu.
another station, the program loops back to the
station statistics screen, continues to the The second option first asks for the number
pressure offset, and is ready for lowering of the of stations. Physical size of the graph paper and
CTD package. Values for the real-time graph, number of pens, i.e. colors, limits this number to
selection of properties for display and the mode five. The sequence of prompting for file name,
of data storage remain as for the previous displaying station statistics and requiring
station. Termination of the program returns the verification is executed for each of the stations.
computer to the HP Pascal operating system. The user must then choose one of the seven
properties to plot ( conductivity, temperature,
Printing of the CTD data requires the user to salinity, sigma-t, sound velocity, fluorescence,
choose an interval over which to average the light transmission ), the size of paper to use,
output. The data may be printed as stored, either and the scaling parameters for the graph ( maximum
raw or averaged over 0.5 m, or averaged over a depth, minimum and maximum property values ). The
larger interval as is frequently done for deeper graph is drawn on the plotter using a different
lowerings. Station statistics are printed at the pen for each station.
top of odd numbered pages. Columns are labeled
from left to right as depth, conductivity, A summary file composed of the station
temperature, salinity, sigma-t, sound velocity, statistics for all CTD stations is made following
light transmission, and fluorescence. Data is each cruise. These are the files that are
averaged as necessary, the derived properties searched when a scientist requests CTD data from a
calculated, and printed in the respective columns. specific location or time. The third option
Multiple copies are possible in case more than one permits the user to make such a file for
scientist requires the data. him/herself as a summary for CTD data collected
during a day, week, entire cruise, etc. He/she
Graphs of the data can be produced on 8 1/2" can choose to send the output of this summary to
X 11" or 11" X 14" paper. A maximum of four the CRT, where only a screen-full at a time is
properties from the list of conductivity, displayed, to the printer or to a text file on
temperature, salinity, sigma-t, sound velocity, disk.

635

More ship users have access to IBM-compatible
than to Hewlett Packard microcomputers. The last
option provides CTD data on IBM compatible mass
storage media. A complementary program must be
executed on one of the IBM compatibles on board
ship. Data is transmitted from the HP 9816
through the serial port to the IBM microcomputer.
The user can select to send the data file by file
or an entire disk of files at a time. Transfer
occurs at 9600 baud using an Xmodem checksum
protocol.

CONCLUSIONS

As an extension of the concept of the R/V
CAPE HATTERAS being an user-friendly ship, CTD
software was developed to permit the scientist to
fashion CTD operations to meet his/her
specifications. It has proven invaluable to
scientists as an aid in determining sample schemes
while at sea. This is an ongoing process and
updates to the CTD software package will continue
to be made as use and need dictate.

ACKNOWLEDGEMENTS

Support for development of the CTD software
was provided by the National Science Foundation
Shipboard Technicians Program, grant number
OCE8615924. Thanks are given to the electronics
technician, Tim Boynton, the head of the DUNCOC
CTD committee, Dr. John Morrison, and the many
ship users over the past three years for their
constructive criticism and helpful suggestions
concerning the CTD operations.

REFERENCES

Fo f ono f f , N. P. and R. C. Millard Jr. (1983)
Algorithms for computation of fundamental
properties of seawater. Unesco technical
papers in marine science No. 44.

636

ONLINE ACCESS TO NODC INFORMATION SERVICES

Douglas Hamilton and Janet Ward

National Oceanographic Data Center
National Environmental Satellite, Data, and Information Service
National Oceanic and Atmospheric Administration
Washington, D. C. 20235

ABSTRACT minimize the time and personnel needed to handle
WOCE data and to maintain up-to-date knowledge
Oceanographic data holdings of the National of the status of WOCE data.,,5 In view of those
Oceanographic Data Center (NODC) are large and needs, NODC developed a prototype on-line
varied. A goal at NODC is to provide timely information system called NOSIE - the NODC Ocean
information about those data to users via online Science Information Exchange.
caqmter aooess. Recently, NODC has developed a
prototype system, called NOSIE (NODC Ocean
Science Information Ewliange), that includes THERMAL DATA
descriptions of NODC data files, helps for
sending or ordering data, interactive inventory GLOBAL DISTRIBUTION
sLmmaries, and bulletin boards. NOSIE is a
mem-driven, modular system to which information 120
resources of various types can easily be added. N@
It was designed using system software for screen 100 ----- ------ ------------------------ ----- - -------- ---------------
management to provide a consistent look-and-feel WT
during each session. Future plans am to expand BD ---------------*--------------- ----- I---------- &19=
inventory smnaries, add detailed inventories, 60 -------- ---------------- ... ......... XBT
and make subsets of ocean files or special data STD
sets directly accessible online. 40 -------- --------
SBT
20 --------------

1. INTMDUCTION 0
194) 1%0 1W
19T 1970
Oceanographic data on water column properties, COLIEC110N YEAR
chenistry, biology, surface waves, and other
parameters are available frarithe National Figure 1. Time distribution of thermal data
Oceanographic Data Center (NODC). Data held by collected by various instnments
NODC span the globe and cover the time period
from 1900 to the present, although most data
were obtained after 1940. For exanple, Figure I Table 1. Major NODC data holdings, January 1988
shows the volume of one type of NODC data -
terperature observations - as a function of
time, collected with various instruments aver DISCIPLINE VOUME
(GBytes)
the years since 1940. A summary of the types -
and volumes of data available from NODC is shown Physical
Buoy (Wind/wave) . . . . . . ... 3.95
in Table 1. The person who wants to order data Current . . . . . . . . . . . . 2.45
from NODC needs to know if the type of data Ocean Station . . . . . . . . . 1.25
wanted are available in the region and time subsurface Temperature . . . . . 0.67
salinity/Temperature/Depth . . . 0.83
period of interest. Therefore, data requestors Other . . . . 0-06
need access to inventories and information about Biological T@TiL'Pff*YS*IC*AL' 9.20
NODC data holdings. Fish/Shellfish ... . . . . . . . 0.12
Benthic/Intertidal Organisms. . 0.09
One of NODC's principal goals is to improve ease ma ine Birds . . . . . . . . . . 0.05
Plankt n . . . . . . . . . . . . 0.03
of access to data and to data inventories. other 0 . . . . . . . 0.03
Those who plan ocean research projects such as T;T@L BIOLOGICAL 0.32
the World ocean Circulation Experiment (WOM) 2 Chemical
Marine chemistry. 0.07
and the Global Ocean Flux Study (GOFS)l have P.llutants/Toxic substances 0.02
TOTAL CHEKICAL 0.09
stated the need for timely access to ocean data.
"Maxiimm use of online summaries, electronic TOTAL DATA BASE . . . . . . 9.61
queries, and electronic data transfer will

637

It became technically possible to develop NOSIE NOSli Menu Structure
in mid-1987, when NODC became a node on the (Top three /a vels)
Space Physics Analysis Network (SPAN), managed
by the National Aeronautics and Space
Aftinistxaticn (VASA). Major oceanographic V@TOSIE
centers connected to SPAN (Figure 2) ommmicate
messages, data, and information about data MENU
directly Unvugh the network. Although NOSIE
resides on an NODC caqmter, anyone with access
to SPAN can tap into the NOSIE information Data Catalogs/ U or
resource. In addition, those with no SPAN General SubmIssion Inventories SeZes,
connection can dial directly into the system.
It is considered by NODC to be the "electronic"
window to information about our ocearKx1raphic
data holdings.

d- les
SPAN Ocean Nodes
Buletin
boards

AOSU
Hot Or&* user
and request
publkations B--V*

UD"-r.
ANSI
ODC
JCSB

JP
SI
yet aveftable; module
UT-i ATAM N FSU under development.
Figure 3. Menu structure of the NODC Science
Infonration Rwhange Syst@

oceans and seas over a time period of several
decades. Inventory summary searches provide the
number of stations available for geographic and
Figure 2. Ocean centers on the Space Physics tire criteria set by the user. The resolution
Analysis Network for searches is month for time, and one-degree
square for geographic area.

b. File Descriptions
2. NOSIE STRUCnM AND OONTMS
For those who need an introduction to the types
NOSIE consists of a series of menus which lead of data stored at NODC, brief text descriptions
to: other merms, to text information, to of NODC oceanographic data files are available
bulletin boards, or to interactive inventory from the Catalogs/Inventories menu. As shown in
summaries. By navigating among appropriate menu Figure 4, users may find the time period and
choices, users can quickly obtain the geographic area coverage of data in a file. The
information desired. The first three menu file size, file structure, and principal
levels are shown in Figure 3. paranve-ters available are also presented
in these file descriptions.
a. Online Inventories
c. Assistance in ordering Data
Someone who wants to order data fran a
particular region of the world, and perhaps from After reading a file description, and checking
a specific time period, will want to know how the inventory summary for amounts of data
many data are available. Because NODC data available, a user may decide to ask for more
holdings are so large and varied, it was not information or order data. nie User Services
practical to provide complete, detailed menu leads to text screens which give ordering
inventories with this prototype. However, procedures. In that section, guidelines for
inventory smmaries are provided for four major ordering data from NODC are presented, and a
data files - CID/STD (conductivity/salinity contact person and phone number are shown.
tenverature depth) data, oceanographic station other segments give information on user charges
data, and expendable and mechanical and how payments can be made. If more
bathythermograph data. Data in each of these information is wanted, a single key-stroke adds
files are available from roost of the earth's the name of the file being reviewed to a

638

Fite: Expendable Bathythermograph (XBT) Data FILE NO. OF NO. OF NO. OF
DATA TYPE ALIAS SOURCE STATIONS RECORDS TAPES
Geographic Coverage: Worldwide Oceans Ocean Stations CIOO MIAS UK; 549 8,321 -
Time Period: 1966 to Present NOAA AOML;
NMFS woods
Fite Organization: cruise and Geographic Order Hole;
Lamont (1)
Fite Size- 718,181 stations; 220 megabytes (July 1, 1988) Analog XBTs C116 usGs woods 84 84
Cruise file - 3 magnetic tapes (6250 bpi) Hole (2);
Geofile - 5 magnetic tapes (6250 bpi) US Navy
Description: Digital XBTs C116 NOS Seattle; 1,497 1,497 1
NMFS@
This file contains temperature-depth profile data obtained using Narragansett
the expendable bathythermograph (XRT) instrument. Standard XSTs (3); PRC (4);
can obtain profiles to depths of either 450 or 760 meters, UK Navy
depending on the model. With special instruments, measurements C/STD F022 NOAA PMEL (4); 2,052 76,696
can be obtained to 1830 meters. Cruise information, position, OR State Univ.
date, and time are reported for each observation. The data record (8); Lamont
comprises pairs of temperature-depth values. Unlike the MST data (1); NMFS Woods
file, in which temperature values are recorded at uniform 5 meter Hole; Univ. AK;
intervals, the XBT Data FiLe contains temperature values at non- NMFS La Jolts
uniform depths. These depths are at a minimum number of points (9)
("inflection points") required to record the temperature curve to Wind/Wave F191 NOAA NDBC 87 449,786
an acceptable degree of accuracy. On output, however, the user may Spectra
request temperature values either at inflection points or [GOSS STD Lev. A NOAA OPC - -
interpolated to uniform depth increments. Sea Level Lev. A Univ. Hi (5) 24 3,043
Chemistry/ Lev. A WHOI 311 7,153
Plankton
Note: Although new XBT instruments can provide digital output on GEOSAT Lev. A US Navy 1.7M 1.7M
magnetic tape, NODC stilt receives paper strip charts with analog (Wind/Wave)
traces that must be digitized. only part of the XBT data that it
archives are digitized by NODC itself. A substantiaL amount of
data are digitized by the U.S. Navy Fleet Numerical Oceanography PROJECTS: (1) Marathon - 7
Center (FNOC), Monterey, Calif.; some data are also digitized by (2) MMS - New England Shetf/Slope
commerical contractors. (3) SOOP
(4) US-PRC Cooperative Study
Fite Structure: (5) Tropical Ocean Global Atmosphere (TOGA)
(6) STACS and Antilles Current
One variabte-tength record (maximum 2,540 characters). (7) EPOCS and Puget Sound LREP
(8) Peru Currents Project
Note: Although they are maintained in separate files, mechanfcaL (9) CaICOFI
bathythermograph (MST) data and expendable bathythermograph (XBT)
data are stored in a common format. Figure 5. Sample 'Taomt Data Acquisition" list

PRINCIPAL PARAMETERS
Depth 3. ACCESS TO NOSIE
Temperature
Figure 4. Sairple file description from NOSIE NOSIE is available in two ways. For those with
access to SPAN, sinply type "SET HOST NODC!", and
type "INFO" at the prompt. For those who have a
"shopping list". At the end of a session, the modem but cannot CoiMeCt to SPAN, dial (202)
shopping list can be reviewed and changed. At 673-5657 (300 baud), or (202) 673-5665 (1200
NODC an information specialist reviews the baud). When the modem is connected, type a
shopping list arid sends the information carriage return. When 11 appears,
requested. type 11C NODC", then "INFO". The opening NOSIE
screen provides instructions for the session.
d. Assistance in Sending Data
4. SYSTEM DESIGN FEATURES
Helps for sending data to NODC are available
from the Data Submission menu. Text screens NOSIE was designed and developed to operate on a
provide guidelines on the types of data to send, Digital Equipment Corporation (DEC) VAX VNS
computer media and suggested data formats, and capiter. In order to meet both present arid
type of documentation to send with the data. A future needs, NODC developed NOSIE based on two
contact arid Phone number for more direct primary design considerations.
assistance are also provided. First, the system would be required to allow the
e. Bulletin Boards addition of new information and interfaces to
other systems without major modifications to the
Bulletin boards of up-to-data information are source code. In order to accoanplish this, a
accessible through the General menu. Lists of modular design featuring a main FoRrRAN program
data sets recently acqaired are updated weekly and subroutines was adopted. As new
(Figure 5)_ Also, special NODC data products, requirerents warrant, subroutine modules can be
such as data sets from the SEQUAT4/FOCAL Project, added, modified, or deleted from NOSIE without a
are advertised in the Recent Data Products significant in-pact on the system's overall
bulletin board. operation.

639

Second, NOSIE was designed to be easy to operate Of oceanographIc data by coordinating with other
(user-friendly) and easy to maintain. To meet on-line systems. For example, standards in
these requirements, NOSIE employs the DEC's session presentation and control could reduce
Terminal Data Management System (TEM) and the confusion to users of several systems. At the
Ommun Data Dictionary (CDD). very least, NOSIE will include pointers to other
The TEM package is used in two ways: (1) to ocean information systems.
display the special interactive menu screens NOSIE is a prototype. It is an experiment by
during a NOSIE session and automatically check NODC to learn if having timely access to data
user-entered responses, and (2) at the VMS level and to information about NODC and its archives
to create and modify the interactive screens. is beneficial to users. You are encouraged to
The CM package is used to store and manage data try NOSIE, to tell us what works and what
record descriptions. doesn't, and to think about other ocean data
services you would like to see available at your
5. FLTILM PIANS computer terminal in an online mode.
At the end of each NOSIE session, users have the 6. REFERENCES
opportunity to enter comments or questions.
Those messages are checked daily at NODC in 1. Collins, E. and K. Hughes. Report of Global
order to respond to requests for data or Ocean Flux Study (GOFS) data management
information. In addition, comments about the meeting, Massachussetts Institute of
system are solicited to guide its future Technology, February 1987.
develcprent. As a result of these comments,
NODC is working on expanding and improving NOSIE 2. Emery, W. Global Circulation Studies: The
in several ways. need for data submission and exchange in
Inventory summaries of other NODC files are wom, Worm NEWSLETTER. 4, WOCE International
PlwudM Office, Godalming, Surrey, UK. May
planned. Also, the possibility is being 1987. p12-13.
considered of providing graphic charts to assist
users in setting inventory search criteria and 3. Hewitt, R., S. Jacobson and C. Meyer.
to show results of the search. CalCOFI On-line Data System Programmer's
I Manual, Administrative Report IJ-88-03,
Some data search questions require more detailed National Marine Fisheries Service, la Jolla,
answers than are given by the inventory CA, 1988.
sunmaries. Although it is not possible to
provide detailed inventories of all NODC data 4. Soreide, N. and S. P. Hayes. EPIC: An
files through NOSIE, such inventories are , Oceanographic Data Archival and Retrieval
planned for selected subsets of the archive or System, Proceedings of the Fourth Working
for data sets of special interest. These will Symposium on Oceanographic Data Systems,
give users the ability to search, for exwple, Computer Society, Washington, D.C.,
for specific measured parameter s, for project 1986.
data, for data from specific ships, and other
criteria. 5. U.S. Science Steerirxj Committee for WOM.
U.S. WOCE nVIEWMATION PLAN, First Draft,
Within existing computer system limitations, it U.S. Planning Office for WOCE, Texas A&M
is planned to place selected subsets or special University, College Station, TX, 1988.
data sets online for users to peruse or upload
to their own caqwter system. As planned 6. Withes, W. and D. Hamilton. Opportunities
projects such as WOCE, GOFS, and others get in Oceanographic Science Offered by New
underway, this is expected to be a quick way for Advances in Data Management, Proceedings of
researchers to have access to small data sets of the Law of the Sea Conference, University of
interest. Larger data sets will be distributed Rhode Island, June 1988.
on conventional magnetic media, and an optical
CD-RCM.6 NODC is now preparing its first ocean
data CD-RCK product to test the feasibility of
distributing very large data.

NOSIE is one of several on-line ocean
information system. Within NCAA, for example,
the Pacific Marine Environmental L-Amratory
developed EPIC, which provides inventories of
and access to data stored at that facility.4
The California Cooperative Fisheries
Inventigations project has provided on-lim
access to its data for several years.3 More
recently, the University of Delaware developed
OCEANIC, which is available either via SPAN or
OMNET, a commercial network system. We are
investigating ways to improve service to users

640

ROBUST SEQUENTIAL W-INTERVAL APPROXIMATION
DETECTORS WITH O-DEPENDENT SAMPLING

Evriclea Voudouri Ludwik Kurz

Manhattan College Polytechnic University
Riverdale# N.Y.10471 Brooklyng N.Y.1120

ABSTRACT contaminated and varying noise environments. The
theory of the MIPA detectors was extended to
In this paper, the theory of a robust sequential include sequential operation for independent
detection scheme is extended to include operation sampling in [5] and for dependent sampling of
with Q-dependent samples of unspecified unspecified bivariate and first order Markov
dependence. The detectors are designed to exhibit dependence in [6]. The above assumptions on the
a near optimum performance for nominal noise type of dependence-can be relaxed if a modified
distributions and maintain high efficiency to version of the Woinsky-Kurz scheme, [7], is
changing noise environments by adapting their introduced to treat the problem of signal
optimum nonlinearity Using an m-interval detection with dependent samples of unspecified
polynomial approximation (MIPA) of it. The Q-dependence, Q>2. In this case, the sampling rate
performance is evaluated for weak signals in can increase 'even further, resulting in an
contaminated and/or varying noise backgrounds improved performance over detectors with
often encountered in underwater acoustics. It is independent or bivariate dependent samples.
shown that the MIPA detectors improve their
efficiency, in certain cases up to Q times, at the Woinsky's idea was first to create a sequence of
expense of a slight increase in their structural independent vector samples of dimension N, N>Q, by
complexity to include operation with dependent properly grouping the dependent samples. !ffis is
sampling. Moreover, the sequential MIPA detectors accomplished by placing N consecutive samples in a
are asymptotically optimum and increase their vector, form an intermediate statistic and then
transmission rate up to four times as compared to skipping Q-1 samples to ensure that the next
their fixed sample size counterparts. vector of N components is independent of the
previous one. A similar approach to the sequential
MIPA detection is analytically tractable only for
1. INTRODUCTION the asymptotic case, (m-->*o). However, without
much loss in efficiency, analytic solutions can be
In recent years, the evolution of high speed obtained for finite m and for small signal
sampling techniques and the fact that data from applications if each sample vector is transformed,
multi-sensor arrays are temporally and spatially via a suboptimum projection, into a single
dependent, due to unwanted noise and interference, variable with a univariate distribution upon which
created a demand for detection procedures with the sequential MIPA test is applied. Hence, the
dependent samples [1]. In addition, it has been theory of sequential MIPA detectors is extended to
shown that noise fields in sonar, radar, and some include processing of Q-dependent samples.
communication channels are non-gaussian and highly
contaminated with impulsive noise [2]. In this In particular, the performance of sequential MIPA
case, parametric detectors based on a gaussian detector with Q-dependent sampling is analyzed in
nominal noise distribution are not appropriate detail for two important classes of detection
since their performance can be significantly. problems: The shift-of-mean alternative which
degraded for other than gaussian and varying noise finds applications in many data communications and
environments. Therefore, robust and adaptive to image processing problems and the change-of-scale
changing noise conditions detection procedures alternative which finds applications in processing
have been introduced [3-5]. Moreover, sequential, signal data or spectral estimation data from,
i.e., with variable sample size, operation of the multi-sensor arrays.
detectors has been used to reduce the average time
to decision considerably (up to four times) as 2. STRUCTURE OF THE DETECTOR
compared to detectors operating in a fixed-sample
size mode [11]. N test samples, z i=[x x2F ...,xN] from a
stationary Q-dependeni random process with p.d.f.
A fixed sample size robust detection scheme was f(xltx x ), N>Q, form an intermediate test
introduced by Tsai and Kurz in [4]. The Tsai-Kurz Q
scheme uses a piecewise approximation to the statistic in the following manner:
optimum nonlinearity function, known as an The samples are projected upon the op-direction of
m-interval polynomial approximation (MIPA), to an N-dimensional space via a linear
gain insensitivity of performance to highly

CH2585-8/88/0000- 641 $1 @1988 IEEE

transformation, namely z J'z. where J' is the the scores of order j, j=0,1,...,p, for,the inter-
iv N :L N , +
transpose vector of the transformation. val A k(p
Let f W be the density of the noise samples Independence between the intermediate testi
projeSted upon the 0-direction statistics of (la) is accomplished by using N
consecutive samples to form T.Jz } and then
fo(P (z IVg o(V (z tv )IA (z + hOT (z )I A 1(z skipping Q-1 samples before obtain"A T ifz
0 0
Since Tifz VI and Ti+,(z V I are independent, the
where gog(.) represents the nominal noise (usually sequential MIPA statistic with dependent sampling
gaussian) and ho(p(.) represents the heavy-tailed is an approximation to the sequential probability
contaminating noise of unknown form. The notation ratio test [81 and all. the expressions for its
IA(zV) represents the indicator function of performance measures, i.e., the average samples
interval A, which is 1 if xr=A and 0 otherwise. number (ASN) and the power function P(6), derived
The set A. represents the interval by Wald are valid. In this case, 9 represents the
(a-, a+) in which the nominal noise distribution parameter of the unknown alternative, and under
0 0
is dominant and A is the complement interval of the hypothesis "o-
A0 + Thus (la) represents the classical cumulative sum
Assume that the parameters Y and a have been es- form for the sequential Q-dependent polynomial
timated from a projected noise sam 0 e [5]. The ex- approximation detector (SQDMIPA) with stopping
Pi boundaries given by a=log[(l-A)/Al and
tension of the robust sequential MIPA to the Q-de- b--log[3/(l-a)1 where *I is the type I error,
pendent samples case yields then and -p is the type II error. Sampling continues if
n b < >Tn< a and one accepts H 0 or Hif if T n < b or
1@ = ET i f ZP) (la) Tn - a, respectively.
i=l The basic structure of the detector is shown in
where T.fz I is an intermediate statistic based on figure 1.
z Ind defined as
ln !IV- (zi!p) Projected Estivation:
T {z I (z ) + Leorning
9 (z A0 iT Y Somple C@ - Y SCORES, PARTITIONS
09 i(P OPTIMUM
PARTITIONS
p + bLuKtb
+ ze C_ z I (z (lb) Projected ao-
k=1 j=o 1 jkiv i(V + ilp Received oly M --i-D
z Somple. a- NO I
i(P TEST
where gof (.) and gl,(.) are the p.d.f.'s of the YES 0
projected sample in the 9-direction and in the PARAMETRIC
middle portion of the distribution under the DETECTOR T1
hypothesis Ho ( noise only conditions) and under
the alternative H, (signal corrupted by noise b <T.<Cl T,
conditions), respectively. Moreovert 81 is a YE T, I
preselected, for design purposes, alternative NO
parameter and may represent the signal level. Tn<b:Ho Tn> o:H1
I+ (ze), k = 0,1, ... Im, are the indicator func-

t
i4 + + + Fig. 1 Structure of the sequential MIPA detector
0 a of the intervals AkV = (a k-l' a kI and with dependent sampling.
Akq) = (a k' ak-l 1, which partition the � trans- Introduce- tT
formation of the sample space z,.l and c are -J @--as- the momeTM-genera-
jkT ting function of T., and assume fi(t) exists and
is finite for any f@oint t, including the origin.
The solution to the fundamental identity of se-
0
@b
TEST

T'
YES t
NO

quential analysis, [8], reduces then to

642

2
t0 -2 E{Ti I/ Var{Ti (2)
9->O 9 919(z
The proof follows from the cumulant expansion for B 04) .29 T 9=0 gw(z 9) dz
OW (Cramer [91), where the assumption was made 9 6; (z
that T i satisfies the strong mixing condition
OT
From the above discussion it follows that the
average sample number for the sequential MIPA It is interesting to note that t (0) is free of
detector with Q-dependent sampling is any dependence and from (6) and 0(3), [51, is
implied that the same thresholds can be used for
E{nJ9J = L(S)/EfT i 11 EfT i ) 0 (3a) near optimum detection under independent and
dependent sampling. However, the optimum
E{nle} = - ab / VarfT il EfT i } 0 (3b) partitions and therefore the scores must be
adjusted to compensate for the amount of
where dependence introduced in any particular
L(S) = [a(l - e It ) + b(e at -1)1/ (e at _e It) environment.
DO For the locally most powerful scores, the optimum
performance in a sequential mode of operation is
Following the same procedure as in the independent achieved if the partitions along the 9-direction,
sampling case, [5], it can -be shown that the a minimize the time to decision, i.e. minimize
locally most powerful scores which maximize the Ae ASN. Following a similar procedure as in [51
efficacy can be expressed in terms of the and under the assumption of small SIN ratio so
partition moments of the projected sample, that powers of 9 1 and e higher than two can be
:t(j) + + + neglected, we conclude:
E as c D eEC For the locally most powerful scores, the ASN
ok4 + jk(p jkq,' kq) achieves its minimum when the expression
where is a (2px2p) determinant with a typical
term 6kwj E�(j+q) I j,q = 01112,...Ip, D + is ob- P? + B 09 (9)
oko + kv is maximized. The expression above, however,
tained from D. by replacing its j-th column defines the efficacy of the fixed sample-size
jk? +(j) test, [51, in which case, the optimum set of
by the column of ekT' while -partitions is obtained by maximizing (9).
Hence, the following theorem is established.
�(j)
E P h (z ) dz and THBOREM. the set of partitions which minimizes the
OkT V 00 1P (V ASN for the sequential MIPA detector with
dependent sampling is identical to the partitions
which maximize the efficacy of the fixed sample
size tests. The optimum scores in both detectors
:L(j) are identical.
e +fzj"' ho4zo)l dz,,.
k(p ;-9 8=0 It can also be -demonstrated, using the same
approach as in [11], that as the type I and Type
III errors approach zero, the asymptotic efficiency
For small 8 1 the expectation, the variance, and of the sequential test compared to its fixed
the parameter t 0(0) reduce to sample size counterpart, is four times higher at
2 9=8 I and 8=8 . The efficiency improves even more
EfT 1 10) = (90 1 - e 1 /2)JE PIP + B 01V (4) for values o? 9 exceeding el.
, e2 C
Var{T + Bor} (5) The superiority in performance of the sequential
pq) MIPA detection sch6me over other robust tests has
t0 (8) 1 26/e 1 (6) been established already in ref. (5]. In this
paper, the performance of the detector is
where evaluated by comparing it to the independent
m p + + +(j), sampling case while taking into account the time
E 2: Z (D-1k /D- ) - to decision. The efficiency of the robust MIPA
PO k=l--j=o 3 T k(, ek,, (7) detector with dependent sampling T (z relative
nD
and to the independent sampling test T W, for
nI
sequential operation is

643

RE = (ASN I/ASND) - (tI/tD) (10) optimum scores are estimated recursively using the
stochastic approximation algorithms proposed in
where t I /t D is defined as the ratio of time [5].
required for independent observations to the total
time needed for a decision interval, including If we let in (11) m-->ao, the efficiency of the
enough delay to assure independent samples SQDMIPA detector with respect to the most powerful
[T i 1zV),T i+l 1z 4) M. detector with independent sampling approaches
Hence, using,(3) and (10), the relative efficiency I D(f;G) NQ
can be expressed in terms of the expectations of RE (12)
the two tests CO I(f.,G) N + Q
ED{TiIG} Q where I(f;O) is the Fisher's Information on the
RE= marginal noise density, and I (f;O) is the
E N + Q D
I {Tile} Fisher's Information for f OT (z (P ) given by
where ED{Ti} is given in (4) and EI(Ti} is the +to )]2 fo,(.
expectation of the Sequential MIPA test with f [f@O(zo) / fo(p (z (P ).dz 1P
independent sampling [5].

3. EXAMPLES
In this section, the performance of the sequential In the case of gaussian noise, z T is also gaussian
MIPA detector with dependent sampling (SQDMIPA) is and the relative efficiency can be expressed in
investigated under the shift-of-mean and change of terms of the covariance matrix Z of the
scale alternativesand the optimum parameters for N-dimensional noise distribution and the variance
zero order ('p=O) and first order (p=l) MIPA are 62 2/
estimated. as RE= {or Jl.'J){NQ/(N-tQ-1)} which for N=2
betectors based on the shift-of-mean alternative and Q=2 is given in terms of the correlation
find applications in processing of known signals coefficient by RE = (4/3)/(l+r).
in additive noise. If the level of the signal is For values,of Q>2, and as N approaches infinity
8, then the alternative is represented in terms of
the asymptotic relative efficiency reduces to
the hypothesis c.d.f. as F(Z) = Fo(Z--e) and Q-1
F(Z).@FO(Z) for any vector of random variables Z RE Q 1 + 2 2: r k
and 0 > 0. N-->oo k=1
where r k is the correlation coefficient between
the noise samples x. and x., such that li-jI = k.
.1 1
It can be shown that the probability density of It should be pointed oat that for uncorrelated
samples, r=O and the relative efficiency
the projected sample reduces to approaches Q. Furthermore, if we also let p,->ao,
f(P (z (P fo4p (z (P -RN 0) where f T(zT) is the p.d.f. we obtain the same result as Dwyer and Kurz
derived for the sequential partition detector in
.of the projected sample under the unknown [121. Namely,
alternative, fo(p(.) is the p.d.f. of the projected Ts
sample under the hypothesis,. and RE Ts2 r(r) d-C
N NIQ ->*0 0
RN 2@ji where T is the sampling interval for uncorrelated
i=1 samples 8 and r(T ) is the correlation coefficient
The hypothesis-alternative separation parameter in terms of the translation of time between two
now becomes A= RNO and the direction 4v is chosen dependent samples.
such that this separation, and therefore the SM For finite values of N and for Q=2, the relative
ratio for the transformed vector z,,/ is maximized. efficiency is shown in figure 2.
Hence, Ji = 4 -N M.
Moreoverr the optimum partitions which minimize It is more tedious to obtain the above results in
the ASN and therefore maximize E are estimated, the case of contaminated noise. For example, a
via a gradient search routine on (7) by letting N-dimensional density with marginal densities of
p=O and p=l. Thus, the partitions for the zero the Huber mixture type is typically represented as
order and the first order MIPA can be obtained. On
the other hand, the partition moments and the

644

N R@@
(N 1.50 . . . .
fZ (z) E i) (1_ 6)i 6N-i f l(i)2(N-i) (z) 1.45-
1=0
(14)
1.35- A, A-A
where Z is a N-dimensional vector of random var- 1.30- CB
iables, Z = [ x1f x 21 ... 1 X N ], with marginal den- 1.25- C C
sities f X (x) = (I- t) f I (x) + ef 2(x) 1.20- P =0 P=1, N=3
.and . f1@i)2N(N-i) Z. is the joint density of N 1.15-
variables out of which i have marginal density 1.10- A A A--
f1 (x) and the remaining (N-i) have marginal I.05- C C
density f 2 W. 1.00- P= 1, N=2
.95- 10
.90 , I C,
0 .8 1.0 1.2 1.'4 1.6 1.8 2.0-

REA p: order of MIPA, N=Q, m--3
1.8- Degree of contamination E: A: 0 & 1, B: 0.01
r=0.05 N--l-o C: 0.9
1.5- Fig. 3. Relative efficiency, Shift alternative
oo The one-sided change of scale alternative emerges
in detection problems dealing with spectral data
1.0- and transmission through dispersive media. The
distributions care one-sided, and the
- change-of-scale effects a stochastic ordering,
i.e.,
.5-
- F(Z).LFO(Z) for any vector Z. Specifically,
16 20 3 40 N F(Z)=FO{Z/(l-te)}, e>O, and the density of the
projected sample reduces to
f'P(Z'@) = fo'@fzq'/(l+q)) / (1+9).
Fig. 2 Relative efficiency of SQDMIPA with As in the shift alternative case, it can be shown
respect to the MIPA detector, Q=2 that the transformation which maximizes the SIN
In the case of gaussian noise contaminated by ratio is
high power noise samples, f 1 (x) = N(0,6?) and Ji =VN / N, i= 1,2,...,N.
f2(x) can be represented by The relative efficiency with respect to the
2 independent sampling detector is obtained by (11)
N(O, ff 2) where C 2 >> C I' and (4), and the asymptotic relative efficiency
(m->oo) reduces to
In this case, the relative efficiency with respect
to the independent sampling detector can only be REO'= (f-D /CI )[Q/(N+Q-1)]
evaluated using numerical methods on (12). For
finite m, in particular, the relative efficiency where E Dis the efficacy of the SQDMIPA detector
is depicted in figure 3 in terms of the as m->aof
parameter a 0 , where a =a + and -a =a It is + Q4 2
shown that for both ?irsot and second order: E f [1+Z {f' (Z )/f (z M foP (z dz
polynomial approximations, bivariate and D -00 (P ov (P O(P T
trivariate dependence, and for several values of
contamination parameter c the relative efficiency and E I is obtained by (15) if foV(z;,) is
is almost independent of the selection of a and
0 . substituted by the marginal density of the noise
maintains near optimum performance even for high
degrees of contamination. samples, fo(x).
B C
C_
C_

P @__Co P

645

Application: The change of scale alternative-can 2.0
model several practical problems including power
spectral analysis and processing of sonar, radar,. 1.9-
and seismic data. Spectral analysis of time series.
data is performed by transformation of time. 1.8-
segments into sequential estimates of the spectrum
of the received signal. The problem is one of 1.7- r=0.05, 0.01
detecting narrow band signals in background noise.
Assuming constant signals in frequency and 1.6-
gaussian noise at the input, the optimum detector
incoherently integrates the spectrum for a period 1.5- r=0.223
of time before detection. The integration is
equivalent to averaging the measurements of the 1.4-
spectrum over time. The sequential MIPA test may
give a useful alternative to integration, 1.3-
providing robustness which averaging cannot
provide. 1.2'- - -
In this case, the spectrum analyzer may be 0 .2 .4 .6 1.0 1.2 1.4 1.6 1.8 2.0
considered to be a bank of narrowband filters of
width 2&Hz, each followed by a squaring device IFig. 4. Relative efficiency vs a
and the SQDMIPA detector. With gaussian noise at scale alternative, m=21 P=O 0
,the input of the filters, the second order
probability density of the output of the squaring On the other hand, if the noise at the input of
device can be expressed in terms of a series the filters is contaminated gaussian of the type
expansion [13]. represented by (14), then the second order
probability density at the output of the squaring
Then, the probability density of the projected
noise samples, for,z >0, is device is obtained using a similar procedure like
0 the one used by Rice in [14]. In this case, the
f (z r212 6 2) expf-z F2/262 efficacy of the SQDMIPA detector has been
Of T T evaluated and its relative efficiency versus the
00 n most powerful detector with independent sampling
(r2) fL (z &121) L (z r2126 2 (M->oo) is shown in fig. 5 as a function of the
2n (P 2n+l contamination parameter e and for several.
n=o correlation coefficients r.
where L n(x) is the Laguerre Polynomial.
If a narrow band gaussian signal is added to
the noise, the change-of-scale alternative may be 1.4
used to model the probability density and 9 may
represent the unknown SM ratio.
The performance of the most powerful detector is 1.3 r:O '& 0.05
obtained by letting m-> w. AS m->oo the efficacy r= 0.1
of the detector becomes 2/(l+r2) and is 1.2 %0.223
independent of the variance of the noise. Hence,
the relative efficiency with respect to the 1.1
optimum detector with independent sampling reduces
to (4/3)/(I+r2 1.0
.05 OJ 0.15 0.2 0.25
For finite m, the optimum design parameters of the,
detector have been estimated following the same
procedure as in the shift-of-mean alternative case
and for several correlation coefficients. The Fig. 5. Relative efficiency vs the contamination
relative efficiency with respect to the parameter Scale alternative.
independent sampling detection scheme is displayed
in figure 4 as a function of a where From figures 2-5 it is concluded that the MIPA@
+ 0 detectors with dependent sampling improve their.
a- E{z + a performance considerably as compared to the MIPA
&L-0=02
2@ 23

0 detectors operating with independent samples.
Especially in the case of the change-of-scale
,alternative, the efficiency increases up to 80%.

646

7. M. Woinsky and L. Kurz, "Nonparametric De-
4. CONCLUSIONS tection Using Dependent Samples,n IEEE
Trans. on Info. -Theory, Vol. IT-16,
In this paper, the effects of dependent sampling PP. 355-358, May 1970.
on sequential m-interval polynomial approximation
(SMIPA) detectors is considered. First, the theory 8. A. Wald, Sequential Analysis, John Wiley
of SMIPA detectors is extended to include and Sons, New York, 1947.
N-dimensional processes by introducing a
suboptimum linear transformation from an 9. H. Cramer, Mathematical Methods of Statistics,
N-dimensional space to one dimensional space. It Princeton Univ. Press, Princeton, N.J. 1964.
is shown that the same partitions and scores lead
,to optimum performance for both the fixed sample 10. M. Rosenblatt, "Independence and Dependence,"
size and the sequential MIPA detectors even though Proceedings of the Fourth Berkeley Symposium
the sequential operation reduces the average time on Mathematical Statistics and Probability,
to decision by a factor of four. Moreover, the Vol. 11, 1961.
decision thresholds remain the same as in the
independent sampling case, and the optimum 11. R. Dwyer, "Sequential Partition Detectors with
partitions and scores are easily adjusted to Applications," PH.D. Dissertation Polytechnic
compensate for the dependence of the test samples. University, 1976
Efficiency calculations have been made for the
shift-of-mean and the change-of-scale 12. R. Dwyer and L. Kurz, "Sequential Partition
alternatives. For contaminated noise densities and Detectors with Dependent Sampling, " Journal
for zero and first order MIPA, numerical results of Cybernetics, Vol. 10, pp. 211-232, 1980.
show that both alternatives give improved -
efficiencies under the SQDMIPA designs as compared 13.. J. Barrett and D. Lampard, "An Expansion for
to the sequential MIPA detection schemes. Some Second-order Probability Distributions
and its Application to Noise Problems, IRE
It should be pointed out that the SQDMIPA detector Trans. Info Theory, Vol. IT-1, pp. 10-15,
can be easily implemented. All the parameters can 1955
be estimated from a noise sample and updated
periodically, via parallel processors, without 14. S.O. Rice, "Mathematical Analysis of Random
interrupting the detection process. Thus, the Noise," Bell Systems Technical Journal,
detector can adapt its performance to slowly Vol. 23, p. 78, 1945.
changing noise environments.
ACKNOWLEDGMENT
REFERENCES
This work was supported in part by the Manhatttan
1. R. Dwyer and L. Kurz,"Characterizing Partition College Faculty Grant and RCA
detectors with stationary and Quasi-stationary
Markov Dependent Data," IEEE Trans. on Info
Theory, Vol. IT-32, pp 471-482, July 1986

2. R. Dwyer, "A Technique for Improving Detection
and Estimation of Signals Contaminated by
Under Ice Noise," J. Acoust. Soc. of Am.,
Vol. 74, pp. 124-130, 1983.

3. P. Huber, "A Robust Version of the Probability
Ratio Test," Annals of Math. Stat., Vol. 36,
pp. 1753-1758, 1965.

4. C. Tsai and L. Kurz, "An Adaptive Robust-
izing Approach to Kalman Filtering,"
Automatica, Vol. 19, No 3, pp. 279-288, 1983.

5. E. Voudouri and L. Kurz, "A Robust Approach
to Sequential Detection," IEEE Trans. on
Acoustics, Speech and Signal Processing,
Vol. ASSP-36, pp 1200-1210, August 1988

6. E. Voudouri and L. Kurz, "Robust m-Interval
Polynomial Approximation Detectors with.
Dependent Sampling" Proceedings of the nine-
teenth Conf. of Information Science and
Systems, pp. 132-137, March 1985.

647

A SHIPBOARD DATA, ACQUISITION, LOGGING AND DISPLAY SYSTEM

Geoffrey Samuels

Rosenstiel School of marine & Atmospheric Science
University of Miami
Miami, Florida 33149

ABSTRACT the VMS operating system that runs on the microVAX.
This system was modelled after a set of CTD data
A shipboard system for acquisition and logging of acquisition and logging routines developed at the
oceanographic data is under development at the University of Rhode Island.
University of Miami utilizing seagoing microvAx
computers running the V14S operating system. The 2. DESCRIPTION OF THE SYSTEM
separate functions of data acquisition, logging and
display are performed by independent programs. one feature of VMS that is used is the ability of
Data are transferred between programs by using a independent programs to share the same data among
common database resident in the computer memory. themselves. This can be accomplished in a number
An example of the system is a working system of of ways: one way is using what is called a
routines to acquire, log and display oceanographic shareable image. A shareable image is a piece of
data collected by a shipboard SAIL (Serial ASCII software code that can be accessed by two or more
Instrumentation Loop) system. separate computer programs. By installing a
FORTRAN named COMMON block or a C external data
structure as a shareable image, a common data
1. INTRODUCTION buffer can be established and accessed by several
programs simultaneously. This has enabled the
one of the requirements of an oceanographic separate functions of data acquisition, data
research cruise is the need to gather data and to logging and data display to be split into separate
make it easily available to investigators. A independent programs; they communicate between
typical cruise may need to collect CTD data, themselves by sharing the same common data buffer.
navigational information, Acoustic Doppler Current The data acquisition program does the data
Profiles and meteorological data. collection, either passively from instruments that
output data automatically or actively by issuing
In the past at the University of Miami many data requests to instruments that must be polled.
different types of data were individually collected The program then stores the data in the common data
by dedicated microcomputers. For example CTD data buffer along with a timestamp showing the time of
acquisition would be done by one microcomputer acquisition. An independent data logging program
while shipboard navigation would be collected by a accesses the data buffer and stores the updated
separate microcomputer. The data acquisition, data on disk. Another program can take the raw
logging and display would be done by a single data in the data buffer, apply calibrations to it
monolithic computer program. There were problems and store the calibrated data in a separate
with this approach: since the programs had to do calibrated common data buffer which can be accessed
several functions they tended to be large and by display and plotting programs. The speed of the
complex. It was difficult to modify and maintain microVAX permits rapid updates of the common data
these large programs and often memory limitations buffer so that the data can be displayed or plotted
of the computers made it impossible to extend the on a near realtime basis. Figure 1 shows a block
program's capabilities. diagram of the process.

Recently, technological advances have introduced Each data source can have its own set of
computers that are comparable to large mainframes acquisition, logging, calibration and display
in speed and memory, yet are physically small routines communicating via common data buffers.
enough to be practical, seagoing computers. At the The rest of this paper will provide an example by
University of Miami Rosenstiel School of marine and describing a system to acquire and log shipboard
Atmospheric Science (RSMAS), microVAX computers are SAIL (Serial ASCII Instrumentation Loop) data.
being introduced for seaboard data acquisition and
logging. These computers are fast enough and have 3. MULTIPROGRAM SAIL DATA ACQUISITION AND LOGGING
enough memory to run data acquisition programs that
previously required *three or four separate The SAIL system allows access to a variety of
microcomputers. At RSMAS, a system of data instrument modules monitoring various shipboard
acquisition, logging, and display routines is being parameters- It i described more fully in the
developed to take advantage of certain features of ANSI/IEEE standardl@. Typical parameters measured

CH2585-8/88/0000-648 $1 @1988 IEEE

on board University of Miami vessels include ship calibration program which reads the raw data in the
speed and heading, anemometer wind speed and common data buffer, applies a set of calibrations
direction, thermosalinograph temperature and to the data and stores the calibrated data in a
conductivity, wet and dry bulb temperatures and separate calibrated common data buffer.
barometric pressure. Each module is independently Calibration constants can be kept and read from
polled by the SAIL controller; upon receipt of an separate calibration files. A SAIL display program.
.attention character and a unique two-digit address, accesses the calibrated common data buffer in much
a module will transmit an ASCII data message the same way the logging program accesses the raw
followed by a message termination character. data buffer and displays the current calibrated
SAIL data on a terminal. In a multi-terminal
On the microVAX system the SAIL system controller situation several copies of the display program may
is the basic data acquisition program. The program be run, making for convenient display of shipboard
(called the SAIL "grabber") maintains a list of navigational and environmental parameters.
active SAIL adresses in the common data buffer. It
polls each module address and stores the received Future implementations of the system will include
ASCII data message along with a timestamp of when realtime data plotting of the calibrated data
the poll was issued. After all addresses in the utilizing terminal graphics or pen plotters.
list have been polled, the grabber program places a
"lock" upon the common data buffer and transfers 4. CONCLUSION
the data it has acquired. It then releases the
lock on the common data buffer and begins another By utilizing interprocess communication through
set of polls. If a module fails to respond after a shareable image common data buffers, the different
specified time, or if invalid data is received, an functions of data acquisition, data logging and
error message is sent to the console of the data display can be accomplished by independent
microVAX and to an error logging program. The programs. This allows flexibility in meeting the
grabber program is maintained as an independently requirements of a particular principal investigator
running process updating the common data buffer at (or investigators) for a research cruise and makes
the end of each set of polls, typically ten to it easier for future expansion of the system.
fifteen seconds, depending upon the number of
selected modules and the length of the data 5. ACKNOWLEDGEMENTS
messages. Communication with the grabber program
is accomplished via the common data buffer. The I wish to thank the National Science Foundation for
list of active SAIL modules can be modified at any its financial support of the project. I also wish
time by an independent program. Also, if it is to thank O.B. Brown, J.W. Brown, R.H. Evans, A. Li
desired to halt the grabber program, a flag value and G. Basham for their many helpful comments and
can be set in the common data buffer by another guidelines. I also especially thank L. Covington
program. for demonstrating the feasibility of the project.

Actual logging of the raw data. messages is 6. REFERENCE
accomplished by a separately running logging 1
program which can be set to log any or all of the IEEE Standards Board, "IEEE Standard Serial ASCII
SAIL modules. The logging program only checks the Instrumentation Loop SAIL Shipboard Data
common data buffer at set intervals; a typical Communication" ANSI/IEEE Std 997-1985, 1985, 12pp.
sampling interval would be five minutes. At that
time the logging program places a lock upon the
common data buf fer and makes a copy of the data.
It then releases the lock so that the grabber
program can continue updating the buf fer. The Data Data
logging program checks this data for each SAIL Source kcquisition
address to be logged. If the module data has been Program
updated since the last logging check, the raw SAIL
data message, along with its timestamp, is written
to a disk (or tape) file. Raw Common
Data Raw Data
Since the data acquisition and data logging pro- File Buffer
grams are independent of each other, communicating
through the common data buffer, more than one
logging program can be run at the same time. One Raw
investigator may only want one subset of all Data Data
available SAIL parameters logged every 30 minutes, Logger Calibration
while another investigator might desire all
parameters logged every five minutes. This
flexibility was one reason for splitting the Realtime
functions of data acquisition and data logging. Display
Program Common
The ASCII data stored in the common data buffer has -.Calibrated
to be calibrated before it can be displayed or Realtime Data
Buffer
@Raw
r
F@
at a
le

plotted. This is accomplished by a separate Plot
Program

Figure I

649

OFFSHORE OCEANOGRAPFUC APPLICATIONS FOR BATTERY-POWERED, HIGH-END
MICROPROCESSORS

Carroll V. Bakerl and Wm. T. Whelan 2

'Skidaway Institute of Oceanography 20cean Communications Systems, Inc.
Post Office Box 13687 2430 Industrial Drive
Savannah, Georgia 31416 Panama City, Florida 32405

ABSTRACT operational amplifiers contributed useful performance and
energy economy improvements as well in the new design.
The availability of the all static CMOS 8OC88 micropro-
cessor and a commercial, single board computer featuring In this paper, we briefly describe three offshore applica-
it has presented the designer of battery-powered, unmanned tions which make use of the 80C88 monoboard computer.
oceanographic systems with a new, powerful building Two of these were developed, solely, by the senior author
block. Three different, oceanic applications are described, and the staff of the Skidaw.ay Institute of Oceanography
which have substantially benefited from this powerful new (SKJO), a unit of the University System of Georgia,
tool. located on Skidaway Island near Savannah. The third,
which is discussed first because of its complexity, has been
a joint development by SKIO and a supporting contractor,
Ocean Communication Systems Inc., of Panama City,
Florida.
1.0 INTRODUCTION

2.0 THE LORAN REPEATER BUOY.
The ubiquitous microprocessor revolution of the past
decade has been rather slow to make inroads into applica- The first application of the 80C88 computer discussed is
tions aboard "stay-behind" instruments used in coastal and that in a free-drifting buoy system, designed to measure
oceanographic sensing. The reasons for this were several, surface or near surface, oceanic or estuarine water circula-
among which include, the electrical energy demands of tion. The buoy in which the computer is installed provides
most devices, and the relatively modest computer power of a minimal wind cross section so its motions will be dom-
the available hardware. Until 1984, most of the "CMOS inated by small water movements. The water drag is
CPU" hardware was approximately 50 percent NMOS tech- enhanced by the use of an underwater 'sail' or current trap
nology. Operations in the "powered-down" mode reduced approximately 4 feet on a side. The trap depth setting is
the battery demand by a relatively small amount when selected for the water depth of concern. This can range
compared to what a 100 percent CMOS microprocessor from surface to about 125 meters. The buoy is 10 feet
could do. The few, true, static CMOS CPUs had modest long. It has a cylindrical electronic housing 20-inches long
computing capacities and required unfamiliar programming and 8-inches in diameter, which is located below the water-
skills. Despite this, there were enough applications made line. Its launch weight, fully ballasted, is 130 lbs.
to indicate the future course of events. The need for an
interrupt driven engine with sufficient computing power to In 1977, the original Loran repeater buoy system was
make a difference and which could provide static sleep developed by the contractor with support from the Geogra-
between interrupts was obvious. Harris Semiconductors phy and Earth Sciences Branches of the U.S. Office of
met the need with its 80C88. For the first time, a powerful Naval Research (ONR). This support was part of ONR's
microprocessor was available which was capable of bring- ongoing commitment to transfer unclassified aspects of a
ing the desk computer to sea and running it (for a long military development program for HF, Over-the Horizon
time) from a handful of flashlight batteries! Onset Com- Radar (OTHR) to the civilian, scientific community (Ref.
puters developed a single board computer system which 1). The present SKIO contractor worked many years in the
was designed to exploit the opportunities offered by a ONR radar program where he was also associated with
powerful processor, stingy of battery capacity, in isolated, development of several large ocean data buoys.
Istand alone' situations. This provided a convenient,
interrupt-driven computer suited to the type of applications 2.1 Communication Considerations
discussed here.
The Model A-8 Loran buoy employs an advanced HF com-
The growing capabilities of new, micropowered, precision munication system over which it periodically broadcasts its

CH2585-8/88/oooo- 650 $1 @1988 IEEE

Loran position. The position information is obtained from 2.2 Data Protection
a minimally-modified, automated Loran Receiver-Navigator
carried aboard the buoy. After many trials we have settled It would be most unusual when the position data from the
on any of several Sitex CMOS models. The particular Loran Buoy could not be delayed for a period of at least a
model used is mission-dependent. The buoy can be few hours until transmission could be made at a better time
received by an unlimited number of properly equipped sites thus avoiding interference limitations. This option has
located within the coverage area of the broadcast. For the been built into the buoy data system in two ways: the first,
electromagnetic surface wave mode of HF which is used at a software solution and the second, a more , complex,
sea, this coverage is a circle approximately 200 miles in hardware approach. Both require the inclusion of static
diameter (Ref. 2). Several dozen buoys can share a single RAM CMOS memory in the buoy. Sixty five thousand
transmission channel. It is noteworthy that the electromag- five hundred and thirty five bytes of SRAM provides
netic surface wave BF propagation employed by the sys- sufficient nonvolatile storage for all of the data collected
tem, termed Mode 1, permits buoy broadcast reception for from an entire cruise of a drifting buoy. With the data
only a few miles inland from the coast (Ref. 3). stored on board, it is a simple matter to protect nighttime
collections by one of the following three approaches. First,
System transmitters and receivers are specifically designed program the buoy to routinely transmit the 'at-risk' night
to process the coherent phase modulation/demodulation and data during daylight of the following day; second, include
the encoding/decoding employed for error control. a command link to a receiver on board the buoy and
Extended operational battery life is obtained through exten- request a repeat of ' specific portions of the data record
sive use of CMOS technology throughout the system and which may have been lost to noise or interference; and
high efficiency, MOSFET, switchmode transmitters. The third, recover any missed data from, the buoy by direct
application of advanced signal processing techniques per- hardwire 'conversation' via the user's programming umbili-
mit the use of only modest levels of radiated power for cal cable attachment when the buoy is recovered aboard
effective cornmunications over the range supported by the ship. Actually, all three options are available in the Loran
HF surface wave mode noted above. Most Loran opera- buoy discussed here. For the command receiver option,
tions use surface wave mode since the modest transmission the receiver cannot be operated continually, since the
ranges not only suit the scale of local problems, but energy penalty would be prohibitive. Therefore, at known,
because the distance between buoys and the intended fixed times, the buoy management computer periodically
recovery ship(s) prudently should not exceed 50 to 100 activates the energy-conservative, onboard command
miles (Ref. 4). receiver. The latter is then available for receipt of instruc-
tions regarding retransmission of specific portions of the
For mode 1, surface wave propagation, the maximum data or changes to the user-installed data collection and
power output provided by the buoy transmitter at the reporting schedule. All stored data is formatted in the
antenna base is approximately 0.8 watt. With the usual packet mode. The packet or "ping-pong" mode of data
efficiency of small antennas at 4 Mhz, the effective radi- transfer, with its immediate error correction capability, is
ated power level is in the range of 30 to 50,milliwatts. also available as an operating option.
This has been judged to be a reasonable compromise
between the buoy's battery operating life and measured 2.3 Other Data-Gathering Capabilities of the Loran-
reliability of 300 baud message receipts at the 100 mile Repeating Buoy
radius of the diffraction zone. Commonly, reception under
these conditions can be anticipated to be* satisfactory The buoy can be provided with a suite of extremely low-
throughout the day and approximately one-half of the dark, powered sensors to report on the environment which they
sunset-to-sunrise hours. During the first half of the average traverse. Optional CMOS ADCs convert (12-bit plus sign)
night, inteference from other world users can limit message the analog data for digital transmission by the system.
copy to ranges of 50 miles or less. The presence of a This option can substantially extend the scope of a circula-
severe thunderstorm, at night, which is located within line- tion study for efforts, such as oil spill studies (Ref. 6).
of-sight of a particular receiving station, can eliminate any
useful reception at that station until the storm dissipates or 2.4 Buoy Data Receiving Terminals
moves on.
Multiple buoys time-share a single radio channel. This
The performance cited in the preceding paragraphs is con- greatly simplifies the cost and complexity of an extended
siderably improved over that of traditional HF (civilian or drift study in, say, the Gulf Stream where relatively large
military) circuits. This improvement reflects application of areas must be covered. The receivers are specialized
a number of modern signal and data processing techniques because of the complex signal processing they must per-
(Ref. 5) combined with an innovative, small antenna on the. form (Ref. 5). Attempts to use commercial receivers are
buoy. generally disappointing, and since the system-specific
receivers are about twice the going commercial price of a
good 'ham' receiver, efforts to minimize the number of
receiving positions are worthwhile. Because of elec-

651

tromagnetic noise produced by ship's machinery, special
@A- it
active, noise-reducing antenna systems are available and
advisable for reception aboard most research vessels.

3.0 A 80C88-BASED COMPUTER APPLICATION
TO COLLECTION OF OFFSHORE
ENVIRONMENTAL DATA
J,
Our second application of a computer-based system
management (Fig. 1A) was for the update of an offshore
sea-air environmental monitoring station. The original sta-
tion, previously described (Ref. 7), is located on an
offshore platform, 17 miles from the data collection center
at'SKIO on Skidaway Island. It recently became necessary
to convert this diesel generator facility to unattended, A
solar-powered operation.
In cooperation with the U.S. Coast Guard (USCG), Skida- Fig. 1. A.) 8CO88 based computer.
way maintained an air-sea sensor system and VHF radio B.) VHF ra
(Fig. 113, 2) data link to the Institute. Commercial grade dio transmitter.
prime power was platform-generated and it supported the
SKIO system. In 1987, the USCG converted to a solar-
powered unmanned operation, and SKIO redesigned its
@,Z
environmental station similarly. Wherever possible, our
@AA,
sensor system was reconfigured to minimize its electrical
load. An added premium was placed on system reliability,
since an 8-hour round trip cruise is required to correct any
4ro kA
I YX
onsite difficulty.

The new system management computer selected was the
same single board, 8OC88 computer employed by the
Loran buoy. The sensors managed and reported by the
new system are included: 1. Anemometer, reporting wind
speed and direction; 2. Ten thermistors; two for air tem-
perature and eight for water temperature at increasing
depths; 3. Subsurface pressure gauge for wave and sea
level measurement; 4. Barometer; 5. Battery terminal vol-
tage and charging current. All sensor outputs are con-
verted to rational engineering units by the onsite computer Alk
prior to transmission to the mainframe at the Institute. Fig. 2. VHF radio antenna.

3.1 Solar Power Supply and Load Details; Communications

A pair of solar panels (Fig. 3), supplied by Solarex Corp.,
provides up to 5 amperes at 15.8 volts under bright, sum-
mer conditions. Each panel area is 2 square feet. They
are directly connected to a pair of large, paralleled, 12 volt,
industrial lead-acid batteries. The combined battery capa-
city is 200 ampere-hours and is sufficient to eliminate the
need for charging regulators. This sytem has proven ade- tit 51,
quate for long winter nights and protracted periods of con-
current, overcast, daytime weather. The power system has
-free since its installation 14 months ago and
been trouble
has shown no signs of deterioration. Repair of minor wind
and wave damage to several of the sensors or their wiring
has been the only required maintenance to date.

The significant load is the 50 watt demand of the VHF
transmitter (Fig. 1B) which operates 6 minutes of each Fig. 3. Solar panels.

652

hour, 24 times each day. The sensor and computer loads mately I square meter in cross section at the open door
are negligible in comparison with the transmitter load. It is and can be sequentially opened and closed on command
interesting that the commercial anemometer package, which from the user's keyboard. The on-deck weight of the sys-
we chose not to modify, draws 100 milliamperes compared tem is approximately 1000 pounds and* it is readily handled
to the "fully-up" computer demand of 20 milliamperes. by the ship's hydraulic hoist gear as a side trawl.

4.0 MICROPROCESSOR APPLICATION 4.1 Sensor-Derived Information Continuously Available on
FOR A SMART PLANKTON NET Data Screen

An instrumented, towed, underwater, plankton-collection When making a towed collection, the investigator normally
net has been developed by SKIO. The system which per- seeks to tow the net at a selected depth from a given start-
mits 'user friendly' control by the scientific investigator on ing position to some approximate terminus along a
the towing vessel has been a long-standing need (Ref. 8). preselected heading. To assist in this his data display
However, the conflicting requirement for maintaining ship's screen is updated with course and ship's speed information
time productivity while still permitting the flexibility of from the ship's navigation and pilotage equipment. An
collection operations, has made this a difficult task. important tow parameter is the water' throughput rate as
Experience suggested that a major improvement would well as the accumulated water flux as the tow is underway.
result if a system could be devised to provide all concerned The net carries a series of sensors which provide the com-
with a clear picture of precisely what is occurring beneath puter with a constant update of raw data from which it can
the surface as the net is towed. The ability to know the calculate and display volumetric data in real time. The
position as well as the course of the net, in three sensors provide the following information: 1. The timing
dimensions, at all times in c Ion-s.i.der.e.d eIs.sential-. ItIis criti- of all door openings and closings and the continuous status
cally important to minimize the number of times the net of all doors; 2. The readout of a precision inclinometer
must be raised for any reason other than legitimate which measures the vertical angle, and its variations, of the
recovery of the collection. This is especially the case net and its doors; 3. The instantaneous flow rate of water
where deep towing is involved. Rugged, dependable door into the door(s); 4. The precise depth of the doors; 5.
latches on the nets and trustworthy, on-deck status displays Water temperature; 6. Shallow (0-100 meters) and deep
of the door conditions are essential to user confidence. (0-1000 meters) hydrostat readouts for precise depth track-
ing; and 7. Continuous monitoring of the watertight
The new net system employs the same 80C88 microcom- integrity for all dry chambers. A micropowered conduc-
puter to manage the array of seven (7) stacked nets in the tivity cell may be added by the time of this publication to
single, towed package (Fig. 4). The nets are each approxi- provide complete CTD profiling capabilities while raising,
lowering or towing the nets. Computation of all the criti@
cal ship's handlers are performed and routed the appropri-:
ate display terminals on the bridge and deck.
A'.@ "PI, 4.2 Communications B
"A etween Ship Deck and Net
The coaxial two-wire, steel tow cable is 2500 meters in
length and is carried on a deck winch. All data communi-
cations signals as well as the continuous, constant recharge
current of 0.4 ampere for a peak-handling, 12-volt battery
in the submerged package, are passed over the total cable
length. The compliance of this 'smart', constant current
charger must be approximately 125 volts to overcome the
resistive drop in the long cable. Full duplex communica-
tion is maintained between the deck user's terminal and the
microprocessor in the net electronics capsule (Fig. 5). Bell
212 protocol is used at 1200 baud data rate. Customized
couplers and drivers are used at each end of the cable.

5.0 SUMMARY

The capabilities of the 100 percent CMOS 80C88
microprocessor (and related brethren, including the 80C86
and 80C286, etc.) make it possible to provide computing
and data manipulation capabilities for battery-powered, ,
-alone ocean systems which are comparable to those
stand
available at our desks. The efficient, logical structures of
Fig. 4. Towed package. these Intel engines are nearly irresistible invitations to

653

Y777@-_"",
tems. Wiley and Son, Inc., 1982, pp. 208-300.
od,
!Aw
6. Fredericks, R. G., Wiseman, W. J., and Whelan, Wm.
T. "An Expendable Telemetry Buoy for Coastal
Oceanography." Proceedings, OCEANS '77; Annual
Conf. of Marine Tech. Soc. and IEEE, 1978, pp.
2Cl-2C5.

7
7. Schwing, F. B., Blanton, J. 0. Lamhut, L. and Baker,
C. "Ocean Circulation and Meterology of the Georgia
Coast." Technical Report Series Number 84-1, 1984,
22 pp.

8. Wibe, P. H., Burt, K. H., Boyd, S. H. and Morton, A.
W. "A Multiple Opening/Closing Net and Environ-
mental Sensing System for Sampling Zooplankton."
Journal of Marine Research, Vol. 34(3), 1976: 313-
Fig. 5. Net power supply and modem. 326.

assembly language programming, for the few who have
submitted to the disciplines and developed the necessary
skills. All SKIO application programs described here are
written in the assembly language. Interestingly, in each of
the systems described, the throughput chokepoint was in
the serial communications link and not the CPU.

6.0 ACKNOWLEDGEMENTS

Thanks are expressed to Dr. Jack 0. Blanton, Dr. Gustav
X. Paffenhofer, Mr. Lee H. Knight, Ms. Dee Peterson, and
.4s. Anna Boyette for their suggestions and cooperation in
the preparation of this paper. Without their help the
preparation of this paper would have been most difficult.

7.0 REFERENCES

1. Abstracts. Coastal Geography Programs. 1978
Annual Contractor's Conference. Office of Naval
Research. Arlington, Va. 78 pp.

2. Barrick, D. E. "Theory of Ground Wave Propagation
Across a Rough Sea at Dekametric Wavelengths."
Battelle Memorial Institute, Columbus Labs. 1970. 93
PP.

3. Barrick, D. E. "Increased Propagation Attenuation
Over Land." Memo to DARPA (U); U.S. Department
of Commerce. NOAA Environmental Research Labs,
Boulder, CO, June 1973. 8 pp.

4. Kirwan, Jr., A. D. and Chang, M. S. "Effects of
Sampling Rate and Random Position Error on
Analysis of Drifter Data." Journal of Physical
Oceanography, 1979, Vol. 9: 382-387.

5. Holmes, J. "Coherent Detection of Coherent Modu-
lated Signals." In: Coherent Spread Spectrum Sys-

654

CIDS - A SHIPBOARD CENTRALIZED DITEGRATED DATA SYSTEM

Richard Findley

Rosenstiel School of Marine & Atmospheric Science
University of Miami
Miami, Florida 33149

ABSTRACT - The system should have flexibility to meet future
needs.
In December of 1986, a microvAX computing system
was installed aboard the University of Miami Based on the above objectives, a modular networked
Rosenstiel School of Marine and Atmospheric science [ETHERNET) system was selected, using Digital
(RSMAS) Research vessel COLUMBUS ISELIN. This paper Equipment Corporation's (DEC) microVAX family of
presents several aspects of the installation, computers.
including design approach, reliability, personal
computer networking, file serving, personal GETTING STARTED
computers as interfaces, instrument interfacing and
low cost disk sharing. The original system was ordered in early December
of 1986, and was delivered during the Christmas
holidays. Because the operating system software
BACKGROUND had not yet been delivered, a system was copied
from an existing VAX on campus; a programmer from
Prior to the installation of the present computer URI installed a CTD data acquisition package; and
system, the computing capability aboard R/V the ship departed for a 90-day CTD cruise on
COLUMBUS ISELIN was limited to several HP-85 January 2. During that cruise, one problem was
computers and an IBM AT compatible. Data was being encountered with the hardware, due to a crimped
collected and stored in a variety of formats. wire inside the chassis. Since that time, there
This resulted in a fragmented data base which have been no other hardware failures at sea.
prevented realtime inspection and analysis of data. Subsequent to the original installation, several
After using these small computers for several years enhancements have been made, but the general
to perform various individual tasks, it became concept remains the same.
apparent that a Centralized Integrated Data System
(CIDS) was required for logging, processing, and SYSTEM WERVIEK
analyzing a wide variety of data sources, with
realtime display for underway decision making and The system consists of two networked microVAX
for post-cruise analysis of surface atmospheric and computers with remote access and shared resources.
oceanographic parametersi A modular approach was the key used to allow for
reliability, flexibility, and reduced acquisition
DESIGN CRITERIA and maintenance costs. Figure I is a block diagram
As a result of the limitations experienced with the CPU
smaller computers, the following criteria were Ram merriory
established:
Ram memory
- system accessibility provided to the scientists Video Driver
on board as well as being able to connect their Video Rani
own computers to the system. Ethernet

- Present shipboard electronics technicians could Buss Extender
maintain the system.
Extended Buss
- A high degree of hardware reliability which would SCSI Adapter
include the ability to recover from hardware
failures with a minimum of expertise. Real Time Clock
8 Line Serial Interface
- Hardware and software compatability with 9 Track Interface
computers at RSMAS so that the existing, in-house IEEE-488 Interface
knowledge and experience base could be used. A/D - D/A
- Aquisition costs, as well as continuing hardware
and software maintenance costs, should be Figure I
reasonable.

CH2585-a/8e/oooo. 655 $1 @1988 IEEE

showing the basic individual computer configuration - SCSI mass storage devices are considerably less
while Figure 2 shows the the overall system. expensive than traditional DEC mass storage
devices.
The computers are powered from an Uninterruptible
Power Supply (UPS), which is monitored by the - A great deal of redundancy and flexibility is
computer. If the ship's power fails the computer achieved as either computer may access the
informs the user of run time remaining on the UPS devices on the SCSI buss.
once per minute, until only ten minutes remain
then the computer shuts itself down gracefully. At this time a Digi-Data 1600-bpi 9-track tape
Upon restoration of power, the computer reboots and drive is connected to one of the microVAX
selected software is restarted. The computers and computers. Future plans call for the this tape.
peripheral mass storage devices are shock-mounted drive to be connected to the SCSI buss so that'
in two fiberglass cabinets having water-resistant either computer may access the drive directly.
covers, which are installed when being loaded or
unloaded from the ship. Terminal Servers
Computers Terminal servers are installed on the Bridge, in
the main lab, and on the lower deck of the R/V
Two computers were installed for redundancy. Each COLUMBUS ISELIN where the scientist staterooms are
is configured in two modular sections, the first as located. These servers provide serial RS-232
a DEC microVAX 3200 diskless workstation, and the communication with the computers; each server has
second as an expansion chassis also manufactured by eight serial lines. Terminal servers have several
DEC. This configuration takes advantage of the advantages over traditional computer-based serial
educational discounts being offered by DEC while interfaces:
obtaining the same circuit boards as those used for
the standard microVAX. The workstation section - Resources, such as printers, plotters and RS-232
contains the CPU, RAM memory, ETHERNET interface, sensors may be used by either of the computers on
video driver, and video RAM. The expansion chassis the network.
which has an identical power supply and backplane
as the first section, contains an 8-line serial - Terminal servers also allow connection to either
interface, a host adapter which connects the computer in the event of computer failure.
microVAX to a Small Computer System interface
(SCSI) buss, a hardware clock, and other boards as - A somewhat reduced CPU load as compared to direct
required. The two sections are connected by a connection to the computer serial ports.
ribbon cable and are readily disconnected.
- Wiring is simplified; instead of running eight
Either computer can be used interchangeably; one is serial lines to the computer, a single cable is
normally used for realtime data collection and run to each server which is located near the
display, the second for data processing and as a terminals of interest.
hot spare for the first.
Printers
Disk and Tape Drives
Two line printers are connected to the main lab
Each computer has a SCSI host adapter, which terminal server,,one ison a print queue the other
physically connects the microVAX computer buss may be allocated for realtime CTD printing. A
CQ-buss) to a SCSI buss, and provides emulation of laser printer on a print queue is also connected to
traditional DEC disk and tape drives. Both the main lab terminal server.
computers are connected to a common SCSI buss.
cable, allowing all the devices. attached to the SYSTEM NETWORKING
buss to be accessed by either computer. TWO
Control Data Corporation WREN IV 300 MB disk drives The two system computers, terminal servers and IBM
and an Exabyte 8 mm tape drive are attached to the compatible computers normally installed aboard the
SCSI buss. The disk drives and tape drive are ship are connected together by ETHERNET. Additional
housed in a separate chassis with independent power computers may be connected using ETHERNET or RS-232
supplies for each drive. The advantages. of the serial connections. A DEC ETHERNET Local Network
SCSI bussed storage peripherals are; Interface (DELNI) allows eight connections with
- Higher performance another connector allowing additional DELNIs to be
daisy-chained. Connection to shore-based computers
is'accomplished through INMARSAT at 2400 baud or
The tape drive stores 2.2 GB of data with a through Malabar, Florida, via ATS-III satellite at
streaming mode throughput of 189 KB per second 300 baud.
as compared to DEC's TK50 tape drive, which
stores 95 MB with a throughput of KB per second. IBM Compatible Networking

* The WREN IV disk drive has a capacity of 300 MB Currently installed on the network is a software
with an access time of 16.0 ms as compared to package from DEC, VAX/VMS Services for DOS. This
DEC's RD54, which has a capacity of 151 MB with software provides IBM compatible computers Virtual
in access time of 38.3 ms. Disk Service, File Service, Printer Service, and
DECnet-DOS Service from the system computers.
Virtual Disk Service provides access to space on a

656

@Dopple, n
"or
EB SAE,

Dry Bulb Terna
"SINBAD" Wet Bulb TgM.
Solar Radialion Inst.
CPU Solar Radiation Integ.
User Supplied Ram memory Fluarometer #1
PC Ram memory Fluommeter #2
Video Driver FluoTometer #3
Video Ran Depth
Bridge Terminal Ethernet Depth Sweep
Server I Sea Surf. Temp.
Dry Lab Tenn. #1 Buss Exten@;-T Sea Surf. Sal.
UPS #1 Ship Speed
UPS #2 Ship Heading
Dry Lab Term #2 Extended Bug Barometric Pressure
Available SCSI Adapt Port Wind Speed
Laser Printer Real Tune Clo Port Wind Direction
8 Line Serial interface Stbd. Wind Speed
Stbd. Wind Direction
Electronics Lab Temp.
Electronics Lab Hurnid.
Lab Terminal Hydro Van Temp.
Server Hydm Van Humid.
Loran #1 Winch I.D.
Loran #2 300 mbyte 300 mbyte Winch Tension
GPS Satnav Loc k1 Ethernet Disk Drive Disk Drive XBT Launch ID
Transit Sathav I iterface 12 &byte
Omega
Winch Oper. Terun. Tape Drive
Available SCSI Chassis
Available

Stateroom Terminal
Server n
Tech Cabm s.1.
SE. Cabin Modem
Sci. Cabin
INMARSAT
MO&M
-96. -Cabin "POPEYE"
Sci. Cabin
Sci. Cabin CPU
I Sci. Cabin _j Ram memory
Ram memory
Video Driver
Video Rm
Ethernet

D/A Buss Extender
Chi wiwh C.Ow
Ch2 -
AM Extended Buss
Chl - Wimb 7@ndoa 9 TrWk T04X
02- SCSI Adapter
03- Real Time Clock Drive
Cb4-
W@ 8 Line Serial Interface
Mr.
9 Track Interface
IEEE-488 interface
AID - D/A

FIGURE 2, CIDS BLOCK DIAGRAM
Serial *1.rfw.

rf.
'I 7nec e
Interface

657

micrdVAX disk for use as a DOS-formatted remote IBM-compatible can be stored directly on the
disk. Application software can be maintained on microvAxs. This allows the data to be collected
this disk and the IBM compatible computer can boot continuously by the IBM-compatible computers and to
from this disk. Performance from this virtual disk be backed up on a regular basis using the standard
exceeds that of a hard disk connected directly to back-up facilities on the microvAX. The files may,
the IBM-compatible computer. File Service provides be accessed read only by the microVAX for realtime
access to files stored on the microVAX. This is data manipulation.
similar to the Virtual Disk Service above but is
slower. The advantage of File Service is that the INSTALLED SOFTWARE
fIles are readable by either the microVAX or the
IBM-compatible. Printer Service provides IBM With microVax workstations sold to educational
printer emulation on networked DEC pirinters for institutions DEC includes an Educational Software
IBM-compatible computers. Finally DECnet-DOS Library (ESL). Although this library contains more
provides access to mail, phone, file transfers, and than 20 software licenses which must be renewed
other network utilities. yearly, it still offers significant savings. , The
following ESL software is presently installed on
Other Computer Networking the system:

Accommodation of other computers not capable of -Operating System: VMS
using DOS or DECnet, such as Apple, Sun, and
Hewlett Packard, etc., may access the system using -Languages: FORTRAN, BASIC, PASCAL, C
public domain software such as Kermit or Xmodem.
(DEC has announced , plans to produce Apple- -Tools: VAX Language Sensitive Editor, VAX
compatible services similar to the VAX/VMS services Performance and Coverage Analyzer, VAX DEC/CMS
for DOS.) (Code Management System), VAX DEC/MS (Module
Management System), VAX DEC/Shell (Unix overlay),
INSTRUMENTATION INTERFACING VAX DEC/Test Manager Source Code Analyzer,

Once the new system was installed, the interfacing Communications: DECnet-VAX, VAX/VMS Services for
?f various shipboard scientific and navigational DOS
instrumentation was undertaken. Previously, much
of the scientific and navigational data was Graphics; VAX GKS (Graphic Kernel System)
collected using an HP-85 computer and Serial ASCII
Instrumentation LOOP (SAIL) modules designed by VAX Information Architecture: VAX DATATRIVE, VAX
Oregon State University (OSU). Previous experience FMS (Forms Management System), VAX Rdb/VMS, VAX
with this system had been good but because of baud Common Data Dictionary
rate limitations of the OSU SAIL modules, it was
impossible to sample individual sensors at high of f ice Tools: WPS-PLUS/VMS (Word Processing
repetition rates. Elimination of SAIL serial System), VAX Notes
sensors with long data strings permitted a ten-fold
improvement (from 60 seconds to 6 seconds) in the Additionally, software licenses have been.purchased
time required to serially poll every sensor for VAX LAB which provide 1/0 routines and dat3a
remaining on the loop. The SAIL system is now processing and graphics routines. Sof tware has
connected directly to one of the microVAXs, which been developed in house to control and display data
performs all loop control and logging functions. f rom, the SAIL loop. CTD data collection software
Most serial devices are now interfaced through I'las been acquired from the University of Rhode
terminal servers. This allows either computer to Island and CTD data processing software has been
access the sensor, with sampling speed only limited acquired from NoRDA.
by the sensor's ability to output the data. A few
serial instruments which are physically located USER INTERFACE
nearby are connected directly to the microVAX
instead of through terminal servers. (This is a The CIDS system is useless unless the scientist on
matter of convenience and not due to any special board has physical access and can learn how to use
requirement.) Analog sensors requiring very high it. Scientists are provided with their own
sampling rates are interfaced directly to the computer accounts. A computer-based instruction
microVAX through an 8-channel A/D board which has a course is on the system so that they may learn the
50-kHz sampling speed. necessary commands to perform the tasks needed to
meet their requirements. A shell is placed on top
Interfacing of Instrumentation Controlled by IBM of the operating system so that users familiar with
Compatible Computers UNIX may use familiar commands. Terminals can be
located in either the labs or cabins, and
Instruments on board, including the RD Instruments user-supplied IBM compatible computers can be
Doppler Current Profiler, and the LI-COR Spectral installed temporarily. This user interface
Radiometer, have IBM-compatible computers as provides the scientist with the capability of
controllers. Because it was not desirable or leaving the ship with a 9-track tape written in
practical to rewrite the software to interface VAX/VMS format or other industry standard formats.
these instruments directly to the microVAX system, Data can be transferred to DOS 5-1/4" floppies of
the IBM-compatible is used as an interface. Using varying densities or scientists may connect their
vAx/vMS services for DOS virtual Disk Service and own computers to the system and transfer data using
File Service, data being collected by an VAX/VMS Services for DOS or public domain software.

658

ACQUISITION AND MAINTENANCE COSTS ACKNOWLEDGMENTS
Figure 3 shows the cost to duplicate the system and I would like to thank the National Science
the annual hardware and software maintainance Foundation and, in particular, Larry Clark of the
costs. office of oceanographic Facilities Support for
providing the funding for this system. I would
Acquis. Yearly like to thank Otis Brown, Robert Evans, Jim Brown,
and Angel Li of the RSMAS Remote Sensing Group, and
Cost Maint. Hank Poor and Grant Basham of the RSMAS Computing
MicroVAX 3200 Workstation Center for their patience and assistance in this
undertaking. Finally, I thank Bill Hahn and his
(Diskless) (2) $27,000 $4,800 technician group at URI for their expertise, which
Expansion Chassis (2) 5,000 260 aided greatly in the design of this system.
9-Track Tape Drive/Interface 6,000 1,164
8-Line Serial Interface (2) 2,500 260 REFERENCE
SCSI Host Adapter (2) 3,200 360
WREN IV Disk Drives (2) 4,800 720 IEEE Standards Board, "IEEE Standard Serial ASCII
Exabyte Tape Drive 3,200 540 Instrumentation Loop SAIL Shipboard DATA
SCSI Peripheral Chassis 1,200 120 Communication" ANSI/IEEE Std 997-1985, 1985, l2pp.
Line Printers (2) .2,600 600
Laser Printer 3,600 600
Terminal Servers (3) 9,600 828
DELNI 1,100 96
miscellaneous cables 1,500 . -
Educational Software Library - 1,000
Additional Software 1,500 450
Documentation Service - 900

TOTAL $72,800 $12,698

one-year warranty on hardware, maintenance
not required first year.

Figure 3

SUMMARY

After almost two years of operation, with virtually
no incidences of failure, reliability has exceeded
expectations. The modular approach to the system
has allowed an upgrade from a MicroVAX If to a
MicroVAX III with no additional capital outlay.
Intra-institute sharing of inhouse-developed
software for this system, and a similar system on
board R/V ENDEAVOR has made it possibleto make the
best use of limited funds available for
programming. In short, scientists and system
operators are pleased with the system.

659

Validation of Computer Model Predictions of the Large-Scale
Transient Dynamic Towing Response of Flexible Cables

John D. Babb

Naval Underwater Systems Center, Newport, RI U2841-5U47

ABSTRACT

The large-scale transient dynamic response of tow point and intermediate locations, was measured
cables is being investigated by the U. S. Navy using commercial and custom-designed submersible
using several large computer models. Dynamic strain-gaged transducers. Experimental measure-
position and tension data gathered during recent ment uncertainties were determined to be + 1.4
full-scale instrumented towing tests are compared meters for transient dynamic position dat7a and +
to analytical predictions for the purpose of U.5 meters during less dynamic steady-state tovii-ng
validation. Comparisons of predicted and experi- conditions. Tension measurement uncertainties are
mental data indicate acceptable validation, with + 45 N pounds for tow point tension and + 1b N for
the exception of slack end effects at the free- fransducers located within the cable.
end. Results are presented for two computer
models and two large-scale transient conditions. DESCRIPTION OF PHYSICAL CONFIGURATION
Analytical shortfalls and future areas of
investigation are discussed. The physical validation experiment was estab-
lished using a weighted cable framework attached
INTRODUCTION to the Systems Measurement Platform (SMP) barge at
Seneca Lake (figures I and 3 ). The instrumented
The dynamics of flexible cables have been of cable was hung in a catenary shape from two elec-
interest to aeronautical and ocean engineers pri- tric releases near the middle of the H-shaped
marily from the standpoint of steady-state or channel. Power and signal transmission was
small-disturbance towing problems. However, the accomplished by connecting the instrumented cable
large-scale transient dynamic response of cables to the tow vessel by a 94.5 meter Kevlar-
is of interest to several U. S. Navy problems -reinforced electromechanical cable.
concerned with array mooring deployment, under-
water buoy release, and underway deployment of The instrumented cable was constrained by de-
tethered vehicles. Several analytical models sign to approximate the position of an operational
exist which are capable of predicting the tran- system which had been deployed (payed out) from a
sient dynamic cable response; however, experi- submarine and was ready for separation (breakaway)
mental data for validation and verification and subsequent towing. The upper end of the in-
purposes has typically been limited to laboratory strumented cable and the electromechanical cable
measurements of shortened cable systems. were located at a depth of 15 meters. The lower
end of the instrumented cable (breakaway or
Transient dynamic position and tension data free-end) was positioned at 30 meters depth and
was acquired during full-scale instrumented approximately 49 meters forward. In reality, the
towing tests at the NUSC Seneca Lake Acoustic cable configuration came to equilibrium at slight-
Test Facility, Dresden, NY (references I and 2). ly lower depths and horizontal separation distance
The objective of the 1986 experiment was to between each end (figure 4). In addition, some
measure the dynamic position and tension response repositioning in the Y-direction (out of the
of an instrumented flexible cable during trans- vertical X-Z plane) occurred as a result of
ient towing. The initial cable configuration unequal cable lengths and weights.
(figure 1) was designed to approximate the
position and behavior of an operational U.S. Navy The scenario for the tow tests was for the tow
system which is deployed and towed from a vessel to commence its run from the center of the
submarine (figure 2). south notch of the SMP (figure 3). The tow vessel
would accelerate, within 5 seconds, to the desired
Transient dynamic position data was acquired speed, towing in a southerly direction. Prior to
with acoustic tracking techniques using custom- the electro-mechanical cable pulling taut, the
designed hardware. Cable tension history, at the instrumented configuration would be

660 United States Government work not protected by copyright

released from the rigging and the cable system steady-state validation but must be processed
would be pulled, upper end over the lower end, first. Attempts to process a 5 meter/second
through the transient course. Following com- transien/steady-state tow run were limited as a
pletion of the transient phase of the tow run, result of monetary and technical difficulties. It
the tow vessel would make a large radius turn and is anticipated that the need for validation data
proceed through the steady-state tow course. In at tow speeds greater than 3.40 meters/second will
this manner, it was believed possible to acquire be sufficient to generate additional funding.
good quality dynamic position data during both
transient and steady-state tow conditions. Each acoustic configuration was instrumented
with a tow point tension cell at the tow vessel
Three instrumented cable configurations were and at the head end of the instrumented cable.
employed during the validation experiment (two The primary configuration additionally had a
for acoustic measured position and one for pressure transducer attached at the head end for
tension measurements). Tables 1, 2, and 3 list depth indication. Problems related to signal
the physical parameters of the tow configura- cross-talk between the acoustic and tension in-
tions. The parameters are based on physical mea- strumentation resulted in an inability to acquire
surements of the tow configurations and are suit- simultaneous dynamic position and tension data.
able for use during computer model predictions. However, a third tow configuration was instru-
mented with tension load cells at the tow point
EXPERIMENTAL TEST DATA and first, fourth and sixth pinger positions for
measurement of tow-induced tension at those loca-
Fifteen instrumented tow tests were conducted tions. This configuration was similar in hydro-
during the 1986 experiment. The primary acoustic dynamic shape and construction as the primary
configuration was employed for only one combined acoustic configuration. Figures 14 and 15 present
transient/steady-state tow run before it experi- the processed tension data for two 2.6
enced electrical failure as a result of over- meter/second transient tow tests. The timing of
straining the electrical connections. The cable relsease and tow speed variations are
acoustic position data was subsequently processed similar to those in Table 4. No data was measured
(figures 5 through 8) with indication that only for the free-end (sixth pinger) position as a
the upper three pingers were operational during result of failure of the transducer wiring.
the tow run. Transient results presented in the
figures show that the tow vessel remained on a COMPUTER MODEL PREDICTIONS
straight course, per test plan. Table 4 presents
a summary of the tow run (TA-1). Steady-state The transient dynamic behavior of the U.S.
tow data consisting of approximately 10 seconds Navy operational system is being modeled using two
(20 contiguous data points) were processed. The large, 3-dimensional, straight segment cable
experimental data indicated a steady-state tow models which were originally developed to solve
speed of 3.48 meters/second and a critical tow other marine and U. S. Navy problems. The Naval
angle of 15.09 degrees in the vertical (X-Z) Civil Engineering Laboratory (NCEL) SEADYN code
plane. (reference 3) is a 20,000 line finite element code
which was originally written for ship mooring
The back-up acoustic configuration was used applications. The University of Cincinnati
for six combined transient/steady-state tow runs UCINCABLE code (reference 4) is a 75UU line rigid
and four steady-state tow runs. Tow speeds body dynamics model which is formulated from
ranged from 2.6 to 7.7 meters/second. The 3.04 Lagrangian dynamics. This code was developed for
meter/second (TB-3) tow run data was processed to mine sweeping applications.
provide initial data for computer validation
(figures 9 through 13). Review of figure 13 in- SEADYN Predictions
dicated that the tow vessel did not complete a
straight tow and therefore created a more The SEADYN cable dynamics model is the most
difficult validation test case than a strictly capable of the two simulations used for this
2-dimensional tow. The timing and general validation effort. Due to the nature of finite
description of the motion of the tow vessel is elements, the model can handle the fixed con-
presented in Table 5 for use during computer straints which are required boundary conditions
model validation. Acoustic problems related to when modeling the validation configuration (figure
inadequate Signal-to-Noise Ratio (SNR) resulted 1) and the moving constraints of the Navy opera-
in the opportunity to track only the first, tional problem (figure 2). The physical
fourth and sixth acoustic sources in the backup parameters of Tables 1 and 2 and the run geometr-
configuration. Attempts to process the steady- ies of Tables 4 and 5 were used for simulation of
state tow portion of test TB-3 were reduced to runs TA-1 and TB-3, respectively. Figures 16
single data points as a result of the SNR and through 22 depict SEADYN predictions of dynamic
Seneca Lake sound propagation characteristics. position and tension for test TA-1 (Run #21B).
This data is not adequate to provide validation This run required approximately 20.5 days of CPU
for steady-state tow response. on a VAX 11/785 and 17 hours of CPU on a CRAY
XMP-28.
Additional back-up acoustic configuration tow
data is available for further transient and

664

Figures 23 through 29 depict SEAUYN predictions configuration closer than the UCINCABLE run.
for test TB-3 at the expense of 25.16 days of CPU However, the UCINCABLE tension predictions compare
on a VAX 11/785. with the tension module instrumented runs TT-1.and
TT-2 better than SEADYN results do@ Both SEADYN
Comparison of the SEADYN predictions and the and UCINCABLE have trouble with the prediction of
experimental data for runs TA-1 and TB-3 indicate free-ended, slack cable response. It should be
that the majority of the dynamic efforts are ac- noted that the intermediate tension module data of
counted for with the exception of the cable re- figures 14 and 1b should be treated with caution.
sponse at the free-end. The slack end response This comment is provided since the head-end and
of SEADYN has been addressed in both references 3 intermediate transducers were separated by 3U.5
and 5. The problems with the numerical model are meters on the 53.8 meter cable. Tension values
due to inadequate discretization of the cable and should be proportional.
the presence of perfectly flexible joints in the
numerical model. The physical validation experi- Steady-state predictions of SEADYN provide
ment configuration contained more opportunity for better position comparison with the experimental
slack cable effects as a result of the reposi- data than UCINCABLE does. This continuing agree-
tioning depicted in figure 4. ment between SEADYN*and the physical experiments
in respect to cable orientation, indicates that
SEADYN predictions for the steady-state por- the SEADYN model is better suited for modeling the
tion of test TA-1 (Run #21BSS) consisted of a U.S. Navy operational system if accuracy is a
vertical plane tow angle of 12.51 degrees and a concern.- The UCINCABLE model is not as accurate
tow point tension of 890 N for the 3.48 meter/ but is capable of prediction behavior much quick-
second tow run. These values compare well with er. This relationship was previously developed in
the experimental data (figure 3U). reference 6 for several different problems.
UCINCABLE Predictions The tension comparisons of UCINGABLE with the
physical experiment and the poor showing of the
The UGINCABLE cable dynamics model is capable SEADYN model provide an interesting comment.
of predicting the towing response of cable sys- Since both models utilize the same hydrodynamic
tems in an "open chain" configuration. This term drag coefficients the differences in tension
implies that no two points on the modeled con- predictions must be caused by the interaction
figuration are constrained and no closed loops of between material elastic moduli and numerical
cable exist. As a result of this problem, SEADYN stability in the SEADYN model.
output at the time of release (9.75 seconds) was
employed to generate the initial UCINCABLE input. The need for additional validation data for
A 2-dimensional preprocessor was used for con- comparison with the two computer models remains a
verting SEADYN data for run #21B (TA-1) to high priority for future research. The SEADYN
UCINCABLE run #32. However, the 3-dimensional model has been installed on the NUSC CRAY XMP-28
tow response of SEADYN run #200 (TB-3) proved to and selected for optimization as a sample test
be a difficult conversion due to the use of case. The UCINCABLE code will be installed in
dextral angle relationships for the UCINCABLE order to take advantage of the available turn
input file. A suitable input file has yet to be around times between runs and to allow additional
generated. comparisons with SEADYN. NUSC efforts continue to
develop and validate computer models which can
The UCINGABLE predictions for run TA-1 (run provide quick, accurate predictions although the
#32) are presented in figures,31 through.37. This primary focus at this time is modeling of the
run was completed in approximately 38.1 hours on towed cable response from a maneuvering submarine
a VAX 11/785. At this date the UCINCABLE model at the conclusion of the transient fetch-up.
has not been converted for solution on the NUSC
CRAY XMP-28. The free-end response of the REFERENCES
modeled configuration has identical problems as
noted in the SEADYN data discussion. The tension 1. J. D. Babb, "Experimental Measurement and
response of the UCINCABLE model at 9.75 seconds Computer Validation of Flex-Hose Large-Scale
should be ignored since it is caused by the need Transient Dynamic Behavior (Interim Report)",
to generate an initial configuration and is not Naval Underwater Systems Center, Newport, RI , TM
representative of physical events. No. 87-2UU6, 19 February 1987.
UCINCABLE predictions of the steady-state 2. J. D. Babb, "Experimental Measurement of the
response of run TA-1 (run #31) indicated that the Large-Scale Transient Dynamic Towing Response of
UCINCABLE model continues to have difficulties Flexible Cables", ASME 3rd Annual Symposium of
with the cable position (figure 3U). UCINCABLE Current Practice and New Technology in Ucean
values of vertical tow angle and tow point ten- Engineering, ULU Vol. 13, January 19TT.
sion are 15.66 degrees and 560 N respectively.
3. R. L. Webster andP. A. Palo, "SEADYN User's
DISCUSSION AND CONCLUSIONS Manual", Naval Civil Engineering Laboratory, Port
Comparison of the experimental, SEADYN and Hueneme, CA TN No. N-1630, April 1982.
UCINCABLE results show that the SEADYN model
predicts the position of the instrumented tow
662

4. J. W. Kamman and R. L. Huston, Users Manual 6. J. W. Kamman and R. L. Huston, "Modeling of
for a Three-dimensional, Finite Segment Computer Submerged Cable Dynamics", University of
Code for Submerged and Partially Submerged Cable Cincinnati, Department of Mechanical and
Systems", University of Cincinnati, Department of Industrial Engineering, UNk-UC-MIE-U7U183-16, July
Mechanical and Industrial Engineering, 1983.
ONR-UC-MIE-050183-15, May 1983.

5. P. A. Palo, "Comparisons between Small-scale
Cable Dynamics Experimental Results and
Simulation Using SEADYN and SNAPLG Computer
Models", Naval Civil Engineering Laboratory, Port
Hueneme, CA. TM No. M-44-79-5, January 1979.

Table 1. Primary Acoustic Instrumen ted Tow Configuration
ELECTRO-MECHANICAL CABLE:
LENGTH 94.5 m
DIAMETER 1.43 cm
UNIT WET WEIGHT 0.85 NIM
YOUNG'S MODULUS X AREA 1.437 x 107 N
NORMAL DRAG COEF 1, ICIENT 1.200
TANGENTIAL DRAG COEFFICIENT 0.0046

INSTRUMENTED FLEX-HOSE:
TOW END TENSION @PINGER I PINGER 2 FINGER 3 PINGER 4 PINGER 5 PINGER 6
MODULE
LENGTH FROM TOW END TO
INSTRUMENTATION 1.07 M 1.47 m 10.74 m 23.11 m 32.36 m 44.75 m 54.07 m

OVERALL LENGTH 54.07 m
FLEX-HOSE DIAMETER 1.59 cm
INSTRUMENTATION ATTACHMENT LENGTH 0.46 m
INSTRUMENTATION DIAMETER -1.75 cm (average over attachment)
UNIT WET WEIGHT 3:.26 N/m
YOUNG'S MODULUS x AREA 1.370 x 10 7 N
NORMAL DRAG COEFFICIENT 1.389
TANGENTIAL DRAG COEFFICIENT 0.0259

Table 2. Backup Acoustic Instrumented Tow Configuration

ELECTRO-MECHANICAL CABLE:
LENGTH 94.5 m
DIAMETER 1.43 cm
UNIT WET WEIGHT 0.85 Nlm
YOUNG'S MODULUS x AREA 1.437 x 10 N
NORMAL DRAG COEFFICIENT 1.200
TANGENTIAL DRAG COEFFICIENT 0.0046

INSTRUMENTED FLEX-HOSE:
TOW END TENSION FINGER PINGER 2 PINGEli 3 PINGER 4 PINGER 5 PINGER 6
MODULE
LENGTH FROM TOW END TO
INSTRUMENTATION 1.32 m 2.01 m 10.54 m 22.60 m 31.82 m 44.01 m 53.08 m

OVERALL LENGTH 53.54 m
FLEX-HOSE DIAMETER 1.59 cm
INSTRUMENTATION ATTACHMENT LENGTH 0.61 m
INSTRUMENTATION DIAMETER 1.90 cm @average over attachment)
UNIT WET WEIGHT 3.41 N/m
YOUNG'S MODULUS x AREA 1.370 x 10 7 N
NORMAL DRAG COEFFICIENT 1.389
TANGENTIAL DRAG COEFFICIENT 0.0259

Table 3. Tension Module Instrumented Tow Configuration
ELECTRO-MECHANICAL CABLE:
LENGTH 94.5 m
DIAMETER 1.43 cm
UNIT WET WEIGHT 0.85 NIM
YOUNG'S MODULUS x AREA 1.437 x 10 7 N
NORMAL DRAG COEFFICIENT 1.200
TANGENTIAL DRAG COEFFICIENT 0.0046

INSTRUMENTED FLEX-HOSE:
TOW END TENSION INTERMEDIATE TENSION TAIL END TENSION
MODULE MODULE MODULE
LENGTH FROM TOW END TO
INSTRUMENTATION 1.37 m 31.88 m, 53.38 m

OVERALL LENGTH 53.79.m
FLEX-HOSE DIAMETER 1.59 cm
INSTRUMENTATION ATTACHMENT LENGTH 0.61 m
INSTRUMENTATION DIAMETER 1.90 cm (average over attachment)
UNIT WET WEIGHT 3.47 N/m
YOUNG'S MODULUS x AREA 1.370 x 107 N
NORMAL DRAG COEFFICIENT 1.389
TANGENTIAL DRAG COEFFICIENT 0.0259

6@63

Table 4. Summary of Primary Acoustic Transient Run TA-1
INITIAL FLEX-HOSE CONFIGURATION:
X Y z
TOW VESSEL TOWPOINT 51.81 m 0.00 M 0.00 M
FLEX-HOSE E/M CABLE JUNCTION
(TOW END) 4.33 m 2.82 m 19.29 m
LOWER END OF FLEX-HOSE
(TAIL END) 42.24 m 1.,29 m 33.94 111

TOW MEDIUM: FRESH WATER T 18.33 C

TOW POINT MOTIONS:

INITIAL VELOCITY (t = 0.0 sec.@ 0.00 M/S 0.00 M/S 0.00 M/S
TOW VESSEL ACCELERATION It . 0.0 sec.)
VELOCITY AFTER TOW VESSEL
ACCELERATION (t = 5.0 sec.) 3.41 m/s 0.00 st/s 0.00 M/S
VELOCITY AT END OF
TRANSIENT RUN (t = 80.00 sec.) 3.41 m/s 0.00 M/S 0.00 M/s
RELEASE TIMES:
RELEASE OF BOTH ENDS t= 9.75 sec.

Table 5. Summary of Backup Acoustic Transient Run TB-3
INITIAL FLEX-HbSE CONFIGURATION:
X Y z
TOW VESSEL TOWPOINT 51.81M 0.00 M 0.00 M
FLEX-HOSE E/M CABLE JUNCTION
(TOW END) 4.32m 2.83 m 19.29 m
LOWER END OF FLEX-HOSE
(TAIL END) 42.16m 1.30 m 33.93 m

TOW MEDIUM: FRESH WATER T 18.33C

TOW POINT MOTIONS:
X Y
INITIAL VELOCITY (t - 0.0 sec-) 0.00 M/S 0.00 M/S 0.00 M/S
TOW VESSEL ACCELERATION (t = 0.0 sec-)
VELOCITY AFTER TOW VESSEL
ACCELERATION (t = 5.0 sec.) 2.64 m/s 1.52 m/s 0.00 M/S
GRADUAL CHANGE IN TOW VESSEL.
DIRECTION (t = 5.0 sec.)
VELOCITY AFTER TOW VESSEL
DIRECTION CHANGE It = 9.75 sec.) 3.05 m/s 0.00 M/s 0.00 M/s
GRADUAL CHANGE IN TOW VESSEL
DIRECTION (t = 9.75 sec.)
VELOCITY AFTER TOW VESSEL
DIRECTION CHANGE (t = 12.25 sec.) 3.05 m/s -0.12 m/s 0.00 M/S
VELOCITY AT END OF
TRANSIENT RUN (t = 80.00 sec.) 3.05 m/s -0.12 m/s 0.00 M/S

RELEASE TIMES:
RELEASE OF TOW END t 9.75 sec.
RELEASE OF TAIL END t 12.25 sec.

4 IN @IUNCH

UPPER RELEASE
POINT BREAKAWA
---- ------------
15 m _-1

30 m
TRANSIENT

INSTRUMENTED
FLEX-HOSE LOWER RELEASE T STEADY STATE TOW
TOWLINE POINT
Figure 1. Transient and Steady-State Towing Range
ENTED
.S
@IN
T' M
FL. H @E OW'
@E H. @E
GUIDANCE WIRE
Profile
Figure 2. U.S. NavyApplication Addressed.by the
Validation Experiment

664

R
E
L 0
A ------ ------
T 15
1 30
V
E 45
D 60
1 75
S
P 90
L 105
A
C 120
E
M 135
E 150
*,,ft N 165 INTERMEDIATE (#3) PINGER X COORO
----- INTERMEDIATE (#3) PINGER. Y COORD
180 INTERMEDIATE (#3) PINGER Z COORD
M I I Ir TI T I I I I I f
0 4 8 12 16 20 24 28 32 36 40 44 48 52 56 60

RELATIVE TIME SEC.
Figure 3. Transient and Steady-State Towing Range Figure 7. Run #TA-1 Intermediate (43).Pinger 1-D
0 Overhead V 0 Track
E 7.5
7.5 R
T is
E/M CABLE 1 22.5
15 C
D A 30
E22.5 L 37.3
P D 45
INSTRUMENTE
0
T CABLE 1 52.5
H 30 S 60
37.5 (DEFORMED) T 67.5
M A
N 75
45 C 82.5
E/M CABLE (DEFORMED) E
.52.5 SEAD83 RUN #21 INITIAL CONFIGURATION W/OjRI2GING 90 PINGER X1 4.75 TO 43.75 SEC.
SEAD83 RUN #21 LIVE CONFIGURATION W/ RIG I N 97, 5 PINGER X2 4.75 TO 41.75 SEC.
60 M i05 PINGER X3- 4.75 TO 34.25 SEC.
I I I I I I I I I Ir r I I I Y
0 7:5 15 22.5 30 37.5 45 52.5 60 67.5 75 0 15 30 45 60 75 90 105 120 135 150

HORIZONTAL DISTANCE M. HORIZONTAL DISTANCE M
Figure 4. Representation of Initial Configuration Figure 8. 2-D (X-Z Plane) Track of Run #TA-1
R
R of Pigging E
E
L 0 L 0 ----------
A .. ......... ........ ....... A ----------------------------------------------- *---------
T 15 T 15 ----------------
1 30 1 30
V V
E 45 E 45
0' 60 0 60
1 75
75 S
S P 90
P 90
L L 105
A 105 A
C 120 C 120
E E 135
M 135 M
E 150
E 150 N
N 0 END (#I) PINGER X COORD
T 165 HEAD END (#I) PINGER x coaRO T 165 HEA
HEAD ENO (#I) PINGER Y COORD HEAD ENO (#I) PINGER Y COORD
180 HEAD ENO (#I) PINGrER JZ COORD
180 HEAD END 1) PINGER Z ICOORD rI
M r I I(#r I II I I I . T-__@ M I I I
0 4 8 12 16 20 24 28 32 36 40 44 48 52 56 60 0 4 8 12 16 20 24 28 32 36 40 44 48 52 56 60
RELATIVE TIME SEC. RELATIVE TIME SEC
Figure 5. Run #TA-1 Head-End (#l) Pinger J-D R Figure 9.. Run #TB-3 Head-End (#l) Pinger
R Track E 1-D Track
E L 0 ....... ........ .......... .......
1. 0 A
A ..... T 15 . ............
T 15 1 30
1 30 V
V E 45
E 45
D 60 0 60
1 75
1 75 S
S P 90
P 90 L 105
L 105 A
A C 120
C 120 E
E M 135
M 135 E 150
E 150 N - INTERMEDIATE (#4) PINGER X COORD
N INTERMEDIATE (#2 PI T 165 - ___.
T 165 )MNGERR X COORD INTERMEDIATE (#4) PINGER Y COORD
-@INSTRENTFD
UM
CABLE

INTERMEDIATE (#2) PI GE Y COORD 180 IrNTERM'DIATE #4),PINGER, Z COORD
COORO T - T. (I T
180 INTEMEYATFE (.#2),PINGER, @ r. M
M r T T T T 0 4 8 12 16 20 24 28 32 36 40 44 48 52 56 60
0 4 0 12 16 20 24 28 32 36 40 44 48 52 56 60
RELATIVE TIME SEC. RELATIVE TIME SEC
Figure 6. Run 4TA-1 Intermediate (#2),Pinger I-D Figure 10. Run *TB-3 Intermediate (#4) Pinger 1-D
Track Track

665

R
E
L 0 ....................... .............................................
A 1300
T 15 1200
1 30 ......... 1100 8900 N CAPACITY LOAD CELL
V ----- TENSION MODULE (HEAD END)
E 45 1000 TENSION MODULE (INTERMEDIATE)
60
0 N 900
1 75 S goo
S 1 700
P 90 0
L 105 N 600
A 500
C 120
E N 400
M 135 300
E 150
N TAIL END (#6) PINCER X COORD 200
T 163 . ..... TAIL END (#6) PINCER Y COORD 100
180 TrAILIEN' (f.6) PINGER Z ICOORD 0
H I _ , I 1 1 1 1 -
0 48 12 16 20 24 28 32 36 40 44 48 52 56 60 048 12 16 20 24 28 32 36 40 44 48 52 56 60 64 68 72 76 80

RELATIVE TIME SEC TIME SEC.
Figure 11. Run #TB-3 Tail-End (#6) Pinger 1-D R Figure 15. Tension History Run #TT"2
Track E
V 0 L 0
E 7.5 A
R T 15
T 15 1 30 ----------------
1 22.5 ...... V -- -----
C
A 30 .... . ... . E 45
L 37.5 0 60
0 45 1 75
52.5
I P 90
S 60 L 105
T 67.5
A C 120
N 75 E
C 82.5 M 135
E E 150
90 N
97.5 PINCER X1 1.75 TO 58.75 SEC T 165 HEAD END (NODE 10) PINCER X COORD
PINCER X4 1.75 TO 58.75 SEC ..... HEAD END (NODE 10) PINCER Y COORD
M 105 PINCER X6 1.75 TO 58.75 SEC - 180 HFADIE'D (Nr"Ej1O)jP'TGER Z COORD
I I r T T I M -1 -7-1 1 1 1
0 15 30 45 60 75 90 105 120 135 150 165 180 048 12 16 20 24 28 32 36 40 44 48 52 56 60

HORIZONTAL DISTANCE M RELATIVE TIME SEC.
Figure 12. 2-D (X-Z,Plane) Track of Run #TB-3 A Figure 16. SEADYN Prediction of Head-End (#l)
T E Pinger Response during Run #TA-1
R 0 L 0
A 15
N 0.75 T ----------
S 1 30 --------------
V 1.5 V
E 2.25 E 45
60
R
S 3
E 3.75 75
S
0 4.5 P 90
I L 105
S 3.25 A
T C 120
A E 135
N 6.75 M
C 7.5 E 150
E N OE 13) PINCER X COORD
PINCER X1 1.75 TO 58.75r SEC T 165 INTERMEDIATE (NO
8.25 ..... INTERMEDIATE (NODE 13) PINCER Y,COORD
PINCER X4 1.75 TO 58.75 SEC 180 INTEqMEDIATE (NODE,13@ PINCER Z COORD
M 9 PINCER X6 1.75 TO 58.75 SEC r r 7
I I I I i I I M I I 1 1 7-1
0 15 30 45 60 75 90 105 120 135 150 165 180 04812 16 20 24 28 32 36 40 44 46 52 56 60
HORIZONTAL DISTANCE M RELATIVE TIME SEC.
Figure 13. 2-D (X-Y Plane) Track of Run #TB-3 R Figure 17. SEADYN Prediction of Intermediate
E (#2) Pinger Response during Run #TA-1
L 0
1300 (A
T 15
1200 1 30 -----------------------
1100 8900 N CAPACITY LOAD CELL V
T 1000 ----- TENSION MODULE (HEAD END) E 45
E TENSION MODULE (INTERMEDIATE) D 60
N 900 1 75
S Boo S
I P 90
0 700 L 105
N 600 A
500 C 120
E
N 400 M 135
300 E 150
200
T 165 INTERMEDIATE (NODE 17) PINCER X COORD
100 ----- INTERMEDIATE (NODE 17) PINCER Y COORD
180 IN
rTEI@E'I'ATF "OD'1'7@ PINCER Z COORD
0 M I I I 1 1 7__@
0 48 12 16 2a 24 28 32 36 40 44 48 52 56 60 64 68 72 76 80 048 12 16 20 24 28 32 36 40 44 48 52 56 60
TIME SEC. RELATIVE TIME SEC.
Figure 14. Tension History Run #TT-1 Figure 18. SEADYN Prediction of Intermediate
03) Pinger Response during Run #TA-1
,J4

666,

R R
E E
L
A L 0
T15 A ...........................
T 15
V30 1 30 ------------ -------
------------ V
E45 E 45
060 0 60
175 1 75
S S
P90 P 90
L105 L 105
A A
C120 C 120
E E
M135 M 135
E150 E 150
N N
T165 TAIL END (NODE 27) PINGER X COORO T 165 HEAD ENO (MODE 10) PINGER X COORD
TAIL ENO (NODE 27) PINGER Y COORD HEAD END (NODE 10) PINGER Y COORD
180 TrAILIE'D, (NP'E,27), PINGER Z,COORD 180 Mr
M T_1 T EAO,EN_0, (NPDE,10),PINGER Z COORD
I I I H r I I. I I
0 4 8 12 16 20 24 28 32 36 40 44 48 52 56 60 0 48 12 16 20 24 28 32 36 40 44 48 5 2 56 60
RELATIVE TIME SEC. RELATIVE TIME SEC.
Figure 19. SEADYN Prediction of Tail-End (#6) RFigure 23. SEADYN Prediction of Head-End 01)
Pinger Response during Run #TA-1 E Pinger Response during Run #T13-3
V0 L 0
A ......................................
E7.5 15
R T
T15 1 30
122.5 V
C30 ... .. E 45
A ------------------
L37.5 - ------- - 0 60
45 75
0 S
152.5 P 90
S60 L 105
T A
A67.5 C 120
N75 E
M 135
C82.3 PINGER X1 9.75 TO 72.15 SEC. E
E90 PlNGER X2 9.75 TO 72.15 SEC. N 150
97.5 P'INGER X3 9.75 TO 72.15 SEC. T 165 INTERMEDIATE (NODE 21) PINGER X COORD
M --- PINGER X6 9.75 TO 72.15 SEC. So ..... INTERMEDIATE (NODE 21) PINGER Y COORD
105 - I INTERMEDIATF (MODE, 2'@ PIINGErR. Z COORDI
I I I I I I I I M II 1 1 T T -1 1 1
0 15 30 45 60 7.5 90 105 120 135 150 0 48 12 16 20 24 28 32 36 40 44 48 52 56 60

HORIZ. DISTANCE M. RELATIVE TIME SEC.
Figure120. SEADYN 2-D (X@Z Plane) Representation Figure 24. SEADYN Prediction,of Interme diate
of Pinger Motions during Run #TA-1 R (#4) Pinger
E Response during Run #TB-3
0 L
............................
A
T 15
30
V - -----
E 45
30
0 0 60
E
P 75
T45 S
H P 90
60 L
RELEASE POSITION T 9.75 SEC A 105
T .15.00 SEC. C L20
75 E
T -20.00 SEC . M 135
T -25.00 SEC. E 15D
90 T 30. DO SEC. N
T 35.00 SEC. T 165 TAIL ENO (NODE 29) PINGER X COORD
--- TAIL END (NODE 29) PINGER Y COORD
105 180 TA
rLIENO (NGOE,29)j PITN.GER Z COURO
M Ir _T I I I I
0 15 30 45 60 75 90 105 120 135 150 .0 48 12 16 20 24 28 32 36 40 44 48 52 56 60

HORIZONTAL DISTANCE M. RELATIVE TIME SEC.
Figure 21., SEADYN Snapshot,Representation of Figure 25. SEADYN Prediction of Tail-End (#6)
Instrumented Cable Tow Response during Run 4TA-1 V Ia Pinger Response during Run #TB-3
1300 E 7.5
1200
T 15
1100 1 22.5 ------------------------
C --- -------
T1000 A 30
E
900 L 37.5
N
Stsuu D 45
700 52.5
0
S
N600 60
T
67.5
500 -- -------------------- ...... ...... A
N 75
N400 C 82.5 PINGER X1 9.75 TO 70.05 SEC.
300 iii:T-1 E 90 PINGER X4 9.75 TO 70.05 SEC.
200 _-'r01fPM N'T TERS TUW 97.5 PINGER X6 9.75 TO 70.05 SEC.
100 ..... HEAD ENO TENSI
M
105
T1, il,,:, INTERMEDIATE H ION
D 2NSE TENS
IT
0 4 8 12 16 20 24 28 32 36 40 44 48 52 56 60 64 68 72 76 80 0 15 30 45 60 75 90 105 120 135 150

TIME SEC. HORIZ. DISTANCE M.
Figure 22. SEADYN Prediction of Cable Tension Figure 26. SEADYN 2-D (X-Z Plane) Representation
History during Run #TA-1 of Pinger Motions during Run #TB-3

667

T
R 0 0
A
N 0.75 R
S 1.5 E
V L 7.5
E 2.25 A
R T
S 3 1 15
V e
---------------
E 3. 75 E e,
D 4.5
I PINGER X1 9.75 TO 70.05 SEC D 22.5
S 5.25 PINGER X4 9.75 TO 70.05 SEC E
T PINGER X6 9.75 TO 70.05 SEC
6 P
A T 30
N 6.75 H
C 7.5
E 8.25 M 37.5 ACOUSTIC TEST TA-1
UCINCABLE RUN #31
M 45 SEAD83 RUN #21BSS
T r- I I I I
0 15 30 45 60 75 90 105 120 135 150 60 50 40 30 20 10 0 -10

HORIZ. DISTANCE M.
RELATIVE HORIZ. DISTANCE M.
Figure 27. SEADYN 2-D (X-Y Plane) Representation Figure 30. COMparison.of Stea 'dy-State Tow Angle
Of Pinger Motions during Run #TB-3 Predictions

R
0 E
L 0
A
15 T 15
I
30
V -------------------------------
0 30 E
E 43
D
P 45
T 1 60
S ---------
H P 75
60 L
RELEASE POSITION T 9.75 SEC A 90
T - 15.00 SEC. C
75 T - 20.00 SEC. E 105
T 25.00 SEC. M
90 T 30.00 SEC. E 120
T 35.00 SEC. N
T 135 HEAD END (NODE 9) PINGER X COORO
105 150 ..... HEAD END (NODE 9) PINGER Z COORD
0 15 30 45 610 M i I I I I I I I I I I I I. I,1
75 90 105 120 135 150 0 4 8 12 16 20 24 28 32 36 40 44 48 52 56 60

HORIZONTAL DISTANCE M. RELATIVE TIME SEC.
Figure 28. SEADYN Snapshot Representation of Figure 31.. UCINCABLE Prediction of Head-End
Instrumented Cable Tow Response during Run M-3 (#l) Pinger Response during Run 4TA-1

R
E
L 0
1300 A
1200 1
1100 V 30
T 15

T1000 45
E 900 D
N 1 60
5 800 S
1 700 P 75
0 L
N 600 A 90
500 C
N 400 -------------- E 105
C41 M 120
300 E
200 135 INTERMEDIATE (NODE 12) PINGER x C.OR.
100 'j___:-_--_-__T_d_WPOINT TENSION T ... INTERMEDIATE (NODE 12) PINGER Z COO.,
..... HEAD END TENSION 150
0 -i-,I'TTERrM "A@TE HOSE TENSAON
_j ET I . @ I I M
0 4 8 12 16 20 24 28 32 36 40 44 48 52 56 60 64 68 72 76 80 0 4 8 12 16 20 24 28 32 36 40 44 48 52 56 60
TIME SEC. RELATIVE TIME SEC.
Figure 29. SEADYN Prediction of Cable Tension Figure 32. UCINCABLE Prediction of Intermea-
History during Run #TB-3 iate (#2) Pinger Response during Run #TA-1

668

R
E
L 0 0
A
T 15
I ............ 15
V
30
E
45 30
D E
1 60
P 45
S T
P 75
L 60
A 90 RELEASE POSITION T 9.75 SEC
C M 0
E 105 1 - 15. 0 SEC.
M 75 T - 20.00 SEC.
E 120 1 25.00 SEC.
90 T 30.00 SEC.
N 135 INTERMEDIATE (NODE 16) PINGER X CO 1 35.00 SEC.
I ORD
150 INTERMEDIATE (NODE 16) PINGER Z COORD 105
M I I I - I I I --F-- I I I I I I I I I I
0 4 8 12 16 20 24 28 32 36 40 44 48 52 56 60 0 15 30 45 60 75 90 105 120 135 150

RELATIVE TIME SEC. HORIZONTAL DISTANCE M.
Figure 33. UCINCABLE Prediction of intermed- Figure 36. UCINCABLE Snapshot Representation
iate (#3) Pinger Response during Run #TA-1 of Instrumented Tow Cable Response during Run #TA-1

R
E
L 0 1300
A
T 15 1200
1 1100
V 30
E T 1000
45 E 900
0 N
1 60 S Boo
S 1 700
P 75 0
L N 600
A 90 500
C 105 N 400
E ......
M 300
E 120
N 200 ........ 7.--,.J.QWPQJNT TENSION
T 135 TAIL END (NODE 26) PINGER X COORD 100 ----- HEAd CWb-f0W@Y66F --------
..... TAIL END (NODE 26) PINGER Z COORD
1
0 0 INTERMEDIATE HOSE TENS:ON
51
0 4 8 12 IIS 20 24 28 32 36 40 44 48 52 56 60 0 4 8 12 16 20 24 28 32 36 40 44 48 52 56 60 64 68 72 76 80

RELATIVE TIME SEC. TIME SEC.

Figure 34. UCINCABLE Prediction of intermed- Figure 37. UCINCABLE Prediction of Cable Tension
iate (#6) Pinger Response during Run #TA-1 History during Pun #TA-1

E 7.5
R
T 15
22.5
A 30
L37.5 41-
D 45
152.5
S 60
T
A67.5
N 75
C82.5 PINGER XI 9.75 TO 79.70 SEC.
E
90 ----- PINGER X2 9.75 TO 79.70 SEC.
97.5 PINGER X3 9.75 TO 79.70 SEC.
M 105 P NGER X6 9.75 TO 79.70 SEC.
r I I I I
0 15 30 45 60 75 90 105 120 135 150

HORIZ. DISTANCE M.
Figure 35. UCINCABLE 2-D (X-Z Plane) Representation
of Pinger Motions during Run.#TA-1

669

SURFACE TET TR ENGINEERING MOORING (STEK)*

Henri 0. Berteaux, Daniel E. Frye,
Peter R. Clay, Edward C. Mellinger

Woods Hole Oceanographic Institution
ABSTRACT which would maintain the buoy on station
and would also provide a reliable signal
A Surface T61emetery Engineering Mooring path between the deep sensors and the
(STEM) has been developed to collect and surface. The mechanical,and electrical
problems of properly connecting these
transmit oceanographic and meteorological cables at the buoy attachment point and at
data via satellite links. Data the points of instrument insertion in the
telemetered included currents (from 50 and mooring line also had to be surmounted.
250 meters), water and air temperature,
wind, relative humidity, barometric The successful completion of the STEM
pressure, and various engineering project also depended on advanced buoy
parameters engineering for classical mooring design
and on specialized mooring logistics for
The unique aspect of the STEM design was deployment, servicing and recovery.
the use of electromechanical cable for
both the strength member of the mooring 2.0 RETROSPECTIVE
and the electrical connection between the
subsurface instruments and the surface The R/V ENDEAVOR, operated by the
buoy. University of Rhode Island, deployed the
The surface mooring was deployed 150 miles STE 'M mooring on November 21, 1987 at
south of Cape Cod in 2700 meters of 39 011N Latitude and 70 0OOW Longitude.
water setout in November 1987 and This location is near the well-known
retrieved in May 1988, it operated Site D, 150 miles due south of Cape Cod
successfully through the harsh (Figure 1). The water depth at the site
N. Atlantic winter. is 2672 -meters. The mooring configuration
is depicted in Figure 2.

1.0 OBJECTIVES
780 W 700 620
The purpose of the Surface Telemetry 460- 1 NOVA
Engineering Mooring (STEM) was to N SCOTIA
demonstrate the feasibility of collecting
and transmitting data via satellite A.
telemetry. STEM is designed to handle an
extensive suite of meteorological, PORTLAND
oceanographic and engineering data
obtained from sensors distributed on the BOSTON >
surface buoy and on the mooring line. The 1000_@
surface mooring was to be deployed well '@00
off-shore in deep waters and had to NEW YORK
survive the harsh environment of the
Bu Farm
wintry Northwest Atlantic.

This ambitious goal required both
electrical and mechanical engineering sit
support. The electrical engineering
effort placed emphasis on the 380-
modification and integration of existing
instruments and sensors with a system
controller. This controller ensures the
timely and sequential interrogation of the /I
7
sensors and the subsequent data processing CAPE HATTE
and transfer to buoy mounted, satellite
transmitters. The innovative mechanical
engineering contribution consisted mainly Figure I Site D Location
of the development and evaluation of
electromechanical cables

CH2sa5_s/88/ooo0_ 670 $1 @1988 IEEE

4/ ARGOS a GOES TRANSMITTERS
MET SENSOR PACKAGE
3m DIA. DISCUS BUOY a CENTRAL CONTROLLER

2m 5/8 chain
5 4 m 50m U19, 1/2"dia. E/M cable
VMCM (w/ TEMPERATURE) #11
200m 309, 1/2 dia. E/M cable
254m---->
_VMCM (w/ TEMPERATURE) #2
_W71
-375m U19, 5/ld'dia. wire rope f.
wag-

-375m 309, 5/16 dia. wire rope

400nn 3xI9, 5/16" dia. wire rope
lip
11-CONTINUITY METER
_50rn 309, 3/8"dia. E/M cable
1457 rn --->
11-ENGINEERING INSTRUMENT #1
100m ftl6l'dia. wire rope
1557 m 3.
500 rn 15/16" dia. nylon rope

2000 m ---->
500M 13/16" dia. nylon rope

2 GLASS BALLS Figure 3 STEM Buoy
120 rn 3/4" dia. nylon rope (adjustable)
E
BACK-UP FLIDTATION (34) 17" GLASS BALLS At 00 00 hours on January 1, 1988, a
zrn 1/2" chain software error prompted the controller to
ENGINEERING INSTRUMENT #2 collect data at an accelerated rate which
2m 1/2" chair, soon jammed the data buffer. On January 7
i--ACOUSTIC RELEASE the problem resulted in the cessation of
BOTTOM 5m 1/211 chain
20m P dia. nylon rope all telemetered data.
DEPTH 5m 1/2; chain
2672 m ANCH@ (6000 Ibs) On February 9,' 1988 a short (2-day) repair
cruise was made with the R/V IDA-Z, a
SITE "D" TEST: 39'10.8'N, 70*00. 55'E converted, fishing vessel from New
Bedford, MA. The buoy was boarded, the
SURFACE TELEMETRY MOORING (STEM) software problem was corrected "in situ",
the cabling polarity problem was fixed,
NOVEMBER 21, 1987 - MAY 2, 1988 and a new wind sensor was installed to
replace the original instrument which had
Figure 2 STEM Mooring damaged bearings. The vessel then
promptly and safely returned to Woods
This six month engineering test mooring Hole. On February 12 a major storm hit
Was equipped with telemetering current Site D.
meters set at 50 and 250 meter nominal
depths and a telemetering meteorological From the time of the repairs to the time
station mounted on a standard WHOI of recovery, done with the help of the R/V
3- meter discus buoy (Berteaux, 1976) OCEANUS on May 2, 1988, data from both
(Figure 3). Telemetry to shore was current meters and most meteorological
performed through ARGOS and GOES satellite sensors were continually received. The
links. The telemetry data stream included quality of these data is further discussed
hourly averages of current, water in a later section of this paper.
temperature, wind, air temperature and
relative humidity, barometric pressure and 3.0 MOORING AND CABLE DESIGN
buoy heading, as well as a number of
engineering housekeeping parameters 3.1 Electromechanical (E/M) Cables
(mooring line tension, cable integrity,
battery voltage, water leaks, etc... ). Numerous specifications for oceanographic
Recording engineering instruments were E/M cables have been published in the
also inserted at depths of 1400 and 2540 literature (A. Driscoll, 1982 and P.
meters respectively to monitor tension and Gibson, 1984), yet very few E/M cables
tilt in the mooring line. (Clay, 1987) have been designed and fabricated
From the time of deployment to specifically for deep sea mooring
December 31, 1987 all data channels were applications. Mechanically these cables
transmitted and received via ARGOS as must provide enough strength to safely
programmed, with the notable exception of maintain the mooring on station. To
the upper current meter. The cause of prevent conductor damage, the cable
this transmission failure was later modulus of elasticity must be high enough
identified as a polarity inversion in the so that the cable exhibits a minimum
cabling to the instrument. elongation when under tension. The
54

254

671

armor must resist corrosion and fatigue, Cables #3 and #4 are of contrahelical
and must be torque balanced to reduce the double armor construction. The armor of
danger of kinking. When deployed in the #3 is made of Galvanized Improved Plow
fishbite zone, typically the upper 2000m Steel (GIPS) whereas the armor of #4 is
of the water column, (Berteaux and made of KEVLAR. These cables are designed
Prindle, 1987) it must provide a fair to be used only on subsurface moorings
degree of protection to the conductors (A. Bocconcelli, 1987).
against fishbite attacks.
The strength member of cables # 1 and # 2
Modern, hard wire telemetry uses low is the well-known 3 X 19 oceanographic
voltage, low power digital signal rope. Over the years this type of rope
transmission. These needs are easily has been successfully used to anchor
accommodated with relatively small gauge hundreds of deep sea surface and
conductors, typically AWG #20 (.812mm) or subsurface moorings. It seems logical to
# 22 (.644mm). Conductor insulation, retain this rope design to produce an E/M
material and thickness must be specified cable suited for surface mooring
to not only satisfy the electrical applications.
requirements but also to ensure perfect
and easy bonding at the termination ends Using the 3 X 19 construction, three
as well as maximum conductor protection conductors can be placed in the center of
from mechanical abuse. the strands, as shown in cable # 2, or in
the valleys between the strands as shown
With these requirements in mind, four in cable # 1. The first approach provides
experimental E/M cables have been excellent mechanical and fishbite
developed and fully to partly evaluated. protection for the conductors. However,
These cables, together with some of their keeping the armor wires from damaging the
characteristics, are shown in Figure 4. conductor insulation during stranding
requires heavy insulation walls and great
c
are.
_X
V

X
AM
3 0,

(3) STRANDED COPPER
012, IRS (AWG #20)
CONDUCTO
0,
Xr
A
INSULATION:
0.015* WALL PVC
0.010* WALL NYLON

'A' A
3/fl" 3x19 TORQUE BALANCED
WIRE ROPE (Mc WHITE)

INNER JACKET.
POLYURETHANE
0.054" WALL
O.D. 0.500"

-4 505 STEEL
4
O.D. 0.51
TAPE @2 LAYERS)
6

OUTER JACKET.
HYTREL
0.047" WALL
O.D. = 0.610"
TION STRENGTH 1b3/1OOO' OU'131DE CONDUCTORS um
Fco;;;U@ (Ibe) - DIL
1 3 X 19 GAPS 14,800 360 0.610 (3) #20 AWG CONSOUDA10
2 3 X 18 GIPS 14.800 248 1 0.460 3 #22 AWO VAcWHYX Figure 5 STEK Cable Detail
3 STEEL ARMOR 10,300 257 0.478 (3) #20 AWC B11K
4 KEVLAR DOUBLE 8,300 94.1 0.475 (3) #20 AWC ILLW.
LAYER

Figure 4 Experimental Electromechanical
Cables

672

Cable manufacturing is somewhat easier in ends, the cable assemblies were tensile
the second approach. However, a fishbite proof tested and thereafter, subjected to
resistant barrier must be added to protect 3000 psi of hydrostatic pressure for four
the conductors now at the periphery of the hours. The conductors were then checked
rope structure. The STEM cable, shown in for continuity and insulation integrity.
detail in Figure 5 had a standard 3/8 inch The assemblies were pronounced fit for sea
3 X 19 strength member (UBS, 14,800 lbs). duty after successful completion of these
Three # 20 stranded copper PVC insulated tests.
conductors were laid between the strands .
and held in place with the help of a Mylar When deployed, the continuity of the two
tape. Polyurethane was then extruded over' upper E/M cable assemblies was monitored
the cable to provide an additional water by telemetry. The condition of the
barrier. Finally two layers of stainless lower assembly was monitored by a
steel tape were wrapped over the recording instrument which periodically
polyurethane jacketed core and a jacket of checked the continuity of its 3
hard Hytrel extruded over the entire conductors.
assembly, with a resulting outside
diameter of 0.610 inch. After retrieval the cable assemblies were
carefully inspected and tested for shorts,
3.2 Electromechanical Terminations opens and residual strength. No sign of
conductor damage or rope deterioration was
Two 50 meter and one 200 meter E/M cable found.
assemblies were deployed in the STEM
mooring (Figure 2). Each length was The four types of E/M cable shown in
terminated at both ends with terminations Figure 4 are scheduled to undergo an 18-
designed and built at WHOI. As shown in month deep sea evaluation, starting
Figure 6, the terminations are of the February 1989. This test will be part of
Clevis type for ease of connection with the Engineering Surface Oceanographic
existing instrumentation. The steel body Mooring (ESOM) program.
is made of two halves. The tapered half
contains the wires of the armor, which are
embedded in an epoxy pour. The Clevis 4. 0 DATA COLLECTION AND TELEMETRY
half screws into the tapered half and is
filled with polyurethane. The three The STEM system was equipped with
conductors are then placed in a protective environmental sensors to monitor ocean
polyurethane sheath and laid to a junction currents and temperatures, surface winds,
splice which connects them to a commercial air temperature, relative humidity and
pigtail. This design provides full cable barometric pressure. Currents were
strength and a double water barrier. measured with EG & G Vector Measuring
Current Meters (VMCMs) at depths of 55 and
ANTIROTATING PIN 257 m. These instruments recorded data
POTTING SOCKET POLYUR THANE internally and were equipped with FSK
CLE (Frequency Shift Keying) modems which
E/M CABLE allowed two-way communication with the
controller using two conductors of the
electromechanical mooring
cable. Average current and water
BENDING STRAIN RELIEF_BOOT EPO temperature values were recorded at 15
2 POLYURETHANE minute intervals and loaded into an output
rONDUCTOR CO E FROM CABLE JACKET buffer. Every 15 minutes, 7.5 minutes
BRANTNER PIGTAIL after a record cycle, the system
3 CONDUCTORS @ontroller interrogated each current meter
/_ individually using the SAIL protocol
(IEEE, 1985) and collected the most recent
MOLDED SPLICE (POLYURETHANE current component and water temperature
values. These 15-minute values were then
scalar averaged to form hourly averages
Figure 6'Experimental Electromechanical and loaded into the ARGOS and GOES
Cable Termination telemetry buffers.

Meteorological data collection was
performed by a Coastal Climate Weatherpak
instrument. This small, self-contained
3.3 Testing unit was modified for interrogation via
The 3 X 19 wire rope was made by McWhyte SAIL protocol in a manner analogous to the
Inc. The conductors and the E/M cable current mieters. The Weatherpak uses an RM
were fabricated by consolidated Products Young Wind Monitor to measure wind speed
Inc. Acceptance testing of the E/M cable, and direction, an Anderaa clamped compass
performed at WHOI, included strength and to monitor buoy heading, a Rotronics
rotation tests, and conductor resistance relative humidity and air temperature
measurements. After terminating both sensor and an AIR barometer.

673

Using the Coastal Climate menu-driven - Average, minimum and maximum voltage in
software, the Weatherpak was programmed to the primary battery (IOV).
collect 10-minute averages of each of the
meteorological parameters once each hour. - Average, minimum and maximum voltage in
It accomplished this by measuring wind the secondary battery (12V).
speed and direction every second for 10-
seconds and then computing a true vector - Average, minimum and maximum voltage in
wind velocity using a compass reading the 12 V charging circuit.
(buoy heading) every 10-seconds to relate
measured winds to true geographic winds. - Average, minimum and maximum current
Sixty of these 10-second averages were consumption..
then used to compute a 10-minute average
which was forwarded to the system - Controller internal temperature.
controller on request, once each hour.
- Communication error counts between the
In addition to the environmental system controller and each of the
measurements, various engineering external devices, i.e. the current
measurements were made on an hourly basis meters, the Weatherpak, the ARGOS PTT
and telemetered once or more per day. and the GOES transmitter.
Included in these engineering measurements
were: Each of these parameters was telemetered
one or more times each day. Table 1
- Average, minimum and maximum mooring illustrates the data format sent through
line tension directly beneath the the ARGOS data collection system. A
surface buoy. similar message containing 6-one hour
averages was telemetered via the GOES
- Continuity of the electrical conductors system.
in the mooring cable.

- Water presence or absence in the
instrument well.

Telemetry Message (hourly): 64hex characters (256 bits)

HHQNNNNEEEETTTTQNNNNEEEETTTT SSSDDDGGGAAATTTTHHHPPPP MMMMDDDD3CCCC

VMCM1 VMCM2 WPAK ENG DATA

Telemetered Data ExRlanation

HH GMT hour during which data was acquired.
Q Number of values included in the hourly average from
vMCM1.
NNNN Averaged value of north component of current from VMCM1.
EEEE Averaged value of each component of current from VMCMI.
TTTT Averaged temperature from VMCM1.
Q Number of values included in the hourly average from
VMCM2.
NNNN Averaged value of north component of current from VMCM2.
EEEE Averaged value of east component of current from VMCM2.
TTTT Averaged temperature from VMCM2.
SSS Average wind velocity.
DDD Average wind direction.
GGG Maximum 4-second wind gust.
AAA Buoy heading.
TTTT Air temperature.
HHH Relative humidity.
PPPP Barometric pressure.
MMMM Engineering value 1.
DDDD Engineering value 2.
3 Constant
CCCC Checksum. L

Table 1. Format of the telemetered data sent via Argos.

674

Telemetry of the data from the surface power usage and Service ARGOS charges.
.buoy to shore was accomplished using both Thus, a total of sixteen 256-bit
ARGOS and GOES transmitters. These messages were transmitted every 8 minutes,
systems were chosen because they are each message consisting of hourly averages
reliable, well,understood, easy to of the scientific parameters previously
implement and inexpensive. They do not mentioned. The engineering data were
require the user to maintain a special reported less frequently (Table 2). A
receiver station,and they provide coverage check-sum was used at the end of each
worldwide (ARGOS). Other telemetry message to edit bad transmissions.
schemes have been evaluated at WHOI under
this project and may be required in The GOES telemetry was handled in a
situations where data throughput similar way, but without modification to
requirements are higher (Briscoe, 1987). the standard system. A Synergetics 40
In this prototype experiment both ARGOS watt GOES transmitter and Master Control
and GOES systems had more than enough Module were interfaced to the system
capacity to handle the required data. controller. once every third hour the
GOES buffer was loaded with the most
In order to send hourly averages from both recent 6 hours of averaged data. Its
current meters, the meteorological package telemetry schedule called for a 1 minute
and various engineering data, it was transmission every 3 hours, so that loot
necessary to modify the standard data redundancy was anticipated.
Synergetics/FORE ARGOS
transmitter. This modification, In addition to the telemetry discussed
performed at WHOI, allowed us to send 16 above, the surface buoy was also equipped
times more data than is possible with a- with a secondary ARGOS transmitter with
stock transmitter. The modification, separate antenna and power supply in case
which is primarily a software revision to of damage to the primary unit. A Vector
the manufacturer's control software, Averaging Wind Recorder (VAWR) was mounted
provides a data buffer which stores eight on the surface buoy to collect weather
. 256-bit messages. Each message in this information for comparison with the
buffer is transmitted sequentially at 1 Weatherpak.
minute intervals and received by the,
satellite during a typical 10 to 15 minute
pass. In addition, the transmitter ID is
changed twice per minute, effectively
making it the equivalent of two
transmitters in terms.of data throughput,

Transmission Time (GMT) Parameters
Channel Channel 2

00 Primary Battery(Min.) VMCM1 Timeouts
01 Primary Battery(Max.) VMCM1 Comm Errors
02 Primary Battery(Aver.) VMCM1 Address Failures
03 1@ V Supply(Min.) VMCM2 Timeouts
04 16 V Supply(Max.) VMCM2 Comm Errors
05 16 V Supply(Aver.) VMCM2 Address Failures
06 Tension(Min.) 0006
07 Tension(Max.) 0007
08 Tension(Aver.) 0008
09 Cable Continuity(Min.) 0009
10 Cable Continuity(Max.) OOOA
11 Cable Continuity(Aver.) OOOB
12 Secondary Battery(Min.) WPAIK Timeouts
13 Secondary Battery(Max.) WPAK Comm Errors
14 Secondary Battery(Aver.) WPAK Address Failures
15 Tension(Min.) PTT Timeouts
16 Tension(Max.) PTT Comm Errors
17 Float Switch PTT Address Failures
18 Snunt Current(Min.) GOES Timeouts
19 Snunt Current(Max.) GOES Comm Errors
20 Snunt Current(Aver.) GOES Address Failures
21 Tension(Min.) Controller Resets
22 Tension(Max.) 0016
23 Controller Temp.(Aver.) Tension(Aver.)

Table 2. STEK Engineering Data

675

5.0 CONTROLLER AND POWER SYSTEM It also controlled the charging of the
secondary rechargeable power supply and
The system controller was designed and monitored the communication link between
built at WHOI to accommodate a wide range the controller and the peripheral devices.
of control applications requiring
collection of data from a variety of In operation the STEM controller performed
sensors, data processing and recording, its data collection, processing and 1/0
and telemetry (Mellinger, et.al, 1986). operations on an hourly cycle. The
It is an 8OC86 based machine designed to current meters were interrogated every 15
fit in a 6-inch ID pressure case for use minutes. These data were converted to
with in situ oceanographic instruments. engineering units, and averaged hourly and
It makes use of a back plane derived from piped to the storage buffer. The
the IEEE - 796 (Multibus) standard (IEEE, Weatherpak was interrogated once each
1983). hour, the data were re formated to
eliminate unnecessary information and
piped to the storage buffer. The various
engineering data (Table 2) were collected
ANTENNA hourly, averaged and stored along with the
external values in the storage buffer.
Every hour the data, were down-loaded to
ANTENNA the ARGOS transmitter for telemetry to
:3 satellite. Every third hour the GOES
transmitter buffer was updated.

The format of the telemetered data sent
- - - - - - - - via ARGOS is shown in Table 1.
CABLE TENSION
[IVSAILIOC SHUNT CURRENT Power for the STEM system was supplied by
- FLOAT SWTCH a 30V alkaline battery installed in the
CONTROLIER - 12V BATTERY instrument well. A DC/DC converter in the
- 16V SUPPLY controller provided 5V to the controller
0
1
< -30V BArTERY and the various engineering sensors and
CABLE RESISTANCE 12V to charge a secondary lead-acid
INTERNAL TEMPERATURE battery which consisted of 6 - Gates cells
in series. The lead-acid battery was
designed to provide the high current
Z required by the 40 - watt GOES transmitter
ILSIR@[email protected] W@@ during its 1-minute transmissions
scheduled every third hour. This 12
volt supply also powered the ARGOS
VMCM1 transmitter and the Weatherpak.
The current meters, the navigation light,
and the secondary ARGOS transmitter were
VWCM2 all powered by their own alkaline
supplies. Power usage by the STEM system
was about 1.1 watts on average exclusive
of the navigation light, current meters,
Figure 7 Block,Diagrain Data Acquisition and secondary ARGOS tranmitter. major
Telemetry System subsytem usage was approximately as
follows:

The controller was equipped with hardware - ARGOS PTT (2-IDs) = 0.33 watts
and software to handle all of the STEM 1/0
requirements (Figure 7). These included: - GOES transmitter (40-watts) = 0.60 watts
- SAIL communication with two current - Controller = 0.04 watts
meters sharing a single pair of
conductors. - Weatherpak = 0.13 watts
- SAIL communication with a modified Total 1.1 watts
ARGOS transmitter.
- SAIL communication with a modified 6.0 RESULTS
Coastal Climate Weatherpak. Two problems became apparent following
- RS232 interface to a synergetics GOES deployment. First, data from the upper
VMCM at 55m was not being telemetered.
7AW

transmitter This turned out to be a cable polarity
problem which was repaired on the repair
- A/D conversion for 8 channels of cruise.
analog input

676

one week after deployment the GOES maximum of about 11 repeats of the same
transmission was lost. A deficiency in hour. Typically each data point was
the GOES transmitter battery recharging collected 4 to 7 times. Use of a check-
circuit proved to be the cause of the sum to eliminate data corrupted by the
transmission loss. Prior to the telemetry link resulted in automatic
occurrence of this problem, when the elimination of about one transmission in
transmitter batteries were still fresh, six.
good quality data were transmitted and
acquired by GOES link. Meteorological data collected by the
Weatherpak was of much poorer quality due
Data quality from the current meters was to sensor,problems. The bearings on the
excellent throughout the deployment. On RM Young Wind Monitor corroded away within
the retrieval the VMCMs showed little wear 10 weeks of deployment. The replacement
and tear and their bearings were in good unit, which used modified bearings faired
shape. An example of the current meter better. but was still somewhat affected by
data is shown in Figures 8 and 9. In increased bearing friction. The
general the level of redundancy used in Weatherpak compass worked reliably as did
the data sent via ARGOS was more than the barometric pressure sensor, but the
enough to ensure that no hourly data were humidity and air temperature sensors
missed, other than one or two instances failed within a few weeks. The air
where Service ARGOS had a system problem. temperature sensor lasted about one week,
The level of redundancy in the messages and then apparently suffered a short
received at WHOI varied from a minimum of circuit. The humidity sensor showed a
two messages (for a single hour) to a gradual increase in humidity with time

STEM - teLemetered data from VMCM-1, depth 55 m.

(TqMPERRTUREI
12
Uj
Ll
cc 41
M 12 1-5 14 15 16 17 IS 19 M 22 . %23 24 25 26 27 28 29 30 31
L3 1 2 3 4 5 a 7 5 9 to 11

10 cm/sec urrent Vectors)

IVECTOR DIRECTIONJ
360
7
U)
W I: a '
W
W

0 19 20 21 22 23 24 25 26' V 28 29 30 31-1
12.345610910111213141516171

tVECTOR SPEED)
so-

40.

(n 30.
A
IL
20. a

10 v
V%'

1 2 3 4 5 6 7 a 9 10 11 32 13 14 25 16 17 IS 19 20 21 22 23 24 25 26 27 28 29 30 31
March 1988

Figure 8 Telemetered Current Meter Data
VMCM-l (55m)

677

STEM LeLemetered data from VMCM-2, depth 257 m.

C@ (TEMPERRTURE)
12

W
w
%o%o%o
LJ 1 2 3 1 5 6 7 8 9 10" 11 '02 13 14 15 is 17 10 19 20 el 22 23 241 25 26 27 28 29 30 31 -7
M

N
10 cm/sec (Current Vectors)

(VECTOR DIRECTION)
360
tj
u
a: 180'
uj
co

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 25 27 28 29 30 31

(VECTOR SPEED)
so.

10-

Lo 30- .1

20-
%

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 36 19 20 21 22 23 24 25 26 27 28 29 30 31
March 1988

Figure 9 Telemetered Current Meter Data
VMCR-2 (257m)

until it read 102% humidity all the time. the STEM design can provide reliable
Figure 10 shows typical telemetered data mooring and telemetry performance for
from the Weatherpak. periods of six months or longer in most
ocean environments. The performance of
Various engineering data were telemetered the meteorological package was
at regular intervals. Figure 11 shows a disappointing,and more sensor testing and
time series of some of the engineering evaluation is required before long-term
data. Of particular interest are the reliability can be assured.
tension measurements which were made The incremental costs for using
directly beneath the surface buoy. telemetering surface moorings rather than
Tension values averaged between 1000 and standard surface moorings with internally
2000 pounds and varied from instantaneous
lows of near zero to values exceeding 3000 recording instruments are not known in
pounds which occurred during periods of detail, but are primarily influenced by
high currents. the additional costs of the cable and
terminations, the controller and its
7.0 CONCLUSIONS special software, and the testing
procedures required to ensure that all
The STEM experiment was the first subsystems operate correctly. More
successful use of a general purpose detailed information on costs and long
electromechanical cable to moor a surface term reliability will not be available
buoy in deep water under very severe until more experience is gained with these
weather conditions. None of the critical systems.
components, namely cables, connectors,
instruments or buoy showed excessive signs
of wear and tear. Thus, it appears that

678

1030 (BarometrLc Pressure)
9901

gkpe-A@. A. jlolel- 'Aar
NRVRS@ - R7, //r/ RR @- 'N'T1 gill Py
10 m/sec (WLnd Vector.

360. (DIRECTION FROM WHICH WIND IS BLOWINGI
(n %ft%& %.:. - -
w
0@ lea-' dw
CD
Li
M
0 rI
1 2 3 4 7" 'ena910 1 @112 '1314 'IS16 a 1 @920 @2122 @232, @2525 @2728 29 3a 31

(VECTOR SPEED)
40-

30-

V7
20'

to-

0
5 %
1 2 -3 6 7
a910 11 12 13 ill 15 18 17 18 19 ZU 21 22 23 24 2S 26 27 2@ 29 30 31
March- I
Figure 10 Telemetered Atmospheric Data

STEM - teLemetered data
3000. (TENSION)
2000-x xxxxxxx xxxxxxxxxxx xx X xx x xx x x x x xx x Xx x xx x xx x xx x xxxx x xxx x x xx X x X X xx X xx x xx x
xxx xxxxxx x a
13 X
1000' 0013 00 a C313 ao13 a a a 013 p 00 a C30 013 (313a a13130coo a 00130
00 +00000013013 13 a 13 M 0 13 001310 00 CIO+ 00013+ a 13 , 013
+ ++4 0 + +4 4 1313 1313 p 130
+ ++ + ++ + ++ + ++ + ++ + + ++ + �+ �4+ ++ + *+ + ++4 + 4+ + ++ + + .. . . . .
0 34 5 a 7 ++ +
12 13 15 16 17 10 11 20 21 Z2 Z3 Z11 25 25 27 28 29 30 31

(CONTINUITY)
2050 x x x x x x x x x x x x x x x x x x x x x x
4
850: xx
51: 13a 0 a 13
a
1 ISO - 13a 13
1050: a a a aa a
850
650 + 0
910 11 12 33 14 15 16 17 18 19 20 21 22 23 24 25 25 27 28 i9 30 it

(16 VOLT SUPPLY)
20"
'n 5-6 1� t 7& 76 is att�bn @s IS nba h t t 8 a 7A 4 ISa6
-3 :+ + + + + + + + + + + + + + + + + + + + + + + + +

1 2 31 5 8 70910 11 12 13 14 15 16 17 Ja 19 20 2t 22 23 24 25 26 27 28 29 30 31

(SECONDARY BATTERY)
20-

-JIS-x x x x x x x x x x x x x x x x x x x x x x x x
CD
> P -P P43 -P 43 43 -P 43 -P 13 -P 43 -P+ -P 43 -P -P -P P -P -P P -P 13 43 -P 43P43
301 1 1 '' ' '''''I II. ''' ' '' ' ' .. I I '''
1 2 34 5 6 78910 it' 12 13 14 IS 16 37 ia 10 20 21 22 23 24 25 26 27 28 29 30 31

(PRIMARY BATTERY)
30-

-j 25-

20
1 2 34 5 6 70910 It 12 13 14 15 16 17 18 19 20 21 22 23 24 25 28 27 28 29 30 31
MARCH 1988
Figure 3.1 Telemetered Engineering Data

679

ACKNOWLEDGEMENTS (7) Briscoe, M.G. and D.E. Frye, June
The authors want to express their 1987. Motivations and Methods for Ocean
gratitude to the engineers and Data Telemetry. Marine Technology Society
technicians, at WHOI and in the Industry, Journal. Vol. 21, No. 2, pp. 42-57.
who provided expertise and support to the
STEM program. In particular we want to (8) Melligner, E.C., K.E. Prada, R.L.
thank S. Kery for detailed mooring Koehler, and K.W. Dohertyi August
analysis, P. 01 Malley and M. Gould for 1986.Instrument Bus An Electronic System
the development of the E/M cable Architecture for Oceanographic
terminations, S. Longworth for the design Instrumentation, Woods Hole Oceanogr.
and preparation of the engineering Inst. Tech. Rept. WHOI # 86-30,
instruments, E. Denton and J. Hardiman
for extensive electronic technical (9) institute of Electrical and
support and D. Payne and S. Tarbell for Electronics Engineers (IEEE). 1983. IEEE
data dissemination. Standard 796-1983, Standard Microcomputer
System Bus, IEEE, 245 East 47th Street,
We are grateful to the skippers and the New York, NY.
crew of the R/Vs ENDEAVOR, IDA-Z, and
OCEANUS for safely deploying, servicing,
and recovering the STEM mooring.
wHOI Contribution No. 6862
We want to thank B. Pratt who provided the
art work and T. Shipley for preparing the
paper.

Finally the continued interest and support
of Dr. S. Ramberg, Dr. A. Weinstein and
Dr. D. Evans of the Office of Naval
Research in the field of telemetry from
moored arrays should be recognized and is
gratefully acknowledged.

The work reported in this paper was
performed under ONR Contract Number
N00014-86-K-0751 and N00014-84-C-0134, NR
083-400

REFERENCES

(1) Berteaux, H.O. Buoy Engineering.
1976. Wiley and Sons, Inc., New York,
N.Y.

(2) Clay, P.R. and H.O. Berteaux. 1987.
The High Performance Oceanographic Mooring
(HIPOM). MTS/IEEE Oceans 187, Halifax
Conference Proceeding.

(3) Driscoll, A., Editor. 1982. Handbook
of Oceanographic Winch, Wire, and Cable
Technology. Graduate School of
Oceanography. University of Rhode Island.
Narragassett, RI.

(4) Gibson, P. 1984. Operational
Characteristics of Electromechanical
Cables. ASNE. Energy Sources Technology
Conference, New Orleans, LA.

(5) Bocconcelli A., Editor. 1987. Real
Time Environmental Arctic Monitoring
(RTEAM) Woods Hole Oceanogr. Inst. Tech.
Rept. WHOI-87-50.

(6) Institute of Electrical and
Electronics Engineers (IEEE). April 1985.
IEEE Standard Serial ASCII instrumentation
Loop (SAIL) Shipboard Data Communication,
IEEE Standard 997-1985. 12p.

680

NEW TECHNOLOGIES AND DEVELOPMENTS
IN NDBC BUOY AND MOORING DESIGN

LT Daniel R. May, USCG

National Data Buoy Center (NDBC)
Stennis Space Center, MS 39529

ABSTRACT mooring systems are less labor intensive and require less ship deployment
time due to the ease in handling.
Since the early 1970's, the National Data Buoy Center (NDBQ has
deployed deep-ocean weather buoys and mooring systems. Starting in Initial engineering concerns were concentrated on the nylon/polypropylene
1980, new deep-ocean, bottom-insensitive, inverse-catenary mooring connection and the long-term performance of polypropylene as a moor-
systems were deployed based on developments of the late 1970's. Since ing component. An integral splice between the nylon and polypropylene
1980, refinements to the basic mooring design have been made incor- segments was the desired approach, although due to the differences in
porating recent technical developments, such as use ofnoncorrosive plastic line diameter size, a special splice was considered. The difference in line
materials. In addition, new developments have been made in mooring diameters occurred when designing an overall line strength requirement
design for reduced mooring costs, in recoverable mooring systems, and for the mooring system. Since nylon line is approximately 80 percent
in prevention ofdamage to mooring linesfromfishbite. During this same stronger than polypropylene, size for size, a much larger diameter
period, buoy hulls have undergone new developments, with the trend be- polypropylene line was required to equal the strength of the nylon
ing toward smaller, lightweight, more durable hulls. Use and develop- segment.
ment of new foam buoys have recently been undertaken for adaptation
in nearshore coastal applications. During the period 1982-1984, research was conducted on the use of
polypropylene as a long-term, deep-water mooring component. Also,
numerous tests were conducted on various types of combination
nylon/polypropylene splices. All results indicated that the polypropylene
1. INTRODUCTION mooring line would perform adequately if an acceptable
nylon/polypropylene splice could be developed. Using 8-strand plaited
The National Data Buoy Center (NDBQ conducts research, development, construction ropes, numerous nylon/polypropylene size combinations
testing, and deployment of buoys and moorings in the deep-ocean en- were spliced together using standard butt splices. During this testing period
vironment. The organization was established by the U.S. Coast Guard over 30 combination rope splices were evaluated with excellent results.
in 1967 and transferred to the National Oceanic and Atmospheric Ad- As expected, a tapered style splice, where the larger polypropylene line
ministration (NOAA) in 1970. Presently, NDBC continues to operate a is slowly tapered as it is spliced into the smaller nylon line, proved to
fleet of moored buoys that provides real-time meteorological data to the be the best overall splice. See Figure 1. For the nylon-to-polypropylene
National Weather Service and other agencies. side of the splice, a standard butt splice was used. A standard butt splice
for both sides of the splice was tested and worked as well as most all splices
The NDBC mooring program has been a unique developmental process tried, but these were cumbersome to make and had an uneven lay to the
that has evolved into a highly regarded asset of NDBC. The early research, polypropylene-to-nylon splice area.
development, and testing, combined with the operational experience of
moored buoys, has led to a well-established, successful system of stan-
dard mooring designs. Refinement and enhancement of these standard
mooring systems in recent years has produced more versatile and less costly 4
mooring systems. A

A system of standard mooring designs was first established in 1978 based
on state-of-the-art technology. A minimum mooring design life goal of
A",
6 years was chosen for deep-ocean mooring systems. Preliminary testing
and development followed, with'operational deployments occurring in
1980-1981 through the present. Time has proved these designs to be suc-
cessful. NDBC currently has one mooring system over 71/2. years old and
several other mooring systems over 61/2 years old. In recent years, the
trend has been to refine these designs based on new technologies. Several
of these new enhancements are discussed in the sections below.

2. COMBINATION NYLON/POLYPROPYLENE
MOORING SYSTEMS

This mooring system, which uses long lengths of nylon and polypropylene
line, was first suggested as an alternative version of the inverse-catenary
design developed in the early 1980's[l]. The polypropylene line provides
the required buoyancy, normally provided by glass spheres, for support
of the chain and hardware suspended off the ocean bottom. Several ad
vantages are achieved by use of this type mooring system. The difference
in material cost is quite substantial over the all nylon/glass sphere flota
tion type mooring since the polypropylene line is relatively inexpensive.
A cost savings of approximately $4-5 K can be realized for a mooring
system deployed at a depth of 3,050 meters. Also, deployment of these Figure 1. NylonlPolypropylene Splice

681 United States Government work not protected by copyright

Deployed in September 1984 off the Hawaiian Islands at a depth of 4,850
meters, the first deep-water combination nylon/polypropylene mooring
system is still in use after over four years service. Additional W,
nylon/polypropylene mooring systems have been deployed at deep-water . ......
locations as shown in Table 1. No failures or problems associated with
the polypropylene line or the combination nylon/polypropylene splices
have been seen to date. Presently, use of these systems has been expand-
ed to shallow water locations and to all other deep water locations
wherever feasible. Shallow water locations (900 to 1500 meters) require
the use of larger diameter polypropylene line (5.72-
centimeter diameter).
5" S
Ile,

10
Table 1. Combination AylonlPolypropylene Mooring Systems

STATION BUOY TYPE DEPTH (M) DEPLOYED RECOVERED
51002 NOMAD 4,850 09/05/84 STILL IN USE
44004 NOMAD 3,230 12/16/84 STILL IN USE
32302 3-METER 4,200 01/19/86 STILL IN USE
41002 NOMAD 3,655 03/26/86 06/03/86
46006 12-METER 3,930 06/28/86 STILL IN USE
46037 3-METER 3,500 08/08/87 09/22/87
46002 NOMAD 3;425 04/29/87 STILL IN USE

3. USE OF PLASTICS IN MOORING SYSTEMS Figure 2. Standard NOMAD Mooring Insulator

One area that NDBC has examined in recent years in regard to im-
provements in mooring systems involves the use of nonmetallic com- internal bolt system and placed through the yoke connection. An over-
ponents. New innovations in the use of high-strength, low-wear plastics sized steel split-key shackle connects by a hole bored through the bolt.
have prompted evaluation of these materials for use as components in See Figure 3. Tests over the past 3 years have indicated this design works
seawater mooring applications. well, although some wear and brittleness is seen in the acetal bushing after
1 V2 to 2 years. Two buoys are currently fitted with this design, but with
One component currently being evaluated for use in long-term, deep-water a new shortened bushing, for further field testing.
mooring systems is the plastic rope thimble. Currently, NDBC uses cast
aluminum-bronze rope thimbles that have a special urethane coating. The Another application utilizes a press-fit type bushing coupled with a stan-
coating helps protect the synthetic mooring line as well as isolate the thim- dard Coast Guard split-key shackle (Figure 4). Several plastic materials
ble from its connection to a carbon steel connecting shackle. The thimbles were examined, including acetal, acetron (a specially engineered acetal)
are heavy, cumbersome, and somewhat expensive after application of the and thordon (a shipboard rudder and shaft bearing material). Thus far,
special urethane coating. The plastic thimbles, manufactured by Sam- the thordon material appears to show the best performance based on ex-
son, use a hard plastic internal spool and a protective urethane cover perimental data. Two NOMAD buoys are currently outfitted with this
shield. They are lightweight, easy to use, and are comparably priced. The type design, one with an acetal bushing and one with a thordon bushing,
spool, which must support the mooring loads by providing a bearing sur- and will be recovered in the late fall of 1988. Based on the results of the
face for the mooring line and connecting steel shackle, is composed of
a specially engineered nylon of high strength with dispersed lubricants
for excellent wear properties. NDBC has used many of these plastic
thimbles in developmental mooring systems over the past 3 to 4 years
with excellent results. Upon recovery and inspection, it was found that 51@
all thimbles had performed exceptionally well without any degradation
to the plastic shield or spool components. In November 1985, two of these
thimbles were deployed in an operational mooring system with a 3-meter
discus buoy in the Hawaiian Islands. The mooring system, located in the
Alenuihaha channel between the islands of Hawaii and Maui at a depth
of 1,010 meters, was recovered in July 1988 after 3 years of service in
a high-current, rough-channel environment. This mooring system is ser-
ving as a long-term evaluation of the plastic thimble for exclusive use
in all NDBC deep-water mooring systems.

In another application, several new plastics are under evaluation as isola-
tion and bushing materials used in the connection of NDBC's NOMAD
buoy to its mooring system. The NOMAD buoy has a 316 stainless steel
mooring yoke that provides good strength and corrosion performance
for long-term seawater applications. Presently, a special stainless steel
shackle combined with a 2-meter spliced loop of 11/4 -inch-diameter nylon
line is used to connect the buoy to its carbon steel mooring chain. Use
of the looped nylon line, called an insulator (Figure 2), isolates the
dissimilar metals from contact in seawater. This design has been in use
for over 10 years with excellent results, but could be improved and made
less costly by using a different isolation/connection technique at the buoy A*
yoke.

Several new isolation/connection techniques are being evaluated using
plastic isolation bushings coupled with a standard carbon steel connec-
ting shackle. In one application, an acetal bushing is threaded onto an Figure 3. Special Purpose Split-Key Shackle Connection

682

Table 2. NDBC Fishbite Mooring Failures

SITE DAYS ON TYPE FAILURE DEPTH
LOCATION DEPTH (FT) STATION MOORING LINE (FT)
25-54.3 N 11,1340 71 1.75'DIA NYLON 545
WIP" 89-42.4 IN
1980
25-00.0 N 10,600 1.118 1.25' DIA NYLON 3.350
88-00.0 W
3 FAILURES -
.4- 26-00.0 N 91996 724 1.625* DIA NYLON 1,350
93
-30.0 W
VM-l, 1981 NONE
34-64.2 N
MIR M 1 FA191LSURE 72-53.5 W 13,880 377 1.75* DIA NYLON 820
T', 1983 29-18 0 N 293 1.125* CIA NYLON 350
77-18.1 W
23-24.0 N
lp 2 FAILURES 10,680 886 1.75' CIA NYLON 2,075
62-18.0 W
1984 NONE
29-18 8 N
'1.1111,1:@,;141'11 @q
77-19.8 W u,z,u_ am 1.125'DIA NYLON 220
32-18.0 N
lq
@@, , I 1 12,396 1.75' DIA NYLON 1,500
il$ 114,01,lili 1985 75-17.4 W
26-54.9 N 10,200 9D9 1.75' DLA NYLON 950
89 42.7 W
32-16.0 N
1986 75.1 4.0 W 12,000 46 1.125" DIA NYLON 1,400
2 FAILURES 29-19.0 N 3,420 377 1,125' DIA NYLON i,800
77-21.OW
1987 25.58.3 N 10,380 2,368 1.75' CIA NYLON 850
1 FAILURE 85-55.9 W

Figure 4. Third Class Split-KeylBushing Combination AVERAGE DEPTH OF FAILURE - 1,350 FEET BELOW SURFACE
AVERAGE TIME ON STATION - 769 DAYS

ongoing field tests, a final plastic configuration NOMAD yoke connec-
tion is anticipated for early next year. evaluations, including tensile break-strength tests, will be performed to
provide an indication of how much strength was lost due to fishbite.
4. FISHBITE-RESISTANT MOORING The goal of this new study by ND13C is to develop a fishbite-resistant,
CABLE DEVELOPMENTS long-term, mooring cable as well as further the scientific knowledge of
moorings and the fishbite phenomenon. Based on the results of the study,
Fishbite of mooring cables is the phenomenon by which mooring lines should a particular cable be identified as successful, it would be strongly
are damaged or severed by the bite of various ocean species, most notably considered for use in NDBC's current mooring design. A final report on
sharks [2]. NDBC first encountered the fishbite problem during the early the results of the study is expected to be completed in January 1989.
1970's during the deployment of its weather data buoys in the Gulf of
Mexico. After early studies indicated no precise explanation for the
phenomenon or any solution for a deep-water, long-term, synthetic moor- 5. RETRIEVABLE SYNTHETIC MOORING SYSTEMS
ing system, NDBC continued with its development for the best overall
mooring system, omitting any specific fishbite protection. This approach Deep-water synthetic mooring systems deployed by NDBC are typically
worked quite well until the early 1980's, when the buoy network was great- not recovered; however, where similar type moorings are used in in-
ly expanded, especially with the use of smaller buoy hulls and smaller termediate water depths (150 to 610 meters) or the Great Lakes, a means
diameter mooring lines. During this period and in the following years, of retrieval can be provided on the synthetic line. Since NDBC buoy opera-
mooring failures due to fishbite began to occur more frequently (Table 2). tions are supported by U.S. Coast Guard buoy tenders (Figure 5) the
means of synthetic line retrieval is designed for use by the lifting boom
Due to this increase, and since new technical advancements had been made located on each ship. Normal operation of the boom is in the lifting and
concerning mooring cables, NDBC began a renewed study of the fishbite pulling of buoy chain. Thus, a series of retrieval loops were designed for
problem. The first portion of the study involved laboratory screening of use on the mooring line so that it could be recovered in the same man-
numerous developmental cables in an attempt to test their "fishbite ner. Due to the location of the boom and available deck space, the max-
resistance." Most of these cables were new parallel fiber construction imum distance retrieval loops can be spaced apart is 6 to 8 meters. Thus,
cables with specially designed jackets or construction techniques to pre- for long lengths of line to be retrieved, numerous loops are required.
vent damage resulting from cutting. The screening tests were conducted Early loop designs were developed when long Tz (temperature vs. depth)
by the Ocean Engineering Department at Woods Hole Oceanographic lines were used as integral members of a buoy mooring system in the ear-
Institution and included investigation of the cutting force required to sever ly 1980's. Cables up to several hundred feet long were installed in the
the cable (both in the slack condition and with the cable under tension),
stabbing force, and cut resistance. near-surface portion of the mooring and required retrieval loops for
routine servicing and removal. The loops consisted of segments of 8-strand
From the results of the initial study, four different type cables showed nylon line braided back onto the Tz cable in an increasing helic: angle
an ability to resist cutting[31, This led to the second phase of the study, or "Chinese finger" fashion. When placed under load, the strands of
which involves a field test of the four cables in an attempt to gain actual the line became taut against each other holding the loop in place. These
fishbite experience. The field test portion of the study is being conducted same type loops were used in a Gulf of Mexico mooring on a single
by Old Dominion University in conjunction with an ongoing study of 305-meter segment of 8-strand nylon line. The loops held just as well on
FAD (Fish Aggregation Device) buoys off the coast of Puerto Rico. The the nylon line as they had on the jacketed Tz cables. The only drawbacks
cables are being deployed during the summer of 1988 on lightweight buoys to these particular type loop designs are the manhours involved in
off the coast of the Island of Culebra. Fish attractors will be placed on installation of the loops and the possibility of some slight slippage dur-
each mooring system to establish an FAD. Previous use of these type FAD ing use. Complete fabrication and installation requires approximately one
systems in these waters showed a significant amount of mooring line
fishbite. The test cables will be physically inspected and their underwater manhour per loop which, when long segments are involved, can be costly
condition videotaped periodically for evidence of fishbite. Cables that and time-consuming.
exhibit a high incidence of fishbite or any unique type of fishbite damage In 1986 efforts began on formulating a standard retrieval loop design
will be removed from the existing mooring for further analysis and testing. based on modification of the previous design. Emphasis was primarily
Additional lengths of cables will be used to replace removed segments on the installation method since this was seen as a deficiency of the early
to continue the evaluation period. At the end of the summer or early fall, design. Three separate retrieval loop designs, all using 8-strand nylon line,
all buoys and mooring systems will be recovered and a final inspection were developed, fabricated, and tested for evaluation. The three designs
and analysis will be made of the four cables' condition. Additional are shown in Figure 6. A variation of the Type A design was also tested
where the loop was spliced back up onto the mooring line. All designs

683

method used the "Chinese finger" type method of attachment using the
four legs of the loop in a braid down the mooring line. The second method
used a loop with much shorter length legs and involved splicing the loop
into the mooring line by following the four pairs of strands with each
leg of the loop. The latter method was the least difficult method and re-
quired less than 10 minutes to install.

Excellent results were achieved on all loops during repeated break tests
a
of both loops. Slight slippage was seen on the "Chinese finger" loops,
p
UIMA;1111@' F' @,Mcj
02" although they held to ultimate break strengths of approximately 9,980
kilograms. (Individual leg strength was rated at 6,530 kilograms.) The
spliced-in loops worked extremely well and, with a single leg strength of
3,900 kilograms, provided a consistent break strength of approximately
7,250 kilograms.

Continued development of synthetic mooring line retrieval loops is
planned with more testing of the spliced-in type Kevlar loops. An increase
in single line size to achieve a final break strength of approximately 9,000
kilograms is the goal. From all development and testing thus far, this
appears to be the best overall design, as well as the quickest and easiest
to install.

6. BUOY HULL DEVELOPMENTS

Development of new, low-cost, smaller, moored buoy hulls has been a
major goal of NDBC during the 1980's. Maintenance and refurbishment
costs of the large 10- and 12-meter discus hulls have reached prohibitive
levels, especially in austere budget years. The excellent seakeeping
NOMAD buoys are ideal for offshore (100-250 nm) and severe-
Figure 5. USCG Buoy Tender environment locations, such as the Gulf of Alaska, but, at $80 to $100 K
per buoy hull, are too costly for nearshore or coastal applications.
worked extremely well, providing over 9,000 kilograms wet break strength In the early 1980's NDBC began investigating an aluminum 3-meter discus
and only requiring about one-half the installation time. No slippage or hull with a design similar to that developed by Woods Hole Oceanographic
excessive elongation was seen during the numerous tests performed. Of Institution. NDBC procured four of these buoy hulls of varying configura-
the two type A designs, the first design appeared to be the better. In the tions and in 1983-1984 began deployments in coastal areas and in the Great
alternate version, when the loop is under load, the top portion of the Lakes. After a successful deployment of one year (1984-85) off the en-
loop fibers cut against each other since the splice is made back up onto trance of the Columbia River, the 3-meter buoy hull was declared opera-
the mooring line. tional, and an additional 15 hulls were procured for replacement of 10-
and 12-meter hulls and for expansion of the buoy network in the coastal
During 1987 another retrieval loop design was developed and tested us- area. The standard configuration 3-meter buoy hull is shown in Figure
ing a special Kevlar fiber as the material for the retrieval loop. The pur- 8. The hull of the buoy is constructed of 5086 marine grade aluminum
pose in developing this new loop design was first to seek additional im- with the tripod mast constructed of 6061 grade aluminum. A rigid, three-
provements over current designs, and second to gain additional knowledge leg, steel bridle is attached to the hull for stability and for the mooring
with the interworking of different synthetic fibers such as nylon and connection. Additional data on the 3-meter buoy are shown in Figure 9.
Kevlar. To construct the loop, two segments of equal length each were
folded over into a loop. A short piece of shrink-tube was used over the Continued success of the 3-meter hull as a moored meteorological plat-
line for chafe protection and to give the loop its rigid shape. A completed form has been achieved with deployments in the South Pacific, Pacific
loop is shown in Figure 7. The loops were attached to 8-strand nylon Northwest, in the straits of the Islands of Hawaii, and increasingly in
mooring line in two different fashions for testing. The first

LOOP DESIGN A LOOP DESIGN 8 LOOP DESIGN C

TAPE TAPE TAPE
END 1-118" NYLON END END

1-118" NYLON 1-3/4" DIAMETER 8-STRAND NYLON

MAKE LOOP 8-101A DIAMETER
SPLICE BACK (EACH SIDE) OFF MAIN LINE 1-1/8" DIAMETER 8-STRAND NYLON
2 DOUBLE TUCKS
3 SINGLE TUCKS

MAKE LOOP 8-10" DIAMETER
OFF MAIN LINE

1-1/8' NYLON

6-7' 6.7

EYE SPLICE TECHNIQUE
6-15" DIAMETER EYE 6-8" DIAMETER EY@E Figure 6. Retrieval Loop Designs
684

8-10'

PLASTIC
SH INK TUBE
OVER FLATBRAID

WHIP EACH SIDE OF LOOP
AND ENTIRE BUNDLE

LIGHTLY COAT WITH FLEXANE
AFTER WHIPPING

3-METER DISCUS

1. (3D01)
6 a. HULL WEIGHT 1,887 LB
6-) b. POWER SYSTEM 610 LB
c. PAYLOAD 100 LB
d. BALLAST 500 LB
e. DEPLOYED 3,300 LB
f. VCG 2.5 FT

2 SEGMENTS (12 LONG EACH) JETSTRAN 2. (3D02 - 3D19)
207 KEVLAR FLAT BRAID
(FOLD OVER TO FORM LOOP) a. HULL WEIGHT 2,600 LB
b. POWER SYSTEM 660 LB
c. PAYLOAD 210 LB
d. BALLAST 0 LB
e. DEPLOYED 3,470 LB
f. VCG 2.4 FT

.t___1APE OR WHIP EACH END

Figure 7 Kevlar Retrieval Loop Figure 9. 3-Meter Buoy Hull Data

New offshore deployments of 3-meter buoy hulls are planned in 1989 and
1990 in the Gulf of Mexico and the North Atlantic Ocean, as well as new
coastal locations. One such coastal application will be approximately 32
kilometers off Cape Canaveral, Florida, for timely weather observations
in support of NASA's Space Shuttle program. An additional quantity
of 3-meter buoy hulls is being procured during 1989 in support of these
and other programs. The cost of the new hulls is approximately $20 K
per buoy.

Also, NDBC has recently developed the use of foam in the construction
of small buoy hull designs. One such buoy was constructed this past winter
for use as a marker buoy off the coast of California. The buoy will sup-
port existing NDBC mooring systems while operational buoy hulls are
refurbished, rebatteried, or outfitted with new calibrated sensors. The
foam material used in the construction of the buoy hull is a special surlyn
material. It is lightweight (48 kgs/m3 density), extremely durable, per-
manantly colored throughout, nonwater absorbant, and requires no
additional protective outer skin. Figure 10 shows the marker buoy con-
figuration prior to shipment to the west coast. It is designed such that
the mast, hull, and counterweight tube can be disassembled for easy
hen required for operational use, the buoy can be assembled
storage. W
in less than 30 minutes and is capable of supporting a 2,268-kilogram
mooring load. A dry-cell-powered Obstruction and Identification Light
provides approximately 6 months on-station time if required.
Figure 8. 3-Meter Discus Buoy Small foam buoy hulls are also being considered for nearshore coastal
applications. This potential application includes wave measurements as
meteorological measurements. This lightweight (less than 340 kilograms),
coastal applications that include offshore Cape Hatteras during the winter low-cost buoy hull could be adapted for use in almost any nearshore loca-
of 1987-1988. The 3-meter deployment in the Pacific Northwest tion as well as in bays and inlet areas. Similar to the 3-meter hull, such
(48* 17'11 " N and 133'45'43 " W) was for a I -year duration and resulted buoys could be carried onboard a vessel or towed to its station.
in the highest significant wave height (over 14 meters) recorded by a
3-meter buoy. A new directional wave system, adapted specifically for
the 3-meter hull, has gained high acceptance from other government agen- 7. CONCLUSIONS
cies and weather forecasters in areas subject to problems associated with
high waves and storm surge. Additionally, a new upper mooring system The National Data Buoy Center is committed to the further development
has been designed, along with a towing pendant, for easier deployment of buoy, mooring, and meteorological systems for timely, accurate, and
and retrieval. In most coastal areas (less than 65 kilometers offshore) the low-cost weather observations. The recent developments in synthetics,
3-meter buoy can be towed to station and deployed by a small vessel. plastics, foams, and new buoy and mooring designs during the 1980's
Retrieval and exchange of buoys can also be performed on station without will provide the basis for further improvements and developments in the
hoisting the buoy from the water through the use of a synthetic mooring 1990's.
retrieval pendant. An experimental exercise in the Gulf of Mexico this
past winter showed the buoy can be towed in seas of less than I meter
by small Coast Guard vessels, such as the 17-meter Aids to Navigation
Boat.

685

J
', @ 77

A'N

Figure 10. Foam Marker Buoy

&REFERENCES

I .NOAA Data Buoy Office, NDBO Buoy Mooring Workshop Report,
Department of Commerce, NOAA, Bay St. Louis, MS, September
1978.
2. Prindle, Bryce and Robert G. Walden, Deep Sea Lines Fishbite Manual,
Department of Commerce, NOAA, Bay St. Louis, MS 1975.

3. Prindle, Bryce, Preliminary Assessment of Fishbite Resistance of Lines
Designed for Use in Deep Sea Mooring, Woods Hole Report No.
400-QANW-6-00458:S-1, Sponsored by National Data Buoy Center,
Department of Commerce, NOAA, Bay St. Louis, MS, May 1987.

686

INTERNATIONAL ICE PATROL APPLIED OCEANOGRAPHY

Stephen R. Osmer and Donald L. Murphy

Commander, international Ice Patrol, U. S. Coast Guard
Avery Point, Groton, CT 06340-6096

ABSTRACT The prevalence of fog, the accumulation Of
icebergs, the severe storm conditions so common in
The United States Coast Guard utilizes this region, the concentration of trans-Atlantic
oceanographic research in accomplishing its shipping, and the presence of oil platforms and
assigned International Ice Patrol mission. The fishing vessels scattered over the Grand Banks
international Ice Patrol is using side-looking make this one Of the potentially most dangerous
airborne radar for improved iceberg detection and areas in the world for marine transportation.
to observe ocean surface features (i.e. fronts,
eddies); is applying current velocity and sea The international Ice Patrol (iiP) operations
surface temperature data from satellite-tracked center, located in Groton, Connecticut, maintains
oceanographic drifters; has instituted using a plot showing the location of all icebergs
aircraft-launched expendable bathythermograpb detected in its operations area (40*N-52*N and
(AXBT) probes; and is using a satellite infrared 39*W-57*W). The primary data source is IIP's own
imagery interpretation system. The International aerial reconnaissance detachment which patrols the
Ice Patrol is striving to improve the quality and north Atlantic for seven consecutive days on
quantity of environmental data collected on the alternate weeks during the six-month ice season
Grand Banks of Newfoundland. The data is used by (March-August). Each flight covers only a small
U. S. Navy and Canadian environmental forecast portion of the region. These iceberg position
centers to provide input data for the Ice Patrol's data are augmented by the Canadian Atmospheric
iceberg drift and deterioration prediction models. Environment Service (AES) patrols, sightings by
the offshore industry, and reports from vessels
transiting the area. An average of 300-400
icebergs are tracked by the IIP each year; during
severe ice years 1000-2000 iceberg might be
1. INTRODUCTION tracked.

The sinking of the luxury passenger liner RKS The iceberg sighting information is entered into a
TITANIC in 1912 prompted maritime communities with computer at the IIP Operations Center along with
ships transiting the Grand Banks of Newfoundland, ocean current, wind, wave, and sea surf ace
Canada, to establish an iceberg patcol in the temperature data. iIP utilizes two models - an
area. Since 1914, the United States Coast Guard iceberg drift and prediction model and an iceberg
has been tasked with the management and operation deterioration model.
of this patrol, known as the International Ice
Patrol. The patrol observes and studies ice Every twelve hours, the predicted iceberg
conditions near the Grand Banks and warns mariners locations are used to estimate the limit of all
of any iceberg threats. The International Ice known ice. This limit, along with a few of the
Patrol is funded by the twenty signatory nations more critical predicted iceberg locations, is
to the Safety Of Life At Sea (SOLAS) Convention, broadcast as an "Ice Bulletin" from radio stations
who reimburse the United States government for in the United States, Canada, and Europe for the
this service. it has proven to be an outstanding benefit of all vessels transiting the North
example of effective international collaboration Atlantic. in addition to this bulletin, a radio
for the preservation of life and property at sea. facsimile chart of the area, visually depicting
the locations of ice, is broadcast once each day.
icebergs, mainly from glaciers in west Greenland,
are carried southward by the cold Labrador Current 2. ICEBERG RECONNAISSANCE
to the vicinity of the Grand Banks and into the
shipping lanes between Europe and the major ports Since 1983, IIP has utilized Coast Guard HC-130
of the northeast United States and Canada. This aircraft equipped with a real aperture, X-band
is also the area where the Labrador Current meets side-looking airborne radar (SLAR) <an AN/APS-135)
the relatively warm Gulf Stream; temperature as its primary reconnaissance tool (1,2,3). The
differences between the two water masses of up to SLAR has an all weather capability which helps to
20*C produce dense fog some 40-50% of the year. negate the prevalent fog and adverse, weather
conditions in the Ice Patrol operations area.

687 United States Government work not protected by copyright

In June of this year, IIP conducted a ten day SLAR The drifters are air-deployed from Coast Guard
experiment of f Newfoundland to evaluate the HC-130 aircraft on routine ice reconnaissance
AN/APS-131 SLAR. This SLAIR' is installed as part flights. They are equipped with a window-shade
of the AIREYE system aboard Coast Guard HU-25 drogue tethered 50 meters below the surface, a
aircraft (Falcon jet). drogue-tension sensor, and a sea surface
temperature (SST) sensor. Service ARGOS processes
The fundamental goal of this research was to all sensor and position data and relays them to
provide Commander, International Ice Patrol (CIIP) IIP via computer link. These data are also shared
with guidance on the ability of the with all major environmental data collection
AIREYE-equipped HU-25 to perform the iceberg centers worldwide through the Global
detection mission of the IIP. Specifically, there Telecommunications System (GTS).
were two objectives:
The wind input for the air-drag and wind-current
1. Provide CIIP with a basis for determining computations is calculated from wind data provided
the best altitude for iceberg searches .- and every twelve hours by the Fleet Numerical
predicting the probability of iceberg detection as Oceanography Center (FLENUMOCEANCEV).
a function of sea state, range, and iceberg size.
Deterioration Predictions
2. Compare the iceberg detection capability
of the AN/APS-131 SLAR in the AIREYE package with A parametric model is used to estimate the
the AN/APS-135 SLAR currently used on the IIP deterioration of an iceberg after its initial
HC-130 reconnaissance aircraft. sighting. This model is used to predict the
removal of a "melted" iceberg from plot.
The Atmospheric Environment Service (AES) of
Canada also had two of its SLAR-equipped ice Deterioration is calculated as a function of
reconnaissance aircraft@ participate. Their SLAR's monthly averaged solar insolation, and daily
were an AN/APS-96 and a CAL-100. Their interest estimates of water temperature, wave height, and
was to evaluate performance. This cooperative wave period (approximately 85% of iceberg
effort is the latest in a long series between IIP deterioration is due to wave effects).
and AES (4). Deterioration from calving and aerial melting are.
ignored (12, 13).
The U. S. Coast Guard icebreaker NORTHWIND acted
as the surface truth platform, recording iceberg Water temperature and wave data are provided by
target information and the environmental FLENUMOCEANCEN.
conditions.
Forecast Limitations
3. OPERATIONAL FORECAST MODELS
The major shortcomings in operational forecasting
Drift Predictions of the iceberg danger are: (1) the accuracy of
the iceberg sighting (position, size, and shape);
once an iceberg has been reported to the IIP (2) computational constraints which require IIP to
operations center, its location, size, and shape place the icebergs into broad categories of size
are entered into a computer drift model that is and shape; (3) the resolution of the current,
used to estimate the subsequent motion of @the temperature, and wave fields; and (4) the quality
iceberg. The use of the model is necessary for of the environmental data.
two reasons: first, so that future sightings of
an iceberg are recognized as resights rather than Using the drifters improves on the mean currents,
as new sightings, and, second, so that the and provides additional sea surface temperature
position of an iceberg can be estimated if no data, but ten to twenty drifters per year cannot
resighting is made. adequately cover the entire operations area for
the six-month ice season.
This dynamic model combines the effects of water
drag, air drag, the Coriolis acceleration, and sea The latter two shortcomings are classic concerns
surface slope to . determine the iceberg of regional forecasts. In the IIP operations
acceleration. Wave forces are ignored (5,6). area, data observations are sparse, seriously
degrading forecasts and requiring the use of
Model input data are derived from several grid-spacings which smooth much of the data useful
sources. The mean geostrophic current field is to IIP.
based on hydrographic surveys conducted by the
Coast Guard from 1936 to 1974. Because this,mean 4. RESEARCH EFFORTS
current field fails to incorporate temporal
variability of the flow field, IIP began In an ef fort to improve the input data, IIP has
air-deploying satellite-tracked TIROS been discussing with FLENUMOCEANCE9, Naval Eastern
Oceanographic Drifters (TOD's) in 1983 and using Oceanography Center (HAVEASTOCEANCEN), *and the
their drifts to modify the historical field Canadian Maritime Command/Meteorological
(7,8,9,10,11). Between ten and twenty buoys are Oceanography Center (MARCOM/METOC) ways to improve
deployed each season. the quality of analysis and quantity of data
collected in the area.

688

During the 1988 ice season, IIP provided drift In 1988, lip also provided real-time SLAR analysis
trajectory data to WAVEASTOCEANCEff for ocean of ocean surface features (i.e. eddies, fronts) to
feature analysis during 1988. WAVEASTOCEANCEN and MARCOM/METOC.

lip is equipping future drifters with barometric Also in 1987, lip conducted a second research
pressure sensors. This data, along with the SST cruise in an effort to expand its limited data on
data, will be available to FLENUMOCEANCE9 and iceberg deterioration from USCGC TAMAROA. six
MARCOM/METOC. icebergs were studied, and their motion compared
to deployed TOD's.
These drifters could also be equipped with other
sensors such as thermistor chains if sponsors lip has put together atlases of the TOD
could be found. The drifters have a life span of trajectories since 1983 and SLAR feature analysis
6 months and become entrained in the North since 1985.
Atlantic Current around 45*W. In the past lip has
ensured the processing and dissemination of 5. SUMMARY
drifter data as drifters crossed the North
Atlantic although their importance of IIP ended at The international Ice Patrol is applying
390W. technology and oceanography to conduct its mission,
in the most efficient manner. Cooperative efforts
The past two years, lip has been involved in a with U. S. and Canadian agencies have given
cooperative project with the Naval Oceanographic excellent results in the previous years and hold
Research and Development Activity (NORDA) to great promise for the future.
evaluate the potential of mini-drifters (sonabuoy
size) (14, 15). in February 1988, Coast Guard - 6. REFERENCES
aircraft deployed eight of these mini-drifters in
the Gulf of Mexico. Based upon the initial (1) Thayer, N. B., 1985. EFFECTS OF SIDE
promising results, two more were deployed by lip LOOKING AIRBORNE RADAR (SLAR) ON ICEBERG
aircraft in mid-March - one in the Labrador DETECTION DURING THE 1983 AND 1984
Current and one in the North Atlantic Current. INTERNATIONAL ICE PATROL SEASONS. Appendix
Two more were deployed in the IIP area in early C to Report Of The International Ice Patrol
April. in the North Atlantic 1984 Season, Bulletin
No. 70.
The past year, IIP has been involved in an
evaluation of installing an aircraft launched (2) Thayer, V. B. and N. C. Edwards, 1987.
expendable bathythermograph (AXBT) system on its ICEBERG/SHIP TARGET DISCRIMINATION WITH
aircraft (16). The evaluations were successful. SIDE-LOOKING AIRBORNE RADAR. Appendix B to
A system was purchased and used during the 1988 Report Of The International Ice Patrol In
ice season. The data, collected on lip flights, The Worth Atlantic 1985 Season, Bulletin
was provided to all three oceanography centers. No. 71.
During the AXBT system evaluations, data were
provided to and used by the Harvard Gulfcast (3) Edwards, N. C. and W. B. Thayer, 1986.
model. AXBT*s should aid in future SLAR SIDE-LOOKIVG AIRBORNE RADAR DETECTION AND
interpretation of ocean surface features. IDENTIFICATION OF ICEBERGS. Proceedings
Of The Canadian East Coast Workshop On Sea
Recent Ice Patrol research has focused on the use Ice January 7-9, 1986.
of remotely sensed data to improve model inputs
(i.e. inferred currents and SST). Satellite (4) Osmer, S. R.. and H. McRuer, 1987. 1987
infrared imagery, used with great success in many PRESEASON ICEBERG SURVEY AND SEASON
parts of the world's oceans to map oceanic fronts PREDICTION. Proceedings OCEANS '87, Vol 1.
and infer circulation, is limited to a few images
per months by the persistent fog and clouds in the (5) Mountain, D. G., 1980. ON PREDICTING
IIP region. However, imaging radars (i.e. SLAR) ICEBERG DRIFT. Cold Regions Science And
can map ocean frontal features through clouds and Technology, Vol 1 (3/4) pp 273-282.
fog. The Ice Patrol has been evaluating the
detection of ocean surface features by SLAR since (6) Murphy, D. L. and I. Anderson, 1987. AN
1985 (17). EVALUATION OF THE INTERNATIONAL ICE PATROL
DRIFT MODEL. Appendix D to Report Of The
In 1986, Ice Patrol used its SLAR to map the International Ice Patrol In The North
boundaries of a warm core eddy located between the Atlantic 1985 Season, Bulletin No. 71.
Labrador and the Worth Atlantic Currents (18).
Concurrent hydrographic survey conducted by USCGC (7) Summy, A. D., 1984. OCEANOGRAPHIC
EVERGREEN and satellite imagery showed that the CONDITIONS ON THE GRAND BANKS DURING THE
SLAR would reliably map the eddy location, 1982 INTERNATIONAL ICE PATROL SEASON.
although interpreting radar imagery of the sea Appendix B to Report Of The International
surface still requires much research. in 1987 Ice Patrol In The North Atlantic 1982
another warm core eddy was studied by SLAR, Season, Bulletin No. 68.
satellite imagery, and surface hydrography
collected from USCGC BITTERSWEET.

689

(8) Anderson, 1., 1984. OCEANOGRAPHIC (13) Anderson, 1., 1984. ICEBERG DETERIORATION
CONDITIONS On THE GRAND BA14KS DURING THE MODEL. Appendix C to Report Of The
1983 INTERNATIONAL ICE PATROL SEASON. International lee Patrol In The North
Appendix B to Report Of The international Atlantic 1983 Season, Bulletin No. 69.
lee Patrol in The Worth Atlantic 1983
Season, Bulletin No. 69. (14) Anderson, 1., 1987. MINI-DRIFTER BUOY
FUNCTIONAL TEST. International lee Patrol
(9) Anderson, 1., 1985. OCEANOGRAPHIC Technical Report 87-2.
CONDITIONS ON THE GRAND BANKS DURING
THE 1984 INTERNATIONAL ICE PATROL SEASON. (15) Thayer, N. B., D. L. Murphy, and W. A.
Appendix B to Report of The International Henry, 1988. TEST AND EVALUATION OF THE
lee Patrol In The North Atlantic 1984 COMPACT METEOROLOGICAL AND OCEANOGRAHIC
Season, Bulletin No. 70. DRIFTER (CMOD). international lee Patrol
Technical Report 88-02.
(10) Anderson, 1., 1987. OCEANOGRAPHIC
CONDITIONS ON THE GRAND BANKS DURING (16) Alfultis, M. A., 1988. INTERNATIONAL ICE
THE 1985 INTERNATIONAL ICE PATROL SEASON. PATROL AXBT EVALUATION. International Ice
Appendix C to Report Of The International Patrol Technical Report 88-05.
Ice Patrol In The North Atlantic 1985
Season, Bulletin No. 71. (17) Thayer, N. B. and D. L. Murphy, 1987.
DETECTION OF OCEAN FRONTS IN THE GULF
(11) Anderson, 1., 1988. TOD OCEANOGRAPHIC STREAM/LABRADOR CURRENT SYSTEM BY SIDE-
DRIFTER TRACKS ON THE GRAND BANKS DURING LOOKING AIRBORNE RADAR. Appendix F to
THE 1986 INTERNATIONAL ICE PATROL SEASON. Report Of The International lee Patrol In
Appendix B to Report Of The International The North Atlantic 1985 Season, Bulletin
Ice Patrol In The North Atlantic 1986 No. 71.
Season, Bulletin No. 72.
(18) Murphy, D. L., and I. Anderson, and N. B.
(12) White, F. M., M. L. Spaulding, and Thayer, 1988. OBSERVATIONS OF AN OCEANIC
L. Gominho, 1980. THEORETICAL ESTIMATES FRONT SOUTH OF FLEMISH PASS. Appendix C to
OF THE VARIOUS MECHANISMS INVOLVED IN Report Of The International lee Patrol In
ICEBERG DETERIORATION IN THE OPEN OCEAN The North Atlantic 1986 Season, Bulletin
ENVIRONMENT. U. S. Coast Guard Research no. 72.
And Development Center Report CG-D-62-80.

690

OCEANOGRAPHY ON EAGLE AUSTRALIA '88 CRUISE

Ross L. Tuxborn, Lieutenant Commander, USCG

International Ice Patrol
Avery Point
Groton, Connecticut 06340
,z
ff-Z
14, r
ABSIRACT rt @C- Al FV
0
HIM 1 3@ @@, , -
f "A @ 1 -1 - A
k9o, i@li'o' F
On 26 January l9ffl Australia celebrated its blCentelnlal W*4
r
with a parade of tall ships in Sydney Harbor. Representing
the United States was the Coast Gjard's EAGLEa three masted
barque used for training the service's future officers.
ht
EAG[ F's voyage to Australia and back took eig
months.
Av
During that period, Umning was conducted for the 290 cadets
who were embarked for the two phases of the cruise. A.course
in meteorology and physical oc 501@,*
4anography was part of the
onboard curriculum.
Additionally F191 I
FAIR served as a platform of opportunity
for a number of projects. NOAA installed SEAS cnboard to J,
facilitate weather and xBr observations. -r Ylir C,
odle projects
j
included marine mamnal observations, plankton taws, and sea
surface pH measurements.
The environmental data was supplied to, forecasters and
researchers worldwide who are trying to comprehend the
interactions between the ocean and abmso-@ere. The hends-on
experience the cadets received in all aspects of observatim Alw k',
3@ I
Q
and data collection was an attarxement to the daily classroom
work and was invaluable in guiding the cadets taards an
understanding of the marine environaent.

1. INTRODUCTION

The Tall Ship EAGLE has long been associated
with the nautical education of Coast Guard
officers. This paper describes a recent expansion
of that role; to not only be a testing ground for
nautical skills, but, also to serve as a floating
classroom for academic courses and provide a Figure 1. USCGC EAGLE (WIX 327).
platform for oceanographic research.
EAGLE is homeported at the Coast Guard a war reparation from Germany. Originally she was
Academy, located on the Thames River in New London, the German naval training ship HORST WESSEL; built
Connecticut. The Academy is a four year college
with an enrollment of approximately 800 cadets. in 1936 by the Blobm and Voss Shipyard, Hamburg,
Young men and women who successfully complete the Germany (Regan and Johnson, 1986).
program graduate as commissioned officers with EAGLE's hull, masts, and deck structures are
constructed of steel. The term "barque" refers to
Bachelor of Science Degrees. About half of the the sail .rigging. Figure 2 shows the sail
Coast Guard's officers come from the Academy. configuration.
Training on EAGLE is an integral part of a In addition to the mission of cadet training,
cadet's professional education. A cruise on the the EAGLE has often participated in celebrations
sailing ship gives cadets the chance to apply the marking important events. EAGLE was the lead ship
professional skills they learn. Normally these for the tall ships parades into New York harbor for
cruises take place during the summer and are less the US Bicentennial in.1976 and for the Statue of
than ten weeks in duration.
New London has been the home for the EAGLE Liberty Centennial in 1986. When the invitation
since 1945, when the vessel was commissioned into for EAGLE to take part in Australia's bicentennial
the Coast Guard following World War II. EAGLE was celebration held in 1988 was accepted by the Reagan

691 United States Government work not protected by copyright

main 0 Er e 7royal
Yal\
Le FK%
opg;
main Ilant
topgalla
gaff 8 main upp .er 5 fore pper
tops i top ail
topsail

6 fore
X ain lower % lower.,
topsai
topsail

spanker 10i
mair@sail foresail 4
7

1. flying jib 4. fore topmast staysail 7. main topmast staysail io. mizzen staysail
2. outer jib 5. main royal staysail 8. mizzen topgallant staysail
3. inner jib 6. main topgallant staysail 9. mizzen topmast staysail
Figure 2. Sail plan of USCGC EAGLE (from EAGLE Seamanship).

Administration, it set in motion events which numbered approximately 200. Cadets, Academy
caused significant departures from the norm in instructors, and personnel assigned full time to
Academy education planning and led to the the ship comprised the crew. A complete change of
unprecedented voyage of EAGLE to Australia. instructors and cadets took place in Australia on
30 December. In all, about 280 cadets were
2. ACADEMIC PLAN involved.
Members of the EAGLE's permanent crew made the
Australia's bicentennial celebration was 26 entire eight month voyage. The forty five officers
January 1988. EAGLE left New London on 10 and enlisted personnel were the backbone of each
September 1987 to travel the 14,000 nautical miles phase's crew for this a 'rduous undertaking.
distance to arrive at Sydney, Australia on time. During the semester at sea two courses were
It was 6 May 1988 before EAGLE returned home to New given to the sophomores, known as "Third Class"
London. cadets (Class of 90) while the seniors, known as
Since cadets made up the majority of EAGLE's "First Class" cadets (1988 year class) received
crew and the voyage occurred during the Academy's one. Third Class cadet courses were:
1987 Fall and 88 Spring semesters, special Organizational Behavior and Meteorology and
arrangements were devised to provide academic Physical Oceanography. The one the First Class
courses. A constraint was that those cadets who cadets received was a 4.7 credit hour nautical
participated still had to'complete their education science course entitled Deck Watch Officer. Its
in four years. topics included: Shiphandling, Sail Handling,
For the first time in the Academy's modern Navigation, Watchkeeping, Communications, and Rules
history academic courses were held in the summer. of the Road.
The entire Class of 1990 and about fifty members of Topics covered by the Meteorology and Physical
the 1988 class attended the summer session to Oceanography Course included: weather elements,
complete two thirds of a normal semester course atmospheric circulation, weather notation, weather
load. The other third of the load was received chart analysis, physical properties of seawater,
onboard EAGLE during the voyage. oceanic circulation, waves, and tides.
The cruise was divided into two phases which This course was very well suited for
coincided with the Fall and Spring semesters back presentation at sea. Everyday, observations of
at the Academy. EAGLE's crew for each phase clouds, sea state, and other environmental

692

AUSTRALIA 1 88
9/10/87-- 5/6/88
60 -Will
4 O%b
NEW
I)LONDON
30

%%
H '7
AWAII

SAM GA TAHITI

TONGA
-30
Y

Phase 1
-60 Phase 2 ---
-120 -150 190 150 120 9 0
figure 3. Australia '88 Cruise Track.

conditions were used for in-class discussions. cutters, i.e. High Endurance Cutters and Polar
Nature cooperated often by providing phenomena to Icebreakers. The following goals could then be
reinforce lessons. For example, Hurricane Emily's met. EAGLE would serve as a platform of
crossing of EAGLE's track in late September 1987 opportunity for collection of environmental data
fit nicely with the lesson on tropical cyclones. along its track. Practical experience in
By having both marine science and nautical observation and data analysis would be passed to
science co'urses, the point was stressed that the cadets. The observations would be used daily
rudimentary knowledge of oceanography, as well as to reinforce the weather and oceans lessons. And,
navigation skills are important to the safe environmental data would be available for input to
operation of an ocean-going vessel. the ongoing operational planning for the ship
(especially essential for a sailing ship).
3. DATA COLLECTION PLAN Projects were planned with the Naval
Underwater Systems Center (NUSC) in New London and
Prior to this voyage EAGLE had no with NOAA. These included synoptic weather
environmental sampling capability other than that observation, expendable bathythermograph (XBT)
which is typical of smaller Coast Guard vessels. casts, collection of surface water samples for pH,
Observation of weather and the transmission of the and marine mammal observations. Also set up was a
encoded data was done, but, only once or twice per cadet project to collect plankton in the South
day because the ship was billeted with only one Pacific.
radio operator.
It was decided in the early stages of cruise 4. EQUIPMENT
planning to install oceanographic and meteorologic
sampling equipment onboard EAGLE to upgrade its Equipment needed for EAGLE's science program
-capabilities to match those of larger Coast Guard came from a variety of sources. Primary

693

B XBT SECTION('00 TAHITI LEG 101S.7 A
T160:S 910 . S
1 1V @iz Y@ 2:W
@4 0 (62. 0 6.
T,

50-
2
100-
2 3
2
150-

2100-

250-

300-

350-

400-

45
rneTers

Figure 4. Temperature cross-section A-B in South Pacific Ocean.
Station 50 is in the Galapagos Island Group.

contributors were: NOAA's National Marine Fishery polar orbitting satellites when they were over the
Service (NMFS) and National Ocean Service (NOS), ship. Two display monitors were connected to the
NUSC, and the Academy's Department of Science. WSR. One located in the Navigator's space, near
NOAA extended tremendous assistance by the bridge, was used by personnel conning the ship.
providing the advanced SEAS (Shipboard The other was in the classroom and was used daily
Environmental Data Acquisition System) for for demonstrations. Between the two pieces of
collecting and transmitting (via satellite) weather equipment, EAGLE was able to keep track of changing
and XBT data - and enough XBT probes for the weather patterns.
entire cruise. SEAS basically consists of IBM Modern methods of navigation, such as SATNAV,
PC-compatible microcomputer, GOES satellite OMEGA, and LORAN C, were taught to the cadets and
transmitter with omni-directional antenna, and XBT @used inconjunction with traditional celestial
controller and launcher. Complete description of navigation. Often, while in the open ocean, the
this system was provided by Szabados, Roman, and electronic equipment would be turned off so*use of
Taylor, 1987. the sextant could be emphasized. The blending of
Essential for obtaining weather information technology with traditional nautical skills proved
for EAGLE's operational planning and for use in the to be an interesting and effective approach to both
science course were a weather facsimile recorder the marine science and nautical science subjects.
(WX FAX) and a weather satellite receiver (WSR).
Weather charts were received over WX FAX from 5. OBSERVATIONS
broadcasting stations which covered the areas of
ocean the ship transited. The WSR received The kinds of environmental observations and
infra-red and visual images from NOAA-9 and NOAA-10 the numbers of each made are shown in Table 1.

694

Table 1: Oceanographic and meteorologic will gra duate to the fleet and again be involved in
observation made during Australia '88 Cruise. XBT casts, weather reporting, and other
oceanographic studies. The program on EAGLE
provided the cadet's valuable hands-on experience
Synoptic Weather 630 observations that they will carry with them throughout their
careers.

XBT 230 casts (T6 and T7) 7. ACKNOWLEDGMENTS

PH 81 observations The author wishes to express his appreciation
to NOS for providing SEAS. I am especially
Plankton tows 18 samples grateful to Gerald Bloom and Dave Pritchard for all
their help with assembly and installation of the
Marine Mammal 41 sighting reports system onboard EAGLE. The help provided by Charley
Brown of NUSC in putting together equipment is also
acknowledged.
Captain Ernst Cummings and the entire crew of
Synoptic Weather. When not in port, weather EAGLE are gratefully acknowledged for their
observations were made four times daily on the invaluable cooperation with the onboard science
synoptic hour. The data was encoded on NOAA Form programs. In particular, ET1 Ray Weidauer is
72-1A and relayed via SEAS to NOAA and US Navy thanked for his technical support.
forecast centers.
XBT. Drops were usually done twice daily, in 8. REFERENCES
the morning and evening, when EAGLE was in depths
greater than four hundred meters. This data too (1) Mellen, R.H., P.M. Scheifele, and D.G.
was sent back to NOAA and Navy users by SEAS. Browning. 1987. Global Model of Sound Absorption in
PH. Sampling the surface waters for PH was Sea Water. Scientific and Engineering Studies,
done once per day while underway from 12 to 29 Naval Underwater Systems Center, US Navy.
September 1987 and 11 November 1987 to 8 February
1988. This data was provided to NUSC researchers (2) Regan, P.M. and P.H. Johnson. 1986. EAGLE
for their ongoing study of the effects of PH on Seamanship; 2nd Ed. Naval Institute Press,
sound attenuation in seawater (Mellen, Scheifele, Anapolis, MD.
and Browning, 1987).
Plankton Tows. Collection of plankton was (3) Szabados, M., R. Charles, and B. Taylor. 1987.
done whenever the opportunity was presented. Which Transmission of Real Time Oceanographic and
was when winds were very light or when the one main Meteorologic Data from Ships. Proceedings of the
diesel engine was down for repairs. A Third Class Oceans `87 Conference.
cadet collected the samples for an upper-level
research project to qualitatively describe plankton
along EAGLE's South Pacific track. Interesting.
instances of diurnal vertical migration and
upwelling productivity were evidenced in
preliminary analysis of the samples.
Marine Mammal Observations. On all occasions
when marine.mammals were seen, sighting reports
were completed. This data was provided to NMFS and
NUSC researchers.

6. CONCLUSION

Real-time XBT and synoptic weather data
collected on this cruise was distributed to
interested users, including forecasters and
researchers studying global ocean-atmospheric
interactions. Data collected in areas of Pacific
Ocean outside the usual sealanes should be of
particular interest.
The skills gained by the cadets in observing
and interpreting weather patterns went beyond what
could be provided in a conventional classroom
setting. On a daily basis EAGLE served as an
ongoing lab for sharpening these skills and
reinforcing the classroom presentations. Due to
the success of this course, plans to conduct a
similar weather course during normal summer cruises
are being developed.
The Coast Guard has a long history of
participation in NOAA/Navy environmental programs.
Many of the cadets who took part in this cruise

695

UNITED SMTES NAVY OPERATIONAL oaEMEGRAPHY: FIGUIW- SNAW WITH OCEWO@ INTELLIGENCE

Lieutenant Ocumander James A. McNitt, United States Navy

Office of the Oceanographer of the Navy
34th and Massachusetts Avenue, BW
Washirigton, DC 20392-1800

ABSTRACT Acquisition of hardware and software must be
inrx:yvative, cost-effective, and affordable and
A key factor in enhancing the U.S. Navy's warfight- lead to nOn-duPlicative but survivable systems.
ing capability in the 1990s is the ability of the This Process is dependent upon a strong
on-scene ccmTander to utilize and exploit environ- oceanography technology base, a closely coordin-
rental information in real-time. In order to ac- ated and focused research and development pro-
cmplish this, Navy scientists and engineers are gram, and full use of available technology and
developing integrated, tactical support systems equipment.
which the fleet can use at sea. These support
system may be tailored for the specific mission of PRESENT OCFANoGRAPHY ARcHrmc=
the platform or battle group as required. The tac-
tical support system meld a combination of real- The present Ncmss architecture overtaxes the
tire data (remotely sensed or in-situ measurements) ccuputer resources at the Fleet Numerical
with historical databases. These data are then Oceanography Center (ENOC). Located in Monterey,
used as input to sophisticated predictive models, California, FNoc generates Navy unique gicbal and
the output of which are used as aids to the Naval regional scale atmospheric and oceanic rcdel
Oceano@@pher supporting the tactical derision output data and products. Products are transmit-
maker. Ea#iasis is placed on usable and under- ted as vector graphics data on the Naval
standable displays and ease of dissemination of Enviramental Data Network (NEEN). Additionally,
products. This capability is further used when FNOC trarLmnits fleet application products directly
@iwirtrmental factors are considered in all Navy to ships by naval message. Fleet applications
weapm, sensor, and cmmmnication systems from products are tailored to specific weapon, ccnmm-
early design tl@rough test and evaluation to full ication, and sensor systems. These products rely
operational capability. on historical and observed data to assess system
effectiveness (e.g., sonar performance).

The Naval oceanographic Office (NAVOCEANO)
INTRODUCTION Operational Oceanography Center, located at the
Stennis Space Center, Mississippi, uses in-house
The U.S. Navy operates, trains, and fights in expertise and global oceanographic data collec-
the global air-ocean envirorment. over the past ticn and assimilation to provide regional and
few years, the operating forces have increased local scale analyses of remtely sensed data to
their awareness of the effects of the air-ocean oceanography centers and fleet users by naval
environment on weapon, cmwinication, and sensor message-
system. The primary goal of the Naval
Oceanographic and Meteorological Support System The three regional oceanography centers and
(NCHSS) is to enable all fleet tactical decision the two oceanography cmmand oextiers use the data
makers to make fall use of and exploit real-time received on the NEEN to provide products to fleet
oceanographic information, including: oceano-- units by three methods: fleet broadcast (radio
graphic; meteorological; and mapping, charting, teletYpe), alphanumeric products by naval message,
and geodesy information. and satellite imagery and graphics by facsimile
broadcast. In addition to the NEDN, the ooeano-
The foundation for a revolutiorkuy NCHSS is graphy centers are connected to FNOC by the Naval
being built. Key to the oceanographic Environmental Satellite Network (NESN). The NESN
architecture of the 1990s are: provides Defense Meteorological Satellite Program
(U,ZP) imagery to the Naval Satellite Display
1. Regionalization. Stations (NSDS). The NSDS is an image processor:
2. Stand-alone battle group capability. 8-bit oriented with a 256 gray-scale and animates
3. Improved use of cmmmications. up to 50 images per loop.1 The oceanography

696 United States Government work not protected by copyright

centers are a igned the following areas of 2. Bathythermal profile of ocean (temperature
responsibility (ACR): versus depth).

1. Naval Eastern Oceanography Center (NECIC) 3. Rawinsande derived atmospheric sounding of
Atlantic Ocean taTeratum, pressure, humidity, wind speed, and
wind direction.
2. Naval Oceanography Ccumand Center Rota
(NOM) Mediterranean Sea 4. Electronic Refractcmeter Set installed
aboard carrier-based early airborne warning
3. Naval Western Oceanograp@y Center (NWOC) aircraft (E-2C) to measure the refractive index.
Eastern Pacific Ocean
Fbr the most part these point measuranents do
4. Naval Oceanography Ccumand Center Guam not accurately account for horizontal changes in
(MCC) Western Pacific and Indian ocean atmospheric and oceanic parameters. The TESS(2)
units are now being upgraded to receive and
5. Naval Oceanography Polar Center (NPOC) process low resolution satellite data from the
Polar Oceans AN/SM2-6. Although TESS(2) enables the naval
oceanog rapher@ to determine sea-surface temperature
Enlisted and officer oceanography specialists gradients in cloud-free area , he still cannot
are assigned to aircraft carriers, amphibious apply it to the applications products.
ships, and battleships. During the last two Additionally, he is still dependent upon the
decades the availability of satellite imagery has oceanography centers for analysis and forecast
dramatically improved the oceanographer's products-
capabilities at sea. The AN/SM-6 Satellite
Receiving Set is the standard satellite receiving Limitations of the present architecture
set for afloat units. It receives and reproduces include:
selected AFT pictures transmitted from NOAA polar
orbiting satellites and other foreign satellites. 1. oceanography centers and the fleet depend
Some AN/SK@-6 units have been modified to receive upon FNOC for all Navy oceanographic and meteoro-
GOES and Eurqpean Space Agency Meteorological logical products and fleet application products.
Satellite (MBOSAT). The AN/M2-10 DMSP oceanography centers have a limited capability to
Reoeiver-Recorder Set for Meteorological Data is generate products or provide input to FNOC
the Navy standard EMP satellite reoei-hng set. products.
It is only available on a limited number of
carriers. Both the AN/SQ-6 and AN/S4(@-10 have 2. Although TESS(2) has streamlined aspects
reached the end of their lift
,Tles. The first of afloat oceanography office routine, the battle
AN/SM@-6 was purchased in 1969. Maintenance and group is still dependent upon significant
repair are no longer cost-effective. quantities of data from the oceanography centers.

All aircraft carriers and aT#iibicus ships
have the Tactical Enviram-ental Support System OCEANOGRAPHY ARCHITECTURE OF TM 19901S
(TESS(2)]; a data processing and display system
hosted on the HP-9020 desktop cmpxter. TESS(2) Two parallel efforts will determine the future
databases include extensive historical oceano- of naval oceanography: regionalization of the
graphic data; weapon, cammunication, and sensor shore-based support network and develcprent of a
system parameters; and envimnmental applications starxi-alone capability for the battle group using
programs- advanced ccnputer technology and assimila- tion of
satellite data in tactical scale nowcast models and
These program provide products such as: applications products.
aircraft icing analysis, tidal prediction, sound
speed profile, acoustic propagation loss, acoustic Regionalization. The Oceanographer of the
raytrace, electrawagnetic propagation summary, Navy is sponsoring the development of a "store and
electromagnetic path loss versus range, forward" capability to increase data ocmmuni-
electromagnetic coverage diagram, and ballistic caticnus from FNOC to the oo&viography centers
and densities corrections. through the distribution of gridded data,
primarily analysis and forecast products. In
Real-time, locally cbserved inputs which can addition to grid field data, the oceanography
be used to provide environmental products using centers will receive raster scan, alphanumerics,
existing processors and resources are 1=ted to: and binary data. The oceanography centers will
1. Surface observations of air and sea-surface generate the products for their ADRs, provide
(sea injection) taq3P_mture, wind speed and input to FNOC's models based on local knowledge,
direction, relative humidity, and atnoqlieric and produce all fleet applications products on a
pressure.

697

co-lor-ated TESS(3).3 Additionally, regionaliza_ Related Efforts. other resource sponsors
tien will reduce the potential for a single point within Navy'are funding system acquisitions which.
of failure in data production. directly involve Naval Oceanography. The AV-8B
TESS(3) and AN/SK@-ll. A minim= of 71 aircraft relies heavily upon digital map data-
bases. The Navy Research, Development, and
-TESS(3) units will be procured starting in 1990 Acquisition policy directs all program to con-
for installation at shore sites and ships with sider appropriate environmental factors in Navy
.oceanography officer and enlisted personnel. The weaporVsensor systems from program initiation
TESS(3) full scale engineering development (ESED) t1uvagh test and evaluation to full operational
contract was awarded to lockheed Missiles and capability.5 Naval oceanographers working within
Space Ocapany, Inc., on 11 July 11988. The FSED the Navy acquisition organization advise program
phase will last only two years. operational managers in this critical area. Many of these
testing on an east coast aircraft carrier is programs will be supported by TESS(3).
scheduled for 1990 with initial operational
capability scheduled for 1991. CCNCIIJDING PjaAkPIZ

The aircraft carrier configuration of TESS(3) Tactical Naval oceanography is more than an
consists of three high-resolution image work architecture. It is an aggressive, innovative,
stations and three coapxter processing units and exciting cmbination of experienced
(CPU), each containing 16 megabytes of memory, professionals and advanced technology. Industry
operating at 20 million instructions per second is playing a key role in that TESS(3) pushes
(KIPS). Satellite imagery can be displayed as available technology to its limits. The Navy
1024 by 1024 by 8 bit images with a 4 bit graphic laboratories are responding to an extraordinary
overlay and a 256 gray-scale. Fiber optic challenge during a period of fiscal austerity.
connections between caq)onents and interfaces will Finally, the oceanographer of the Navy has placed
prevent electromagnetic interference and enhance tactical scale support at the top of his priority
electrcmagnetic hardening. TESS(3) makes full use list in order to achieve a near-term inprovement
Of distributed processing. It is capable of in services which is sure to have a widespread and
receiving and processing satellite data thrvugh the dramatic effect in the fleet.
AN/SM-11 while generatinj up to three applications
products, including both image and graphic REFERENCES
processing- The data management functions will
automatically run the forecasting, application 1. CNOC. Satellite Data Utilization Plan, 1988.
program control, and observer functions. Camronder, Naval Oceanography command, Stennis
TESS(3) automatically monitors and nkinages, Space Center, NS 39529-5000.
satellite data received from the AN/SMQ-11, which 2. CNOC. Naval Oceanography Equipment plan,
replaces the AN/SK@-6 and AN/SM-10 receivers. Vol. 1, 1987. Meteorological Equipnent.
The AN/SMQ-11 receives direct transmissions from Ocamander, Naval Oceanography Camend, Stennis
polar orbiting satellites providing coverage of Space Center, NS 39529-5000.
approximately 1,600 nautical miles around the
receiver. These data include: TIMS operational 3. CNOC. Concept of operations for the NEDN
Vertical Samder (TOVS) and Advanced Vexy High Oceanographic Data Distribution and Expansion
Resolution Radiometer (AMM) data from the TIROS- System, 1988. Comakinder, Naval Oceanography
N/NOAA satellites and operational Linescan System COUMM-1d, Stennis Space Center, NS 39529-5000.
(OLS), Special Sensor Microwave Imager (SSM/I),
and Special Sensor Microwave Temperature Soundex 4. COMSPAMUMSOCM. contract Specification for
(SSMIT) data from the EMP satellites. It also Tactical Environmental Support System [TESS(3)],
receives WUAX data from GOES. The receiver and Vol. 1. System level Performance and Design
processing hardware are housed in twin 19 inch Requirements, 1987. Coamander, Space and Naval
racks. The trainable antenna system consists of Warfare Systems Command (PM-141), Washington, DC
two planar arrays.4 20363-5100.

The Naval Environmental Prediction Research 5. OPNAVINST 5000.42C. Research, Development,
Facility (NEPRF) and Naval ocean Research and and Acquisition Procedures, 1986. Chief of Naval
Development Activity (NOMA) are developing Operations, Washington, DC 20350-2000.
predictive models for TESS(3). Tactical scale
ocean circulation and thermal structure models and
atmospheric forecast models will require
initialization data from the oceanography centers.
innovative and technologically advanced
assimilation techniques are being developed to
enable TESS (3) to create howcasts from data
received frcm the AN/S1,1Q-11. This capability will
enable the battle group oceanographer to provide
briefings and tactical reconwendaticris when data
are root available from the oceanography centers.

698

WINDROSE, PC SOFT`WARE FOR WIND DATA ANALYSIS

Kyle Monkelien, Thomas L Murrell

U. S. Minerals Management Service, Alaska OCS Region

ABSTRACT CRITERIA USED IN DESIGNING THE PROGRAM

The Minerals Management Service (MMS) requires that Several criteria were established for the development of
lessees collect and submit meteorological and this program:
oceanographic data for offshore oil and gas o t'
oyera ions in
frontier areas. The collection and anal ses these data 1. The pro ram would have to be able to use
are necessary to establish a baseline =@Ich can be used informat from several databases and spreadsheet
to evaluate Critical Operations and Curtailment Plans files.
and other regulatory requirements. To help in the
analyses of @yind data the MMS Alaska Outer 2. The program would need to handle variable amounts
Continental Shelf (OC�) Region has developed a of data.
computer program which will produce both a tabular 3. The rograrn would have to run on IBM PC-
a
m7 and agaphic summary in the form of a rose
sdui am8 Ta The 12 depicts the relationshi atidle hardware for ease of use and flexibility.
ayam _p of the comp
percentage of in five magnitude intervals to the
direction of occurrence. The program is designed to 4 *The program would have to be user friendly and menu
work with an IBM personal computer, or compatible, driven.
with an Enhanced Graphics Adapter (EGA). Hardcopy 5. The program would need to provide hi h-quality
of the rose diagram can be obtained through the use of a
Hewlett Packard (HP) plotter and a tabular summary graphs ffir use in publications and also provUe tibular
Can be output through the use of the print screen summaries of the analysis.
function.
Figure 1.
A review of current data-base and Wind data.
spreadsheet software established
that an ASCII format could be
utilized to extract previously stored 20 12
data for use with the rogram. A 40 14
WHY THE PROGRAM WAS DEVELOPED suitable format for thke Cata was 170 15
selected: the data were arranged
into two columns with the wind 230 22
The Minerals Management Service requires that lessees direction in the First column and 240 15
collect and submit meteorological and oceanographic the wind speed in the second 220 9
column (figure 1). Empt
data for OCS oil and gas operations in frontier areas. y lines 200 17
The data collected by the operators includes data on (rows) or page breaks in the data 200 17
wind speed and dire n, air and water temperature, field were unacceptable. ASCII
_@Oibjlity, current, and, where files could be produced by several
relative humidity, cris 200 18
applicable, data on ice coverage and ice movement. commercial software rograms, 200 21
This is used to establish a baselifie for conditions which including EXCEI, eymphony 210 21
can be encountered in frontier areas such as the DBASEIII Plus, as well as standard 200 18
Beaufort Sea, the Bering Sea, and the Chukchi Sea. The editors and word-processin 200 20
baseline is used to evaluate Critical Operations and programs. A simple program m 9
Curtailment Plans and other information which is could be used to enter wind data 210 22
submitted in support of exploratory activities. The without the use of commercial 210 18
WINDROSE program was designed to analyze the wind database and spreadsheet roWrams; 11o 22
component ol the data supplied by lessees. Previous was also written. The WRID OSE 100 23
analysis of the wind data was accomplished through f] roNram was designed for use on,an 110 25
manual calculations and the hand drafting, of rose -compatible computer running
diagrams. This method of analysis was becoming DOS with an EGA monitor. Aii
inefficient due to the large, amounts of data being HP @o= any standard printer
submitted by the companies. When we were can e producing a hard copy. Microsoft Quick
unsuccessful in our attempt to find a commercial Basic was used as the programming lanwe, and the
program which would meet the needs of our office, it drafting commands were written in Graphics
was decided to write a program inhouse. Language.

699 United States Government work not protected by copyright

HOW THE PROGRAM ANALYZES DATA Figure 3. Screen 2 - main menu.

The program provides an analysis of the percentage
occurrence of @vind speeds from eight direction ranges. OPTIONS LIST FOR FILE - A. SAMIPLEI.PRN
The wind speeds are divided into six ranges: (1 cal
R2) greater than 0 knots but less than or eqD to 1% 1. PRINT OBSERVATION SUMMARY
ots, (3) eater than 10 knots but less than or equal to
21 knots, 41) reater than 21 knots but less than or equal 2. VIEW WINDROSE
to 33 knots M greater than 33 knots but less than or
equal to 47 knots, and (6) reater than 47 knots. The 3. SELECT NEW FILE
wind directions aregiven in 5egrees and are divided into
eight equal intervals. The program first examines the
data to determine the number &f occurrences per wind 4. QUIT
direction. It then calculates the percentages from each MAKE SELECTION: 7
direction and uses that figure to draw a rose diagram.
The program also will produce a summary of the. rose
diagram in tabular form which can be printed using a
print screen function.

USER INTERFACE AND PROGRAM OUTPUTS

Screen 1
The first screen (figure 2) rompts, the user to input the Figure 4. Printout of observation summary-- hardcopy
name of the file to be analyzed. At this point, the user output under Option 1.
can also ask to view a diiectory listing or to end the
program. OBSERVATIONS SUMKARY FOR LEASE SAMPLE1 PRN WELL NO. 01
BEGINNING - 00/OD/88 ENDING 00/00/;9
Figure 2. Screen 1. KAGNITUDES, CALM - S - 0 1 - O<S<-10 2 - 10<S<-21
(S-KNOTS) 3 - 21@S-33 4 - 33@S<47 5 - S>t.7

KhGNITUDE DIRECTION OBSERVATIONS

I N (338-22) 4
I HE (23-67) 1:
ENTER DRCVE:PATH\FILENAME], [D] FOR DIRECTORY, OR [E] TO END:? I B (68-112 7
1 BE (113-157) 8
I S (158-202 ) I
I SW (203-247) 2
1 W (24 a2
9;2 9 1 12
1 NN (2 -337) 6

2 M (33 "-22) 6
2 ME (23-67 )
2 E (68_112) 29
2 BE (113-157) 13
2 S (158-202) 41
2 Sw (203-247) 37
2 W (24 a-292) 19
2 RW (2 93-3 37) 10

3 N (338-22) 0
3 ME (23-67) 0
3 E (68-112 ) 35
3 SE (113-157) 6
3 S (158-202) 10
3 SW (203-247) 27
3 W (248-292) 19
3 KW (2 93_33 7) 2

4 N (33 1-22) 0
4 ME (2 3_67 0
4 z f68-112 0
4 SE ( 113-157) 0
Screen 2: Main Menu 4 S (158-202) 0
4 SW (203-247) 4
4 W (24:-292))
The second screen (figure 3) lists four main options. We 4 NW (2 3-337 0.
will describe the options briefly here then in the next
section, we will discuss Option 2 -- the most powerful 5 N (338-22) 0
and versatile part of the program -- in detail. 5 ME (23-67) 0
5 R ( 68-1121 0
5 BE (113-157) 0
r, S f158-202)
5 Sw (203-247) 0
5 W 1248-292) 0
Option 1, PRINT OBSERVATION SUMMARY, sends 5 MW (293-337) 0
a summary to the printer. This summary lists the number CALK 3
of observations for each wind-magnitude interval as well
as the total number of observations (figure 4) . TOTAL OBSERVATIONS - 375

700

Option 2, VIEW WINDROSE, allows the user to view Figure 6. Windrose screen display under Option 2.
the windrose; to view a summary of the percentages; and
to plot the windrose, either alone or with the percentage
summary. 'Unm to CmTlmq

.30

Option 3, SELECT NEW FILE, enables the user to
select a new data file for analysis.

Option 4, QUIT, terminates the program and returns the
user to DOS.

Option 2: View WINDROSE
As mentioned above Option 2 allows the user to view
the windrose, to view the percentage summary, and to
plot the windrose either alone or @vith the percentage
summary.

When option 2 is selected, the user is first asked to enter
lease and well numbers and observation dates figure 5).
This information is not required; it is slEqply i2flyng
th , i Figure 7. Percentage summary -- screen display under
information that, if enter@T, will appear in t e title me -
of the percentage summary. Option 2.

Figure 5. Screen display under Option 2.
COMIATIVE WIND SPEED SUNKhRY FOR LEASE SAXPIZI-M WELL NO- 01
BEGINNING - 00/00/88 ENDING- 00100/89
NACNITUDS 1 2 3 4 5 TOTAL
PLEASE ENTER LEASE AND WELL INFORMATION. MAGNITUDES
NORTH 1.07 1.60 0.00 0.00 0.00 2.67
1: D<S<-IO
NORTHEAST 4.00 5.07 0.00 0.00 0.00 9.07
2t 10<S<-21
EAST 1.87 7.73 9.33 0.00 0.00 18.93
LEASE NUMBER: ? 3: 21<S<-33
SOUTHEAST 2.13 3.47 1.60 0.00 Q.DO 7.20
4: 33<S<-47
WELL NUMBER: ? SOUTH 5.07 10.93 2.67 0.00 0.00 18.67
5: S>47
SOUTHWEST 6.40 9.87 7.20 1.07 O.CD 24.53
BEGINNING DATE OF OBSERVATIONS USED IN PLOTTING WIND ROSE - ?
WEST 3.20 5.07 5.07 0.00 0.00 13.33

ENDING DATE OF OBSERVATIONS USED IN PLOTTING WIND ROSE - ? NORTHWEST 1.60 2.67 0.53 0.00 0.00 4.80

TOTAL 25.33 46.40 26.40 1.07 0100 99.20

PERCENT WINDS 99.20 PERCENT CALK 0.80 TOTAL 100.000
DO YOU WISH TO VIEW THE WIND ROSE ArAIN (Y OR NR I

At this point, the program offers several options:
The user can view the windrose once more, either
Next, the windrose is disDlayed on the screen (figure 6) . at the same size or enlarged. (An enlarged
In the windrose, the wind magnitude is represented by windrose is useful in some instances when the wind
the width of the boxes, and the percent occurrence by is uniformly distributed.)
the box length. With a color moffitor the boxes are also
differentiated by color. A circle scale is provided to
assist in evaluating the percentages shown on the
@iagram. The percent of 6din. observations is displayed The user can plot the windrose either with or
in numeric form in the center of the diagram. The cWm without a percentage summary. (figure 8 shows a
percentage is rounded to the nearest whole number. windrose plotted With a. percentage summary.)
The user can also add a title and/or narrative
description.

Next the user presses the return key and the percentage
summary is displayed on the screen (figure 7). The user can return to the main menu.

701

WINDROSE WITH SUMMARY
THIS PLOT WAS DEVELOPED USING DATA CONTAINED IN FILE
SAMPLEI.PRN

tN CUMULATIVE WIND SPEED SUMMARY (PERCENT)

MAGNITUDE 1 2 3 4 5 TOTAL

NORTH 1.1 1.6 0.0 0.0 0.0 2.7
NORTHEAST 4.0 5.1 0.0 0.0 0.0 9.1
ix EAST 1.9 7.7 9.3 0.0 0.0 18.9
SOUTHEAST 2.1 3.5 1.6 0.0 0.0 7.2
SOUTH 5.1 10.9 2.7 0.0 0.0 18.7
SOUTHWEST 6.4 9.9 7.2 1.1 0.0 24.5
WEST 3.2 5.1 5.1 0.0 0.0 0.3
NORTHWEST 1.6 2.7 .5 0.0 0.0 4.8
TOTAL 25.3 46.4 26.4 1.1 0.0

n WINDS 99.2
CALM .8
TOTAL 100.0
0 10 20 30 40 50 60 70 80 90 100
WIND SPEED SUMMARY

MAGNITUDE:CALM - S - 0 : I O<S<-10 : 2 iO<S<-21
(S-KNOTS) 3 - 21<S<-33 4 - 33<S<-47 5 - S>47
CALM 1 2 3 4 5

Figure 8.
Plot of windrose with percentage summary -- hardcopy output under Option 2.

CONCLUSION Copies of the program and source code are available
through the:
The WINDROSE pro am was developed to provide an Mineral's Management Service
easy method for anglyzing wind data submitted by
0 erators conducting @exp oration activities on the OCS. ATTN: Librarian
e r summarizes the observations and provides
* sta istical. an is of the data in either a rose diagram 949 E. 36th Ave., Suite 110
* as a table su ary. It will enhance the establishment
of the baseline for conditions which can be expected for Anchorage, AK 99508
operations conducted on the OCS. The program could
easily be modified, by an
,y programmer faihiliar with
BASIC, to process other data requiring similar graphic
depiction.
r

702

MEASUREMENT OF LUMINANCE DISTRIBUTION ON THE SEA SURFACE
FOR COMFORTABLE LIVING SPACE

Maki Enomoto, Toshimasa Kawanishi and Wataru Kato

Department of Oceanic Architecture & Engineering
College of Science & Technology, Nihon University
7-24-1 Narashinodai Funabashi-City Chiba, Japan 274

ABSTRUCT
The purpose of this report is to obtain a Sky Light

basic data of the lighting and daylight
Di t nligh
plan to design the comfortable indoor

lighting environment.

The light incident in a room through a
Reflected Light
window in a structure has three kinds of of Sea Surface

lights, direct sunlight, sky light, and
Fig. 1 The Incident Light to The Structure
reflected light of ground. in Coast and Offshore

A reflected light of sea surface to the 1. INTRODUCTION

structure in a coast or offshore zone is T h e amount of i n d o o rd a y 1 i g h t

bigger than the light of ground. illumination from a sea surface is mainly

The special consideration needs to be influenced by a sun.elevation, a sea

paid for lighting and daylight plan to surface condition, and other elements.

design the comfortable living space in the If the weather is fine, the incident

ocean optical environment. light from a sea surface is great.
We have seldom come up to these studie SI.)2) Fig.1 indicates the incident light of the

In this report, we will discuss ab out the structure in coast and offshore. Fig.2

measurement of horizontal and vertical indicates the view of reflected light of

distribution of sea surface luminance sea surface.

against a sun elevation. The way to treat reflected light of the

To carry out our study, we developed a sea surface would be difficult because of
digital image processing system using a the changes in the sea surface conditions.
Charge Coupled Device(CCD) camera as a Fig.3(a) indicates a reflected light of
Di@. 0',Sunii.

measurement system. the smooth sea surface. And it is a

specular reflection which reflects

703 United States Government work not protected by copyright-

Video Signal
Video Tape Recorder
CCD Camara
Video Signal

Video Signal Luminance Signal

Image Processing Board
47-

Fig. 2 The Reflected Ligbt of Sea surface Monitor Television Personal Computer

Fig. 4 Block Diagram of Measurement System
Direct Sunlight Table. 1 Machine Parts a .nd Specifications

CCD Camara Camara Tube Sposification Solid State Image Specificat-
ion (Type. of Interline)
Camera Tube Area 8.BXS.6[=j (size of 2/31n.)
Effective Picture Elements 766(10 x493(V)
SIN Ratio More Than 501dBI
(a) Smooth Sea Surface Hough Sea Surface Gum Corrector 7

l6mm Standard Lens Horizontal Angle of View 30.40*
Fig. 3 The Condition of Sea Surface Vertical Angle of View 23.18*
exceedingly enormous amount of light in image Processing Bo- Resolution 256 X 256
ard A/D Converter 6 Bits 64 Tone
regular reflectionofthe sunlight. Sampling ?ims Image Tim 1/60 [Sac]

Fig.3(b) indicates reflect condition on analog digital converter o f i m a g e

the rough surface. The light which processing board "FDM-4"(PHOTORON-MADE).

reflects diffusely by the trend of wave This converted image is analyzed by a

become much less than the amount of light personal computer "PC-98O1VM-21"(NEC-
by the specular reflectiorl@) MADE).

2. MEASUREMENT SYSTEM 3. CHARACTER AND CORRECTION OF

Fig.4 indicates the block diagram of EACH SYSTEM

measurement system and table.1 indicates The spectral responsivity of CCD camera

details of each equipment. largelydiffers from the standardrelative

The luminance distribution appears as an luminous efficiency V(X) w h i c h is

illuminance distribution on image space on indicated in Fig.5. Therefore, we used

CCD camera "XC-77"(SONY-MADE) through the a color correction filter "No. 102"

16(mm) standard lens. (KODAK-MADE) and an infrared-rays-cut

A v i d e o s i g n a I w h i c h is directly -filter "IR."

p r o p o r t i o n a 1 to an illuminance As a r e s u 1 t , t h e r e is o n 1 y 5-9%

distribution is transmitted from a CCD difference between the results of this

camera. And the video signal is converted correction and the standard relative
F

into the luminance signal of 64 tone by luminous efficiency.

704

100 o CCD
0 IR The condition of the weather and sea
A 102
r > A Corrected during the measurement was sunny, 1 cloud
Value
or
U V(A)
9: amount, 2 Beaufort wind scale, 2 wind wave
0 60
A
scale, and 0 swell scale.
Id
It becomes cloudy and windy after 16:00
20
U and the swell scale increased to 1.

Table.2 indicates meteological a n d
0 400 Soo 600 700 800
Wavelength(nm)
oceanographic conditions.
Fig. 5 Relative Luminous Efficienty curve
10 The amount of a erosol in the air was

enormous; we could not even observe the

0
sun at sunset.

0.1 5. METHODS OF MEASUREMENT
10, 10, 10,
Luminance lcd/m') We fixed the optic axis of lens at the
Fig. 6 Ralation of Luminance and Reference Voltage
'height of human eyes and took photos of

Fig.6 shows t h e relativity of the the reflected light of sea surface which

incident luminance and the reference varied from time to time. Also, we tried

voltage on the imag e processing board T*k'*
according to the F-number of a camera. Table. 2 Meteorological and
The 1 i g h t source f o r fig.6 w a s the Oceanographic Conditions Tky. Bay
F-1--ci,7.d7T..d 5-11 1..T7 T,.d S;.
lambertian surface made out of a white Ti.o A-1. F-. S..J. T... T. 1. xl"Btio*
S-1. S.fl. (11/S
147- _T_
.00 2 1 10 54. 5,
acrytic board lighted by an incandescen
__7 I10 2 -1--- 0_11
1 2 1 a I so. 11,
lamp from the rear. As a r e s u 1 t t 1, i r _. 3,0 2 1 .... ...
40 3 1 0 1 48. 6,
50 3 1 3.0 44. 6,
relation was proportional. _I5._Q_0 _3 -0 1- 4. 1- -4 2. 6' 0 to 20 30 K- P-1fk 0C...
10 3 TO 5 _- -
20 3 (a)
_L_ _LL 38.-!'--
4 2 __0 I_� 7_ _36. 5-
4. OUTLINE OF MEASUREMENT 41 3 2 1 24. 1 34. 5, Funabashi-Shi
11 1 1 1 23.8 3:. S, Tokyo L8:30. . hrashi-Shi
11A* 3 1 0 23. 5 30. 3" 19:00.
We measured at a breakwater 5 meters 10 3 2 _0 24. 0 29. 4* 17:00
16:00 Hakuhari
24. 0 26. 4, 15:00
above the sea level in the seashore of 30 a 2 1 24. 1 24. 2, Chiba-Shi
48 3 1 3 1 34. S 22. 4'
so 3 1 3 j 3" 20. r- 1_@,
Makuhari Beach, Narasino-City, Chiba I I "1 1 3 1 3 Tokyo-Bay
10 3 __2_ 1 3S. a 16. V
20 3 3 1 3
Pref., Japan on May 18th, 1988. so 3 3 1 3S. 2 12. 3*
40 3 3 1 34. 6 10. 4 .
We recorded the luminance for one minute so 3 2 1 35. 0 a. 5*
18: 00 a 3 1 34. a 5. 5, Sodegaura
to 3 3 1 1 34. 7 1.
every ten minutes from-14:00 to 18:30. - . 4.-
1 201 3 1 36 2.8
1 30 13 1 1 3 1 $.a 0. 9* (b)
Fig.7(a) and (b) show the points where we
Fig. 7 Measurement Point
IX"
A
c

v
0

measured.

.705

to catch the variation of the reflected o Sea Surface Luminance
light of sea surface at the very center of A Sky Light Luminance

the photo. The image of the ref lected
Z-
a ..
light was recorded in video tapes. We A-AA, 0
U
brought them back to our labolatory and cu l
u

analyzed t h e i m a g e w i t h a personal

computer. 4
10

While some of us o'perating the video
camera, other scanned and obtained the 10
60* 40' 20* 0.
data of the maximum luminance of the sea
Sun Elevation
surface 5 degrees below horizon and the Fig. 8 Variation of Sea Surface Luminance
sky 5 degrees above horizon. and Sky Light Luminance

u o e
a 6%
M o Makuhari o
o
6. RESULT AND CONSIDERATION
a Jagashima o
We observed 2 trends from the measured o o
o

data. One is reflectance of the sea

surface by luminance meter. The others w o
r
o
No
i s luminance distribution of horizontal oo
coo
z oo
and vertical on the sea surface. o o
is variation area of sun glitter on the 60* 40* 20* 0.
Sun Elevation (*)
sea surface. Fig.9 Variation of Sea Surface
Reflectance

(1) VARIATION OF REFLECTANCE Fig.9 we assign the value 1 for the

Fig.8 indicates variation of sea surface maximum none dimentional reflectance in

and skylight luminance according to sun Makuhari when the sun elevation is 18.3

elevation and fig.9 indicates variation of degrees. The calculation for fig.9 was

sea surface reflectance with variable sun done under the assumption of observing
elevation in Makuhari Beach and J6gashima smoothsea surface.
in Kanagawa Pref( Oct.25th, 1987. It was From 54.5 degrees to 40.5 degrees sun
clear, 0 cloud amount, and 2 Beaufort Wind e I e v a t i o n , t h e n o n e dimensional
Scale)." reflectance increased gently because of

The value of sea surface reflectance is the increase of the wave angle that makes

obtained from a division of maximum regular ref lection to the measurement

luminance of sea surface by theoretical point.
value of horizontal luminance of sunlight. The reflectance rapidly increased after

706

36.5 degrees sun elevation -and reached the 11.61 11.6'
Vertical Line@ 9;
maximum at 18.3 degrees and started

Horizontal 3;
to decrease suddenly after 14.3 degreees. I Line
_10' 3' 0 16.
3' -3-
As a result, we discovered that the 3'
-6
decrease reflectance at 36.5 degrees was -9.
-12*
caused by the increasing amount of cloud
(a) b
at the time of observation. Fig. 10 Horizontal Line and Vertical Line

As a reference, the increased scale of

cloud amount at 36.5 degrees on table.2
6.5*
and the decreased skylightluminance at Sun Elevation,[11 18.3 12.3*
same degrees on fig.8 can be used. 24.3*
30.3'
The maximum reflectance at sun elevation 36.5'

42.6*
o f 18.3 degrees was caused by the 8- 48.6*
6-
occurance of the strongest r e g u 1 a r 4-
2.
0
reflection toward the measurement point of 5' 0. 5. 7.
r
5 meters above the sea level. Scanning Angle

The sudden decrease of the reflectance of
Fig. 11 Luminance Distribution on Horizontal Line
fig.9 is a resemblance of the sea surface

reflectance of C.Cox and W.H.Munk in 4
6.5'
Beaufort Wind Scale. 12. 3* 1
18. 3*
Cox and Munk 5) discovered t h a t t h i s Sun ElevaItion P1 24.3'
30.3'
decrease was caused by the wave angle; a
36. 5'
42.6*
therefore, we assumed the same reason for lo-
8- 48.6'
this decrease. 6-
4-
U 2- Sky Horizon Sea
The maximum sea surface reflectance in o
6' 3' 0' -3o -6o -9o
J6gashima is 14.3 degrees sun elevation. Scanning Angle

But we could get the similar result from
@Fig. 12 Luminance Distribution on Vertical Line
these two different measurement.
degrees below horizon and the vertical

(2) VARIATION OF LUMINANCE line appears at the center of the image,

DISTRIBUTION and fig.10(b) indicates the co-ordinates

In order to obtain the graphical data for on the horizontal and vertical line.

luminance distribution, we established two We processed the image into the graphic

lines on the i m a g e as indicated in data recorded of every hour from 14:00.

fig.10(a) the horizontal line appears at 3 Fig.11 indecates the horizontal line,
Elaat.-n[-]
18
24.3 .1
2

3
n' @@l 2.
24.

3
36.A5
6

303
_3 .5
2 6

486

SyKorionSea

707

fig.12 indicates the vertical line.. and started to decrease suddenly after

In both figures, the horizontal axis 14.3 degrees.

indicates scanning angle of camera, the 2) At 30.3 degrees sun elevation, the high

vertical one indicates luminance, and the luminance surface grew and expanded to 10

other one is sun elevation. degrees of scanning angle on the luminance

At 54.5 degrees sun elevation, t h e distribution of the horizontal line.

luminance is constant and sea surface 3) At 18.3 degrees, high luminance surface

luminance is s m a 11 e r than s k y 1 i g h t m o v e s onshore on t h e luminance

luminance. distribution of the vertical line.

At 42.6 degrees sun elevation, high

luminance points increased a 1 i t t 1 e

because of the increased amount.of wave

angle which can reflect the sunlight to REFERENCES
the measurement point for the maximum. I)ARCHITECTURAL INSTITUTE OF J A P A N
At 36.5 degrees sun e 1 e v a t i o n , s e a Architectural D e s i g n D a t a B o o k 1
surface luminance is bigger than skylight (Environmental Section). Maruzen, 1985.
luminance in the figure of vertical line. 2 ) M a s a o I n u i a n d Y o s h i k i Nakamura:
And at 30.3 degrees the high luminance Luminance Distribution in Natural and
surface, in the figure of horizontal line, A r t i f i c i a 1 Landscapes. Journal of
which is more than 4.0 X 10' [cd/m' I grew Architecture, Planning Environmental

and expanded to 10 degrees ofscanning Engineering. No.384, p36 - 43, Feb, 1987.
angle. 3)Motoaki Kishino: Some Ploblems of Light

At 18.3 degrees, high luminance surf ace near Sea Surface. Marine Science/Monthly.

moves onshore in the fig.12 and luminance Vol.4, No.9, p 11 - 16, Sep, 1972.

is maximum value for both fig.11 and 12. 4)Maki Enomoto: A Basic Studies on

The width of high luminance surface is Measurement of Optical Environment in

reduced to a half of the high luminance Coas@al and Offshore Structure. Graduation

surface. At 6.5 degrees, luminance is Thesis of Dept of Oceanic Architecture &

suddenly decreased and the width of high Engineering C o 11 e g e of S c i e n c e &

luminance surface.is also reduced. Technology, Nihon University, Feb, 1986.

7. CONCLUTION 5)Cbx and Munk: Statistics of the sea
1) The v a r i a t i o n of reflectance is surface derived from sun glittre. Journal
increased after 36.5 degrees sun elevation of Marine Reseach, 13-2, p198 227, 1958.

and reached the maximum at 18.3 degrees

708

USE OF A MACRO-HYBRID CAMERA AT NATIONAL GEOGRAPHIC

EMORY KRISTOF, JOSEPH STANCAMPLANO & ALVIN CHANDLER

NATIONAL GEOGRAPHIC SOCIETY

ABSTRACT The most important decision was which SLR to
chop up. For reasons of size, retained structural
Still cameras and television cameras are used integrity after alterations, and its mechanical
separately and in concert to produce underwater shutter, we chose a Nikon FM2. We retained the
images. It is generally accepted that pictures shutter, the mirror box, the penta-prism, and the
done on film still cameras are sharper and have lens mount. What was left established the inter-
higher resolution that those produced with video nal diameter of the housing which is five inches.
cameras. The video cameras are capable of much The macro-lens is a Nikon auto-focus 55mm. The
greater sensitivity to light and most importantly auto-focus feature is not used, but the internal
feed back information at a real-time rate. focusing of the lens keeps the changes to the
length of the lens to a minimum. Since the lens
There are several commercially available is looking through flat glass the one third magni-
cameras that have the film and the electronic fication rule in water comes into play and the
imager sharing the same optics. This is usually lens is equal to a 73mm lens in air.
accomplished by use of a mirror that moves back
and forth between the film and imager. Cameras The remains of the Nikon camera body and the
have been built that have a video camera looking lens were mated to a Benthos 378 camera chassis
through the optics of a 35mm SLR. The Geographic and film drive. The film is handled normally in a
designed and built its own hybrid camera along Benthos 50 foot 400 exposure cassette. A Sony
these lines to accomplish high quality remotely XC-77 CCD camera with a 16mm relay lens focuses
done macro-photography from submersibles and ROVs. on the groundglass through the Nikon eyepiece.
The camera has been successfully used in the field The Sony sees about two-thirds of the total 35mm
from both types of platforms. The paper wi 11 be frame. The focusing is accomplished remotely with
both on the design and construction of our camera a motor. The first edition of the camera didn't
and the techniques employed during its use. have remote aperture setting. Color Negative
film was relied upon to handle differences in
exposure.

Some of the most striking and informative For its initial use on the Cayman Wall , the
underwater still pictures produced by scuba diving camera was first mounted on the RSL Perry PC 1802.
photographers, such as David Doubilet, have been A pair of 300 watt second strobes provided illumi-
made with macro-lenses and SLR cameras in under- nation. The aperture was preset at f1l and the
water housings. Most of the deep water (beyond film used was Kodak Vericolor 160 ASA. When
scuba depth) still photographs produced for placed on a DOE Phantom a single 100 watt second
National Geographic are from cameras mounted strobe and a setting of f5.6 were used.
externally on submersibles and ROVS. These are
special 35mm cameras, such as the Benthos, with Shooting pictures from the submersible was
wide angle optics placed in tube-shaped deep sea easier than from the ROV, as could be expected.
housings. Focus and aperture settings have to be The biggest problems with shooting from the ROV
preset. In most cases an independent video camera are coordinating the wide angle television view
with the same angle of view is used as a with the magnified macro picture, and the magnifi-
11viewfinder." This is a simple approach to pho- cation of the ROV's motion by the macro camera.
tography that is adequate for many subjects. The picture goes in and out focus very quickly
if the ROV is being flown in mid-water and is not
Macro -ph otogr aphy is best done with a SLR. sitting on something or otherwise braced. I found
The design objectives for our camera were to take it was best to pre-focus the camera and when the
the best features of a good SLR and macro-lens, ROV swam into the proper focus to shoot it--fast.
combine them with the extended film capacity and
drive of an existing deep sea camera, install a The mod II camera has a remotely adjustable
small TV camera with a relay lens to look thru the aperture control and a second slaved strobe light
shooti'ng lens, and package the results in as small that can be placed in the manipulator claw. This
a pressure-proof housing as possible. Since the makes possible a more controlled and moveable
camera was to be used part of the time on our ROV light. These improvements are to be tried out
the small size was especially critical. this summer in the Monterey Canyon. The camera

CH2585-8188/0000- 709 $1 @1988 IEEE

has already proved to be a success for the
National Geographic and the first picture will be
published in the November 1988 Magazine. Other
scientists and industrial users might find this
type of special camera to be a good tool.
40,

Stalked crinoid at 750 feet (photo from sub).

710

UNDERWATER HYBRID 35mm/TV CAMERA
U NIKON FE-@ HIRROR BOX
CCD CAMERA

35irn, 2.8 AF NIKKOR

SHUTM
BENTHOS 400 ET CASSET SPEED CONT.

APE

SONY XC37 VIDEO CX-fM POWER SUPPLY

Diagram of macro-hybrid camera.

NOW *0 4MM

7
wFFM 'k

Macro-hybrid camera mounted on RSL Perry 1802, overtop a pair of video cameras providing stereo.
Emory is pictured
Author, Kristof through bubble.

711

(Above): Brain coral in 200 feet of water and (below) close-up of brain coral (both photos from ROV).

712

3-D AS AN UNDERWATER TObL

EMORY KRISTOF & AL;VIN CHANDLER, DR. WILLIAM HAMNER

NATIONAL GEOGRAPHIC SOCIETY, UCLA

The National Geographic Society has used its video system, the number of fields per second is
first set of 3-D video cameras to produce tapes of doubled. We achieve this by doubling the ver-
deep sea creatures from down further that a mile. tical scanning rate, thereby producing 120 fields
We are presently building a second miniaturized instead of 60 fields per second. Thus the number
set of cameras to be mounted on ROVs. The tapes of lines per field is halved, but each eye sees a
are being introduced to the scientific community proper twofold interlaced image. When displayed
and used by the Geographic in educational on a suitably prepared 120 HZ monitor, the two
displays. The StereoGraphic system that we employ images are displayed in sequence."
will display both video or computer CAD material.
We are using this second ability to produce 3-D One of the real advantages of this system is
video stereo of a wire model of a shipwreck sur- that the signal can be recorded on an unmodified
veyed in situ with a SHAPPS system. We are NTSC VCR. A major improvement was the introduc-
looking at taking this data and producing a tion of LCD technology to replace the PLZT
hologram. The StereoGraphic system is the first to goggles. This made possible lower cost, lower
allow the user to electronically set the point of voltage viewing glasses with much high.er light
convergence between the two cameras. This means transmission rates. Large LCD polarized switching
the system can be used like a rangefinder to plates are now available to place in front of
instantly establish the distance to an object and monitors and video projectors. This allows
then its dimensions. The Geographic is working to multiple viewers to experience the 3-D effect
computerize this ability. William Hamner has had wearing only inexpensive polarized glasses.
a 3-D motion analysis program created for a Sun
work station. He is now able to pull off speed The Universal Camera Controller or UCC will
and vector information from 3-D video tape. 3-D process the video from any pair of genlockable
has many applications in the underwater world and cameras. A feature of the UCC I found really use-
it is a technology whose time has finally arrived. ful was the ability to electronically align and
The Geographic is in the forefront of its uses. converge the cameras at any time. Where the con-
vergence point is placed in a 3-0 image can be
very important to the quality and the perception
The National Geographic Society has been of that image. This feature also allows the
experimenting with 3-D video for five years. The system to function like an optical rangefinder,
initial reason was to better control manipulator making possible instant depth and size measure-
functions on its ROV. For this purpose we ments with a coupled PC.
obtained a mirror-stereopticon dual TV viewing
hood from ARD Corp. This granted the operator of The Geographic was impressed enough with the
the ROV real-time 3-0 viewing from a pair of JVC StereoGraphic system to purchase it. The initial
nine inch monitors. We were experimenting with use was to produce first-time 3-D video of deep
the hood and different pairs of cameras in the sea animals from depths down to greater than a
lab, when the decision was made to pursue a system mile. We have been baiting these animals into
that would allow us to record and present stereo manned submersibles during a multi-year study
video material to the public. known as the BEEBE PROJECT. In the summer of 1988
the StereoGraphic UCC, the NTSC Display
Over the last decade we followed the develop- Controller, a 13" 120 HZ monitor, two LCD shut-
ment of time-multiplexed stereoscopic video tered glasses and a Sony BVW-25 Betacam VCR were
systems utilizing PLZT ceramic goggles. Typified installed in the International Underwater
by the Honeywell system, video fields alternated Contractor submersible PISCES VI. Mounted exter-
between "left eye" and "right eye" information. nally on a pan and tilt were a pair of Sony single
The resulting picture flickered because there were chip CCD cameras with 4.8mm lenses. The 4,500 psi
only 30 fields per second per eye. The picture rated camera housings were built by Deep Sea Power
could only be viewed by looking at the monitor and Light.
through an expensive pair of PLZT goggles that
alternately shuttered each eye. The location for the dives was eight miles
N.W. of Bermuda. Many animals, including dif-
The further development of the technology was ferent species of deep water sharks, were seen for
undertaken by Lenny Lipton and Lhary Meyer of the the first time in their natural habitat at the
StereoGraphics Corp. "In the StereoDimensional feeding stations. The Project scientists all

CH2585-8/88/0000. 713 $1 @1988 IEEE

like the 3-D video tapes and felt they were reaso- of the StereoGraphic image. You feel that you are
nable simulations of the actual diving experience. looking with your own eyes through a face mask.
They are now looking at ways to pull additional This is the beginning of true visual telepresence.
data from the tapes. Animal counts will be easier
to perform from 3-D video tapes and determination To retain the quality of the original, we
of sizes will be possible. have been recording the video on two time code
linked Sony BVW-35 Betacam recorders. The
The Geographic decided to build a 3-D theater finished show will be presented by two computer
for our museum to show the results of the Project. controlled laser-disc players. Chris has built
It was patterned to look like an underwater habi- electronic convergence control into his viewing
tat, with five viewing "portholes." Each system. He has demonstrated the ability to use
"porthole" contained a 13" 1000 x 1000 graphic this function as a rangefinder. Coupled with a
monitor set in 2011 from the glass of the video measuring system he has sized objects remo-
"porthole." The glass is a special 5 x 7 inch LCO tely. A protype system demonstrating a completely
shutter supplied by StereoGraphics. The viewer integrated 3-D video system including cameras,
looks through the port and sees the three minute viewers, recorders, and measuring instrumentation
show in 3-D without having to wear special will be shown at OCEANS 88.
glasses. The video tape is converted to RGB and
line-doubled to make up for the reduced vertical After establishing the distance and size of a
resolution. We have been very pleased with the desired object, all that is left to measure is the
performance of the equipment which has been speed and vector of the object if it is relevant.
trouble free since we turned it on in January. Dr. William Hamner had software created for a Sun
The public has liked the show. work station by Motion Analysis of Santa Rosa,
California that allows him to take that infor-
Presently we are using a pair of finger sized mation from his dual camera 3-D system. This was
Panasonic TV cameras mounted behind the dome of a created as part of Bill"s NSF funded study of the
W II MiniRover by Chris Nicholson of Deep Sea schooling of fish.
Systems. This work in Monterey Canyon is our
first blue water trial of 3-D video on a ROV. The Geographic has plans to do a high resolu-
Chris has built a dual nine inch monitor viewing tion acoustic survey of a shipwreck that we can
system that uses a half-silvered mirror, present as 3-D computer graphic information in our
polarizing filters on both monitors and polarized museum theater. 3-D video has many applications
glasses. This is an old idea, but as optimized by in the underwater world, and it is a technology
Chris, and taking advantage of the new small CCD whose time has finally arrived. The Geographic is
cameras the results have been impressive. The in the forefront of its uses.
resolution of the image is more than double that

Stereo pair of deep
sea video cameras
Fit,, wow"
mounted on the
front of PISCES VI, a
35mm still camera
is mounted
underneath.

714

@_' @73
R_@

R
g 4i)F".

KF;

A-1-4 -,0-4f

@M@@ _4

Ce"

(Top): Emory Kristof operating 3-D video system a mile deep in the Atlantic inside PISCES VI. Viewing
is through LCD shuttered glasses. (Bottom'left): Rack mounted StereoGraphic system is seen over Dr.
Eugenie Clark inside of PISCES VI. (Bottom right): Scenes from the 3-D show include two types of deep
sea sharks and several large crabs at the bait cage at a depth of 3,000 feet.

715

0 @-Q -2

0
0

C' 0

0 00

0 0 0

coo Q 0 0-0

.10 0 0 0r

0 0 00
0 0.0

7`4
P
PL
L

*.,,>, r-I
> r

tAq"-

1: 'n

Drawing of exterior of Geographic 3-D theater (top), and drawing of interior of Geographic 3-D
theater (bottom).

716

rv
L

q I

A-

Ci'NTTZ

National Geographic 3-D System
on IUC PISCES VI

"41

Universal Camara Controller

SONY BVW-25
Betacarn VCR
(unmodified)

4.8inm lenibs
Cut-away of Geographic 3-D theater (top). Many
eels eating the bait at a depth of 5,100 feet
NISC Display Controller
(right).

SONY CCD
Genlockable cameras
(unmodified)

Sync Generator 13" 120 HZ Monitor

Viewingglasw

717

@eP
Iff,

Val
#f
A
ef
I'M
14 J,
46

W"N
iL illi4
IT

1#44-

tv
F,

Emory Kristof
graphic 3 inside Ceo-
-D theater viewing
,show thr. LCD shuttered
.P.rthoj_e,@gh
in front (top) and Emory
Of Of Geographic
37D theater (right).

718

DATE DUE

3 6668 14106 1152