[Federal Register Volume 83, Number 121 (Friday, June 22, 2018)]
[Proposed Rules]
[Pages 29212-29310]
From the Federal Register Online via the Government Publishing Office [www.gpo.gov]
[FR Doc No: 2018-12906]
[[Page 29211]]
Vol. 83
Friday,
No. 121
June 22, 2018
Part III
Department of Commerce
-----------------------------------------------------------------------
National Oceanic and Atmospheric Administration
-----------------------------------------------------------------------
50 CFR Part 217
Taking and Importing Marine Mammals; Taking Marine Mammals Incidental
to Geophysical Surveys Related to Oil and Gas Activities in the Gulf of
Mexico; Proposed Rule
Federal Register / Vol. 83 , No. 121 / Friday, June 22, 2018 /
Proposed Rules
[[Page 29212]]
-----------------------------------------------------------------------
DEPARTMENT OF COMMERCE
National Oceanic and Atmospheric Administration
50 CFR Part 217
[Docket No. 110811494-7925-01]
RIN 0648-BB38
Taking and Importing Marine Mammals; Taking Marine Mammals
Incidental to Geophysical Surveys Related to Oil and Gas Activities in
the Gulf of Mexico
AGENCY: National Marine Fisheries Service (NMFS), National Oceanic and
Atmospheric Administration (NOAA), Commerce.
ACTION: Proposed rule; request for comments.
-----------------------------------------------------------------------
SUMMARY: NMFS has received a petition for an incidental take regulation
(ITR) from the Bureau of Ocean Energy Management (BOEM). The requested
ITR would govern the authorization of take of small numbers of marine
mammals over the course of five years incidental to geophysical survey
activities conducted by industry operators in Federal waters of the
U.S. Gulf of Mexico (GOM). BOEM submitted the petition in support of
oil and gas industry operators, who would conduct the activities. A
final ITR would allow for the issuance of Letters of Authorization
(LOA) to the aforementioned industry operators over a five-year period.
As required by the Marine Mammal Protection Act (MMPA), NMFS requests
comments on its proposed rule, including the following; the proposed
regulations, several alternatives to the proposed regulations described
in the ``Proposed Mitigation'' and ``Alternatives for Consideration''
sections of the preamble, two baselines against which to evaluate the
incremental economic impacts of the proposed regulations (addressed in
the ``Economic Baseline'' section), and, two sections with broader
implications: A clarification of NMFS's interpretation and application
of the ``small numbers'' standard (see the ``Small Numbers'' section of
the preamble); and an alternative method for assessing Level B
harassment from exposure to anthropogenic noise (see the ``Estimated
Take'' section of the preamble).
DATES: Comments and information must be received no later than August
21, 2018.
ADDRESSES: You may submit comments on this document, identified by
NOAA-NMFS-2018-0043, by any of the following methods:
Electronic submission: Submit all electronic public
comments via the Federal e-Rulemaking Portal. Go to
www.regulations.gov/#!docketDetail;D=NOAA-NMFS-2018-0043, click the
``Comment Now!'' icon, complete the required fields, and enter or
attach your comments.
Mail: Submit written comments to Jolie Harrison, Chief,
Permits and Conservation Division, Office of Protected Resources,
National Marine Fisheries Service, 1315 East West Highway, Silver
Spring, MD 20910.
Comments regarding any aspect of the collection of information
requirement contained in this proposed rule should be sent to NMFS via
one of the means provided here and to the Office of Information and
Regulatory Affairs, NEOB-10202, Office of Management and Budget, Attn:
Desk Officer, Washington, DC 20503, [email protected].
Instructions: Comments sent by any other method, to any other
address or individual, or received after the end of the comment period,
may not be considered by NMFS. All comments received are a part of the
public record and will generally be posted for public viewing on
www.regulations.gov without change. All personal identifying
information (e.g., name, address), confidential business information,
or otherwise sensitive information submitted voluntarily by the sender
will be publicly accessible. NMFS will accept anonymous comments (enter
``N/A'' in the required fields if you wish to remain anonymous).
Attachments to electronic comments will be accepted in Microsoft Word,
Excel, or Adobe PDF file formats only.
FOR FURTHER INFORMATION CONTACT: Ben Laws, Office of Protected
Resources, NMFS, (301) 427-8401. Electronic copies of the application
and supporting documents, as well as a list of the references cited in
this document, may be obtained online at: www.fisheries.noaa.gov/national/marine-mammal-protection/incidental-take-authorizations-oil-and-gas. In case of problems accessing these documents, please call the
contact listed above.
SUPPLEMENTARY INFORMATION:
Purpose and Need for Regulatory Action
This proposed rule would establish a framework under the authority
of the MMPA (16 U.S.C. 1361 et seq.) to allow for the authorization of
take of marine mammals incidental to the conduct of geophysical survey
activities in the GOM. We received a petition from BOEM requesting the
five-year regulations. Subsequent LOAs would be requested by industry
operators. Take would occur by Level A and/or Level B harassment
incidental to use of active acoustic sound sources. Please see the
``Background'' section below for definitions of harassment.
Legal Authority for the Proposed Action
Section 101(a)(5)(A) of the MMPA (16 U.S.C. 1371(a)(5)(A)) directs
the Secretary of Commerce to allow, upon request, the incidental, but
not intentional taking of small numbers of marine mammals by U.S.
citizens who engage in a specified activity (other than commercial
fishing) within a specified geographical region for up to five years
if, after notice and public comment, the agency makes certain findings
and issues regulations that set forth permissible methods of taking
pursuant to that activity and other means of effecting the ``least
practicable adverse impact'' on the affected species or stocks and
their habitat (see the discussion below in the ``Proposed Mitigation''
section), as well as monitoring and reporting requirements. Section
101(a)(5)(A) of the MMPA and the implementing regulations at 50 CFR
part 216, subpart I provide the legal basis for issuing this proposed
rule containing five-year regulations, and for any subsequent LOAs. As
directed by this legal authority, this proposed rule contains
mitigation, monitoring, and reporting requirements.
Summary of Major Provisions Within the Proposed Rule
Following is a summary of the major provisions of this proposed
rule regarding geophysical survey activities. These measures include:
Standard detection-based mitigation measures, including
use of visual and acoustic observation to detect marine mammals and
shut down acoustic sources in certain circumstances;
Time-area restrictions designed to avoid effects to
certain species of marine mammals in times and/or places believed to be
of greatest importance;
Vessel strike avoidance measures; and
Monitoring and reporting requirements.
Background
Section 101(a)(5)(A) of the MMPA (16 U.S.C. 1361 et seq.) directs
the Secretary of Commerce (as delegated to NMFS) to allow, upon
request, the incidental, but not intentional, taking of small numbers
of marine mammals by U.S. citizens who engage in a specified activity
(other
[[Page 29213]]
than commercial fishing) within a specified geographical region if
certain findings are made, regulations are issued, and notice is
provided to the public.
An authorization for incidental takings shall be granted if NMFS
finds that the taking will have a negligible impact on the species or
stock(s), will not have an unmitigable adverse impact on the
availability of the species or stock(s) for subsistence uses (where
relevant), and if the permissible methods of taking and requirements
pertaining to the mitigation, monitoring and reporting of such takings
are set forth.
NMFS has defined ``negligible impact'' in 50 CFR 216.103 as an
impact resulting from the specified activity that cannot be reasonably
expected to, and is not reasonably likely to, adversely affect the
species or stock through effects on annual rates of recruitment or
survival.
The MMPA states that the term ``take'' means to harass, hunt,
capture, or kill, or attempt to harass, hunt, capture, or kill any
marine mammal.
Except with respect to certain activities not pertinent here, the
MMPA defines ``harassment'' as: Any act of pursuit, torment, or
annoyance which (i) has the potential to injure a marine mammal or
marine mammal stock in the wild (Level A harassment); or (ii) has the
potential to disturb a marine mammal or marine mammal stock in the wild
by causing disruption of behavioral patterns, including, but not
limited to, migration, breathing, nursing, breeding, feeding, or
sheltering (Level B harassment).
National Environmental Policy Act
To comply with the National Environmental Policy Act of 1969 (NEPA;
42 U.S.C. 4321 et seq.) and NOAA Administrative Order (NAO) 216-6A,
NMFS must evaluate the proposed action (i.e., the promulgation of
regulations and subsequent issuance of incidental take authorizations)
and alternatives with respect to potential impacts on the human
environment.
In August 2017, BOEM produced a final Programmatic Environmental
Impact Statement (PEIS) to evaluate potential significant environmental
effects of geological and geophysical (G&G) activities on the Outer
Continental Shelf (OCS) of the GOM, pursuant to requirements of NEPA.
These activities include geophysical surveys in support of hydrocarbon
exploration and development, as are described in the petition for ITR
before NMFS. The PEIS is available online at: www.boem.gov/Gulf-of-Mexico-Geological-and-Geophysical-Activities-Programmatic-EIS/. NMFS
participated in development of the PEIS as a cooperating agency and
believes it is appropriate to adopt the analysis in order to assess the
impacts to the human environment of issuance of the subject ITR and any
subsequent LOAs. Information in the petition, BOEM's PEIS, and this
document collectively provide the environmental information related to
proposed issuance of this ITR for public review and comment.
Summary of Request
BOEM was formerly known as the Minerals Management Service (MMS)
and, later, the Bureau of Ocean Energy Management, Regulation, and
Enforcement (BOEMRE). On December 20, 2002, MMS petitioned NMFS for
rulemaking under Section 101(a)(5)(A) of the MMPA to authorize take of
sperm whales (Physeter macrocephalus) incidental to conducting
geophysical surveys during hydrocarbon exploration and development
activities in the GOM. On March 3, 2003, NMFS published a notice of
receipt of MMS's application and requested comments and information
from the public (68 FR 9991). MMS subsequently submitted a revised
petition on September 30, 2004, to include a request for incidental
take authorization of additional species of marine mammals. On April
18, 2011, BOEMRE submitted a revision to the petition, which
incorporated updated information and analyses. NMFS published a notice
of receipt of this revised petition on June 14, 2011 (76 FR 34656). In
order to incorporate the best available information, BOEM submitted
another revision to the petition on March 28, 2016, which was followed
on October 17, 2016, by a revised version that was deemed adequate and
complete based on NMFS's implementing regulations at 50 CFR 216.104. In
the interim period, BOEM, with NMFS representing NOAA as a cooperating
agency, prepared a PEIS for the GOM OCS Proposed G&G Activities.
On December 8, 2016 (81 FR 88664), we published a notice of receipt
of the petition in the Federal Register, requesting comments and
information related to the request. This 30-day comment period was
extended to January 23, 2017 (81 FR 92788), for a total review period
of 45 days. The comments and information received during this public
review period informed development of the proposed ITR discussed in
this document, and all comments received are available online at
www.fisheries.noaa.gov/national/marine-mammal-protection/incidental-take-authorizations-oil-and-gas.
Geophysical surveys are conducted in support of hydrocarbon
exploration and development in the GOM, typically by companies that
provide such services to the oil and gas industry. Broadly, these
surveys include (1) deep penetration surveys using large airgun arrays
as the acoustic source, (2) shallow penetration surveys using a small
airgun array, single airgun, or subbottom profiler as the acoustic
source, and (3) high-resolution surveys, which may use a variety of
acoustic sources. Generally speaking, these surveys may occur within
Federal territorial waters and waters of the U.S. Exclusive Economic
Zone (EEZ) (i.e., to 200 nautical miles (nmi)) within the GOM, and
corresponding with BOEM's Western, Central, and Eastern GOM OCS
planning areas. The use of these acoustic sources is expected to
produce underwater sound at levels that have the potential to result in
harassment of marine mammals. Cetacean species with the potential to be
present in the GOM are described below.
This proposed rule would establish a framework under the authority
of the MMPA (16 U.S.C. 1361 et seq.) and NMFS's implementing
regulations (50 CFR 216.101 et seq.) to allow for the authorization,
through LOAs, of take of marine mammals incidental to the conduct of
geophysical surveys for oil and gas activities in the GOM. The
requested regulations would be valid for five years.
Description of the Specified Activity
Overview
The specified activity consists of geophysical surveys conducted by
industry operators for a variety of reasons related to hydrocarbon
exploration, development, and production. These operators are typically
companies that provide geophysical services, such as data acquisition
and processing, to the oil and gas industry, including exploration and
production companies. The petition describes a five-year period of
geophysical survey activity and provides estimates of the amount of
effort by survey type and location. BOEM's PEIS (BOEM, 2017) describes
a range of potential survey effort. The levels of effort in the
petition (which form the basis for the modeling effort described later
in the ``Estimated Take'' section) are the high-end estimates. Actual
total amounts of effort by survey type and location would not be known
in advance of receiving LOA requests from industry operators.
Geophysical surveys are conducted to obtain information on marine
seabed
[[Page 29214]]
and subsurface geology for a variety of reasons, including to: (1)
Obtain data for hydrocarbon and mineral exploration and production; (2)
aid in siting of oil and gas structures, facilities, and pipelines; (3)
identify possible seafloor or shallow depth geologic hazards; and (4)
locate potential archaeological resources and benthic habitats that
should be avoided. In addition, geophysical survey data inform Federal
government decisions. For example, BOEM uses such data for resource
estimation and bid evaluation to ensure that the government receives a
fair market value for OCS leases, as well as to help to evaluate worst-
case discharge for potential oil-spill analysis and to evaluate sites
for potential hazards prior to drilling.
Deep penetration seismic surveys using airgun arrays as an acoustic
source (sound sources are described in the ``Detailed Description of
Activities'' section) are a primary method of obtaining geophysical
data used to characterize subsurface structure. These surveys are
designed to illuminate deeper subsurface structures and formations that
may be of economic interest as a reservoir for oil and gas
exploitation. A deep penetration survey uses an acoustic source suited
to provide data on geological formations that may be thousands of
meters (m) beneath the seafloor, as compared with a shallow penetration
or high resolution geophysical (HRG) survey that may be intended to
evaluate shallow subsurface formations or the seafloor itself (e.g.,
for hazards).
Deep penetration surveys may be two-dimensional (2D) or three-
dimensional (3D) (see Figure 1-2 of the petition), and there are a
variety of survey methodologies designed to provide the specific data
of interest. 2D surveys are designed to acquire data over large areas
(thousands of square miles) in order to screen for potential
hydrocarbon prospectivity, and provide a cross-sectional image of the
structure. In contrast, 3D surveys may use similar acoustic sources but
are designed to cover smaller areas with greater resolution (e.g., with
closer survey line spacing), providing a volumetric image of underlying
geological structures. Repeated 3D surveys are referred to as four-
dimensional (4D), or time-lapse, surveys that assess the depletion of a
reservoir.
Shallow penetration and high-resolution surveys are designed to
highlight seabed and near-surface potential obstructions, archaeology,
and geohazards that may have safety implications during rig
installation or well and development facility siting. Shallow
penetration surveys may use a small airgun array, single airgun, or
subbottom profiler, while high-resolution surveys (which are limited to
imaging the seafloor itself) may use single or multibeam echosounders
or side-scan sonars.
Dates and Duration
The specified activities may occur at any time during the five-year
period of validity of the proposed regulations. Actual dates and
duration of individual surveys are not known. Survey activities are
generally 24-hour operations. However, BOEM estimates that a typical
seismic survey experiences approximately 20 to 30 percent of non-
operational downtime due to a variety of factors, including technical
or mechanical problems, standby for weather or other interferences, and
implementation of mitigation measures.
Specified Geographical Region
The proposed survey activities would occur off the Gulf of Mexico
coast of the United States, within BOEM's Western, Central, and Eastern
GOM OCS planning areas (approximately within the U.S. EEZ; Figure 1).
U.S. waters of the GOM include only the northern GOM. BOEM manages
development of U.S. Federal OCS energy and mineral resources within OCS
regions, which are divided into planning areas. Within planning areas
are lease blocks, on which specific production activities may occur.
Geophysical survey activities may occur on scales ranging from entire
planning areas to multiple or specific lease blocks, or could occur at
specific potential or existing facilities within a lease block.
In addition to general knowledge and other citations contained
herein, this section relies upon the descriptions found in Sherman and
Hempel (2009), Wilkinson et al. (2009), and BOEM (2017).
The GOM is a deep marginal sea--the largest semi-enclosed coastal
sea of the western Atlantic--bordered by Cuba, Mexico, and the United
States and encompassing more than 1.5 million square kilometers
(km\2\). The GOM is distinctive in physical oceanography and freshwater
influx, with major, persistent currents and a high nutrient load.
Oceanic water enters from the Yucatan Channel and exits through the
Straits of Florida, creating the Loop Current. The Loop Current--the
GOM's most dominant oceanographic feature--flows clockwise between Cuba
and the Yucatan Peninsula, Mexico, and circulates into the eastern GOM
before exiting as the Florida Current, where it ultimately joins the
Gulf Stream in the Atlantic. Small-scale, ephemeral currents known as
eddies form off the Loop Current and may enter the western GOM. The
eastern edge of the Loop Current interacts with the shallow shelf to
create zones of upwelling and onshore currents--nutrient-rich events
promoting high phytoplankton growth and supporting high productivity.
The distribution of plankton in the deeper waters of the GOM,
especially the northern and eastern parts of the Gulf, is controlled by
the Loop Current (Mullin and Fulling, 2004). The temporal movement of
all organisms, including marine mammals and their prey, may be affected
by upwelling of nutrient rich cold water eddies (Davis et al., 2002).
However, habitat use appears to be more directly correlated with static
features such as water depth, bottom gradient, and longitude (Mullin
and Fulling, 2004). Temporal fluctuation near the surface can cause
changes in diurnal movement patterns in squid, which prefer colder
water, but does not substantially affect cetaceans feeding on squid in
deeper waters.
[[Page 29215]]
[GRAPHIC] [TIFF OMITTED] TP22JN18.000
The northern GOM is characterized as semi-tropical, with a seasonal
temperature regime influenced mainly by tropical currents in the summer
and continental influences during the winter. The GOM is
topographically diverse, with an extensive continental shelf
(comprising about 30 percent of the total area), a steep continental
slope, and distinctive bathymetric and morphologic processes and
features. These include the Flower Garden Banks, which are surface
expressions of salt domes that host the northernmost coral reefs in the
U.S. The northern GOM also has a small section of the larger abyssal
plain of the greater GOM. The GOM has about 60 percent of U.S. tidal
marshes, hosting significant nursery habitat for fish and other marine
species. A major climatological feature is tropical storm activity,
including hurricanes. Sea surface temperature ranges from 14-24 [deg]C
in the winter and 28-30 [deg]C in the summer. The area is considered to
be of moderately high productivity (referring to fixated carbon (i.e.,
g C/m\2\/yr), which relates to the carrying capacity of an ecosystem).
Muddy clay-silts and muddy sands dominate bottom substrates of the
region offshore Texas and Louisiana, transitioning to sand, gravel, and
shell from Alabama to Florida. The shelf off Florida is a carbonate
limestone substrate overlain with sand and silt, supporting extensive
seagrass beds, and interspersed with gravel-rock and coral reefs. The
continental shelf in the western GOM is broadest (up to 135 miles) off
Houston, Texas, and east to offshore the Atchafalaya Delta, Louisiana.
It reaches its narrowest point (approximately 12 miles) near the mouth
of the Mississippi River southeast of New Orleans, Louisiana. The
continental shelf is narrow offshore Mobile Bay, Alabama, but broadens
significantly offshore Florida to almost 200 miles wide.
Topography of the continental slope off the Florida panhandle is
relatively smooth and featureless aside from the De Soto Canyon,
whereas the slope off western Florida is distinguished by steep
gradients and irregular topography. In the central and western GOM, the
continental slope is characterized by canyons, troughs, mini-basins,
and salt structures (e.g., small diapiric domes) with higher relief
than surrounding areas. The Sigsbee Escarpment defines the southern
limit of the Texas-Louisiana slope and was formed by a large system of
salt ridges that underlie the region. In addition to De Soto Canyon off
the coast of Florida, the northern GOM contains four significant
canyons on or near the Texas-Louisiana continental slope: Mississippi
Canyon, located southwest of the Mississippi River Delta; Alaminos
Canyon, located on the western end of the Sigsbee Escarpment; Keathley
Canyon, also located on the western end of the Sigsbee Escarpment; and
Rio Perdido Canyon, located between the Texas-Louisiana continental
slope and the East Mexico continental slope.
The GOM is strongly influenced by freshwater input from several
rivers, most importantly the Mississippi River and its tributary, the
Atchafalaya River. The Mississippi River and its tributaries drain a
large portion of the continental United States and carry large amounts
of freshwater into the GOM along with sediment and a variety of
nutrients and pollutants. The highest volume of
[[Page 29216]]
freshwater from the Mississippi River flows into the GOM from May
through November, when large volumes of turbid water become entrained
in a westward-flowing longshore current. The delivery and deposition of
increased loads of terrestrial organic material, including significant
industrial and agricultural discharge, have often resulted in severe
oxygen depletions in bottom waters and the appearance of a so-called
``dead zone,'' where large numbers of benthic fauna die. This is the
largest zone of coastal hypoxia in the western hemisphere.
Wetlands in the GOM have experienced severe loss and degradation,
due in part to interference with normal erosional/depositional
processes, sea level rise, and coastal subsidence. Wetlands are
converted to open water when accretion is insufficient to compensate
for natural subsidence, while large areas of wetlands have been drained
for industrial, urban, and agricultural development. Increasing
salinity due to saltwater intrusion accompanies these changes, which
further exacerbates the loss of coastal flora. This loss of wetlands
ultimately increases erosion due to waves and tides, with the whole
issue exacerbated by sea level rise.
The northern GOM hosts a vigorous complex of offshore hydrocarbon
exploration, extraction, shipping, service, construction, and refining
industries, resulting in additional impacts to coastal wetlands as well
as large- and small-scale petroleum discharges and oil spills. Of
particular note, in 2010 the Macondo discovery blowout and explosion
aboard the Deepwater Horizon drilling rig (also known as the Deepwater
Horizon explosion, oil spill, and response; hereafter referred to as
the DWH oil spill) caused oil, natural gas, and other substances to
flow into the GOM for 87 days before the well was sealed. Total oil
discharge was estimated at 3.19 million barrels (134 million gallons),
resulting in the largest marine oil spill in history (DWH NRDA
Trustees, 2016). In addition, the response effort involved extensive
application of dispersants at the seafloor and at the surface, and
controlled burning of oil at the surface was also used extensively as a
response technique. The oil, dispersant, and burn residue compounds
present ecological concerns in the region. We discuss the impacts of
the DWH oil spill on marine mammals in greater detail later in our
``Description of Marine Mammals in the Area of the Specified Activity''
section.
The GOM is also known for having many natural hydrocarbon seeps
that contribute to a background level of chemicals in the environment.
Chemosynthetic communities with aerobic bacterial components typically
are associated with natural oil seeps. These naturally occurring seeps
are common in deep slope waters, and there are hundreds of known,
constant seeps that produce perennial slicks of oil at consistent
locations (Kvenvolden and Cooper, 2003). DWH NRDA Trustees (2016)
provided an estimate of the total amount of natural oil seepage in the
GOM of between 9 and 23 million gallons per year. Although there is
much uncertainty in attempting to estimate seepage rates (Kvenvolden
and Cooper, 2003), it is clear that natural seepage is not comparable
to the DWH oil spill release; about six to 15 times more oil was
released from a single location in 87 days as is typically slowly
released in a year from thousands of seeps across the entire GOM.
In addition to being a major area for activities associated with
the oil and gas industry, the GOM hosts significant amounts of
commercial fishing and tourism activities and has two of the world's
busiest shipping fairways and top-ranking ports for container and
passenger vessel traffic, all of which are noise-producing activities.
The underwater environment is typically loud due to ambient sound,
which is defined as environmental background sound levels lacking a
single source or point (Richardson et al., 1995). The sound level of a
region is defined by the total acoustical energy being generated by
known and unknown sources. These sources may include physical (e.g.,
wind and waves, earthquakes, ice, atmospheric sound), biological (e.g.,
sounds produced by marine mammals, fish, and invertebrates), and
anthropogenic (e.g., vessels, dredging, construction) sound. A number
of sources contribute to ambient sound, including wind and waves, which
are a main source of naturally occurring ambient sound for frequencies
between 200 hertz (Hz) and 50 kilohertz (kHz) (Mitson, 1995) (for
description of metrics related to underwater sound, please see the
``Description of Sound Sources'' section later in this document). In
general, ambient sound levels tend to increase with increasing wind
speed and wave height. Precipitation can become an important component
of total sound at frequencies above 500 Hz, and possibly down to 100 Hz
during quiet times. Marine mammals can contribute significantly to
ambient sound levels, as can some fish and snapping shrimp. The
frequency band for biological contributions is from approximately 12 Hz
to over 100 kHz. Sources of ambient sound related to human activity
include transportation (surface vessels), dredging and construction,
oil and gas drilling and production, geophysical surveys, sonar, and
explosions. Vessel noise typically dominates the total ambient sound
for frequencies between 20 and 300 Hz. In general, the frequencies of
anthropogenic sounds are below 1 kHz and, if higher frequency sound
levels are created, they attenuate rapidly.
The sum of the various natural and anthropogenic sound sources that
comprise ambient sound at any given location and time depends not only
on the source levels (as determined by current weather conditions and
levels of biological and human activity) but also on the ability of
sound to propagate through the environment. In turn, sound propagation
is dependent on the spatially and temporally varying properties of the
water column and sea floor, and is frequency-dependent. As a result of
the dependence on a large number of varying factors, ambient sound
levels can be expected to vary widely over both coarse and fine spatial
and temporal scales. Sound levels at a given frequency and location can
vary by 10-20 decibels (dB) from day to day (Richardson et al., 1995).
Estabrook et al. (2016) measured underwater noise at seven sites in
the northern GOM, within three frequency bands (10-500 Hz (LF); 500-
1,000 Hz (MF); 1,000-3,150 Hz (HF)). The authors found that the GOM is
a spectrally, temporally, and spatially dynamic ambient noise
environment, and that, while abiotic and other anthropogenic noise
sources contributed significantly to the ambient noise environment,
noise from geophysical surveys dominated the noise environment during
the study period (2010-2012) and chronically elevated noise levels
across several marine habitats. Specifically, although wind was a
significant noise source at higher frequencies (i.e., 500-3,550 Hz),
these levels were relatively low compared to those of anthropogenic
noise in the low-frequency band (10-500 Hz). Previous studies had
identified anthropogenic sound as a major noise contributor in the GOM
(e.g., Newcomb et al., 2003); however, Estabrook et al. (2016) found
that sound levels from shipping activity were not nearly as pronounced
as those from geophysical surveys, which, in many cases, persisted for
months. As described below, typical airgun surveys fire pulses
approximately every 10-20 seconds but, in addition, the resulting
multipath propagation and reverberation from airgun pulses can exceed
ambient levels during the interpulse interval (Guerra et
[[Page 29217]]
al., 2011; Guan et al., 2015). Estabrook et al. (2016) found that, in
some instances, there were near-continuous elevated noise levels and
that airgun noise propagated over large spatial scales of several
hundred kilometers. Background noise, considered to be the noise level
that is present in the absence of notable anthropogenic, biological,
and meteorological sound sources, was measured across all sites as
follows: 102 dB (LF), 84 dB (MF), and 85 dB (HF). The median equivalent
continuous sound pressure level across all sites was: 112 dB (LF), 90
dB (MF), and 93 dB (HF). Finally, the median equivalent continuous
sound pressure level for a five-day interval when airgun pulses were
present was: 124 dB (LF), 91 dB (MF), and 92 dB (HF).
Wiggins et al. (2016) also monitored the northern GOM soundscape
over a comparable time period (2010-2013), conducting measurements at
five locations and monitoring frequencies from 10-1,000 Hz. The authors
made similar findings, i.e., that average ambient noise levels at low
frequencies in the northern GOM are among the highest measured in the
world's oceans, and geophysical surveys dominate these high noise
levels. In fact, Wiggins et al. (2016) found that during passage of a
hurricane, low frequency sound pressure levels actually decreased due
to the absence of survey activity. Although shipping noise was
observed, the duration was typically shorter (approximately one hour
versus more than 12 hours), and was masked by airgun noise at lower
frequencies.
Detailed Description of Activities
An airgun is a device used to emit acoustic energy pulses into the
seafloor, and generally consists of a steel cylinder that is charged
with high-pressure air. There are different types of airguns;
differences between types of airguns are generally in the mechanical
parts that release the pressurized air, and the bubble and acoustic
energy released are effectively the same. Airguns are typically
operated at a firing pressure of 2,000 pounds per square inch (psi).
Release of the compressed air into the water column generates a signal
that reflects (or refracts) off the seafloor and/or subsurface layers
having acoustic impedance contrast. Individual airguns are available in
different volumetric sizes and, for deep penetration seismic surveys,
are towed in arrays (i.e., a certain number of airguns of varying sizes
in a certain arrangement) designed according to a given company's
method of data acquisition, seismic target, and data processing
capabilities.
Airgun arrays are typically configured in subarrays of 6-12 airguns
each. Towed hydrophone streamers (described below) may follow the array
by 100-200 m and can be 5-12 kilometer (km) long. The airgun array and
streamers are typically towed at a speed of approximately 4.5 to 5
knots (kn). BOEM notes that arrays used for deep penetration surveys
typically have between 20-80 individual elements, with a total volume
of 1,500-8,460 in\3\. However, BOEM's permitting records show that
during one recent year, over one-third of arrays in use had volumes
greater than 8,000 in\3\. The output of an airgun array is directly
proportional to airgun firing pressure or to the number of airguns, and
is expressed as the cube root of the total volume of the array.
Airguns are considered to be low-frequency acoustic sources,
producing sound with energy in a frequency range from less than 10 Hz
to 2 kHz (though there may be energy in the signal at frequencies up to
5 kHz), with most energy radiated at frequencies below 500 Hz.
Frequencies of interest to industry are below approximately 100 Hz. The
amplitude of the acoustic wave emitted from the source is equal in all
directions (i.e., omnidirectional) for a single airgun, but airgun
arrays do possess some directionality due to phase delays between guns
in different directions. Airgun arrays are typically tuned to maximize
functionality for data acquisition purposes, meaning that sound
transmitted in horizontal directions and at higher frequencies is
minimized to the extent possible.
When fired, a brief (~0.1 second) pulse of sound is emitted by all
airguns in an array nearly simultaneously, in order to increase the
amplitude of the overall source pressure signal. The combined signal
amplitude and directivity is dependent on the number and sizes of
individual airguns and their geometric positions within the array. The
airguns are silent during the intervening periods, with the array
typically fired on a fixed distance (or shot point) interval. The
intervals are optimized for water depth and the distance of important
geological features below seafloor, but a typical interval in
relatively deep water might be approximately every 10-20 s (or 25-50 m,
depending on vessel speed). The return signal is recorded by a
listening device, and later analyzed with computer interpretation and
mapping systems used to depict the subsurface. There must be enough
time between shots for the sound signals to propagate down to and
reflect from the feature of interest, and then to propagate upward to
be received on hydrophones or geophones. Reverberation of sound from
previous shots must also be given time to dissipate. The receiving
hydrophones can be towed behind or in front of the airgun array (may be
towed from the source vessel or from a separate receiver vessel), or
geophone receivers can be deployed on the seabed. Receivers may be
displaced several kilometers horizontally away from the source, so
horizontal propagation time is also considered in setting the interval
between shots.
Sound levels for airgun arrays are typically modeled or measured at
some distance from the source and a nominal source level then back-
calculated. Because these arrays constitute a distributed acoustic
source rather than a single point source (i.e., the ``source'' is
actually comprised of multiple sources with some pre-determined spatial
arrangement), the highest sound levels measurable at any location in
the water will be less than the nominal source level. A common analogy
is to an array of light bulbs; at sufficient distance--in the far
field--the array will appear to be a single point source of light but
individual sources, each with less intensity than that of the whole,
may be discerned at closer distances (Caldwell and Dragoset (2000)
define the far field as greater than 250 m). Therefore, back-calculated
source levels are not typically considered to be accurate indicators of
the true maximum amplitude of the output in the far field, which is
what is typically of concern in assessing potential impacts to marine
mammals. In addition, the effective source level for sound propagating
in near-horizontal directions (i.e., directions likely to impact most
marine mammals in the vicinity of an array) is likely to be
substantially lower (e.g., 15-24 dB; Caldwell and Dragoset, 2000) than
the nominal source level applicable to downward propagation because of
the directional nature of the sound from the airgun array. The
horizontal propagation of sound is reduced by noise cancellation
effects created when sound from neighboring airguns on the same
horizontal plane partially cancel each other out.
Survey protocols generally involve a predetermined set of survey,
or track, lines. The seismic acquisition vessel(s) (source vessel) will
travel down a linear track for some distance until a line of data is
acquired, then turn and acquire data on a different track. In some
cases, data is acquired as the source vessel(s) turns continuously
rather than moving on a linear track (i.e., coil surveys). The spacing
between track lines and the length of track lines can vary greatly,
depending on the objectives of a survey.
[[Page 29218]]
In addition to the line over which data acquisition is desired, full-
power operation may include run-in and run-out. Run-in is approximately
1 km of full-power source operation before starting a new line to
ensure equipment is functioning properly, and run-out is additional
full-power operation beyond the conclusion of a trackline (e.g., half
the distance of the acquisition streamer behind the source vessel, when
used) to ensure that all data along the trackline are collected by the
streamer. Line turns can require two to six hours when towed
hydrophones are used, due to the long trailing streamers, but may be
much faster when streamers are not used. Spacing and length of tracks
varies by survey. Survey operations often involve the source vessel(s),
supported by a chase vessel. Chase vessels typically support the source
vessel(s) by protecting the long hydrophone streamer from damage (e.g.,
from other vessels) (when used) and otherwise lending logistical
support (e.g., returning to port for fuel, supplies, or any necessary
personnel transfers). Chase vessels do not deploy acoustic sources for
data acquisition purposes; the only potential effects of the chase
vessels are those associated with normal vessel operations.
The general activities described here could occur pre- or post-
leasing and/or on- or off-lease. Pre-lease surveys are more likely to
involve larger-scale activity designed to explore or evaluate geologic
formations. Post-lease activities may also include deep penetration
surveys, but would be expected to be smaller in spatial and temporal
scale as they are associated with specific leased blocks. Shallow
penetration and HRG surveys are more likely to be associated with
specific leased blocks and/or facilities, with HRG surveys used along
pipeline routes and to search for archaeological resources and/or
benthic communities. Specific types of surveys are described below
(summarized from the petition); for full detail please refer to
sections 1.2 and 1.3 of the petition.
While these descriptions reflect existing technologies and current
practice, new technologies and/or uses of existing technologies may
come into practice during the period of validity of these proposed
regulations. NMFS will evaluate any such developments on a case-
specific basis to determine whether expected impacts on marine mammals
are consistent with those described or referenced in this document and,
therefore, whether any anticipated take incidental to use of those new
technologies or practices is appropriately authorized under what would
be the existing regulatory framework. We also note here that activities
that may result in incidental take of marine mammals, and which would
therefore appropriately require authorization under the MMPA, are not
limited to those activities requiring permits from BOEM. Operators
should be aware that there may be some activities previously
unpermitted by BOEM, such as certain ancillary activities, that would
appropriately be subject to the requirements of this proposed rule and
they should consult NMFS regarding the need to obtain a LOA under this
rule prior to conducting such activities. Unauthorized taking of marine
mammals is a violation of the MMPA.
2D and 3D Surveys (Deep Penetration Surveys)--As discussed, deep
penetration surveys use an airgun array(s) as the acoustic source and
may be 2D or 3D (with repeated 3D surveys termed 4D). Surveys may be
designed as either multi-source (i.e., multiple arrays towed by one or
more source vessel(s)) or single source. Surveys may also be
differentiated by the way in which they record the return signals using
hydrophones and/or geophones. Hydrophones may be towed in streamers
behind a vessel (either the source vessel(s) or a separate vessel) or
in some cases may be placed in boreholes (called vertical seismic
profiling) or spaced at various depths on vertical cables in the water
column. Sensors may also be incorporated into ocean-bottom cables (OBC)
or autonomous ocean-bottom nodes (OBN) and placed on the seafloor--
these surveys are referred to generally as ocean-bottom seismic (OBS).
Autonomous nodes can be tethered to coated lines and deployed from
ships or remotely-operated vehicles, with current technology allowing
use in water depths to approximately 3,000 m. OBS surveys are most
useful to acquire data in shallow water and obstructed areas, as well
as for acquisition of four-component survey data (i.e., including
pressure and 3D linear acceleration collected via geophone). For OBS
surveys, one or two vessels usually are needed to lay out and pick up
cables, one ship is needed to record data, one ship tows an airgun
array, and two smaller utility boats support survey operations. The
size of the OBS receiver grid is usually limited by the amount of
equipment available; however, to efficiently conduct a survey,
approximately 500 nodes or 100 km of cable are needed.
We described previously the basic differences between 2D and 3D
surveys. A typical 2D survey deploys a single array covering an area
approximately 12.5-18 m long and 16-36 m wide behind the source vessel,
whereas a 3D vessel may deploy multiple source arrays and/or streamers,
with a potentially much larger width behind the vessel. A 3D vessel
usually will tow 8-14 streamers (but as many as 24), each 3-8 km long.
For example, an array containing ten streamers could have a total swath
width behind the vessel of 675-1,350 m. Among 3D surveys in particular,
there are a variety of survey designs employed to acquire the specific
data of interest. These survey types may differ in the number of
vessels used (for source or receiver), sound sources deployed, and the
location or type of hydrophones. Conventional, single-vessel 3D surveys
are referred to as narrow azimuth (NAZ) surveys. Other 3D survey
techniques include wide-azimuth (WAZ), multi-azimuth (MAZ), rich-
azimuth (RAZ), and full-azimuth (FAZ) surveys. Please see Figures 1-10
and 1-11 in the petition for depictions of these survey geometries.
In conventional 3D seismic surveys involving a single source
vessel, only a subset of the reflected wave field can be obtained
because of the narrow range of source-receiver azimuths (thus called
NAZ surveys). Newer survey techniques, as well as improvements in data
processing, provide better data quality than that achievable using
traditional NAZ surveys, including better illumination, higher signal-
to-noise ratios, and higher resolution. This is useful in imaging
subsurface areas containing complex geologic structures, particularly
those beneath salt bodies with irregular geometries.
Offset refers to the distance between a source and a particular
receiver, while azimuth refers to the angles covered by the various
directions between a source and individual receiving sensors. With NAZ
surveys, the width (crossline dimension) of the nominal area imaged
when the source is fired one time will be less than half the length
(inline dimension). The aspect ratio (crossline divided by inline) of
this nominal area is much less than 0.5 (see Figure 1-10 of the
petition).
To achieve wider azimuthal coverage, multiple source vessels are
deployed in order to achieve greater crossline dimension of the nominal
area imaged. Different WAZ methods using multiple source vessels and,
in some cases, multiple receiver vessels, are depicted in Figure 1-11
of the petition. A basic method used to acquire MAZ data involves a
single source and streamer vessel, using conventional 3D survey
methodology, covering transects on the same area multiple times along
different azimuthal directions (Figure 1-11D of
[[Page 29219]]
the petition). A combination of WAZ and MAZ geometries provides either
RAZ or FAZ results. Acquisition of RAZ data requires using multiple
passes of one source-and-streamer vessel and two source-only vessels.
Making two passes at right angles to each other with a specific WAZ
configuration would produce 180[deg] azimuth (i.e., FAZ) coverage. New
survey designs will likely continue to be tested as the industry works
to make WAZ, MAZ, RAZ, and FAZ shooting more efficient and less costly.
Another development is synchronized discharge of airgun arrays being
towed by different vessels (advances in data processing can separate
the energy from synchronized sources using differences in source-to-
receiver offset distances). While this increases the level of sound in
the ensonified water volume, it also reduces the length of time that
the water volume is ensonified.
In summary, 3D survey design involves a vessel with one or more
acoustic sources covering an area of interest with relatively tight
spatial configuration. In order to provide richer, more useful data,
particularly in areas with more difficult geology, survey designs
become more complicated with additional source and/or receiver vessels
operating in potentially increasingly complicated choreographies. The
time required to complete one pass of a trackline for a single NAZ
vessel and the time required for one pass by a multi-vessel entourage
conducting a WAZ survey will be essentially the same. Turn times will
be somewhat longer during multi-vessel surveys to ensure that all
vessels are properly aligned prior to beginning the next trackline.
Turn times depend mostly on the vessels and the equipment they are
towing (as in conventional 3D surveys); however, the number of vessels
towing streamers in the entire entourage is the main determinant of the
turn time. The MAZ technique, where multiple passes are made, increases
the time needed for a survey in proportion to the number of passes that
will be made within an area. The reduction in the number of passes is
one of the most significant driving factors in continued efforts to
design more efficient surveys. Coil surveys, described previously,
reduce the total survey time due to elimination of the trackline-turn
methodology.
Borehole Seismic Surveys--The placement of seismic sensors in a
drilled well or borehole is another way data can be acquired. These
surveys, typically referred to as vertical seismic profiles (VSP),
provide information about geologic structure, lithology, and fluids
that is intermediate between that obtained from sea surface surveys and
well-log scale information (well logging is the process of recording
various physical, chemical, electrical, or other properties of the
rock/fluid mixtures penetrated by drilling a borehole). VSP surveying
is conducted by placing receivers such as geophones at many (50-200)
depths in a wellbore and recording both direct-arriving and reflection
energy from an acoustic source. The acoustic source usually is a single
airgun or small airgun array hung from a platform or deployed from a
source vessel. The airguns used for VSPs may be the same or similar to
those used for 2D and 3D towed-streamer surveys; however, the number of
airguns and the total volume of an array used are less than those used
for towed-streamer surveys. Less sound energy is required for VSP
surveys because the seismic sensors are in a borehole, which is a much
quieter environment than that for sensors in a towed streamer, and
because the VSP sensors are located nearer to the targeted reflecting
horizons. Some VSP surveys take less than a day, and most are completed
in a few days. Borehole seismic surveys include 2D VSPs, 3D VSPs,
checkshot surveys, and seismic while drilling (SWD).
Types of 2D VSPs are defined by source location, as follows: (1)
Zero-offset VSPs involve a single source position that is close to the
well (often deployed from a platform) compared to the depths where the
sensors are placed (thereby causing the sensors to receive mostly
vertically propagating energy); (2) offset VSPs involve a stationary
vessel-based source position (or multiple positions) that is far enough
away from the well that the recorded waveforms have a significant
amount of horizontally-propagating energy; (3) walkaway VSPs involve a
moving vessel and multiple source positions along a line away from the
well; and (4) deviated-well VSPs involve source positions placed
vertically above a well path. See Figure 1-12 of the petition for
depictions.
3D VSPs involve use of multi-level sensor strings, allowing 1,500
to 3,000 m to be instrumented within a well. As with 2D VSPs,
individual airguns and arrays used are generally similar to those used
in towed-streamer surveys. The data acquisition design could involve
typical 3D rectangular survey vessel track patterns, or spiral track
patterns with the source vessel moving away from the well. For 3D VSPs,
the distance from the well covered by the source vessel will
approximately equal the depth of the well (see Figure 1-13 in the
petition).
Checkshot surveys are similar to zero-offset VSPs but are less
complex. The purpose of a checkshot survey is to estimate the velocity
of sound in rocks penetrated by the well, and these surveys are
typically conducted quickly. These surveys involve a single source
typically hung from a platform and a sensor placed at a few depths in
the well, where only the first energy arrival is recorded.
SWD refers to the acquisition of borehole data, using an airgun
array as an acoustic source, while there is downtime from the actual
drilling operation. SWD surveys are run intermittently for weeks up
until the well completion depth.
Shallow Penetration/HRG Surveys--These surveys are conducted to
provide data informing initial site evaluation, drilling rig
emplacement, and platform or pipeline design and emplacement.
Identification of geohazards (e.g., gas hydrates, buried channels) is
necessary to avoid drilling and facilities emplacement problems, and
operators are required to identify and avoid archaeological resources
and certain benthic communities. In most cases, conventional 2D and 3D
deep penetration surveys do not have the correct resolution to provide
the required information. Although HRG surveys may use a single airgun
source, they generally use electromechanical sources such as side-scan
sonars, shallow- and medium-penetration subbottom profilers, and
single-beam echosounders or multibeam echosounders. Non-airgun HRG
sources are often used in combination in order to acquire necessary
data during a single deployment. HRG surveys are sometimes conducted
using autonomous underwater vehicles (AUV) equipped with multiple
acoustic sources.
HRG surveys may be conducted using airguns as the acoustic source.
These typically use one or two airguns that are the same as those
described for use in arrays during deep penetration surveys. However,
the total volume is typically only approximately 40-400 in\3\, the
streamers are shorter, and the shot intervals are shorter. The intent
is typically to image the shallow subsurface (less than 1,000 m below
the seafloor). Including vessel turns at the end of lines, the time
required to survey one OCS lease block is approximately 36 hours. These
surveys are sometimes conducted using 3D techniques, e.g., multiple
sources and/or streamers.
Electromechanical sources are generally considered to be relatively
[[Page 29220]]
mid- to high-frequency sources, and produce acoustic signals by
creating an oscillatory overpressure through rapid vibration of a
surface, using either electromagnetic forces or the piezoelectric
effect of some materials. A vibratory source based on the piezoelectric
effect is commonly referred to as a transducer, which may be designed
to excite an acoustic wave of a specific frequency, often in a highly
directive beam. The directional capability increases with increasing
operating frequency.
Subbottom profiling surveys are typically used for high-resolution
imaging of the shallow subsurface. These surveys may use a variety of
acoustic sources, commonly referred to as ``boomers,'' ``sparkers,'' or
``chirps.'' A sparker uses electricity to vaporize water, creating
collapsing bubbles that produce a broadband (50 Hz to 4 kHz),
omnidirectional pulse of sound that can penetrate a few hundred meters
into the subsurface. Short hydrophone arrays towed near the sparker
receive the return signal; typically, the sparker is towed on one side
of the vessel and the hydrophone array is towed on the other side. A
boomer consists of a circular piston moved by electromagnetic force,
generating a broadband acoustic pulse (300 Hz to 3 kHz, though
adjustments to the applied electrical impulse may increase the
frequency). Boomer systems can penetrate as deep as 200 m in soft
sediments, though a more typical penetration may be 25-50 m. Boomer
sources show some directionality, which increases with the acoustic
frequency; at frequencies below 1 kHz they can usually be considered
omnidirectional. Boomers are typically sled-mounted and towed behind
the vessel, with short hydrophone arrays used to receive the return
signal. The characteristics of the acoustic wave emitted by the boomer
source are comparable to those emitted by the sparker source.
Chirp (Compressed High-Intensity Radiated Pulse) sources operate
differently, sending a continuous sweep of frequencies (e.g., 500 Hz to
24 kHz) approximately every 0.5 to 1 seconds. Some chirp systems work
in multiple frequency bands simultaneously (e.g., 3.5/12/200 kHz).
Beamwidth will vary depending on the frequency, but is approximately
10-30[deg]. Because this continuous sweep of frequencies provides a
much wider range of information, chirp systems are able to create a
much clearer, higher-resolution image while achieving the same or
better depth of penetration. Chirps are typically towed behind the
vessel or deployed on an AUV.
Side-scan sonars and echosounders do not penetrate the surface of
the seabed, using reflections of sound pulses to locate, image, and aid
in the identification of objects in the water column and on the
seafloor, and to determine water depth. Echosounders typically emit
short, single-frequency signals, with frequency decreasing as water
depth increases. A deep-water system might operate at approximately 3-
12 kHz, while a shallow-water system might operate at 200 kHz or
greater. Multibeam echosounder systems use an array of transducers that
project a fan-shaped beam under the hull of a vessel and perpendicular
to the direction of motion, producing a swath of depth measurements to
ensure full coverage of an area. Echosounders are typically hull-
mounted or deployed on AUVs. Side-scan sonar systems produce shaded
relief images of the ocean bottom by recording the intensity and timing
of signals reflected off the seafloor, and consist of two transducers
on the sides of the towed sonar body that are oriented perpendicularly
to the towing direction. The signals are typically single-frequency,
with a highly directional beam that is wide across-track and narrow in
the direction of travel. Due to the transducer placement, side-scan
sonars may not effectively image the area directly beneath the vessel
and are often used in conjunction with echosounders. Side-scan sonars
are typically high-frequency sources and therefore have a limited range
(50-200 m). In deeper water, the source may be towed at greater depth
or deployed on an AUV.
Representative Sound Sources
Because the specifics of acoustic sources to be used would not be
known in advance of receiving LOA requests from industry operators, it
is necessary to define representative acoustic source parameters, as
well as representative survey patterns. BOEM determined realistic
representative proxy sound sources and survey patterns, which are used
in the modeling and more broadly to support the analysis, after
discussions with individual geophysical companies.
Representative sources include a single airgun, an airgun array,
and multiple electromechanical sources: Boomer, chirp, multibeam
echosounder, and side-scan sonar. Two major survey types were
considered: Large-area seismic and small-area, high-resolution
geotechnical. Large-area seismic surveys are assumed to cover more than
1,000 mi\2\ (2,590 km\2\) and include 2D, 3D NAZ, 3D WAZ, and coil
types. Geotechnical study surveys are assumed to cover an area less
than 100 mi\2\ (259 km\2\) and use small airguns and/or high-frequency
electromechanical sources installed on an AUV. VSP surveys, assuming a
single source vessel with one 8,000 in\3\ array, were also modeled.
The nominal airgun sources used for analysis of the proposed action
include a small single airgun (90 in\3\ Sercel airgun) towed at 4 m
depth and a large airgun array (8,000 in\3\) towed at 8 m depth.
Airguns are assumed to fire simultaneously at 2,000 psi. The airgun
array was assumed to consist of 72 elements (Bolt 1900 LLXT airguns)
arranged in six sub-arrays of 12 airguns each with 9 m in-line
separations. Individual elements range from 40 to 250 in\3\. The layout
of the modeled array (i.e., airgun distribution in the horizontal
plane) is shown in Figure 11 of Zeddies et al. (2015). For the single
airgun, modeled source levels were 227.7 dB 0-peak (pk) sound pressure
level (SPL) and 207.8 dB sound exposure level (SEL) (for description of
metrics related to underwater sound, please see ``Description of Sound
Sources,'' later in this document). Modeled source levels for the array
range from 248.1 (broadside, i.e., perpendicular to the tow direction)
to 255.2 (endfire; i.e., parallel to the tow direction) dB 0-pk SPL and
from 225.7 (broadside) to 231.8 (endfire) dB SEL. Zeddies et al. (2015,
2017a), ``Acoustic Propagation and Marine Mammal Exposure Modeling of
Geological and Geophysical Sources in the Gulf of Mexico'' and
``Addendum to Acoustic Propagation and Marine Mammal Exposure Modeling
of Geological and Geophysical Sources in the Gulf of Mexico,'' are
hereafter referred to as ``the modeling report.'' The reports are
available online at: www.fisheries.noaa.gov/national/marine-mammal-protection/incidental-take-authorizations-oil-and-gas. Below, we
outline the representative operational parameters of the different
survey types that were used in the modeling simulations to predict the
exposure of marine mammals to different received levels of sound.
Source vessels are assumed to travel at an average speed of 4.5-5
kn (i.e., 200-220 linear km per day), and airgun arrays were assumed to
be off during turns. The run-in and run-out sections were 1 km long.
Each large-area survey (excluding coil surveys) was assumed to cover an
area of 10 x 30 lease blocks, equivalent to 48 x 145 km or
approximately 6,960 km\2\. Coil surveys are assumed to cover a smaller
area of 12 x 12 lease blocks, equivalent to 58 x 58 km or approximately
3,364 km\2\.
2D surveys were simulated by assuming use of a single 8,000 in\3\
array,
[[Page 29221]]
with transect lines offset laterally by 4.8 km. The production lines
were filled in with a racetrack fill-in method, skipping two tracks on
the left side turn (15 km wide turn) and transitioning onto the
adjacent line on the right side turn (5 km wide turn) (see Figure 105
of the modeling report). The vessel speed was 4.5 kts and the shot
interval was 21.6 s (approximately every 50 m).
3D NAZ surveys were simulated by assuming use of two source vessels
towing identical arrays. Sources at each vessel produce seismic pulses
simultaneously. Both vessels follow the same track, but were separated
along the track by 6 km. The production lines were laterally spaced by
1 km (see Figure 106 of the modeling report). The production lines were
filled via a racetrack fill-in method with eight loops in each
racetrack (7-8 km wide turn). Forty-nine lines were required to fully
cover the survey area. The vessel speed was 4.9 kn and the shot
interval was 15 s (approximately every 37.5 m) for each vessel.
3D WAZ surveys were simulated by assuming use of four source
vessels towing identical arrays. Sources at each vessel produce seismic
pulses sequentially. The tracks of each vessel had the same geometry
and had 1.2 km lateral offset. The vessels also had 500 m offset along
the track (see Figure 107 of the modeling report). The production lines
were filled in with a racetrack fill-in method with two loops in each
racetrack (9.6 km wide turn). Forty lines were required to fully cover
the survey area. The vessel speed was 4.5 kn, with individual vessel
shot interval of 86.4 s (approximately every 200 m)--equivalent to 21.6
s for the group.
Coil surveys are performed by multiple vessels that sail a series
of circular tracks with some angular separation while towing acoustic
sources. These surveys were simulated by assuming use of four source
vessels towing identical arrays. Sources at each vessel produce seismic
pulses simultaneously. Tracks consist of a series of circles with 12.5
km diameter (see Figure 108 of the modeling report). Once each vessel
completes a full circle, it advances to the next one along a tangential
connection segment. The offset between the center of one circle and the
next, either along-swath or between swaths, was 5 km. The full survey
geometry consisted of two tracks with identical configuration with 1.2
km and 600 m offsets along X and Y directions, respectively. Two of the
four vessels followed the first track with 180[deg] separation; the
other two vessels followed the second track with 180 [deg] separation
relative to each other and 90 [deg] separation relative to the first
pair. One hundred circles per vessel pair were required to fully cover
the survey area. The vessel speed was 4.9 kn and the shot interval was
20 s (approximately every 50 m) for each vessel.
For small-area, high-resolution geotechnical surveys, we described
the proxy single airgun source above. The representative boomer system
was the Applied Acoustics AA301, based on a single plate with
approximately 40 cm baffle diameter. The input energy for the AA301
boomer plate was up to 350 joules (J) per pulse or 1,000 J per second.
The width of the pulse was 0.15-0.4 milliseconds (ms). A source
verification study performed on a similar system by Martin et al.,
(2012) showed that the broadband source level for the system was 203.3
dB root mean square (rms) SPL over a 0.2 ms window length and 172.6 dB
SEL. These data were used for modeling the boomer source with a -4.6 dB
correction applied to account for differences in input energy between
the two systems.
As noted above, certain high-resolution acoustic sources may be
deployed together and used concurrently. Here, the modeling assumes
that a multibeam echosounder, side-scan sonar, and chirp subbottom
profiler are operated concurrently and deployed on an AUV. Towing depth
of the AUV was assumed to be 4 m below the sea surface when the water
depth was less than 100 m and 40 m above the seafloor where water depth
was more than 100 m. The representative multibeam echosounder (MBES)
system was the Simrad EM2000 (manufactured by Kongsberg Maritime AS).
According to manufacturer specifications, this device operates at 200
kHz and is equipped with a transducer head that produces a single beam
17 [deg] x 88 [deg] wide. The nominal source level was 203 dB rms SPL,
with per-pulse SEL dependent on the pulse length (160-175 dB). Pulse
width is 0.04-1.3 ms. The representative side-scan sonar is the
EdgeTech 2200 IM, which works at two frequencies simultaneously (120
and 410 kHz). The beam angle produced by two side-mounted transducers
was 70 [deg] x 0.8 [deg] at 120 kHz and 70 [deg] x 0.5 [deg] at 410
kHz. At 120 kHz, the estimated peak source level is 210 dB with pulse
length of 8.3 ms; at 410 kHz these values are 216 dB and 2.4 ms. The
chirp subbottom profiler uses the same side-scan sonar system, which is
designed as a modular system for installation on an AUV, and adds the
DW-424, a full spectrum chirp subbottom profiler that produces a sweep
signal in the frequency range from 4 to 24 kHz. The projected beamwidth
varies from 15 [deg] to 25 [deg] depending on the emitted frequency,
with estimated source level of 200 dB and pulse length of 10 ms.
For these HRG surveys, the same survey pattern was assumed
regardless of the source. Total survey area was assumed to be an area
of 1 x 3 lease blocks, equivalent to 5 x 14.5 km or approximately 72.5
km\2\. A single source vessel towing the appropriate source (i.e.,
single airgun, boomer, or AUV with concurrently operated MBES, side-
scan sonar, and chirp) was assumed. Production lines were laterally
spaced 30 m (see Figure 109 of the modeling report) then filled in with
a racetrack fill-in method where each racetrack has 20 loops (1.2 km
wide turn). One hundred and sixty lines were required to fully cover
the survey area. The vessel speed was 4 kn and, for surveys using the
single airgun, the shot interval was 10 seconds(s) (approximately every
20 m).
Estimated Levels of Effort
As noted previously, actual total amounts of effort by survey type
and location would not be known in advance of receiving LOA requests
from industry operators. Therefore, BOEM provided projections of survey
level of effort for the different survey types for a 10-year period
(note that this proposed rule covers only a 5-year period). In order to
construct a realistic scenario for future geophysical survey effort,
BOEM evaluated recent trends in permit applications as well as industry
estimates of future survey activity. BOEM also accounted for
restrictions under the Gulf of Mexico Energy Security Act (GOMESA; Pub.
L. 109-432), which precludes leasing, pre-leasing, or any related
activity (though not geophysical surveys that have been permitted) in
the GOM east of 86[deg]41' W, in BOEM's Eastern Planning Area (EPA) and
within 125 mi (201 km) of Florida, or in BOEM's Central Planning Area
(CPA) and within 100 mi of Florida (and according to certain other
detailed stipulations). These leasing restrictions, which will to some
degree influence geophysical survey effort, are in place until June 30,
2022.
In order to provide some spatial resolution to the projections of
survey effort and to provide reasonably similar areas within which
acoustic modeling might be conducted, the geographic region was divided
into seven zones, largely on the basis of water depth, seabed slope,
and defined BOEM planning area boundaries. Shelf regions typically
extend from shore to approximately 100-200 m water depths where
bathymetric relief is gradual (off Florida's west coast, the shelf
extends
[[Page 29222]]
approximately 150 km). The slope starts where the seabed relief is
steeper and extends into deeper water; in the GOM water deepens from
100-200 m to 1,500-2,500 m over as little as a 50 km horizontal
distance. As the slope ends, water depths become more consistent,
though depths can vary from 2,000-3,300 m. Three primary bathymetric
areas were defined as shelf (0-200 m water depth), slope (200-2,000 m),
and deep (>2,000 m).
Available information regarding cetacean density in the GOM (e.g.,
Roberts et al., 2016) shows that, in addition to water depth, animal
distribution tends to vary from east to west in the GOM and appears
correlated with the width of shelf and slope areas from east to west.
The western region is characterized by a relatively narrow shelf and
moderate-width slope. The central region has a moderate-width shelf and
moderate-width slope, and the eastern region has a wide shelf and a
very narrow slope. Therefore, BOEM's western, central, and eastern
planning area divisions provide appropriate longitudinal separations
for the shelf and slope areas. Due to relative consistency in both
physical properties and predicted animal distribution, the deep area
was not subdivided. As shown in Figure 2, Zones 1-3 represent the shelf
area (from east to west), Zones 4-6 represent the slope area (from east
to west), and Zone 7 is the deep area (note that other features of
Figure 2 are described in the ``Estimated Take'' section). Table 1
displays BOEM's 10-year estimated levels of effort, estimated as 24-hr
survey days, including annual totals by survey type and by zone for
deep penetration and shallow penetration surveys, respectively.
[GRAPHIC] [TIFF OMITTED] TP22JN18.001
Table 1--Projected Levels of Effort in 24-Hr Survey Days for Ten Years, by Zone and Survey Type \1\
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Shallow Total
Year Zone \2\ 2D \3\ 3D NAZ \3\ 3D WAZ \3\ Coil \3\ VSP \3\ Total hazards Boomer \4\ HRG \4\ (shallow)
(deep) \3\ \4\ \4\
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
1................................... 1................................. 0 0 0 0 0 0 0 0 1 1
2................................. 0 243 0 0 0 243 2 0 19 21
3................................. 0 30 0 0 0 30 0 0 4 4
4................................. 0 0 0 0 0 0 0 0 0 0
5................................. 56 389 192 82 2 721 0 0 26 26
6................................. 0 186 49 21 0 256 0 0 10 10
7................................. 69 515 248 106 2 940 0 0 34 34
-----------------------------------------------------------------------------------------------------------------------------------------------------------
Total........................... .................................. 125 1,363 489 209 4 2,190 2 0 94 96
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
2................................... 1................................. 0 0 0 0 0 0 0 0 1 1
2................................. 0 364 43 19 0 426 2 0 19 21
3................................. 0 0 0 0 0 0 0 0 4 4
4................................. 33 0 0 0 0 33 0 0 0 0
5................................. 0 389 192 82 2 665 0 0 26 26
6................................. 0 99 0 0 0 99 0 0 11 11
7................................. 30 502 241 103 2 878 0 0 34 34
-----------------------------------------------------------------------------------------------------------------------------------------------------------
Total........................... .................................. 63 1,354 476 204 4 2,101 2 0 95 96
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
[[Page 29223]]
3................................... 1................................. 0 0 0 0 0 0 0 0 1 1
2................................. 0 243 0 0 0 243 2 0 18 20
3................................. 0 0 0 0 0 0 0 0 4 4
4................................. 0 0 0 0 0 0 0 0 1 1
5................................. 0 342 160 69 2 573 0 0 27 27
6................................. 0 186 49 21 0 256 0 0 12 12
7................................. 0 456 208 89 2 755 0 0 36 36
-----------------------------------------------------------------------------------------------------------------------------------------------------------
Total........................... .................................. 0 1,227 417 179 4 1,827 2 0 99 101
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
4................................... 1................................. 0 0 0 0 0 0 0 0 0 0
2................................. 0 364 43 19 0 426 2 1 16 19
3................................. 0 30 0 0 0 30 0 0 3 3
4................................. 66 61 21 9 0 157 0 0 1 1
5................................. 28 247 96 41 2 414 0 0 27 27
6................................. 0 99 0 0 0 99 0 0 12 12
7.............................. 94 380 140 60 2 676 0 0 36 36
-----------------------------------------------------------------------------------------------------------------------------------------------------------
Total........................... .................................. 188 1,181 300 129 4 1,802 2 1 95 98
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
5................................... 1................................. 0 0 0 0 0 0 0 0 0 0
2................................. 0 243 0 0 0 243 0 0 20 20
3................................. 0 0 0 0 0 0 0 0 3 3
4................................. 0 92 0 0 0 92 0 0 0 0
5................................. 0 295 192 82 2 571 2 1 25 28
6................................. 0 99 0 0 0 99 0 0 13 13
7................................. 0 467 241 103 3 814 3 2 34 39
-----------------------------------------------------------------------------------------------------------------------------------------------------------
Total........................... .................................. 0 1,196 433 185 5 1,819 5 3 95 103
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
6................................... 1................................. 0 0 0 0 0 0 0 0 0 0
2................................. 0 364 43 19 0 426 0 0 18 18
3................................. 0 0 0 0 0 0 0 0 2 2
4................................. 0 92 0 0 0 92 0 0 1 1
5................................. 0 247 160 69 2 478 0 0 30 30
6................................. 0 186 49 21 0 256 0 0 13 13
7................................. 0 421 208 89 3 721 0 0 40 40
-----------------------------------------------------------------------------------------------------------------------------------------------------------
Total........................... .................................. 0 1,310 460 198 5 1,973 0 0 104 104
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
7................................... 1................................. 0 0 0 0 0 0 0 0 0 0
2................................. 0 243 0 0 0 243 0 0 16 16
3................................. 0 30 0 0 0 30 0 0 2 2
4................................. 33 61 21 9 0 124 0 0 1 1
5................................. 28 247 160 69 2 506 0 0 32 32
6................................. 0 99 0 0 0 99 0 0 13 13
7................................. 64 380 220 94 3 761 0 0 43 43
-----------------------------------------------------------------------------------------------------------------------------------------------------------
Total........................... .................................. 125 1,060 401 172 5 1,763 0 0 107 107
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
8................................... 1................................. 0 0 0 0 0 0 0 0 0 0
2................................. 0 364 43 19 0 426 0 0 16 16
3................................. 0 0 0 0 0 0 0 0 2 2
4................................. 11 61 0 0 0 72 0 0 1 1
5................................. 9 247 128 55 2 441 0 0 35 35
6................................. 0 99 0 0 0 99 0 0 13 13
7................................. 21 380 160 69 3 633 0 0 46 46
-----------------------------------------------------------------------------------------------------------------------------------------------------------
Total........................... .................................. 41 1,151 331 143 5 1,671 0 0 113 113
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
9................................... 1................................. 0 0 0 0 0 0 0 0 0 0
2................................. 0 243 0 0 0 243 0 0 16 16
3................................. 0 0 0 0 0 0 0 0 2 2
4................................. 0 61 0 0 0 61 0 0 1 1
5................................. 0 200 192 82 2 476 0 0 35 35
6................................. 0 99 0 0 0 99 0 0 14 14
7................................. 0 321 241 103 3 668 0 0 47 47
-----------------------------------------------------------------------------------------------------------------------------------------------------------
Total........................... .................................. 0 924 433 185 5 1,547 0 0 115 115
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
10.................................. 1................................. 0 0 0 0 0 0 0 0 0 0
2................................. 0 364 43 19 0 426 0 0 13 13
3................................. 0 30 0 0 0 30 0 0 2 2
4................................. 5 61 0 0 0 66 0 0 1 1
5................................. 0 200 160 69 2 431 0 0 37 37
6................................. 0 99 0 0 0 99 0 0 14 14
7................................. 5 321 200 86 3 615 0 0 49 49
-----------------------------------------------------------------------------------------------------------------------------------------------------------
[[Page 29224]]
Total........................... .................................. 10 1,075 403 174 5 1,667 0 0 116 116
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
\1\ Projected levels of effort in 24-hr survey days.
\2\ Zones follow the zones depicted in Figure 2.
\3\ Deep penetration survey types include 2D, which uses one source vessel with one large array (8,000 in\3\); 3D NAZ, which uses two source vessels using one large array each; 3D WAZ and
coil, each of which uses four source vessels using one large array each (but with differing survey design); and VSP, which uses one source vessel with a large array. ``Deep'' refers to
survey type, not to water depth.
\4\ Shallow penetration/HRG survey types include shallow hazards surveys, assumed to use a single 90 in\3\ airgun, subbottom profiling using a boomer, and high-resolution surveys using the
MBES, side-scan sonar, and chirp systems concurrently. ``Shallow'' refers to survey type, not to water depth.
Table 2 provides a summary of the projected levels of effort. Very
little effort is predicted in the EPA, with no deep penetration surveys
expected in Zone 1 and an annual average of 63 survey days predicted in
Zone 4. Similarly, very little overall effort is expected in western
shelf waters. The vast majority of effort is expected to occur in the
CPA, in all water depths. For deep penetration surveys, 3D NAZ is
expected to be the most common survey type (in terms of total survey
says) with approximately 65 percent of the total. 3D WAZ surveys
represent approximately 22 percent of total survey days. Shallow
penetration surveys overall represent an insignificant addition to the
projected deep penetration effort, reflecting the smaller amount of
effort associated with these survey types.
Year 1 provides an example of what might be a high-effort year in
the GOM, while Year 9 is representative of a low-effort year. A
moderate level of effort in the GOM, according to these projections,
would be similar to the level of effort projected for Year 4. However,
per-zone ranges can provide a different outlook than does an assessment
of total year projected effort across zones. For example, in the
``high'' effort annual scenario (Year 1; considering total projected
survey days across zones), there are 263 projected survey days in Zone
2, while the ``moderate'' effort annual scenario (Year 4) projects 446
survey days in Zone 2. Projected levels of effort presented here
represent expected maxima, and it is possible that actual levels of
effort will be lower, whether due to effects of the economy on industry
activities or other reasons. Please see Figure 3.2-1 of BOEM's PEIS
(BOEM, 2017) for projected potential ranges of survey activity. The
ranges of projected activity level include an upper bound based on
industry capacity in the GOM and a lower bound that accounts for a
number of things that could affect these activities (e.g., marketplace
changes, adjustment of schedules for closures).
Table 2--Summary of Projected Levels of Effort in 24-Hr Survey Days
----------------------------------------------------------------------------------------------------------------
Deep penetration surveys Shallow penetration/HRG surveys
Zone/region -----------------------------------------------------------------------------
Min Mean Max Min Mean Max
----------------------------------------------------------------------------------------------------------------
1 (Shelf east).................... 0 0 0 0 0 1
2 (Shelf central)................. 243 304 426 13 18 21
3 (Shelf west).................... 0 11 30 2 3 4
4 (Slope east).................... 0 63 157 0 1 1
5 (Slope central)................. 414 480 721 26 30 37
6 (Slope west).................... 99 133 256 10 13 14
7 (Deep).......................... 615 678 940 34 40 49
-----------------------------------------------------------------------------
Total......................... 1,547 1,669 2,190 96 105 116
----------------------------------------------------------------------------------------------------------------
Proposed mitigation, monitoring, and reporting measures are
described in detail later in this document (please see ``Proposed
Mitigation'' and ``Proposed Monitoring and Reporting'').
Description of Marine Mammals in the Area of the Specified Activity
Sections 3 and 4 of the petition summarize available information
regarding status and trends, distribution and habitat preferences, and
behavior and life history of the potentially affected species. We refer
the reader to these descriptions, to descriptions of the affected
environment in Appendix E of BOEM's PEIS, as well as to NMFS's Stock
Assessment Reports (SAR; www.fisheries.noaa.gov/national/marine-mammal-protection/marine-mammal-stock-assessments), incorporated here by
reference, instead of reprinting the information. Additional general
information about these species (e.g., physical and behavioral
descriptions) may be found on NMFS's website (www.fisheries.noaa.gov/find-species), the U.S. Navy's Marine Resource Assessment for the GOM
(DoN, 2007a) (available online at: www.navfac.navy.mil/products_and_services/ev/products_and_services/marine_resources/marine_resource_assessments.html), or W[uuml]rsig (2017).
Table 3 lists all species with expected potential for occurrence in
the Gulf of Mexico and summarizes information related to the population
or stock. For taxonomy, we follow Committee on Taxonomy (2017). While
no mortality or serious injury is anticipated or proposed for
authorization, potential biological removal (PBR; defined in the MMPA
as the maximum number of animals, not including natural mortalities,
that may be removed from a marine mammal stock while allowing that
stock to reach or maintain its optimum sustainable population) and
annual serious injury and mortality from anthropogenic sources are
included here as gross indicators of the status of the species and
other threats (as described in NMFS's SARs).
Species that could potentially occur in the proposed survey areas,
but are not reasonably expected to have potential to
[[Page 29225]]
be affected by the specified activity, are described briefly but
omitted from further analysis. These include extralimital species,
which are species that do not normally occur in a given area but for
which there are one or more occurrence records that are considered
beyond the normal range of the species. For status of species, we
provide information regarding U.S. regulatory status under the MMPA and
Endangered Species Act (ESA).
Marine mammal abundance estimates presented in this document
represent the total number of individuals that make up a given stock or
the total number estimated within a particular study area. NMFS's stock
abundance estimates for most species represent the total estimate of
individuals within the geographic area, if known, that comprises that
stock. All managed stocks in this region are assessed in NMFS's U.S.
Atlantic SARs (e.g., Hayes et al., 2017). All values presented in Table
3 are the most recent available at the time of publication and are
available in the 2016 SARs (Hayes et al., 2017) or draft 2017 SARs
(www.fisheries.noaa.gov/national/marine-mammal-protection/draft-marine-mammal-stock-assessment-reports).
In some cases, species are treated as guilds. In general ecological
terms, a guild is a group of species that have similar requirements and
play a similar role within a community. However, for purposes of stock
assessment or abundance prediction, certain species may be treated
together as a guild because they are difficult to distinguish visually
and many observations are ambiguous. For example, NMFS's GOM SARs
assess stocks of Mesoplodon spp. and Kogia spp. as guilds. Here, we
consider beaked whales and Kogia spp. as guilds. In the following
discussion, reference to ``beaked whales'' includes the Cuvier's,
Blainville's, and Gervais beaked whales, and reference to ``Kogia
spp.'' includes both the dwarf and pygmy sperm whale.
Twenty-one species (with 25 managed stocks) have the potential to
co-occur with the proposed survey activities. Extralimital species or
stocks unlikely to co-occur with survey activity include 31 estuarine
bottlenose dolphin stocks (discussed below), the blue whale
(Balaenoptera musculus), fin whale (B. physalus), sei whale (B.
borealis), minke whale (B. acutorostrata), humpback whale (Megaptera
novaeangliae), North Atlantic right whale (Eubalaena glacialis), and
the Sowerby's beaked whale (Mesoplodon bidens). All mysticete species
listed here are considered only of accidental occurrence in GOM and are
generally historically known only from a very small number of
strandings and/or sightings (W[uuml]rsig et al., 2000; W[uuml]rsig,
2017). The blue whale is known from two stranding records, the fin
whale from five strandings and rare sightings, and the sei whale from
five strandings (W[uuml]rsig, 2017). Although North Atlantic right
whales are well known from the east coast of Florida, that area
represents the southern limit of their range; W[uuml]rsig (2017)
reports one stranding and one sighting of two whales in the GOM.
Occasional minke whale strandings and rare sightings near the Florida
Keys show a winter-spring pattern, which may be indicative of
northward-migrating whales from the Caribbean becoming disoriented
(W[uuml]rsig et al., 2000). In 1997, a single group of six humpback
whales was observed approximately 250 km east of the Mississippi River
delta in deep water; however, this sighting as well as other occasional
strandings and rare sighting records are believed to represent vagrants
from the Caribbean (W[uuml]rsig et al., 2000). A Sowerby's beaked whale
was found stranded in western Florida in 1984, a record representing
the lowest known latitude for the species (Bonde and O'Shea, 1989). We
also note here that Hildebrand et al. (2015) report acoustic detections
of an ``as yet unidentified species of beaked whale'' from three sites.
At the three sites--Mississippi Canyon, Green Canyon, and Dry
Tortugas--vocal encounters of the unknown species represented four,
three, and 0.1 percent of total beaked whale vocal encounters. The same
acoustic echolocation signature was previously reported near Hawaii
(but without simultaneous visual and acoustic detection), and would
presumably be a species with tropical distribution (Hildebrand et al.,
2012; McDonald et al., 2009). Nothing else is known of this potential
new species.
Roberts et al. (2016) developed a stratified density model for the
fin whale in the GOM, on the basis of one observation during an aerial
survey in the early 1990s. None of the other extralimital species
listed here were observed during NMFS shipboard or aerial survey effort
from 1992-2009. The fin whale is the second-most frequently reported
mysticete in the GOM (after the Bryde's whale), though with only a
handful of stranding and sighting records, and is considered here as a
rare and likely accidental migrant. As noted by the model authors,
while the probability of a chance encounter is not zero, the single
sighting during NMFS survey effort should be considered extralimital
(Roberts et al., 2015a).
Estuarine stocks of bottlenose dolphin primarily inhabit inshore
waters of bays, sounds, and estuaries (BSE), and stocks are defined
throughout waters adjacent to the specified geographical region.
However, estuarine stock ranges are generally described as including
coastal waters (i.e., waters adjacent to shore, barrier islands, or
presumed outer bay boundaries and outside of typical inshore ranges) to
approximately 1-3 km. For example, bottlenose dolphins that were
captured in Texas and outfitted with radio transmitters largely
remained within the bays, with three individuals tracked to 1 km
offshore (Lynn and W[uuml]rsig, 2002). Radio-tracking of dolphins in
the St. Joseph Bay, Florida area showed that most dolphins stayed
within the bay and that, although some individuals ranged more than 40
km along the coastline from the study site, they never ventured outside
of immediate nearshore waters (Balmer et al., 2008). More recently,
dolphins captured in Barataria Bay, Louisiana were fitted with
satellite-linked transmitters, showing that most dolphins remained
within the bay, while those that entered nearshore coastal waters
remained within 1.75 km (Wells et al., 2017). Therefore, these stocks
would not generally be expected to be impacted by the described
geophysical surveys. If a deep penetration seismic survey were
occurring in nearshore Federal waters (i.e., at least 3 miles from
shore but 9 miles from shore off Texas and Florida), it is possible
that a dolphin belonging to a BSE stock could be affected. However,
such surveys are expected to be rare in such shallow waters, and given
the fact that BSE dolphins in sheltered inshore waters would largely
not be impacted by noise generated offshore, we believe that impacts
from the described activities that could potentially be considered as a
``take'' (as defined by the MMPA) should be considered discountable.
In addition, the West Indian manatee (Trichechus manatus
latirostris) may be found in coastal waters of the GOM. However,
manatees are managed by the U.S. Fish and Wildlife Service and are not
considered further in this document.
[[Page 29226]]
Table 3--Marine Mammals Potentially Present in the Specified Geographical Region
--------------------------------------------------------------------------------------------------------------------------------------------------------
NMFS stock
ESA/ MMPA abundance (CV, Predicted mean (CV)/ Annual M/
Common name Scientific name Stock status; Nmin, most recent maximum abundance PBR SI (CV)
strategic (Y/ abundance survey) 2 \3\ \4\
N) \1\ 8
--------------------------------------------------------------------------------------------------------------------------------------------------------
Order Cetartiodactyla--Cetacea--Superfamily Mysticeti (baleen whales)
--------------------------------------------------------------------------------------------------------------------------------------------------------
Family Balaenopteridae
(rorquals):
Bryde's whale................ Balaenoptera edeni. Gulf of Mexico..... - \5\; Y 33 (1.07; 16; 2009) 44 (0.27)/n/a...... 0.03 0.7
--------------------------------------------------------------------------------------------------------------------------------------------------------
Superfamily Odontoceti (toothed whales, dolphins, and porpoises)
--------------------------------------------------------------------------------------------------------------------------------------------------------
Family Physeteridae:
Sperm whale.................. Physeter GOM................ E/D; Y 763 (0.38; 560; 2,128 (0.08)/2,234. 1.1 0
macrocephalus. 2009).
Family Kogiidae:
Pygmy sperm whale............ Kogia breviceps.... GOM................ -; N 186 (1.04; 90; 2,234 (0.19)/6,117 0.9 0.3 (1.0)
2009) \6\. \6\.
Dwarf sperm whale............ K. sima............ GOM................ -; N ................... ................... ....... .........
Family Ziphiidae (beaked whales):
Cuvier's beaked whale........ Ziphius cavirostris GOM................ -; N 74 (1.04; 36; 2009) 2,910 (0.16)/3,958 0.4 0
\6\.
Gervais beaked whale......... Mesoplodon GOM................ -; N 149 (0.91; 77; ................... 0.8 0
europaeus. 2009) \6\.
Blainville's beaked whale.... M. densirostris.... GOM................ -; N ................... ................... ....... .........
Family Delphinidae:
Rough-toothed dolphin........ Steno bredanensis.. GOM................ -; N 624 (0.99; 311; 4,853 (0.19)/n/a... 3 0.8 (1.0)
2009).
Common bottlenose dolphin.... Tursiops truncatus GOM Oceanic........ -; N 5,806 (0.39; 4,230; 138,602 (0.06)/ 42 6.5
truncatus. 2009). 192,176 \6\. (0.65)
GOM Continental -; N 51,192 (0.10; ................... 469 0.8
Shelf. 46,926; 2011-12).
GOM Coastal, -; N 12,388 (0.13; ................... 111 1.6
Eastern. 11,110; 2011-12).
GOM Coastal, -; N 7,185 (0.21; 6,044; ................... 60 0.4
Northern. 2011-12).
GOM Coastal, -; N 20,161 (0.17; ................... 175 0.6
Western. 17,491; 2011-12).
Clymene dolphin.............. Stenella clymene... GOM................ -; N 129 (1.00; 64; 11,000 (0.16)/ 0.6 0
2009). 12,115.
Atlantic spotted dolphin..... S. frontalis....... GOM................ -; N 37,611 (0.28; 47,488 (0.13)/ Undet. 42 (0.45)
29,844; 2000-01) 85,108.
\7\.
Pantropical spotted dolphin.. S. attenuata GOM................ -; N 50,880 (0.27; 84,014 (0.06)/ 407 4.4
attenuata. 40,699; 2009). 108,764.
Spinner dolphin.............. S. longirostris GOM................ -; N 11,441 (0.83; 13,485 (0.24)/ 62 0
longirostris. 6,221; 2009). 31,341.
Striped dolphin.............. S. coeruleoalba.... GOM................ -; N 1,849 (0.77; 1,041; 4,914 (0.17)/5,323. 10 0
2009).
Fraser's dolphin............. Lagenodelphis hosei GOM................ -; N 726 (0.7; 427; 1996- 1,665 (0.73)/n/a... Undet. 0
2001) \7\.
Risso's dolphin.............. Grampus griseus.... GOM................ -; N 2,442 (0.57; 1,563; 3,137 (0.10)/4,153. 16 7.9
2009). (0.85)
Melon-headed whale........... Peponocephala GOM................ -; N 2,235 (0.75; 1,274; 6,733 (0.30)/7,105. 13 0
electra. 2009).
Pygmy killer whale........... Feresa attenuata... GOM................ -; N 152 (1.02; 75; 2,126 (0.30)/n/a... 0.8 0
2009).
False killer whale........... Pseudorca GOM................ -; N 777 (0.56; 501; 3,204 (0.36)/n/a... Undet. 0
crassidens. 2003-04) \7\.
Killer whale................. Orcinus orca....... GOM................ -; N 28 (1.02; 14; 2009) 185 (0.41)/n/a..... 0.1 0
Short-finned pilot whale..... Globicephala GOM................ -; N 2,415 (0.66; 1,456; 1,981 (0.18)/n/a... 15 0.5 (1.0)
macrorhynchus. 2009).
--------------------------------------------------------------------------------------------------------------------------------------------------------
\1\ ESA status: Endangered (E)/MMPA status: Depleted (D). A dash (-) indicates that the species is not listed under the ESA or designated as depleted
under the MMPA. Under the MMPA, a strategic stock is one for which the level of direct human-caused mortality exceeds PBR or which is determined to be
declining and likely to be listed under the ESA within the foreseeable future. Any species or stock listed under the ESA is automatically designated
under the MMPA as depleted and as a strategic stock.
\2\ NMFS marine mammal stock assessment reports online at: www.fisheries.noaa.gov/national/marine-mammal-protection/marine-mammal-stock-assessments. CV
is coefficient of variation; Nmin is the minimum estimate of stock abundance.
\3\ This information represents species- or guild-specific abundance predicted by habitat-based cetacean density models (Roberts et al., 2016). These
models provide the best available scientific information regarding predicted density patterns of cetaceans in the U.S. Gulf of Mexico, and we provide
the corresponding abundance predictions as a point of reference. Total abundance estimates were produced by computing the mean density of all pixels
in the modeled area and multiplying by its area.
\4\ These values, found in NMFS's SARs, represent annual levels of human-caused mortality plus serious injury from all sources combined (e.g.,
commercial fisheries, ship strike). A CV associated with estimated mortality due to commercial fisheries is presented in some cases.
\5\ NMFS has proposed to list the GOM Bryde's whale as an endangered species under the ESA (81 FR 88639; December 8, 2016).
\6\ Abundance estimates are in some cases reported for a guild or group of species when those species are difficult to differentiate at sea. Similarly,
the habitat-based cetacean density models produced by Roberts et al. (2016) are based in part on available observational data which, in some cases, is
limited to genus or guild in terms of taxonomic definition. NMFS's SARs present pooled abundance estimates for Kogia spp. and Mesoplodon spp., while
Roberts et al. (2016) produced density models to genus level for Kogia spp. and as a guild for beaked whales (Ziphius cavirostris and Mesoplodon
spp.). Finally, Roberts et al. (2016) produced a density model for bottlenose dolphins that does not differentiate between oceanic, shelf, and coastal
stocks.
\7\ NMFS's abundance estimates for these species are greater than eight years old and not considered current. PBR is therefore considered undetermined,
as there is no current minimum abundance estimate for use in calculation. We nevertheless present the most recent abundance estimate.
\8\ We note that Dias and Garrison (2016) present abundance estimates for oceanic stocks that were calculated for use in DWH oil spill injury
quantification. For most stocks, these estimates are based on pooled observations from shipboard surveys conducted in 2003, 2004, and 2009 and
corrected for detection bias. Estimates for beaked whales and Kogia spp. were based on density estimates derived from passive acoustic data collection
(Hildebrand et al., 2012). The abundance estimate for Bryde's whales incorporated the results of additional shipboard surveys conducted in 2007, 2010,
and 2012. Here we retain NMFS's official SARs information for comparison with model-predicted abundance (Roberts et al., 2016).
[[Page 29227]]
For the majority of species potentially present in the specified
geographical region, NMFS has designated only a single generic stock
(i.e., ``Gulf of Mexico'') for management purposes, although there is
currently no information to differentiate the stock from the Atlantic
Ocean stock of the same species, nor information on whether more than
one stock may exist in the GOM (Hayes et al., 2017).
During aerial and ship-based cetacean surveys, the most commonly
sighted species in the GOM are bottlenose dolphins, pantropical spotted
dolphins, Atlantic spotted dolphins, Risso's dolphins, sperm whales,
and Kogia spp. (Baumgartner et al., 2001; Mullin and Fulling, 2004;
Mullin et al., 2004, Maze-Foley and Mullin, 2006; Mullin, 2007; Dias
and Garrison, 2016). Short-finned pilot whales, striped dolphins,
Clymene dolphins, spinner dolphins, and beaked whales are somewhat
commonly observed during surveys and have different rates of detection
(Mullin et al., 2004; Mullin and Fulling, 2004; Dias and Garrison,
2016). Rarely recorded species include melon-headed whales, false
killer whales, killer whales, and pygmy killer whales (Dias and
Garrison, 2016). Bryde's whales are also infrequently seen and are the
only species of baleen whale recurrently seen in the GOM (Baumgartner
et al., 2001; Mullin and Fulling, 2004; Mullin et al., 2004, Maze-Foley
and Mullin, 2006; Mullin, 2007; Dias and Garrison, 2016). Fraser's
dolphins are present in the GOM, but there are very few detections
during marine mammal surveys (Mullin and Fulling, 2004; Dias and
Garrison, 2016).
For the bottlenose dolphin, NMFS defines an oceanic stock, a
continental shelf stock, and three coastal stocks. As in the
northwestern Atlantic Ocean, there are two general bottlenose dolphin
ecotypes: ``coastal'' and ``offshore.'' These ecotypes are genetically
and morphologically distinct (Hoelzel et al., 1998; Waring et al.,
2016), though ecotype distribution is not clearly defined and the
stocks are delineated primarily on the basis of management rather than
ecological boundaries. The offshore ecotype is assumed to correspond to
the oceanic stock, with the stock boundary (and thus the de facto
delineation of offshore and coastal ecotypes) defined as the 200-m
isobath. All genetic samples collected during 1994-2008 in waters
greater than 200 m were of the offshore ecotype (Waring et al., 2016).
The continental shelf stock is defined as between two typical survey
strata: the 20- and 200-m isobaths. While the shelf stock is assumed to
consist primarily of coastal ecotype dolphins, offshore ecotype
dolphins may also be present. There is expected to be some overlap with
the three coastal stocks as well, though the degree is unknown and it
is not thought that significant mixing or interbreeding occurs between
them (Waring et al., 2016). The coastal stocks are defined as being in
waters between the shore, barrier islands, or presumed outer bay
boundaries out to the 20-m isobath and, as a working hypothesis, NMFS
has assumed that dolphins occupying habitats with dissimilar climatic,
coastal, and oceanographic characteristics might be restricted in their
movements between habitats, thus constituting separate stocks (Waring
et al., 2016). Shoreward of the 20-m isobath, the eastern coastal stock
extends from Key West, FL to 84[deg] W longitude; the northern coastal
stock from 84[deg] W longitude to the Mississippi River delta; and the
western coastal stock from the Mississippi River delta to the Mexican
border. The latter is assumed to be a trans-boundary stock, though no
information is available regarding abundance in Mexican waters. Genetic
studies have shown significant differentiation between inshore stocks
and the adjacent coastal stock (Sellas et al., 2005) and among dolphins
living in coastal and shelf waters (Waring et al., 2016), suggesting
that despite spatial overlap there may be mechanisms reducing
interbreeding among coastal stocks and between coastal stocks and BSE
stocks (Waring et al., 2016). Continued studies are necessary to
examine the current stock boundaries delineated in coastal, shelf, and
oceanic waters (Waring et al., 2016).
In Table 3 above, we report two sets of abundance estimates: those
from NMFS's SARs and those predicted by Roberts et al. (2016)--for the
latter we provide both the annual mean and the monthly maximum (where
applicable). Please see footnotes 2-3 for more detail. NMFS's SAR
estimates are typically generated from the most recent shipboard and/or
aerial surveys conducted. GOM oceanography is dynamic, and the spatial
scale of the GOM is small relative to the ability of most cetacean
species to travel. As an example, no groups of Fraser's dolphins were
observed during dedicated cetacean abundance surveys during 2003-2004
or 2009, yet NMFS states that it is probable that Fraser's dolphins
were present in the northern GOM but simply not encountered, and
therefore declines to present an abundance estimate of zero (Waring et
al., 2013). U.S. waters only comprise about 40 percent of the entire
GOM, and 65 percent of GOM oceanic waters are south of the U.S. EEZ.
Studies based on abundance and distribution surveys restricted to U.S.
waters are unable to detect temporal shifts in distribution beyond U.S.
waters that might account for any changes in abundance within U.S.
waters. NMFS's SAR estimates also typically do not incorporate
correction for detection bias. Therefore, they should generally be
considered as underestimates, especially for cryptic or long-diving
species (e.g., beaked whales, Kogia spp., sperm whales). Dias and
Garrison (2016) state, for example, that current abundance estimates
for Kogia spp. may be considerably underestimated due to the cryptic
behavior of these species and difficulty of detection in Beaufort sea
state greater than one, and density estimates for certain species
derived from long-term passive acoustic monitoring are much higher than
are estimates derived from visual observations (Mullin and Fulling,
2004; Mullin, 2007; Hildebrand et al., 2012).
The Roberts et al. (2016) abundance estimates represent the output
of predictive models derived from multi-year observations and
associated environmental parameters and which incorporate corrections
for detection bias. Incorporating more data over multiple years of
observation can yield different results in either direction, as the
result is not as readily influenced by fine-scale shifts in species
habitat preferences or by the absence of a species in the study area
during a given year. NMFS's abundance estimates show substantial year-
to-year variability in some cases. For example, NMFS-reported estimates
for the Clymene dolphin vary by a maximum factor of more than 100 (2009
estimate of 129 versus 1996-2001 estimate of 17,355), indicating that
it may be more appropriate to use the model prediction versus a point
estimate, as the model incorporates data from 1992-2009. The latter
factor--incorporation of correction for detection bias--should
systematically result in greater abundance predictions. For these
reasons, we expect that the Roberts et al. (2016) estimates are
generally more realistic and, for these purposes, represent the best
available information. For purposes of assessing estimated exposures
relative to abundance--used in this case to understand the scale of the
predicted takes compared to the population--we generally believe that
the Roberts et al. (2016) abundance predictions are most appropriate
because they were used to generate the exposure estimates and therefore
[[Page 29228]]
provide the most relevant comparison. Roberts et al. (2016) represents
the best available scientific information regarding marine mammal
occurrence and distribution in the Gulf of Mexico.
As a further illustration of the distinction between the SARs and
model-predicted abundance estimates, the current NMFS stock abundance
estimates for most GOM species are based on direct observations from
shipboard surveys conducted in 2009 (from the 200-m isobath to the edge
of the U.S. EEZ) and not corrected for detection bias, whereas the
exposure estimates presented herein for those species are based on the
abundance predicted by a density surface model informed by observations
from surveys conducted over approximately 20 years and covariates
associated at the observation level. To directly compare the estimated
exposures predicted by the outputs of the Roberts et al. (2016) model
to NMFS's SAR abundance would therefore not be meaningful.
Biologically Important Areas (BIA)--As part of our description of
the environmental baseline, we discuss any known areas of importance as
marine mammal habitat. These areas may include designated critical
habitat for ESA-listed species (as defined by section 3 of the ESA) or
other known areas not formally designated pursuant to any statute or
other law. Important areas may include areas of known importance for
reproduction, feeding, or migration, or areas where small and resident
populations are known to occur.
Although there is no designated critical habitat for marine mammal
species in the specified geographical region, BIAs for marine mammals
are recognized. For example, the GOM Bryde's whale is a very small
population that is genetically distinct from other Bryde's whales and
not genetically diverse within the GOM (Rosel and Wilcox, 2014).
Further, the species is typically observed only within a narrowly
circumscribed area within the eastern GOM. Therefore, this area is
described as a year-round BIA by LaBrecque et al. (2015). Although
survey effort has covered all oceanic waters of the U.S. GOM, whales
were observed only between approximately the 100- and 300-m isobaths in
the eastern GOM from the head of the De Soto Canyon (south of
Pensacola, Florida) to northwest of Tampa Bay, Florida (Maze-Foley and
Mullin, 2006; Waring et al., 2016; Rosel and Wilcox, 2014; Rosel et
al., 2016). NOAA subsequently conducted a status review of the GOM
Bryde's whale. The review, described in a technical memorandum (Rosel
et al. (2016)), expanded this description by stating that, due to the
depth of some sightings, the area is more appropriately defined to the
400-m isobath and westward to Mobile Bay, Alabama, in order to provide
some buffer around the deeper sightings and to include all sightings in
the northeastern GOM. However, the recorded Bryde's whale shipboard and
aerial survey sightings between 1989 and 2015 have mainly fallen within
the BIA described by LaBreque et al. (2015).
LaBrecque et al. (2015) also described eleven year-round BIAs for
small and resident BSE bottlenose dolphin populations in the GOM.
Additional study would likely allow for identification of additional
BIAs associated with other GOM BSE dolphin stocks.
Unusual Mortality Events (UME)--A UME is defined under Section
410(6) of the MMPA as ``a stranding that is unexpected; involves a
significant die-off of any marine mammal population; and demands
immediate response.'' From 1991 to the present, there have been twelve
formally recognized UMEs affecting marine mammals in the region and
involving species under NMFS's jurisdiction. These have primarily
impacted coastal bottlenose dolphins, with multiple UMEs determined to
have resulted from biotoxins and one from infectious disease. None of
these involve ongoing investigation. Most significantly, a UME
affecting multiple cetacean species in the northern GOM occurred from
2010-2014.
The northern GOM UME was determined to have begun in March 2010 and
extended through July 2014. The event included all cetaceans stranded
during this time in Alabama, Mississippi, and Louisiana and all
cetaceans other than bottlenose dolphins stranded in the Florida
Panhandle (Franklin County through Escambia County), with a total of
1,141 cetaceans stranded or reported dead offshore. For reference, the
same area experienced a normal average of 75 strandings per year from
2002-09 (Litz et al., 2014). The majority of stranded animals were
bottlenose dolphins, though at least ten additional species were
reported as well. Since not all cetaceans that die wash ashore where
they may be found, the number reported stranded is likely a fraction of
the total number of cetaceans that died during the UME. There was also
an increase in strandings of stillborn and newborn dolphins (Colegrove
et al., 2016).
The UME investigation and the Deepwater Horizon Natural Resource
Damage Assessment (described below) determined that the DWH oil spill
is the most likely explanation of the persistent, elevated stranding
numbers in the northern GOM after the 2010 spill. The evidence to date
supports that exposure to hydrocarbons released during the DWH oil
spill was the most likely explanation of adrenal and lung disease in
dolphins, which has contributed to increased deaths of dolphins living
within the oil spill footprint and increased fetal loss. The longest
and most prolonged stranding cluster was in Barataria Bay, Louisiana in
2010-11, followed by Mississippi and Alabama in 2011, consistent with
timing and spatial distribution of oil, while the number of deaths was
not elevated for areas that were not as heavily oiled.
However, increased dolphin strandings occurred in Louisiana and
Mississippi before the DWH oil spill, and identified stranding clusters
within the UME suggest that the event may involve different additional
contributing factors varying by location, time, and population (Venn-
Watson et al., 2015a). Some previous GOM cetacean UMEs had included
environmental influences (e.g., low salinity due to heavy rainfall and
associated runoff of land-based pesticides, low temperatures) as
possible contributing factors (Litz et al., 2014). Low air and water
temperatures occurred in the spring of 2010 throughout the GOM prior to
and during the start of the UME, and a portion of the pre-spill
atypical strandings occurred in Lake Pontchartrain, Louisiana,
concurrent with lower than average salinity (Mullin et al., 2015).
Therefore, a large part of the pre-spill increased dolphin strandings
may have been due to a combination of cold temperatures and low
salinity (Litz et al., 2014).
Subsequent health assessments of live dolphins from Barataria Bay
and comparison to a reference population found significantly increased
adrenal disease, lung disease, and poor health, while histological
evaluations of samples from dead stranded animals from within and
outside the UME area found that UME animals were more likely to have
lung and adrenal lesions and to have primary bacterial pneumonia, which
caused or contributed significantly to death (Schwacke et al., 2014a,
2014b; Venn-Watson et al., 2015b). In order to diagnose health, dolphin
capture-release health assessments were conducted in Barataria Bay,
during which physical examinations, including weighing and morphometric
measurements, were conducted, routine biological samples (e.g., blood,
tissue) were obtained, and animals were examined with ultrasound.
Veterinarians then reviewed
[[Page 29229]]
the findings and determined an overall prognosis for each animal (e.g.,
favorable outcome expected, outcome uncertain, unfavorable outcome
expected). Almost half of the examined animals were given a guarded or
worse prognosis, and 17 percent were not expected to survive (Schwacke
et al., 2014a).
The prevalence of brucellosis and morbillivirus infections was low
and biotoxin levels were low or below the detection limit, meaning that
these were not likely primary causes of the UME (Venn-Watson et al.,
2015b; Fauquier et al., 2017). Subsequent study found that persistent
organic pollutants (e.g., polychlorinated biphenyls), which are
associated with endocrine disruption and immune suppression when
present in high levels, are likely not a primary contributor to the
poor health conditions and increased mortality observed in these GOM
populations (Balmer et al., 2015). The chronic adrenal gland and lung
diseases identified in stranded UME dolphins are consistent with
exposure to petroleum compounds (Venn-Watson et al., 2015b). Colegrove
et al. (2016) found that the increase in perinatal strandings resulted
from late-term pregnancy failures and development of in utero
infections likely caused by chronic illnesses in mothers who were
exposed to oil.
While the number of dolphin mortalities in the area decreased after
the peak from March 2010-July 2014, it does not indicate that the
effects of the oil spill on these populations have ended. Researchers
still saw evidence of chronic lung disease and adrenal impairment four
years after the spill (in July 2014) and saw evidence of failed
pregnancies in 2015 (Smith et al., 2017). These follow-up studies found
a yearly mortality rate for Barataria Bay dolphins of roughly 13
percent (as compared to annual mortality rates of 5 percent or less
that have been previously reported for other dolphin populations), and
found that only 20 percent of pregnant dolphins produced viable calves
(compared with 83 percent in a reference population) (Lane et al.,
2015; McDonald et al., 2017). Research into the long-term health
effects of the spill on marine mammal populations is ongoing. For more
information on the UME, please visit www.nmfs.noaa.gov/pr/health/mmume/cetacean_gulfofmexico.htm.
Prior UMEs averaged six months in duration and involved
significantly fewer mortalities. In most of these relatively localized
events, dolphin morbillivirus or brevetoxicosis was confirmed or
suspected as a causal factor (Litz et al., 2014). One other recent UME
occurred during 2011-12 for bottlenose dolphins in Texas. Investigators
were not able to determine a cause for the UME, though findings
included lung infection, poor body condition, and discoloring of teeth.
No connection has been identified between this event and the 2010-14
event described above. For more information on UMEs, please visit:
www.fisheries.noaa.gov/national/marine-life-distress/marine-mammal-unusual-mortality-events.
Deepwater Horizon Oil Spill
We introduced the DWH oil spill--which includes the impacts of the
spill as well as the response efforts--previously in our description of
the ``Specified Geographical Region.'' Here we provide additional
description of the potential effects of the spill on the marine mammals
that may be affected by the activities that are the subject of this
proposed rule. The summary provided below is an incorporation by
reference of relevant information from DWH NRDA Trustees (2016) and DWH
MMIQT (2015); more detail on the DWH oil spill and its effects on
marine mammals is available in these documents. Additional technical
reports relating to the assessment of marine mammal injury due to the
DWH oil spill are available online at: www.doi.gov/deepwaterhorizon/adminrecord. A brief overview of injury assessment activities and
associated findings is provided by Wallace et al., (2017).
On April 20, 2010, the Deepwater Horizon offshore drilling
platform, a semi-submersible exploratory drilling rig operating on the
exploratory Macondo well (within BOEM's Mississippi Canyon lease
block), exploded and subsequently sank in 1,522 m of water in the GOM,
approximately 81 km off the coast of Louisiana. This incident resulted
in the release of an estimated 3.19 million barrels (134 million
gallons) of oil from the compromised well. In addition, approximately
1.84 million gallons of chemical dispersants were applied to the waters
of the spill area. The release of oil continued for 87 days, with an
average of more than 1.5 million gallons of fresh oil entering the
ocean per day--essentially creating a new major oil spill every day for
nearly 3 months, equivalent to the 1989 Exxon Valdez oil spill re-
occurring in the same location every week for the duration. Response
techniques included deployment of containment booms, physical removal
of oil, controlled burning of oil on the surface, major releases of
fresh water to keep the oil offshore, beach and fishery closures,
construction of berms, wildlife rehabilitation and relocation (e.g.,
Wilkin et al., 2017), and application of chemical dispersants on the
surface and at the wellhead on the seafloor (with the goal of breaking
the oil into small droplets). For more information about the DWH oil
spill, please visit response.restoration.noaa.gov/deepwater-horizon-oil-spill and www.deepwaterhorizoneconomicsettlement.com/docs.php.
An estimated 7.7 billion standard cubic feet of natural gas was
released in association with the oil; bacteria proliferated, consumed
the gas, and died. Mucus produced by bacteria, as well as some of the
bacterial mass itself, agglomerated with brown-colored oil droplets and
settled through the water column--this phenomenon is referred to as
``marine oil snow.'' Oil, released from the well-head approximately
1,500 m deep, moved with currents, creating a plume of oil within the
deep sea; oil and associated ``marine oil snow'' also settled on the
sea floor. More buoyant oil traveled up through the water column and
formed large surface slicks; at its maximum extent, oil covered over
40,000 km\2\ of ocean. Cumulatively, over the course of the spill, oil
was detected on over 112,000 km\2\ of ocean. Figure 3 shows the
cumulative area of detectable surface oil slick during the DWH oil
spill. Currents, winds, and tides carried these surface oil slicks to
shore, fouling more than 2,100 km of shoreline, including beaches,
bays, estuaries, and marshes from eastern Texas to the Florida
Panhandle. In addition, some lighter oil compounds evaporated from the
slicks, exposing air-breathing organisms like marine mammals to noxious
fumes at the sea surface. Air pollution resulted from compounds in the
oil that evaporated into the air and from fires purposely started to
burn off oil at the ocean surface. The oil released during the event
was a complex mixture containing thousands of individual chemical
compounds--many of which are known to be toxic to biota--which then
changed as they were subject to natural processes such as mixing with
air and water, microbial degradation, and exposure to sunlight. DWH oil
has a specific chemical signature that, together with other lines of
evidence, allowed investigators to determine which oil-derived
contaminants found in the environment originated from the spill.
Dispersants are chemicals that reduce the tension between oil and
water, leading to the formation of oil droplets that more readily
disperse within the water column. A main purpose of using dispersants
is to enhance the rate at
[[Page 29230]]
which bacteria degrade the oil in order to prevent oil slicks from
fouling sensitive shoreline habitats. The large-scale use of
dispersants raised concerns about the potential for toxic effects of
dispersed oil in the water column, as well as the potential for hypoxia
due to bacterial consumption of dispersed oil. The surface application
of dispersants increased exposure of near-surface biota, such as marine
mammals, to oil that re-entered the water column.
[GRAPHIC] [TIFF OMITTED] TP22JN18.002
The DWH oil spill was subject to the provisions of the Oil
Pollution Act (OPA) of 1990 (33 U.S.C. 2701 et seq.), which addresses
prevention, response, and compensation for oil pollution incidents in
navigable waters, adjoining shorelines, and the U.S. EEZ. Under the
authority of OPA, a council of Federal and state trustees was
established, on behalf of the public, to assess natural resource
injuries resulting from the incident and work to make the environment
and public whole for those injuries. As required under OPA, the
trustees conducted a natural resource damage assessment (NRDA), finding
that the injuries resulting from the DWH oil spill affected such a wide
array of linked resources over such an enormous area that the effects
must be described as constituting an ecosystem-level injury. OPA
regulations (15 CFR part 990) establish a process for conducting a NRDA
that require, in part, the assessment of potential injuries to relevant
resources, here including marine mammals and habitats they rely upon.
OPA regulations define injury as an observable or measurable adverse
change in a natural resource that may occur directly or indirectly.
Types of injuries include adverse changes in survival, growth, and
reproduction; health, physiology and biological condition; behavior;
community composition; ecological processes and functions; and physical
and chemical habitat quality or structure.
The injury assessment first requires a determination of whether an
incident injured natural resources. Trustees must establish that a
pathway existed from the oil discharge to the resource, confirm that
resources were exposed to the discharge, and evaluate the adverse
effects that occurred as a result of the exposure (or response
activities). Subsequently, the assessment requires injury
quantification (including degree and spatiotemporal extent),
essentially by comparing the post-event conditions with the pre-event
baseline. For a fuller overview of the injury assessment process in
this case, please see Takeshita et al. (2017). Because of the vast
scale of the incident, the trustees evaluated injuries to a set of
representative habitats, communities, species, and ecological
processes, with studies conducted at many scales. Key findings are as
follows: (1) Oil flowed within deep ocean water currents hundreds of
miles away from the well and moved upwards and across a very large area
of the ocean surface, affecting vast areas overall (e.g., approximately
112,000 km\2\ of ocean surface; 2,100 km of shoreline; and between
1,000-1,900 km\2\ of seafloor), including every type of habitat
occupied by marine mammals in the northern GOM as well as habitat for
all stocks of marine mammals in the northern GOM; (2) the oil that was
released was toxic to a wide range of organisms, including marine
mammals; (3) oil came into contact with and injured a wide range of
organisms, including marine mammals; (4)
[[Page 29231]]
response activities had collateral impacts on the environment; and (5)
exposure to oil and response activities resulted in extensive injuries
to multiple habitats, species, and ecological functions, across broad
geographic regions. Critical pathways of exposure for marine mammals
included the contaminated water column, where they swim and capture
prey; the surface slick at the air to water interface, where they
breathe, rest, and swim; and contaminated sediment, where they forage
and capture prey. Response workers and scientists witnessed 85
instances of marine mammals (with a total of 1,394 individuals)
swimming in surface oil or with oil on their bodies; these instances
represented a minimum of 11 species, including dolphins, sperm whales,
Kogia spp., and a beaked whale.
The marine mammal injury assessment synthesized data from NRDA
field studies, stranded carcasses collected by the Southeast Marine
Mammal Stranding Network, historical data on marine mammal populations,
NRDA toxicity testing studies, and the published literature. DWH oil
was found to cause problems with the regulation of stress hormone
secretion from adrenal cells and kidney cells, which will affect an
animal's ability to regulate body functions and respond appropriately
to stressful situations, thus leading to reduced fitness. Bottlenose
dolphins living in habitats contaminated with DWH oil showed signs of
adrenal dysfunction, and dead, stranded dolphins from areas
contaminated with DWH oil had smaller adrenal glands (Schwacke et al.,
2014a; Venn-Watson et al., 2015b). Limited cetacean exposure studies
have demonstrated that bottlenose dolphins may sustain liver damage and
that bottlenose dolphins and sperm whales may develop skin lesions
(Engelhardt, 1983). Field and laboratory studies and other data
analysis were designed to explicitly examine other potential
explanations for marine mammal injuries, including biotoxins,
infectious diseases, human and fishery interactions, and other
unrelated potential contaminants. Each of these other factors was ruled
out as a primary cause for the high prevalence of adverse health
effects, reproductive failures, and disease in stranded animals. When
all of the data are considered together, the DWH oil spill is the only
reasonable cause for the full suite of observed adverse health effects.
Findings related to bottlenose dolphins living in heavily oiled
nearshore habitats were described previously in the UME discussion. Due
to the difficulty of investigating marine mammals in pelagic
environments and across the entire region impacted by the event, the
injury assessment focused on health assessments conducted on bottlenose
dolphins in nearshore habitats (i.e., Barataria Bay and Mississippi
Sound) and used these populations as case studies for extrapolating to
coastal and oceanic populations that received similar or worse exposure
to DWH oil, with appropriate adjustments made for differences in
behavior, anatomy, physiology, life histories, and population dynamics
among species. Based on direct observation, injuries were quantified
for four BSE stocks of bottlenose dolphin, e.g., for the Barataria Bay
stock, the DWH oil spill caused 35 percent (CI 15-49) excess mortality,
46 percent (CI 21-65) excess failed pregnancies, and a 37 percent (CI
14-57) higher likelihood that animals would have adverse health
effects. The process for assigning a health prognosis (Schwacke et al.,
2014a) was described previously in the UME discussion. Two dolphins
having received the lowest grade died within 6 months, and the
percentage of the population with the two lowest prognoses (17 percent
poor and grave) essentially predicted the percentage of dolphins that
disappeared and presumably died the following year based on photo-
identification surveys.
Investigators then used a population modeling approach to capture
the overlapping and synergistic relationships among the three metrics
for injury, and to quantify the entire scope of DWH marine mammal
injury to populations into the future, expressed as ``lost cetacean
years'' due to the DWH oil spill (which represents years lost due to
premature mortality as well as the resultant loss of reproductive
output). This approach allowed for consideration of long-term impacts
resulting from immediate losses and reproductive failures in the few
years following the spill, as well as expected persistent impacts on
survival and reproduction for exposed animals well into the future
(Takeshita et al., 2017). For example, lost cetacean years were
estimated for the Barataria Bay stock of bottlenose dolphins, leading
to an estimated 51 percent (CI 32-72) maximum reduction in population
size and a time to recovery of 39 years (CI 24-80) in the absence of
potential benefits of restoration activities. For a more detailed
overview of the injury quantification for these stocks and their post-
DWH population trajectory, please see Schwacke et al. (2017), and for
full details of the overall injury quantification, see DWH MMIQT
(2015).
To calculate the increase in percent mortality for the shelf and
oceanic marine mammal stocks, the Barataria Bay percent mortality was
applied to the percentage of animals in each stock that was exposed to
oil. This percentage was calculated assuming that animals experiencing
a level of cumulative surface oiling similar to or greater than that in
Barataria Bay would have been likely to suffer a similar or greater
degree and magnitude of injury. This is likely a conservative estimate
of impacts, because: (1) Shelf and oceanic species experienced long
exposures (up to 90 days) to very high concentrations of fresh oil and
a diverse suite of response activities, while estuarine dolphins were
not exposed until later in the spill period and to weathered oil
products at lower water concentrations; (2) oceanic cetaceans dive
longer and to deeper depths, and it is possible that the types of lung
injuries observed in estuarine dolphins may be more severe for oceanic
cetaceans; and (3) cetaceans in deeper waters were exposed to very high
concentrations of volatile gas compounds at the water's surface near
the wellhead.
As an example of the calculation, 47 percent of the spinner dolphin
stock range in the northern GOM experienced oiling equal to or greater
than Barataria Bay, and, therefore, was assumed to have experienced a
rate of mortality increase equal to that calculated for Barataria Bay
(35 percent). Thus, the entire northern GOM spinner dolphin stock is
assumed to have experienced a 16 percent mortality increase (0.35 x
0.47 = 0.16). Similarly, the percentage of females with reproductive
failure in Barataria Bay and Mississippi Sound (46 percent; stocks
pooled for sample size considerations) is considered to be the best
estimate of excess failed pregnancies for other marine mammals in the
oil spill footprint, and the percentage of the population with a
guarded or worse health prognosis--compared with dolphins sampled in a
healthy reference population--from Barataria Bay (37 percent) was
applied to other stocks.
The population modeling approach used in the injury quantification
allows consideration of long-term impacts resulting from individual
losses, adverse reproductive effects, and persistent impacts on
survival for exposed animals. The model was run using baseline
mortality and reproductive parameters to determine what the population
trajectory of each stock would have been if the DWH spill had not
happened. The same model was then run a second time, with estimates for
excess mortality, reproductive
[[Page 29232]]
failures, and adverse health effects due to the DWH oil spill. The
number of years predicted for the DWH oil-impacted population to
recover (without active restoration) is the number of years until the
DWH oil-injured population trajectory reaches 95 percent of the
baseline population trajectory, reported as years to recovery. The
output from the population model also predicts the largest proportional
decrease in population size (i.e., the difference between the two
population trajectories when the DWH oil-impacted trajectory is at its
lowest point). A separate population model is run for each stock, with
inputs for the models restricted to the available data for each stock.
For inputs without empirical data, the values are extrapolated from
other stocks or incorporate additional modeling efforts. For bottlenose
dolphins, uncertainty in model output was evaluated by drawing from the
distributions for model input parameters to execute 10,000 simulations,
producing distributions for each of the model outputs. For other
species, because there was insufficient information to construct
informed input parameter distributions, only a single model scenario
was run using point estimates for input parameter values and
simulations were not conducted to explore the effects of uncertainty in
the model parameters.
The results of these calculations for each affected shelf and
oceanic stock, and for northern and western coastal stocks of
bottlenose dolphin, are presented in Table 4. The eastern coastal stock
of bottlenose dolphin was considered to be not affected by the DWH oil
spill, as the cumulative footprint of oil did not overlap the stock's
range. Results for BSE dolphin stocks are not presented here. No
analysis was performed for Fraser's dolphins or killer whales; although
they are present in the GOM, sightings are rare and there were no
historical sightings in the oil spill footprint during the surveys used
in the quantification process. These stocks were likely injured, but no
information is available on which to base a quantification effort.
Table 4--Summary of Modeled Effects of DWH Oil Spill
----------------------------------------------------------------------------------------------------------------
%
% % Females Population
Population % with with % Maximum Years to
Common name exposed to Population reproductive adverse population recovery
oil (95% killed (95% failure (95% health reduction (95% CI)
CI) CI) CI) effects (95% CI) \b\
(95% CI)
----------------------------------------------------------------------------------------------------------------
Bryde's whale................... 48 (23-100) 17 (7-24) 22 (10-31) 18 (7-28) -22 69
Sperm whale..................... 16 (11-23) 6 (2-8) 7 (3-10) 6 (2-9) -7 21
Kogia spp....................... 15 (8-29) 5 (2-7) 7 (3-10) 6 (2-9) -6 11
Beaked whales................... 12 (7-22) 4 (2-6) 5 (3-8) 4 (2-7) -6 10
Rough-toothed dolphin........... 41 (16-100) 14 (6-20) 19 (9-26) 15 (6-23) -17 54
Bottlenose dolphin, oceanic..... 10 (5-10) 3 (1-5) 5 (2-6) 4 (1-6) -4 n/a
Bottlenose dolphin, northern 82 (55-100) 38 (26-58) 37 (17-53) 30 (11-47) -50 (32-73) 39 (23-76)
coastal........................
Bottlenose dolphin, western 23 (16-32) 1 (1-2) 10 (5-15) 8 (3-13) -5 (3-9) n/a
coastal........................
Shelf dolphins \a\.............. 13 (9-19) 4 (2-6) 6 (3-8) 5 (2-7) -3 n/a
Clymene dolphin................. 7 (3-15) 2 (1-4) 3 (2-5) 3 (1-4) -3 n/a
Pantropical spotted dolphin..... 20 (15-26) 7 (3-10) 9 (4-13) 7 (3-11) -9 39
Spinner dolphin................. 47 (24-91) 16 (7-23) 21 (10-30) 17 (6-27) -23 105
Striped dolphin................. 13 (8-22) 5 (2-7) 6 (3-9) 5 (2-8) -6 14
Risso's dolphin................. 8 (5-13) 3 (1-4) 3 (2-5) 3 (1-4) -3 n/a
Melon-headed whale.............. 15 (6-36) 5 (2-7) 7 (3-10) 6 (2-9) -7 29
Pygmy killer whale.............. 15 (7-33) 5 (2-8) 7 (3-10) 6 (2-9) -7 29
False killer whale.............. 18 (7-48) 6 (3-9) 8 (4-12) 7 (3-11) -9 42
Short-finned pilot whale........ 6 (4-9) 2 (1-3) 3 (1-4) 2 (1-3) -3 n/a
----------------------------------------------------------------------------------------------------------------
Modified from DWH NRDA Trustees (2016).
CI = confidence interval. No CI was calculated for population reduction or years to recovery for shelf or
oceanic stocks.
\a\ ``Shelf dolphins'' includes Atlantic spotted dolphins and the shelf stock of bottlenose dolphins (20-200 m
water depth). These two species were combined because the abundance estimate used in population modeling was
derived from aerial surveys and the species could not generally be distinguished from the air.
\b\ It is not possible to calculate YTR for stocks with maximum population reductions of less than or equal to 5
percent.
Coastal and oceanic marine mammals were injured by exposure to oil
from the DWH spill; nearly all of the stocks that overlap with the oil
spill footprint have demonstrable, quantifiable injuries, and the
remaining stocks (for which there is no quantifiable injury) were also
likely injured, though there is not currently enough information to
make a determination. Injuries included elevated mortality rates,
reduced reproduction, and disease. Due to these effects, affected
populations may require decades to recover absent successful efforts at
restoration (e.g., DWH NRDA Trustees, 2017). Tens of thousands of
marine mammals were exposed to the DWH surface slick, where they
inhaled, aspirated, ingested, and came into contact with oil components
(Dias et al., 2017). The oil's physical and toxic effects damaged
tissues and organs, leading to a constellation of adverse health
effects, including reproductive failure, adrenal disease, lung disease,
and poor body condition, as observed in bottlenose dolphins (De Guise
et al., 2017; Kellar et al., 2017). Coastal and estuarine bottlenose
dolphin populations were some of the most severely injured (Hohn et
al., 2017; Rosel et al., 2017; Thomas et al., 2017), as described
previously in relation to the UME, but oceanic species were also
exposed and experienced increased mortality, increased reproductive
failure, and a higher likelihood of other adverse health effects.
Due to the scope of the spill, the magnitude of potentially injured
populations, and the difficulties and limitations of working with
marine mammals, it is impossible to quantify injury without
uncertainty. Wherever possible, the quantification results represent
ranges of values that encapsulate the uncertainty inherent in the
underlying datasets. The population model outputs shown in Table 4 best
represent the temporal magnitude of the injury and the potential
recovery time from the injury.
Aside from the heavily impacted stocks of bottlenose dolphin, two
species of particular concern are the sperm whale and Bryde's whale.
For the Bryde's whale, it was estimated that 48 percent of the
population was impacted by DWH oil, resulting in an estimated 22
percent maximum decline in population size that will require 69 years
to recovery. However, small populations are highly susceptible to
[[Page 29233]]
stochastic, or unpredictable, processes and genetic effects that can
reduce productivity and resiliency to perturbations. The population
models do not account for these effects, and, therefore, the capability
of the Bryde's whale population to recover from this injury is unknown.
For the sperm whale, a 7 percent maximum decline in population size
requiring 21 years to recovery was predicted. However, little is known
about the fate and transport of DWH deep-sea oil plumes in relation to
deep-diving marine mammals, such as sperm whales, and the results
should be viewed with caution. Other stocks with particularly
concerning results include the rough-toothed dolphin and spinner
dolphin (Table 4).
In the absence of active (and effective) restoration, marine mammal
stocks across the northern GOM will take many years to recover (Table
4). Marine mammals are slow to reach reproductive maturity, only give
birth to a single offspring every 3 to 5 years, and are generally long
lived (with lifespans up to 80 years). Two populations of killer whales
suffered losses of 33 and 41 percent in the year following the Exxon
Valdez oil spill in Alaska, and recovery of both populations has been
unexpectedly slow (Matkin et al., 2008). Persistent pollutant exposure
(Ylitalo et al., 2001), decline of a primary prey source (Ver Hoef and
Frost, 2003), and disruption of social groups (Matkin et al., 2008;
Wade et al., 2012) may be contributing factors. Populations of dolphins
depleted as the result of tuna fishery bycatch in the eastern tropical
Pacific also demonstrated slower than expected rates of recovery, which
may be due in part to the continued effects of stressful interactions
with the fishery (Gerrodette and Forcada, 2005). The ability of the
stocks to recover and the length of time required for that recovery are
tied to the carrying capacity of the habitat, and to the degree of
other population pressures. We treat the effects of the DWH oil spill
as part of the environmental baseline in considering the likely
resilience of these populations to the effects of the activities
considered in this proposed regulatory framework.
In addition to injuries from direct exposure to DWH oil, marine
mammal habitat was degraded. Exposure to oil at or near the surface
occurred in an area of high biological abundance and high productivity
during a time of year (spring and summer) that corresponds with peaks
in seasonal productivity in the northern GOM. Developing fish larvae
exposed to the surface slick suffered almost 100 percent mortality, and
oil concentrations at different levels in the water column exceeded
levels known to cause mortality and sub-lethal effects to fish--this is
expected to have caused the loss of millions to billions of fish that
would have reached one year of age. However, though damage to fish and
invertebrate populations was likely significant during the time oil was
present, populations of directly affected fish and invertebrate species
appear not to have suffered a lasting impact. Although marine mammals
were harmed through the effects of DWH oil on plankton, fish, and
invertebrate populations, it is difficult to interpret any long-term
impacts on marine mammal populations resulting from significant short-
term impacts on prey populations. Prey reductions, when they occur, can
have cascading effects on larger species. Animals in the wild live in a
dynamic relationship with their environment and available resources,
balancing energy expenditures and nutritional uptake in order to
survive, remain healthy, and reproduce. Any impact that shifts that
balance by diminishing food resources or requiring unusual expenditures
of energy--whether to acquire prey, avoid predators, fight disease and
infection, or successfully reproduce--is inherently harmful to the
species. Additionally, as noted previously, injury due to the DWH oil
spill is considered an ecosystem-level event, which will impact marine
mammals in particular due to their long lives and position as apex
predators reliant upon a healthy ecosystem (e.g., Moore, 2008; Bossart,
2011).
Marine Mammal Hearing
Hearing is the most important sensory modality for marine mammals
underwater, and exposure to anthropogenic sound can have deleterious
effects. To appropriately assess the potential effects of exposure to
sound, it is necessary to understand the frequency ranges marine
mammals are able to hear. Current data indicate that not all marine
mammal species have equal hearing capabilities (e.g., Richardson et
al., 1995; Wartzok and Ketten, 1999; Au and Hastings, 2008). To reflect
this, Southall et al. (2007) recommended that marine mammals be divided
into functional hearing groups based on directly measured or estimated
hearing ranges on the basis of available behavioral response data,
audiograms derived using auditory evoked potential techniques,
anatomical modeling, and other data. Note that no direct measurements
of hearing ability have been successfully completed for mysticetes
(i.e., low-frequency cetaceans). Subsequently, NMFS (2016) described
generalized hearing ranges for these marine mammal hearing groups.
Generalized hearing ranges were chosen based on the approximately 65 dB
threshold from the normalized composite audiograms, with an exception
for lower limits for low-frequency cetaceans where the result was
deemed to be biologically implausible and the lower bound from Southall
et al. (2007) retained. The functional groups and the associated
frequencies are indicated below (note that these frequency ranges
correspond to the range for the composite group, with the entire range
not necessarily reflecting the capabilities of every species within
that group):
Low-frequency cetaceans (mysticetes): Generalized hearing
is estimated to occur between approximately 7 Hz and 35 kHz, with best
hearing estimated to be from 100 Hz to 8 kHz;
Mid-frequency cetaceans (larger toothed whales, beaked
whales, and most delphinids): Generalized hearing is estimated to occur
between approximately 150 Hz and 160 kHz, with best hearing from 10 to
less than 100 kHz;
High-frequency cetaceans (porpoises, river dolphins, and
members of the genera Kogia and Cephalorhynchus; including two members
of the genus Lagenorhynchus, on the basis of recent echolocation data
and genetic data): Generalized hearing is estimated to occur between
approximately 275 Hz and 160 kHz.
For more detail concerning these groups and associated frequency
ranges, please see NMFS (2016) for a review of available information.
Twenty-one species of cetacean have the reasonable potential to co-
occur with the proposed survey activities. Please refer to Table 3. Of
the cetacean species that may be present, one is classified as a low-
frequency cetacean (i.e., the Bryde's whale), 18 are classified as mid-
frequency cetaceans (i.e., all delphinid and ziphiid species and the
sperm whale), and two are classified as high-frequency cetaceans (i.e.,
Kogia spp.).
Potential Effects of the Specified Activity on Marine Mammals and Their
Habitat
This section includes a summary and discussion of the ways that
components of the specified activity may impact marine mammals and
their habitat. The ``Estimated Take'' section later in this document
includes a quantitative analysis of the number of individuals that are
expected to be taken by this activity. The ``Negligible Impact Analysis
and Determination'' section considers the content of this section and
[[Page 29234]]
the material it references, the ``Estimated Take'' section, and the
``Proposed Mitigation'' section, to draw conclusions regarding the
likely impacts of these activities on the reproductive success or
survivorship of individuals and how those impacts on individuals are
likely to impact marine mammal species or stocks. In the following
discussion, we provide general background information on sound before
considering potential effects to marine mammals from the specified
activities (i.e., sound, ship strike, and contaminants).
Background on Sound and Acoustic Metrics
This section contains a brief technical background on sound, on the
characteristics of certain sound types, and on metrics used in this
proposal inasmuch as the information is relevant to other sections of
this document. For general information on sound and its interaction
with the marine environment, please see, e.g., Au and Hastings (2008);
Richardson et al. (1995); Urick (1983).
Sound travels in waves, the basic components of which are
frequency, wavelength, velocity, and amplitude. Frequency is the number
of pressure waves that pass by a reference point per unit of time and
is measured in Hz or cycles per second. Wavelength is the distance
between two peaks or corresponding points of a sound wave (length of
one cycle). Higher frequency sounds have shorter wavelengths than lower
frequency sounds, and typically attenuate (decrease) more rapidly,
except in certain cases in shallower water. Amplitude is the height of
the sound pressure wave or the ``loudness'' of a sound and is typically
described using the relative unit of the dB. A sound pressure level
(SPL) in dB is described as the ratio between a measured pressure and a
reference pressure (for underwater sound, this is 1 microPascal
([mu]Pa)), and is a logarithmic unit that accounts for large variations
in amplitude; therefore, a relatively small change in dB corresponds to
large changes in sound pressure. The source level (SL) represents the
SPL referenced at a distance of 1 m from the source (referenced to 1
[mu]Pa), while the received level is the SPL at the listener's position
(referenced to 1 [mu]Pa).
When underwater objects vibrate or activity occurs, sound-pressure
waves are created. These waves alternately compress and decompress the
water as the sound wave travels. Underwater sound waves radiate in a
manner similar to ripples on the surface of a pond and may be either
directed in a beam or beams or may radiate in all directions
(omnidirectional sources), as is nominally the case for sound produced
by airguns (though when grouped in arrays there is some
directionality). The compressions and decompressions associated with
sound waves are detected as changes in pressure by aquatic life and
man-made sound receptors such as hydrophones.
Sounds are often considered to fall into one of two general types:
Pulsed and non-pulsed (defined in the following). The distinction
between these two sound types is important because they have differing
potential to cause physical effects, particularly with regard to
hearing (e.g., Ward, 1997 in Southall et al., 2007). Please see
Southall et al. (2007) for an in-depth discussion of these concepts.
The distinction between these two sound types is not always obvious, as
certain signals share properties of both pulsed and non-pulsed sounds.
A signal near a source could be categorized as a pulse, but due to
propagation effects as it moves farther from the source, the signal
duration becomes longer (e.g., Greene and Richardson, 1988).
Pulsed sound sources (e.g., airguns, explosions, gunshots, sonic
booms, impact pile driving) produce signals that are brief (typically
considered to be less than one second), broadband, atonal transients
(ANSI, 1986, 2005; Harris, 1998; NIOSH, 1998; ISO, 2003) and occur
either as isolated events or repeated in some succession. Pulsed sounds
are all characterized by a relatively rapid rise from ambient pressure
to a maximal pressure value followed by a rapid decay period that may
include a period of diminishing, oscillating maximal and minimal
pressures, and generally have an increased capacity to induce physical
injury as compared with sounds that lack these features.
Non-pulsed sounds can be tonal, narrowband, or broadband, brief or
prolonged, and may be either continuous or intermittent (ANSI, 1995;
NIOSH, 1998). Some of these non-pulsed sounds can be transient signals
of short duration but without the essential properties of pulses (e.g.,
rapid rise time). Examples of non-pulsed sounds include those produced
by vessels, aircraft, machinery operations such as drilling or
dredging, vibratory pile driving, and active sonar systems. The
duration of such sounds, as received at a distance, can be greatly
extended in a highly reverberant environment.
Root mean square (rms) is the quadratic mean sound pressure over
the duration of an impulse. Root mean square is calculated by squaring
all of the sound amplitudes, averaging the squares, and then taking the
square root of the average (Urick, 1983). Root mean square accounts for
both positive and negative values; squaring the pressures makes all
values positive so that they may be accounted for in the summation of
pressure levels (Hastings and Popper, 2005). The length of the time
window used for the purpose of the rms SPL calculation can be selected
using different approaches. This value is commonly defined as the 90
percent energy pulse duration, containing the central 90 percent (from
5 to 95 percent of the total) of the cumulative square pressure (or
sound exposure level) of the pulse. However, as was the case in the
modeling performed for this effort, a fixed time window may be used.
Here, a sliding window was used to calculate rms SPL values for a
series of fixed window lengths within the pulse. The maximum value of
rms SPL over all time window positions is taken to represent the rms
SPL of the pulse. This measurement is often used in the context of
discussing behavioral effects, in part because behavioral effects,
which often result from auditory cues, may be better expressed through
averaged units than by peak pressures. Energy equivalent SPL (denoted
Leq) is the measure of the average amount of energy carried
by a time-dependent pressure wave over a period of time. The
Leq is numerically equal to the rms SPL of a steady sound
that has the same total energy as the sound measured over the given
time window. Conceptually, the difference between the two metrics is
that the rms SPL is computed over short time periods, usually one
second or less, and tracks the fluctuations of a non-steady acoustic
signal, whereas the Leq reflects the average SPL of an
acoustic signal over tens of seconds or longer.
Sound exposure level (SEL; represented as dB re 1 [mu]Pa\2\-s)
represents the total energy in a stated frequency band over a stated
time interval or event, and considers both intensity and duration of
exposure. The per-pulse SEL is calculated over the time window
containing the entire pulse (i.e., 100 percent of the acoustic energy).
SEL is a cumulative metric; it can be accumulated over a single pulse,
or calculated over periods containing multiple pulses. Cumulative SEL
represents the total energy accumulated by a receiver over a defined
time window or during an event.
Peak sound pressure (also referred to as zero-to-peak sound
pressure or 0-pk) is the maximum instantaneous sound
[[Page 29235]]
pressure measurable in the water at a specified distance from the
source, and is represented in the same units as the rms sound pressure.
Another common metric is peak-to-peak sound pressure (pk-pk), which is
the algebraic difference between the peak positive and peak negative
sound pressures. Peak-to-peak pressure is typically approximately 6 dB
higher than peak pressure (Southall et al., 2007).
Airguns produce pulsed signals, with energy in a frequency range
from about 10-2,000 Hz, and most energy radiated at frequencies below
200 Hz. Larger airguns, with larger internal air volume, produce higher
broadband sound levels with sound energy spectrum shifted toward the
lower frequencies. The amplitude of the acoustic wave emitted from the
source is equal in all directions (i.e., omnidirectional), but when
used in arrays, airguns do possess some directionality due to different
phase delays between guns in different directions. Airgun arrays are
typically tuned to maximize functionality for data acquisition
purposes, meaning that more sound energy is focused downwardly than
horizontally, and sound transmitted in horizontal directions and at
higher frequencies is minimized to the extent possible.
Acoustic sources used for HRG surveys generally produce higher
frequency signals with highly directional beam patterns. These sources
are generally considered to be intermittent, with typically brief
signal durations, and temporal characteristics that more closely
resemble those of impulsive sounds than non-impulsive sounds. Boomers
generate a high-amplitude broadband (100 Hz-10 kHz) acoustic pulse with
high downward directivity, though may be considered omnidirectional at
frequencies below 1 kHz. Subbottom profiler systems generally project a
chirp pulse spanning an operator-selectable frequency band, usually
between 1 to 20 kHz, with a single beam directed vertically down.
Multibeam echosounders use an array of transducers that project a high-
frequency, fan-shaped beam under the hull of a survey ship and
perpendicular to the direction of motion. Side-scan sonars use two
transducers to project high-frequency beams that are usually wide in
the vertical plane (50[deg]-70[deg]) and very narrow in the horizontal
plane (less than a few degrees).
Vessel noise, produced largely by cavitation of propellers and by
machinery inside the hull, is considered a non-pulsed sound. Sounds
emitted by survey vessels are low frequency and continuous, but would
be widely dispersed in both space and time. Survey vessel traffic is of
low density compared to traffic associated with commercial shipping,
industry support vessels, or commercial fishing vessels, and would
therefore be expected to represent an insignificant incremental
increase in the total amount of anthropogenic sound input to the marine
environment. For these reasons, we do not consider vessel traffic noise
further in this analysis.
Potential Effects of Underwater Sound
Note that, in the following discussion, we refer in many cases to a
review article concerning studies of noise-induced hearing loss
conducted from 1996-2015 (i.e., Finneran, 2015). For study-specific
citations, please see that work. Anthropogenic sounds cover a broad
range of frequencies and sound levels and can have a range of highly
variable impacts on marine life, from none or minor to potentially
severe responses, depending on received levels, duration of exposure,
behavioral context, and various other factors. The potential effects of
underwater sound from active acoustic sources can potentially result in
one or more of the following: Temporary or permanent hearing
impairment, non-auditory physical or physiological effects, behavioral
disturbance, stress, and masking (Richardson et al., 1995; Gordon et
al., 2004; Nowacek et al., 2007; Southall et al., 2007; G[ouml]tz et
al., 2009). The degree of effect is intrinsically related to the signal
characteristics, received level, distance from the source, and duration
of the sound exposure. In general, sudden, high level sounds can cause
hearing loss, as can longer exposures to lower level sounds. Temporary
or permanent loss of hearing will occur almost exclusively for noise
within an animal's hearing range. We first describe specific
manifestations of acoustic effects before providing discussion specific
to the use of airgun arrays.
Richardson et al. (1995) described zones of increasing intensity of
effect that might be expected to occur, in relation to distance from a
source and assuming that the signal is within an animal's hearing
range. First is the area within which the acoustic signal would be
audible (potentially perceived) to the animal, but not strong enough to
elicit any overt behavioral or physiological response. The next zone
corresponds with the area where the signal is audible to the animal and
of sufficient intensity to elicit behavioral or physiological
responsiveness. Third is a zone within which, for signals of high
intensity, the received level is sufficient to potentially cause
discomfort or tissue damage to auditory or other systems. Overlaying
these zones to a certain extent is the area within which masking (i.e.,
when a sound interferes with or masks the ability of an animal to
detect a signal of interest that is above the absolute hearing
threshold) may occur; the masking zone may be highly variable in size.
We describe more severe effects (i.e., certain non-auditory
physical or physiological effects) only briefly as we do not expect
that use of airgun arrays are reasonably likely to result in such
effects (see below for further discussion). Potential effects from
impulsive sound sources can range in severity from effects such as
behavioral disturbance or tactile perception to physical discomfort,
slight injury of the internal organs and the auditory system, or
mortality (Yelverton et al., 1973). Non-auditory physiological effects
or injuries that theoretically might occur in marine mammals exposed to
high level underwater sound or as a secondary effect of extreme
behavioral reactions (e.g., change in dive profile as a result of an
avoidance reaction) caused by exposure to sound include neurological
effects, bubble formation, resonance effects, and other types of organ
or tissue damage (Cox et al., 2006; Southall et al., 2007; Zimmer and
Tyack, 2007; Tal et al., 2015). The survey activities considered here
do not involve the use of devices such as explosives or mid-frequency
tactical sonar that are associated with these types of effects.
When a live or dead marine mammal swims or floats onto shore and is
incapable of returning to sea, the event is termed a ``stranding'' (16
U.S.C. 1421h(3)). Marine mammals are known to strand for a variety of
reasons, such as infectious agents, biotoxicosis, starvation, fishery
interaction, ship strike, unusual oceanographic or weather events,
sound exposure, or combinations of these stressors sustained
concurrently or in series (e.g., Geraci et al., 1999). However, the
cause or causes of most strandings are unknown (e.g., Best, 1982).
Combinations of dissimilar stressors may combine to kill an animal or
dramatically reduce its fitness, even though one exposure without the
other would not be expected to produce the same outcome (e.g., Sih et
al., 2004). For further description of specific stranding events see,
e.g., Southall et al., 2006, 2013; Jepson et al., 2013; Wright et al.,
2013.
Use of military tactical sonar has been implicated in multiple
investigated stranding events, although one stranding event was
contemporaneous with and reasonably associated spatially
[[Page 29236]]
with the use of seismic airguns. This event occurred in the Gulf of
California, coincident with seismic reflection profiling by the R/V
Maurice Ewing operated by Columbia University's Lamont-Doherty Earth
Observatory and involved two Cuvier's beaked whales (Hildebrand, 2004).
The vessel had been firing an array of 20 airguns with a total volume
of 8,500 in\3\ (Hildebrand, 2004; Taylor et al., 2004). Most known
stranding events have involved beaked whales, though a small number
have involved deep-diving delphinids or sperm whales (e.g., Mazzariol
et al., 2010; Southall et al., 2013). In general, long duration (~1
second) and high-intensity sounds (235 dB SPL) have been implicated in
stranding events (Hildebrand, 2004). With regard to beaked whales, mid-
frequency sound is typically implicated (when causation can be
determined) (Hildebrand, 2004). Although seismic airguns create
predominantly low-frequency energy, the signal does include a mid-
frequency component.
Threshold Shift--Marine mammals exposed to high-intensity sound, or
to lower-intensity sound for prolonged periods, can experience hearing
threshold shift (TS), which is the loss of hearing sensitivity at
certain frequency ranges (Finneran, 2015). TS can be permanent (PTS),
in which case the loss of hearing sensitivity is not fully recoverable,
or temporary (TTS), in which case the animal's hearing threshold would
recover over time (Southall et al.,, 2007). Repeated sound exposure
that leads to TTS could cause PTS. In severe cases of PTS, there can be
total or partial deafness, while in most cases the animal has an
impaired ability to hear sounds in specific frequency ranges (Kryter,
1985).
When PTS occurs, there is physical damage to the sound receptors in
the ear (i.e., tissue damage), whereas TTS represents primarily tissue
fatigue and is reversible (Southall et al., 2007). In addition, other
investigators have suggested that TTS is within the normal bounds of
physiological variability and tolerance and does not represent physical
injury (e.g., Ward, 1997). Therefore, NMFS does not consider TTS to
constitute auditory injury.
Relationships between TTS and PTS thresholds have not been studied
in marine mammals, and there is no PTS data for cetaceans, but such
relationships are assumed to be similar to those in humans and other
terrestrial mammals. PTS typically occurs at exposure levels at least
several decibels above (a 40-dB threshold shift approximates PTS onset;
e.g., Kryter et al., 1966; Miller, 1974) that inducing mild TTS (a 6-dB
threshold shift approximates TTS onset; e.g., Southall et al. 2007).
Based on data from terrestrial mammals, a precautionary assumption is
that the PTS thresholds for impulse sounds (such as airgun pulses as
received close to the source) are at least 6 dB higher than the TTS
threshold on a peak-pressure basis and PTS cumulative sound exposure
level thresholds are 15 to 20 dB higher than TTS cumulative sound
exposure level thresholds (Southall et al., 2007). Given the higher
level of sound or longer exposure duration necessary to cause PTS as
compared with TTS, it is considerably less likely that PTS could occur.
For mid-frequency cetaceans in particular, potential protective
mechanisms may help limit onset of TTS or prevent onset of PTS. Such
mechanisms include dampening of hearing, auditory adaptation, or
behavioral amelioration (e.g., Nachtigall and Supin, 2013; Miller et
al., 2012; Finneran et al., 2015; Popov et al., 2016; Nachtigall et
al., 2017).
TTS is the mildest form of hearing impairment that can occur during
exposure to sound (Kryter, 1985). While experiencing TTS, the hearing
threshold rises, and a sound must be at a higher level in order to be
heard. In terrestrial and marine mammals, TTS can last from minutes or
hours to days (in cases of strong TTS). In many cases, hearing
sensitivity recovers rapidly after exposure to the sound ends. Few data
on sound levels and durations necessary to elicit mild TTS have been
obtained for marine mammals.
Marine mammal hearing plays a critical role in communication with
conspecifics, and interpretation of environmental cues for purposes
such as predator avoidance and prey capture. Depending on the degree
(elevation of threshold in dB), duration (i.e., recovery time), and
frequency range of TTS, and the context in which it is experienced, TTS
can have effects on marine mammals ranging from discountable to
serious. For example, a marine mammal may be able to readily compensate
for a brief, relatively small amount of TTS in a non-critical frequency
range that occurs during a time where ambient noise is lower and there
are not as many competing sounds present. Alternatively, a larger
amount and longer duration of TTS sustained during time when
communication is critical for successful mother/calf interactions could
have more serious impacts.
Finneran et al. (2015) measured hearing thresholds in three captive
bottlenose dolphins before and after exposure to ten pulses produced by
a seismic airgun in order to study TTS induced after exposure to
multiple pulses. Exposures began at relatively low levels and gradually
increased over a period of several months, with the highest exposures
at peak SPLs from 196 to 210 dB and cumulative (unweighted) SELs from
193-195 dB. No substantial TTS was observed. In addition, behavioral
reactions were observed that indicated that animals can learn behaviors
that effectively mitigate noise exposures (although exposure patterns
must be learned, which is less likely in wild animals than for the
captive animals considered in the study). The authors note that the
failure to induce more significant auditory effects was likely due to
the intermittent nature of exposure, the relatively low peak pressure
produced by the acoustic source, and the low-frequency energy in airgun
pulses as compared with the frequency range of best sensitivity for
dolphins and other mid-frequency cetaceans.
Currently, TTS data only exist for four species of cetaceans
(bottlenose dolphin, beluga whale (Delphinapterus leucas), harbor
porpoise (Phocoena phocoena), and Yangtze finless porpoise (Neophocoena
asiaeorientalis)) exposed to a limited number of sound sources (i.e.,
mostly tones and octave-band noise) in laboratory settings (Finneran,
2015). In general, harbor porpoises have a lower TTS onset than other
measured cetacean species (Finneran, 2015). Additionally, the existing
marine mammal TTS data come from a limited number of individuals within
these species. There are no data available on noise-induced hearing
loss for mysticetes.
Critical questions remain regarding the rate of TTS growth and
recovery after exposure to intermittent noise and the effects of single
and multiple pulses. Data at present are also insufficient to construct
generalized models for recovery and determine the time necessary to
treat subsequent exposures as independent events. More information is
needed on the relationship between auditory evoked potential and
behavioral measures of TTS for various stimuli. For summaries of data
on TTS in marine mammals or for further discussion of TTS onset
thresholds, please see Southall et al. (2007), Finneran and Jenkins
(2012), Finneran (2015), and NMFS (2016).
Behavioral Effects--Behavioral disturbance may include a variety of
effects, including subtle changes in behavior (e.g., minor or brief
avoidance of an area or changes in vocalizations), more conspicuous
changes in similar behavioral activities, and more
[[Page 29237]]
sustained and/or potentially severe reactions, such as displacement
from or abandonment of high-quality habitat. Behavioral responses to
sound are highly variable and context-specific and any reactions depend
on numerous intrinsic and extrinsic factors (e.g., species, state of
maturity, experience, current activity, reproductive state, auditory
sensitivity, time of day), as well as the interplay between factors
(e.g., Richardson et al., 1995; Wartzok et al., 2003; Southall et al.,
2007; Weilgart, 2007; Archer et al., 2010). Behavioral reactions can
vary not only among individuals but also within an individual,
depending on previous experience with a sound source, context, and
numerous other factors (Ellison et al., 2012), and can vary depending
on characteristics associated with the sound source (e.g., whether it
is moving or stationary, number of sources, distance from the source).
Please see Appendices B-C of Southall et al. (2007) for a review of
studies involving marine mammal behavioral responses to sound.
Habituation can occur when an animal's response to a stimulus wanes
with repeated exposure, usually in the absence of unpleasant associated
events (Wartzok et al., 2003). Animals are most likely to habituate to
sounds that are predictable and unvarying. It is important to note that
habituation is appropriately considered as a ``progressive reduction in
response to stimuli that are perceived as neither aversive nor
beneficial,'' rather than as, more generally, moderation in response to
human disturbance (Bejder et al., 2009). The opposite process is
sensitization, when an unpleasant experience leads to subsequent
responses, often in the form of avoidance, at a lower level of
exposure. As noted, behavioral state may affect the type of response.
For example, animals that are resting may show greater behavioral
change in response to disturbing sound levels than animals that are
highly motivated to remain in an area for feeding (Richardson et al.,
1995; NRC, 2003; Wartzok et al., 2003). Controlled experiments with
captive marine mammals have showed pronounced behavioral reactions,
including avoidance of loud sound sources (Ridgway et al., 1997).
Observed responses of wild marine mammals to loud pulsed sound sources
(typically airguns or acoustic harassment devices) have been varied but
often consist of avoidance behavior or other behavioral changes
suggesting discomfort (Morton and Symonds, 2002; see also Richardson et
al., 1995; Nowacek et al., 2007). However, many delphinids approach
acoustic source vessels with no apparent discomfort or obvious
behavioral change (e.g., Barkaszi et al., 2012).
Available studies show wide variation in response to underwater
sound; therefore, it is difficult to predict specifically how any given
sound in a particular instance might affect marine mammals perceiving
the signal. If a marine mammal does react briefly to an underwater
sound by changing its behavior or moving a small distance, the impacts
of the change are unlikely to be significant to the individual, let
alone the stock or population. However, if a sound source displaces
marine mammals from an important feeding or breeding area for a
prolonged period, impacts on individuals and populations could be
significant (e.g., Lusseau and Bejder, 2007; Weilgart, 2007; NRC,
2005). However, there are broad categories of potential response, which
we describe in greater detail here, that include alteration of dive
behavior, alteration of foraging behavior, effects to breathing,
interference with or alteration of vocalization, avoidance, and flight.
Changes in dive behavior can vary widely, and may consist of
increased or decreased dive times and surface intervals as well as
changes in the rates of ascent and descent during a dive (e.g., Frankel
and Clark, 2000; Ng and Leung, 2003; Nowacek et al.; 2004; Goldbogen et
al., 2013a, 2013b). Variations in dive behavior may reflect
interruptions in biologically significant activities (e.g., foraging)
or they may be of little biological significance. The impact of an
alteration to dive behavior resulting from an acoustic exposure depends
on what the animal is doing at the time of the exposure and the type
and magnitude of the response.
Disruption of feeding behavior can be difficult to correlate with
anthropogenic sound exposure (but see discussion of impacts to sperm
whale foraging behavior below and in ``Proposed Mitigation''), so it is
usually inferred by observed displacement from known foraging areas,
the appearance of secondary indicators (e.g., bubble nets or sediment
plumes), or changes in dive behavior. As for other types of behavioral
response, the frequency, duration, and temporal pattern of signal
presentation, as well as differences in species sensitivity, are likely
contributing factors to differences in response in any given
circumstance (e.g., Croll et al., 2001; Nowacek et al.; 2004; Madsen et
al., 2006a; Yazvenko et al., 2007). A determination of whether foraging
disruptions incur fitness consequences would require information on or
estimates of the energetic requirements of the affected individuals and
the relationship between prey availability, foraging effort and
success, and the life history stage of the animal.
Visual tracking, passive acoustic monitoring, and movement
recording tags were used to quantify sperm whale behavior prior to,
during, and following exposure to airgun arrays at received levels in
the range 140-160 dB at distances of 7-13 km, following a phase-in of
sound intensity and full array exposures at 1-13 km (Madsen et al.,
2006a; Miller et al., 2009). Sperm whales did not exhibit horizontal
avoidance behavior at the surface. However, foraging behavior may have
been affected. The sperm whales exhibited 19 percent less vocal (buzz)
rate during full exposure relative to post exposure, and the whale that
was approached most closely had an extended resting period and did not
resume foraging until the airguns had ceased firing. The remaining
whales continued to execute foraging dives throughout exposure;
however, swimming movements during foraging dives were 6 percent lower
during exposure than control periods (Miller et al., 2009). These data
raise concerns that airgun surveys may impact foraging behavior in
sperm whales, although more data are required to understand whether the
differences were due to exposure or natural variation in sperm whale
behavior (Miller et al., 2009). We discuss these findings in greater
detail under ``Proposed Mitigation.''
Variations in respiration naturally vary with different behaviors
and alterations to breathing rate as a function of acoustic exposure
can be expected to co-occur with other behavioral reactions, such as a
flight response or an alteration in diving. However, respiration rates
in and of themselves may be representative of annoyance or an acute
stress response. Various studies have shown that respiration rates may
either be unaffected or could increase, depending on the species and
signal characteristics, again highlighting the importance in
understanding species differences in the tolerance of underwater noise
when determining the potential for impacts resulting from anthropogenic
sound exposure (e.g., Kastelein et al., 2001, 2005, 2006; Gailey et
al., 2007; Gailey et al., 2016).
Marine mammals vocalize for different purposes and across multiple
modes, such as whistling, echolocation click production, calling, and
singing. Changes in vocalization behavior in response to anthropogenic
noise can
[[Page 29238]]
occur for any of these modes and may result from a need to compete with
an increase in background noise or may reflect increased vigilance or a
startle response. For example, in the presence of potentially masking
signals, humpback whales and killer whales have been observed to
increase the length of their songs (Miller et al., 2000; Fristrup et
al., 2003; Foote et al., 2004), while right whales have been observed
to shift the frequency content of their calls upward while reducing the
rate of calling in areas of increased anthropogenic noise (Parks et
al., 2007). In some cases, animals may cease sound production during
production of aversive signals (Bowles et al., 1994).
Cerchio et al. (2014) used passive acoustic monitoring to document
the presence of singing humpback whales off the coast of northern
Angola and to opportunistically test for the effect of seismic survey
activity on the number of singing whales. Two recording units were
deployed between March and December 2008 in the offshore environment;
numbers of singers were counted every hour. Generalized Additive Mixed
Models were used to assess the effect of survey day (seasonality), hour
(diel variation), moon phase, and received levels of noise (measured
from a single pulse during each ten minute sampled period) on singer
number. The number of singers significantly decreased with increasing
received level of noise, suggesting that humpback whale communication
was disrupted to some extent by the survey activity.
Castellote et al. (2012) reported acoustic and behavioral changes
by fin whales in response to shipping and airgun noise. Acoustic
features of fin whale song notes recorded in the Mediterranean Sea and
northeast Atlantic Ocean were compared for areas with different
shipping noise levels and traffic intensities and during an airgun
survey. During the first 72 hours of the survey, a steady decrease in
song received levels and bearings to singers indicated that whales
moved away from the acoustic source and out of the study area. This
displacement persisted for a time period well beyond the 10-day
duration of airgun activity, providing evidence that fin whales may
avoid an area for an extended period in the presence of increased
noise. The authors hypothesize that fin whale acoustic communication is
modified to compensate for increased background noise and that a
sensitization process may play a role in the observed temporary
displacement.
Seismic pulses at average received levels of 131 dB re 1 [mu]Pa\2\-
s caused blue whales to increase call production (Di Iorio and Clark,
2010). In contrast, McDonald et al. (1995) tracked a blue whale with
seafloor seismometers and reported that it stopped vocalizing and
changed its travel direction at a range of 10 km from the acoustic
source vessel (estimated received level 143 dB pk-pk). Blackwell et al.
(2013) found that bowhead whale call rates dropped significantly at
onset of airgun use at sites with a median distance of 41-45 km from
the survey. Blackwell et al. (2015) expanded this analysis to show that
whales actually increased calling rates as soon as airgun signals were
detectable before ultimately decreasing calling rates at higher
received levels (i.e., 10-minute cumulative sound exposure level (cSEL)
of ~127 dB). Overall, these results suggest that bowhead whales may
adjust their vocal output in an effort to compensate for noise before
ceasing vocalization effort and ultimately deflecting from the acoustic
source (Blackwell et al., 2013, 2015). These studies demonstrate that
even low levels of noise received far from the source can induce
changes in vocalization and/or behavior for mysticetes.
Avoidance is the displacement of an individual from an area or
migration path as a result of the presence of a sound or other
stressors, and is one of the most obvious manifestations of disturbance
in marine mammals (Richardson et al., 1995). For example, gray whales
are known to change direction--deflecting from customary migratory
paths--in order to avoid noise from airgun surveys (Malme et al.,
1984). Humpback whales showed avoidance behavior in the presence of an
active airgun array during observational studies and controlled
exposure experiments in western Australia (McCauley et al., 2000a).
Avoidance may be short-term, with animals returning to the area once
the noise has ceased (e.g., Bowles et al., 1994; Goold, 1996; Stone et
al., 2000; Morton and Symonds, 2002; Gailey et al., 2007). Longer-term
displacement is possible, however, which may lead to changes in
abundance or distribution patterns of the affected species in the
affected region if habituation to the presence of the sound does not
occur (e.g., Bejder et al., 2006; Teilmann et al., 2006).
Forney et al. (2017) detail the potential effects of noise on
marine mammal populations with high site fidelity, including
displacement and auditory masking, noting that a lack of observed
response does not imply absence of fitness costs and that apparent
tolerance of disturbance may have population-level impacts that are
less obvious and difficult to document. As we discuss in describing our
proposed mitigation later in this document, avoidance of overlap
between disturbing noise and areas and/or times of particular
importance for sensitive species may be critical to avoiding
population-level impacts because (particularly for animals with high
site fidelity) there may be a strong motivation to remain in the area
despite negative impacts. Forney et al. (2017) state that, for these
animals, remaining in a disturbed area may reflect a lack of
alternatives rather than a lack of effects. The authors discuss several
case studies, including western Pacific gray whales, which are a small
population of mysticetes believed to be adversely affected by oil and
gas development off Sakhalin Island, Russia (Weller et al., 2002;
Reeves et al., 2005). Western gray whales display a high degree of
interannual site fidelity to the area for foraging purposes, and
observations in the area during airgun surveys has shown the potential
for harm caused by displacement from such an important area (Weller et
al., 2006; Johnson et al., 2007). As we discuss below in ``Proposed
Mitigation,'' similar concerns exist in relation to the potential for
survey activity in the resident habitat of the GOM's small population
of Bryde's whales. Forney et al. (2017) also discuss beaked whales,
noting that anthropogenic effects in areas where they are resident
could cause severe biological consequences, in part because
displacement may adversely affect foraging rates, reproduction, or
health, while an overriding instinct to remain could lead to more
severe acute effects.
A flight response is a dramatic change in normal movement to a
directed and rapid movement away from the perceived location of a sound
source. The flight response differs from other avoidance responses in
the intensity of the response (e.g., directed movement, rate of
travel). Relatively little information on flight responses of marine
mammals to anthropogenic signals exist, although observations of flight
responses to the presence of predators have occurred (Connor and
Heithaus, 1996). The result of a flight response could range from
brief, temporary exertion and displacement from the area where the
signal provokes flight to, in extreme cases, marine mammal strandings
(Evans and England, 2001). However, it should be noted that response to
a perceived predator does not necessarily invoke flight (Ford and
Reeves, 2008), and
[[Page 29239]]
whether individuals are solitary or in groups may influence the
response.
Behavioral disturbance can also impact marine mammals in more
subtle ways. Increased vigilance may result in costs related to
diversion of focus and attention (i.e., when a response consists of
increased vigilance, it may come at the cost of decreased attention to
other critical behaviors such as foraging or resting). These effects
have generally not been demonstrated for marine mammals, but studies
involving fish and terrestrial animals have shown that increased
vigilance may substantially reduce feeding rates (e.g., Beauchamp and
Livoreil, 1997; Fritz et al., 2002; Purser and Radford, 2011). In
addition, chronic disturbance can cause population declines through
reduction of fitness (e.g., decline in body condition) and subsequent
reduction in reproductive success, survival, or both (e.g., Harrington
and Veitch, 1992; Daan et al., 1996; Bradshaw et al., 1998). However,
Ridgway et al. (2006) reported that increased vigilance in bottlenose
dolphins exposed to sound over a five-day period did not cause any
sleep deprivation or stress effects.
Many animals perform vital functions, such as feeding, resting,
traveling, and socializing, on a diel cycle (24-hour cycle). Disruption
of such functions resulting from reactions to stressors such as sound
exposure are more likely to be significant if they last more than one
diel cycle or recur on subsequent days (Southall et al., 2007).
Consequently, a behavioral response lasting less than one day and not
recurring on subsequent days is not considered particularly severe
unless it could directly affect reproduction or survival (Southall et
al., 2007). Note that there is a difference between multi-day
substantive behavioral reactions and multi-day anthropogenic
activities. For example, just because an activity lasts for multiple
days does not necessarily mean that individual animals are either
exposed to activity-related stressors for multiple days or, further,
exposed in a manner resulting in sustained multi-day substantive
behavioral responses.
Stone (2015a) reported data from at-sea observations during 1,196
airgun surveys from 1994 to 2010. When large arrays of airguns
(considered to be 500 in\3\ or more) were firing, lateral displacement,
more localized avoidance, or other changes in behavior were evident for
most odontocetes. However, significant responses to large arrays were
found only for the minke whale and fin whale. Behavioral responses
observed included changes in swimming or surfacing behavior, with
indications that cetaceans remained near the water surface at these
times. Cetaceans were recorded as feeding less often when large arrays
were active. Behavioral observations of gray whales during an airgun
survey monitored whale movements and respirations pre-, during-, and
post-seismic survey (Gailey et al., 2016). Behavioral state and water
depth were the best `natural' predictors of whale movements and
respiration and, after considering natural variation, none of the
response variables were significantly associated with survey or vessel
sounds.
Stress Responses--An animal's perception of a threat may be
sufficient to trigger stress responses consisting of some combination
of behavioral responses, autonomic nervous system responses,
neuroendocrine responses, or immune responses (e.g., Seyle, 1950;
Moberg, 2000). In many cases, an animal's first and sometimes most
economical (in terms of energetic costs) response is behavioral
avoidance of the potential stressor. Autonomic nervous system responses
to stress typically involve changes in heart rate, blood pressure, and
gastrointestinal activity. These responses have a relatively short
duration and may or may not have a significant long-term effect on an
animal's fitness.
Neuroendocrine stress responses often involve the hypothalamus-
pituitary-adrenal system. Virtually all neuroendocrine functions that
are affected by stress--including immune competence, reproduction,
metabolism, and behavior--are regulated by pituitary hormones. Stress-
induced changes in the secretion of pituitary hormones have been
implicated in failed reproduction, altered metabolism, reduced immune
competence, and behavioral disturbance (e.g., Moberg, 1987; Blecha,
2000). Increases in the circulation of glucocorticoids are also equated
with stress (Romano et al., 2004).
The primary distinction between stress (which is adaptive and does
not normally place an animal at risk) and ``distress'' is the cost of
the response. During a stress response, an animal uses glycogen stores
that can be quickly replenished once the stress is alleviated. In such
circumstances, the cost of the stress response would not pose serious
fitness consequences. However, when an animal does not have sufficient
energy reserves to satisfy the energetic costs of a stress response,
energy resources must be diverted from other functions. This state of
distress will last until the animal replenishes its energetic reserves
sufficiently to restore normal function.
Relationships between these physiological mechanisms, animal
behavior, and the costs of stress responses are well-studied through
controlled experiments and for both laboratory and free-ranging animals
(e.g., Holberton et al., 1996; Hood et al., 1998; Jessop et al., 2003;
Krausman et al., 2004; Lankford et al., 2005). Stress responses due to
exposure to anthropogenic sounds or other stressors and their effects
on marine mammals have also been reviewed (Fair and Becker, 2000;
Romano et al., 2002b) and, more rarely, studied in wild populations
(e.g., Romano et al., 2002a). For example, Rolland et al. (2012) found
that noise reduction from reduced ship traffic in the Bay of Fundy was
associated with decreased stress in North Atlantic right whales. These
and other studies lead to a reasonable expectation that some marine
mammals will experience physiological stress responses upon exposure to
acoustic stressors and that it is possible that some of these would be
classified as ``distress.'' In addition, any animal experiencing TTS
would likely also experience stress responses (NRC, 2003).
Auditory Masking--Sound can disrupt behavior through masking, or
interfering with, an animal's ability to detect, recognize, or
discriminate between acoustic signals of interest (e.g., those used for
intraspecific communication and social interactions, prey detection,
predator avoidance, navigation) (Richardson et al., 1995; Erbe et al.,
2016). Masking occurs when the receipt of a sound is interfered with by
another coincident sound at similar frequencies and at similar or
higher intensity, and may occur whether the sound is natural (e.g.,
snapping shrimp, wind, waves, precipitation) or anthropogenic (e.g.,
shipping, sonar, seismic exploration) in origin. The ability of a noise
source to mask biologically important sounds depends on the
characteristics of both the noise source and the signal of interest
(e.g., signal-to-noise ratio, temporal variability, direction), in
relation to each other and to an animal's hearing abilities (e.g.,
sensitivity, frequency range, critical ratios, frequency
discrimination, directional discrimination, age or TTS hearing loss),
and existing ambient noise and propagation conditions.
Under certain circumstances, marine mammals experiencing
significant masking could also be impaired from maximizing their
performance fitness in survival and reproduction. Therefore, when the
coincident (masking) sound is man-made, it may be considered harassment
when disrupting or altering critical behaviors. It is important to
[[Page 29240]]
distinguish TTS and PTS, which persist after the sound exposure, from
masking, which occurs during the sound exposure. Because masking
(without resulting in TS) is not associated with abnormal physiological
function, it is not considered a physiological effect, but rather a
potential behavioral effect.
The frequency range of the potentially masking sound is important
in determining any potential behavioral impacts. For example, low-
frequency signals may have less effect on high-frequency echolocation
sounds produced by odontocetes but are more likely to affect detection
of mysticete communication calls and other potentially important
natural sounds such as those produced by surf and some prey species.
The masking of communication signals by anthropogenic noise may be
considered as a reduction in the communication space of animals (e.g.,
Clark et al., 2009; Matthews et al., 2016) and may result in energetic
or other costs as animals change their vocalization behavior (e.g.,
Miller et al., 2000; Foote et al., 2004; Parks et al., 2007; Di Iorio
and Clark, 2009; Holt et al., 2009). Masking can be reduced in
situations where the signal and noise come from different directions
(Richardson et al., 1995), through amplitude modulation of the signal,
or through other compensatory behaviors (Houser and Moore, 2014).
Masking can be tested directly in captive species (e.g., Erbe, 2008),
but in wild populations it must be either modeled or inferred from
evidence of masking compensation. There are few studies addressing
real-world masking sounds likely to be experienced by marine mammals in
the wild (e.g., Branstetter et al., 2013).
Masking affects both senders and receivers of acoustic signals and
can potentially have long-term chronic effects on marine mammals at the
population level as well as at the individual level. Low-frequency
ambient sound levels have increased by as much as 20 dB (more than
three times in terms of SPL) in the world's ocean from pre-industrial
periods, with most of the increase from distant commercial shipping
(Hildebrand, 2009). All anthropogenic sound sources, but especially
chronic and lower-frequency signals (e.g., from vessel traffic),
contribute to elevated ambient sound levels, thus intensifying masking.
Ship Strike
Vessel collisions with marine mammals, or ship strikes, can result
in death or serious injury of the animal. Wounds resulting from ship
strike may include massive trauma, hemorrhaging, broken bones, or
propeller lacerations (Knowlton and Kraus, 2001). An animal at the
surface may be struck directly by a vessel, a surfacing animal may hit
the bottom of a vessel, or an animal just below the surface may be cut
by a vessel's propeller. Superficial strikes may not kill or result in
the death of the animal. These interactions are typically associated
with large whales, which are occasionally found draped across the
bulbous bow of large commercial ships upon arrival in port. Although
smaller cetaceans are more maneuverable in relation to large vessels
than are large whales, they may also be susceptible to strike. The
severity of injuries typically depends on the size and speed of the
vessel, with the probability of death or serious injury increasing as
vessel speed increases (Knowlton and Kraus, 2001; Laist et al., 2001;
Vanderlaan and Taggart, 2007; Conn and Silber, 2013). Impact forces
increase with speed, as does the probability of a strike at a given
distance (Silber et al., 2010; Gende et al., 2011).
Pace and Silber (2005) also found that the probability of death or
serious injury increased rapidly with increasing vessel speed.
Specifically, the predicted probability of serious injury or death
increased from 45 to 75 percent as vessel speed increased from 10 to 14
kn, and exceeded 90 percent at 17 kn. Higher speeds during collisions
result in greater force of impact, but higher speeds also appear to
increase the chance of severe injuries or death through increased
likelihood of collision by pulling whales toward the vessel (Clyne,
1999; Knowlton et al., 1995). In a separate study, Vanderlaan and
Taggart (2007) analyzed the probability of lethal mortality of large
whales at a given speed, showing that the greatest rate of change in
the probability of a lethal injury to a large whale as a function of
vessel speed occurs between 8.6 and 15 kn. The chances of a lethal
injury decline from approximately 80 percent at 15 kn to approximately
20 percent at 8.6 kn. At speeds below 11.8 kn, the chances of lethal
injury drop below 50 percent, while the probability asymptotically
increases toward 100 percent above 15 kn.
In an effort to reduce the number and severity of strikes of the
endangered North Atlantic right whale, NMFS implemented speed
restrictions in 2008 (73 FR 60173; October 10, 2008). These
restrictions require that vessels greater than or equal to 65 ft (19.8
m) in length travel at less than or equal to 10 kn near key port
entrances and in certain areas of right whale aggregation along the
U.S. eastern seaboard. Conn and Silber (2013) estimated that these
restrictions reduced total ship strike mortality risk levels by 80 to
90 percent.
For vessels used in geophysical survey activities, vessel speed
while towing gear is typically only 4-5 kn. At these speeds, both the
possibility of striking a marine mammal and the possibility of a strike
resulting in serious injury or mortality are discountable. At average
transit speed, the probability of serious injury or mortality resulting
from a strike is less than 50 percent. However, the likelihood of a
strike actually happening is again unlikely. Ship strikes, as analyzed
in the studies cited above, generally involve commercial shipping,
which is much more common in both space and time than is geophysical
survey activity. Jensen and Silber (2004) summarized ship strikes of
large whales worldwide from 1975-2003 and found that most collisions
occurred in the open ocean and involved large vessels (e.g., commercial
shipping). Commercial fishing vessels were responsible for three
percent of recorded collisions, while no such incidents were reported
for geophysical survey vessels during that time period.
It is possible for ship strikes to occur while traveling at slow
speeds. For example, a hydrographic survey vessel traveling at low
speed (5.5 kn) while conducting mapping surveys off the central
California coast struck and killed a blue whale in 2009. The State of
California determined that the whale had suddenly and unexpectedly
surfaced beneath the hull, with the result that the propeller severed
the whale's vertebrae, and that this was an unavoidable event. The
strike represented the only such incident in approximately 540,000
hours of similar coastal mapping activity (p = 1.9 x 10-6;
95% CI = 0-5.5 x 10-6; NMFS, 2013). In addition, a research
vessel reported a fatal strike in 2011 of a dolphin in the Atlantic,
demonstrating that it is possible for strikes involving smaller
cetaceans to occur. In that case, the incident report indicated that an
animal apparently was struck by the vessel's propeller as it was
intentionally swimming near the vessel. While indicative of the type of
unusual events that cannot be ruled out, neither of these instances
represents a circumstance that would be considered reasonably
foreseeable or that would be considered preventable.
Although the likelihood of vessels associated with geophysical
surveys striking a marine mammal are low, we require a robust ship
strike avoidance protocol (see ``Proposed Mitigation''), which we
believe eliminates any
[[Page 29241]]
foreseeable risk of ship strike. We anticipate that vessel collisions
involving seismic data acquisition vessels towing gear, while not
impossible, represent unlikely, unpredictable events for which there
are no preventive measures. Given the required mitigation measures, the
relatively slow speeds of vessels towing gear, the presence of bridge
crew watching for obstacles at all times (including marine mammals),
the presence of marine mammal observers, and the small number of
seismic survey cruises relative to commercial ship traffic, we believe
that the possibility of ship strike is discountable and, further, that
were a strike of a large whale to occur, it would be unlikely to result
in serious injury or mortality. No incidental take resulting from ship
strike is anticipated or proposed for authorization, and this potential
effect of the specified activity will not be discussed further in the
following analysis.
Other Potential Impacts
Here, we briefly address the potential risks due to entanglement
and contaminant spills. We are not aware of any records of marine
mammal entanglement in towed arrays such as those considered here, and
we address measures designed to eliminate the potential for
entanglement in gear used by OBS surveys in ``proposed Mitigation.''
The discharge of trash and debris is prohibited (33 CFR 151.51-77)
unless it is passed through a machine that breaks up solids such that
they can pass through a 25-mm mesh screen. All other trash and debris
must be returned to shore for proper disposal with municipal and solid
waste. Some personal items may be accidentally lost overboard. However,
U.S. Coast Guard and Environmental Protection Act regulations require
operators to become proactive in avoiding accidental loss of solid
waste items by developing waste management plans, posting informational
placards, manifesting trash sent to shore, and using special
precautions such as covering outside trash bins to prevent accidental
loss of solid waste. Any permits issued by BOEM would include guidance
for the handling and disposal of marine trash and debris, similar to
BSEE's Notice to Lessees 2015-G03 (``Marine Trash and Debris Awareness
and Elimination'') (BSEE, 2015; BOEM, 2017). We believe entanglement
risks are essentially eliminated by the proposed requirements, and
entanglement risks are not discussed further in this document.
Marine mammals could be affected by accidentally spilled diesel
fuel from a vessel associated with proposed survey activities.
Quantities of diesel fuel on the sea surface may affect marine mammals
through various pathways: Surface contact of the fuel with skin and
other mucous membranes, inhalation of concentrated petroleum vapors, or
ingestion of the fuel (direct ingestion or by the ingestion of
contaminated prey) (e.g., Geraci and St. Aubin, 1980, 1985, 1990).
However, the likelihood of a fuel spill during any particular
geophysical survey is considered to be remote, and the potential for
impacts to marine mammals would depend greatly on the size and location
of a spill and meteorological conditions at the time of the spill.
Spilled fuel would rapidly spread to a layer of varying thickness and
break up into narrow bands or windrows parallel to the wind direction.
The rate at which the fuel spreads would be determined by the
prevailing conditions such as temperature, water currents, tidal
streams, and wind speeds. Lighter, volatile components of the fuel
would evaporate to the atmosphere almost completely in a few days.
Evaporation rate may increase as the fuel spreads because of the
increased surface area of the slick. Rougher seas, high wind speeds,
and high temperatures also tend to increase the rate of evaporation and
the proportion of fuel lost by this process (Scholz et al., 1999). We
do not anticipate potentially meaningful effects to marine mammals as a
result of any contaminant spill resulting from the proposed survey
activities, and contaminant spills resulting from the specified
activity are not discussed further in this document.
Anticipated Effects on Marine Mammal Habitat
Physical Disturbance--Sources of seafloor disturbance related to
geophysical surveys that may impact marine mammal habitat include
placement of anchors, nodes, cables, sensors, or other equipment on or
in the seafloor for various activities. Equipment deployed on the
seafloor has the potential to cause direct physical damage and could
affect bottom-associated fish resources. Several NTLs detail the
mitigation measures used to prevent adverse impacts (``Biologically-
sensitive Underwater Features and Areas'' (NTL 2009-G39), ``Deepwater
Benthic Communities'' (NTL 2009-G40), and ``Shallow Hazards Program''
(NTL 2008-G05) (MMS, 2008; 2009a; 2009b)).
Placement of equipment, such as nodes, on the seafloor could damage
areas of hard bottom where direct contact with the seafloor occurs and
could crush epifauna (organisms that live on the seafloor or surface of
other organisms). Damage to unknown or unseen hard bottom could occur,
but because of the small area covered by most bottom-founded equipment,
the patchy distribution of hard bottom habitat, BOEM's review process,
and BOEM's application of avoidance conditions of approval, contact
with unknown hard bottom is expected to be rare and impacts minor.
Seafloor disturbance in areas of soft bottom can cause loss of small
patches of epifauna and infauna due to burial or crushing, and bottom-
feeding fishes could be temporarily displaced from feeding areas.
Overall, any effects of physical damage to habitat are expected to be
minor and temporary.
Effects to Prey--Sound may affect marine mammals through impacts on
the abundance, behavior, or distribution of prey species (e.g.,
crustaceans, cephalopods, fish, zooplankton). Marine mammal prey varies
by species, season, and location and, for some, is not well documented.
Here, we describe studies regarding the effects of noise on known
marine mammal prey.
Fish utilize the soundscape and components of sound in their
environment to perform important functions such as foraging, predator
avoidance, mating, and spawning (e.g., Zelick et al., 1999; Fay, 2009).
Depending on their hearing anatomy and peripheral sensory structures,
which vary among species, fishes hear sounds using pressure and
particle motion sensitivity capabilities and detect the motion of
surrounding water (Fay et al., 2008). The potential effects of airgun
noise on fishes depends on the overlapping frequency range, distance
from the sound source, water depth of exposure, and species-specific
hearing sensitivity, anatomy, and physiology. Key impacts to fishes may
include behavioral responses, hearing damage, barotrauma (pressure-
related injuries), and mortality.
Fish react to sounds which are especially strong and/or
intermittent low-frequency sounds, and behavioral responses such as
flight or avoidance are the most likely effects. Short duration, sharp
sounds can cause overt or subtle changes in fish behavior and local
distribution. The reaction of fish to airguns depends on the
physiological state of the fish, past exposures, motivation (e.g.,
feeding, spawning, migration), and other environmental factors.
Hastings and Popper (2005) identified several studies that suggest fish
may relocate to avoid certain areas
[[Page 29242]]
of sound energy. Several studies have demonstrated that airgun sounds
might affect the distribution and behavior of some fishes, potentially
impacting foraging opportunities or increasing energetic costs (e.g.,
Fewtrell and McCauley, 2012; Pearson et al., 1992; Skalski et al.,
1992; Santulli et al., 1999; Paxton et al., 2017). However, some
studies have shown no or slight reaction to airgun sounds (e.g., Pena
et al., 2013; Wardle et al., 2001; Jorgenson and Gyselman, 2009; Cott
et al., 2012). More commonly, though, the impacts of noise on fish are
temporary. Investigators reported significant, short-term declines in
commercial fishing catch rate of gadid fishes during and for up to five
days after survey operations, but the catch rate subsequently returned
to normal (Engas et al, 1996; Engas and Lokkeborg, 2002); other studies
have reported similar findings (Hassel et al., 2004). However, even
temporary effects to fish distribution patterns can impact their
ability to carry out important life-history functions (Paxton et al.,
2017).
SPLs of sufficient strength have been known to cause injury to fish
and fish mortality and, in some studies, fish auditory systems have
been damaged by airgun noise (McCauley et al., 2003; Popper et al.,
2005; Song et al., 2008). However, in most fish species, hair cells in
the ear continuously regenerate and loss of auditory function likely is
restored when damaged cells are replaced with new cells. Halvorsen et
al. (2012a) showed that a TTS of 4-6 dB was recoverable within 24 hours
for one species. Impacts would be most severe when the individual fish
is close to the source and when the duration of exposure is long. No
mortality occurred to fish in any of these studies.
Injury caused by barotrauma can range from slight to severe and can
cause death, and is most likely for fish with swim bladders. Barotrauma
injuries have been documented during controlled exposure to impact pile
driving (an impulsive noise source, as are airguns) (Halvorsen et al.,
2012b; Casper et al., 2013). For geophysical surveys, the sound source
is constantly moving, and most fish would likely avoid the sound source
prior to receiving sound of sufficient intensity to cause physiological
or anatomical damage.
Invertebrates appear to be able to detect sounds (Pumphrey, 1950;
Frings and Frings, 1967) and are most sensitive to low-frequency sounds
(Packard et al., 1990; Budelmann and Williamson, 1994; Lovell et al.,
2005; Mooney et al., 2010). Available data suggest that cephalopods are
capable of sensing the particle motion of sounds and detect low
frequencies up to 1-1.5 kHz, depending on the species, and so are
likely to detect airgun noise (Kaifu et al., 2008; Hu et al., 2009;
Mooney et al., 2010; Samson et al., 2014). Cephalopods have a
specialized sensory organ inside the head called a statocyst that may
help an animal determine its position in space (orientation) and
maintain balance (Budelmann, 1992). Packard et al. (1990) showed that
cephalopods were sensitive to particle motion, not sound pressure, and
Mooney et al. (2010) demonstrated that squid statocysts act as an
accelerometer through which particle motion of the sound field can be
detected. Auditory injuries (lesions occurring on the statocyst sensory
hair cells) have been reported upon controlled exposure to low-
frequency sounds, suggesting that cephalopods are particularly
sensitive to low-frequency sound (Andre et al., 2011; Sole et al.,
2013). Behavioral responses, such as inking and jetting, have also been
reported upon exposure to low-frequency sound (McCauley et al., 2000b;
Samson et al., 2014).
Impacts to benthic communities from impulsive sound generated by
active acoustic sound sources are not well documented. There are no
published data that indicate whether threshold shift injuries or
effects of auditory masking occur in benthic invertebrates, and there
are little data to suggest whether sounds from seismic surveys would
have any substantial impact on invertebrate behavior (Hawkins et al.,
2014), though some studies have indicated showed no short-term or long-
term effects of airgun exposure (e.g., Andriguetto-Filho et al., 2005;
Payne et al., 2007; 2008; Boudreau et al., 2009). Exposure to airgun
signals was found to significantly increase mortality in scallops, in
addition to causing significant changes in behavioral patterns during
exposure (Day et al., 2017). However, the implications of this finding
are not straightforward, as the authors state that the observed levels
of mortality were not beyond naturally occurring rates.
There is little information concerning potential impacts of noise
on zooplankton populations. However, one recent study (McCauley et al.,
2017) investigated zooplankton abundance, diversity, and mortality
before and after exposure to airgun noise, finding that the exposure
resulted in significant depletion for more than half the taxa present
and that there were two to three times more dead zooplankton after
airgun exposure compared with controls for all taxa. The majority of
taxa present were copepods and cladocerans; for these taxa, the range
within which effects on abundance were detected was up to approximately
1.2 km. In order to have significant impacts on r-selected species such
as plankton, the spatial or temporal scale of impact must be large in
comparison with the ecosystem concerned (McCauley et al., 2017).
Therefore, the large scale of effect observed here is of concern--
particularly where repeated noise exposure is expected--and further
study is warranted.
Prey species exposed to sound might move away from the sound
source, experience TTS, experience masking of biologically relevant
sounds, or show no obvious direct effects. Mortality from decompression
injuries is possible in close proximity to a sound, but only limited
data on mortality in response to airgun noise exposure are available
(Hawkins et al., 2014). The most likely impacts for most prey species
in a given area would be temporary avoidance of the area. Surveys using
towed airgun arrays move through an area relatively quickly, limiting
exposure to multiple impulsive sounds. In all cases, sound levels would
return to ambient once a survey ends and the noise source is shut down
and, when exposure to sound ends, behavioral and/or physiological
responses are expected to end relatively quickly (McCauley et al.,
2000b). The duration of fish avoidance of a given area after survey
effort stops is unknown, but a rapid return to normal recruitment,
distribution, and behavior is anticipated. While the potential for
disruption of spawning aggregations or schools of important prey
species can be meaningful on a local scale, the mobile and temporary
nature of most surveys and the likelihood of temporary avoidance
behavior suggest that impacts would be minor.
Acoustic Habitat--Acoustic habitat is the soundscape--which
encompasses all of the sound present in a particular location and time,
as a whole--when considered from the perspective of the animals
experiencing it. Animals produce sound for, or listen for sounds
produced by, conspecifics (communication during feeding, mating, and
other social activities), other animals (finding prey or avoiding
predators), and the physical environment (finding suitable habitats,
navigating). Together, sounds made by animals and the geophysical
environment (e.g., produced by earthquakes, lightning, wind, rain,
waves) make up the natural contributions to the total acoustics of a
place. These acoustic conditions, termed acoustic habitat, are one
attribute of an animal's total habitat.
[[Page 29243]]
Soundscapes are also defined by, and acoustic habitat influenced
by, the total contribution of anthropogenic sound. This may include
incidental emissions from sources such as vessel traffic, or may be
intentionally introduced to the marine environment for data acquisition
purposes (as in the use of airgun arrays). Anthropogenic noise varies
widely in its frequency content, duration, and loudness and these
characteristics greatly influence the potential habitat-mediated
effects to marine mammals (please also see the previous discussion on
masking in the ``Acoustic Effects'' subsection), which may range from
local effects for brief periods of time to chronic effects over large
areas and for long durations. Depending on the extent of effects to
habitat, animals may alter their communications signals (thereby
potentially expending additional energy) or miss acoustic cues (either
conspecific or adventitious). Problems arising from a failure to detect
cues are more likely to occur when noise stimuli are chronic and
overlap with biologically relevant cues used for communication,
orientation, and predator/prey detection (Francis and Barber, 2013).
For more detail on these concepts see, e.g., Barber et al., 2009;
Pijanowski et al., 2011; Francis and Barber, 2013; Lillis et al., 2014.
The term ``listening area'' refers to the region of ocean over
which sources of sound can be detected by an animal at the center of
the space. Loss of communication space concerns the area over which a
specific animal signal, used to communicate with conspecifics in
biologically-important contexts (e.g., foraging, mating), can be heard,
in noisier relative to quieter conditions (Clark et al., 2009). Lost
listening area concerns the more generalized contraction of the range
over which animals would be able to detect a variety of signals of
biological importance, including eavesdropping on predators and prey
(Barber et al., 2009). Such metrics do not, in and of themselves,
document fitness consequences for the marine animals that live in
chronically noisy environments. Long-term population-level consequences
mediated through changes in the ultimate survival and reproductive
success of individuals are difficult to study, and particularly so
underwater. However, it is increasingly well documented that aquatic
species rely on qualities of natural acoustic habitats, with
researchers quantifying reduced detection of important ecological cues
(e.g., Francis and Barber, 2013; Slabbekoorn et al., 2010) as well as
survivorship consequences in several species (e.g., Simpson et al.,
2014; Nedelec et al., 2015).
Specific to the GOM and the activities considered here, Matthews et
al. (2016, 2017) developed a first-order cumulative and chronic effects
assessment for noise produced by oil and gas exploration activities in
the U.S. GOM. The 2016 report was originally presented as Appendix K in
BOEM (2017), with an addendum to the report produce in 2017; both are
available online at: www.fisheries.noaa.gov/national/marine-mammal-protection/incidental-take-authorizations-oil-and-gas. Here, we
summarize the study and its findings (referred to here as ``the CCE
report''). For full methodological details and results, please see the
report.
As discussed previously in this section, direct exposure to the
pulses produced by airguns can result in acute impacts at close ranges.
However, low-frequency dominant airgun noise undergoes multiple
reflections at the ocean bottom and surface and refraction through the
water column, both of which cause prolonged decay time of the original
acoustic signals (Urick, 1984). Extended decay time can lead to high
sound levels lasting from one impulse to the onset of the next,
elevating ambient noise levels (Guan et al., 2015). In addition, low-
frequency energy from airgun surveys, with access to conductive
propagation conditions (e.g., deeper waters), has been documented to
travel long distances, contributing to increased background noise over
very large areas (Nieukirk et al., 2012). Implications for acoustic
masking and reduced communication space resulting from noise produced
by airgun surveys are expected to be particularly heightened for
animals that actively produce low frequency sounds or whose hearing is
attuned to lower frequencies. Bryde's whales are the only GOM species
classified within the low-frequency hearing group, producing calls that
span a low frequency range that directly overlaps the dominant energies
produced by airguns. However, impacts associated with cumulative noise
within the frequencies of the Matthews et al. (2016) study (10-5,000
Hz), are relevant to the majority of cetacean species in the GOM. In
the addendum to the CCE report (Matthews et al., 2017), the same
methods for calculating changes in communication space were applied to
sperm whales (based on male sperm whale slow-clicks; Madsen et al.,
2002b).
Acoustic modeling was conducted for ten locations (``receiver
sites'') within the study area to examine aggregate noise produced over
a full year. The locations of the receiver sites are given in Table 5
and shown in the map of Figure 4. These sites were chosen to reflect
areas of biological importance to cetaceans, (e.g., LaBrecque et al.,
2015), areas of high densities of cetaceans (Roberts et al., 2016), and
areas of key biological diversity (e.g., National Marine Sanctuaries).
The study area was divided into six ``activity zones'' (Figure 4) (note
that these zones are different from those used for acoustic exposure
modeling and described below in the ``Estimated Take'' section).
Table 5--Modeled Receiver Site Locations, Water Depths, and Selection Basis
--------------------------------------------------------------------------------------------------------------------------------------------------------
Water depth
Site Receiver site Latitude Longitude (m) Selection basis
--------------------------------------------------------------------------------------------------------------------------------------------------------
1................................. Western GOM.......... 27.01606[deg] N........... 95.7405[deg] W............ 842 Higher density
cryptic deep diving
and social pelagic
cetaceans.
2................................. Florida Escarpment... 25.95807[deg] N........... 84.6956[deg] W............ 693 Higher density
multiple cetacean
species shelf break
and slope.
3................................. Midwestern GOM....... 27.43300[deg] N........... 92.1200[deg] W............ 830 Higher density
multiple cetacean
species shelf break
and slope.
4................................. Sperm whale site..... 24.34771[deg] N........... 83.7727[deg] W............ 1,053 Higher density sperm
whales and cryptic
deep diving
cetaceans.
5................................. Deep offshore........ 27.64026[deg] N........... 87.0285[deg] W............ 3,050 Location of NOAA
noise reference
station.
6................................. Mississippi Canyon... 28.15455[deg] N........... 89.3971[deg] W............ 1,106 Higher density sperm
whales and cryptic
deep diving
cetaceans.
7................................. Bryde's whale site... 28.74043[deg] N........... 85.7302[deg] W............ 212 Bryde's whale
biologically
important area.
[[Page 29244]]
8................................. De Soto Canyon....... 29.14145[deg] N........... 87.1762[deg] W............ 919 Higher density sperm
whales and cryptic
deep diving
cetaceans.
9................................. Flower Garden Banks 27.86713[deg] N........... 93.8259[deg] W............ 88 National Marine
National Marine Sanctuary.
Sanctuary.
10................................ Bottlenose dolphin 29.40526[deg] N........... 93.3247[deg] W............ 12 Bottlenose dolphin
site. biologically
important area.
--------------------------------------------------------------------------------------------------------------------------------------------------------
Note that ``closure areas'' depicted in Figure 4 represent those
described in Chapter 2.8 of BOEM (2017), which are in some cases
different from those described in this document (see the ``Proposed
Mitigation'' section). Matthews et al. (2016, 2017) analyzed multiple
scenarios, including a baseline scenario (referred to in the CCE report
as ``Alternative A'') in which no geophysical surveys are conducted and
noise consists of natural sounds and a minimum estimate of commercial
vessel noise; a survey activity scenario (referred to in the CCE report
as ``Alternative C'') in which projected activities were uniformly
distributed throughout the study area, with the exception of the
coastal waters restriction from February to May (as described below in
the ``Proposed Mitigation'' section); and a closure scenario (referred
to in the CCE report as ``Alternative F1'') in which no activities are
conducted in the restriction areas, 25 percent of the activity that
would have occurred in the restriction areas is redistributed into non-
restriction areas of the same activity zone (Figure 4), and 75 percent
of the activities that would have occurred in the restriction areas are
not conducted at all. Matthews et al. (2016, 2017) also assessed
additional scenarios not relevant to this proposed rulemaking; these
are not discussed here.
[GRAPHIC] [TIFF OMITTED] TP22JN18.003
Several simplifying assumptions were necessary. Changes in the
distribution of survey activities would result in differences in the
relative amount of noise accumulating at different receiver sites, and
that variance was not examined. Instead, results associated with zone-
varying densities of activity types but homogenous distributions of
activities of each type within zones were presented. The approach
applied accounts for spatial variance in resulting cumulative noise due
to factors affecting sound propagation (e.g., topography, bottom type)
among locations of key management interest in the region. However, it
does not produce results for additional locations (e.g., a uniform
map).
The average of the projected annual amounts of survey activities
for ten
[[Page 29245]]
years in each zone (Table 1) was calculated from the total survey line
length within the respective zones. These average activity levels were
modified by implementing area restrictions. Two representative acoustic
sources were modeled and applied to five total activity types: Various
configurations of one or more 8,000 in\3\ airgun arrays were used to
simulate 2D, 3D NAZ, 3D WAZ, and coil surveys, and a single 90 in\3\
airgun was used to simulate boomer and sparker type sources used for
geotechnical surveys (see Table 2 in the CCE report for full details of
these assumptions). Since the specific location of each type of
activity was unknown, the survey source pulses were uniformly
distributed throughout the activity zones according to the projected
amount of each type of survey activity. In order to account for the
seasonal closure of coastal waters, Zones 1, 3, and 5 were separated
into waters occurring within coastal vs. deeper waters at the 20-m
isobath. The numbers of pulses occurring annually within the coastal
versus deeper portions of the zone were titrated to account for only
eight months per year of survey activity within the coastal portion.
The acoustic fields at the receiver sites were modeled at
frequencies from 10 Hz to 5 kHz, for sources up to 500 km away. Results
are provided for three depths as available at each receiver location:
5, 30, and 500 m. Annual cumulative SELs and time-averaged equivalent
SPLs (Leq) at the selected receiver sites were calculated
for all survey activity. A feature of underwater sound propagation is
that nearby sources contribute substantially more SEL than more distant
sources, since the exposure levels decay approximately with the square
of distance from the source. This causes cumulative SEL received from
spatially distributed and moving sources to be dominated by the sources
closest to a receiver. However, the duration of exposures from very
close sources is typically quite short. While exposures from nearby
sources are important for assessing acute effects, their inclusion in a
chronic effects assessment can be misleading. To overcome this issue,
this approach excluded the highest shot exposures received during a
fraction (10 percent) of the total study time period. Thus, the
effective accumulation period was 90 percent of a year. The cumulative
levels estimated using the approach applied in the study are accurate
when the cell dimensions are small, relative to the source-receiver
separation. This approach could have led to errors when survey lines
approached within a few kilometers from the receiver locations;
however, the close range cells where this could have been a problem
were automatically excluded by the removal of the top 10 percent of
pulse noise contributions. Marine mammal hearing frequency weighting
filter coefficients were applied to the received levels, and results
are presented both with and without weighting. Results relevant to this
proposed rule for cumulative SEL (Tables 8 and 10 in the CCE report)
and Leq (Tables 12 and 16 in the CCE report) calculations
are presented in the CCE report.
A baseline ambient noise level must be assumed to estimate lost
listening area and changes in communication space for various levels of
activity. Here, ambient noise levels were defined as some contribution
of commercial shipping noise in the 50-800 Hz band and noise from
natural sounds (produced mainly by wind and waves). The commercial
shipping noise levels were obtained from products available at
cetsound.noaa.gov/sound-index, which provide commercial shipping noise
levels over the GOM region in one third-octave frequency bands between
50-800 Hz (shipping noise was neglected outside this range). Natural
ambient noise levels were calculated from the formulas of Wenz (1962)
and Cato (2008) for a wind speed of 8.5 kn. The natural noise levels
were added to the vessel noise levels to generate composite one third-
octave band ambient levels between 10 Hz and 5 kHz. Broadband ambient
levels varied between 94.3 and 102.3 dB, depending on the receiver
location and depth (Table 7 in the CCE report). Estimates were assigned
to each receiver site based on proximity and matched by water depth.
Tables 13 and 17 in the CCE report present relevant results for modeled
Leq above ambient at each receiver site with and without
frequency weighting.
The lost listening area assessment method has been applied to in-
air noise (Barber et al., 2009) and in soundscape management contexts
(NPS, 2010). Sound sources considered by this method can be from the
same species (as discussed for communication space), a different
species (e.g., predator or prey), natural sounds, or anthropogenic
sounds. The lost listening area method applied by Barber et al. (2009)
calculates a fractional reduction in listening area due to the addition
of anthropogenic noise to ambient noise. It does not provide absolute
areas or volumes of space; however, a benefit of the listening area
method is that it does not rely on source levels of the sounds of
interest. Instead, the method depends on the rate of sound transmission
loss. Such results can be considered with frequency weightings, which
represent the hearing sensitivity variations of three marine mammal
species groups and transmission loss variations with range, or more
generally without weighting. Results are presented as a percentage of
the original listening area remaining due to the increase in noise
levels relative to no activity and between activity scenarios. Relevant
results are presented in Tables 20, 22, and 25 of the CCE report.
The communication space assessment was performed for Bryde's whales
and sperm whales using methods previously implemented for examining
anthropogenic noise effects on whales (Clark et al., 2009; Hatch et
al., 2012). Communication space represents the area within which whales
can detect calls from other whales. For Bryde's whales, all
calculations were performed in the single one third-octave frequency
band centered at 100 Hz, representing the highest received sound levels
for the calls attributed to Bryde's whales in the GOM (Rice et al.,
2014; Sirovic et al., 2014). A one third-octave band sound level of 152
dB at 1 m was specified. An estimate of 12.36 dB signal processing gain
(which accounts for the animal's ability to not only detect but
recognize a signal from an animal of the same species) was applied. The
areas of communication space at each receiver for the Bryde's whale
calls under ambient conditions and under each relevant activity
scenario are presented in Tables 28, 29, and 31 of the CCE report.
Relative losses of communication space (in both areas and percentages)
between the activity scenarios are presented in Table 34 of the CCE
report.
For sperm whales, calculations were performed in the third-octave
frequency band centered at 3,150 Hz, with a specified sound level of
181 dB at 1 m (Madsen et al., 2002b). Sperm whales produce at least
four types of clicks: Usual clicks, buzzes (also called creaks), codas
(patterns of 3-20 clicks), and slow-clicks (or clangs). Sperm whales on
feeding grounds emit slow-clicks in seemingly repetitive temporal
patterns (Oliveira et al., 2013), supporting the hypothesis that their
function is long range communication between males, possibly relaying
information about individual identity or behavioral states. These calls
were chosen for the analysis since they have a lower frequency emphasis
and longer duration than other sperm whale clicks (the center frequency
of usual clicks and buzzes is 15 kHz; Madsen et al., 2002b). Since the
[[Page 29246]]
frequency band of slow-clicks is closest to that of the airgun
activity, these calls are the most affected in the context of the
study. In addition, low-frequency sounds generally propagate farther
than high-frequency ones. Thus, low-frequency communication is
generally more affected by distant noise sources than high-frequency
communication. The signal processing gain was estimated at 3.0 dB,
based on a median frequency bandwidth of 4 kHz and call length of 500
[mu]s (Madsen et al., 2002b). Results for sperm whales are shown in
Table 2 of the CCE report addendum.
In the 3,150 Hz band, noise contribution from airgun survey
activities in the GOM was estimated between 82.0 and 82.1 dB for all
sites and all alternatives, levels similar to the estimated baseline
levels of 82.0 dB at all sites. Therefore, the analysis shows that the
survey activities do not significantly contribute to the soundscape in
the 3,150 Hz band, and that there will be no significant change in
communication space for sperm whales under the modeled alternatives.
Because other sperm whale calls are higher-frequency, they would not be
expected to be affected. However, we must be clear that this analysis
is in reference to potential chronic effects resulting from changes to
effective communication space, and that acute expects, as discussed
elsewhere in this preamble, remain of concern for sperm whales. The
remaining discussion that follows is in reference to the findings for
Bryde's whales and to general findings for other hearing groups.
The lost listening area and communication space metrics do not
reflect variance in an individual animal's experience of the noise
produced by the modeled activities from one moment to the next. With
both sources of noise and animals moving, the time-series of an
individual's noise exposure will show considerable variation. The
methods used by Matthews et al. (2016, 2017) were meant to average the
conditions generated by low-frequency dominant noise sources throughout
a full year, during which animals of key management interest rely on
habitats within the study area. Considered as a complement to
assessments of the acute effects of the same types of noise sources in
the same region (discussed below in the ``Estimated Take'' section),
the CCE assessment estimates noise produced by the same sources over
much larger spatial scales, and considers how the summation of noise
from these sources relates to levels without the proposed activity
(ambient). Approaches such as the communication space estimation
include approximation for the evolved ability of many acoustically
active animals, such as Bryde's whales, to hear the calls of
conspecifics in the presence of some overlapping noise.
At most sites, lost listening area was greater for deeper waters
than for shallower waters, which is attributed to the downward-
refracting sound speed profile near the surface, caused by the
thermocline, which steers sound to deeper depths. The winter sound
speed profile applied in the CCE modeling (February) was considered to
be conservative relative to summer, as it includes a surface sound
channel at certain sites that are conducive to sound propagation from
shallow sound sources. Shallow water noise levels were reduced due to
surface interactions that increase transmission loss, particularly for
low frequencies. Listening area reductions were also generally most
severe when weighted for low-frequency hearing cetaceans. Filters that
more heavily weighted the mid-frequencies modeled in this study (150
Hz-5 kHz) often reduced estimates of lost listening area. Canyon areas
in the central and eastern GOM saw significant loss of listening area.
Both low- and mid-frequency weighted losses were high in the
Mississippi Canyon, while only low-frequency weighted values were high
for the De Soto Canyon. Both of these sites are considered important to
sperm whales as well as other deep diving odontocetes. Other areas
relevant to sperm whales, including site 4 off the Dry Tortugas, also
saw heavy reductions in listening area. Additional heavily affected
sites were those chosen to represent locations with predicted high
densities of cryptic deep divers (e.g., site 1 in the far western GOM).
Though most of these species are classified as having mid-frequency
hearing sensitivity, many have shown sensitivity to airgun noise, with
sperm whales the most well documented in the GOM. These modeling
results suggest that accumulations of noise from survey activities
below 5 kHz and often heightened at depth could be degrading the
availability of animals that forage at great depths in the GOM to use
acoustic cues find prey as well as to maintain conspecific contact.
Comparison between results provided for the two metrics applied in
the CCE report highlights important interpretive differences for
evaluating the biological implications of background noise. The
strength of the communication space approach is that it evaluates
potential contractions in the availability of a signal of documented
importance to a population of animals of key management interest in the
region. In this case, losses of communication space for Bryde's whales
were estimated to be higher in eastern and central GOM canyons and
shelf break areas. The maintenance of listening area and communication
space at site 7 is of particular interest because the location is
within the area of designated biological importance to the Bryde's
whale. The apparent protection of listening area and communication
space within the calling frequencies utilized by the Bryde's whale
appears to take advantage of both local propagation conditions and the
predicted lower levels of survey activity in the shallower portions of
the Eastern Planning Area, which more strongly affect noise levels at
this site. However, the significant loss of low-frequency listening
area and communication space for their calls estimated for in
additional locations, including just off the shelf in the eastern GOM,
is of concern for this population.
The effectiveness of time-area restrictions for maintaining
communication space or listening area were highly variable among
locations. This assessment evaluated the implications of displacing a
portion (25 percent) of the activity that would have taken place within
a restriction area to within the remaining area outside the
restriction. Thus, sites that were within large restriction areas
(sites 6 and 8) experienced reduced cumulative noise levels and
improved listening and communication conditions when those restrictions
were in effect. Conditions at sites within restrictions designed around
biologically important areas (sites 7 and 10) were not improved solely
because they were not degraded under non-restriction conditions. In
contrast, some sites outside restrictions, particularly those located
in deeper water zones that correspond with denser projected levels of
survey activity (sites 1, 3, and 5) experienced higher noise levels
with time-area restrictions, due to activity that was displaced to
within their propagation vicinity. Finally, the methods used in this
assessment to remove 10 percent of shots from survey activity closest
to the receiver locations are likely to have reduced the relative
difference between accumulated energy resulting from smaller
restrictions (which further eliminated shots that would have taken
place within the 160 dB buffered restriction areas). This loss of
resolution between restriction and non-restriction results does not
adequately capture the reduction in acute noise exposure that could be
experienced by animals through implementation of a restriction.
[[Page 29247]]
The CCE report is described here in order to present information
regarding potential longer-term and wider-range noise effects from
sources such as airguns. The metrics applied in this study do not, in
and of themselves, document the consequences of lost listening area or
communication space for the survivorship or reproductive success of
individual animals. However, they do translate a growing body of
scientific evidence for concern regarding the degradation of the
quality of high-value acoustic habitats into quantifiable attributes
that can related to baseline conditions, including those to which
animals have evolved.
In general, losses of broadband listening area far exceeded losses
of communication space when evaluated at the same locations and under
the same activity levels. This is appropriate to the interpretive role
of the lost listening space calculation, which is to provide a more
conservative estimate of the areas over which animals have access to a
variety of acoustic cues of importance to their survival and
reproductive success. Acoustic cues provide particularly important
information in areas where other sensory cues are diminished (e.g.,
dark) and where navigation is challenging (e.g., complex coastlines and
topography). Documentation of such cues (e.g., Barber et al., 2009;
Slabbekoorn et al., 2010) indicate that they can be well outside of the
frequencies that animals use to communicate with conspecifics, are
often of lower source levels than conspecific calls and in many cases
cannot benefit from evolved capacity to compensate for noise (e.g.,
gain applied to communication space calculations), due to the absence
of a mechanism for natural selection to act (e.g., most eavesdropping
contexts). The results of the CCE study highlight the need for further
long-term monitoring in the GOM.
Estimated Take
This section provides an estimate of the number and type of
incidental takes that may be expected to occur under the proposed
activity, which will inform NMFS's negligible impact determination.
Realized incidental takes would be determined by the actual levels of
activity at specific times and places that occur under any issued LOAs.
Harassment is the only type of take expected to result from these
activities. Except with respect to certain activities not pertinent
here, section 3(18) of the MMPA defines ``harassment'' as: Any act of
pursuit, torment, or annoyance which (i) has the potential to injure a
marine mammal or marine mammal stock in the wild (Level A harassment);
or (ii) has the potential to disturb a marine mammal or marine mammal
stock in the wild by causing disruption of behavioral patterns,
including, but not limited to, migration, breathing, nursing, breeding,
feeding, or sheltering (Level B harassment).
Incidental takes would primarily be expected to be by Level B
harassment, as use of the described acoustic sources has the potential
to result in disruption of behavioral patterns for individual marine
mammals. There is also some potential for auditory injury (Level A
harassment) to result for mysticetes and high frequency species due to
the size of the predicted auditory injury zones for those species.
Auditory injury is less likely to occur for mid-frequency species, due
to their relative lack of sensitivity to the frequencies at which the
primary energy of an airgun signal is found, as well as such species'
general lower sensitivity to auditory injury as compared to high-
frequency cetaceans. As discussed in further detail below, we do not
expect auditory injury for mid-frequency cetaceans. The proposed
mitigation and monitoring measures are expected to minimize the
severity of such taking to the extent practicable. No mortality is
anticipated as a result of these activities.
Acoustic Thresholds
Using the best available science, NMFS has developed acoustic
thresholds that identify the received level of underwater sound above
which exposed marine mammals would be reasonably expected to exhibit
behavioral disruptions (equated to Level B harassment) or to incur PTS
of some degree (equated to Level A harassment).
Level B Harassment--Although available data are consistent with the
basic concept that louder sounds evoke more significant behavioral
responses than softer sounds, defining sound levels that disrupt
behavioral patterns is difficult because responses depend on the
context in which the animal receives the sound, including an animal's
behavioral mode when it hears sounds (e.g., feeding, resting, or
migrating), prior experience, and biological factors (e.g., age and
sex). Some species, such as beaked whales, are known to be more highly
sensitive to certain anthropogenic sounds than other species. Other
contextual factors, such as signal characteristics, distance from the
source, and signal to noise ratio, may also help determine response to
a given received level of sound. Therefore, levels at which responses
occur are not necessarily consistent and can be difficult to predict
(Southall et al., 2007; Ellison et al., 2012; Bain and Williams, 2006).
Based on the practical need to use a relatively simple threshold
based on available information that is both predictable and measurable
for most activities, NMFS has historically used a generalized acoustic
threshold based on received level to estimate the onset of Level B
harassment. This approach was developed based on the 1997 High-Energy
Seismic Survey Workshop (HESS, 1999) and a 1998 NMFS workshop on
acoustic criteria, and assumed a step-function threshold. A step-
function threshold assumes that animals receiving SPLs that exceed the
threshold will always respond in a way that constitutes behavioral
harassment, while those receiving SPLs below the threshold will not.
This approach assumes that the responses of marine mammals would not be
affected by differences in acoustic conditions; differences between
species and populations; differences in gender, age, reproductive
status, or social behavior; or the prior experience of the individuals
(or any other contextual factor). For impulsive sources, such as
airguns, a threshold of 160 dB rms SPL was selected on the basis of
measured avoidance responses observed in whales. Specifically, the
threshold was initially derived from data for mother-calf pairs of
migrating gray whales (Malme et al., 1983, 1984) and bowhead whales
(Richardson et al., 1985, 1986) responding when exposed to airguns.
Subsequent data collection has not suggested that the 160-dB value is
generally unrepresentative, inasmuch as a single-value threshold used
to predict behavioral responses across multiple taxa and contexts can
be adequately representative. This threshold was historically
unweighted, meaning that the assessment of potential for behavioral
disturbance does not account for differential hearing sensitivity
across species.
However, most marine mammals exposed to impulse noise demonstrate
responses of varying magnitude in the 140[hyphen]180 dB rms exposure
range (Southall et al., 2007), including the whales studied by Malme et
al. (1983, 1984), and potential disturbance levels at SPLs above 140 dB
rms were also highlighted by HESS (1999). Studies of marine mammals in
the wild and in experimental settings do not support the assumptions
described above for the single step approach--different species of
marine mammals and different individuals of the same species respond
differently to noise exposure. Further,
[[Page 29248]]
studies of animal physiology suggest that gender, age, reproductive
status, and social behavior, among other variables, probably affect how
marine mammals respond to noise exposures (e.g., Wartzok et al., 2003;
Southall et al., 2007; Ellison et al., 2012).
Southall et al. (2007) did not suggest any specific new criteria
due to lack of convergence in the data, instead proposing a severity
scale that increases with sound level as a qualitative scaling
paradigm. Lack of controls, precise measurements, appropriate metrics,
and context dependency of responses all contribute to variability.
Subsequently, Wood et al. (2012) proposed a probabilistic response
function at which 10 percent, 50 percent, and 90 percent of individuals
exposed are assumed to produce a behavioral response at exposures of
140, 160, and 180 dB rms, respectively. It is important to note that
the probabilities associated with the steps identify the proportion of
an exposed population that is likely to respond to an exposure, rather
than an individual's probability of responding. This function is
shifted for species (or contexts) assumed to be more behaviorally
sensitive, e.g., for beaked whales, 50 percent and 90 percent response
probabilities were assumed to occur at 120 and 140 dB rms,
respectively.
In assessing the potential for behavioral response as a result of
sonar exposure, the U.S. Navy has developed, with NMFS, acoustic risk
functions (or ``dose-response'' functions) that relate an exposure to
the probability of response. These assume that the probability of a
response depends first on the ``dose'' (in this case, the received
level of sound) and that the probability of a response increases as the
``dose'' increases (e.g., Dunlop et al., 2017). Based on observations
of various animals, including humans, the relationship represented by
an acoustic risk function is a more robust predictor of the probable
behavioral responses of marine mammals to noise exposure. Similar
approaches are commonly used for assessing the effects of other
``pollutants''. However, no such function has yet been developed for
exposure to noise from acoustic sources other than military sonar.
Defining such a function is difficult due to the complexity resulting
from the array of potential social, environmental, and other contextual
effects described briefly above, as well as because it requires
definition of a ``significant'' response (i.e., one rising to the level
of ``harassment''), which is not well-defined.
NMFS acknowledges that the 160-dB rms step-function approach is
simplistic, and that an approach reflecting a more complex
probabilistic function is better reflective of available scientific
information. Such an approach takes the fundamental step of
acknowledging the potential for Level B harassment at exposures to
received levels below 160 dB rms (as well as the potential that animals
exposed to received levels above 160 dB rms will not respond in ways
constituting behavioral harassment). Zeddies et al. (2015) assessed the
potential for behavioral disturbance of marine mammals as a result of
the specified activities described herein against both the 160 dB rms
step-function and the Wood et al. (2012) approach described above.
Although Wood et al. (2012) also used a modified risk function for
migrating baleen whales due to assumed heightened sensitivity when in
that behavioral state, this approach was deemed not relevant for the
GOM as the only baleen whale present is resident. The modified risk
function for sensitive species was used for beaked whales. While there
has been no direct evaluation of beaked whale sensitivity to noise from
airguns, there is significant evidence of sensitivity by beaked whales
to mid-frequency sonar (Tyack et al., 2011; DeRuiter et al., 2013;
Stimpert et al., 2014; Miller et al., 2015), as well as to vessel noise
(Aguilar Soto et al., 2006; Pirotta et al., 2012).
The approach described by Wood et al. (2012), which we are using
here, also accounts for differential hearing sensitivity by
incorporating frequency-weighting functions. The analysis of Gomez et
al. (2016) indicates that behavioral responses in cetaceans are best
explained by the interaction between sound source type and functional
hearing group. Southall et al. (2007) proposed auditory weighting
functions for species groups based on known and assumed hearing ranges
(Type I). Finneran and Jenkins (2012) developed newer weighting
functions based on perceptual measure of subjective loudness, which
better match the onset of hearing impairment than the original
functions (Type II). However, because data for the equal-loudness
contours do not cover the full frequency range of the Type I filters, a
hybrid approach was proposed. Subsequently, Finneran (2016) recommended
new auditory weighting functions (Type III) which were adopted by NMFS
(2016). While Type III filters are better designed to predict the onset
of auditory injury, as a conservative measure Type I filters were
retained for use in evaluating potential behavioral disturbance in
conjunction with the Wood et al. (2012) probabilistic response
function.
NMFS is currently evaluating available information towards
development of guidance for assessing the effects of anthropogenic
sound on marine mammal behavior. For this specified activity we have
determined it appropriate to use the Zeddies et al. (2015) exposure
estimates produced using the Wood et al. (2012) approach as our basis
for estimating take and considering the effects of the specified
activity on marine mammal behavior.
While we believe that the general approach of Wood et al. (2012)--a
probabilistic risk function that allows for the likelihood of
differential response probability at given received levels on the basis
of multiple factors, including behavioral context, distance from the
source, and particularly sensitive species--is appropriate, we
acknowledge that there is some element of professional judgment
involved in defining the particular steps at which specific response
probabilities are assumed to occur and that this remains a relatively
simplistic approach to a very complex matter. However, we believe that
the Wood et al. (2012) function is consistent with the best available
science, and is therefore an appropriate approach. We are aware of the
recommendations of Nowacek et al. (2015)--i.e., a similar scheme, but
shifted downward with the 50 percent response probability midpoint at
140 dB rms--but disagree that these recommendations are justified by
the available scientific evidence. In fact, our preliminary analysis of
data presented in available studies describing behavioral response to
intermittent sound sources (including airguns and sonar) (e.g., Malme
et al., 1984, 1988; Houser et al., 2013; Antunes et al., 2014; Moretti
et al., 2014), conducted using a non-parametric regression method,
indicates that the 50 percent midpoint is very close to 160 dB rms
(i.e., 159 dB rms). While there may be other recommended iterations of
this basic approach, we address the differences between Wood et al.
(2012) and Nowacek et al. (2015) below.
Both the Wood et al. (2012) and Nowacek et al. (2015) functions
acknowledge that Level B harassment is not a simple one-step function
and responses can occur at received levels below 160 dB rms. The
relevant series of step functions provided within Wood et al. (2012)
for beaked whales (50 percent at 120 dB; 90 percent at 140 dB) and all
other species (10 percent at 140 dB; 50 percent at 160 dB; 90 percent
at
[[Page 29249]]
180 dB) attempt to provide a more realistic behavioral paradigm, which
is probabilistic and acknowledges that not all exposures are expected
to yield similar responses for every species and/or behavioral context,
as described above. The differences between Wood et al. (2012) and
Nowacek et al. (2015) stem from how probabilities at corresponding
received level are assigned, with both methodologies seemingly relying
upon professional judgment in interpreting available data to make these
decisions.
Regarding mysticetes, changes in vocalization associated with
exposure to airgun surveys within migratory and non-migratory contexts
have been observed (e.g., Castellote et al., 2012; Blackwell et al.,
2013; Cerchio et al., 2014). The potential for anthropogenic sound to
have impacts over large spatial scales is not surprising for species
with large communication spaces, like mysticetes (e.g., Clark et al.,
2009), although not every change in a vocalization would necessarily
rise to the level of a take. Additionally, because of existing acoustic
monitoring techniques, detecting changes in vocalizations at further
distances from the source is more likely, as opposed to observing other
types of responses (e.g., visible changes in behavior) at these
distances. However, the consideration of these observed vocal responses
is not contrary to Wood et al. (2012). Specifically, Blackwell et al.
(2013) report the onset of changes in vocal behavior for migratory
bowhead whales at received levels that are consistent with those
provided in the Wood et al. (2012) function for migrating mysticete
species (which are not present in the GOM). Cerchio et al. (2014)
observed the number of singing humpback whales in a breeding habitat
decrease in the presence of increasing background received levels
during airgun surveys. However, because the study was opportunistic,
specific information on distances between singers and source vessels,
as well as received levels at the singing whales, could not be
obtained. Nevertheless, some probability of these vocal responses would
likely be captured by the Wood et al. (2012) function for all other
species/behaviors. Moreover, a decision about the appropriateness of a
particular function should be based on how well it reflects the best
available information, rather than on how it affects the resulting
number of takes.
We also acknowledge concern regarding the differences between sperm
whales and other cetaceans in the mid-frequency group, i.e., sperm
whales are believed to be somewhat more sensitive to low-frequency
sound, and Miller et al. (2009) conclude that exposure to noise from
airguns may impact sperm whale foraging behavior. While the available
information provides a basis for concern regarding the effects of
airguns on sperm whales, the onset of changes in buzz rates (i.e.,
indicators of foraging behavior) occur at received levels that are
consistent with the probabilities predicted by the Wood et al. (2012)
function for all other species/behaviors. Moreover, the probabilistic
function recommended by Nowacek et al. (2015) likewise does not make
distinctions between any species or species groups, including sperm
whales (i.e., Nowacek et al. (2015) offers a single function for all
species and contexts). Therefore, Nowacek et al. (2015) offers no
advantage in this regard.
Additionally, the application of the Nowacek et al. (2015) approach
disregards the important role that distance from a source plays in the
likelihood that an animal will respond to a given received level from
that source type in a particular manner. By assuming, for example, a 50
percent midpoint at 140 dB rms, the approach implies an unrealistically
high probability of marine mammal response to signals received at very
far distances from a source (e.g., greater than 50 km). DeRuiter et al.
(2013) found that beaked whales exposed to similar received levels
responded when the sound was coming from a closer source and did not
respond to the same level received from a distant source. Although the
Wood et al. (2012) approach does not specifically include a distance
cut-off, the distances at which marine mammals are predicted to respond
better comport with the distances at which behavioral responses have
been detected and reported in the literature.
Finally, other than providing the 50 percent midpoint, Nowacek et
al. (2015) offer minimal detail on how their recommended probabilistic
function should be derived and/or implemented, and provide no
quantitative recommendations for acknowledging that behavioral
responses can vary by species group and/or behavioral context. For
example, relying upon Nowacek et al. (2015), in comparison with Wood et
al. (2012), does not adequately acknowledge that beaked whales are
known to be particularly sensitive and behavioral impacts would be
underestimated. The behavioral harassment criteria upon which the
analysis presented herein is based are presented in Table 6.
Table 6--Behavioral Exposure Criteria
----------------------------------------------------------------------------------------------------------------
Probability of response to frequency-weighted rms
SPL
Group ---------------------------------------------------
120 140 160 180
----------------------------------------------------------------------------------------------------------------
Beaked whales............................................... 50% 90% n/a n/a
All other species........................................... n/a 10% 50% 90%
----------------------------------------------------------------------------------------------------------------
Level A Harassment--NMFS's Technical Guidance for Assessing the
Effects of Anthropogenic Sound on Marine Mammal Hearing (NMFS, 2016)
identifies dual criteria to assess the potential for auditory injury
(Level A harassment) to occur for different marine mammal groups (based
on hearing sensitivity) as a result of exposure to noise. The technical
guidance identifies the received levels, or thresholds, above which
individual marine mammals are predicted to experience changes in their
hearing sensitivity for all underwater anthropogenic sound sources, and
reflects the best available science on the potential for noise to
affect auditory sensitivity by:
Dividing sound sources into two groups (i.e., impulsive
and non-impulsive) based on their potential to affect hearing
sensitivity;
Choosing metrics that best address the impacts of noise on
hearing sensitivity, i.e., peak sound pressure level (peak SPL)
(reflects the physical properties of impulsive sound sources to affect
hearing sensitivity) and cumulative sound exposure level (cSEL)
(accounts for not only level of exposure but also duration of
exposure); and
Dividing marine mammals into hearing groups and developing
auditory weighting functions based on the
[[Page 29250]]
science supporting that not all marine mammals hear and use sound in
the same manner.
The premise of the dual criteria approach is that, while there is
no definitive answer to the question of which acoustic metric is most
appropriate for assessing the potential for injury, both the received
level and duration of received signals are important to an
understanding of the potential for auditory injury. Therefore, peak SPL
is used to define a pressure criterion above which auditory injury is
predicted to occur, regardless of exposure duration (i.e., any single
exposure at or above this level is considered to cause auditory
injury), and cSEL is used to account for the total energy received over
the duration of sound exposure (i.e., both received level and duration
of exposure) (Southall et al., 2007; NMFS, 2016). As a general
principle, whichever criterion is exceeded first (i.e., results in the
largest isopleth) would be used as the effective injury criterion
(i.e., the more precautionary of the criteria). Note that cSEL acoustic
threshold levels incorporate marine mammal auditory weighting
functions, while peak pressure thresholds do not (i.e., flat or
unweighted). Weighting functions for each hearing group (e.g., low-,
mid-, and high-frequency cetaceans) are described in NMFS (2016).
NMFS (2016) recommends 24 hours as a maximum accumulation period
relative to cSEL thresholds. These thresholds were developed by
compiling and synthesizing the best available science, and are provided
in Table 7 below. The references, analysis, and methodology used in the
development of the thresholds are described in NMFS (2016), which is
available online at: www.nmfs.noaa.gov/pr/acoustics/guidelines.htm.
Table 7--Exposure Criteria for Auditory Injury for Impulsive Sources
------------------------------------------------------------------------
Cumulative sound exposure
Peak level \2\
Hearing group pressure ---------------------------
\1\ Non-
Impulsive impulsive
------------------------------------------------------------------------
Low-frequency cetaceans........ 219 dB..... 183 dB...... 199 dB
Mid-frequency cetaceans........ 230 dB..... 185 dB...... 198 dB
High-frequency cetaceans....... 202 dB..... 155 dB...... 173 dB
------------------------------------------------------------------------
\1\ Referenced to 1 [mu]Pa; unweighted within generalized hearing range.
\2\ Referenced to 1 [mu]Pa\2\-s; weighted according to appropriate
auditory weighting function. All airguns and the boomer are treated as
impulsive sources; other HRG sources are treated as non-impulsive.
The technical guidance was classified as a Highly Influential
Scientific Assessment and, as such, underwent three independent peer
reviews, at three different stages in its development, including a
follow-up to one of the peer reviews, prior to its dissemination by
NMFS. Details of each peer review are included within the technical
guidance, and specific peer reviewer comments and NMFS's responses are
available online at: www.nmfs.noaa.gov/pr/acoustics/guidelines.htm. In
addition, there were three separate public comment periods. Responses
to public comments were provided in a previous Federal Register notice
(81 FR 51694; August 4, 2016). At this time, NMFS considers the
technical guidance to represent the best available scientific
information. Therefore, we are not soliciting and will not respond to
comments concerning the contents of the technical guidance, as such
comments are outside the scope of this proposed rule. NMFS recently
provided a fourth opportunity for review of the technical guidance (82
FR 24950; May 31, 2017) for the specific purpose of soliciting input to
assist in review of the technical guidance pursuant to Executive Order
13795.
Modeling Overview
Zeddies et al. (2015, 2017a) (i.e., ``the modeling report'')
provides estimates of the annual marine mammal acoustic exposure caused
by sounds from geophysical survey activity in the GOM for ten years of
notional activity levels (Table 1). Here we provide a brief overview of
key modeling elements, with more detail provided in the following
sections. Significant portions of the following discussion represent
incorporation by reference of Zeddies et al. (2015) and, for full
details of the modeling effort, the interested reader should see the
report (available online at: www.fisheries.noaa.gov/national/marine-mammal-protection/incidental-take-authorizations-oil-and-gas). The
original modeling report (Zeddies et al., 2015) evaluated the potential
for auditory injury using criteria described by Southall et al. (2007)
and Finneran and Jenkins (2012), with some appropriate modifications.
Following completion of NMFS's technical guidance (NMFS, 2016), the
original exposure modeling results for auditory injury were updated
using the frequency-weighting functions and associated thresholds
described in NMFS (2016) (Zeddies et al., 2017a).
A modeling workshop was held in 2014 as a collaborative effort
between the American Petroleum Institute (API) and the International
Association of Geophysical Contractors (IAGC), NMFS, and BOEM. The
objectives of the workshop were to identify: (1) Gaps in modeling sound
fields from airgun arrays and other active acoustic sources, including
data requirements and performance in various contexts, (2) gaps in
approaches to integration of modeled sound fields with biological data
to estimate marine mammal exposures, and (3) assumptions and
uncertainties in approaches and resultant effects on exposure
estimates. This workshop aided BOEM and NMFS's development of a Request
for Proposals, Statement of Work, and, ultimately, the methodologies
undertaken in the modeling project.
The project was divided into two phases. Each phase produced
exposure estimates computed from modeled sound levels as received by
simulated animals (animats) in a specific modeling area. In Phase I
(described below under ``Test Scenarios;'' all other discussion here
refers to Phase II), a typical 3D WAZ survey was simulated at two
locations in order to establish the basic methodological approach and
to provide results used to evaluate test scenarios that could influence
exposure estimates. Results from the test scenarios were then used to
guide the main modeling effort of Phase II. In Phase II, the GOM was
divided into seven modeling zones with six survey types simulated
within each zone to estimate the potential effects of each survey.
The zones were designed as described previously (``Description of
the Specified Activity;'' Figure 2)--shelf and slope waters were
divided into eastern, central and, western zones, plus a single deep-
water zone--to account for both the geospatial dependence of acoustic
fields and the geographic variations of animal distributions. The
selected boundaries considered sound propagation conditions and species
distribution to create regions of optimized uniformity in both acoustic
environment and animal density. Survey types included deep penetration
surveys using a large airgun array (2D, 3D NAZ, 3D WAZ, and coil),
shallow penetration surveys using a single airgun, and high resolution
surveys concurrently using side-scan sonar, subbottom profiler, and
multibeam echosounder. The results from each zone were summed to
provide GOM-wide estimates of take for each marine mammal species for
each survey type
[[Page 29251]]
for each notional year. To get these annual aggregate exposure
estimates, 24-hr average exposure estimates from each survey type were
multiplied by the number of expected survey days from BOEM's effort
projections. Because these projections are not season-specific, surveys
were assumed to be equally likely to occur at any time of the year and
at any location within a given zone.
Sound Field Modeling
Acoustic source emission levels and directivity of a single airgun
and an airgun array were modeled using JASCO Applied Sciences' Airgun
Array Source Model (AASM). Source levels for high-resolution sources
were obtained from manufacturer's specifications for representative
sources. The AASM accounts for the physics of oscillation and radiation
of airgun bubbles (Ziolkowski, 1970) and nonlinear pressure
interactions between airguns, port throttling, bubble damping, and
generator-injector gun behavior (Dragoset, 1984; Laws et al., 1990;
Landro, 1992). The model was originally fit to a large library of
empirical airgun data, consisting of measured signatures of Bolt 600/B
airguns ranging in volume from 5 to 185 in\3\. Airgun signatures have a
random component at higher frequencies that cannot be predicted using a
deterministic model; therefore, AASM uses a stochastic simulation to
predict the high-frequency components based on a statistical analysis
of a large collection of airgun source signature data (maintained by
the International Association of Oil and Gas Producers' Joint Industry
Programme). AASM is capable of predicting airgun source levels at
frequencies up to 25 kHz, and produces a set of notional signatures for
each array element based on array layout; volume, tow depth, and firing
pressure for each element; and interactions between different elements
in the array. The signatures are summed to obtain the far-field source
signature of the entire array in the horizontal plane, which is then
filtered into one third-octave frequency bands to compute the source
levels of the array as a function of frequency band and azimuthal angle
in the horizontal plane (at the source depth), after which it is
considered to be an azimuth-dependent directional point source in the
far field. Electromechanical sources were modeled on the basis of
transducer beam theory, which is often used to estimate beam pattern of
the source in the absence of field measurements, and which is described
in detail in the modeling report.
It should be noted that source modeling for the boomer source was
compared to that for the single airgun. Results of the comparison
indicate that the acoustic field modeling results for the airgun
adequately approximate the ones for the boomer. Considering the
negligible fraction of total surveys conducted using boomers and that
the estimated impact from the single airgun is always greater than for
the boomer, the single airgun results were used as a conservative
substitute for the boomer.
Underwater sound propagation (i.e., transmission loss) as a
function of range from each source was modeled using JASCO Applied
Sciences' Marine Operations Noise Model (MONM) for multiple propagation
radials centered at the source to yield 3D transmission loss fields in
the surrounding area. The MONM computes received per-pulse SEL for
directional sources at specified depths. MONM uses two separate models
to estimate transmission loss.
At frequencies less than 2 kHz, MONM computes acoustic propagation
via a wide-angle parabolic equation (PE) solution to the acoustic wave
equation (Collins, 1993) based on a version of the U.S. Naval Research
Laboratory's Range-dependent Acoustic Model (RAM) modified to account
for an elastic seabed (Zhang and Tindle, 1995). MONM-RAM incorporates
bathymetry, underwater sound speed as a function of depth, and a
geoacoustic profile based on seafloor composition, and accounts for
source horizontal directivity. The PE method has been extensively
benchmarked and is widely employed in the underwater acoustics
community (Collins et al., 1996), and MONM-RAM's predictions have been
validated against experimental data in several underwater acoustic
measurement programs conducted by JASCO (e.g., Aerts et al., 2008; Funk
et al., 2008; Ireland et al., 2009; Blees et al., 2010; Warner et al.,
2010). At frequencies greater than 2 kHz, MONM accounts for increased
sound attenuation due to volume absorption at higher frequencies
(Fisher and Simmons, 1977) with the widely-used BELLHOP Gaussian beam
ray-trace propagation model (Porter and Lui, 1994). This component
incorporates bathymetry and underwater sound speed as a function of
depth with a simplified representation of the sea bottom, as subbottom
layers have a negligible influence on the propagation of acoustic waves
with frequencies above 1 kHz. MONM-BELLHOP accounts for horizontal
directivity of the source and vertical variation of the source beam
pattern. Both propagation models account for full exposure from a
direct acoustic wave, as well as exposure from acoustic wave
reflections and refractions (i.e., multi-path arrivals at the
receiver).
These propagation models effectively assume a continuous wave
source, which is an acceptable assumption for a pulse in the case of
the SEL metric because the energy in the various multi-path arrivals is
summed. When significant multi-path arrivals cause broadening of the
pulse, the continuous wave assumption breaks down for pressure metrics
such as rms SPL. Multipath arrivals can have very different temporal
and spectral properties when received by marine mammals (Madsen et al.,
2006b).
Models are more efficient at estimating SEL than rms SPL.
Therefore, conversions may be necessary to derive the corresponding rms
SPL. Propagation was modeled for a subset of sites using a full-wave
RAM PE model (FWRAM), from which broadband SEL to SPL conversion
factors were calculated using a sliding 100 ms integration window. This
window was selected to represent the shortest expected temporal
integration time for the mammalian ear (Plomp and Bouman, 1959;
MacGillivray et al., 2014). The FWRAM required intensive calculation
for each site, thus a representative subset of modeling sites were used
to develop azimuth-, range-, and depth-dependent conversion factors.
These conversion factors were used to calculate the broadband rms SPL
from the broadband SEL prediction at all the modeling sites. Conversion
factors were calculated for each modeling location.
For electromechanical source and single airgun propagation
modeling, a fixed conversion difference of +10 dB from SEL to rms SPL
was applied at all receiver positions, because there was little
variability over the range of propagation for these sources. This
approach is accurate at distances where the pulse duration is less than
100 ms, and conservative for longer distances. Most of the effects of
these sources occur at relatively short distances where the pulse
durations are short so this approach is not expected to be overly
conservative even for lower-level effects. This is a conservative but
reasonable approximation to simplify the variability across all HRG
sources, effectively assuming that an HRG transmission is on for only
1/10 of a second for any given second.
As described below, in order to accurately estimate exposure a
simulation must adequately cover the various location- and season-
specific environments. The surveys may be conducted at any location
within the planning area and occur at any time of
[[Page 29252]]
the year, so simulations must adequately cover each area and time
period. We previously introduced the seven zones within which potential
exposures were modeled, corresponding with shelf and slope environments
subdivided into western, central, and eastern areas, as well as a
single and deep zone (Figure 2). The subdivision depth definitions are:
Shelf, 0-200 m; slope, 200-2,000 m; and deep, greater than 2,000 m.
Within each of the seven zones, a set of representative survey-
simulation rectangles for each of the survey types was defined, with
larger areas for the ``large-area'' surveys (i.e., deep penetration
airgun) and smaller areas for the ``small-area'' surveys (i.e., shallow
penetration airgun and HRG). In Figure 2, the smaller numbered boxes
represent the survey area extents for the different survey types. The
stars represent acoustic modeling sites along western, central, and
eastern transects (Figure 2).
A set of 30 sites was selected to calculate acoustic propagation
loss grids as functions of source, range from the source, azimuth from
the source, and receiver depth. These were then used as inputs to the
acoustic exposure model. Geographic coordinates and water column depth
of each acoustic modeling site are listed in Table 48 of the modeling
report. The environmental parameters and acoustic propagation
conditions represented by these 30 modeling sites were chosen to be
representative of the prevalent acoustic propagation conditions within
the survey extents. Inputs are as follows:
Water depths throughout the modeled area were obtained
from the National Geophysical Data Center's U.S. Coastal Relief Model
l. Bathymetry data have a horizontal resolution of approximately 80 x
90 m.
The top sections of the sediment cover in the GOM are
represented by layers of unconsolidated sediments at least several
hundred meters thick, with grain size of the surficial sediments
following the general trend for sedimentary basins (decreasing with the
distance from the shore). For the shelf zone, the general surficial
bottom type was assumed to be sand, for the slope zone silt, and for
the deep zone clay. In constructing a geoacoustic model for input to
MONM, a median grain size value was generally selected. Assumed
geoacoustic properties for each zone as a function of depth are
presented in Tables 52-55 of the modeling report.
The sound speed profiles for the modeled sites were
derived from temperature and salinity profiles from the U.S. Naval
Oceanographic Office's Generalized Digital Environmental Model V 3.0
(GDEM). GDEM provides an ocean climatology of temperature and salinity
for the world's oceans on a latitude-longitude grid with 0.25[deg]
resolution, with a temporal resolution of one month, based on global
historical observations from the U.S. Navy's Master Oceanographic
Observational Data Set. The GDEM temperature-salinity profiles were
converted to sound speed profiles.
Variation in the sound speed profile throughout the year was
investigated and a set of 12 sound speed profiles produced, each
representing one month in the shelf, slope, and deep zones. The set was
divided into four seasons and, for each zone, one month was selected to
represent the propagation conditions in the water column in each
season. Acoustic fields were modeled using sound speed profiles for
winter (January-March) and summer (July-September). Profiles for Season
1 (February) provided the most conservative propagation environment
because a surface duct, caused by upward refraction in the top 50-75 m
(of sound above 500 and 250 Hz, respectively), was present. Ducting of
the sound above the relevant frequency cutoffs is important as most
marine mammals are sensitive to these sounds and the horizontal far-
field acoustic projection from the airgun array sources do have
significant energy in this part of the spectrum. Profiles for Season 3
(August or September) provided the least conservative results because
they have weak to no sound channels at the surface and are strongly
downward refracting in the top 200 m. Only the top 100 m of the water
column are affected by the seasonal variation in the sound speed.
Many assumptions are necessary in modeling complex scenarios. When
possible, the most representative data or methods were used. When
necessary, the choices were made to be conservative so as not to
ultimately underestimate potential marine mammal exposures to noise.
Assumptions related to acoustic modeling include:
The environmental input parameters used for
transmission loss modeling were from databases that provide averaged
values with limited spatial and temporal resolution. Sound speed
profiles are averaged seasonal values taken from many sample
locations. Geoacoustic parameters (including sediment type,
thickness, and reflectivity coefficients) and bathymetric grids are
smoothed and averaged to characterize large regions of the seafloor.
Local variability, which can be affected by weather, daily
temperature cycles, and small-scale surface and sediment details,
generally increases signal transmission loss, but was removed by
these averaging processes. As a result, the transmission loss could
in some cases be underestimated and, therefore, the received levels
would be overestimated.
The acoustic propagation model, MONM, used the
horizontal-direction source level for all vertical angles. This may
slightly underestimate the true sound levels in the vertical
directional beam of the array that ensonifies a zone directly under
the array. This is expected to be a minor effect given the small
volume over which the reduction occurs. Additionally, there is a
steep angle limitation in the PE model used in MONM that also leads
to slightly reduced levels directly under the array. The wide-angle
PE that is used in MONM is accurate to at least 70 degrees. The
reduced-level zone is a cone within a few degrees of vertical, which
represents a relatively small water volume that should not
significantly affect results.
Seasons modeled: To account for seasonal variation in
propagation, winter (most conservative) and summer (least
conservative) were both used to calculate exposure estimates.
Propagation during spring and fall was found to be almost identical
to the results for summer, so those seasons were represented with
the summer results. The primary seasonal influence on transmission
loss is the presence of a sound channel, or duct, near the surface
in winter.
Marine Mammal Density Information
The best available scientific information was considered in
conducting marine mammal exposure estimates (the basis for estimating
take). Historically, distance sampling methodology (Buckland et al.,
2001) has been applied to visual line-transect survey data to estimate
abundance within large geographic strata (e.g., Fulling et al., 2003;
Mullin and Fulling, 2004). Design-based surveys that apply such
sampling techniques produce stratified abundance estimates and do not
provide information at appropriate spatiotemporal scales for assessing
environmental risk of a planned survey. To address this issue of scale,
efforts were developed to relate animal observations and environmental
correlates such as sea surface temperature in order to develop
predictive models used to produce fine-scale maps of habitat
suitability (e.g., Waring et al., 2001; Hamazaki, 2002; Best et al.,
2012). However, these studies generally produce relative estimates that
cannot be directly used to quantify potential exposures of marine
mammals to sound, for example. A more recent approach known as density
surface modeling couples traditional distance sampling with
multivariate regression modeling to produce density maps predicted from
fine-scale environmental covariates (e.g., DoN, 2007b; Becker et al.,
2014; Roberts et al., 2016).
[[Page 29253]]
Roberts et al. (2016) provided several key improvements over
information previously available for the GOM, by incorporating NMFS
aerial and shipboard survey data collected over the period 1992-2009;
controlling for the influence of sea state, group size, availability
bias, and perception bias on the probability of making a sighting; and
modeling density from an expanded set of eight physiographic and 16
dynamic oceanographic and biological covariates. There are multiple
reasons why marine mammals may be undetected by observers. Animals are
missed because they are underwater (availability bias) or because they
are available to be seen, but are missed by observers (perception and
detection biases) (e.g., Marsh and Sinclair, 1989). Negative bias on
perception or detection of an available animal may result from
environmental conditions, limitations inherent to the observation
platform, or observer ability. Therefore, failure to correct for these
biases may lead to underestimates of cetacean abundance (as is the case
for NMFS's SARs abundance estimates for the GOM). Additional data was
used to improve detection functions for taxa that were rarely sighted
in specific survey platform configurations. The degree of
underestimation would likely be particularly high for species that
exhibit long dive times or are cryptic, such as sperm whales, beaked
whales, or Kogia spp. In summary, consideration of additional survey
data and an improved modeling strategy allowed for an increased number
of taxa modeled and better spatiotemporal resolutions of the resulting
predictions. More information concerning the Roberts et al. (2016)
models, including the model results and supplementary information for
each model, is available online at seamap.env.duke.edu/models/Duke-EC-GOM-2015/.
In the GOM, there are clear differences in marine mammal
distribution by water depth, i.e., from shelf to slope and from slope
to deep. Division of the modeling area into zones was chosen so that
nominal marine mammal densities remain relatively constant over the
resulting depth intervals. Density of several species varies within the
shelf and slope areas, seemingly correlated with the orientation and
differences in the widths of these areas over the east-west extent of
the project area. Therefore, shelf and slope zones were divided in
western, central, and eastern areas according to BOEM's planning area
boundaries (Figure 2). The minimum, maximum, and mean (and standard
deviation of the mean) zone-specific marine mammal density estimates,
derived from Roberts et al. (2016), are shown in Tables 62-68 of the
modeling report (with density seeding adjustments). Although sperm
whales are sometimes encountered in shallower water, they were depth
restricted in the model to waters greater than 1,000 m. Females are
rarely seen in waters less than 1,000 m (Taylor et al., 2008), and
Wursig (2017) reports a mean encounter depth of 1,732 m, so this is a
reasonable restriction. It is important to note that the Zone 6
densities for Bryde's whales (Table 67 in the modeling report) reflect
the output of an earlier iteration of the Bryde's whale density model.
This earlier iteration predicted the presence of Bryde's whales in Zone
6 (western GOM slope), an area where they are not currently believed to
occur, on the basis of two ambiguous Balaenoptera spp. sightings from
1992. Subsequently, Roberts et al. (2016) revised the model by changing
the modeling period from 1992-2009 to 1994-2009 so that those sightings
were not included, and also added a bivariate smooth of XY to the
model, to concentrate density where sightings were reported (Roberts et
al., 2015c). Based on the results of this revised model, Bryde's whales
would not be expected to occur in Zone 6 and, on this basis, we have
discounted the predicted exposures of Bryde's whales in that zone.
Animal Movement Modeling and Exposure Estimates
The sound received by an animal when near a sound source is a
function of the animal's position relative to the source, and both
source and animals may be moving. To a reasonable approximation, we
know, predict, or specify the location of the sound source, a 3D sound
field around the source, and the expected occurrence of animals within
100 km\2\ grid cells (Roberts et al., 2016). However, because the
specific location of animals within the modeled sound field is unknown,
agent-based animal movement modeling is necessary to complete the
assessment of potential acoustic exposure. Realistic animal movement
within the sound field can be simulated, and repeated random sampling
(Monte Carlo)--achieved by simulating many animals within the
operations area--used to estimate the sound exposure history of animals
during the operation. Animats are randomly placed, or seeded, within
the simulation boundary at a specified density, and the probability of
an event's occurrence is determined by the frequency with which it
occurs in the simulation. Higher densities provide a finer resolution
for an estimate of the probability distribution function (PDF), but
require greater computational resources. To ensure good representation
of the PDF, the animat density is set as high as is practical, with the
resulting PDF then scaled using the real-world animal density (Roberts
et al., 2016) to obtain the real-world number of individuals affected.
Several models for marine mammal movement have been developed
(e.g., Frankel et al., 2002, Gisiner et al., 2006; Donovan et al.,
2013). Animats transition from one state to another, with user-
specified parameters representing simple states, such as the speed or
heading of the animal, or complex states, such as likelihood of an
animal foraging, playing, resting, or traveling. This analysis uses the
Marine Mammal Movement and Behavior (3MB) model (Houser, 2006). 3MB
controls animat movement in horizontal and vertical directions using
sub-models. Travel sub-models determine horizontal movement, including
sub-models for the animats' travel direction and the travel rate (speed
of horizontal movement). Dive sub-models determine vertical movement.
Diving behavior sub-models include ascent and descent rates, maximum
dive depth, bottom following, reversals, and surface interval. Bottom
following describes the animat's behavior when it reaches the seafloor,
for example during a foraging dive. Reversals simulate foraging
behavior by defining the number of vertical excursions the animat makes
after it reaches its maximum dive depth. The surface interval is the
amount of time an animat spends at the surface before diving again. 3MB
allows a user to define multiple behavioral states, which distinguish
between specific subsets of behaviors like shallow and deep dives, or
more general behavioral states such as foraging, resting, and
socializing. The transition probability between these states can be
defined as a probability value and related to the time of day. The
level of detail included depends on the amount of data available for
the species, and on the temporal and spatial framework of the
simulation.
Parameter values to control animat movement are typically
determined using available species-specific behavioral studies, but the
amount and quality of available data varies by species. While available
data often provides a detailed description of the proximate behavior
expected for real individual animals, species with more available
information must be used as surrogates for those without sufficient
available information. In this study, pantropical spotted dolphins are
used as a surrogate for Clymene, spinner, and
[[Page 29254]]
striped dolphins; short-finned pilot whales are surrogates for Fraser's
dolphins, Kogia spp., and melon-headed whales; and rough-toothed
dolphins are surrogates for false killer whales and pygmy killer
whales. Observational data for all remaining species in the study were
sufficient to determine animat movement. The use of surrogate species
is a reasonable assumption for the simulation of proximate or
observable behavior, and it is unlikely that this choice adds more
uncertainty about location preference. Species-specific parameter
values are given in Tables D-1 to D-18 of the modeling report.
Species-specific animats were created with programmed behavioral
parameters describing dive depth, surfacing and dive durations,
swimming speed, course change, and behavioral aversions (e.g., water
too shallow). The programmed animats were then randomly distributed
over a given bounded simulation area; boundaries extend at least one
degree of latitude or longitude beyond the extent of the vessel track
to ensure an adequate number of animats in all directions, and to
ensure that the simulation areas extend beyond the area where
substantial behavioral reactions might be anticipated. Because the
exact positions of sound sources and animals are not known in advance
for proposed activities, multiple runs of realistic predictions are
used to provide statistical validity to the simulated scenarios. Each
species-specific simulation was seeded with approximately 0.1 animats/
km\2\ which, in most cases, represents a higher density of animats in
the simulation than occurs in the real environment. A separate
simulation was created and run for each combination of location, survey
movement pattern, and marine mammal species. Representative survey
patterns were described under ``Detailed Description of Activities.''
During all simulations in this modeling effort, any animat that
left the simulation area as it crossed the simulation boundary was
replaced by a new animat traveling in the same direction and entering
at the opposite boundary. For example, an animat heading north and
crossing the northern boundary of the simulation was replaced by a new
animat heading north and entering at the southern boundary. By
replacing animats in this manner, the animat modeling density remained
constant. Animats were only allowed to be `taken' once during a 24-hr
evaluation period. That is, an animat whose received level exceeds the
peak SPL threshold more than once during an evaluation period was only
counted once. Energy accumulation for SEL occurred throughout the 24-hr
integration period and was reset at the beginning of each period.
Similarly, the maximum received rms SPL was determined for the entirety
of the evaluation period and reset at the beginning of each period.
In Figure 2, the large transparent boxes represent the seven
defined modeling areas (animal simulation extents) within the seven
zones. During the survey simulations, the source was moved within the
smaller survey area extents, but the sound output would ensonify a
larger area (represented by the animal simulation extents). These
animat simulation boxes set the geographic limits of the 3MB
simulation.
For the large-area surveys, injury simulation boxes extend outward
(north, south, east, and west) by 10 km from the survey limits, a
distance over which the unweighted received levels drop below 160 dB
SEL for a single shot. The behavior simulation boxes, on the other
hand, extend outward by 50 km from the survey limits, a distance
necessary to ensure that the animat movement modeling extends out to
where the weighted received levels drop to 120 dB rms SPL or lower, and
below 160 dB SEL for unweighted received levels. Geographic extent of
the boxes is shown in Tables 59-60 of the modeling report.
The received levels for the single airgun and electromechanical
sources drop off much more quickly with range than for the airgun array
sources discussed above. Consequently, the 3MB simulation boxes for the
small-area surveys were extended to 10 km from the center of the survey
in each cardinal direction, a much larger distance than that required
for the received level conditions, but one that supports more realistic
animal movements. Geographic extent of the boxes is shown in Table 61
of the modeling report.
The JASCO Exposure Modeling System (JEMS) combines animal movement
data (i.e., the output from 3MB), with pre-computed acoustic fields.
The JEMS output was the time-history of received levels and slant
ranges (the three dimensional distance between the animat and the
source) for all animats of the 3MB simulation. Animat received levels
and slant ranges are used to determine the risk of acoustic exposure.
JEMS can use any acoustic field data provided as a 3D radial grid.
Source movement and shooting patterns can be defined, and multiple
sources and sound fields used. For impulsive sources, a shooting
pattern based on movement can be defined for each source, with shots
distributed along the vessel track by location (or time). Because the
acoustic environment varies with location, acoustic fields are pre-
computed at selected sites in the simulation area and JEMS chooses the
closest modeled site to the source at each time step. There were many
animats in the simulations and together their received levels represent
the probability, or risk, of exposure for each survey.
All survey simulations were for 7 days and a sliding 4-hr window
approach was used to get the average 24-hr exposure. In this sliding-
windows approach, 42 exposure estimate samples are obtained for each
seven-day simulation, with the mean value then used as the 24-hr
exposure estimate for that survey. The 24-hr exposure levels were then
scaled by the projected level of effort for each survey type (i.e.,
multiplied by the number of days) to calculate associated annual
exposure levels. The number of individual animals expected to exceed
threshold during the 24-hr window is the number of animats exposed to
levels exceeding threshold multiplied by the ratio of real-world animal
density to model animat density.
As described above for acoustic modeling, assumptions and choices
must be made when modeling complex scenarios:
Social grouping: Marine mammals often form social
groups, or pods, that may number in the hundreds of animals.
Although it was found that group size affects the distribution of
the exposure estimates (see Test Scenario 2, below), the mean value
of the exposure estimate was, generally, unchanged. Because the
annual exposure estimates are meant to represent the aggregate of
many surveys conducted in many locations at various times throughout
the year, it is the mean exposure estimates that are most relevant.
For this reason, social group size was not included in the exposure
estimates.
Mitigation procedures, such as shutting down an airgun
array when animals are detected within an established exclusion
zone, can reduce the injury exposure estimates. Mitigation
effectiveness was found to be influenced by several factors, most
importantly the ability to detect the animals within the exclusion
zone. Some species are more easily detected than others, and
detection probability varies with weather and observational set-up.
Weather during any seismic survey is unknown beforehand and
detection probabilities are difficult to predict, so the effects of
mitigation were not included in the exposure estimates (see Test
Scenario 3, below).
Aversion is a context-dependent behavioral response
affected by biological factors, including energetic and reproductive
state, sociality, and health status of individual animals. Animals
may avoid loud or annoying sounds, which could reduce exposure
levels. The effect of aversion itself
[[Page 29255]]
can be considered as a take (Level B harassment) that results in
avoidance of potential for more serious take (Level A harassment).
Currently, too little is known about the factors that lead to
avoidance (or attraction) of sounds to quantify aversive behavior
for these activities when modeling marine mammal exposure to sound
(see Test Scenario 4, below). However, we include an aversion factor
in defining the level of take that may occur, as compared with the
modeled exposure estimates.
Injury--To evaluate the likelihood an animal might be injured as a
result of accumulated sound energy, the cSEL for each animat in the
simulation was calculated. To obtain that animat's cSEL, the SEL an
animat received from each source over the 24-hr integration window was
summed, and the number of animats whose cSEL exceeded the specified
thresholds (Table 7) during the integration window was counted. To
evaluate the likelihood an animal might be injured via exposure to peak
SPL, the range at which the specific peak SPL threshold occurs (Table
7) for each source based on the broadband peak SPL source level was
estimated. For each 24-hr integration window, the number of animats
that came within this range of the source was counted.
Behavior--To evaluate the likelihood an animal might experience
disruption of behavioral patterns (i.e., a ``take''), the number of
animats that received a maximum rms SPL exposure within the specified
step ranges (Table 6) was calculated. The number of animats with a
maximum rms SPL received level categorized into each bin of the step
function was multiplied by the probability of the behavioral response
specific to that range (Table 6). Specifically, 10 percent of animals
exposed to received levels from 140-159 dB rms would be assumed as
``takes,'' while 50 percent exposed to levels between 160-179 dB rms
and 90 percent exposed to levels of 180 dB rms and above would be. The
totals within each bin were then summed as the total estimated number
of exposures above behavioral harassment thresholds. This process was
repeated for each 24-hr integration window.
Potential for disruption of behavioral patterns was also evaluated
using NMFS's standard 160 dB rms criterion. To evaluate this
likelihood, the exposure simulation was set to use unweighted rms SPL
acoustic fields. The number of animats that received an exposure
greater than 160 dB was counted as the number of behavioral responses.
However, note that the modeling report also separately evaluated
exposures at received levels exceeding 180 dB rms; therefore, the true
number of exposures greater than 160 dB rms would be the sum of
separately calculated exposures between 160 and 180 dB and greater than
180 dB. As with the other criteria, the animat received level was reset
at the beginning of each 24-hr integration window. Please see Zeddies
et al. (2015) for exposure results relating to the 160-dB rms
criterion. The methods did not account for potential habituation,
whereby severity of behavioral reactions to a stimulus may be reduced
due to reduced sensitivity in individual animals from repeated exposure
over time. However, we are not aware of any literature suggesting that
marine mammals in the wild and away from areas with consistent
industrial activity (e.g., ports) become habituated to noise or of any
method by which such theoretical habituation could be modeled.
Test Scenarios
As described above, Phase I of the modeling effort involved
preliminary modeling of a typical 3D WAZ survey (all survey parameters
were described under ``Detailed Description of Activities''), which was
simulated at two locations in order to establish the basic
methodological approach and to provide results used to evaluate test
scenarios that could influence exposure estimates. We provide a summary
of each of the six evaluated test scenarios below. For all test
scenarios, please see the modeling report for full details.
Locations considered were both near the Mississippi Canyon,
including a site centered on the slope of the continental shelf break
and a site centered on the deep ocean plain (please see Figure 10 in
Zeddies et al. (2015)). A reduced suite of six representative species
were included in the Phase I effort: Bryde's whale, sperm whale,
Cuvier's beaked whale, bottlenose dolphin, dwarf sperm whale, and
short-finned pilot whale. Bryde's whales and dwarf sperm whales were
chosen as the only low-frequency species in the GOM and as the
representative high-frequency species, respectively. The four mid-
frequency species were chosen to represent various other aspects of
diving and hearing sensitivity. Cuvier's beaked whales are deep-diving
and behaviorally sensitive to sound, while sperm whales are also deep-
diving and are a unique species in the GOM behaviorally. Short-finned
pilot whales and bottlenose dolphins both represent the swimming
behavior of smaller cetaceans with different preferred water depths.
Note that, for this preliminary modeled scenario, density estimates
were obtained from DoN (2007b), as Roberts et al. (2016) was not yet
available. Full details of the preliminary modeling are available in
the modeling report.
To evaluate potential behavioral response, 30-day simulations of
the hypothetical 3D WAZ survey were run at both sites for each of the
species evaluated. The boundaries of the simulation were determined
from transmission loss calculations, and were set at 50 km from the
source.
Test Scenario 1 (Long-duration Surveys and Scaling Methods)--Some
surveys operate (nearly) continuously for months. Evaluating the
potential impacts due to underwater sound exposures from these extended
operations is challenging because assumptions about parameters that are
valid for short-duration simulations may become less valid, or more
varied, as the time period increases. Treating parameters such as sound
velocity profile or large-scale animal movement as constant over longer
durations, as is typically done in shorter duration simulations, could
lead to errors. However, there is no information indicating that
species migrate regularly on a large-scale in the GOM; thus, large-
scale movement was not integrated into the animal movement model.
Therefore, a test scenario was used to evaluate possible systematic
bias in the modeling process, and methods for scaling results from
shorter-duration simulations to longer duration operations were
suggested.
Exposure estimates from 30-day and 5-day simulations, using
different animat seeding values (0.1 and 2.0 animats/km\2\,
respectively), were determined in subsets using a `sliding window' to
find the number of exposures as a function of time. The 30-day
simulation was used to evaluate exposures against the rms SPL criteria,
and the 5-day simulation was used to evaluate exposures against the
peak SPL and cSEL criteria. The length of the sliding window was 24 hr,
advanced by 4 hr, resulting in 174 samples from the 30-day simulation
and 25 samples from the 5-day simulation. A sliding window of 7 days
advancing by 1 day for the 30-day simulation was also evaluated. Bias
in the model was expected to manifest itself as a trend in the exposure
levels as a function of time.
To investigate potential systematic, and possibly unknown, biases
in the modeling procedure, behavioral exposure estimates were
determined for subsets of the simulations. Behavioral exposure
estimates were determined as a function of time by finding the number
of exposures occurring in 24-hr subsets using a sliding window that
advanced in 4-hr increments. Trends
[[Page 29256]]
were evident, particularly at the slope site, but the trends appeared
to be the consequence of survey design, such as changing sound fields
as the vessels move into different acoustic zones. For sperm whales,
there was an additional bias due to their general avoidance of water
depths less than 1000 m. The area of the slope site began at a location
with water depth approximately 1,500 m, but proceeds to depths less
than 200 m. Therefore, fewer sperm whale animats were within exposure
range of the source later in the simulation. To determine if undesired,
and unknown, systematic biases exist in the modeling procedure,
simulations were run with the source stationary and with no limiting
bathymetric constraints. No clear trends were found, indicating that
undesired systematic biases in the modeling procedure, if present, were
small relative to the survey design and would not affect scaling up the
results in time, if applied.
The number of animats exposed to levels exceeding threshold for 24-
hr time periods multiplied by the number of days in the simulations was
compared to the number of animats exposed to levels exceeding threshold
for the entire duration of the simulations. Given that an animat
represents an individual marine mammal, scaling up the 24-hr average
SPL exposure estimates to 30 days greatly overestimates the number of
individual marine mammals exposed to levels exceeding threshold when
determined over the entire simulation (although the estimated instances
of exposure are reasonably accurate). This occurs because animats were
commonly exposed to levels exceeding these thresholds and the
relatively short reset period of 24-hr means that individual animats
were, in effect, counted several times during the scale-up (i.e., on
multiple days) that would only have been counted once when evaluating
over the entire simulation. Comparison between the full-duration
estimate (obtained through modeling the full survey duration) and the
estimate developed through ``scaling'' the 24-hr exposure estimate
allows for better interpretation of the exposure estimates, e.g.,
through a refined estimate of the number of individuals exposed above
behavioral harassment criteria (versus instances of exposure) and the
average number of days on which those exposures occur (described below
in ``Description of Exposure Estimates''). Because SEL is an
accumulation of energy, evaluating over a longer period (e.g., summing
accumulation over 30 days) could result in more animats exposed to
levels exceeding SEL thresholds than when evaluated over a shorter
period (unlike as described above for SPL metrics).
The systematic trends evident in the modeling procedure indicated
that survey design can affect exposure estimates when scaling is used.
Therefore, the minimum duration of a simulation should include all of
the acoustic environments likely to be encountered during the
operation. The test scenario produced the following recommendations,
which were employed during the Phase II modeling effort: (1) Identify
the shortest large-scale animal movement time-period (e.g., seasonal
migration); (2) Identify acoustic environments over which the survey
will occur (e.g., shallow, slope, deep, and associated geoacoustic
parameters); (3) Identify the minimum period of validity for the
acoustic model (e.g., month due to changing sound velocity profile);
(4) Break the survey into parts that are shorter in duration than both
large-scale animal movement times and the period of acoustic model
validity; (5) Create animal movement simulations for acoustic exposure
with adequate duration to meaningfully sample the exposure-estimating
parameter (e.g., for a 24-hr reset period, enough samples should be
obtained to get a reliable mean value given the various acoustic
environments); (6) If the simulation time is less than the duration of
the survey parts determined in Step 4, then scale the results by the
ratio of survey duration to simulation time (e.g., if the simulation
time is one week, but the survey division is 28 days, then multiply the
simulation exposure results by four); and (7) Sum, or aggregate, the
results from the survey parts to calculate exposures for the entire
survey.
This test scenario also illustrated that knowing the amount of time
that animals are exposed to levels exceeding the threshold criteria can
provide additional information about the potential impacts of the
activity. For example, the amounts of time that animats were exposed to
levels exceeding 160 dB rms SPL over the 30-day duration were
approximately twice as long as the average times in a 24-hr window, as
it was common for the threshold to be exceeded on multiple separate
occasions. Two factors contributed to the total time thresholds were
exceeded--the amount of time per occasion (i.e., how long an animat was
near the source) and the number of occasions that occur (i.e., how many
times an animat was near a source). The number of occasions was,
essentially, the same item determined when finding the number of
animats with exposures exceeding threshold criteria (the typical use of
the threshold criteria). The number of occasions scales with the
duration of the evaluation period, but the time per occasion does not,
and is specific to how an individual animat interacted with a source.
Information provided through this investigation was used to derive
scaler values (described below in ``Description of Exposure
Estimates'') for use in determining the expected number of individuals
represented by a sum total of exposures generated through the scaling
of 24-hr exposures up to match the total duration of a modeled survey.
Test Scenario 2 (Sources and Effects of Uncertainty)--The modeling
process requires the use of simplifying assumptions about oceanographic
parameters, seabed parameters, and animal behaviors. These assumptions
carry some uncertainty, which may lead to uncertainty in the form of
variance or error in individual model outputs and in the final
estimates of marine mammal acoustic exposures. For example, acoustic
propagation models assume a specific shape of the sound speed profile
in the ocean (speed of sound versus depth) for each season. We know,
however, that the real sound speed profile regularly changes and that
substantial variation within a season is possible. The assumption that
a single profile represents the environment through a full season
approximates real-world cases but can, to some degree, cause errors.
The uncertainty in model outputs caused by approximations like this can
be investigated by examining how much the outputs change when the
inputs are purposely offset. ``Parametric uncertainty analysis''
provides a means to characterize the accuracy, or uncertainty, of the
model results in light of errors in model inputs and can also be used
to characterize the expected variability in model results due to
natural variations in some of the input parameters. Use of resampling
techniques can quantify the effects of uncertainty in exposure
estimates due to uncertainty in acoustic and animal movement models.
Uncertainty related to acoustic modeling can be introduced through
source characterization modeling; acoustic propagation modeling; and
selection of inputs for sound speed profiles, geoacoustic parameters,
bathymetry, and sea state. Uncertainty in animal modeling can be
introduced through incomplete knowledge regarding animal locations and
behavioral/motivational states. Both the uncertainty in acoustic
modeling and uncertainty in the animal modeling
[[Page 29257]]
contribute to overall uncertainty in the exposure estimates. Please see
the modeling report for full details of these investigations.
Zeddies et al. (2015) describe uncertainties in the acoustic field
as representing a multi-dimensional envelope that can be wrapped around
the main modeling results. This envelope is meant to enclose the
modeled acoustic field and the real world acoustic field. The
uncertainties in the different dimensions of this envelope (sound speed
profile, geoacoustics, bathymetry, and sea state) cannot be summed to
yield a ``total'' uncertainty as this would be a meaningless quantity.
The overall uncertainty is measured for the volume of the multi-
dimensional uncertainty envelope, but this is a difficult concept to
use in operational planning. The best way to visualize the overall
uncertainty is in terms of the different dimensions of the uncertainty
envelope, which range from inconsequential (e.g., effects of sea state)
to greater than 10 dB between median and maximum propagation scenarios
in the shelf zone due to uncertainty in the sound speed profile.
With regard to uncertainty relating to animal movement parameters,
comparisons between animals generally resulted in similar exposure
estimates when the same filtering and thresholds were applied. The
exposure estimates for bottlenose dolphins, short-finned pilot whales
and, to some extent, sperm whales were similar. For sperm whales,
however, the behavioral depth restriction for this species (animats do
not enter water depths less than 1,000 m) resulted in differences.
Sperm whales also showed greater potential of behavioral response to
noise exposure than other species with the same auditory thresholds.
Sperm whales are deep divers; in this downward refracting environment
they appear to receive consistently greater exposures relative to
shallow diving species.
In order to address overall uncertainty in the exposure estimates
resulting from combined uncertainty due to both acoustic and animal
modeling, a ``bootstrap'' resampling process was used in which relevant
uncertainty could be added to animats' received levels. For example,
for potential auditory injury, the primary acoustic uncertainty was the
source level variance. Airguns are designed to have low inter-shot
variability and predicted source levels within 3 dB. A conservative
estimate of 3 dB standard deviation was used to investigate
the effects of source level variance on SEL injury exposure estimates.
While the mean number of animats above SEL threshold increased relative
to the expected value, the exposure estimate distributions did not
change much. For potential behavioral disturbance, propagation
uncertainty (due to the greater ranges involved) also contributes to
the uncertainty in the acoustic modeling predictions; therefore, 6 dB
was chosen as a test to include both the source variance plus
uncertainty due to propagation. The mean behavioral disruption
estimates and the distribution ranges stayed approximately the same
when 6dB of acoustic variability was included. During
resampling, acoustic uncertainty can be combined with real-world
density (mean standard deviation) and social group size
(mean standard deviation). In general, the uncertainty
associated with the animals (density and group size) does not change
the mean exposure estimate, but can affect the exposure estimate
distribution.
Test Scenario 3 (Mitigation Effectiveness)--With reference to
detection-based mitigation, effectiveness at reducing marine mammal
exposure to potentially injurious sound levels is unknown. Mitigation
effectiveness corresponds with the ability to detect an animal in the
relevant zone. Detectability, and consequently mitigation efficacy,
depends on the species, potentially individual animal characteristics,
survey configuration, and environmental conditions. Mitigation
effectiveness was evaluated using a modeling approach to quantify the
potential reduction in the numbers of exposures at or above Level A
harassment thresholds for selected species by comparing acoustic
exposure estimates with and without mitigation (array shutdown). For
each of the six species considered in the preliminary modeling, a range
of detection probabilities (i.e., g(0)) was considered. The positions
of animats in the simulation are known and reported in short time
steps. The detection probability, however, is the probability of
detecting an animal along the trackline as the survey passes through an
area, rather than for an individual time step. For this evaluation,
g(0) is used as estimate of the detection probability for animats near
the surface and close to the vessel.
Level A harassment exposure estimates associated with the 5-day
survey simulation were calculated with and without a mitigation
procedure. Exposure estimates were computed relative to SEL and peak
SPL exposure criteria. Airgun shutdown was modeled by zeroing all
animat received levels when an animat was detected within an exclusion
zone, with detection registered when the horizontal range of an animat
from the source was less than 500 m, its depth was less than 50 m, and
a random draw from a uniform distribution between 0 and 1 indicated
detection. If the random value was less than the assumed g(0), the
detection was registered, the time of the closest point of approach
(CPA) was found, and the received levels for all animats were zeroed
for 30 minutes before and after the CPA. For the purposes of the
simulation, it was assumed that portions of the survey line missed
during shutdown were re-surveyed (i.e., shutdowns result in an increase
in the overall survey duration in order to keep the distance surveyed
the same as the unmitigated case). Shutdown was assumed to occur only
for the source array around which the animat was detected. Other
sources present in the simulation continued operating. Model
simulations were run for detection probabilities of 0.05 to 0.45
(increments of 0.05) and 0.5 to 0.9 (increments of 0.1) to simulate a
reasonable range of probabilities for cryptic species and other
species, respectively.
The inclusion of mitigation procedures in the simulations reduced
the numbers of exposures based on peak SPL criteria for five out of six
species and detection probabilities considered, even though an
extension in the survey period due to line re-shoot was taken into
account. The exception was Bryde's whales, due to low real-world
density values. Mitigation effectiveness, expressed as the reduction in
the number of individual animals exposed, was generally related to
animal densities; species with higher densities were more often exposed
and the reduction in the number of exposures from mitigation was
greater. As expected, the percentage reduction in exposures for species
with relatively high detection probability was higher than the
percentage reduction for species with relatively low detection
probability.
The usefulness of mitigation depends on species characteristics and
environmental conditions, meaning that there is a high degree of
inherent variability (and potential error) involved in attempting to
predict some reduction in potential exposures resulting from mitigation
effectiveness. Reductions due to mitigation for easily-detected species
with large populations may be large in terms of percentage decrease
(assuming shutdown is a required measure) while, for low-density
species that are difficult to detect in rough seas, there may be little
realistic mitigation effect. Further, for deep-diving species with
unreliable
[[Page 29258]]
vocal rates, a very conservative estimate of mitigation effectiveness
should be used. Ultimately, on the basis of these findings,
quantification of mitigation effectiveness was not incorporated into
the Phase II modeling effort (i.e., is not reflected in the modeled
exposure estimates).
Test Scenario 4 (Effects of Aversion)--Animal behavior in response
to sound exposure may vary widely, but if sounds are perceived as a
threat or an annoyance, animals might temporarily or permanently avoid
the area near the source (e.g., Southall et al., 2007; Ellison et al.,
2012)--a phenomenon referred to as aversion. Aversive responses to
sounds are of particular interest here because such behavior could
decrease the number of injuries that result from acoustic exposure in
the real world. If aversion occurs at a received level lower than that
considered an injurious exposure, a decrease in the corresponding
number of estimated exposures above Level A harassment criteria can be
assumed. The degree of aversion and level of onset for aversion,
however, are poorly understood.
As for mitigation effectiveness, a test scenario was investigated
using a modeling approach to quantify the potential reduction in injury
exposure estimates due to aversion. Aversion is simulated as a
reduction in received levels and, because little is known about the
received levels at which animals begin to avert, the sound levels and
probabilities used to evaluate potential behavioral disturbance are
used to approximate aversion. However, it is possible that aversion
could occur at greater or lesser received sound levels, depending on
the context and/or motivation of the animal. It is important to note
that, as considered here, aversion itself can represent a behavioral
disruption; therefore, aversion is only meaningful in reducing the
potential for injury, i.e., those animals that avert may have avoided
Level A harassment, but would have nevertheless experienced Level B
harassment.
Injury exposure estimates associated with the 5-day 3D WAZ
simulation were determined with and without aversion. The difference in
the mean value of the exposure estimate distributions with and without
aversion indicates the effect of aversion on the injury exposure
estimates. Each animat sampled during the bootstrap resampling process
has an associated exposure history, i.e., a time series of received
sound levels arising from relative motion of the source and animat.
These exposure histories were computed assuming the animats' behaviors
were otherwise unaffected by their received sound levels. Each exposure
history was then modified based on received-level dependent
probabilities of averting:
Step 1: For each bootstrap sample, the occurrence of
aversion was determined probabilistically based on the exposure
level and the probability of aversion defined according to the
function described previously (Table 6) for both SEL and peak SPL.
An iteration-specific aversion efficacy was also chosen randomly
from a uniform distribution in the range of 2-10 dB.
Step 2: Animats for which aversion occurred in Step 1
had their received levels adjusted as described in the following
steps. The received levels were unchanged for animats that did not
avert.
Step 3: For an animat entering an averted state, the
aversion level excesses (the levels above the threshold that
prompted aversion) until the end of the aversion episode were
calculated from the difference between the received level at the
start of aversion and the threshold level at which aversion began up
to a maximum of 5 dB.
Step 4: The adjusted received level during aversion was
set to the greater of two quantities: (1) The received level minus
the aversion efficacy (from Step 1), or (2) the threshold level plus
the aversion level excess at the start of aversion (from Step 3).
Adjusted exposure histories were computed separately for each
source, animat, and episode of aversion; each occurrence of aversive
behavior was thus independent. Although the probability of aversion was
defined in terms of the rms SPL, exposure histories were recorded in
terms of the per-pulse SEL. A nominal conversion offset of +10 dB from
SEL to rms SPL was used so the two metrics could be compared.
Cumulative SELs over the 5-day simulation, were weighted using Type I
filters for Bryde's whales and Type II filters for mid- and high-
frequency cetaceans, but behavioral effects were estimated using Type I
filters for all species, with appropriate adjustments made to the 5-day
SEL exposure histories. The mean time spent in an averted state for
four of six species were approximately 18 and 4 min for the slope and
deep sites, respectively. For beaked whales, the means were 41 and 19
min. Too few Bryde's whale animats exceeded threshold to obtain a
reliable statistical measure.
Aversion in the simulations reduced the numbers of exposures based
on peak SPL criteria for most species. Aversion effectiveness, as
measured by the percentage reduction in the exposure estimates, could
be high: Approximately 85 percent for bottlenose dolphins, Cuvier's
beaked whales, short-finned pilot whales, and sperm whales, and 40
percent for dwarf sperm whales. Bryde's whales, whose real-world
densities were so low that no exposures were modeled even in the
absence of aversion, were the exception. The numbers of exposures based
on SEL criteria were near zero for most species even without aversion.
The reduction in exposures was influenced by the criteria used to
estimate exposures and by the assumptions made with respect to aversion
probability. For example, although the real-world densities of dwarf
sperm whales (a high-frequency cetacean) are similar to those for
Cuvier's beaked whales (a mid-frequency cetacean), exposure estimates
and the decrease in number of exposure estimates arising from aversion
were different. The differences in aversion effectiveness reflect
differences in injury threshold criteria and aversion probability.
Ultimately, the effects of aversion were not quantified in the Phase II
modeling due to lack of information regarding species-specific degree
of aversion and level of onset.
Test Scenarios 5-6 (Separation Distance and Simultaneous Source
Firing)--Geophysical surveys using airgun arrays may use survey designs
that involve multiple source vessels separated by tens of meters to
several kilometers, while newer technology has allowed for different
surveys to be performed closer together than previously. Due to the
possibility that the combined sound pressure levels of multiple airgun
arrays operated close to one another could lead to increased noise
effects than would occur with a single source, these scenarios were
designed to address the issue of the aggregate noise produced by
multiple airgun arrays and the potential for those signals to combine
and lead to larger effects.
The investigations found that while SEL increases for overlapping
surveys, injury due to accumulated energy is a rare event, and
threshold exceedance resulted from a few high-level exposures near a
source rather than an accumulation of many lower-level exposures. The
range to injury assessed by peak SPL is up to a few hundred meters and
does not accumulate. Injury in typical airgun surveys, therefore,
occurs mainly because of a close encounter with a single airgun array.
There are practical limits to how close two acquisition lines can be
without one survey source interfering with the other survey's
recordings. Depending on the survey type and the propagation
environment of the area, the stand-off distance between fully
concurrent surveys operating independently may be several tens of
kilometers. If two surveys are conducted in closer proximity, then the
operators will generally agree to
[[Page 29259]]
``time-sharing'' strategies whereby, for example, one survey acquires a
line while the other completes a line turn with the source inactive, or
similar ways of minimizing the amount of missed effort. Effects of
overlapping surveys on injury exposure estimates are unlikely.
For potential behavioral disturbance, overlapping surveys may
affect exposure estimates, but the effect is either small or
potentially negative (reducing the overall number of estimated
exposures). Because coincident reception in which the sound level
increases appreciably occurs only in small portions of the ensonified
volume, overlapping survey sound fields do not generally result in
higher maximum received sound pressure levels. And, because animals may
only be ``taken'' once within a 24-hr window, animals exposed in more
than one survey are only counted once in the aggregate of the surveys.
This does not preclude possible behavioral effects of animals spending
more time above threshold, but such effects are not addressed by
existing criteria.
From an energetic perspective, the relative firing pattern of
different arrays does not matter. The same SEL will be registered when
two arrays are alternated or fired simultaneously. For the pressure-
based metrics, peak SPL and rms SPL, simultaneous firing can increase
the received levels, but in only a small portion of the ensonified
volume. Because the maximum received levels are rarely increased, the
exposure estimates based on SPL are rarely increased. The most likely
place for meaningful summation to occur is very near the source, and in
that case the firing pattern would be included in the simulation and
therefore in the exposure estimates.
In summary, neither separation distance nor simultaneous firing is
of significant concern when estimating exposures using the current
criteria.
Modeling Issues
NMFS is aware of criticism that the modeling results are
unrealistic or overly conservative (e.g., ``biased modeling based on
flawed assumptions''). For example, we received public comment in
response to our Federal Register notice of receipt of the petition from
the IAGC, API, National Ocean Industries Association, and Offshore
Operators Committee (hereafter referred to as ``the Associations'').
The Associations quote certain statements made by BOEM in its draft
Programmatic EIS (e.g., ``an overly conservative upper limit,''
exposure estimates are ``higher than BOEM expects would actually occur
in a real world environment,'' modeling results represent a ``worst-
case scenario''). NMFS strongly disagrees with these characterizations.
While the modeling required that a number of assumptions and choices be
made by subject matter experts, some of these are purposely
conservative to minimize the likelihood of underestimating the
potential impacts on marine mammals represented by the level of effort
specified by the applicant. The modeling effort incorporated
representative sound sources and projected survey scenarios (both based
on the best available information obtained through BOEM's consultation
with members of industry as well as historical permit application
data), physical and geological oceanographic parameters at multiple
locations within the GOM and during different seasons, the best
available information regarding marine mammal distribution and density,
and available information regarding known behavioral patterns of the
affected species. Current scientific information and state-of-the-art
acoustic propagation and animal movement modeling were used to
reasonably estimate potential exposures to noise. NMFS's position is
that the results of the modeling effort represent a conservative but
reasonable best estimate, not a ``worst-case scenario.''
We call attention to our own public comments submitted to BOEM
following review of the draft PEIS: ``[NMFS] disagrees that the PEIS
analysis is based on the `upper limit' of potential marine mammal
exposures to sound produced by [survey] activities. The PEIS provides
no reasonable justification as to why the exposure estimates [. . .]
should be considered as `conservative upper limits', represent an
`overestimate,' or are `unrealistically high.' [NMFS] believes that the
exposure estimates represent a conservative but reasonable best
estimate [. . . .] [NMFS] disagrees that `each of the inputs into the
models is purposely developed to be conservative.' Although it may be
correct that conservativeness accumulates throughout the analysis, BOEM
has not adequately described the nature of conservativeness associated
with model inputs or to what degree (either quantitatively or
qualitatively) such conservativeness `accumulates.' While exposure
modeling is inherently complex, complexity does not inherently result
in overestimation of exposures [. . . .] [NMFS] strongly disagrees that
the exposure estimates are `overly conservative,' are `upper limits,'
or that these estimates are in some way differentiated from what might
actually be expected to occur.'' Finally, we note that BOEM's final
PEIS removed erroneous statements and provided additional clarification
regarding descriptions of the modeling results to more accurately
describe the nature of the results as a conservative but reasonable
best estimate, consistent with NMFS's comments on the draft PEIS.
IAGC and API contracted with JASCO Applied Sciences, who performed
the modeling effort, to conduct additional analysis regarding the
effect that various acoustic model parameters or inputs have on the
outputs used to estimate numbers of animals exposed to threshold levels
of sound from geophysical sources used in the GOM (``Gulf of Mexico
Acoustic Exposure Model Variable Analysis;'' Zeddies et al., 2017b).
The results of this analysis were not made available to NMFS in time to
fully consider them in preparing these proposed regulations. However,
the report is available online for public review
(www.fisheries.noaa.gov/national/marine-mammal-protection/incidental-take-authorizations-oil-and-gas) and we expect to consider these
results as appropriate in developing a final rule. The primary finding
of Zeddies et al. (2017b) is that use of appropriate acoustic injury
criteria (i.e., NMFS, 2016) and quantitative consideration of animal
aversion and mitigation effectiveness decrease predictions of injurious
exposure. As described herein, we have used acoustic criteria for both
Level A and Level B harassment that reflect the best available science,
and have incorporated reasonable correction for animal aversion.
Here, we address some specific issues regarding the modeling
assumptions and briefly address the results provided by Zeddies et al.
(2017b):
Representative large array. The Associations state that
the selected array (8,000 in\3\) is unrealistically large, resulting in
an overestimation of likely source levels and, therefore, size of the
sound field with which marine mammals would interact. Zeddies et al.
(2017b) evaluated the use of a substitute 4,130 in\3\ array, finding
that reduction in array volume reduces the number of predicted
exposures. Use of a smaller airgun array volume with lower source level
creates a smaller ensonified area resulting in fewer numbers of animals
expected to exceed exposure thresholds.
The particular array was selected as a realistic representative
proxy after BOEM's discussions with individual geophysical companies.
An 8,000-in\3\ array was considered reasonable, as it falls within the
range of typical airgun
[[Page 29260]]
arrays currently used in the GOM, which are roughly 4,000-8,400 in\3\
(BOEM, 2017). According to BOEM's permitting records, approximately
one-third of arrays used in a recent year were 8,000 in\3\ or greater.
More importantly, the horizontal modeling of the 8,000-in\3\ array
should give sound pressure results similar to other configurations. The
output of an airgun array is directly proportional to the firing
pressure and to the number of elements. However, the sound pressure
(peak amplitude) generated by the array is not linear but instead is
proportional to the cube root of the volume of that array. For example,
doubling the size of the airgun array from 4,000 to 8,000 in\3\ would
be expected to add approximately 3 dB to the source pressure level.
Thus, an 8,000 in\3\ array produces only about twice the loudness of a
1,000 in\3\ array, assuming similar parameters such as the number of
elements and the spatial dimensions of the array. This volume to
loudness ratio holds for the sizes of single elements as well, e.g., a
240-in\3\ element only generates twice the peak pressure level of a 30-
in\3\ element (not eight times the level). It is primarily the
frequency components of the source signals that differ with size, i.e.,
larger elements produce more low-frequency sound. It should also be
noted that airgun arrays are configured geometrically so as to direct
energy downward into the seafloor (known as tuning the array); the
model fully recognizes this directionality and accounts for the lower
sound energy radiated at shallower angles and at specific bearings in
computing the exposure levels.
The exact configuration of the 4,130 in\3\ array evaluated by
Zeddies et al. (2017b) is not provided. Assuming that it is roughly
symmetrical to the 8,000 in\3\ array modeled by Zeddies et al. (2015,
2017a), and using the scaling laws where only total volume applies, the
larger array would be expected to be about 2 dB louder. Contrary to
this estimate, Zeddies et al. (2017b) report a 7.3 dB difference in
source levels, a result that cannot be completely understood given the
information provided by Zeddies et al. (2017b). One identified issue is
that the source level for the smaller array (247.9 dB) is for a
broadside prediction, while the source level for the larger array
(255.2 dB) is for the endfire prediction. The broadside source level
for the larger array is predicted to be 248.1 dB, which is reasonably
close to that of the smaller array (i.e., within 2 dB difference). The
broadside value may be a better representation of source level for the
main beams which are directed downward, while the endfire is applicable
for a smaller range of horizontal bearing from the array. Ultimately,
differences in the array geometry may be significant, and the lack of
transparency in disclosing this information for the smaller array
problematic to a meaningful comparison of results. Overall, the 8,000-
in\3\ array used by Zeddies et al. (2015, 2017a) remains a reasonable
representation of the arrays that may be used in the future, without
being overly conservative.
Sound propagation modeling. Acoustic propagation in the
GOM is complex and routinely changing due to variations in the Loop
Current (and its eddies) and weather (including hurricanes).
Additionally, propagation modeling needs to address a wide range of
water depths (i.e., shelf, slope, and deep waters) as well as strong
freshwater runoff from the Mississippi River and other rivers. In order
to capture this variability, the acoustic propagation modeling examined
the historic sound velocity profiles (SVP) for the entire U.S. GOM
throughout the entire year. As summarized earlier, these SVPs were
analyzed for similarities and ultimately grouped into seven zones or
areas with SVPs of similar structure or characteristics. These seven
zones also included consideration of bathymetric, oceanographic, and
biological factors in their definition. The SVP analysis also
identified the need to capture seasonal variations by modeling the
summer and winter seasons, which represent the bounds of reasonable
environmental variability, rather than ``extremes.'' The profiles
selected to model each of these seven zones are reasonable
representatives of the family of SVPs for that zone and reflect an
average of feasible conditions. Within each of the geographic
boundaries for each modeled zone, multiple sites were selected to serve
as the actual acoustic location for a modeled source, in order to
capture the propagation for that zone. The sites selected for these
locations included consideration of the overall characteristic of the
zone (i.e., it should be representative of the zone and not an extreme
case), the proximity of the adjacent zones, the location of important
bathymetric or oceanographic features, and, if possible, any important
information on biologically important factors (e.g., migratory routes,
animal concentrations). Finally, the 3D propagation fields for each of
the zones were examined by modeling multiple azimuthal planes radiating
out from the source location. For additional detail, see the modeling
report.
Mitigation and aversion. As discussed in further detail
above, the effects of mitigation and aversion on exposure estimates
were investigated via Test Scenarios. We acknowledge that both of these
factors would lead to a reduction in likely injurious exposure to some
degree. However, these factors were ultimately not quantified in the
modeling because, in summary, there is too much inherent uncertainty
regarding the effectiveness of detection-based mitigation to support
any reasonable quantification of its effect in reducing injurious
exposure and there is too little information regarding the likely level
of onset and degree of aversion to justify its use in the modeling.
Zeddies et al. (2017b) found that incorporation of aversion into the
modeling process appears to reduce the number of predicted injurious
exposures, though the magnitude of the effect was variable. The authors
state that this variability is likely because there are few samples of
injurious exposure exceedance, meaning that the statistical variability
of re-running simulations is evident. While aversion and mitigation
implementation would be expected to reduce somewhat the modeled levels
of injurious exposure, they would not be expected to result in any
meaningful reduction in assumed exposures resulting in behavioral
disturbance. However, we incorporated a reasonable adjustment to
modeled Level A exposure estimates to account for aversion for low- and
high-frequency species and, as described below, we do not believe that
Level A harassment is likely to occur for mid-frequency cetaceans.
In conclusion, and as stated by BOEM (2017), the results of the
modeling are expected to incorporate a reasonable margin of
conservatism, and they represent use of the most credible, science-
based methodologies and information available at this time. We believe
it appropriate to incorporate conservatism to a reasonable extent in
order to produce take estimates that would be sufficient to address the
likely impacts of the activity and to allow for issuance of
authorizations that would cover the expected requests by operators over
the course of 5 years.
Take Estimates
In order to provide an estimate of takes of marine mammals that
could occur as a result of a reasonably expected level of geophysical
survey activity in the GOM over the course of 5 years, we evaluated
BOEM's 10-year level of effort predictions and the
[[Page 29261]]
associated modeled exposures provided by Zeddies et al. (2015, 2017a).
The acoustic exposure history of many simulated animals (animats)
allows for the estimation of takes due to operations. These modeled
takes are summed and represent the aggregate takes expected to result
from future surveys given the specified levels of effort for each
survey type in each year, and may vary according to the statistical
distribution associated with these mean annual exposures. We use the
scaling factors derived from the results of Test Scenario 1 to
differentiate between the total number of predicted instances of take
and the likely number of individual marine mammals to which the takes
occur. This information--total number of takes (with Level A harassment
takes based on assumptions relating to mid-frequency cetaceans in
general as well as aversion, as described below) and individuals, on an
annual basis for five hypothetical years representing three different
potential levels of survey effort--provide a partial basis for our
negligible impact analysis, as well as the bounds within which
incidental take authorizations would be issued in association with this
proposed regulatory framework.
In summary, BOEM provided estimated levels of effort for
geophysical survey activity in the GOM for a notional ten-year period.
Exposure estimates were then computed from modeled sound levels
received by animats for several representative types of geophysical
surveying. Because animals and acoustic sources move relative to the
environment and each other, and the sound fields generated by the
sources are shaped by various physical parameters, the sound levels
received by an animal are a complex function of location and time. The
basic modeling approach was to use acoustic models to compute the 3D
sound fields and their variations in time. Animats were modeled moving
through these fields to sample the sound levels in a manner similar to
how real animals would experience these sounds. From the time histories
of the received sound levels of all animats, the numbers of animals
exposed to levels exceeding effects threshold criteria were determined
and then adjusted by the number of animals expected in the area, based
on density information, to estimate the potential number of real-world
marine mammal exposures to levels above the defined criteria.
With the overall modeling goal to estimate exposure levels from
future survey activity whose individual details such as exact location
and duration are unknown, a primary concern was how to account for
different survey types, locations and spatial extents, and durations.
In Test Scenario 1, issues arising when estimating impacts during long-
duration surveys were investigated and a method was suggested. The
defined 24-hr integration window, or reset period, creates a scaling
time-basis for impact analysis, and 24 hours is short relative to most
surveys. Test Scenario 1 demonstrated that while scaling (multiplying)
the average 24-hr exposure estimate by the number of days of a survey
is appropriate for estimating the number of instances of exposure above
threshold, this same number is likely an overestimate of the number of
individual marine mammals exposed above threshold during that time
period. The associated 30-day model runs resulted in lower numbers of
animats exposed to levels exceeding the threshold because individual
animats were only counted once in the 30-day period even when exposed
above the threshold across multiple days, which allows for a more
refined consideration of individual animal takes, i.e., comparison
between the results of these two methods (24-hr exposure estimate
scaled to 30 days versus 30-day exposure estimate) allows for a more
realistic understanding of the likely numbers of individuals exposed
within a 30-day period (as well as a better understanding of which
species are likely taken across more days). However, while this
correction helps account for the difference in estimates of individuals
taken between the primary modeling method (24-hr modeled exposures
multiplied by total number of survey days) and a 30-day modeled event,
these remain somewhat of an overestimate, as evidenced by the total
predicted takes versus the population abundance. Reasons include that
many of the surveys will likely be significantly longer than 30 days,
and that this correction does not address the fact that individuals
could be taken by multiple surveys within a given year. In conclusion,
while the exposure estimates presented in the modeling report identify
instances of anticipated take, the ``corrected'' take numbers identify
a closer approximation, and relative comparison, of the numbers of
individuals affected. However, this method of correction still
overestimates the numbers of individuals affected across the year, as
it does not consider the additional repeated takes of individuals
during surveys that are longer than 30 days or by multiple surveys.
The parameters governing animal movement were obtained from short-
duration events, such as several dives, and for this modeling effort
did not include long-duration behavior like migration or periodically
revisiting an area as part of a circulation pattern. These behaviors
could be modeled, but there are no data available currently to support
detailed modeling of this type of behavior in the GOM. Seven-day
simulations were chosen to ensure differing environments would be
sampled.
With any modeling exercise, uncertainty in the input parameters
results in uncertainty in the output. Sources of uncertainty and their
effects on exposure estimates were investigated in Test Scenario 2. The
primary source of uncertainty in this project was the location of the
animals at the times of the surveys, which drives the choice of using
an agent-based modeling approach and Monte Carlo sampling. Density
estimates assume a uniform, static distribution of animals over a
survey area, although real world animal densities can fluctuate
significantly. However, assuming many surveys will be conducted in many
locations, the variations in density are expected to average toward the
mean. Sources of uncertainty in the other modeling parameters were
found to affect the variance of the modeling results, as opposed to
their mean, and the use of mean input parameters is therefore justified
by the same argument as using mean animal densities: With many surveys
occurring over many locations, variations are expected to average
toward the mean. The effects of the variability in many of the modeling
parameters on exposure estimates were quantified using a resampling
technique. It was found that uncertainty in parameters such as animal
density and social group size had a profound effect on the distribution
of the exposure estimates, but not on the mean exposure. That is, the
distribution shape and range of the number of animals above threshold
changed, but the mean number of animals above threshold remained the
same.
We previously presented BOEM's 10-year activity projections under
``Detailed Description of Activities'' (Table 1), and identified
representative ``high,'' ``moderate,'' and ``low'' effort years. Level
of effort is currently significantly reduced in the GOM. A decrease in
permit applications was seen over the 2016 calendar year and the trend
in reduced exploration activity continued in 2017. However, BOEM states
that they assume that future levels will return to previous levels.
Therefore, the existing scenario levels, which contain projections
based on BOEM's
[[Page 29262]]
analysis by subject matter experts of past activity levels and trends
as well as industry-projected activity levels, remain valid (BOEM,
2017). BOEM's projected activity levels must be viewed as notional
years. While they are based on expert professional judgment as informed
by historical data and the best available information, it would be
inappropriate to view them as literal representations of what would
definitively happen in a given year. Therefore, in order to provide the
best reasonable basis for conducting a negligible impact analysis, and
in recognition of the current economic downturn as it relates to oil
and gas industry exploratory activity, we select one ``high-activity''
year, two separate ``moderate-activity'' years, and two separate ``low-
activity'' years as the basis for our assessment (corresponding with
the detailed per-survey type effort projections given in Table 1 for
Years 1, 4, 5, 8, and 9, respectively). Exposure estimates above Level
A and Level B harassment criteria, developed by Zeddies et al. (2015,
2017a) in association with the activity projections for these year
scenarios, are presented here (Table 8). Exposure estimates were
generated based on the specific modeling scenarios (including source
and survey geometry), i.e., 2D survey (1 x 8,000 in\3\ array), 3D NAZ
survey (2 x 8,000 in\3\ array), 3D WAZ survey (4 x 8,000 in\3\ array),
coil survey (4 x 8,000 in\3\ array), shallow penetration survey (either
single 90 in\3\ airgun or boomer), and HRG surveys (side-scan sonar,
multibeam echosounder, and subbottom profiler). Here, we present
scenario-based pooled exposure estimates by species.
Table 8--Estimated Exposures by Survey Scenario
[Zeddies et al., 2015, 2017a] 1
--------------------------------------------------------------------------------------------------------------------------------------------------------
Survey effort scenario \2\
-------------------------------------------------------------------------------------------------------------
Species High Moderate #1 Moderate #2 Low #1 Low #2
-------------------------------------------------------------------------------------------------------------
A B A B A B A B A B
--------------------------------------------------------------------------------------------------------------------------------------------------------
Bryde's whale............................. 15 560 11 413 14 498 11 386 11 402
Sperm whale............................... 45 43,504 29 27,271 38 33,340 30 26,651 32 27,657
Kogia spp................................. 3,640 16,189 2,375 11,428 3,180 13,644 2,358 10,743 2,811 11,165
Beaked whale.............................. 52 235,615 38 162,134 47 190,777 37 151,708 38 156,584
Rough-toothed dolphin..................... 150 37,666 114 30,192 128 31,103 112 28,663 105 26,315
Bottlenose dolphin........................ 1,940 653,405 2,797 977,108 1,783 596,824 2,679 938,322 1,718 579,403
Clymene dolphin........................... 469 110,742 312 72,913 380 87,615 304 69,609 310 72,741
Atlantic spotted dolphin.................. 331 133,427 423 174,705 290 116,698 397 164,824 269 109,857
Pantropical spotted dolphin............... 2,924 606,729 2,048 419,738 2,535 511,037 1,987 399,581 2,032 419,824
Spinner dolphin........................... 262 82,779 195 59,623 246 73,013 189 56,546 195 59,253
Striped dolphin........................... 194 44,038 133 29,936 164 36,267 130 28,522 133 29,890
Fraser's dolphin.......................... 52 13,858 36 9,654 44 11,394 35 9,127 35 9,391
Risso's dolphin........................... 103 27,062 73 18,124 91 21,914 71 17,309 74 18,092
Melon-headed whale........................ 252 68,900 171 47,548 213 56,791 169 44,842 170 46,631
Pygmy killer whale........................ 83 18,029 57 12,278 71 14,788 56 11,677 57 12,141
False killer whale........................ 111 25,511 77 17,631 94 20,828 75 16,774 76 17,163
Killer whale.............................. 5 1,493 3 1,031 4 1,258 3 984 3 1,036
Short-finned pilot whale.................. 68 19,258 43 12,155 51 14,163 42 11,523 42 11,900
--------------------------------------------------------------------------------------------------------------------------------------------------------
\1\ A and B refer to estimated exposures above Level A and Level B harassment criteria, respectively. For all species other than the Bryde's whale,
exposures above Level A harassment criteria were predicted by the peak SPL metric. For the Bryde's whale, exposures above Level A harassment criteria
were predicted by the cSEL metric.
\2\ High survey effort scenario corresponds with level of effort projections given previously for Year 1 (Table 1). Moderate #1 and #2 and Low #1 and #2
correspond with Years 4, 5, 8, and 9, respectively.
For all mid-frequency cetaceans, i.e., all species other than the
Bryde's whale and Kogia spp., we do not expect Level A harassment to
actually occur. For all species other than low-frequency cetaceans
(i.e., Bryde's whale), the estimates of exposure above Level A
harassment criteria are based on the peak pressure metric and, for mid-
frequency cetaceans, no exposures above Level A harassment criteria
were predicted for airgun surveys on the basis of the cSEL metric.
However, the estimated zone size for the 230 dB peak threshold for mid-
frequency cetaceans is only 18 m and, while in a theoretical modeling
scenario it is possible for animats to engage with a zone of 18 m
radius around a notional point source and, subsequently, for these
interactions to scale to predictions of real world exposures given a
sufficient number of predicted 24-hr survey days in confluence with
sufficiently high predicted real world animal densities, this is not a
realistic outcome. The source level of the array is a theoretical
definition assuming a point source and measurement in the far field of
the source. The 230 dB isopleth was within the near field of the array
where the definition of source level breaks down, so actual locations
within the 18 m of the array center where the sound level exceeds 230
dB peak SPL would not necessarily exist. Further, our proposed
mitigation (see discussion in ``Proposed Mitigation'' would require a
power-down for small dolphins within a 500-m exclusion zone (and a
shutdown for other mid-frequency cetaceans). During the power-down
procedure, a single airgun would remain firing. The output of a single
airgun would not be expected to exceed the peak pressure injury
threshold for mid-frequency cetaceans. Therefore, we expect the
potential for Level A harassment of mid-frequency cetaceans to be de
minimis, even before the likely moderating effects of aversion are
considered. When considering potential for aversion, we do not believe
that Level A harassment is a likely outcome for any mid-frequency
cetacean.
For other species (i.e., Bryde's whales and Kogia spp.), we believe
that while some amount of Level A harassment is likely, the lack of
aversion within the animal movement modeling process results in
overestimates of potential injurious exposure. Although there was not
sufficient information to inform a precise quantification of aversion
within the modeling (Test Scenario 4), we believe that sufficient
information exists to inform a reasonable, conservative approximation
of aversion and apply an offset method accordingly (Southall et al.,
2017). Ellison et al. (2016) demonstrated that animal movement models
where no aversion probability was used overestimated the potential for
high levels of exposure required for PTS by about five times.
Accordingly, total
[[Page 29263]]
estimated exposures above Level A harassment criteria (without
accounting for behavioral aversion) were multiplied by 0.2 to
reasonably obtain a more realistic estimate of potential injurious
exposure. Adjusted total scenario-specific and mean annual take
estimates are given in Table 9.
Table 9--Scenario-Specific Expected Take Numbers and Mean Annual Take Level 1
--------------------------------------------------------------------------------------------------------------------------------------------------------
Survey effort scenario \2\
-----------------------------------------------------------------------------------------------------------------------
Species High Moderate #1 Moderate #2 Low #1 Low #2 Mean annual take
-----------------------------------------------------------------------------------------------------------------------
A B A B A B A B A B A B
--------------------------------------------------------------------------------------------------------------------------------------------------------
Bryde's whale................... 3 560 2 413 2 498 2 386 2 402 2 452
Sperm whale..................... 0 43,504 0 27,271 0 33,340 0 26,651 0 27,657 0 31,685
Kogia spp....................... 728 16,189 475 11,428 636 13,644 472 10,743 562 11,165 575 12,634
Beaked whale.................... 0 235,615 0 162,134 0 190,777 0 151,708 0 156,584 0 179,364
Rough-toothed dolphin........... 0 37,666 0 30,192 0 31,103 0 28,663 0 26,315 0 30,788
Bottlenose dolphin.............. 0 653,405 0 977,108 0 596,824 0 938,322 0 579,403 0 749,012
Clymene dolphin................. 0 110,742 0 72,913 0 87,615 0 69,609 0 72,741 0 82,724
Atlantic spotted dolphin........ 0 133,427 0 174,705 0 116,698 0 164,824 0 109,857 0 139,902
Pantropical spotted dolphin..... 0 606,729 0 419,738 0 511,037 0 399,581 0 419,824 0 471,382
Spinner dolphin................. 0 82,779 0 59,623 0 73,013 0 56,546 0 59,253 0 66,243
Striped dolphin................. 0 44,038 0 29,936 0 36,267 0 28,522 0 29,890 0 33,731
Fraser's dolphin................ 0 13,858 0 9,654 0 11,394 0 9,127 0 9,391 0 10,685
Risso's dolphin................. 0 27,062 0 18,124 0 21,914 0 17,309 0 18,092 0 20,500
Melon-headed whale.............. 0 68,900 0 47,548 0 56,791 0 44,842 0 46,631 0 52,942
Pygmy killer whale.............. 0 18,029 0 12,278 0 14,788 0 11,677 0 12,141 0 13,783
False killer whale.............. 0 25,511 0 17,631 0 20,828 0 16,774 0 17,163 0 19,581
Killer whale.................... 0 1,493 0 1,031 0 1,258 0 984 0 1,036 0 1,160
Short-finned pilot whale........ 0 19,258 0 12,155 0 14,163 0 11,523 0 11,900 0 13,800
--------------------------------------------------------------------------------------------------------------------------------------------------------
\1\ A and B refer to expected scenario-based instances of take by Level A and Level B harassment, respectively. For the Bryde's whale and Kogia spp.,
expected Level A takes represent modeled exposures adjusted to account for aversion.
\2\ High survey effort scenario correspond level of effort projections given previously for Year 1 (Table 1). Moderate #1 and #2 and Low #1 and #2
correspond with Years 4, 5, 8, and 9, respectively.
Economic Baseline
This proposed rule has been designated as significant under
Executive Order 12866. Accordingly, a draft regulatory impact analysis
(RIA) has been prepared and is available for review online at:
www.fisheries.noaa.gov/national/marine-mammal-protection/incidental-take-authorizations-oil-and-gas. The RIA evaluates the potential costs
and benefits of these proposed incidental take regulations, as well as
a more stringent alternative, against two baselines. The two baselines
correspond with: (1) Regulatory requirements associated with management
of geophysical survey activity in the GOM prior to 2013 pursuant to
permits that were issued by BOEM under its authorities in the Outer
Continental Shelf Lands Act but that did not address statutory
requirements of the MMPA administered by NOAA; and (2) conditions in
place since 2013 pursuant to a settlement agreement, as amended through
stipulated agreement, involving a stay of litigation (NRDC et al. v.
Zinke et al., Civil Action No. 2:10 cv-01882 (E.D. La.)). Under the
settlement agreement (which expires in November 2018), industry trade
groups representing operators agreed to include certain mitigation
requirements for geophysical surveys in the GOM. Appendix B of the RIA
provides an initial regulatory flexibility analysis (IRFA), while
Appendix C addresses other compliance requirements.
Office of Management and Budget (OMB) Circular A-4 directs that the
baseline for regulatory analysis should be the agency's best assessment
of the state of the world in the absence of the proposed action. A-4
also provides that agencies may present multiple baselines where this
would provide additional useful information to the public on the
projected effects of the regulation. We are presenting two baselines
for public information and comment, consistent with the A-4 provision
allowing agencies to present multiple baselines. Thus, in addition to a
baseline that reflects current assumed industry practices as agreed
upon in the 2013 settlement agreement, NMFS is also presenting a
baseline corresponding with geophysical activities in the GOM as
carried out prior to the 2013 settlement agreement but without
authorization from NMFS under the MMPA.
Estimated direct costs of the measures in the proposed regulations,
relative to both baselines, are presented in Table 10. Details
regarding cost estimation are available in the RIA. A qualitative
evaluation of indirect costs related to the proposed regulations is
also provided in the RIA. Note that these costs would be diffused
across all operators receiving LOAs.
Table 10--Quantified Direct Compliance Costs by Baseline
------------------------------------------------------------------------
Annualized costs, millions \1\
---------------------------------------
Mitigation measure Pre-stay agreement Stay agreement
baseline (prior to baseline (2013-
2013) present)
------------------------------------------------------------------------
Mitigation requirements for $3.9-$49.7 $3.9-$49.7
dolphins: Shutdowns for large
dolphins in the exclusion zone
and power downs for small
dolphins in the exclusion zone.
Expanded observer requirements $0.02-$2.1 $0
and mitigation in shallow
waters: Shutdowns for all
``whale'' species in the
exclusion zone for airgun
surveys in water depths less
than 200 m in the Central and
Western Planning Areas.........
Additional mitigation $1.1-$3.0 $1.1-$3.0
requirements: Shutdowns for
Bryde's/beaked/Kogia whales
outside of exclusion zone for
deep penetration airgun surveys
[[Page 29264]]
Acoustic monitoring and $43.9-$127 $21.9-$65.8
associated mitigation:
Shutdowns for all non-delphinid
detections for deep penetration
airgun surveys.................
Observer requirements for non- $0.12-$0.39 $0.12-$0.39
airgun HRG surveys and
associated mitigation:
Shutdowns for whale and large
dolphin observations in the
exclusion zone.................
Remove minimum separation n/a ($37.9)-($266)
distance requirements for deep
penetration airgun surveys: The
stay agreement baseline
includes minimum separation
distances. Costs reflect the
downtime associated with
maintaining the minimum
separation distance from other
surveys. This mitigation
measure is not included in the
proposed rule, thus creating a
benefit (negative cost) of the
proposed rule relative to the
stay agreement baseline........
---------------------------------------
Proposed Rule Total Direct $49-$182 \2\ ($10.8)-($147)
Compliance Costs...........
------------------------------------------------------------------------
\1\ Costs are presented in terms of 2016 U.S. dollars and are annualized
over the five-year timeframe applying a 7% discount rate. Annualized
costs applying a 3% discount rate are provided in Appendix D of the
RIA.
\2\ Estimates within parentheses indicate negative costs, or cost
savings. The proposed rule total direct compliance costs relative to
the stay agreement baseline reflect new costs of $27-$119 less cost
savings of $38-$266.
Proposed Mitigation
Under Section 101(a)(5)(A) of the MMPA, NMFS must set forth the
permissible methods of taking pursuant to such activity, and other
means of effecting the least practicable adverse impact on such species
or stock and its habitat, paying particular attention to rookeries,
mating grounds, and areas of similar significance, and on the
availability of such species or stock for taking for certain
subsistence uses (``least practicable adverse impact''). Consideration
of the availability of marine mammal species or stocks for taking for
subsistence uses pertains only to Alaska, and is therefore not relevant
here. NMFS does not have a regulatory definition for ``least
practicable adverse impact.'' However, NMFS's implementing regulations
require applicants for incidental take authorizations to include
information about the availability and feasibility (economic and
technological) of equipment, methods, and manner of conducting such
activity or other means of effecting the least practicable adverse
impact upon the affected species or stocks and their habitat (50 CFR
216.104(a)(11)). It is important to note that in some cases, certain
mitigation may be necessary in order to ensure a ``negligible impact''
on an affected species or stock, which is a fundamental requirement of
issuing an authorization--in these cases, consideration of
practicability may be a lower priority for decision-making if impacts
to marine mammal species or stocks would be greater than negligible in
the measure's absence.
In evaluating how mitigation may or may not be appropriate to
ensure the least practicable adverse impact on species or stocks and
their habitat, we carefully consider two primary factors:
(1) The manner in which, and the degree to which, implementation of
the measure(s) is expected to reduce impacts to marine mammal species
or stocks, their habitat, and their availability for subsistence uses
(when relevant). This analysis will consider such things as the nature
of the potential adverse impact (such as likelihood, scope, and range),
the likelihood that the measure will be effective if implemented, and
the likelihood of successful implementation.
(2) The practicability of the measure for applicant implementation.
Practicability of implementation may consider such things as cost,
impact on operations, personnel safety, and practicality of
implementation.
While the language of the least practicable adverse impact standard
calls for minimizing impacts to affected species or stocks, we
recognize that the reduction of impacts to those species or stocks
accrues through the application of mitigation measures that limit
impacts to individual animals. Accordingly, our analysis focuses on
measures designed to avoid or minimize impacts on marine mammals from
activities that are likely to increase the probability or severity of
population-level effects, including auditory injury or disruption of
important behaviors, such as foraging, breeding, or mother/calf
interactions. See also 82 FR 19460 (April 27, 2017) and 83 FR 10954
(March 13, 2018) (discussion of least practicable adverse impact
standard in proposed incidental take rule for Navy's Surveillance Towed
Array Sensor System Low Frequency Sonar activities and Atlantic Fleet
Testing and Training activities, respectively).
NMFS is aware of public statements that there is no scientific
evidence that geophysical survey activities have caused adverse
consequences to marine mammal stocks or populations, and that there are
no known instances of injury to individual marine mammals as a result
of such surveys. For example, BOEM stated publicly that ``there has
been no documented scientific evidence of noise from airguns . . .
adversely affecting marine animal populations'' (BOEM, 2014;
www.boem.gov/BOEM-Science-Note-August-2014/). On their face, these
carefully worded statements are not incorrect; however, they are easily
misconstrued and, as used in arguments against certain proposed
mitigation measures, represent a common logical fallacy (i.e., that a
proposition is false because it has not yet been proven true). In
reality, conclusive statements regarding population-level consequences
of acoustic stressors cannot be made due to insufficient investigation,
as such studies are exceedingly difficult to carry out and no
appropriate study and reference populations have yet been established.
For example, a recent report from the National Academy of Sciences
noted that, while a commonly-cited statement from the National Research
Council (``[n]o scientific studies have conclusively demonstrated a
link between exposure to sound and adverse effects on a marine mammal
population'') remains true, it is largely because such impacts are very
difficult to demonstrate (NRC, 2005; NAS, 2017).
Population[hyphen]level effects are inherently difficult to assess
because of high variability, migrations, and multiple factors affecting
the populations.
[[Page 29265]]
The MMPA defines ``take'' to include Level B (behavioral)
harassment, which has been documented numerous times for marine mammals
in the presence of airguns (in the form of avoidance of areas, notable
changes in vocalization or movement patterns, or other shifts in
important behaviors), as well as auditory injury (Level A harassment),
for which there is also evidence from loud sound sources (e.g.,
Southall et al., 2007). Further, there is growing scientific evidence
demonstrating the connections between sub-lethal effects, such as
behavioral disturbance, and population-level effects on marine mammals
(e.g., Lusseau and Bedjer, 2007; New et al., 2014). Disruptions of
important behaviors, in certain contexts and scales, have been shown to
have energetic effects that can translate to reduced survivorship or
reproductive rates of individuals (e.g., feeding is interrupted, so
growth, survivorship, or ability to bring young to term is
compromised), which in turn can adversely affect populations depending
on their health, abundance, and growth trends. As BOEM stated in a
follow-up to the above-referenced Science Note, ``[we] should not
assume that lack of evidence for adverse population-level effects of
airgun surveys means that those effects may not occur.'' (BOEM, 2015;
www.boem.gov/BOEM-Science-Note-March-2015/).
While direct evidence of impacts to species or stocks from a
specified activity is rarely available, and additional study is still
needed to describe how specific disturbance events affect the fitness
of individuals of certain species, there have been improvements in
understanding the process by which disturbance effects are translated
to the population. With recent scientific advancements (both marine
mammal energetic research and the development of energetic frameworks),
the relative likelihood or degree of impacts on species or stocks may
often be inferred given a detailed understanding of the activity, the
environment, and the affected species or stocks. This same information
is used in the development of mitigation measures and helps us
understand how mitigation measures contribute to lessening effects (or
the risk thereof) to species or stocks. We also acknowledge that there
is always the potential that new information, or a new recommendation
that we had not previously considered, becomes available and
necessitates reevaluation of mitigation measures (which may be
addressed through adaptive management) to see if further reduction of
population impacts are possible and practicable.
In the evaluation of specific measures, the details of the
specified activity will necessarily inform each of the two primary
factors discussed above (expected reduction of impacts and
practicability), and will be carefully considered to determine the
types of mitigation that are appropriate under the least practicable
adverse impact standard. Analysis of how a potential mitigation measure
may reduce adverse impacts on a marine mammal stock or species and
practicability of implementation are not issues that can be
meaningfully evaluated through a yes/no lens. The manner in which, and
the degree to which, implementation of a measure is expected to reduce
impacts, as well as its practicability in terms of these
considerations, can vary widely. For example, a time/area restriction
could be of very high value for decreasing population-level impacts
(e.g., avoiding disturbance of feeding females in an area of
established biological importance) or it could be of lower value (e.g.,
decreased disturbance in an area of high productivity but of less
firmly established biological importance). Regarding practicability, a
measure might involve operational restrictions that completely impede
the operator's ability to acquire necessary data (higher impact), or it
could mean additional incremental delays that increase operational
costs but still allow the activity to be conducted (lower impact). A
responsible evaluation of ``least practicable adverse impact'' will
consider the factors along these realistic scales. Expected effects of
the activity and of the mitigation as well as status of the stock all
weigh into these considerations. Accordingly, the greater the
likelihood that a measure will contribute to reducing the probability
or severity of adverse impacts to the species or stock, the greater the
weight that measure is given when considered in combination with
practicability to determine the appropriateness of the mitigation
measure, and vice versa. We discuss consideration of these factors in
greater detail below.
1. Reduction of Adverse Impacts to Marine Mammal Species and Stocks and
Their Habitat
The emphasis given to a measure's ability to reduce the impacts on
a species or stock considers the degree, likelihood, and context of the
anticipated reduction of impacts to individuals as well as the status
of the species or stock. The ultimate impact on any individual from a
disturbance event (which informs the likelihood of adverse species- or
stock-level effects) is dependent on the circumstances and associated
contextual factors, such as duration of exposure to stressors. Though
any proposed mitigation needs to be evaluated in the context of the
specific activity and the species or stocks affected, measures with the
following types of goals are often applied to reduce the likelihood or
severity of adverse species- or stock-level impacts: Avoiding or
minimizing injury or mortality; limiting interruption of known feeding,
breeding, mother/calf, or resting behaviors; minimizing the abandonment
of important habitat (temporally and spatially); minimizing the number
of individuals subjected to these types of disruptions; and limiting
degradation of habitat. Mitigating these types of effects is intended
to reduce the likelihood that the activity will result in energetic or
other types of impacts that are more likely to result in reduced
reproductive success or survivorship. It is also important to consider
the degree of impacts that were expected in the absence of mitigation
in order to assess the added value of any potential measures. Finally,
because the least practicable adverse impact standard authorizes NMFS
to weigh a variety of factors when evaluating appropriate mitigation
measures, it does not compel mitigation for every kind of individual
take, even when practicable for implementation by the applicant.
The status of the species or stock is also relevant in evaluating
the appropriateness of certain mitigation measures in the context of
least practicable adverse impact. The following are examples of factors
that may (either alone, or in combination) result in greater emphasis
on the importance of a mitigation measure in reducing impacts on a
species or stock: The stock is known to be decreasing or status is
unknown, but believed to be declining; the known annual mortality (from
any source) is approaching or exceeding the PBR level; the affected
species or stock is a small, resident population; or the stock is
involved in a UME or has other known vulnerabilities, such as
recovering from an oil spill.
Habitat mitigation, particularly as it relates to rookeries, mating
grounds, and areas of similar significance, is also relevant to
achieving the standard and can include measures such as reducing
impacts of the activity on known prey utilized in the activity area or
reducing impacts on physical habitat. As with species- or stock-related
mitigation, the emphasis given to a measure's ability to reduce impacts
on a species or stock's habitat considers the degree, likelihood,
[[Page 29266]]
and context of the anticipated reduction of impacts to habitat. Because
habitat value is informed by marine mammal presence and use, in some
cases there may be overlap in measures for the species or stock and for
use of habitat.
We consider available information indicating the likelihood of any
measure to accomplish its objective. If evidence shows that a measure
has not typically been effective or successful, then either that
measure should be modified or the potential value of the measure to
reduce effects is lowered.
2. Practicability
Factors considered may include those such as cost, impact on
operations, personnel safety, and practicality of implementation. In
carrying out the MMPA's mandate, we apply the previously described
context-specific balance between the manner in which and the degree to
which measures are expected to reduce impacts to the affected species
or stocks and their habitat and practicability for the applicant. The
effects of concern, addressed previously in the ``Potential Effects of
the Specified Activity on Marine Mammals and Their Habitat'' section,
include auditory injury, severe behavioral reactions, disruptions of
critical behaviors, and potentially detrimental chronic and/or
cumulative effects to acoustic habitat (see discussion of this concept
in the ``Anticipated Effects on Marine Mammal Habitat'' section). Here,
we focus on measures with proven or reasonably presumed ability to
avoid or reduce the intensity of acute exposures that may potentially
result in these effects with an understanding of the drawbacks of these
requirements, while also evaluating time-area restrictions that would
avoid or reduce both acute and chronic impacts. To the extent of the
information available to us, we consider practicability concerns, as
well as potential undesired consequences of the measures, e.g.,
extended periods using the acoustic source due to the need to reshoot
lines. We also recognize that instantaneous protocols, such as shutdown
requirements, are not capable of avoiding all acute effects, and are
not suitable for avoiding many cumulative or chronic effects and do not
provide targeted protection in areas of greatest importance for marine
mammals. Therefore, in addition to a basic suite of seismic mitigation
protocols, we also consider measures that may not be appropriate for
other activities (e.g., time-area restrictions specific to the proposed
surveys discussed here) but that are warranted here given the scope of
these specified activities and associated higher potential for
population-level effects and/or a large magnitude of take of
individuals of certain species, in the absence of such mitigation.
In order to satisfy the MMPA's least practicable adverse impact
standard, we propose a suite of basic mitigation protocols that are
required regardless of the status of a stock. Additional or enhanced
protections are proposed for species whose stocks are in poor health
and/or are subject to some significant additional stressor that lessens
that stock's ability to weather the effects of the specified activity
without worsening its status. We reviewed the mitigation measures
proposed in the petition, the requirements specified in BOEM's PEIS,
seismic mitigation protocols required or recommended elsewhere (e.g.,
HESS, 1999; DOC, 2013; IBAMA, 2005; Kyhn et al., 2011; JNCC, 2017;
DEWHA, 2008; BOEM, 2016; DFO, 2008; GHFS, 2015; MMOA, 2015; Nowacek et
al., 2013; Nowacek and Southall, 2016), and the available scientific
literature. We also considered recommendations given in a number of
review articles (e.g., Weir and Dolman, 2007; Compton et al., 2008;
Parsons et al., 2009; Wright and Cosentino, 2015; Stone, 2015b). The
suite of mitigation measures proposed here differs in some cases from
the measures proposed in the petition and/or those specified by BOEM in
the preferred alternative identified in their PEIS in order to reflect
what we believe to be the most appropriate suite of measures to satisfy
the requirements of the MMPA.
For purposes of defining mitigation requirements, we differentiate
here between requirements for two classes of airgun survey activity:
Deep penetration and shallow penetration, with surveys using arrays
greater than 400 in\3\ total airgun volume considered deep penetration.
We consider this a reasonable cutoff as most arrays or single airguns
of this size or smaller will typically be purposed for shallow
penetration surveys--BOEM states in the petition that airgun sources
used for shallow penetration surveys typically range from 40-400 in\3\,
while the Associations state in their comments on the petition that
deep penetration array volumes used in the GOM range from approximately
2,000 to 8,400 in\3\. We also consider a third general class of
surveys, referred to here as HRG surveys and including those surveys
using the non-airgun sources described previously. HRG surveys are
treated differentially on the basis of water depth, with 200 m as the
divider between shallow and deep HRG. We use this as an indicator for
surveys (shallow) that should be expected to have less potential for
impacts to marine mammals, because HRG sources used in shallow waters
are typically higher-frequency, lower power, and/or having some
significant directionality to the beam pattern. Finally, HRG surveys
using only sources operating at frequencies greater than or equal to
200 kHz would be exempt from the mitigation requirements described
herein, with the exception of adherence to vessel strike avoidance
protocols. We do not make any distinction in standard required
mitigations on the basis of BOEM's planning areas (i.e., Western
Planning Area (WPA), CPA, EPA).
As described previously in the ``Marine Mammal Hearing'' section,
the upper limit of hearing for marine mammals is approximately 160 kHz;
therefore, they would not be expected to detect signals from systems
operating at frequencies of 200 kHz and greater. Sounds that are above
the functional hearing range of marine animals may be audible if
sufficiently loud (e.g., M[oslash]hl, 1968). However, the typical
relative output levels of these sources mean that they would
potentially be detectable to marine mammals at maximum distances of
only a few meters, and are highly unlikely to be of sufficient
intensity to result in Level B harassment. Sources operating at high
frequencies also generally have short duration signals and highly
directional beam patterns, meaning that any individual marine mammal
would be unlikely to even receive a signal that would almost certainly
be inaudible.
We are aware of two studies (Deng et al., 2014; Hastie et al.,
2014) demonstrating some behavioral reaction by marine mammals to
acoustic systems operating at user-selected frequencies above 200 kHz.
These studies generally indicate only that sub-harmonics could be
detectable by certain species at distances up to several hundred
meters. However, this detectability is in reference to ambient noise,
not to thresholds for assessing the potential for incidental take for
these sources. Source levels of the secondary peaks considered in these
studies--those within the hearing range of some marine mammals--range
from 135-166 dB, meaning that these sub-harmonics would either be below
levels likely to result in Level B harassment or would attenuate to
such a level within a few meters. Therefore, acoustic sources operating
at frequencies greater than or equal to 200 kHz are not expected to
have any effect on marine mammals. Further, recent sound source
verification testing of these and other similar systems did not observe
any sub-
[[Page 29267]]
harmonics in any of the systems tested under controlled conditions
(Crocker and Fratantonio, 2016). While this can occur during actual
operations, the phenomenon may be the result of issues with the system
or its installation on a vessel rather than an issue that is inherent
to the output of the system. We do not discuss these surveys further
and none of the requirements described below (other than vessel strike
avoidance procedures) would apply to these surveys.
Our consideration of the two major points described above (i.e.,
ability of the measure to reduce the probability or severity of adverse
impacts on marine mammal species or stocks and their habitat and
practicability for the applicant) points to the need for a basic system
of mitigation protocols that reasonably may be expected to achieve the
following outcomes: (1) Avoid or minimize effects of concern that
otherwise could accrue in a way that could cause or appreciably
increase the risk of population-level impacts; (2) be easily
implemented in the field; (3) reduce subjective decision-making for
observers to the extent possible; and, (4) appropriately weigh a range
of potential outcomes from sound exposure in determining what should be
avoided or minimized where possible. Subsequently, we describe measures
specific to the GOM in relation to specific contextual concerns.
Mitigation-Related Monitoring
Avoidance or minimization of acute exposure is first and foremost
dependent upon detection of animals present in the vicinity of the
survey activity. Requirements necessary to adequately detect marine
mammals incur costs, which we consider in scaling mitigation-related
monitoring requirements relative to the expected effects of the
specific activity (as described above, we bin activity types and detail
below the proposed monitoring requirements associated with each).
Visual monitoring is a critical component of any detection system, as
evidenced by the inclusion of visual monitoring requirements in every
set of protocols and recommendations we reviewed, and has long been
accepted as such. However, visual monitoring is only effective during
periods of good visibility and when animals are available for detection
(i.e., at the surface).
Acoustic monitoring is an equally critical component of an
effective detection system, supplanting visual monitoring during
periods of poor visibility and supplementing during periods of good
visibility. There are multiple explanations of how marine mammals could
be in a shutdown zone and yet go undetected by observers. Animals are
missed because they are underwater (availability bias) or because they
are available to be seen, but are missed by observers (perception and
detection biases) (e.g., Marsh and Sinclair, 1989). Negative bias on
perception or detection of an available animal may result from
environmental conditions, limitations inherent to the observation
platform, or observer ability. Species vary widely in the inherent
characteristics that inform expected bias on their availability for
detection or the extent to which availability bias is convolved with
detection bias (e.g., Barlow and Forney (2007) estimate probabilities
of detecting an animal directly on a transect line (g(0)), ranging from
0.23 for small groups of Cuvier's beaked whales to 0.97 for large
groups of dolphins). Typical dive times range widely, from just a few
minutes for Bryde's whales (Alves et al., 2010) to more than 45 minutes
for sperm whales (Jochens et al., 2008; Watwood et al., 2006), while
g(0) for cryptic species such as Kogia spp. declines more rapidly with
increasing Beaufort sea state than it does for other species (Barlow,
2015). Barlow and Gisiner (2006) estimated that when weather and
daylight considerations were taken into account, visual monitoring
would detect fewer than two percent of beaked whales that were directly
in the path of the ship. PAM can be expected to improve on that
performance, and has been used effectively as a mitigation tool by
operators in the GOM since at least 2012. BOEM highlighted the
importance of PAM to detection-based mitigation protocols in the
petition for rulemaking, submitted to NMFS in support of industry, and
we agree. However, we do not agree that use of 24-hr PAM should be
limited to the Mississippi Canyon and De Soto Canyon lease blocks (as
proposed by BOEM). Species that are difficult to detect but vocally
active are present in significant numbers outside those areas, and PAM
should be a standard component of detection-based mitigation anywhere
such species are expected to be present.
PAM does have limitations, e.g., animals may only be detected when
vocalizing, species making directional vocalizations must vocalize
towards the array to be detected, and species identification and
localization may be difficult. However, for certain species and in
appropriate environmental conditions it is an indispensable complement
to visual monitoring during good sighting conditions and it is the only
meaningful monitoring technique during periods of poor visibility;
without PAM, there can be no expectation that any animal would be
detected at night, and even during good conditions many deep-diving
and/or cryptic species would go undetected much of the time. In the
GOM, beaked whales and sperm whales (both vocally active) are two taxa
of greatest concern; beaked whales would rarely be detected by visual
means alone (an analysis of six years of GOM survey data found only 11
records for beaked whales; Barkaszi et al., 2012), and, while commonly
observed when they are at the surface, sperm whales spend significant
amounts of time in locations where they are unavailable for visual
detection. However, acoustic monitoring imposes additional costs on
operators and, as discussed by Nowacek et al. (2013), we consider this
in relation to the anticipated effects of the survey type. Thus, while
PAM should be required during the deep penetration airgun surveys of
greatest concern, we do not propose to require it for other survey
types.
Note that, although we propose requirements related only to
observation of marine mammals, we hereafter use the generic term
``protected species observer'' (PSO). Monitoring by dedicated, trained
marine mammal observers is required in all water depths and, for
certain surveys, observers must be independent. Additionally, for some
surveys, we propose to require that some PSOs have prior experience in
the role. Independent observers are employed by a third-party observer
provider; vessel crew may not serve as PSOs when independent observers
are required. Dedicated observers are those who have no tasks other
than to conduct observational effort, record observational data, and
communicate with and instruct the geophysical survey operator (i.e.,
vessel captain and crew) with regard to the presence of marine mammals
and mitigation requirements. Communication with the operator may
include brief alerts regarding maritime hazards. We are proposing to
define trained PSOs as having successfully completed an approved PSO
training course (see the ``Proposed Monitoring and Reporting''
section), and experienced PSOs as having additionally gained a minimum
of 90 days at-sea experience working as a PSO, with no more than 18
months having elapsed since the conclusion of the relevant at-sea
experience. Training and experience is specific to either visual or
acoustic PSO duties (where
[[Page 29268]]
required). Furthermore, we propose that an experienced visual PSO must
have completed approved, relevant training and must have gained the
requisite experience working as a visual PSO. An experienced acoustic
PSO must have completed a passive acoustic monitoring (PAM) operator
training course and must have gained the requisite experience working
as an acoustic PSO. Hereafter, we also refer to acoustic PSOs as PAM
operators, whereas when we use ``PSO'' without a qualifier, the term
refers to either visual PSOs or PAM operators (acoustic PSOs).
NMFS expects to provide informal approval for specific training
courses in consultation with BOEM and the Bureau of Safety and
Environmental Enforcement (BSEE) as needed to approve PSO staffing
plans. NMFS does not propose to formally administer any training
program or to sanction any specific provider, but will approve courses
that meet the curriculum and trainer requirements specified herein (see
the ``Proposed Monitoring and Reporting'' section). We propose this in
context of the need to ensure that PSOs have the necessary training to
carry out their duties competently while also approving applicant
staffing plans quickly. In order for PSOs to be approved, we propose
that NMFS must review and approve PSO resumes accompanied by a relevant
training course information packet that includes the name and
qualifications (i.e., experience, training completed, or educational
background) of the instructor(s), the course outline or syllabus, and
course reference material as well as a document stating the PSO's
successful completion of the course. Although we are proposing that
NMFS must affirm PSO approvals, third-party observer providers and/or
companies seeking PSO staffing should expect that observers having
satisfactorily completed approved training and with the requisite
experience (if required) will be quickly approved and, if NMFS does not
respond within one week of having received the required information, we
propose that such PSOs shall be considered to be approved. A PSO may be
trained and/or experienced as both a visual PSO and PAM operator and
may perform either duty, pursuant to scheduling requirements. Where
multiple PSOs are required and/or PAM operators are required, we
propose that PSO watch schedules shall be devised in consideration of
the following restrictions: (1) A maximum of two consecutive hours on
watch followed by a break of at least one hour between watches for
visual PSOs (periods typical of observation for research purposes and
as used for airgun surveys in certain circumstances (Broker et al.,
2015)); (2) a maximum of four consecutive hours on watch followed by a
break of at least two consecutive hours between watches for PAM
operators; and (3) a maximum of 12 hours observation per 24-hour
period. Further information regarding PSO requirements may be found in
the ``Proposed Monitoring and Reporting'' section, later in this
document. NMFS has discussed the PSO requirements specified herein with
BSEE and with third-party observer providers; these parties have
indicated that the requirements should not be expected to result in any
labor shortage. For example, a significantly greater amount of survey
activity was occurring in the GOM during 2013-2015 than at present
(i.e., as many as 30 source vessels) with requirements similar to those
described here. No labor shortage was experienced. We request comment
on this assumption. We also invite comment on the proposed definitions
of trained and experienced PSOs, requirements for PSO approval by NMFS,
and watch schedule for visual PSO and PAM operators.
Deep Penetration Airgun--During deep penetration airgun survey
operations (e.g., any day on which use of the acoustic source is
planned to occur; whenever the acoustic source is in the water, whether
activated or not), we propose the additional requirement that a minimum
of two independent PSOs must be on duty and conducting visual
observations at all times during daylight hours (i.e., from 30 minutes
prior to sunrise through 30 minutes following sunset) and 30 minutes
prior to and during nighttime ramp-ups of the airgun array (see ``Ramp-
ups'' below). PSOs should use NOAA's solar calculator
(www.esrl.noaa.gov/gmd/grad/solcalc/) to determine sunrise and sunset
times at their specific location. We recognize that certain daytime
conditions (e.g., fog, heavy rain) may reduce or eliminate
effectiveness of visual observations; however, on-duty PSOs shall
remain alert for marine mammal observational cues and/or a change in
conditions.
We propose that all source vessels must carry a minimum of one
experienced visual PSO, who shall be designated as the lead PSO,
coordinate duty schedules and roles, and serve as primary point of
contact for the operator. Experience is critical to best performance of
the PSO team (e.g., Stone, 2015b), e.g., Mori et al. (2003) found that
observers classed as having limited experience were significantly less
successful in detecting animals than were experienced observers. A
survey of professional PSOs and other experts (GHFS, 2015) highlighted
the importance of experience as a best practice in selecting PSOs, both
for improved performance in detecting animals but also due to the
unique challenges a PSO faces while charged with implementing required
mitigations onboard a working survey vessel. Experience breeds the
confidence and professionalism necessary to maintain positive relations
with the vessel operator while making sometimes difficult decisions
regarding implementation of mitigation. However, while it is desirable
for all PSOs to be qualified through experience, we are also mindful of
the need to expand the workforce by allowing opportunity for newly
trained PSOs to gain experience. Therefore, the lead PSO shall devise
the duty schedule such that experienced PSOs are on duty with trained
PSOs (i.e., those PSOs with appropriate training but who have not yet
gained relevant experience) to the maximum extent practicable in order
to provide necessary mentorship.
With regard to specific observational protocols, we are proposing
to largely follow those described in Appendix B of BOEM's PEIS (BOEM,
2017). The lead PSO shall determine the most appropriate observation
posts that will not interfere with navigation or operation of the
vessel while affording an optimal, elevated view of the sea surface;
these should be the highest elevation available on each vessel, with
the maximum viewable range from the bow to 90 degrees to port or
starboard of the vessel. PSOs shall coordinate to ensure 360[deg]
visual coverage around the vessel, and shall conduct visual
observations using binoculars and the naked eye while free from
distractions and in a consistent, systematic, and diligent manner. All
source vessels must be equipped with pedestal-mounted ``bigeye''
binoculars that will be available for PSO use. Within these broad
outlines, the lead PSO and PSO team will have discretion to determine
the most appropriate vessel- and survey-specific system for
implementing effective marine mammal observational effort. Any
observations of marine mammals by crew members aboard any vessel
associated with the survey, including receiver or chase vessels, should
be relayed to the source vessel and to the PSO team.
We are proposing that all source vessels must use a towed PAM
system for potential detection of marine mammals at all times when
operating the sound source in waters deeper than 100 m. In shallower
waters, only two
[[Page 29269]]
species are typically present (bottlenose and Atlantic spotted dolphin;
rough-toothed dolphins are the only other species potentially
encountered in shelf waters but are typically found in deep water
(Davis et al., 1998; Fulling et al., 2003; Maze-Foley and Mullin,
2006)). While dolphins may be detected using PAM, we are not proposing
to require shutdowns of the source for dolphin presence (described
below); therefore, the mitigation would be of low value relative to the
estimated cost of equipment and additional personnel.
We are proposing that the system must be monitored at all times
during use of the acoustic source, and acoustic monitoring must begin
at least 30 minutes prior to ramp-up. PAM operators must be
independent. Because the role of PAM operator is more technically
complex than is the role of visual PSO, experience is more important
(D. Epperson, BSEE, pers. comm.) and we are proposing that all source
vessels shall carry a minimum of two experienced PAM operators, which
is a stricter requirement than for visual PSOs. PAM operators shall
communicate all detections to visual PSOs, when visual PSOs are on
duty, including any determination by the PSO regarding species
identification, distance, and bearing and the degree of confidence in
the determination. Further detail regarding PAM system requirements may
be found in the ``Proposed Monitoring and Reporting'' section, later in
this document. The effectiveness of PAM depends to a certain extent on
the equipment and methods used and competency of the PAM operator, but
no established standards are currently in place. We do offer some
specifications later in this document and would require that applicants
follow any standards that are established in the future.
Visual monitoring must begin at least 30 minutes prior to ramp-up
(described below) and must continue until one hour after use of the
acoustic source ceases or until 30 minutes past sunset. If any marine
mammal is observed at any distance from the vessel, a PSO would record
the observation and monitor the animal's position (including latitude/
longitude of the vessel and relative bearing and estimated distance to
the animal) until the animal dives or moves out of visual range of the
observer. A PSO would continue to observe the area to watch for the
animal to resurface or for additional animals that may surface in the
area. Visual PSOs shall communicate all observations to PAM operators,
including any determination by the PSO regarding species
identification, distance, and bearing and the degree of confidence in
the determination.
As noted previously, all source vessels must carry a minimum of one
experienced visual PSO and two experienced PAM operators. The observer
designated as lead PSO (including the full team of visual PSOs and PAM
operators) must have experience as a visual PSO. The applicant may
determine how many additional PSOs are required to adequately fulfill
the requirements specified here. To summarize, these requirements are:
(1) 24-Hour acoustic monitoring during use of the acoustic source in
waters deeper than 100 m; (2) visual monitoring during use of the
acoustic source by two PSOs during all daylight hours, with one visual
PSO on-duty during nighttime ramp-ups; (3) maximum of two consecutive
hours on watch followed by a minimum of one hour off watch for visual
PSOs and a maximum of four consecutive hours on watch followed by a
minimum of two consecutive hours off watch for PAM operators; and (4)
maximum of 12 hours of observational effort per 24-hour period for any
PSO, regardless of duties. We invite comment on the mitigation-related
monitoring requirements proposed for deep penetration airgun survey
operations.
Shallow Penetration Airgun--We are proposing that shallow
penetration airgun surveys (those using a total volume of airguns less
than or equal to 400 in\3\) follow the same requirements described
above for deep penetration surveys, with one notable exception. The use
of PAM is not required, except to begin use of the airgun(s) at night
in waters deeper than 100 m. A nighttime start-up must follow the same
protocol described above for deep-penetration surveys: Monitoring of
the PAM system during a 30-minute pre-clearance period and during the
ramp-up period (if applicable). If a PAM system is used during a
shallow penetration survey, the PAM operator must have prior experience
and training but may be a crew member, and the PAM system does not need
to be monitored during full-power firing.
Non-Airgun HRG Surveys--HRG surveys would differ from the
previously described protocols for airgun surveys and, as described
previously, we differentiate between deep-water (greater than 200 m)
and shallow-water HRG. Water depth in the GOM provides a reliable
indicator of the marine mammal fauna that may be encountered and,
therefore, the complexity of likely observations and concern related to
potential effects on deep-diving and/or sensitive species. We are
proposing to generally follow the HRG protocol described in Appendix B
of BOEM's PEIS (BOEM, 2017), with some differences.
Deep-water HRG surveys would be required to employ a minimum of one
independent visual PSO during all daylight operations, in the same
manner as was described for airgun surveys. Shallow-water HRG surveys
would be required to employ a minimum of one visual PSO, which may be a
crew member. PSOs employed during shallow-water HRG surveys would only
be required during a pre-clearance period. PAM would not be required
for any HRG survey.
PAM Malfunction--Emulating sensible protocols described by the New
Zealand Department of Conservation for airgun surveys conducted in New
Zealand waters (DOC, 2013), we are proposing that survey activity may
continue for brief periods of time when the PAM system malfunctions or
is damaged. Activity may continue for 30 minutes without PAM while the
PAM operator diagnoses the issue. If the diagnosis indicates that the
PAM system must be repaired to solve the problem, operations may
continue for an additional two hours without acoustic monitoring under
the following conditions:
Daylight hours and sea state is less than or equal to
Beaufort sea state (BSS) 4;
No marine mammals (excluding delphinids) detected
solely by PAM in the exclusion zone (see below) in the previous two
hours;
NMFS is notified via email as soon as practicable with
the time and location in which operations began without an active
PAM system; and
Operations with an active acoustic source, but without
an operating PAM system, do not exceed a cumulative total of four
hours in any 24-hour period.
Practicability--As discussed above, both visual and acoustic
monitoring capabilities are critical components of any detection-based
mitigation plan, and are routine requirements around the world. Without
the use of acoustic monitoring, even during periods of good visibility,
species projected to bear the greatest consequences of effects from the
specified activity (e.g., beaked whales and sperm whales; see
``Negligible Impact Analysis and Preliminary Determination'') would go
undetected much of the time. In addition, the data collected through
both visual and acoustic monitoring comprises a majority of the
separate monitoring requirements proposed here to satisfy the
requirements of the MMPA (see ``Proposed Monitoring and Reporting'').
[[Page 29270]]
The use of visual observers has historically been required by BOEM;
therefore, the RIA does not assess the costs associated with our
proposal to continue this requirement. The use of PAM came into use in
the GOM via an incentive scheme introduced in MMS's 2007 Notice to
Lessees concerning ``Implementation of Seismic Survey Mitigation
Measures and Protected Species Observer Program'' (NTL No. 2007-G02),
which allowed nighttime start-ups conditional upon use of PAM. More
recently, use of PAM in the GOM was expanded pursuant to the terms of
the 2013 settlement agreement (as amended and extended through
stipulated agreements) referenced above, in which industry parties
agreed to use PAM in water depths greater than 100 m during times of
reduced visibility. The RIA considers the likely incremental costs of
our proposal to require the use of PAM at all times in waters greater
than 100 meters in depth and associated shutdowns for detections of
``whales'' (i.e., sperm whales, baleen whales, beaked whales, and Kogia
spp.), reflecting the increased costs associated with hardware,
software, personnel, and additional shutdowns due to acoustic
detections relative to both pre-2013 settlement agreement and post-2013
settlement agreement. The range of costs shown in Table 10 reflects the
range of projected activity levels provided by BOEM. Please see the RIA
for full details. Operationally, use of PAM should not present
meaningful difficulty to operators because PAM has been used in some
form in the GOM for many years.
In consideration of the expected benefits of the expanded PAM
requirements in reducing the probability or severity of impacts to
marine mammals species or stocks and the practicability for applicant
implementation (e.g., in light of the costs and historical use), we
preliminarily determine these measures are warranted. We invite comment
on the costs for the additional observer and monitoring requirements
and our interpretation of the analysis for determining what measures
are warranted.
Exclusion Zone and Buffer Zone
For deep penetration airgun surveys, we are proposing that the PSOs
shall establish and monitor a 500-m exclusion zone and additional 500-m
buffer zone (total 1 km) during the pre-clearance period and a 500-m
exclusion zone during the ramp-up and operational periods. PSOs should
focus their observational effort within this 1-km zone, although
animals observed at greater distances should be recorded and mitigation
action taken as necessary (see below). For shallow penetration airgun
surveys, we are proposing that the PSO shall establish and monitor a
200-m exclusion zone with additional 200-m buffer (total 400 m zone)
during the pre-clearance period and a 200-m exclusion zone during the
ramp-up (for small arrays only, versus single airguns) and operational
periods. These zones would be based upon radial distance from any
element of the airgun array or from a single airgun (rather than being
based on the center of the array or around the vessel itself). During
use of the acoustic source, occurrence of marine mammals within the
buffer zone (but outside the exclusion zone) would be communicated to
the operator to prepare for the potential shutdown of the acoustic
source. Use of the buffer zone in relation to ramp-up is discussed
under ``Ramp-up.'' Further detail regarding the exclusion zone and
shutdown requirements is given under ``Exclusion Zone and Shutdown
Requirements.''
For deep-water non-airgun HRG surveys, the PSO would establish and
monitor a 400-m zone during the pre-clearance period and a 200-m
exclusion zone during the operational periods (the latter as required
under BOEM's HRG protocol). For shallow-water non-airgun HRG surveys,
the PSO would establish and monitor and 200-m pre-clearance zone (no
shutdowns required during operational periods).
Ramp-Up
Ramp-up of an acoustic source is intended to provide a gradual
increase in sound levels, enabling animals to move away from the source
if the signal is sufficiently aversive prior to its reaching full
intensity. We are proposing that ramp-up is required for all airgun
surveys (unless using only one airgun), but is not required for non-
airgun HRG surveys, as the types of acoustic sources used in such
surveys are not typically amenable to ``ramping up'' the acoustic
output in the way that multi-element airgun surveys are. We infer on
the basis of behavioral avoidance studies and observations that this
measure results in some reduced potential for auditory injury and/or
more severe behavioral reactions. Stone (2015a) reported on behavioral
observations during airgun surveys from 1994-2010, stating that
detection rates of cetaceans during ramp-up were significantly lower
than when the airguns were not firing and on surveys with large arrays
(defined in that study as greater than 500 in\3\), more cetaceans were
observed avoiding or traveling away from the survey vessel during the
ramp-up than at any other time. Dunlop et al. (2016) studied the effect
of ramp-up during an airgun survey on migrating humpback whales,
comparing ramp-up versus use of a constant source level operating at a
higher level than the initial ramp-up stage but lower than at full
power. Although behavioral response indicating potential avoidance was
observed, there was no evidence that audibly increasing levels during
ramp-up was more effective in this experimental context at causing
aversion than was a constant source. Regardless, the majority of whale
groups did avoid the source vessel at distances greater than the radius
of most mitigation zones (Dunlop et al., 2016). Von Benda-Beckmann et
al. (2013), in a study of the effectiveness of ramp-up for sonar, found
that ramp-up procedures reduced the risk of auditory injury for killer
whales, and that extending the duration of ramp-up did not have a
corresponding effect on mitigation benefit. Although this measure is
not proven and some arguments have been made that use of ramp-up may
not have the desired effect of aversion (which is itself a potentially
negative impact assumed to be better than the alternative), ramp-up
remains a relatively low-cost, common-sense component of standard
mitigation for airgun surveys. Ramp-up is most likely to be effective
for more sensitive species (e.g., beaked whales) (e.g., Tyack et al.,
2011; DeRuiter et al., 2013; Miller et al., 2015).
The ramp-up procedure involves a step-wise increase in the number
of airguns firing and total array volume until all operational airguns
are activated and the full volume is achieved. Ramp-up would be
required at all times as part of the activation of the acoustic source
(including source tests; see ``Miscellaneous Protocols'' for more
detail) and may occur at times of poor visibility, assuming appropriate
acoustic monitoring with no detections in the 30 minutes prior to
beginning ramp-up. Acoustic source activation should only occur at
night where operational planning cannot reasonably avoid such
circumstances. For example, a nighttime initial ramp-up following port
departure is reasonably avoidable and may not occur. Ramp-up may occur
at night following acoustic source deactivation due to line turn or
mechanical difficulty. The operator must notify a designated PSO of the
planned start of ramp-up as agreed-upon with the lead PSO; the
notification time should not be less than 60 minutes prior to the
planned ramp-up. A designated
[[Page 29271]]
PSO must be notified again immediately prior to initiating ramp-up
procedures and the operator must receive confirmation from the PSO to
proceed.
We are proposing that ramp-up procedures follow the recommendations
of IAGC (2015). Ramp-up would begin by activating a single airgun
(i.e., array element) of the smallest volume in the array. Ramp-up
continues in stages by doubling the number of active elements at the
commencement of each stage, with each stage of approximately the same
duration. Total duration should be not less than approximately 20
minutes but is not prescribed and will vary depending on the total
number of stages. There will generally be one stage in which doubling
the number of elements is not possible because the total number is not
even. This should be the last stage of the ramp-up sequence. We are
proposing that the operator would be required to provide information to
the PSO documenting that appropriate procedures were followed, and
request comment on how this information would best be documented. Ramp-
ups should be scheduled so as to minimize the time spent with source
activated prior to reaching the designated run-in. We are proposing to
adopt this approach to ramp-up (increments of array elements) because
we believe it is relatively simple to implement for the operator as
compared with more complex schemes involving activation by increments
of array volume, or activation on the basis of element location or
size. Such approaches may also be more likely to result in irregular
leaps in sound output due to variations in size between individual
elements within an array and their geometric interaction as more
elements are recruited. It may be argued whether smooth incremental
increase is necessary, but stronger aversion than is necessary should
be avoided. The approach proposed here is intended to ensure a
perceptible increase in sound output per increment while employing
increments that produce similar degrees of increase at each step. We
request comment on the proposed ramp-up procedures and requirements.
During deep penetration airgun surveys, we are proposing that PSOs
must monitor a 1,000-m zone (or to the distance visible if less than
1,000 m) for a minimum of 30 minutes prior to ramp-up (i.e., pre-
clearance) or start-up (for single airgun or non-airgun surveys). While
the delineation of zones is typically associated with shutdown, the
period during which use of the acoustic source is being initiated is
critical, and in order to avoid more severe behavioral reactions it is
important to be cautionary regarding marine mammal presence in the
vicinity when the source is turned on. This requirement has broad
acceptance in other required protocols: The Brazilian Institute of the
Environment and Natural Resources requires a 1,000-m pre-clearance zone
(IBAMA, 2005), the New Zealand Department of Conservation requires that
a 1,000-m zone be monitored as both a pre-clearance and a shutdown zone
for most species (DOC, 2013), and the Australian Department of the
Environment, Water, Heritage and the Arts requires an even more
protective scheme, in which a 2,000-m ``power down'' zone is maintained
for higher-power surveys (DEWHA, 2008). Broker et al. (2015) describe
the use of a precautionary 2-km exclusion zone in the absence of sound
source verification (SSV), with a minimum zone radius of 1 km
(regardless of SSV results). We believe that the simple doubling of the
proposed exclusion zone described here is appropriate for use as a pre-
clearance zone. Thus, the pre-clearance zone would be 1,000 m for deep
penetration airgun surveys, 400 m for shallow penetration airgun
surveys or deep-water HRG surveys, and 200 m for shallow-water HRG
surveys. We request comment on this interpretation of a pre-clearance
zone which would provide the appropriate protections for the different
survey types.
The pre-clearance period may occur during any vessel activity
(i.e., transit, line turn). Ramp-up must be planned to occur during
periods of good visibility when possible; operators may not target the
period just after visual PSOs have gone off duty. Following
deactivation of the source for reasons other than mitigation, the
operator must communicate the near-term operational plan to the lead
PSO with justification for any planned nighttime ramp-up. Any suspected
patterns of abuse must be reported by the lead PSO to be investigated
by NMFS. Ramp-up may not be initiated if any marine mammal is within
the designated 1,000-m zone. If a marine mammal is observed within the
zone during the pre-clearance period, ramp-up may not begin until the
animal(s) has been observed exiting the zone or until an additional
time period has elapsed with no further sightings. We suggest an
appropriate elapsed time period should be 15 minutes for small
odontocetes and 30 minutes for all other species, and request comment
on this proposal. PSOs will monitor the 500-m exclusion zone during
ramp-up, and ramp-up must cease and the source shut down upon
observation of marine mammals within or approaching the zone.
Exclusion Zone and Shutdown Requirements
Deep Penetration Airgun--An exclusion zone is a defined area within
which occurrence of a marine mammal triggers mitigation action intended
to reduce potential for certain outcomes, e.g., auditory injury, more
severe disruption of behavioral patterns. For deep penetration airgun
surveys, we propose that PSOs must establish a minimum exclusion zone
with a 500-m radius as a perimeter around the outer extent of the
airgun array (rather than being delineated around the center of the
array or the vessel itself). If a marine mammal appears within or
enters this zone, the acoustic source would be shut down (i.e., power
to the acoustic source must be immediately turned off). If a non-
delphinid marine mammal is detected acoustically, the acoustic source
would be shut down, unless the PAM operator is confident that the
animal detected is outside the exclusion zone or that the detected
species is not subject to the shutdown requirement.
The 500-m radial distance of the standard exclusion zone is
expected to contain sound levels exceeding peak pressure injury
criteria for all hearing groups other than, potentially, high-frequency
cetaceans, while also providing a consistent, reasonably observable
zone within which PSOs would typically be able to conduct effective
observational effort. Although significantly greater distances may be
observed from an elevated platform under good conditions, we believe
that 500 m is likely regularly attainable for PSOs using the naked eye
during typical conditions. In addition, an exclusion zone is expected
to be helpful in avoiding more severe behavioral responses. Behavioral
response to an acoustic stimulus is determined not only by received
level but by context (e.g., activity state) including, importantly,
proximity to the source (e.g., Southall et al., 2007; Ellison et al.,
2012; DeRuiter et al., 2013). Ellison et al. (2012) describe a
qualitative, 10-step index for the severity of behavioral response on
the basis of the observed physical magnitude of the response (e.g.,
minor change in orientation, change in respiration rate, fleeing the
area) and its potential biological significance (e.g., cessation of
vocalizations, abandonment of feeding, separation of mother and
offspring). In prescribing an exclusion zone, we seek not only to avoid
most potential auditory injury but also to reduce the likely severity
of the behavioral
[[Page 29272]]
response at a given received level of sound.
Use of monitoring and shutdown or power-down measures within
defined exclusion zone distances is inherently an essentially
instantaneous proposition--a rule or set of rules that requires
mitigation action upon detection of an animal. This indicates that
definition of an exclusion zone on the basis of cumulative sound
exposure level (cSEL) thresholds, which require that an animal
accumulate some level of sound energy exposure over some period of time
(e.g., 24 hours), has questionable relevance as a standard protocol. A
PSO aboard a mobile source will typically have no ability to monitor an
animal's position relative to the acoustic source over relevant time
periods for purposes of understanding whether auditory injury is likely
to occur on the basis of cumulative sound exposure and, therefore,
whether action should be taken to avoid such potential.
Cumulative SEL thresholds are more relevant for purposes of
modeling the potential for auditory injury than they are for dictating
real-time mitigation, though they can be informative (especially in a
relative sense). We recognize the importance of the accumulation of
sound energy to an understanding of the potential for auditory injury
and that it is likely that, at least for low-frequency cetaceans, some
potential auditory injury is likely impossible to mitigate and should
be considered for authorization.
Considering both the dual-metric thresholds described previously
(and shown in Table 7) and hearing group-specific marine mammal
auditory weighting functions in the context of the airgun sources
considered here, auditory injury zones indicated by the peak pressure
metric are expected to be predominant for both mid- and high-frequency
cetaceans, while zones indicated by cSEL criteria are expected to be
predominant for low-frequency cetaceans. Assuming a source level of
255.2 dB 0-pk SPL for the notional 8,000 in\3\ array and spherical
spreading propagation, distances for exceedance of group-specific peak
injury thresholds are as follows: 65 m (LF), 18 m (MF), and 457 m (HF)
(for high-frequency cetaceans, although the notional source parameters
indicate a zone less than 500 m, we recognize that actual isopleth
distances will vary based on specific array characteristics and site-
specific propagation characteristics, and that it is therefore possible
that a real-world distance to the injury threshold could exceed 500 m).
Assuming a source level of 227.7 dB 0-pk SPL for the notional 90 in\3\
single airgun and spherical spreading propagation, these distances
would be 3 m (LF) and 19 m (HF) (the source level is lower than the
threshold criterion value for mid-frequency cetaceans).
Consideration of auditory injury zones based on cSEL criteria are
dependent on the animal's applied hearing range and how that overlaps
with the frequencies produced by the sound source of interest in
relation to marine mammal auditory weighting functions (NMFS, 2016). As
noted above, these are expected to be predominant for low-frequency
cetaceans because their most susceptible hearing range overlaps the low
frequencies produced by airguns, while the modeling indicates that
zones based on peak pressure criteria dominate for mid- and high-
frequency cetaceans. In order to evaluate notional zone sizes and to
incorporate the technical guidance's weighting functions over a seismic
array's full acoustic band, we obtained unweighted spectrum data
(modeled in 1 Hz bands) for a reasonably equivalent acoustic source
(i.e., a 36-airgun array with total volume of 6,600 in\3\). Using these
data, we made adjustments (dB) to the unweighted spectrum levels, by
frequency, according to the weighting functions for each relevant
marine mammal hearing group. We then converted these adjusted/weighted
spectrum levels to pressures (micropascals) in order to integrate them
over the entire broadband spectrum, resulting in broadband weighted
source levels by hearing group that could be directly incorporated
within NMFS's User Spreadsheet (i.e., override the spreadsheet's more
simple weighting factor adjustment). Using the User Spreadsheet's
``safe distance'' methodology for mobile sources (described by Sivle et
al., 2014) with appropriate dB adjustments derived from the methodology
described above, and inputs assuming a 231.8 dB SEL source level for
the notional 8,000 in\3\ array, spherical spreading propagation, a
source velocity of 4.5 kn, pulse duration of 100 ms, and a 25-m shot
interval (shot intervals may vary, with longer shot intervals resulting
in smaller calculated zones), distances for group-specific threshold
criteria are as follows: 574 m (LF), 0 m (MF), and 1 m (HF).
We also assessed the potential for injury based on the accumulation
of energy resulting from use of the single airgun and, assuming a
source level of 207.8 dB SEL, there would be no realistic zone within
which injury would occur. On the basis of this finding as well as the
potential zone sizes based on the peak pressure criteria described
above, we do not expect any reasonable potential for auditory injury
resulting from use of the single airgun. No potential injurious
exposures were predicted for single airgun surveys (Zeddies et al.,
2015, 2017a).
We expect that the proposed 500-m exclusion zone would typically
contain the entirety of any potential injury zone for mid-frequency
cetaceans (realistically, there is no such zone), while the zones
within which injury could occur may be larger for high-frequency
cetaceans (on the basis of peak pressure and depending on the specific
array) and for low-frequency cetaceans (on the basis of cumulative
sound exposure). These findings indicate that auditory injury is
unlikely for mid-frequency cetaceans.
In summary, our intent in prescribing a standard exclusion zone
distance is to (1) encompass zones for most species within which
auditory injury could occur on the basis of instantaneous exposure; (2)
provide additional protection from the potential for more severe
behavioral reactions (e.g., panic, antipredator response) for marine
mammals at relatively close range to the acoustic source; (3) provide
consistency and ease of implementation for PSOs, who need to monitor
and implement the exclusion zone; and (4) to define a distance within
which detection probabilities are reasonably high for most species
under typical conditions. Our use of 500 m as the zone is not based
directly on any quantitative understanding of the range at which
auditory injury would be entirely precluded or any range specifically
related to disruption of behavioral patterns. Rather, we believe it is
a reasonable combination of factors. This zone has been proven as a
feasible measure through past implementation by operators in the GOM.
In summary, a practicable criterion such as this has the advantage of
familiarity and simplicity while still providing in most cases a zone
larger than relevant auditory injury zones, given realistic movement of
source and receiver. Increased shutdowns, without a firm idea of the
outcome the measure seeks to avoid, simply displace survey activity in
time and increase the total duration of acoustic influence as well as
total sound energy in the water (due to additional ramp-up and overlap
where data acquisition was interrupted). The shutdown requirement
described here would be required for most marine mammals, with the
exception of small delphinoids, described in the following section; and
Bryde's whales, any large whale observed with calf, sperm whales,
beaked whales, and Kogia spp.,
[[Page 29273]]
described in the subsequent section entitled ``Other Shutdown
Requirements.'' We request comment on our interpretation of the data,
proposed standard exclusion zone, and shutdown requirements for most
species (see subsequent proposed exceptions) during deep penetration
airgun surveys.
Dolphin Exception--As defined here, the small delphinoid group is
intended to encompass those members of the Family Delphinidae most
likely to voluntarily approach the source vessel for purposes of
interacting with the vessel and/or airgun array (e.g., bow riding).
This exception to the shutdown requirement applies solely to specific
genera of small dolphins--Steno, Tursiops, Stenella, and Lagenodelphis
(see Table 3)--and applies under all circumstances, regardless of what
the perception of the animal(s) behavior or intent may be. Variations
of this measure that include exceptions based on animal behavior--e.g.,
``bow-riding'' dolphins, or only ``traveling'' dolphins, meaning that
the intersection of the animal and exclusion zone may be due to the
animal rather than the vessel--have been proposed by both NMFS and BOEM
and have been criticized, in part due to the subjective on-the-spot
decision-making this scheme would require of PSOs. If the mitigation
requirements are not sufficiently clear and objective, the outcome may
be differential implementation across surveys as informed by individual
PSOs' experience, background, and/or training. The proposal here is
based on several factors: The lack of evidence of or presumed potential
for the types of effects to these species of small delphinoid that our
shutdown proposal for other species seeks to avoid, the uncertainty and
subjectivity introduced by such a decision framework, and the
practicability concern presented by the operational impacts. While
there may be some potential for adverse impacts to dolphins--Gray and
Van Waerebeek (2011) report an observation of a pantropical spotted
dolphin exhibiting severe distress in close proximity to an airgun
survey, examine other potential causes for the display, and ultimately
suggest a cause-effect relationship--we are not aware of other such
incidents despite a large volume of observational effort during airgun
surveys in the GOM, where dolphin shutdowns have not previously been
required. Dolphins have a relatively high threshold for the onset of
auditory injury (i.e., permanent threshold shift) and more severe
adverse behavioral responses seem less likely given the evidence of
purposeful approach and/or maintenance of proximity to vessels with
operating airguns.
The best available scientific evidence indicates that auditory
injury as a result of airgun sources is extremely unlikely for mid-
frequency cetaceans, primarily due to a relative lack of sensitivity
and susceptibility to noise-induced hearing loss at the frequency range
output by airguns (i.e., most sound below 500 Hz) as shown by the mid-
frequency cetacean auditory weighting function (NMFS, 2016). Criteria
for temporary threshold shift (TTS) in mid-frequency cetaceans for
impulsive sounds were derived by experimental measurement of TTS in
beluga whales exposed to pulses from a seismic watergun; dolphins
exposed to the same stimuli in this study did not display TTS (Finneran
et al., 2002). Moreover, when the experimental watergun signal was
weighted appropriately for mid-frequency cetaceans, less energy was
filtered than would be the case for an airgun signal. More recently,
Finneran et al. (2015) exposed bottlenose dolphins to repeated pulses
from an airgun and measured no TTS.
While dolphins are observed voluntarily approaching source vessels
(e.g., bow-riding or interacting with towed gear), the reasons for the
behavior are unknown. In context of an active airgun array, the
behavior cannot be assumed to be harmless. Although bow-riding
comprises approximately 30 percent of behavioral observations in the
GOM, there is a much lower incidence of the behavior when the acoustic
source is active (Barkaszi et al., 2012), and this finding was
replicated by Stone (2015a) for surveys occurring in United Kingdom
waters. There appears to be strong evidence of aversive behavior by
dolphins during firing of airguns. Barkaszi et al. (2012) found that
the median closest distance of approach to the acoustic source was at
significantly greater distances during times of full-power source
operation when compared to silence, while Stone (2015a) and Stone and
Tasker (2006) reported that significant behavioral responses, including
avoidance and changes in swimming or surfacing behavior, were evident
for dolphins during firing of large arrays. Goold and Fish (1998)
described a ``general pattern of localized disturbance'' for dolphins
in the vicinity of an airgun survey. However, while these general
findings--typically, dolphins will display increased distance from the
acoustic source, decreased prevalence of ``bow-riding'' activities, and
increases in surface-active behaviors--are indicative of adverse or
aversive responses that may be construed as ``take'' (as defined by the
MMPA), they are not indicative of any response of a severity such that
the need to avoid it outweighs the impact on practicability for the
industry and operators.
Additionally, increased shutdowns resulting from such a measure
would require source vessels to revisit the missed track line to
reacquire data, resulting in an overall increase in the total sound
energy input to the marine environment and an increase in the total
duration over which the survey is active in a given area.
Instead of shutdown, if a dolphin of the indicated genera (Steno,
Tursiops, Stenella, and Lagenodelphis) appears within or enters the
500-m exclusion zone, or is acoustically detected and localized within
the zone, we present two alternatives.
Proposal 1: The acoustic source would be powered down to
the smallest single element of the array. The power-down is intended to
minimize potential disturbance to dolphins in a practicable way, by
reducing the acoustic output while maintaining what should be an
aversive stimulus. Power-down conditions would be maintained until the
animal(s) is observed exiting the exclusion zone or for 15 minutes
beyond the last observation of the animal, following which full-power
operations may be resumed without ramp-up. A source vessel traveling at
a typical speed of approximately 4.5 kn would transit approximately 2
km during this period. We expect that the resulting gap in data
acquisition would be sufficiently small as to not require reshooting
for infill; therefore, increased time over which acoustic energy is
output, as well as significant operational impacts, would be avoided
while maintaining reasonable protections for dolphins.
Proposal 2: No shutdown or power-down would be required.
We described above the information that supports our preliminary
decision that an exception to the general shutdown requirement is
warranted for small dolphins, as well as the information that we
believe indicates that a power-down requirement is warranted in lieu of
shutdown. However, members of the public may interpret this information
as supporting an exception to the shutdown requirement with no power-
down requirement.
We request comment on both proposals and other variations of these
proposals, including our interpretation of the data and any other data
that support the necessary findings regarding small dolphins for no
shutdown and no power-down or no shutdown but a power-down.
[[Page 29274]]
Although other mid-frequency hearing specialists (e.g., large
delphinoids) are considered no more likely to incur auditory injury
than are small delphinoids, they are much less likely to approach
vessels. Therefore, we have evaluated that retaining a shutdown
requirement for large delphinoids would not have similar impacts in
terms of either practicability for the applicant or corollary increase
in sound energy output and time on the water. We do anticipate some
benefit for a shutdown requirement for large delphinoids in that it
simplifies somewhat the total array of decision-making for PSOs and may
preclude any potential for physiological effects other than to the
auditory system as well as some more severe behavioral reactions for
any such animals in close proximity to the source vessel. The
variations in regulatory text for these proposals can be found in
``Alternative Regulatory Text,'' later in this preamble, and in the
regulatory text at the end of the document.
Practicability--The requirement to use a generalized 500-m
exclusion zone and to require shutdown upon observation of whales
within that zone has historically been required by BOEM. Here, we
assess practicability for possible dolphin shutdowns (described in full
in the RIA). The IAGC provided information in response to a 2014 survey
regarding the costs of survey activities including, by survey type,
average survey duration, mobilization and pre-mobilization costs, and
vessel operating costs per day, allowing for estimates of total average
survey costs. IAGC also provided information relating to estimated
average shutdown time following marine mammal observations in the
exclusion zone and typical additional hours required to reshoot the
areas missed during the shutdown period. For the latter, estimates
ranged from 1-2 additional hours up to 12 hours (for 3D WAZ surveys).
Barkaszi et al. (2012) found that small dolphins were observed within
the exclusion zone on 5.7 percent of days, and that large dolphins were
observed in the exclusion zone on 1.2 percent of days (unidentified
delphinid species were observed on an additional 1.2 percent of days).
The cost of shutdowns for dolphins in the exclusion zone is a function
of the total number of days added to a survey, which accrue via (1)
total time from shutdown until resuming data acquisition (1.6-2 hours)
and (2) time required to reshoot an interrupted survey line (1-12
hours, depending on the survey type). To quantify this cost, the total
number of added days is multiplied by the daily vessel operating cost
for each survey type that uses airguns, with resulting annualized costs
for shutdowns due to dolphins in the exclusion zone depending on actual
level of activity (see RIA for cost estimates). In consideration of the
preceding discussion of expected benefit from shutdowns for dolphins in
context with these impacts on operations, we do not consider full
shutdown for small dolphins in the exclusion zone to be warranted. The
alternative presented requiring power-down for small dolphins in the
exclusion zone is expected to cost less because of the ability to start
back up without a ramp-up and the potentially reduced need to reshoot
lines. The same would hold true for the alternative presented requiring
no power-down based on there being no need to modify the survey at all.
Operationally, we have attempted to minimize the potential for
subjective and potentially inconsistent decision-making by PSOs. NMFS
expects that large delphinoids (e.g., false killer whales, melon-headed
whales) in general are easily distinguished from small delphinoids
(e.g., spotted dolphins, Clymene dolphins) in general by trained,
experienced observers on the basis of differences in size, color, and
cranial/dorsal morphology, and requests any information relating to
this assumption. Based on the protective value of the described measure
and the understanding of practicability, we preliminarily determine the
power-down measures are warranted.
Other Shutdown Requirements--We are proposing that shutdown of the
acoustic source should also be required in the event of certain other
observations regardless of the defined exclusion zone. It must be noted
up front that any such observations would still be within range of
where behavioral disturbance of some form and degree would be likely to
occur, e.g., Zeddies et al. (2015) estimated unweighted mean 95 percent
range to 160 dB rms threshold (i.e., the 50 percent midpoint for
behavioral disturbance) levels across water depths and seasons at
approximately 13 km (range 7.7-21.8 km) for the 8,000 in\3\ array
(Zeddies et al., 2015). Thus, for the species or situations listed
below, we present two alternatives:
Proposal 1: Shutdown of the acoustic source would occur in
the circumstances listed below, with no distance limit (i.e., at any
distance from the source). While visual PSOs would focus observational
effort within the vicinity of the acoustic source and vessel (i.e.,
approximately 1 km radius), this does not preclude them from periodic
scanning of the remainder of the visible area, and we do not have a
reason to believe that such periodic scans by professional PSOs would
hamper the ability to maintain observation of areas closer to the
source and vessel.
Proposal 2: Shutdown of the acoustic source would occur in
the circumstances listed below, only within 1 km of the source
(measured as the radial distance from any element of the airgun array).
We request comment on both proposals and other variations of these
proposals, including our interpretation of the data and any other data
that support the necessary findings regarding initiating shutdown for
certain circumstances at any distance or within 1 km. The variations in
regulatory text for these proposals can be found in ``Alternative
Regulatory Text,'' later in this preamble, and in the regulatory text
at the end of the document.
Circumstances triggering Proposal 1 or Proposal 2 include:
Upon detection (visual or acoustic) of a Bryde's whale. On
the basis of the findings of NMFS's status review (described in a NOAA
technical memorandum; Rosel et al., 2016), NMFS has proposed to list
the GOM Bryde's whale as an endangered species pursuant to the ESA (81
FR 88639; December 8, 2016). These whales form a small and resident
population in the northeastern GOM, with a highly restricted geographic
range and a very small population abundance (fewer than 100)--recently
determined by a status review team to be ``at or below the near-
extinction population level'' (Rosel et al., 2016). The review team
stated that, aside from the restricted distribution and small
population, the whales face a significant suite of anthropogenic
threats, one of which is noise produced by geophysical surveys. We
believe it appropriate to eliminate potential effects to individual
Bryde's whales to the extent practicable. As described previously,
there may be rare sightings of vagrant baleen whales of other species
in the GOM; if identification of the observed whale is inconclusive the
shutdown must be implemented.
Upon observation of a large whale (i.e., sperm whale or
any baleen whale) with calf, with ``calf'' defined as an animal less
than two-thirds the body size of an adult observed to be in close
association with an adult. Groups of whales are likely to be more
susceptible to disturbance when calves are present (e.g., Bauer et al.,
1993), and disturbance of cow-calf pairs could potentially result in
separation of
[[Page 29275]]
vulnerable calves from adults. McCauley et al. (2000a) found that
groups of humpback whale females with calves consistently avoided a
single operating airgun, while male humpbacks were attracted to it,
concluding that cow-calf pairs are more likely to exhibit avoidance
responses to unfamiliar sounds and that such responses should be a
focus of management. Behavioral disturbance has been implicated in
mother-calf separations for odontocete species as well (Noren and
Edwards, 2007; Wade et al., 2012). Separation, if it occurred, could be
exacerbated by airgun signals masking communication between adults and
the separated calf (Videsen et al., 2017). Absent separation, airgun
signals can disrupt or mask vocalizations essential to mother-calf
interactions. Given the status of large whales in the GOM, the
consequences of potential loss of calves, as well as the functional
sensitivity of the mysticete whales to frequencies associated with the
subject geophysical survey activity, we believe this measure is
warranted by the MMPA's least practicable adverse impact standard.
Upon acoustic detection of a sperm whale. Sperm whales are
not necessarily expected to display physical avoidance of sound sources
(e.g., Madsen et al., 2002a; Jochens et al., 2008; Winsor et al.,
2017). Although Winsor et al. (2017) report that distances and
orientations between tagged whales and active airgun arrays appeared to
be randomly distributed with no evidence of horizontal avoidance, it
must be noted that their study was to some degree precipitated by an
earlier observation of significantly decreased sperm whale density in
the presence of airgun surveys (Mate et al., 1994). However, effects on
vocal behavior are common (e.g., Watkins and Schevill, 1975; Watkins et
al., 1985). In response to a low-frequency tone, sperm whales were
observed to cease vocalizing (vocalizations detected during 24 percent
of a baseline period and not detected during transmission;
vocalizations resumed at most 36 hours post-transmission). Although the
signal characteristics in this study were dissimilar to airgun signals,
the authors also note that an airgun survey was being conducted
simultaneously with signals exceeding background noise by 10-15 dB
(Bowles et al., 1994). The sperm whale's primary means of locating prey
is echolocation (Miller et al., 2004), and multiple studies have shown
that noise can disrupt feeding behavior and/or significantly reduce
foraging success for sperm whales at relatively low levels of exposure
(e.g., Miller et al., 2009, 2012; Isojunno et al., 2016; Sivle et al.,
2012; Cure et al., 2016). Effects on energy intake with no immediate
compensation, as is suggested by disruption of foraging behavior
without corollary movements to new locations, would be expected to
result in bioenergetics consequences to individual whales. Farmer et
al. (2018) developed a stochastic life-stage structured bioenergetic
model to evaluate the consequences of reduced foraging efficiency in
sperm whales, finding that individual resilience to foraging
disruptions is primarily a function of size (i.e., reserve capacity)
and daily energetic demands, and that the ultimate effects on
reproductive success and individual fitness are largely dependent on
the duration and frequency of disturbance.
Sperm whales in the GOM spend the majority of their time foraging,
engaging in dive cycles consisting of deep dives of approximately 45
minutes followed by shorter surface intervals (resting bouts) of
approximately 10 minutes (Watwood et al., 2006). Sperm whales alternate
between shallow and deep dives over periods of several hours, targeting
predominantly epipelagic prey during shallow dives and benthopelagic
prey during deep dives (Fais et al., 2015). During the search phase of
their dive, whales emit regular clicks with high directionality, high
source levels, and frequencies around 15 kHz, suitable for long-range
sonar (M[oslash]hl et al., 2003). During the capture phase, interclick
interval, amplitude, and signal duration decrease dramatically,
providing rapid updates on the location of prey during capture,
creating a sound termed as either a creak or a buzz (Madsen et al.,
2002b; Miller et al., 2004). On the basis of observed echolocation
during the ascent phase, Fais et al. (2015) concluded that sperm whale
decisions about where to forage during subsequent dives may be based on
both prior foraging success and information gathered during ascent,
suggesting that sperm whales can perform auditory stream segregation of
multiple targets when echolocating, simultaneously tracking several
targets for sequential capture and perceptually organizing a multi-
target auditory scene. As stated by Farmer et al. (2018), this complex
information-gathering allows sperm whales to efficiently locate and
access prey resources in a dark, patchy, and vast environment while
leaving whales vulnerable to reduction in sensory volume and/or
interference with complex auditory stream signal processing (Fais et
al., 2015). Such effects, which may result from increased noise in the
environment, can increase search effort required to locate resources
and ultimately reduce foraging efficiency (e.g., Zollner and Lima,
1999). As deep-diving animals, sperm whales may be expected to be more
consistently exposed to elevated sound levels in the downward-
refracting acoustic environment.
Miller et al. (2009) showed that GOM sperm whales are susceptible
to disruption of foraging behavior upon exposure to relatively moderate
sound levels at distances greater than contemplated for our proposed
general exclusion zone. Although tagged whales did not change
behavioral state during exposure or show horizontal avoidance, they
increased energy put into swimming and their buzz rates (a proxy for
attempts to capture prey) were approximately 20 percent lower (though
not a statistically significant result). One whale, despite not showing
avoidance behavior, engaged in an unusually long resting bout of 265
minutes (compared with typical duration of approximately 10 min),
representing a significant delay in foraging effort (Miller et al.,
2008, 2009). This finding is of particular importance, as it indicates
that sperm whales may not be as likely to show avoidance of active
sound sources which would then leave them more vulnerable to subsequent
foraging disruption--an effect of greater significance. Analysis
conducted by Jochens et al. (2008) suggested that, for these whales, a
20 percent decrease in foraging activity was more likely than no change
in foraging activity, with one whale showing a statistically
significant decrease of 60 percent.
The income breeding strategy used by sperm whales requires stable
or predictable environments that enable continuous energy acquisition
throughout the year, at rates of up to thousands of kilograms of prey
per day (Irvine et al., 2017; Clarke et al., 1993; Farmer et al.,
2018). On days when sperm whale foraging is impaired, whales would
likely compensate for the caloric deficit by depleting carbohydrate
reserves and, secondarily, lipid and protein reserves (Lockyer, 1991;
Castellini and Rea, 1992; Farmer et al., 2018). Energy reserves are
available from carbohydrates in the blubber and muscle; lipids in the
blubber, muscle, and viscera; and proteins in the muscle and viscera.
However, physiological evidence suggests that sperm whales are poorly
adapted to handle periods of food shortage, as the energy density of
sperm whale blubber is much lower than that of baleen whales; sperm
whales do not exhibit appreciable
[[Page 29276]]
changes in blubber thickness relative to body length, even during
lactation; and the vast majority of blubber lipids are stored in a form
that helps to conserve oxygen during metabolism but is less accessible
as a source of energy (Lockyer, 1981; Koopman, 2007; Farmer et al.,
2018). If total energy reserves are depleted below critical levels, an
individual's body condition would be expected to decline over time and,
for pregnant or lactating females, fetus abortion or calf abandonment
could occur (e.g., New et al., 2013). In this way, responses to airgun
survey noise can accrue towards population-level impacts (e.g., New et
al., 2014; King et al., 2015; Fleishman et al., 2016).
Sperm whales in the northern GOM have a relatively small population
abundance, and with a relatively narrow distribution that overlaps
almost completely with areas of current and future geophysical survey
activity and other oil and gas industry activity. Further, most
resident female sperm whale movements in the GOM range within smaller
areas--approximately 200 km around a core home range--although larger
individual and group movements were also observed (Jochens et al.,
2008). The bioenergetic simulations of Farmer et al. (2018) show that
frequent disruptions in foraging, as might be expected when large
amounts of survey activity overlap with areas of importance for sperm
whales, can have potentially severe fitness consequences. Even partial
disturbances of foraging, if sufficiently frequent, may lead to lower
body condition, with potential indirect effects of delayed sexual
maturation or reduced reproductive fitness (Farmer et al., 2018). It is
also unlikely that any ``hunger response'' following disruption of
foraging would result in increases in daily growth rate that could be
expected to offset the effects of sustained foraging disruption (Farmer
et al., 2018). While the modeling exercise conducted by Farmer et al.
(2018) shows that terminal starvation is an unlikely outcome--though
possible in mature whales repeatedly exposed to sound levels that
result in reduced foraging ability over periods of weeks to months--
minor disruptions can cause substantial reductions in available
reserves over time.
Multiple lines of evidence indicate that sperm whales in the
northern GOM are somewhat isolated from global sperm whale populations
(Jochens et al., 2008). The estimated annual rate of increase from
reproduction for GOM sperm whales is less than one percent per year,
while Chiquet et al. (2013) found that reducing the survivorship rate
of mature female sperm whales by as little as 2.2 percent or the
survivorship rate of mothers by as little as 4.8 percent would drop the
asymptotic growth rate of the northern GOM sperm whale population below
one, i.e., a declining population. NOAA estimates that the DWH oil
spill may have caused reproductive failure in 7 percent of female sperm
whales (DWH MMIQT, 2015). Separately, NOAA estimates that 16 percent of
the sperm whale population was exposed to high concentrations of oil
both at the surface and sub-surface, high concentrations of volatile
gases that could be inhaled at the surface, and response activities
including increased vessel operations, dispersant applications, and oil
burns (DWH MMIQT, 2015). Independent of other factors, the DWH oil
spill may have a long-term impact of reducing the GOM sperm whale
population by up to 7 percent, with an estimated time to recovery of 21
years (DWH MMIQT, 2015). Therefore, even in the absence of other future
stressors, the environmental baseline for the GOM sperm whale
population requires that meaningful measures be taken to minimize
disruption of foraging behavior. Such measures are all the more
important, as we have considered but eliminated a time-area restriction
for sperm whales (described below).
We also considered requirement of shutdown upon visual detection of
sperm whales. Here, we assume that acoustic detections of sperm whales
would most likely be representative of the foraging behavior we intend
to minimize disruption of, while visual observations of sperm whales
would represent resting between bouts of such behavior. Occurrence of
resting sperm whales at distances beyond the exclusion zone may not
indicate a need to implement shutdown. We consider these assumptions in
conjunction with an assessment of the costs and operational feasibility
of these measures in ``Practicability,'' below.
Upon observation (visual or acoustic) of a beaked whale or
Kogia spp. These species are behaviorally sensitive deep divers and it
is possible that disturbance could provoke a severe behavioral response
leading to injury (e.g., Wursig et al., 1998; Cox et al., 2006). Unlike
the sperm whale, we recognize that there are generally low detection
probabilities for beaked whales and Kogia spp., meaning that many
animals of these species may go undetected. Barlow (1999) estimates
such probabilities at 0.23 to 0.45 for Cuvier's and Mesoplodont beaked
whales, respectively. However, Barlow and Gisiner (2006) predict a
roughly 24-48 percent reduction in the probability of detecting beaked
whales during seismic mitigation monitoring efforts as compared with
typical research survey efforts, and Moore and Barlow (2013) noted a
decrease in g(0) for Cuvier's beaked whales from 0.23 at BSS 0 (calm)
to 0.024 at BSS 5. Similar detection probabilities have been noted for
Kogia spp., though they typically travel in smaller groups and are less
vocal, thus making detection more difficult (Barlow and Forney, 2007).
As discussed previously in this document (see the ``Estimated Take''
section), there are high levels of predicted exposures for beaked
whales in particular. Because it is likely that only a small proportion
of beaked whales and Kogia spp. potentially affected by the proposed
surveys would actually be detected, it is important to avoid potential
impacts when practicable. Additionally for Kogia spp.--the one species
of high-frequency cetacean likely to be encountered--auditory injury
zones relative to peak pressure thresholds are significantly greater
than for other cetaceans--approximately 500 m from the acoustic source,
depending on the specific real world array characteristics (NMFS,
2016).
Practicability--In the bulleted subsections above, we evaluated the
importance of offering expanded protections via shutdown for these
species/circumstances and, as discussed, we find that avoidance to
extent practicable of acute impacts for Bryde's whales, sperm whales,
beaked whales, and Kogia spp., as well as for large whales with calves,
is important to a reduction of effects for these species. In the RIA,
we evaluate the annualized incremental costs of these expanded measures
(note that the costs of additional shutdowns based on acoustic
detections is included in our previous discussion of costs associated
with expanded use of PAM). Additional requirements for shutdowns based
on visual detections outside the exclusion zone result in a small cost
relative to the benefits afforded by the measures. Additionally, due to
the rarity of visual observations of these species groups, we do not
believe that the expanded shutdowns would cause any undue operational
burden.
In the GOM, we expect that the optimum detection range of sperm
whales in low-noise conditions is likely to be approximately 2-3 km.
This relatively short detection range is likely due to the propagation
conditions resulting when a relatively warmer mixed surface layer
provides a strong negative sound velocity profile, causing strong
downward refraction of acoustic rays. While the maximum detection
[[Page 29277]]
range of vocalizing marine mammals continues to be a challenging area
in use of PAM for mitigation monitoring, basic signal detection theory
dictates that received levels have to exceed certain noise levels in
order for the signal to be detected. We consider the following sonar
equations:
EL = SL-TL (1)
SNR = EL-NR (2)
SE = SNR-DT (3)
where EL is the received level, SL the source level, TL the
transmission loss, SNR the signal-to-noise ratio, NR the received noise
spectral density, SE the signal excess, and DT the detection threshold.
As the signal (in this case, a sperm whale click) propagates from
its source (the whale) through the environment to a receiver (a
hydrophone), its intensity (acoustic power within a unit area) is
reduced due to acoustic energy divergence and attenuation (absorption
and scattering). By the time the whale click reaches the hydrophone,
its received intensity level is greatly reduced from its original
source level. In addition, for the received level to be detected by the
hydrophone, the signal-to-noise ratio (received level minus the
background noise spectral density) must be above a certain detection
threshold, i.e., there must be a positive signal excess.
Based on various studies (Madsen and Mohl, 2000; Mohl et al., 2000;
Thode et al., 2002; Zimmer et al., 2005), the source levels of sperm
whale clicks fall between 202 and 223 dB re 1 [micro]Pa, with a
pronounced directionality and significant energy above 10 kHz. However,
these values are selected from the most intense clicks from each
sequence so they are likely to have been recorded close to the acoustic
axis (Mohl et al., 2000). Considering all recordings, Mohl et al.
(2000) suggest that sperm whale click maximum source levels are in the
range of 175 to 200 dB re 1 [micro]Pa. By using a middle range of the
maximum source level of 188 dB re 1 [micro]Pa with a 50 percent
detection range at 4 km, and assume an ambient noise spectral density
at 75 dB with a detection threshold of 6 dB, the transmission loss at
this range would be 107 dB. By simply applying a geometric spreading
model, it can be shown that the transmission loss (TL) follows TL =
29.7log10(R), where R is the distance from the source in
meters. Please note that this approximation is based on a very low
ambient noise spectrum density (Wenz, 1962).
In the presence of an airgun survey, the background noise level is
expected to be significantly increased as a result of the reverberant
field generated from intense pulses (Guerra et al., 2011; Guan et al.,
2015). It has been shown that the level of elevated inter-pulse noise
levels can be as high as 20 dB within 1 km of an active firing airgun
array of 640 in\3\ (Guan et al., 2015) to 30-45 dB for a 3,147 cu\3\
airgun array (Guerra et al., 2011). Given that towing hydrophones for
PAM used for marine mammal monitoring would be within 1 km from the
airgun source, the received noise spectral density is expected to be
very high. Using a relatively low 25 dB increase from the inter-pulse
noise level to compute detection with the otherwise the same parameters
from the above example in the quiet environment, one would find that a
50 percent detection probability is quickly reduced to 576 m. If, given
the unfavorable signal propagation conduction in the GOM in comparison
to the more favorable conditions in the North Pacific (Barlow and
Taylor, 2005), a 50 percent detection probability at 3 km in quiet
conditions would be reduced to 462 m during the active airgun survey. A
50 percent detection probability at 2 km in quiet conditions would
further reduce the detection range to 339 m.
However, we recognize that the addition of sperm whale shutdowns
based on visual detections beyond the exclusion zone would result in a
larger estimated additional cost per year. Based on these costs, and
our previous discussion of assumptions related to acoustic versus
visual detections of sperm whales, we preliminarily do not believe the
addition of shutdowns for sperm whales based on visual detections at
any distance to be warranted, and request any information from the
public that would be relevant to this determination. For this proposed
rule, we preliminarily determine that the addition of the proposed
shutdown measures described above are warranted when their likely
ability to reduce the probability or severity of impacts on species or
stocks and their habitat is considered along with their practicability.
Other Surveys--Shutdowns for shallow penetration airgun surveys or
deep-water non-airgun HRG surveys would be similar to those described
for deep penetration airgun surveys, except that the exclusion zone
would be defined as a 200-m radial distance around the perimeter of the
acoustic source, in keeping with BOEM's exclusion zone requirements for
their ``HRG survey protocol.'' The special circumstance shutdowns
described above for deep penetration airgun surveys would not be
required. The dolphin exception described for deep penetration airgun
surveys would apply; if the survey is using a small airgun array (i.e.,
less than or equal to 400 in \3\, versus a single airgun), then power-
down should be implemented as described for deep penetration airgun
surveys. As described previously, no shutdowns would be required for
shallow-water non-airgun HRG surveys.
Shutdown Implementation Protocols--Any PSO on duty has the
authority to delay the start of survey operations or to call for
shutdown of the acoustic source. When shutdown is called for by a PSO,
the acoustic source must be immediately deactivated and any dispute
resolved only following deactivation. The operator must establish and
maintain clear lines of communication directly between PSOs on duty and
crew controlling the acoustic source to ensure that shutdown commands
are conveyed swiftly while allowing PSOs to maintain watch; hand-held
UHF radios are recommended. When both visual PSOs and PAM operators are
on duty, all detections must be immediately communicated to the
remainder of the on-duty team for potential verification of visual
observations by the PAM operator or of acoustic detections by visual
PSOs and initiation of dialogue as necessary. When there is certainty
regarding the need for mitigation action on the basis of either visual
or acoustic detection alone, the relevant PSO(s) must call for such
action immediately.
Upon implementation of shutdown, the source may be reactivated
after the animal(s) has been observed exiting the exclusion zone or
following a 30-minute clearance period with no further observation of
the animal(s). Where there is no relevant zone (e.g., shutdowns at any
distance), a 30-minute clearance period must be observed following the
last detection of the animal(s).
If the acoustic source is shut down for reasons other than
mitigation (e.g., mechanical difficulty) for brief periods (i.e., less
than 30 minutes), it may be activated again without ramp-up if PSOs
have maintained constant visual and acoustic observation and no visual
detections of any marine mammal have occurred within the exclusion zone
and no acoustic detections have occurred. We define ``brief periods''
in keeping with other clearance watch periods and to avoid unnecessary
complexity in protocols for PSOs. For any longer shutdown (e.g., during
line turns), pre-clearance watch and ramp-up are required. For any
shutdown at night or in periods of poor visibility (e.g., BSS 4 or
greater), ramp-up is required but if the shutdown period was brief and
[[Page 29278]]
constant observation maintained, pre-clearance watch is not required.
Power-Down
Power-down, as defined here, refers to reducing the array to a
single element as a substitute for full shutdown. We address use of a
single airgun as a ``mitigation source'' below. In a power-down
scenario, it is assumed that reducing the size of the array to a single
element reduces the ensonified area such that an observed animal is
outside of any area within which injury or more severe behavioral
reactions could occur. Zeddies et al. (2015) modeled the 95 percent
ranges for a single airgun as 360 m to the 160-dB rms SPL threshold and
42 m to the 180-dB rms SPL threshold. As proposed here, power-down to
the single smallest array element is required when a small dolphin
enters the defined EZ, but is not allowed for any other reason (e.g.,
to avoid pre-clearance and/or ramp-up). Our rationale is that this is a
necessary corollary to the dolphin exception described previously. As
described previously, use of the acoustic source at full power may
resume following visual observation of the animal(s) exiting the
exclusion zone or 15 minutes following the last observation of the
animal. If ramp-up were required, it is likely that infill of the
missed line would be necessary, thereby reducing the benefit of the
dolphin exception.
Mitigation Source
Mitigation sources may be separate individual airguns or may be an
airgun of the smallest volume in the array, and have historically been
used when the full array is not being used (e.g., during line turns) in
order to allow ramp-up during poor visibility. The difference between
use of a single airgun in a power-down scenario and as a ``mitigation
source'' is that the power-down scenario is conditional upon the
presence of animals in the exclusion zone, whereas the mitigation
source was historically used during times when the array would
otherwise not be in use at all. The general premise is that this lower-
intensity source, if operated continuously, would be sufficiently
aversive to marine mammals to ensure that they are not within an
exclusion zone, and therefore, ramp-up may occur at times when pre-
clearance visual watch is minimally effective. There is no information
to suggest that this is an effective protective strategy, yet we are
certain that this technique involves input of extraneous sound energy
into the marine environment, even when use of the mitigation source is
limited to some maximum time period. For these reasons, we do not
believe use of the mitigation source is appropriate and propose not to
allow its use. However, as noted above, ramp-up may occur under periods
of poor visibility assuming that no acoustic or visual detections are
made during a 30-minute pre-clearance period. This is a change from how
mitigation sources have been considered in the past in that the visual
pre-clearance period was typically assumed to be highly effective
during good visibility conditions and viewed as critical to avoiding
auditory injury and, therefore, maintaining some likelihood of aversion
through use of mitigation sources during poor visibility conditions was
deemed valuable.
In light of the available information, we think it more appropriate
to acknowledge the limitations of visual observations--even under good
conditions, not all animals will be observed and cryptic species may
not be observed at all--and recognize that while visual observation is
a common sense measure it should not be determinative of when survey
effort may occur. Given the lack of proven efficacy of visual
observation in preventing auditory injury, we do not believe that its
absence should imply such potentially detrimental impacts on marine
mammals. Therefore, use of a mitigation source is not a sensible
substitute component of seismic mitigation protocols. We also believe
that consideration of mitigation sources in the past has reflected an
outdated balance, in which the possible prevention of relatively few
instances of auditory injury is outweighed by many more instances of
unnecessary behavioral disturbance of animals and degradation of
acoustic habitat.
Miscellaneous Protocols
The acoustic source must be deactivated when not acquiring data or
preparing to acquire data, except as necessary for testing. Unnecessary
use of the acoustic source should be avoided. Firing of the acoustic
source at any volume above the stated production volume would not be
authorized; the operator must provide information to the lead PSO at
regular intervals confirming the firing volume.
Testing of the acoustic source involving all elements requires
normal mitigation protocols (e.g., ramp-up). Testing limited to
individual source elements or strings does not require ramp-up but does
require pre-clearance.
We encourage the applicant companies and operators to pursue the
following objectives in designing, tuning, and operating acoustic
sources: (1) Use the minimum amount of energy necessary to achieve
operational objectives (i.e., lowest practicable source level); (2)
minimize horizontal propagation of sound energy; and (3) minimize the
amount of energy at frequencies above those necessary for the purpose
of the survey. However, we are not aware of available specific measures
by which to achieve such certifications. In fact, an expert panel
convened by BOEM to determine whether it would be feasible to develop
standards to determine a lowest practicable source level has determined
that it would not be reasonable or practicable to develop such metrics
(see Appendix L in BOEM, 2017). Minimizing production of sound at
frequencies higher than are necessary would likely require design,
testing, and use of wholly different airguns than are proposed for use
by the applicants. At minimum, notified operational capacity (not
including redundant backup airguns) must not be exceeded during the
survey, except where unavoidable for source testing and calibration
purposes. All occasions where activated source volume exceeds notified
operational capacity must be noticed to the PSO(s) on duty and fully
documented for reporting. The lead PSO must be granted access to
relevant instrumentation documenting acoustic source power and/or
operational volume. BOEM currently requires applicants for permits to
conduct geophysical surveys to submit statements indicating that
existing data are not available to meet the data needs identified for
the applicant's survey (i.e., non-duplicative survey statement) and
that the operations are using the minimal source array size/power
necessary to meet the survey goals and that the array is tuned to
maximize radiation of the emitted energy toward the seafloor.
Restriction Areas
Below we provide discussion of various restriction areas that were
considered during development of the proposed regulations. Because the
purpose of these areas is to reduce the likelihood of exposing animals
within the designated areas to noise from airgun surveys that is likely
to result in harassment (i.e., 50 percent midpoint of the Level B
harassment risk probability function), we are proposing to require that
source vessels maintain minimum standoff distances (i.e., buffers) from
the areas. Sound propagation modeling results for a notional large
airgun array were provided by Matthews et al. (2016), specific to each
of the potential time-area restrictions evaluated therein, in order to
exclude SPLs exceeding 160
[[Page 29279]]
dB rms from those areas. Those distances are proposed for use here and
are described in each section below.
Coastal Restriction--We are proposing that no airgun surveys may
occur shoreward of a line indicated by the 20-m isobath, buffered by 13
km (Matthews et al., 2016), during the months of February through May
(Area 1; Figure 5). Waters shoreward of the 20-m isobath, where coastal
dolphin stocks occur, represent the areas of greatest abundance for
bottlenose dolphins (Roberts et al., 2016).
The restriction is intended specifically to avoid additional
stressors to bottlenose dolphin populations during the time period
believed to be of greatest importance as a reproductive period. BOEM
proposed a similar coastal restriction on airgun survey effort in the
petition submitted in support of industry, and NMFS agrees that this is
appropriate. Coastal dolphin stocks, particularly the northern coastal
stock, were heavily impacted by the DWH oil spill. As described
previously, NOAA estimates that potentially 23 percent of western
coastal dolphins and 82 percent of northern coastal dolphins were
exposed to DWH oil, resulting in an array of long-term health impacts
(including reproductive failure) and possible population reductions of
5 percent and 50 percent for the western and northern stocks,
respectively (DWH MMIQT, 2015). For the northern coastal stock, it is
estimated that these population-level impacts could require 39 years to
recovery, in the absence of other additional stressors.
NMFS's subject matter experts identified a reasonable range that in
their professional judgment encompasses an important reproductive
period for bottlenose dolphins in these coastal waters. Expert
interpretation of the long-term data for neonate strandings is that
February-April are the primary months that animals are born in the
northern GOM, and that fewer but similar numbers are born in January
and May. This refers to long-term averages and in any particular year
the peak reproductive period can shift earlier or later. While pregnant
mothers may be susceptible to the impacts of noise, we believe that
neonates and/or calves are likely most susceptible, because behavioral
disruption could have more severe energetic effects for lactating
mothers and/or lead to disruption of mother-calf bonding and ultimate
effects on rates of neonate and/or calf survivorship. Therefore, we
believe that February through May represents a reasonable best estimate
of the time period of most sensitivity for bottlenose dolphins in
coastal waters.
While none of the dolphin strandings or deaths have been attributed
to airgun survey activities, stocks in the area are stressed, and
studies have shown that marine mammals react to underwater noise.
Behavioral disturbance or stress may reduce fitness for individual
animals and/or may exacerbate existing declines in reproductive health
and survivorship. For example, stressors such as noise and pollutants
can induce responses involving the neuroendocrine system, which
controls reactions to stress and regulates many body processes (NAS,
2017), and there is strong evidence that petroleum-associated chemicals
can adversely affect the endocrine system, providing a potential
pathway for interactions with other stressors (Mohr et al., 2008,
2010). Romano et al., (2004) found that upon exposure to noise from a
seismic watergun, bottlenose dolphins had significantly elevated levels
of a stress-related hormone and, correspondingly, a decrease in immune
cells. Population-level impacts related to energetic effects or other
impacts of noise are difficult to determine, but the addition of other
stressors can add considerable complexity due to the potential for
interaction between the stressors or their effects (NAS, 2017). When a
population is at risk, as is the case for these bottlenose dolphin
populations, NAS (2017) recommends identifying those stressors that may
feasibly be mitigated. We cannot undo the effects of the DWH oil spill,
but the potentially synergistic effects of noise due to the activities
that are the subject of this proposed rule may be mitigated. The post-
DWH oil spill baseline condition of these populations requires caution,
and this restriction may reasonably be anticipated to provide
additional protection to these populations during their peak
reproductive activity. Note that, in reference to the findings of
Matthews et al., (2016), this proposed time-area restriction would also
reduce impacts to stocks of marine mammals occurring within the
restriction area through reducing effects to listening area. We request
comment on our proposed seasonal closure in Area 1.
Practicability--Given survey operators' ability to plan around
these seasonal restrictions, we believe it is unlikely that the
restrictions will affect oil and gas productivity in the GOM.
Therefore, when this practicability factor is considered in light of
the expected ability of these measures to reduce the probability or
severity of impacts on species or stocks and their habitat, we
preliminarily determine these restrictions are warranted. We request
comment on our interpretation of the impact of the proposed seasonal
closure for Area 1.
[[Page 29280]]
[GRAPHIC] [TIFF OMITTED] TP22JN18.004
Bryde's Whale--We examined the appropriateness of restricting
survey effort such that particular areas of expected importance for
Bryde's whales are not ensonified by levels of sound above 160 dB rms
SPL (the 50 percent midpoint for behavioral harassment) (Area 3; Figure
5). We analyzed a year-round closure of the area described herein; we
request comment on this and several other alternatives. The variations
in regulatory text for these proposals can be found in ``Alternative
Regulatory Text,'' later in this preamble, and in the regulatory text
at the end of the document. Matthews et al. (2016) specified a buffer
distance of 5.4 km for the De Soto Canyon area, which we round to 6 km.
As described previously, NOAA's status review team determined the
status of the GOM Bryde's whale is considered to be precarious
(described in the status review technical memorandum (Rosel et al.
(2016)). On the basis of these findings, NMFS has proposed to list the
GOM Bryde's whale as an endangered species pursuant to the ESA (81 FR
88639; December 8, 2016). These whales form a small and resident
population in the northeastern GOM, with a highly restricted geographic
range and a very small population abundance--recently determined by a
status review team to be ``at or below the near-extinction population
level'' (Rosel et al., 2016). The review team stated that, aside from
the restricted distribution and small population, the whales face a
significant suite of anthropogenic threats, one of which is noise
produced by geophysical surveys.
While various population abundance estimates are available (e.g.,
Waring et al., 2016; Roberts et al., 2016; Dias and Garrison, 2016),
the population abundance was almost certainly less than 100 prior to
the DWH oil spill. NOAA estimated that, as a result of that event, 48
percent of the population may have been exposed to DWH oil, with 17
percent killed and 22 percent of females experiencing reproductive
failure. The best estimate for maximum population reduction was 22
percent, with an estimated 69 years to recovery (to the precarious
status prior to the DWH oil spill) (DWH MMIQT, 2015). It is considered
likely that Bryde's whale habitat previously extended to shelf and
slope areas of the western and central GOM similar to where they are
found now in the eastern GOM, and that anthropogenic activity--largely
energy exploration and production--concentrated in those areas could
have resulted in habitat abandonment (Reeves et al., 2011; Rosel and
Wilcox, 2014). Further, the population exhibits very low levels of
genetic diversity and significant genetic mitochondrial DNA divergence
from other Bryde's whales worldwide (Rosel and Wilcox, 2014). Based on
this review and further consultation with the Society for Marine
Mammalogy's Committee on Taxonomy, NMFS has proposed to list the GOM
Bryde's whale as an endangered species pursuant to the ESA (81 FR
88639; December 8, 2016).
The small population size, restricted range, and low genetic
diversity alone place these whales at significant risk of extinction
(IWC, 2017), which has been exacerbated by the effects of the DWH oil
spill. Additionally, Bryde's whale dive and foraging behavior places
them at heightened risk of being struck by vessels and/or entangled in
fishing gear (Soldevilla et al., 2017). It is in consideration of this
environmental baseline and risk profile that we analyzed a year-round
restriction.
LaBrecque et al. (2015) described a biologically important area for
GOM Bryde's whales as between the 100- and
[[Page 29281]]
300-m isobaths in the eastern GOM, from the head of De Soto Canyon to
an area northwest of Tampa Bay. The recorded Bryde's whale shipboard
and aerial survey sightings between 1989 and 2015 have mainly fallen
within this area (see the NOAA's status review technical memorandum
(Rosel et al. (2016)). We are proposing to expand this area for
protection of Bryde's whales following the recommendations of NOAA's
status review (described in the status review technical memorandum
(Rosel et al. (2016)), which stated that due to the depth of some
sightings, the BIA for Bryde's whales in the GOM is more appropriately
defined to the 400-m isobath and westward to Mobile Bay, Alabama, in
order to provide some buffer around the deeper sightings and to include
all sightings in the northeastern GOM. The average depth of Bryde's
whale sightings is 226 m (SE = 7.9; range 199-302 m; Maze-Foley &
Mullin 2006). Rice et al. (2014) detected sounds associated with
Bryde's whales in waters south of Panama City, FL, and there are
sightings of Bryde's whales along the shelf break to Tampa Bay (about
28.0[deg] N). Bryde's whales were also detected acoustically in this
area by Hildebrand et al. (2012). Additionally, because of past survey
design, survey effort in waters less than 200 m water depth has not
been as thorough as that for waters greater than 200 m; therefore,
Bryde's whales may use water depths between 100-200m more regularly
than we currently know. The Bryde's whale restriction is designated as
the area between the 100- and 400-m isobaths, from 87.5[deg] W to
27.5[deg] N (Area 3; Figure 5). This area largely covers the home range
(i.e., 95 percent of predicted abundance) predicted by Roberts et al.
(2016). The designated area would then be buffered by 6 km. The
restriction area would also provide benefit to any other marine mammals
present there--primarily Atlantic spotted dolphins and bottlenose
dolphins, but possibly also including other species that may occur
there in slope waters. Reporting preliminary results from a passive
acoustic monitoring study, Hildebrand et al. (2012) found a
significantly higher detection rate and a more steady presence for
delphinids at this site than at four other sites (three deep-water and
one shallow). Note that, in reference to the findings of Matthews et
al. (2016), a time-area restriction would also reduce impacts to stocks
of marine mammals occurring within the restriction area through
reducing effects to communication space and listening area.
Given the likely condition of this population, and in the absence
of a full habitat characterization and more knowledge about why Bryde's
whales occur where they do, we analyzed a year-round restriction that
covered the full area of Bryde's whale sightings. We request comment on
our interpretation of the data and our evaluated alternative of year-
round restrictions on airgun surveys in Area 3 (Figure 5). In addition,
we present three less-restrictive alternatives, including seasonal
restrictions and no restrictions for Area 3 with differing requirements
for monitoring. We request comment on all proposals and other
variations of these proposals, including our interpretation of the data
and any other data that support the necessary findings regarding time-
area restrictions for Bryde's whales.
Proposal 1: A year-round restriction on airgun surveys in
Area 3, as described above.
Proposal 2: A three-month seasonal restriction on airgun
surveys in Area 3. In addition to public comment on the proposal and
information that may support the necessary findings in consideration of
this proposal, we request information regarding the proposed duration
and/or timing of such a seasonal closure, if sufficient. We note that
this proposal is reflected in our proposed regulatory text, at the end
of this document.
Proposal 3: A three-month seasonal restriction, such as
what is described just previously, but with the addition of a
requirement for BOEM and/or members or representatives of the oil and
gas industry to ensure real-time detection of Bryde's whales across the
area of potential impact including real-time communication of
detections to survey operators. This real-time detection would be used
to initiate shutdowns to ensure that survey operations do not take
place when a Bryde's whale is within 6 km of the acoustic source. We do
not consider towed passive acoustic monitoring to be sufficient to
ensure detection of the Bryde's whale and, for the three-month
restriction, we propose use of a moored listening array. In addition to
public comment on the proposal and information that may support the
necessary findings in consideration of this proposal, as well as on the
appropriate timing and/or duration of a seasonal restriction, we
request information regarding appropriate alternative technologies for
real-time detection of Bryde's whales.
Proposal 4: No restriction, but with the addition of a
requirement for BOEM and/or members or representatives of the oil and
gas industry to ensure real-time detection of Bryde's whales across the
area of potential impact including real-time communication of
detections to survey operators. As with the previous seasonal closure
with monitoring proposal, we do not consider towed passive acoustic
monitoring to be sufficient to ensure detection of the Bryde's whale
and seek comment on appropriate technologies for real-time detection.
We request public comment on the proposal and information that may
support the necessary findings in consideration of this proposal, as
well as regarding appropriate alternative technologies for real-time
detection of Bryde's whales.
The variations in regulatory text for these proposals can be found
in ``Alternative Regulatory Text,'' later in this preamble, and in the
regulatory text at the end of the document.
Practicability--There is a moratorium on leasing pursuant to GOMESA
(through June 2022, or almost the entirety of the period of validity
for these proposed regulations). Further, BOEM has projected very low
activity levels in this area over the next 10 years (Table 1). There
are two active leases in this proposed restriction area (though no
platforms), and an exception to the year-round restriction requirements
would be made in accordance with existing rights associated with those
active leases. The RIA indicates that there is potential for effects on
oil and gas productivity given delays in the ability to conduct
exploratory surveys in advance of the end of the existing GOMESA
moratorium (if not continued) and a year-round restriction may be
warranted. As described just previously, we invite the public to
evaluate and comment on the presented alternatives.
Dry Tortugas--This proposed restriction area is expected to benefit
resident sperm and beaked whales. Beaked whales are acoustically
sensitive, with a correspondingly high magnitude of predicted
exposures, while noise from airgun surveys may have an outsize impact
on sperm whale populations due to disruption of foraging behavior (as
detailed previously). While the predicted impacts on these species are
based on projected levels of activity elsewhere in the GOM, we
acknowledge the potential importance of this area to these species and
propose the restriction to ensure that this habitat is not impacted.
Sightings of both beaked whales and sperm whales are very dense in
this area, and it is possible--based on unpublished observations of
calves here--that sperm whales use this area as a calving area (K.
Mullin, pers. comm.).
[[Page 29282]]
Hildebrand et al. (2012, 2015) conducted passive acoustic monitoring
over more than 3 years (2010-2013) at three deep-water sites on the GOM
slope, including within this area. In contrast with reported visual
observations of sperm whales in the area, preliminary results reported
by Hildebrand et al. (2012) showed relatively low rates of acoustic
detection for sperm whales, and corresponding density estimates were
lower at the Dry Tortugas site than at the other sites (i.e.,
Mississippi Canyon and Green Canyon). However, four species of beaked
whale, including an unidentified species, were detected. As reported by
Hildebrand et al. (2015), Cuvier's beaked whale was the dominant
species presence (61 percent of vocal encounters), but Gervais' beaked
whales also appear to be present in significant numbers (39 percent).
No Blainville's beaked whales were detected. Average densities for
Cuvier's and Gervais' beaked whales were derived from vocal click
counting. Combined density for the two species was very high at the Dry
Tortugas site (approximately 29 whales/1,000 km\2\). At two other sites
where beaked whales are expected to be present in significant numbers
and were detected (Mississippi Canyon and Green Canyon), the combined
density value was approximately 4 whales/1,000 km\2\, at both
locations. Both species had a strong and consistent presence throughout
the monitoring period (Hildebrand et al., 2015).
The area aligns well with a portion of the predicted 25 percent
core abundance area for beaked whales in the GOM, and overlaps with
portions of the sperm whale 25 percent core abundance area (Roberts et
al., 2016; core abundance areas are explained in greater detail below
in ``Central Planning Area''). The restriction area would also provide
benefit to any other marine mammals present there--including other
species expected to occur in deep slope waters. Hildebrand et al.
(2012) estimated the density of Kogia spp. in this area at 5.9 animals/
1,000 km\2\. The proposed year-round restriction area includes waters
bounded by the 200- to 2,000-m isobaths from the northern border of
BOEM's Howell Hook leasing area to 81.5[deg] W (Area 4; Figure 5). The
defined area would be buffered by 9 km (rounded up from the 8.4 km
distance provided by Matthews et al. (2016) for the Dry Tortugas area).
Note that, in reference to the findings of Matthews et al. (2016), this
proposed time-area restriction would also reduce impacts to stocks of
marine mammals occurring within the restriction area through reducing
effects to listening area. We invite the public to comment on our
interpretation of the data and proposal of year-round restrictions on
airgun surveys in Area 4 (Figure 5). We are interested in public
comment on this proposal, including any data that may support the
necessary findings regarding this proposal, including modifications
that could vary the length of closure from what we proposed.
Practicability--BOEM has projected no survey activity in this area
over the next 10 years. There are no active leases, and the area is
subject to the GOMESA moratorium, so we do not expect that there would
be any impact on industry operators. We seek comment on this
assumption.
Central Planning Area (CPA)--We evaluated the possibility of
implementing a restriction area in this portion of the GOM for sperm
whales and for beaked whales (Area 2; Figure 5). Sperm whales, an
endangered species, are considered to be acoustically sensitive and
potentially subject to significant disturbance of important foraging
behavior as detailed earlier in this document. Beaked whales are also
considered to be behaviorally sensitive to noise exposure and are
predicted to sustain a high magnitude of exposures to noise above
criteria for Level B harassment. A potential CPA restriction had
already been identified in BOEM (2017) on the basis of sightings data
and animal telemetry studies (for sperm whales).
Based on satellite tracking studies conducted by Jochens et al.
(2008), the home range of tagged sperm whales within the northern GOM
is broad, comprising nearly the entire GOM in waters deeper than 500 m.
Home range is defined as an area over which an animal or group of
animals regularly travels in search of food or mates that may overlap
with those of neighboring animals or groups of the same species. By
contrast, the composite core area (defined as a section of the home
range that is utilized more thoroughly and frequently as primary
locales for activities such as feeding) of GOM sperm whales generally
includes the Mississippi Canyon, Mississippi River Delta, and, to a
lesser extent, the Rio Grande Slope (Jochens et al., 2008). These data
support the fact that sperm whales aggregate in the Mississippi Canyon
area, but regularly move across the northern GOM continental slope.
Reporting preliminary data from a passive acoustic monitoring study,
Hildebrand et al. (2012) found that among three deep-water sites in the
GOM, the Mississippi Canyon area was home to the greatest density of
sperm whales.
Beaked whales are typically deep divers, foraging for mesopelagic
squid and fish, and are often found in deep water near high-relief
bathymetric features, such as slopes, canyons, and escarpments where
these prey are found (e.g., Madsen et al., 2014; MacLeod and D'Amico,
2006; Moors-Murphy, 2014). In the GOM, all reported sightings have
occurred over the continental slope or the abyss (Roberts et al.,
2015b). Movements or seasonal migrations of beaked whales are not
known, though it is likely that their distributional patterns depend on
the movement of mesoscale hydrographic features. The CPA, including
waters from the slope to 2,000 m and approximately between BOEM's
Atwater Valley and De Soto Canyon leasing areas, is believed to support
relatively high densities of sperm whales and beaked whales (K. Mullin,
pers. comm.).
In order to quantitatively evaluate this large area and produce a
more refined prospective restriction area, we considered the outputs of
habitat-based predictive density models (Roberts et al., 2016) by
creating core abundance areas, i.e., an area that contains some
percentage of predicted abundance for a given species or species group.
Please see ``Marine Mammal Density Information,'' previously in this
document, for a full description of the density models. The purpose of
a core abundance area is to represent the smallest area containing some
percentage of the predicted abundance of each species. Summing all the
cells (pixels) in the species distribution product gives the total
predicted abundance. Core area is calculated by ranking cells by their
abundance value from greatest to least, then summing cells with the
highest abundance values until the total is equal to or greater than
the specified percentage of the total predicted abundance. For example,
if a 50 percent core abundance area is produced, half of the predicted
abundance falls within the identified core area, and half occurs
outside of it.
To determine core abundance areas, we follow a three-step process:
Determine the predicted total abundance of a species/
time period by adding up all cells of the density raster (grid) for
the species/time period. For the Roberts et al. (2016) density
rasters, density is specified as the number of animals per 100 km\2\
cell.
Sort the cells of the species/time period density
raster from highest density to the lowest.
Sum and select the raster cells from highest to lowest
until a certain percentage of the total abundance is reached.
[[Page 29283]]
The selected cells represent the smallest area that represents a
given percentage of abundance. We created a range of core abundance
areas for sperm and beaked whales, and found that there was good
agreement between the outputs of the two models at a range of
approximately 15 to 20 percent core abundance for sperm whales in
concert with a 25 percent core abundance threshold for beaked whales.
On this basis, we defined a restriction area for evaluation as follows,
in two adjacent but distinct areas (which would likely be joined from
an operational perspective): (1) An area bounded by 90[deg] W and
88[deg] W (E-W) and the 500- and 1,000-fathom isobaths (N-S), and (2)
an area bounded by five sets of coordinates (Area 2, Figure 5).
Practicability--We provided a description of this area for
evaluation in the RIA associated with this rule. This analysis found
that our proposed CPA restriction area overlaid approximately 21
percent of active GOM leases (including 95 active production platforms)
and that a significant number of wells have been spudded in the CPA
restriction area in the past five years. These leases accounted for
approximately 50 and 24 percent of total GOM production of oil and gas,
respectively, from 2012-2016. A significant amount of the projected
survey activity considered herein would be conducted in the potential
CPA restriction area. Compliance costs, in terms of operational
mitigation protocols such as shutdown requirements, generally would not
be expected to reduce the level of oil and gas development in the GOM,
given that the costs of survey activities are relatively minor compared
to expenditures on drilling, engineering, installation of platforms,
and production operations. However, in contrast to the findings related
to operational mitigation protocols, area restrictions may lead to
reductions in leasing and exploration activity. The length of time
associated with the restriction is a key concern; the longer the
restriction period, the more difficult for operators to plan surveys to
comply and increasing the likelihood that some portion of planned
surveys are delayed to future years. There is no information available
in the GOM on which to base a definition of seasonality for the CPA
restriction area that we evaluated. The analysis suggests the
possibility that closing the CPA area could affect the broader
contribution of the GOM to U.S. oil and gas activity, with shifts in
effort potentially reducing domestic oil and gas production, industry
income, and employment, ultimately concluding that the economic impact
on the regional economy could be significant. Given that the evaluated
area restrictions account for an estimated 57 percent of oil reserves
and 37 percent of gas reserves, these areas account for a sizable
contribution to regional economic productivity and employment. On the
basis of this analysis, and in consideration of other mitigation
required with regard to sperm whales (i.e., expanded shutdown
requirements), we preliminarily find that implementation of this
restriction area is not warranted when the potential benefits to marine
mammals species or stocks and their habitat are weighed against the
significant costs and impracticality. We request comment on this,
preliminary determination, including our interpretation of the data,
our preliminary finding that inclusion of this measure is not warranted
due to the significant costs and impracticality, and any other data
that may support the necessary findings.
Entanglement Avoidance
We are not aware of any records of marine mammal entanglement in
towed arrays, streamers, or other towed acoustic sources. Therefore, we
do not believe there is evidence to indicate that there is any
meaningful entanglement risk posed by those activities. However, the
use of OBNs or similar equipment requiring the use of tethers or
connecting lines does pose a meaningful entanglement risk. Multiple
marine taxa are susceptible to entanglement in underwater lines and, in
2014, an Atlantic spotted dolphin was entangled in a nylon nodal tether
line and killed during a GOM OBN survey.
In order to avoid the reasonable potential for entanglement in such
lines, one must generally seek to apply common sense, including use of
stiffer lines that are taut and are not positively-buoyant, and are
therefore less likely to wrap or loop around animals, and secure bottom
lines. Specifically, we propose that operators conducting OBN surveys
adhere to the following requirements: (1) Use negatively buoyant coated
wire-core tether cable (e.g., \3/4\'' polyurethane-coated cable with
\1/2\'' wire core); (2) retrieve all lines immediately following
completion of the survey; (3) attach acoustic pingers directly to the
coated tether cable; acoustic releases should not be used; and (4)
employ a third-party PSO aboard the node retrieval vessel in order to
document any unexpected marine mammal entanglement. No unnecessary
release lines or lanyards may be used and nylon rope may not be used
for any component of the OBN system. Pingers must be attached directly
to the nodal tether cable via shackle, with cables retrieved via
grapnel. If a lanyard is required it must be as short as possible and
made as stiff as possible, e.g., by placing inside a hose sleeve.
Similar measures, including the commonly referred to ``orange coated
rope,'' have been required by BOEM as permit conditions and have proven
successful in preventing further entanglements.
Vessel Strike Avoidance
These proposed measures generally follow those described in BOEM's
PEIS (BOEM, 2017). These measures apply to all vessels associated with
any proposed survey activity (e.g., source vessels, streamer vessels,
chase vessels, supply vessels); however, we note that these
requirements do not apply in any case where compliance would create an
imminent and serious threat to a person or vessel or to the extent that
a vessel is restricted in its ability to maneuver and, because of the
restriction, cannot comply. The proposed measures include the
following:
1. Vessel operators and crews must maintain a vigilant watch for
all marine mammals and slow down or stop their vessel or alter course,
as appropriate and regardless of vessel size, to avoid striking any
marine mammal. A visual observer aboard the vessel must monitor a
vessel strike avoidance zone around the vessel, according to the
parameters stated below, to ensure the potential for strike is
minimized. Visual observers monitoring the vessel strike avoidance zone
can be either third-party observers or crew members, but crew members
responsible for these duties must be provided sufficient training to
distinguish marine mammals from other phenomena and broadly to identify
a marine mammal as a baleen whale, sperm whale, or other marine mammal.
2. All vessels, regardless of size, must observe a 10 kn speed
restriction within the EPA restriction area described previously. It is
critically important to avoid vessel strike of a Bryde's whale, as
single mortalities over time can be devastating for such small
populations. Further, Bryde's whales engage in shallow nocturnal
diving, spending significant amounts of time near the surface at night
and increasing the risk of strike when vessels are transiting Bryde's
whale habitat (Soldevilla et al., 2017).
3. Vessel speeds must also be reduced to 10 kn or less when mother/
calf pairs, pods, or large assemblages of cetaceans are observed near a
vessel. A single cetacean at the surface may indicate the presence of
submerged animals in the
[[Page 29284]]
vicinity of the vessel; therefore, precautionary measures should be
exercised when an animal is observed.
4. All vessels must maintain a minimum separation distance of 500
yards (yd) (457 m) from baleen whales. Our intention is to be
precautionary in prescribing avoidance measures to avoid the potential
for strike of Bryde's whales--the only baleen whale that would be
expected with any regularity in the GOM--but we do not expect that crew
members standing watch would be able to reliably identify baleen whales
to species in the GOM. The following avoidance measures should be taken
if a baleen whale is within 500 yd of any vessel:
a. While underway, the vessel operator should steer a course away
from the whale at 10 kn or less until the minimum separation distance
has been established.
b. If a whale is spotted in the path of a vessel or within 500 yd
of a vessel underway, the operator should reduce speed and shift
engines to neutral. The operator should re-engage engines only after
the whale has moved out of the path of the vessel and is more than 500
yd away. If the whale is still within 500 yd of the vessel, the vessel
should select a course away from the whale's course at a speed of 10 kn
or less. The recommendation to shift engines to neutral does not apply
to any vessel towing gear due to safety concerns.
c. This procedure should also be followed if a whale is spotted
while a vessel is stationary. Whenever possible, a vessel should remain
parallel to the whale's course while maintaining the 500-yd distance as
it travels, avoiding abrupt changes in direction until the whale is no
longer in the area.
5. All vessels must maintain a minimum separation distance of 100
yd (91 m) from sperm whales. The following avoidance measures should be
taken if a sperm whale is within 100 yd of any vessel:
a. The vessel underway should reduce speed and shift the engine to
neutral, and should not engage the engines until the whale has moved
outside of the vessel's path and the minimum separation distance has
been established. This does not apply to any vessel towing gear.
b. If a vessel is stationary, the vessel should not engage engines
until the whale has moved out of the vessel's path and beyond 100 yd.
6. All vessels must attempt to maintain a minimum separation
distance of 50 yd (46 m) from all other marine mammals, with an
exception made for those animals that approach the vessel. If an animal
is encountered during transit, a vessel should attempt to remain
parallel to the animal's course, avoiding excessive speed or abrupt
changes in course.
Marine Debris
Any permits issued by BOEM would include guidance for the handling
and disposal of marine trash and debris, similar to BSEE NTL 2015-G03
(``Marine Trash and Debris Awareness and Elimination'') (BSEE, 2015;
BOEM, 2017). If there were an LOA applicant for an activity not
requiring a BOEM permit, NMFS would also require adherence to this
guidance.
Table 11--Summary of Mitigation Measures With Alternatives for Consideration
----------------------------------------------------------------------------------------------------------------
Proposal preliminarily
determined to support
``least practicable Proposal included in
Measure Proposal adverse impact'' and proposed regulatory text?
``negligible impact''
findings?
----------------------------------------------------------------------------------------------------------------
Dolphin shutdown exception......... Power-down............ Yes................... Yes.
No power-down......... No.................... No.
Extended distance shutdown in Shutdown for Yes................... Yes.
certain circumstances. detections at any
distance.
Shutdown for No.................... No.
detections within 1
km.
Time-area restriction for Bryde's Year-round............ Yes................... No.
whales.
Seasonal.............. No.................... Yes.
Seasonal with real- No.................... No.
time detection.
No restriction with No.................... No.
real-time detection.
----------------------------------------------------------------------------------------------------------------
Based on our evaluation of the mitigation measures described in
this section, as well as other measures considered by NMFS, we have
preliminarily determined those mitigation measures that provide the
means of effecting the least practicable adverse impact on the affected
species or stocks and their habitat, paying particular attention to
rookeries, mating grounds, and areas of similar significance. We
request comment on all proposals and other variations of these
proposals, including our interpretation of the data and any other data
that support the necessary findings.
Proposed Monitoring and Reporting
In order to issue an LOA for an activity, Section 101(a)(5)(A) of
the MMPA states that NMFS must set forth requirements pertaining to the
monitoring and reporting of the authorized taking. NMFS's MMPA
implementing regulations further describe the information that an
applicant should provide when requesting an authorization (50 CFR
216.104(a)(13)), including the means of accomplishing the necessary
monitoring and reporting that will result in increased knowledge of the
species and the level of taking or impacts on populations of marine
mammals.
Section 101(a)(5)(A) allows that incidental taking may be
authorized only if the total of such taking contemplated over the
course of five years will have a negligible impact on affected species
or stocks (a finding based on impacts to annual rates of recruitment
and survival) and, further, section 101(a)(5)(B) requires that
authorizations issued pursuant to 101(a)(5)(A) be withdrawn or
suspended if the total taking is having, or may have, more than a
negligible impact (or such information may inform decisions on requests
for LOAs under the specific regulations). Therefore, it is clear that
the necessary requirements pertaining to monitoring and reporting must
address the total annual impacts to marine mammal species or stocks.
Effective reporting is critical both to compliance as well as ensuring
that the most value is obtained from the required monitoring.
These proposed requirements are described below under ``Data
Collection'' and ``LOA Reporting.'' Additional comprehensive reporting,
across LOA-holders on an annual basis,
[[Page 29285]]
is also proposed and is described below under ``Comprehensive
Reporting.''
More specifically, monitoring and reporting requirements should
contribute to improved understanding of one or more of the following:
Occurrence of marine mammal species in action area
(e.g., presence, abundance, distribution, density).
Nature, scope, or context of likely marine mammal
exposure to potential stressors/impacts (individual or cumulative,
acute or chronic), through better understanding of: (1) Action or
environment (e.g., source characterization, propagation, ambient
noise); (2) affected species (e.g., life history, dive patterns);
(3) co-occurrence of marine mammal species with the action; or (4)
biological or behavioral context of exposure (e.g., age, calving or
feeding areas).
Individual marine mammal responses (behavioral or
physiological) to acoustic stressors (acute, chronic, or
cumulative), other stressors, or cumulative impacts from multiple
stressors.
How anticipated responses to stressors impact either:
(1) Long-term fitness and survival of individual marine mammals; or
(2) populations, species, or stocks.
Effects on marine mammal habitat (e.g., marine mammal
prey species, acoustic habitat, or important physical components of
marine mammal habitat).
Mitigation and monitoring effectiveness.
PSO Eligibility and Qualifications
All PSO resumes must be submitted to NMFS, and PSOs must be
approved by NMFS after a review of their qualifications. NMFS expects
to maintain a list of approved PSOs, which will minimize review time
for previously approved PSOs with current experience. These
qualifications include whether the individual has successfully
completed the necessary training (see ``Training,'' below) and, if
relevant, whether the individual has the requisite experience (and is
in good standing). PSOs should provide a current resume and information
related to PSO training; submitted resumes should not include
superfluous information. Information related to PSO training should
include (1) a course information packet that includes the name and
qualifications (e.g., experience, training, or education) of the
instructor(s), the course outline or syllabus, and course reference
material; and (2) a document stating the PSO's successful completion of
the course. PSOs must be trained biologists, with the following minimum
qualifications:
A bachelor's degree from an accredited college or
university with a major in one of the natural sciences and a minimum
of 30 semester hours or equivalent in the biological sciences and at
least one undergraduate course in math or statistics;
Experience and ability to conduct field observations
and collect data according to assigned protocols (may include
academic experience; required for visual PSOs only) and experience
with data entry on computers;
Visual acuity in both eyes (correction is permissible)
sufficient for discernment of moving targets at the water's surface
with ability to estimate target size and distance; use of binoculars
may be necessary to correctly identify the target (required for
visual PSOs only);
Experience or training in the field identification of
marine mammals, including the identification of behaviors (required
for visual PSOs only);
Sufficient training, orientation, or experience with
the survey operation to ensure personal safety during observations;
Writing skills sufficient to prepare a report of
observations (e.g., description, summary, interpretation, analysis)
including but not limited to the number and species of marine
mammals observed; marine mammal behavior; and descriptions of
activity conducted and implementation of mitigation;
Ability to communicate orally, by radio or in person,
with survey personnel to provide real-time information on marine
mammals observed in the area as necessary; and
Successful completion of relevant training (described
below), including completion of all required coursework and passing
(80 percent or greater) a written and/or oral examination developed
for the training program.
The educational requirements may be waived if the PSO has acquired
the relevant skills through alternate experience. Requests for such a
waiver must include written justification, and prospective PSOs granted
waivers must satisfy training requirements described below. Alternate
experience that may be considered includes, but is not limited to, the
following:
Secondary education and/or experience comparable to PSO
duties;
Previous work experience conducting academic,
commercial, or government-sponsored marine mammal surveys; and
Previous work experience as a PSO; the PSO should
demonstrate good standing and consistently good performance of PSO
duties.
Training--NMFS expects to provide informal approval for specific
training courses in consultation with BOEM and BSEE as needed to
approve PSO staffing plans. NMFS does not propose to formally
administer any training program or to sanction any specific provider,
but will approve courses that meet the curriculum and trainer
requirements specified herein. These requirements adhere generally to
the recommendations provided by Baker et al. (2013). Those
recommendations include the following topics for training programs:
Life at sea, duties, and authorities;
Ethics, conflicts of interest, standards of conduct,
and data confidentiality;
Offshore survival and safety training;
Overview of oil and gas activities (including
geophysical data acquisition operations, theory, and principles) and
types of relevant sound source technology and equipment;
Overview of the MMPA and ESA as they relate to
protection of marine mammals;
Mitigation, monitoring, and reporting requirements as
they pertain to geophysical surveys;
Marine mammal identification, biology and behavior;
Background on underwater sound;
Visual surveying protocols, distance calculations and
determination, cues, and search methods for locating and tracking
different marine mammal species (visual PSOs only);
Optimized deployment and configuration of PAM equipment
to ensure effective detections of cetaceans for mitigation purposes
(PAM operators only);
Detection and identification of vocalizing species or
cetacean groups (PAM operators only);
Measuring distance and bearing of vocalizing cetaceans
while accounting for vessel movement (PAM operators only);
Data recording and protocols, including standard forms
and reports, determining range, distance, direction, and bearing of
marine mammals and vessels; recording GPS location coordinates,
weather conditions, Beaufort wind force and sea state, etc.;
Proficiency with relevant software tools;
Field communication/support with appropriate personnel,
and using communication devices (e.g., two-way radios, satellite
phones, internet, email, facsimile);
Reporting of violations, noncompliance, and coercion;
and
Conflict resolution.
PAM operators should regularly refresh their detection skills
through practice with simulation-modeling software, and should keep up
to date with training on the latest software/hardware advances.
Visual Monitoring
The lead PSO is responsible for establishing and maintaining clear
lines of communication with vessel crew. The vessel operator shall work
with the lead PSO to accomplish this and shall ensure any necessary
briefings are provided for vessel crew to understand mitigation
requirements and protocols. While on duty, PSOs will continually scan
the water surface in all directions around the acoustic source and
vessel for presence of marine mammals, using a combination of the naked
eye and high-quality binoculars (bigeye binoculars must be provided
during deep penetration airgun surveys; see below), from optimum
vantage points for unimpaired visual observations with minimum
distractions. PSOs will collect observational data for all marine
mammals observed, regardless of distance from the vessel, including
species, group size, presence of calves,
[[Page 29286]]
distance from vessel and direction of travel, and any observed behavior
(including an assessment of behavioral responses to survey activity).
Upon observation of marine mammal(s), a PSO will record the observation
and monitor the animal's position (including latitude/longitude of the
vessel and relative bearing and estimated distance to the animal) until
the animal dives or moves out of visual range of the observer, and a
PSO will continue to observe the area to watch for the animal to
resurface or for additional animals that may surface in the area. PSOs
will also record environmental conditions at the beginning and end of
the observation period and at the time of any observations, as well as
whenever conditions change significantly in the judgment of the PSO on
duty.
For all deep penetration airgun surveys and deep-water surveys
(i.e., water depths greater than 200 m) generally, the vessel operator
must provide bigeye binoculars (e.g., 25 x 150; 2.7 view angle;
individual ocular focus; height control) of appropriate quality (i.e.,
Fujinon or equivalent) solely for PSO use. These should be pedestal-
mounted on the deck at the most appropriate vantage point that provides
for optimal sea surface observation, PSO safety, and safe operation of
the vessel. The operator must also provide a night-vision device suited
for the marine environment for use during nighttime ramp-up pre-
clearance, at the discretion of the PSOs. NVDs may include night vision
binoculars or monocular or forward-looking infrared device (e.g.,
Exelis PVS-7 night vision goggles; Night Optics D-300 night vision
monocular; FLIR M324XP thermal imaging camera or equivalents). At
minimum, the device should feature automatic brightness and gain
control, bright light protection, infrared illumination, and optics
suited for low-light situations. This equipment is not required for
shallow penetration airgun surveys or non-airgun HRG surveys that occur
in shallow water.
Other required equipment, which should be made available to PSOs by
the third-party observer provider, includes reticle binoculars (e.g., 7
x 50) of appropriate quality (i.e., Fujinon or equivalent), GPS,
digital single-lens reflex camera of appropriate quality (i.e., Canon
or equivalent), compass, and any other tools necessary to adequately
perform the tasks described above, including accurate determination of
distance and bearing to observed marine mammals.
Individuals implementing the monitoring protocol will assess its
effectiveness using an adaptive approach. Monitoring biologists will
use their best professional judgment throughout implementation and seek
improvements to these methods when deemed appropriate. Any
modifications to protocol will be coordinated through an adaptive
management process.
Acoustic Monitoring
Use of PAM is required for deep penetration airgun surveys.
Monitoring of a towed PAM system is required at all times, from 30
minutes prior to ramp-up and throughout all use of the acoustic source.
Towed PAM systems generally consist of hardware (e.g., hydrophone
array, cables) and software (e.g., data processing and monitoring
system). Some type of automated detection software must be used; while
not required, we recommend use of industry standard software (e.g.,
PAMguard, which is open source). Hydrophone signals are processed for
output to the PAM operator with software designed to detect marine
mammal vocalizations. Current PAM technology has some limitations
(e.g., limited directional capabilities and detection range, masking of
signals due to noise from the vessel, source, and/or flow,
localization) and there are no formal guidelines currently in place
regarding specifications for hardware, software, or operator training
requirements. However, a working group (led by A.M. Thode) is
developing formal standards under the auspices of the Acoustical
Society of America's (ASA) Accredited Standards Committee on Animal
Bioacoustics (ANSI S3/SC1/WG3; ``Towed Array Passive Acoustic
Operations for Bioacoustics Applications''). While no formal standards
have yet been completed, a ``roadmap'' was developed during a 2016
workshop held for the express purpose of continuing development of such
standards. A workshop report (Thode et al., 2017) provides a highly
detailed preview of what the scope and structure of the standard would
be, including operator training, planning, hardware, real-time
operations, localization, and performance validation. NMFS expects that
LOA applicants will incorporate these considerations in developing or
refining PAM plans (described below), as appropriate. NMFS proposes to
adopt such standards in governing the development of PAM plans
following finalization.
Our requirement to use PAM refers to the use of calibrated
hydrophone arrays with full system redundancy to detect, identify and
estimate distance and bearing to vocalizing cetaceans, to the extent
possible. Multi-hydrophone (i.e., more than four) arrays are required
to allow for potential determination of bearing and range to detected
animals. With regard to calibration, the PAM system should have at
least one calibrated hydrophone, sufficient for determining whether
background noise levels on the towed PAM system are sufficiently low to
meet performance expectations. Additionally, if multiple hydrophone
types occur in a system (i.e., monitor different bandwidths), then one
hydrophone from each such type should be calibrated, and whenever sets
of hydrophones (of the same type) are sufficiently spatially separated
such that they would be expected to experience ambient noise
environments that differ by 6 dB or more across any integrated species
cluster bandwidth, then at least one hydrophone from each set should be
calibrated. The arrays should incorporate appropriate hydrophone
elements (1 Hz to 180 kHz range) and sound data acquisition card
technology for sampling relevant frequencies (i.e., to 360 kHz). This
hardware should be coupled with appropriate software to aid monitoring
and listening by a PAM operator skilled in bioacoustics analysis and
computer system specifications capable of running appropriate software.
In the absence of a formally defined set of prescriptions
addressing any of these three facets of PAM technology, all applicants
must provide a PAM plan including description of the hardware and
software proposed for use prior to proceeding with any survey where PAM
is required. As recommended by Thode et al. (2017), the plans should,
at minimum, adequately address and describe (1) the hardware and
software planned for use, including a hardware performance diagram
demonstrating that the sensitivity and dynamic range of the hardware is
appropriate for the operation; (2) deployment methodology, including
target depth/tow distance; (3) definitions of expected operational
conditions, used to summarize background noise statistics; (4) proposed
detection-classification-localization methodology, including
anticipated species clusters (using a cluster definition table), target
minimum detection range for each cluster, and the proposed localization
method for each cluster; (5) operation plans, including the background
noise sampling schedule; (6) array design considerations for noise
abatement; and (7) cluster-specific details regarding which real-time
displays and automated detectors the operator would monitor. Where
relevant, the plan should address the potential for PAM deployment on a
[[Page 29287]]
receiver vessel or other associated vessel separate from the acoustic
source.
In coordination with vessel crew, the lead PAM operator will be
responsible for deployment, retrieval, and testing and optimization of
the hydrophone array. While on duty, the PAM operator must diligently
listen to received signals and/or monitoring display screens in order
to detect vocalizing cetaceans, except as required to attend to PAM
equipment. The PAM operator must use appropriate sample analysis and
filtering techniques and, as described below, must report all cetacean
detections. While not required prior to development of formal standards
for PAM use, we recommend that vessel self-noise assessments are
undertaken during mobilization in order to optimize PAM array
configuration according to the specific noise characteristics of the
vessel and equipment involved, and to refine expectations for distance/
bearing estimations for cetacean species during the survey. Copies of
any vessel self-noise assessment reports must be included with the
summary trip report.
Data Collection
PSOs must use standardized data forms, whether hard copy or
electronic. PSOs will record detailed information about any
implementation of mitigation requirements, including the distance of
animals to the acoustic source and description of specific actions that
ensued, the behavior of the animal(s), any observed changes in behavior
before and after implementation of mitigation, and if shutdown was
implemented, the length of time before any subsequent ramp-up of the
acoustic source to resume survey. If required mitigation was not
implemented, PSOs should submit a description of the circumstances. We
require that, at a minimum, the following information be reported:
Vessel names (source vessel and other vessels
associated with survey) and call signs;
PSO names and affiliations;
Dates of departures and returns to port with port name;
Dates and times (Greenwich Mean Time) of survey effort
and times corresponding with PSO effort;
Vessel location (latitude/longitude) when survey effort
begins and ends; vessel location at beginning and end of visual PSO
duty shifts;
Vessel heading and speed at beginning and end of visual
PSO duty shifts and upon any line change;
Environmental conditions while on visual survey (at
beginning and end of PSO shift and whenever conditions change
significantly), including wind speed and direction, Beaufort sea
state, Beaufort wind force, swell height, weather conditions, cloud
cover, sun glare, and overall visibility to the horizon;
Factors that may be contributing to impaired
observations during each PSO shift change or as needed as
environmental conditions change (e.g., vessel traffic, equipment
malfunctions);
Survey activity information, such as acoustic source
power output while in operation, number and volume of airguns
operating in the array, tow depth of the array, and any other notes
of significance (i.e., pre-ramp-up survey, ramp-up, shutdown,
testing, shooting, ramp-up completion, end of operations, streamers,
etc.) (if the survey is a non-airgun survey, information relevant to
the acoustic source used should be provided);
If a marine mammal is sighted, the following
information should be recorded:
[cir] Watch status (sighting made by PSO on/off effort,
opportunistic, crew, alternate vessel/platform);
[cir] PSO who sighted the animal;
[cir] Time of sighting;
[cir] Vessel location at time of sighting;
[cir] Water depth;
[cir] Direction of vessel's travel (compass direction);
[cir] Direction of animal's travel relative to the vessel;
[cir] Pace of the animal;
[cir] Estimated distance to the animal and its heading relative
to vessel at initial sighting;
[cir] Identification of the animal (e.g., genus/species, lowest
possible taxonomic level, or unidentified); also note the
composition of the group if there is a mix of species;
[cir] Estimated number of animals (high/low/best);
[cir] Estimated number of animals by cohort (adults, yearlings,
juveniles, calves, group composition, etc.);
[cir] Description (as many distinguishing features as possible
of each individual seen, including length, shape, color, pattern,
scars or markings, shape and size of dorsal fin, shape of head, and
blow characteristics);
[cir] Detailed behavior observations (e.g., number of blows,
number of surfaces, breaching, spyhopping, diving, feeding,
traveling; as explicit and detailed as possible; note any observed
changes in behavior);
[cir] Animal's closest point of approach (CPA) and/or closest
distance from the acoustic source;
[cir] Platform activity at time of sighting (e.g., deploying,
recovering, testing, shooting, data acquisition, other); and
[cir] Description of any actions implemented in response to the
sighting (e.g., delays, shutdown, ramp-up, speed or course
alteration, etc.); time and location of the action should also be
recorded; and
If a marine mammal is detected while using the PAM
system, the following information should be recorded:
[cir] An acoustic encounter identification number, and whether
the detection was linked with a visual sighting;
[cir] Time when first and last heard;
[cir] Types and nature of sounds heard (e.g., clicks, whistles,
creaks, burst pulses, continuous, sporadic, strength of signal,
etc.); and
[cir] Any additional information recorded such as water depth of
the hydrophone array, bearing of the animal to the vessel (if
determinable), species or taxonomic group (if determinable),
spectrogram screenshot, and any other notable information.
LOA Reporting
PSO effort, survey details, and sightings data should be recorded
continuously during surveys and reports prepared each day during which
survey effort is conducted. These reports would include amount and
location of line-kms surveyed, all marine mammal observations with
closest approach distance, and corrected numbers of marine mammals
``taken.'' We propose submission of such reports to NMFS within 90 days
of survey completion or following expiration of an issued LOA. In the
event that an LOA is issued for a period exceeding one year, annual
reports would be submitted during the period of validity.
There are multiple reasons why marine mammals may be present and
yet be undetected by observers. Animals are missed because they are
underwater (availability bias) or because they are available to be
seen, but are missed by observers (perception and detection biases)
(e.g., Marsh and Sinclair, 1989). Negative bias on perception or
detection of an available animal may result from environmental
conditions, limitations inherent to the observation platform, or
observer ability. In this case, we do not have prior knowledge of any
potential negative bias on detection probability due to observation
platform or observer ability. Therefore, observational data corrections
must be made with respect to assumed species-specific detection
probability as evaluated through consideration of environmental factors
(e.g., f(0)). In order to make these corrections, we propose a method
recommended by the Marine Mammal Commission for estimating the number
of cetaceans in the vicinity of geophysical surveys based on the number
of groups detected.
This method incorporates f(0) and BSS-specific g(0) values from
Barlow (2015) that were derived using Distance sampling methods
(Buckland et al., 2001) and sightings data. If we know that we have
detected n groups, and the probability of detecting each group is p, a
standard way to estimate the total number of groups is n/p. We know n
for each species from the data collected during each survey, so the
problem is to find p for each species. During scientific marine mammal
surveys, p is estimated from the data collected on each survey as part
of a line-transect analysis. The probability p for each species depends
[[Page 29288]]
principally on the distance of the animals from the observer, but may
also depend on other factors such as group size and sea state.
In the absence of a line-transect analysis, the Commission suggests
taking estimates of p from other studies which use ships of similar
size and searching methods. For line-transect analysis, p is a product
of the probability of detecting a group of animals directly on the
trackline (g(0)) and the probability of detecting a group of animals
within the half-strip width on each side of the trackline (m/w, where w
is the transect truncation distance beyond which data are not recorded
and m is the effective strip half-width). The effective strip half-
width also may be expressed as m = 1/f(0), where f(0) is the estimated
probability density function of observed perpendicular distances y
evaluated at y = 0.
The species discussed in Barlow (2015) may be different from those
observed during a geophysical survey, but data from similar species can
be used. Since g(0) and f(0) values for each species or genera depend
on group size, BSS, swell height and other factors, those factors
should be taken into account if possible.
The probability of detecting a group of cetaceans can therefore be
expressed as:
[GRAPHIC] [TIFF OMITTED] TP22JN18.005
If there are n sightings of a species along a section of trackline,
the estimated number of Groups for a given BSS, within a perpendicular
distance w on each side of the trackline, and within the Level B
harassment zone is:
[GRAPHIC] [TIFF OMITTED] TP22JN18.006
and the estimated number of individual animals in that given BSS then
is:
[GRAPHIC] [TIFF OMITTED] TP22JN18.007
where S is the mean group size for the species.
The number of animals seen within each BSS should be summed for
each Level B harassment zone. That total number then must be scaled by
the distance to the Level B harassment threshold relative to the
truncation distance to estimate the total number of animals potentially
taken during a given survey. Examples of the application of this
process are given in the Commission's letter, relevant portions of
which are available online at: www.fisheries.noaa.gov/national/marine-mammal-protection/incidental-take-authorizations-oil-and-gas.
As noted, a draft report must be submitted to NMFS within 90 days
of the completion of survey effort or following expiration of the LOA
(whichever comes first), or annually (if a multi-year LOA is issued),
and must include all information described above under ``Data
Collection.'' The report will describe the operations conducted and
sightings of marine mammals near the operations. The report will
provide full documentation of methods, results, and interpretation
pertaining to all monitoring. The report will summarize the dates and
locations of survey operations, and all marine mammal sightings (dates,
times, locations, activities, associated survey activities);
information regarding locations where the acoustic source was used must
be provided. The LOA-holder shall provide geo-referenced time-stamped
vessel tracklines for all time periods in which airguns (full array or
single) were operating. Tracklines should include points recording any
change in airgun status (e.g., when the airguns began operating, when
they were turned off, or when they changed from full array to single
gun or vice versa). GIS files shall be provided in ESRI shapefile
format and include the UTC date and time, latitude in decimal degrees,
and longitude in decimal degrees. All coordinates should be referenced
to the WGS84 geographic coordinate system. In addition to the report,
all raw observational data shall be made available to NMFS. This report
must also include a validation document concerning the use of PAM (if
PAM was required), which should include necessary noise validation
diagrams and demonstrate whether background noise levels on the PAM
deployment limited achievement of the planned detection goals.
The report will also include estimates of the number of takes based
on the observations and in consideration of the detectability of the
marine mammal species observed (as described above). Applicants must
provide an estimate of the number (by species) of marine mammals that
may have been exposed (based on observational data and accounting for
animals present but unavailable for sighting) to the survey activity
within areas associated with the relevant frequency-weighted sound
fields (i.e., 140/160/180 dB rms). The draft report must be accompanied
by a certification from the lead PSO as to the accuracy of the report.
A final report must be submitted within 30 days following resolution of
any comments on the draft report.
Comprehensive Reporting
Individual LOA-holders will be responsible for collecting and
submitting monitoring data to NMFS, as described above. In addition, on
an annual basis, LOA holders will also collectively be responsible for
compilation and analysis of those data for inclusion in subsequent
annual synthesis reports. Individual LOA-holders may collaborate to
produce this report or may elect to have their trade associations
support the production of such a report. These reports would summarize
the data presented in the individual LOA-holder reports, provide
analysis of these synthesized results, discuss the implementation of
required mitigation, and present any recommendations. This
comprehensive annual report would be the basis of an annual adaptive
management process (described below in ``Adaptive Management''). The
following topics should be described in comprehensive reporting:
Summary of geophysical survey activity by survey type,
geographic zone (i.e., the seven zones described in the modeling
report), month, and acoustic source status (e.g., inactive, ramp-up,
full-power, power-down);
Summary of monitoring effort (on-effort hours and/or
distance) by acoustic source status, location, and visibility
conditions (for both visual and acoustic monitoring);
Summary of mitigation measures implemented (e.g.,
delayed ramp-ups, shutdowns, course alterations for vessel strike
avoidance) by survey type and location;
Sighting rates of marine mammals during periods with
and without acoustic source activities and other variables that
could affect detectability of marine mammals, such as:
[cir] Initial sighting distances of marine mammals relative to
source status;
[cir] Closest point of approach of marine mammals relative to
source status;
[cir] Observed behaviors and types of movements of marine
mammals relative to source status;
[cir] Distribution/presence of marine mammals around the survey
vessel relative to source status;
[cir] Analysis of the effects of various factors influencing the
detectability of marine mammals (e.g., wind speed, sea state, swell
height, presence of glare or fog); and
[cir] Estimates of the number of marine mammals taken by
harassment, corrected for animals potentially missed by observers;
Summary and conclusions from monitoring in previous
year; and
Recommendations for adaptive management.
Each annual comprehensive report should cover one full year of
monitoring effort and must be submitted for review by October 1 of each
year. Therefore, to allow for adequate preparation, each
[[Page 29289]]
report should analyze survey and monitoring effort described in reports
submitted by individual LOA-holders from July 1 of one year through
June 30 of the next. Of necessity, the first annual report may cover a
different period of time, e.g., from the date of issuance of a rule
until October 1 of the next year.
Reporting Injured or Dead Marine Mammals
In the event that the specified activity clearly causes the take of
a marine mammal in a manner not permitted by the authorization (if
issued), such as a serious injury or mortality, the LOA-holder shall
immediately cease the specified activities and immediately report the
take to NMFS. The report must include the following information:
Time, date, and location (latitude/longitude) of the
incident;
Name and type of vessel involved;
Vessel's speed during and leading up to the incident;
Description of the incident;
Status of all sound source use in the 24 hours
preceding the incident;
Water depth;
Environmental conditions (e.g., wind speed and
direction, Beaufort sea state, cloud cover, and visibility);
Description of all marine mammal observations in the 24
hours preceding the incident;
Species identification or description of the animal(s)
involved;
Fate of the animal(s); and
Photographs or video footage of the animal(s) (if
equipment is available).
The LOA-holder shall not resume its activities until NMFS is able
to review the circumstances of the prohibited take. NMFS would work
with the LOA-holder to determine what is necessary to minimize the
likelihood of further prohibited take and ensure MMPA compliance. The
LOA-holder may not resume their activities until notified by NMFS.
In the event that the LOA-holder discovers an injured or dead
marine mammal, and the lead PSO determines that the cause of the injury
or death is unknown and the death is relatively recent (i.e., in less
than a moderate state of decomposition as we describe in the next
paragraph), the LOA-holder will immediately report the incident to
NMFS. The report must include the same information identified in the
paragraph above this section. Activities may continue while NMFS
reviews the circumstances of the incident. NMFS would work with the
LOA-holder to determine whether modifications to the activities are
appropriate.
In the event that the LOA-holder discovers an injured or dead
marine mammal, and the lead PSO determines that the injury or death is
not associated with or related to the specified activities (e.g.,
previously wounded animal, carcass with moderate to advanced
decomposition, or scavenger damage), the LOA-holder would report the
incident to NMFS within 24 hours of the discovery. The LOA-holder would
provide photographs or video footage (if available) or other
documentation of the animal to NMFS.
Negligible Impact Analysis and Preliminary Determination
NMFS has defined negligible impact as an impact resulting from the
specified activity that cannot be reasonably expected to, and is not
reasonably likely to, adversely affect the species or stock through
effects on annual rates of recruitment or survival (50 CFR 216.103). A
negligible impact finding is based on the lack of likely adverse
effects on annual rates of recruitment or survival (i.e., population-
level effects). An estimate of the number of takes alone is not enough
information on which to base an impact determination. In addition to
considering estimates of the number of marine mammals that might be
``taken,'' NMFS considers other factors, such as the type of take
(e.g., mortality, injury), the likely nature of any responses (e.g.,
intensity, duration), the context of any responses (e.g., critical
reproductive time or location, migration), as well as effects on
habitat, and the likely effectiveness of the mitigation. We also assess
the number, intensity, and context of estimated takes by evaluating
this information relative to population status. Consistent with the
1989 preamble for NMFS's implementing regulations (54 FR 40338;
September 29, 1989), the impacts from other past and ongoing
anthropogenic activities are incorporated into this analysis via their
impacts on the environmental baseline (e.g., as reflected in the
regulatory status of the species, population size and growth rate where
known, ongoing sources of human-caused mortality, or ambient noise
levels).
For each potential activity-related stressor, we consider the
potential impacts on affected marine mammals and the likely
significance of those impacts to the affected stock or population as a
whole. Potential risk due to vessel collision and related mitigation
measures as well as potential risk due to entanglement and contaminant
spills were addressed under ``Proposed Mitigation'' and ``Potential
Effects of the Specified Activity on Marine Mammals'' and are not
discussed further, as there are minimal risks expected from these
potential stressors.
The ``specified activity'' for these regulations is a broad program
of geophysical survey activity that could occur at any time of year in
U.S. waters of the GOM. In recognition of the broad scale of this
activity in terms of geographic and temporal scales, we propose use of
a new analytical framework--first described by Ellison et al. (2015)--
through which an explicit, systematic risk assessment methodology is
applied to evaluate potential effects of aggregated discrete acoustic
exposure events (i.e., proposed geophysical survey activities) on
marine mammals. We believe the approach described here addresses the
scope and scale of potential impacts to marine mammal populations from
these activities. Development of the approach was supported
collaboratively by BOEM and NMFS, which together provided guidance to
an expert working group (EWG) in terms of application to relevant
regulatory processes. The framework and preliminary results are
described by Southall et al. (2017), which is available online at:
www.fisheries.noaa.gov/national/marine-mammal-protection/incidental-take-authorizations-oil-and-gas. That document is a companion to this
analysis, and is referred to hereafter as the ``EWG report.'' The risk
assessment framework described below was developed and preliminarily
implemented by Southall et al. (2017) in relation to the specified
activity described herein; we incorporate the framework and its results
into our analysis as appropriate.
As described previously, Zeddies et al. (2015, 2017a) provided
marine mammal noise exposure estimates based on BOEM-provided
projections of future survey effort and based on best available
modeling of sound propagation, animal distribution, and animal
movement. This provided a conservative but reasonable best estimate of
potential acute noise exposure events that may result from the
described suite of activities. The primary goal in this new analytical
effort was to develop a systematic framework that would use those
modeling results to put into biologically-relevant context the level of
potential risk of injury and/or disturbance to marine mammals. The
framework considers both the aggregation of acute effects as well as
the broad temporal and spatial scales over which chronic effects may
occur. Previously, Wood et al. (2012) conducted an analysis of a
proposed airgun survey, in which they derived a qualitative risk
assessment method of
[[Page 29290]]
considering the biological significance of exposures predicted to be
consistent with the onset of physical injury and behavioral disturbance
(the latter determined according to the same approach used here).
Subsequently, Ellison et al. (2015) described development of a more
systematic and (in some cases) quantitative basis for a risk-assessment
approach to assess the biological significance and potential population
consequences of predicted noise exposures. The approach here, which
incorporated the results of Zeddies et al. (2015, 2017a) as an input,
includes certain modifications to and departures from the conceptual
approach described by Ellison et al. (2015). These are described in
greater detail in the EWG report.
Generally, this approach is a relativistic risk assessment that
provides an interpretation of the exposure estimates within the context
of key biological and population parameters (e.g., population size,
life history factors, compensatory ability of the species, animal
behavioral state, aversion), as well as other biological,
environmental, and anthropogenic factors. The analysis is performed
specifically on a species-specific basis for each effort scenario
(``high,'' ``moderate,'' and ``low'') within each modeling zone (Figure
2). The end result provides an indication of the biological
significance of these exposure numbers for each affected marine mammal
stock (i.e., yielding the severity of impact and vulnerability of
stock/population information), as well as forecasting the likelihood of
any such impact. This result is expressed as relative impact ratings of
overall risk that couple potential severity of effect on a stock and
likely vulnerability of the population to the consequences of those
effects, given biologically relevant information (e.g., compensatory
ability).
Spectral, temporal, and spatial overlaps between survey activities
and animal distribution are the primary factors that drive the type,
magnitude, and severity of potential effects on marine mammals, and
these considerations are integrated into both the severity and
vulnerability assessments. In discussion with BOEM and NMFS, the EWG
developed a strategic approach to balance the weight of these
considerations between the two assessments, specifying and clarifying
where and how the interactions between potential disturbance and
species within these dimensions are evaluated. Overall ratings are then
considered in conjunction with our proposed mitigation strategy (and
any additional relevant contextual information) to ultimately inform
our preliminary determinations. Elements of this approach are
subjective and relative within the context of this program of projected
actions and, overall, the analysis necessarily requires the application
of professional judgment.
Severity of Effect
Level A Harassment--In order to evaluate the potential severity of
the expected potential takes by Level A harassment (Table 9) on the
species or stock, the EWG report uses a PBR-equivalent metric. As
described previously, PBR is defined by the MMPA as the maximum number
of animals, not including natural mortalities, that may be removed from
a marine mammal stock while allowing that stock to reach or maintain
its optimum sustainable population. To be clear, NMFS does not expect
any of the potential occurrences of injury (i.e., PTS) that may be
authorized under this rule to result in mortality of marine mammals,
nor do we believe that Level A harassment should be considered a
``removal'' in the context of PBR when used to inform a negligible
impact determination. PTS is not appropriately considered equivalent to
serious injury. However, PBR can serve as a gross indicator of the
status of the species and a good surrogate for population
vulnerability/health and, accordingly, PBR or a related metric can be
used appropriately to inform a separate analysis to evaluate the
potential relative severity to the population of a permanent impact
such as PTS on a given number of individuals. This analysis is used to
assess relative risks to populations as a result of PTS; NMFS does not
expect that Level A harassment could directly result in mortality and
our use of the PBR metric in this context should not be interpreted as
such.
However, because habitat-based density models (Roberts et al.,
2016) were used to predict cetacean distribution and abundance in the
GOM, exposure estimates cannot appropriately be directly related to the
PBR values found in NMFS's SARs. Therefore, a modified PBR value was
derived on the basis of the typical pattern for NMFS's PBR values,
where the value varies between approximately 0.6-0.9 percent of the
minimum population abundance depending upon population confidence
limits (higher with increasing confidence). For endangered species, PBR
values are typically \1/5\ of the values for non-endangered species due
to assumption of a lower recovery factor--endangered species are
typically assigned recovery factors of 0.1, while species of unknown
status relative to the optimum sustainable population level (i.e., most
species) are typically assigned factors of 0.5. This basic relationship
of population size relative to PBR (e.g., considered equivalent to
estimated X percent of PBR) was used to define the following relative
risk levels due to Level A harassment.
Very high--Level A takes greater than 1.5 or 0.3
percent (the latter figure is used for endangered species) of zone-
specific estimated population abundance.
High--0.75-1.5 or 0.15-0.3 percent of zone-specific
population.
Moderate--0.375-0.75 or 0.075-0.15 percent of zone-
specific population.
Low--0.075-0.375 or 0.015-0.075 percent of zone-
specific population.
Very low--less than 0.075 or less than 0.015 percent of
zone-specific population.
Relative severity scores by zone (Figure 2) and species for high,
moderate, and low annual activity scenarios are shown in Tables 4-7 of
the EWG report. However, as described previously, we do not believe
that Level A harassment is likely to actually occur for mid-frequency
cetaceans and therefore do not predict any take by Level A harassment
for these species. The risk presented by Level A harassment to mid-
frequency species is therefore expected to be none to very low.
Due to the combination of density estimates and effort projections,
the predicted takes by Level A harassment (accounting for aversion) for
both Bryde's whale and Kogia spp. are expected to represent a ``very
high'' risk for the moderate and low effort scenarios in Zone 4 (note
that the ``high'' effort scenario, while including the most survey days
when aggregating across the entire GOM, includes no projected survey
days in Zone 4). For Kogia spp. only, all three effort scenarios
represent a ``very high'' risk in Zones 6 and 7. All other combinations
of effort and zone result in overall evaluated risk of none to low for
these species. We note that regardless of the relative risk assessed in
this framework, because of the anticipated received levels and duration
of sound exposure expected for any marine mammals exposed above Level A
harassment criteria, no individuals of any species or stock are
expected to receive more than a relatively minor degree of PTS, which
would not be expected to meaningfully increase the likelihood or
severity of any potential population-level effects.
Level B Harassment--As described above in ``Estimated Take,'' a
significant model assumption was that populations of animals were reset
for each 24-hr period. Exposure estimates for the 24-hr period were
then aggregated across all
[[Page 29291]]
assumed survey days as completely independent events, assuming
populations turn over completely within each large zone on a daily
basis. While the modeling provides reasonable estimates of the total
number of instances of exposure exceeding Level B harassment criteria,
it is likely that it leads to substantial overestimates of the numbers
of individuals potentially disturbed, given that all animals within the
areas modeled are unlikely to be completely replaced on a daily basis.
Therefore, in assuming an increased number of individuals impacted,
these results would lead to an overestimation of the potential
population-level consequences of the estimated exposures. In order to
evaluate modeled daily exposures and determine more realistic exposure
probabilities for individuals across multiple days, we use information
on species-typical movement behavior to determine a species-typical
offset of modeled daily exposures, using the exploratory analysis
discussed under ``Estimated Take'' (i.e., Test Scenario 1). In this
test scenario, modeled results were compared for a 30-day period versus
the aggregation of 24-hr population reset intervals. When conducting
computationally-intensive modeling over the full assumed 30-day survey
period (versus aggregating the smaller 24-hr periods for 30 days),
results showed about 10-45 percent of the total number of takes
calculated using a 24-hr reset of the population, with differences
relating to species-typical movement and residency patterns. Given that
many of the evaluated survey activities occur for 30-day or longer
periods, particularly some of the larger surveys for which the majority
of the modeled exposures occur, using such a scaling process is
appropriate in order to evaluate the likely severity of the predicted
exposures. However, as noted earlier, even with this correction factor
the resulting number of predicted takes of individuals is still an
overestimate because individuals are expected to be exposed to multiple
surveys in a year and many surveys are longer than 30 days. This
approach is also discussed in more detail in the EWG report.
The test scenario modeled six representative GOM species/guilds:
Bryde's whale, sperm whale, beaked whales, bottlenose dolphin, Kogia
spp., and short-finned pilot whale. For purposes of this analysis,
bottlenose dolphin was used as a proxy for other small dolphin species,
and short-finned pilot whale was used as a proxy for other large
delphinids. Tables 22-23 in the modeling report provide information
regarding the number of modeled animals receiving exposure above
criteria for average 24-hr sliding windows scaled to the full 30-day
duration and percent change in comparison to the same number evaluated
when modeling the full 30-day duration. This information was used to
derive 30-day scalar ratios which, when applied to the total instances
of exposure given in Table 9, captures repeated takes of individuals at
a 30-day sampling level. Scalar ratios are as follows: Bryde's whale,
0.189; sperm whale, 0.423; beaked whales, 0.101; bottlenose dolphin,
0.287; Kogia spp., 0.321; and short-finned pilot whale, 0.295.
Application of the re-scaling method reduced the overall magnitude of
modeled takes for all species by slightly more than double to up to
ten-fold. This output was used in a severity assessment.
Table 12--Scenario-Specific Expected Take Numbers, Instances and Individuals \1\
--------------------------------------------------------------------------------------------------------------------------------------------------------
Survey effort scenario \2\
-------------------------------------------------------------------------------------------------------------
Species High Moderate #1 Moderate #2 Low #1 Low #2
-------------------------------------------------------------------------------------------------------------
Ins. Ind. Ins. Ind. Ins. Ind. Ins. Ind. Ins. Ind.
--------------------------------------------------------------------------------------------------------------------------------------------------------
Bryde's whale............................. 560 106 413 78 498 94 386 73 402 76
Sperm whale............................... 43,504 18,395 27,271 11,531 33,340 14,097 26,651 11,269 27,657 11,694
Kogia spp................................. 16,189 5,189 11,428 3,663 13,644 4,373 10,743 3,443 11,165 3,579
Beaked whale.............................. 235,615 23,704 162,134 16,311 190,777 19,193 151,708 15,262 156,584 15,753
Rough-toothed dolphin..................... 37,666 10,793 30,192 8,651 31,103 8,912 28,663 8,213 26,315 7,540
Bottlenose dolphin........................ 653,405 187,222 977,108 279,974 596,824 171,010 938,322 268,860 579,403 166,018
Clymene dolphin........................... 110,742 31,731 72,913 20,892 87,615 25,105 69,609 19,945 72,741 20,843
Atlantic spotted dolphin.................. 133,427 38,231 174,705 50,059 116,698 33,438 164,824 47,228 109,857 31,478
Pantropical spotted dolphin............... 606,729 173,848 419,738 120,269 511,037 146,429 399,581 114,493 419,824 120,293
Spinner dolphin........................... 82,779 23,719 59,623 17,084 73,013 20,921 56,546 16,202 59,253 16,978
Striped dolphin........................... 44,038 12,618 29,936 8,578 36,267 10,392 28,522 8,172 29,890 8,564
Fraser's dolphin.......................... 13,858 3,971 9,654 2,766 11,394 3,265 9,127 2,615 9,391 2,691
Risso's dolphin........................... 27,062 7,754 18,124 5,193 21,914 6,279 17,309 4,960 18,092 5,184
Melon-headed whale........................ 68,900 20,355 47,548 14,047 56,791 16,777 44,842 13,247 46,631 13,776
Pygmy killer whale........................ 18,029 5,326 12,278 3,627 14,788 4,369 11,677 3,450 12,141 3,587
False killer whale........................ 25,511 7,536 17,631 5,209 20,828 6,153 16,774 4,955 17,163 5,070
Killer whale.............................. 1,493 441 1,031 305 1,258 372 984 291 1,036 306
Short-finned pilot whale.................. 19,258 5,689 12,155 3,591 14,163 4,184 11,523 3,404 11,900 3,516
--------------------------------------------------------------------------------------------------------------------------------------------------------
\1\ Instances of take (``Ins.'') reflects expected scenario-based takes by Level B harassment given previously in Table 9. Scalar ratios were applied as
described in preceding text to derive expected numbers of individuals taken (``Ind.'').
\2\ High survey effort scenario correspond level of effort projections given previously for Year 1 (Table 1). Moderate #1 and #2 and Low #1 and #2
correspond with Years 4, 5, 8, and 9, respectively.
As was done in evaluating severity of Level A harassment, the
scaled Level B harassment takes were rated through a population-
dependent binning system. For each species, scaled takes were divided
by the zone-specific predicted abundance, and these proportions were
used to evaluate the relative severity of modeled exposures based on
the distribution of values across species to evaluate behavioral risk
across species--a simple, logical means of evaluating relative risk
across species and areas. Relative risk ratings using percent of area
population size were defined as follows:
Very high--Adjusted behavioral takes greater than 800
percent of zone-specific population;
High--Adjusted behavioral takes 400-800 percent of
zone-specific population;
Moderate--Adjusted behavioral takes 200-400 percent of
zone-specific population;
Low--Adjusted behavioral takes 100-200 percent of zone-
specific population; and
Very low--Adjusted behavioral takes less than 100
percent of zone-specific population.
Results of severity ranking for Level B harassment are shown in
Tables 8-10 of Southall et al. (2017). Note that these have been
adjusted here to account for
[[Page 29292]]
the erroneous density value that underlies the exposure predictions
given by Zeddies et al. (2015, 2017b) for Bryde's whales in Zone 6.
Vulnerability of Affected Population
Vulnerability rating seeks to evaluate the relative risk of a
predicted effect given species-typical and population-specific
parameters (e.g., species-specific life history, population factors)
and other relevant interacting factors (e.g., human or other
environmental stressors). The assessment includes consideration of four
categories within two overarching risk factors (species-specific
biological and environmental risk factors). These values were selected
to capture key aspects of the importance of spatial (geographic),
spectral (frequency content of noise in relation to species-typical
hearing and sound communications), and temporal relationships between
sound and receivers. Explicit numerical criteria for identifying
severity scores were specified where possible, but in some cases
qualitative judgments based on a reasonable interpretation of given
aspects of the proposed activity and how it relates to the species in
question and the environment within the specified area were required.
Factors considered in the vulnerability assessment were detailed in
Southall et al. (2017) and are reproduced here (Table 13); note that
the effects of the DWH oil spill are accounted for through the non-
noise chronic anthropogenic risk factor identified below, while the
effects to acoustic habitat and on individual animal behavior via
masking described in ``Potential Effects of the Specified Activity on
Marine Mammals and Their Habitat'' are accounted for through the
masking chronic anthropogenic noise risk factors. Species-specific
vulnerability scoring according to this scheme is shown in Table 14.
Based on the range in vulnerability assessment scoring, an overall
vulnerability rating was selected from the zone- and species-specific
aggregate vulnerability score as shown in Table 15.
Table 13--Vulnerability Assessment Factors
------------------------------------------------------------------------
Score
------------------------------------------------------------------------
Masking: Degree of spectral overlap between biologically
important acoustic signals and predominant noise source
of proposed activity (max: 7 out of 30):
Communication masking: Predominant noise energy +3/+1
directly/partially overlaps \1\ species-specific
signals utilized for communication.................
Foraging masking: Predominant noise energy directly/ +2/+1
partially overlaps \1\ species-specific signals
utilized in foraging (including echolocation and
other foraging coordination signals)...............
Navigation/Orientation signal masking: Predominant +2/+1
noise energy directly/partially overlaps \1\
signals likely utilized in spatial orientation to
which species is well capable of hearing...........
Species population: Stock status, trend, and size (max:
7 out of 30):
Population status: Endangered (ESA) and/or depleted +3/0
(MMPA) (Y/N).......................................
Trend rating: Decreasing/unknown or data deficient/ +2/+1/0/-1
stable (i.e., within 5 percent)/increasing (last
three SARs for which new population estimates were
updated)...........................................
Population size: Small (less than 2,500)............ +2
Species habitat use and compensatory abilities: Degree
to which activity within a specified area \2\ overlaps
with species habitat and distribution (max: 7 out of
30):
Habitat use: Survey area contains greater than 30/15- +4/+2/+1/0
30/5-15/less than 5 percent of total region-wide
estimated..........................................
population (during defined survey period)...........
Temporal sensitivity: Survey overlaps temporally Up to +3
with well-defined species-specific biologically-
important period (e.g., calving)...................
Other (chronic) noise and non-noise stressors: Magnitude
of other potential sources of disturbance or other
stressors that may influence a species response to
additional noise and disturbance of the proposed
activity (max: 9 out of 30):
Chronic anthropogenic noise: Species subject to high/ +2/+1
moderate degree of current or known future
(overlapping activity) chronic anthropogenic noise.
Chronic anthropogenic risk factors (non-noise): Up to +4/+2
Species subject to high/moderate degree of current
or known future risk from other chronic, non-noise
anthropogenic activities (e.g., fisheries
interactions, ship strike).........................
Chronic biological risk factors (non-noise): Known Up to +3
presence of disease, parasites, prey limitation, or
high predation pressure............................
------------------------------------------------------------------------
\1\ Direct or partial overlap means that the predominant spectral
content of received noise exposure from activity specific sources is
expected to occur at identical frequencies as signals of interest, or
that secondary (lower-level) spectral content of received noise
exposure from activity specific sources is expected to occur at
identical frequencies as signals of interest.
\2\ This is the area over which a specified activity is evaluated and a
local population is determined, in this case the seven modeling zones.
BILLING CODE 3510-22-P
[[Page 29293]]
[GRAPHIC] [TIFF OMITTED] TP22JN18.008
BILLING CODE 3510-22-C
Table 15--Vulnerability Rating Scheme
------------------------------------------------------------------------
Risk
probability
Total score (% of Vulnerability rating
total)
------------------------------------------------------------------------
24-30.............................. 80-100 Very high
18-23.............................. 60-79 High
12-17.............................. 40-59 Moderate
6-11............................... 20-39 Low
0-5................................ 0-19 Very low
------------------------------------------------------------------------
Risk
In the final step of the framework, severity and vulnerability
ratings are integrated to provide relative impact ratings of overall
risk. The likely severity of effect was assessed as the percentage of
total population affected based on scaled modeled Level B harassment
takes relative to zone population size. There is no risk when there is
no survey activity in a given zone for a given effort scenario, and
zones predicted to contain abundance of less of five or less
individuals of a species were also considered to have de minimis risk.
Severity and vulnerability assessments each produce a numerical rating
(1-5) corresponding with the qualitative rating (i.e., very low, low,
moderate, high, very high). A matrix is then used to integrate these
two scores to provide an overall risk assessment. The matrix is shown
in Table 2 of Southall et al. (2017). Please see Tables 8-10 of the EWG
report for species- and zone-specific severity and vulnerability
ratings for each of three activity scenarios. Tables 16-17 provide
relative impact ratings by zone, and Table 18 provides GOM-wide
relative impact ratings, for overall risk associated with predicted
takes by Level B harassment, for each of three activity scenarios.
[[Page 29294]]
Table 16--Overall Evaluated Risk by Zone and Activity Scenario
[Zones 1-4]
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Zone 1 \1\ Zone 2 Zone 3 Zone 4 \1\
Species -----------------------------------------------------------------------------------------------------------------------------------------------------------------
High High Moderate Low High Moderate Low Moderate Low
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Bryde's whale................. Low............. n/a............. n/a............. n/a............. n/a............. n/a............. n/a............. Moderate........ Moderate.
Sperm whale................... n/a............. n/a............. n/a............. n/a............. n/a............. n/a............. n/a............. Moderate........ Low.
Kogia spp..................... Low............. n/a............. n/a............. n/a............. n/a............. n/a............. n/a............. Low............. Low.
Beaked whale.................. n/a............. n/a............. n/a............. n/a............. n/a............. n/a............. n/a............. High............ Low.
Rough-toothed dolphin......... Low............. Moderate........ High............ High............ Very low........ Very low........ Very low........ Low............. Very low.
Bottlenose dolphin............ Low............. Low............. High............ Moderate........ Very low........ Very low........ Very low........ Very low........ Very low.
Clymene dolphin............... n/a............. n/a............. n/a............. n/a............. n/a............. n/a............. n/a............. Moderate........ Low.
Atlantic spotted dolphin...... Low............. Moderate........ High............ High............ Very low........ Very low........ Very low........ Very low........ Very low.
Pantropical spotted dolphin... Low............. n/a............. n/a............. n/a............. n/a............. n/a............. n/a............. Very low........ Very low.
Spinner dolphin............... Very low........ n/a............. n/a............. n/a............. n/a............. n/a............. n/a............. Very low........ Very low.
Striped dolphin............... n/a............. n/a............. n/a............. n/a............. n/a............. n/a............. n/a............. Low............. Very low.
Fraser's dolphin.............. Low............. Low............. High............ Moderate........ n/a............. n/a............. n/a............. Low............. Very low.
Risso's dolphin............... Low............. n/a............. n/a............. n/a............. n/a............. n/a............. n/a............. Very low........ Very low.
Melon-headed whale............ n/a............. n/a............. n/a............. n/a............. n/a............. n/a............. n/a............. Moderate........ Moderate.
Pygmy killer whale............ n/a............. n/a............. n/a............. n/a............. n/a............. n/a............. n/a............. Low............. Very low.
False killer whale............ Low............. Low............. Moderate........ Moderate........ Very low........ Very low........ Very low........ Very low........ Very low.
Killer whale.................. n/a............. n/a............. n/a............. n/a............. n/a............. n/a............. n/a............. Low............. Very low.
Short-finned pilot whale...... n/a............. n/a............. n/a............. n/a............. n/a............. n/a............. n/a............. Low............. Very low.
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
n/a = no activity projected for zone or five or less individuals predicted in zone.
\1\ No activity is projected in Zone 1 under the moderate and low activity scenarios, and no activity is projected in Zone 4 under the high activity scenario.
Table 17--Overall Evaluated Risk by Zone and Activity Scenario
[Zones 5-7]
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Zone 5 Zone 6 Zone 7
Species -----------------------------------------------------------------------------------------------------------------------------------------------------------------
High Moderate Low High Moderate Low High Moderate Low
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Bryde's whale................. Very high....... Very high....... Very high....... n/a............. n/a............. n/a............. n/a............. n/a............. n/a.
Sperm whale................... Very high....... Very high....... Very high....... Very high....... Very high....... High............ Moderate........ Moderate........ Moderate.
Kogia spp..................... High............ High............ Moderate........ Moderate........ Moderate........ Low............. Moderate........ Low............. Low.
Beaked whale.................. Very high....... Very high....... Very high....... High............ Moderate........ Moderate........ High............ High............ High.
Rough-toothed dolphin......... High............ High............ Moderate........ Moderate........ Low............. Low............. Low............. Low............. Low.
Bottlenose dolphin............ High............ High............ Moderate........ Low............. Very low........ Very low........ Low............. Very low........ Very low.
Clymene dolphin............... High............ High............ Moderate........ Moderate........ Low............. Low............. Low............. Low............. Low.
Atlantic spotted dolphin...... High............ High............ High............ Moderate........ Low............. Low............. n/a............. n/a............. n/a.
Pantropical spotted dolphin... High............ High............ Moderate........ Moderate........ Low............. Low............. Low............. Low............. Low.
Spinner dolphin............... High............ High............ Moderate........ Low............. Very low........ Very low........ Low............. Very low........ Very low.
Striped dolphin............... High............ High............ Moderate........ Moderate........ Low............. Low............. Low............. Low............. Low.
Fraser's dolphin.............. High............ High............ Moderate........ Moderate........ Low............. Low............. Low............. Low............. Low.
Risso's dolphin............... High............ High............ High............ Low............. Very low........ Very low........ Very low........ Very low........ Very low.
Melon-headed whale............ High............ High............ Moderate........ Moderate........ Low............. Low............. Moderate........ Low............. Low.
Pygmy killer whale............ High............ High............ Moderate........ Moderate........ Low............. Low............. Low............. Low............. Low.
False killer whale............ High............ High............ Moderate........ Low............. Very low........ Very low........ Low............. Low............. Low.
Killer whale.................. High............ High............ High............ Moderate........ Low............. Low............. Low............. Low............. Low.
Short-finned pilot whale...... High............ High............ Moderate........ Moderate........ Low............. Moderate........ Moderate........ Low............. Low.
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
n/a = no activity projected for zone or five or less individuals predicted in zone.
Table 18--Overall Evaluated Risk by Activity Scenario, GOM-Wide
----------------------------------------------------------------------------------------------------------------
Moderate activity
Species High activity scenario scenario Low activity scenario
----------------------------------------------------------------------------------------------------------------
Bryde's whale........................ Moderate............... Moderate............... Moderate.
Sperm whale.......................... Very high.............. High................... High.
Kogia spp............................ Moderate............... Low.................... Low.
Beaked whale......................... Very high.............. High................... High.
Rough-toothed dolphin................ Moderate............... Low.................... Low.
Bottlenose dolphin................... Low.................... Moderate............... Low.
Clymene dolphin...................... Moderate............... Low.................... Low.
Atlantic spotted dolphin............. Low.................... Low.................... Low.
Pantropical spotted dolphin.......... Moderate............... Low.................... Low.
Spinner dolphin...................... Low.................... Low.................... Low.
Striped dolphin...................... Moderate............... Low.................... Low.
Fraser's dolphin..................... Moderate............... Low.................... Low.
Risso's dolphin...................... Moderate............... Low.................... Low.
Melon-headed whale................... Moderate............... Moderate............... Moderate.
Pygmy killer whale................... Moderate............... Low.................... Low.
False killer whale................... Moderate............... Low.................... Low.
Killer whale......................... Moderate............... Low.................... Low.
Short-finned pilot whale............. Moderate............... Low.................... Low.
----------------------------------------------------------------------------------------------------------------
[[Page 29295]]
Overall, the results of the risk assessment show that (as
expected), risk is highly correlated with effort and density. Areas
where little or no survey activity is predicted to occur or areas
within which few or no animals of a particular species are believed to
occur have very low or no potential risk of negatively affecting marine
mammals, as seen across activity scenarios in Zones 1, 3, and 4. Areas
with consistently high levels of effort (Zones 2, 5, 6, and 7) are
generally predicted to have higher overall evaluated risk across all
species. However, fewer species of animals are expected to be present
in Zone 2, where we primarily expect shelf species such as bottlenose
and Atlantic spotted dolphins. In Zone 7, animals are expected to be
subject to less other chronic noise and non-noise stressors, which is
reflected in the vulnerability scoring for that zone. Therefore,
despite consistently high levels of projected effort, overall rankings
for that zone are lower than for Zones 5 and 6.
Zones 5 and 6 were the only zones with ``very high'' levels of risk
due to behavioral disturbance, identified for three species of
particular concern in Zone 5 (Bryde's, beaked, and sperm whales) and
two in Zone 6 (beaked and sperm whales). Projected effort levels were
sufficiently high in Zone 5 that the rankings were not generally
sensitive to activity scenario, while in Zone 6 the highest rankings
were associated with the high activity scenario. As particularly
sensitive species, beaked whales and sperm whales consistently receive
relatively high severity scores. Bryde's whales receive very high
vulnerability scoring across zones, due in large part to the
differential susceptibility to masking, while sperm whales were also
typically ranked as being highly vulnerable. Relatively high levels of
risk were also identified for other species in some contexts, and these
are generally explained by the interaction of specific factors related
to survey effort concentration and areas of heightened geographic
distribution or specific factors related to population trends or zone-
related differences in vulnerability. When considered across the entire
GOM and all activity scenarios, the only species considered to have
relatively high risk are the sperm whale and beaked whales, while the
Bryde's whale and melon-headed whales have relatively moderate risk.
Although the scores generated by the EWG framework, and further
aggregated across zones as described by NMFS above, are species-
specific, additional stock-specific information can be gleaned through
the zone-specific nature of the analysis in that, for example with
bottlenose dolphins, the zones align with stock range edges. These
species-specific risk scores are broadly applied in NMFS's negligible
impact analysis to all of the multiple stocks that are analyzed in this
rule (Table 3), however, NMFS is also considering additional stock-
specific information in our analysis, where appropriate, as indicated
in our ``Description of Marine Mammals in the Area of the Specified
Activity,'' ``Potential Effects of the Specified Activity on Marine
Mammals and Their Habitat,'' and ``Proposed Mitigation'' sections
(e.g., coastal bottlenose dolphins were heavily impacted by the DWH oil
spill and we have therefore recommended a time/area restriction to
reduce impacts).
In order to more fully place the predicted amount of take into
meaningful context, it is useful to understand the duration of exposure
at or above a given level of received sound, as well as the likely
number of repeated exposures across days. While a momentary exposure
above the criteria for Level B harassment counts as an instance of
take, that accounting does not make any distinction between fleeting
exposures and more severe encounters in which an animal may be exposed
to that received level of sound for a longer period of time. However,
this information is meaningful to an understanding of the likely
severity of the exposure, which is relevant to the negligible impact
evaluation, and is not directly incorporated into the risk assessment
framework described above. For example, for bottlenose dolphin exposed
to noise from 3D WAZ surveys in Zone 6, the modeling report shows that
approximately 72 takes (Level B harassment) would be expected to occur
in a 24-hr period. However, each animat modeled has a record or time
history of received levels of sound over the course of the modeled 24-
hr period. The 50th percentile of the cumulative distribution function
indicates that the time spent exposed to levels of sound above 160 dB
rms SPL (i.e., the 50 percent midpoint for behavioral harassment) would
be only 1.8 minutes--a minimal amount of exposure carrying little
potential for significant disruption of behavioral activity. We provide
summary information regarding the total time in a 24-hr period that an
animal would spend in this received level condition in Table 19.
Additionally, as we discussed in the ``Estimated Take'' section for
Test Scenario 1, by comparing exposure estimates generated by
multiplying 24-hr exposure estimates by the total number of survey days
versus modeling for a full 30-day survey duration for six
representative species, we were able to refine the exposure estimates
to better reflect the number of individuals exposed above threshold.
Using this same comparison and scalar ratios described above, we are
able to predict an average number of days each of the representative
species modeled in the test scenario were exposed above the Level B
harassment thresholds. As with the duration of exposures discussed
above, the number of repeated exposures is important to our
understanding of the severity of effects. Specifically, for example,
the ratio for beaked whales indicates that the 30-day modeling showed
that approximately 10 percent as many individual beaked whales could be
expected to be exposed above harassment thresholds as was reflected in
the results given by multiplying average 24-hr exposure results by the
survey duration (i.e., 30 days). However, the approach of scaling up
the 24-hour exposure estimates appropriately reflects the instances of
exposure above threshold (which cannot be more than 1 in 24 hours), so
the inverse of the scalar ratio suggests the average number of days in
the 30-day modeling period that beaked whales are exposed above
threshold is approximately ten. It is important to remember that this
is an average and that it is likely some individuals would be exposed
on fewer days and some on more. Table 19 reflects the average days
exposed above threshold for the indicated species having applied the
scalar ratios described previously.
[[Page 29296]]
Table 19--Time in Minutes (Per Day) Spent Above 160 dB rms SPL (50th Percentile) and Average Number of Days
Individuals Exposed Above Threshold During 30-Day Survey
----------------------------------------------------------------------------------------------------------------
Survey type and time (min/day) above 160 dB rms Average number
---------------------------------------------------------------- of days
exposed above
threshold
Species during 30-day
2D 3D NAZ 3D WAZ Coil survey
---------------
5.3
----------------------------------------------------------------------------------------------------------------
Bryde's whale................... 5.1 11.8 4.6 19.5 2.4
Sperm whale..................... 4.7 9.5 4.0 17.2 3.1
Kogia spp....................... 3.3 8.0 3.0 16.3 9.9
Beaked whale.................... 4.8 10.1 4.0 20.3 3.5
Rough-toothed dolphin........... 3.6 7.8 3.1 14.2 3.5
Bottlenose dolphin.............. 3.3 8.4 2.9 15.1 3.5
Clymene dolphin................. 3.2 7.9 2.9 13.7 3.5
Atlantic spotted dolphin........ 5.5 12.8 5.0 23.6 3.5
Pantropical spotted dolphin..... 3.2 7.9 2.9 13.7 3.5
Spinner dolphin................. 3.2 7.9 2.9 13.7 3.5
Striped dolphin................. 3.2 7.9 2.9 13.7 3.5
Fraser's dolphin................ 3.3 8.0 3.0 16.3 3.5
Risso's dolphin................. 4.5 10.9 3.9 18.6 3.5
Melon-headed whale.............. 3.3 8.0 3.0 16.3 3.1
Pygmy killer whale.............. 3.6 7.7 3.1 14.2 3.1
False killer whale.............. 3.6 7.7 3.1 14.2 3.1
Killer whale.................... 9.3 23.3 8.0 35.4 3.1
Short-finned pilot whale........ 3.3 8.0 3.0 14.7 3.1
----------------------------------------------------------------------------------------------------------------
We expect that Level A harassment could occur for low-frequency
species (i.e., Bryde's whale)--due to these species' heightened
sensitivity to frequencies in the range output by airguns, as shown by
their auditory weighting function--and for high-frequency species, due
to their heightened sensitivity to noise in general (as shown by their
lower threshold for the onset of PTS) (NMFS, 2016). However, to the
extent that Level A harassment occurs it will be in the form of PTS,
and the degree of injury is expected to be mild. If hearing impairment
occurs, it is most likely that the affected animal would lose a few dB
in its hearing sensitivity, which in most cases is not likely to affect
its ability to survive and reproduce. Hearing impairment that occurs
for these individual animals would be limited to at and slightly above
the dominant frequency of the noise sources, i.e., in the low-frequency
region below 2-4 kHz. Therefore, the degree of PTS is not likely to
affect the echolocation performance of the Kogia spp., which use
frequencies between 60-120 kHz (Wartzok and Ketten, 1999). Further,
modeled exceedance of Level A harassment criteria typically resulted
from being near an individual source once rather than accumulating
energy from multiple sources. Overall, the modeling indicated that
exceeding the SEL threshold is a rare event and having four vessels
close to each other (350 m between tracks) did not cause appreciable
accumulation of energy at the ranges relevant for injury exposures.
Accumulation of energy from independent surveys is expected to be
negligible. For Kogia spp., because of expected sensitivity, we expect
that aversion may play a stronger role in avoiding exposures above the
peak pressure threshold than we have accounted for. For these reasons,
and in conjunction with our proposed mitigation plan, we do not believe
that Level A harassment will play a meaningful role in the overall
degree of impact experienced by marine mammal populations as a result
of the projected survey activity.
We consider the relative impact ratings described above in
conjunction with our proposed mitigation and other relevant contextual
information in order to produce a final assessment of impact to the
stock or species, i.e., our preliminary negligible impact
determination. Annual levels of human-caused mortality are less than
PBR for all GOM stocks aside from the Bryde's whale and, for most
species, are zero (Hayes et al., 2017). The effects of the DWH oil
spill, which is not reflected in NMFS's published values for annual
human-caused mortality, are accounted for through our vulnerability
scoring (Table 14). We developed mitigation requirements, including
time-area restrictions, designed specifically to provide benefit to
certain populations for which we predict a relatively high amount of
risk in relation to exposure to survey noise. The proposed time-area
restrictions, described in detail in ``Proposed Mitigation'' and
depicted in Figure 5, are designed specifically to provide benefit to
the bottlenose dolphin, Bryde's whale, and beaked and sperm whales,
with additional benefits to Kogia spp., which are often found in higher
densities in the same locations of greater abundance for beaked and
sperm whales. In addition, we expect these areas to provide some
subsidiary benefit to additional species that may be present. The
Atlantic spotted dolphin would also benefit from the coastal
restriction proposed for bottlenose dolphins, and multiple shelf-break
associated species would benefit from both the Bryde's whale and Dry
Tortugas restrictions. The output of the Roberts et al. (2016) models,
as used in core abundance area analyses (described in detail in
``Proposed Mitigation''), provides information about species most
likely to derive subsidiary benefit from the proposed restrictions.
Notably, high densities of Kogia spp. are predicted in the area of the
Dry Tortugas restriction. Other shelf-break/pelagic species that are
abundant in the eastern GOM include the melon-headed whale, Risso's
dolphin, and rough-toothed dolphin, but numerous other species would be
expected to be present in varying numbers at various times.
These proposed measures benefit both the primary species for which
they were designed and the species that may benefit secondarily by
likely reducing
[[Page 29297]]
the number of individuals exposed to survey noise and, for resident
species in areas where seasonal restrictions are proposed, reducing the
numbers of times that individuals are exposed to survey noise. However,
and perhaps of greater importance, we expect that these restrictions
will reduce disturbance of these species in the places most important
to them for critical behaviors such as foraging and socialization. The
Bryde's whale area is the only known habitat of the species in the GOM,
while the Dry Tortugas area is assumed to be an area important for
beaked whale foraging and sperm whale reproduction. The coastal
restriction would provide protection for the bottlenose dolphin
populations most severely impacted by the DWH oil spill during a time
of importance for reproduction. Further detail regarding rationale for
these restrictions is provided under ``Proposed Mitigation.''
The endangered sperm whale and the Bryde's whale received special
consideration in our development of proposed mitigation. The
alternative of a year-round closure alternative with a 6-km buffer is
designed to avoid impacts to the Bryde's whale by completely avoiding
known habitat. Survey activities must avoid all areas where the Bryde's
whale is found, and we propose to require shutdown of the acoustic
source upon observation of any Bryde's whale at any distance. The
Bryde's whale is proposed for listing as endangered, has a very low
population size, is more sensitive to the low frequencies output by
airguns, and faces significant additional stressors. Therefore,
regardless of impact rating, we believe that the year-round closure
alternative and 6-km buffer described previously would allow us to make
the necessary negligible impact finding. We preliminarily find, were
this alternative finalized, that the total potential marine mammal take
from the projected survey activities will have a negligible impact on
the Bryde's whale.
While the economic analysis accompanying this proposed rule
indicates that a CPA restriction benefiting sperm whales would not be
practicable, we propose to require a shutdown of the acoustic source
upon any acoustic detection of sperm whales. We also propose shutdown
requirements upon any detection of beaked whales or Kogia spp.
(although these two species are rarely detected visually). If the
observed animal is within the behavioral harassment zone, it would
still be considered to have experienced harassment, but by immediately
shutting down the acoustic source the duration and degree of disruption
is minimized and the significance of the harassment event reduced as
much as possible. Therefore, in consideration of the proposed
mitigation, we preliminarily find that the total potential marine
mammal take from the projected survey activities will have a negligible
impact on the sperm whale, beaked whales, and Kogia spp.
The risk assessment process rates impacts as moderate or less for
all other affected species. Therefore, in consideration of the proposed
mitigation, we preliminarily find that the total potential marine
mammal take from the projected survey activities will have a negligible
impact on all other affected species, including all affected stocks of
bottlenose dolphin.
In summary and as described above, the following factors primarily
support our preliminary determination that the impacts resulting from
this activity are not expected to adversely affect the affected species
or stocks through effects on annual rates of recruitment or survival:
No mortality is anticipated or authorized;
Level A harassment not expected for species other than
Bryde's whale and Kogia spp., and not expected to be a meaningful
source of harm for these species;
Risk assessment process rates impacts as moderate or
less, for most species in most places and higher risk species have
associated mitigation to lessen impacts;
Known habitat for Bryde's whales protected;
Shutdown requirements for species of concern (Bryde's
whale, sperm whale, beaked whales, Kogia spp.); and
Modeling resulted in daily exposures totaling 3-35
minutes, which, in most situations, is likely insufficient time to
result in disruptions of behavior that raise concerns about fitness
consequences.
Based on the analysis contained herein of the likely effects of the
specified activity on marine mammals and their habitat, and taking into
consideration the implementation of the proposed monitoring and
mitigation measures, with a year-round closure in Bryde's whale habitat
(Area 3; Figure 5), we preliminarily find that the total marine mammal
take from the proposed activity will have a negligible impact on all
affected marine mammal species or stocks.
Small Numbers
What are small numbers?
The MMPA does not define ``small numbers.'' NMFS's and the U.S.
Fish and Wildlife Service's joint 1989 implementing regulations defined
small numbers as a portion of a marine mammal species or stock whose
taking would have a negligible impact on that species or stock. This
definition was invalidated in Natural Resources Defense Council v.
Evans, 279 F.Supp.2d 1129 (2003) (N.D. Cal. 2003), based on the court's
determination that the regulatory definition of small numbers was
improperly conflated with the regulatory definition of ``negligible
impact,'' which rendered the small numbers standard superfluous. As the
court observed, ``the plain language indicates that small numbers is a
separate requirement from negligible impact.'' Since that time, NMFS
has not applied the definition found in its regulations. Rather,
consistent with Congress' pronouncement that small numbers is not a
concept that can be expressed in absolute terms (House Committee on
Merchant Marine and Fisheries Report No. 97-228 (September 16, 1981)),
NMFS now makes its small numbers findings based on an analysis of
whether the number of individuals taken annually from a specified
activity is small relative to the stock or population size. The Ninth
Circuit has upheld a similar approach. See Center for Biological
Diversity v. Salazar, No. 10-35123, 2012 WL 3570667 (9th Cir. Aug. 21,
2012). However, we have not previously indicated what we believe the
upper limit of small numbers is. Here, we provide additional
information and clarification regarding our consideration of small
numbers pursuant to paragraphs (A) and (D) of section 101(a)(5) of the
MMPA.
To maintain an interpretation of small numbers as a proportion of a
species or stock that does not conflate with negligible impact, we
propose the following framework. A plain reading of ``small'' implies
as corollary that there also could be ``medium'' or ``large'' numbers
of animals from the species or stock taken. We therefore propose a
simple approach that establishes three equal bins corresponding to
small, medium, and large numbers of animals: Small is comprised of 1-33
percent, medium 34-66 percent, and large 67-100 percent of the
population abundance.
NMFS's practice for making small numbers determinations is to
compare the number of individuals estimated to be taken against the
best available abundance estimate for that species or stock. Although
NMFS's implementing regulations require applications for incidental
take to include an estimate of the marine mammals to be taken, there is
nothing in paragraphs (A) or (D) of section 101(a)(5) that requires
NMFS to quantify or estimate numbers of marine mammals to be taken for
purposes of evaluating whether the number is small.
[[Page 29298]]
While it can be challenging to predict the numbers of individual marine
mammals that will be taken by an activity (many models calculate
instances of take and are unable to account for repeated exposures of
individuals), in some cases we are able to generate a reasonable
estimate utilizing a combination of quantitative tools and qualitative
information. When it is possible to predict with relative confidence
the number of individual marine mammals of each species or stock that
are likely to be taken, we recommend the small numbers determination be
based directly upon whether or not these estimates exceed one third of
the stock abundance. In other words, as in past practice, when the
estimated number of animals is up to, but generally not greater than,
one third of the species or stock abundance, NMFS will determine that
the numbers of marine mammals of a species or stock are small.
When sufficient quantitative information is not available to
estimate the number of individuals that might be taken (typically due
to insufficient information about presence, density, or daily or
seasonal movement patterns of the species in an area), we consider
other factors, such as the spatial scale of the specified activity
footprint as compared with the range of the affected species or stock
and/or the duration of the activity in order to infer the relative
proportion of the affected species or stock that might reasonably be
expected to be taken by the activity. For example, an activity that is
limited to a small spatial scale (e.g., a coastal construction project
or HRG survey) and relatively short duration might not be expected to
result in take of more than a small number of a comparatively wider-
ranging species. Unlike direct quantitative modeling of a number of
individuals taken, this comparison may necessitate the presentation of
some additional information and logical inferences to make a small
numbers determination.
Another circumstance in which NMFS considers it appropriate to make
a small numbers finding in the absence of a quantitative estimate is in
the case of a species or stock that may potentially be taken but is
either rarely encountered or only expected to be taken on rare
occasions. In that circumstance, one or two assumed encounters with a
group of animals (meaning a group that is traveling together or
aggregated, and thus exposed to a stressor at the same approximate
time) could reasonably be considered small numbers, regardless of
consideration of the proportion of the stock (if known), as rare brief
encounters resulting in take of one or two groups should be considered
small relative to the range and distribution of any stock.
In summary, when quantitative take estimates of individual marine
mammals are available or inferable through consideration of additional
factors, and the number of animals taken is one third or less of the
best available abundance estimate for the species or stock, NMFS would
consider it to be of small numbers. When quantitative take estimates
are not available, NMFS will examine other factors, such as the spatial
extent of the take zone compared to the species or stock range and/or
the duration of the activity to determine if the take will likely be
small relative to the abundance of the affected species or stocks.
Last, NMFS may appropriately find that one or two predicted group
encounters will result in small numbers of take relative to the range
and distribution of a species, regardless of the estimated proportion
of the abundance.
How is the small numbers standard evaluated within the structure of the
section 101(a)(5)(A) process?
Neither the MMPA nor NMFS's implementing regulations address
whether the small numbers determination should be based upon the total
annual taking for all activities occurring under incidental take
regulations or to individual LOAs issued thereunder. The MMPA does not
define small numbers or explain how to apply the term in either
paragraph (A) or (D) of section 101(a)(5), including how to apply the
term in a way that allows for consistency between those two very
similar provisions. NMFS has not previously made a clear and deliberate
policy choice or specifically explored applying the small numbers
finding to each individual LOA under regulations that cover multiple
concurrent LOA holders. Here we propose a reasonable interpretation of
how to make a small numbers determination based on a permissible
interpretation of the statute.
Specifically, section 101(a)(5)(A)(i)(I) explicitly states that the
negligible impact determination for a specified activity must take into
account the total taking over the five-year period, but the small
numbers language is not tied explicitly to the same language. As the
provision is structured, the small numbers language is not framed as a
standard for the issuance of the authorization, but rather appears in
the chapeau as a limitation on what the Secretary may allow. The
regulatory vehicle for authorizing (i.e., allowing) the take of marine
mammals is the LOA.
Given NMFS's discretion in light of the ambiguities in the statute
regarding how to apply the small numbers standard, and the clear
benefits of application as described here, we have determined that the
small numbers finding should be applied to the annual take authorized
in each LOA. To demonstrate why this approach is preferred, we first
describe below why it is beneficial to NMFS, the public, and the
resource (marine mammals) to utilize section 101(a)(5)(A) for multiple
activities, where possible.
From a resource protection standpoint, it is more
protective to conduct a comprehensive negligible impact analysis
that considers all of the activities covered under the rule and
ensures that the total combined taking from those activities will
have a negligible impact on the affected marine mammal species or
stocks and no unmitigable adverse impact on subsistence uses.
Furthermore, mitigation and monitoring are more effective when
considered across all activity and years covered under regulations.
From an agency resource standpoint, it ultimately will
save significant time and effort to cover multi-year activities
under a rule instead of multiple incidental harassment
authorizations (IHAs). While regulations require more analysis up
front, additional public comment and internal review, and additional
time to promulgate compared to a single IHA, they are effective for
up to five years and can cover multiple actors within a year. The
process of issuing individual LOAs under incidental take regulations
utilizes the analysis, public comment, and review that was conducted
for the regulations, and takes significantly less time than it takes
to issue an IHA.
From an applicant standpoint, incidental take
regulations offer more regulatory certainty than IHAs (five years
versus one year) and significant cost savings, both in time and
environmental compliance analysis and documentation, especially for
situations like here, where multiple applicants will be applying for
individual LOAs under regulations. In the case of this proposed
rule, the certainty afforded by the promulgation of a regulatory
framework (e.g., by using previously established take estimates,
mitigation and monitoring requirements, and procedures for
requesting and obtaining an LOA) is a significant benefit for
prospective applicants.
A review of IHAs we have issued suggests that bundling together two
or three IHAs that might be ideal subjects for a combined incidental
take regulation (e.g., for ongoing maintenance construction activities,
or seismic surveys in the Arctic) would very often result in greater
than small numbers of one or more species being taken if we were to
apply the small numbers standard across all activity contemplated by
the regulation in a
[[Page 29299]]
year, thereby precluding the use of section 101(a)(5)(A) in many cases.
Application of the small numbers standard across the total annual
taking covered by regulations, inasmuch as potential applicants can see
that the total take may exceed one third of species or stock abundance,
creates an incentive for applicants to pursue individual IHAs, and will
often preclude the ability to gain the benefits of regulations outlined
above.
Our conclusion is that NMFS can appropriately elect to make a
``small numbers'' finding based on the estimated annual take in
individual LOAs issued under the rule. This approach does not affect
the negligible impact analysis, which is the biologically relevant
inquiry and based on the total annual estimated taking for all
activities the regulations will govern. Making the small numbers
finding based on the estimated annual take in individual LOAs allows
NMFS to take advantage of the associated administrative and
environmental benefits of utilizing section 101(a)(5)(A) that would be
precluded in many cases if small numbers were required to be applied to
the total annual taking under the regulations.
Although this application of small numbers may be argued as being
less protective of marine mammals, NMFS disagrees. As specifically
differentiated from the negligible impact finding, the small numbers
standard has little biological relevance. The negligible impact
determination, which does have biological significance, is still
controlling, and the total annual taking authorized across all LOAs
under an incidental take regulation still could not exceed the overall
amount analyzed for the negligible impact determination. Moreover, to
the extent that this process is perceived as less protective than
applying the small numbers standard across all activity occurring
annually under the regulations (in that the small numbers standard can
be met more readily under our proposed approach), that perception
ignores the fact that applicants could always opt to pursue an IHA to
circumvent a more restrictive approach to applying small numbers under
section 101(A)(5)(A) (in cases where there is no serious injury or
mortality).
How will small numbers be evaluated under this proposed GOM rule?
In this proposed rule, up-to-date species information is available,
and sophisticated models have been used to estimate take in a manner
that will allow for quantitative comparison of the take of individuals
versus the best available abundance estimates for the species or
stocks. Specifically, while the modeling effort utilized in the rule
enumerates the estimated instances of takes that will occur across days
as the result of the operation of certain survey types in certain
areas, the modeling report also includes the evaluation of a test
scenario that allows for a reasonable modification of those generalized
take estimates to better estimate the number of individuals that will
be taken within one survey. LOA applicants using modeling results from
the rule to inform their applications will be able to reasonably
estimate the number of marine mammal individuals taken by their
proposed activities. LOA applications that do not use the modeling
provided in the rule to estimate take for their activities will need to
be independently reviewed, and applicants will be required to ensure
that their estimates adequately inform the small numbers finding.
Additionally, if applicants use the modeling provided by this rule to
estimate take, additional public input will not be deemed necessary
(unless other conditions necessitating public review exist, as
described in the ``Letters of Authorization'' section); if they do not,
however, NMFS will publish a notice in the Federal Register soliciting
public comment. The estimated take of marine mammals for each species
will then be compared against the best available scientific information
on species or stock abundance estimate as determined by NMFS, and
estimates that do not exceed one-third of that estimate will be
considered small numbers.
Adaptive Management
The regulations governing the take of marine mammals incidental to
geophysical survey activities would contain an adaptive management
component. The comprehensive reporting requirements associated with
this proposed rule (see the ``Proposed Monitoring and Reporting''
section) are designed to provide NMFS with monitoring data from the
previous year to allow consideration of whether any changes are
appropriate. The use of adaptive management allows NMFS to consider new
information from different sources to determine (with input from the
LOA-holders regarding practicability) on an annual or biennial basis if
mitigation or monitoring measures should be modified (including
additions or deletions). Mitigation measures could be modified if new
data suggests that such modifications would have a reasonable
likelihood of reducing adverse effects to marine mammal species or
stocks or their habitat and if the measures are practicable. The
adaptive management process and associated reporting requirements would
serve as the basis for evaluating performance and compliance.
The following are some of the possible sources of applicable data
to be considered through the adaptive management process: (1) Results
from monitoring reports, as required by MMPA authorizations; (2)
results from general marine mammal and sound research; and (3) any
information which reveals that marine mammals may have been taken in a
manner, extent, or number not authorized by these regulations or
subsequent LOAs or that the specified activity may be having more than
a negligible impact on affected stocks.
Under this proposed rule, NMFS proposes an annual adaptive
management process involving BOEM, BSEE, and industry operators
(including geophysical companies as well as exploration and production
companies). Industry operators may elect to be represented in this
process by their respective trade associations. NMFS, BOEM, and BSEE
(i.e., the regulatory agencies) and industry operators who have
conducted or contracted for survey operations in the GOM in the prior
year (or their representatives) will provide an agreed-upon description
of roles and responsibilities, as well as points of contact, in advance
of each year's adaptive management process. The foundation of the
adaptive management process would be the annual comprehensive reports
produced by LOA-holders (or their representatives), as well as the
results of any relevant research activities, including research
supported voluntarily by the oil and gas industry and research
supported by the Federal government. Please see the ``Monitoring
Contribution Through Other Research'' section below for a description
of representative past research efforts. The outcome of the annual
adaptive management process would be an assessment of effects to marine
mammal populations in the GOM relative to NMFS's determinations under
the MMPA and ESA, recommendations related to mitigation, monitoring,
and reporting, and recommendations for future research (whether
supported by industry or the regulatory agencies).
Data collection and reporting by individual LOA-holders would occur
on an ongoing basis, per the terms of issued LOAs. In a given annual
cycle, we propose that the comprehensive annual report would summarize
and synthesize the LOA-specific reports received from
[[Page 29300]]
July 1 of one year through June 30 of the next, with report development
(supported through collaboration of individual LOA-holders or by their
representatives) occurring from July 1 through September 30 of a given
year. Review and revision of the report, followed by a joint meeting of
the parties, would occur between October 1 and December 31 of each
year. Any agreed-upon modifications would occur through the process for
modifications and/or adaptive management described in the proposed
regulatory text following this preamble.
Monitoring Contribution Through Other Research
NMFS's MMPA implementing regulations require that applicants for
incidental take authorizations describe the suggested means of
coordinating research opportunities, plans, and activities relating to
reducing incidental taking and evaluating its effects (50 CFR
216.104(a)(14)). Such coordination can serve as an effective supplement
to the monitoring and reporting required pursuant to issued LOAs and/or
incidental take regulations. We expect that relevant research efforts
will inform the annual adaptive management process describe above, and
that levels and types of research efforts will change from year to year
in response to identified needs and evolutions in knowledge, emerging
trends in the economy and available funding, and available scientific
and technological resources. Here, we describe examples of relevant
research efforts, which may not be predictive of any future levels and
types of research efforts. Research occurring in locations other than
the GOM may be relevant to understanding the effects of geophysical
surveys on marine mammals or marine mammal populations or the
effectiveness of mitigation.
Industry--In 2006, several exploration and production (E&P)
companies and industry associations began a multi-year research program
known as the E&P Sound and Marine Life Joint Industry Program (JIP).
The aim of the program was to advance scientific understanding of the
effects of sound generated by offshore oil and gas industry operations
on living marine resources, including marine mammals. Since its
inception, the JIP, the largest nongovernmental funder of research on
this topic, has allocated $55 million to fund a wide range of different
projects. The JIP website (www.soundandmarinelife.org) hosts a database
of available products funded partially or fully through the JIP. As of
June 2017, this database contained records for 133 JIP data products,
including 41 project reports and 83 peer-reviewed publications, as well
as the other notable products mentioned below. JIP policies stipulate
that the research results be shared in public reports and submitted to
peer-reviewed scientific journals to ensure maximum transparency and
value to the wider research, stakeholder, and regulatory communities.
JIP-funded projects and products are organized into six research
categories: (1) Sound source characterization; (2) physical and
physiological effects and hearing; (3) behavioral reactions and
biologically significant effects; (4) mitigation and monitoring; (5)
research tools; and (6) communication. Below, we summarize certain key
studies as well as additional initiatives that are planned or underway
(note that this is a small sample of studies and that not all of the
initiatives described below have been funded through the JIP).
Analyses of existing PSO data: The GOM is one of three
regions currently being reviewed under a JIP contract, initiated in
2016, to assess the utility of existing PSO data. Visual PSO and PAM
data through 2015 are being examined for quality and consistency,
and assessments will be made about the data's utility in the
validation of risk modeling, assessing behavioral responses, and the
potential for deriving animal density and distribution information.
This work will complement and reinforce similar efforts by BOEM (see
below). An earlier JIP study resulted in standardizing the basic
data recording formats used by vessel operators in the UK and other
jurisdictions (jncc.defra.gov.uk/page-1534).
Acoustic measurements and modeling: The JIP has funded
measurement of the acoustic output of both single airgun sources as
well as airgun arrays that help increase confidence in the source
and propagation models used in the GOM. These include extensive
near-field, mid-field, and far-field in-water acoustic measurements
(conducted in Norwegian waters in 2007-2010) of the most commonly
used single-source and two-element configurations over a range of
volumes, depths, and pressures with the objective of measuring
acoustic output at higher frequencies up to 50 kHz. More recently,
measurements of the sound field from a fully operational airgun
array in the GOM have been completed, with fully analyzed data
products anticipated in 2018. Additionally, the JIP is funding work
into the development of standard procedures for underwater noise
measurements for activities related to offshore oil and gas
exploration and production, to ensure that processing of selected
acoustic metrics used to describe the characteristics of a sound
signal propagating in water can be analyzed in a consistent and
systematic manner, and is funding a review of available marine
acoustic propagation models.
PAMGuard: Industry has funded ongoing development and
at-sea testing of this now-standard, open source real-time PAM
software to improve mitigation capabilities during operations. More
information and the software itself is available online at
www.pamguard.org.
Alternative technology: Pursuant to the terms of a
settlement agreement (as amended) concerning pending litigation
between the Natural Resources Defense Council et al. and the
Department of Interior (joined by industry as intervenor-defendants)
(NRDC et al. v. Zinke et al., Civil Action No. 2:10 cv-01882 (E.D.
La.)), industry has conducted a study of vibroseis technology,
including construction and testing of prototypes. Development of
vibroseis technology is promising in terms of reducing potential
harm to marine mammals because the system outputs lower peak
amplitude, and consequently less high-frequency energy, while
maintaining the main bandwidth necessary for seismic data
acquisition.
Advanced dive behavior tag technology development: The
JIP co-funded, with BOEM's predecessor agency (MMS) and the U.S.
Navy's Office of Naval Research (ONR), initial development of
advanced dive behavior tracking technology that has been used to
study sperm whale diving and foraging behavior in the GOM.
Effects of sound on marine mammal hearing: The JIP
funds multiple hearing research projects specifically focused on
defining the impacts of seismic sound sources on the hearing systems
of various marine mammal species, e.g., TTS, TTS growth, and masking
in bottlenose dolphins and harbor porpoise. For example, the JIP
funded research by the U.S. Navy's Marine Mammal Program that
specifically examined the physiological effect of airgun sound on
hearing in bottlenose dolphins by measuring TTS after exposure to
multiple seismic pulses (Finneran et al., 2015). New and ongoing
studies are aimed at developing an understanding of the role of
hearing recovery between exposures from intermittent sound sources,
like airguns, in the process of TTS generation, as well as
developing TTS growth functions to better refine TTS/PTS threshold
relationships. The JIP has also funded research into modeling work
to better estimate baleen whale hearing.
Behavioral response study: The JIP and BOEM jointly
funded a study examining how humpback whales respond to airgun sound
in general and to the ramp-up procedure specifically (Behavioral
Response of Australian Humpback Whales to Seismic Surveys (BRAHSS)).
The experimental design progressed from using a single airgun source
to a fully operational commercial array with a ramp-up procedure,
and involved treatment and control groups, a pre-trial statistical
power analysis, a range of exposures, and a four-stage ramp-up
design. For more details of the study and results, please see Cato
et al. (2013) and Dunlop et al. (2013, 2015, 2016, 2017).
BOEM--BOEM's Environmental Studies Program (ESP) develops, funds,
and manages scientific research to inform policy decisions regarding
OCS resource development. These environmental studies cover a broad
range of disciplines, including physical
[[Page 29301]]
oceanography, biology, protected species, and the environmental impacts
of energy development. Through the ESP, BOEM is a leading contributor
to the growing body of scientific knowledge about the marine and
coastal environment. BOEM and its predecessor agencies have funded more
than $1 billion in research since the studies program began in 1973.
Technical summaries of more than 1,200 BOEM-sponsored environmental
research projects and more than 3,400 research reports are publicly
available online through the Environmental Studies Program Information
System (ESPIS). Below, we summarize certain key studies, as well as
additional initiatives that are planned or underway. For the latest
information on BOEM's ongoing environmental studies work, please visit
www.boem.gov/studies.
Analyses of existing PSO data: MMS previously funded an
analysis of GOM PSO data from 2002-2008 (Barkaszi et al., 2012), and
BOEM has currently contracted for additional analyses of PSO data
from 2009-2015.
Development of PAM standards: As discussed in
``Proposed Monitoring and Reporting,'' BSEE is working with Scripps
Institute of Oceanography to develop standards for towed PAM
systems.
Passive acoustic monitoring: BOEM is funding a fixed
PAM array for 5 years. Hydrophones will be deployed, maintained, and
redeployed on a regular schedule throughout the GOM. Placement will
include shelf, slope and deep water depths as well as all planning
areas in order to gather a comprehensive data set representative of
the entire GOM. This program is expected to establish a relative
baseline for ambient noise in the GOM against which to evaluate
potential future noise impacts from permitted activities as well as
characterize the sound budget from other kinds of noise already
occurring in the GOM (e.g., shipping). In addition, acoustic
recorders will be able to detect vocalizing marine mammals,
providing both spatial and temporal information about cetacean
species in the GOM.
Sperm whale studies: The Sperm Whale Acoustic
Monitoring Program (SWAMP) began in 2000 with joint support from
MMS, ONR, and NMFS and laid the groundwork for future study by
developing new methods for studying sperm whale behavior and their
responses to sound. Subsequently, the Sperm Whale Seismic Study
(SWSS) began in 2002 to evaluate potential effects of geophysical
exploration on sperm whales in the GOM (e.g., Jochens et al., 2008).
SWSS included support from MMS, ONR, the National Science Foundation
(NSF), and a coalition of industry funders. In 2009, MMS (through an
interagency agreement with NMFS) began the Sperm Whale Acoustic Prey
Study (SWAPS), which studied how airgun noise may affect sperm whale
prey species (e.g., squid and small pelagic fish).
GoMMAPPS: BOEM is supporting a multi-year, multi-
disciplinary study of marine protected species in the GOM (Gulf of
Mexico Marine Assessment Program for Protected Species (GoMMAPPS)),
which is patterned after the successful Atlantic Marine Assessment
Program for Protected Species (AMAPPS) that began in 2010 and has
provided valuable information on the seasonal distribution and
abundance of protected species in U.S. waters of the Atlantic Ocean.
The overall goals are to improve our understanding of living marine
resource abundance, distribution, habitat use, and behavior in the
GOM to facilitate appropriate mitigation and monitoring of potential
impacts from human activities, including geophysical survey
activities. The study will utilize a variety of methods, depending
on target species, including aerial surveys, shipboard surveys,
satellite tagging and tracking, and genetic analyses. GoMMAPPS is a
joint partnership of BOEM, NMFS, the U.S. Fish and Wildlife Service,
and the U.S. Geological Survey. More information is available online
at (www.boem.gov/GOMMAPPS/).
Workshops: BOEM has funded various workshops, including
a 2012 workshop focused on mitigation and monitoring associated with
seismic surveys and a 2013 workshop concerning quieting technologies
for reducing noise during seismic surveying (BOEM, 2014).
Impact on Availability of Affected Species for Taking for Subsistence
Uses
There are no relevant subsistence uses of marine mammals implicated
by these actions. Therefore, we have determined that the total taking
of affected species or stocks would not have an unmitigable adverse
impact on the availability of such species or stocks for taking for
subsistence purposes.
Endangered Species Act (ESA)
Section 7(a)(2) of the Endangered Species Act of 1973 (16 U.S.C.
1531 et seq.) requires that each Federal agency insure that any action
it authorizes, funds, or carries out is not likely to jeopardize the
continued existence of any endangered or threatened species or result
in the destruction or adverse modification of designated critical
habitat. To ensure ESA compliance for the promulgation of regulations
and potential issuance of LOAs, NMFS consults internally whenever we
propose to authorize take for ESA-listed marine mammal species. The
sperm whale is listed as endangered under the ESA, and the GOM Bryde's
whale has been proposed to be listed as endangered. Consultation under
section 7 of the ESA will be concluded prior to issuance of any final
incidental take regulations.
Letters of Authorization
Under issued incidental take regulations, industry operators would
be able to apply for and obtain LOAs, as described in NMFS's MMPA
implementing regulations (50 CFR 216.106). LOAs may be issued for
multiple years, depending on the degree of specificity with which an
operator can describe their planned survey activities. Because the
specified activity described herein does not provide actual specifics
of the timing, location, and survey design for activities that would be
the subject of issued LOAs, such requests must include, at minimum, the
information described at 50 CFR 216.104(a)(1 and 2), and should include
an affirmation of intent to adhere to the mitigation, monitoring, and
reporting requirements described in the regulations. The level of
effort proposed by an operator would be used to develop an LOA-specific
take estimate based on the results of Zeddies et al. (2015, 2017a). The
annual estimated take, per zone and per species, would serve as a cap
on the number of authorizations that could be issued. Applicants may
choose to present additional information in a request for LOA, e.g.,
independent exposure estimates, description of proposed mitigation and
monitoring (if more stringent than the requirements in issued
regulations). However, such additional information would be subject to
NMFS review and approval as well as public review via a 30-day comment
period prior to issuance. Any substantive departure from the activity
and exposure estimation parameters described here and which form the
basis for our preliminary determinations would be subject to public
review.
Technologies continue to evolve to meet the technical,
environmental, and economic challenges of oil and gas development. The
use of ``new and unusual technologies'' (NUT), i.e., technologies other
than those described herein, would be evaluated on a case-by-case basis
and may require public review. Some seemingly new technologies proposed
for use by operators are often extended applications of existing
technologies and interface with the environment in essentially the same
way as well-known or conventional technologies. For such evaluations,
we propose to follow the existing process used by BOEM, by using the
following considerations:
Has the technology or hardware been used previously or
extensively in the U.S. GOM under operating conditions similar to
those anticipated for the activities proposed by the operator? If
so, the technology would not be considered a NUT;
Does the technology function in a manner that
potentially causes different impacts to the environment than similar
equipment or procedures did in the past? If
[[Page 29302]]
so, the technology would be considered a NUT;
Does the technology have a significantly different
interface with the environment than similar equipment or procedures
did in the past? If so, the technology would be considered a NUT;
and
Does the technology include operating characteristics
that are outside established performance parameters? If so, the
technology would be considered a NUT.
We would consult with BOEM as well as with NMFS's Endangered
Species Act Interagency Cooperation Division regarding the level of
review necessary for issuance of an LOA in which a NUT is proposed
for use.
Alternative Regulatory Text
Please see Table 11 for a summary of mitigation measures with
alternatives for consideration, for which alternative regulatory text
is presented here.
Area Restriction
Based on our analyses-to-date (``Proposed Mitigation''
and ``Negligible Impact Analysis and Preliminary Determination''),
we evaluated a year-round restriction on airgun surveys in Area 3
(Figure 5), and our preliminary finding of negligible impact on the
Gulf of Mexico stock of Bryde's whale is based on a year-round
restriction in this area. Alternative regulatory text at Sec.
217.184(e)(2) for this proposal would read: ``No use of airguns may
occur within the area bounded by the 100- and 400-m isobaths, from
87.5[deg] W to 27.5[deg] N (buffered by 6 km).''
For our proposals of no restriction or a seasonal restriction, but
with the addition of a requirement for BOEM and/or members or
representatives of the oil and gas industry to ensure real-time
detection of Bryde's whales across the area of potential impact
including real-time communication of detections to survey operators,
which would be used to initiate shutdowns to ensure that survey
operations do not take place when a Bryde's whale is within 6 km of the
acoustic source, the proposed regulatory text would be the following.
For the three-month restriction, we are proposing using a moored
listening array and thus the alternative regulatory text at Sec.
217.184(e)(2) would read: ``No use of airguns may occur within the area
bounded by the 100- and 400-m isobaths, from 87.5[deg] W to 27.5[deg] N
(buffered by 6 km), during June through August. During September
through May, LOA-holders conducting airgun surveys must monitor the
area of potential impact using a moored passive listening array and may
not use airguns when Bryde's whales are detected within 6 km of the
acoustic source.'' For no restriction plus a requirement of real-time
detection using the moored array in the area of impact alone,
alternative regulatory text at Sec. 217.184(e)(2) would read: ``In the
area bounded by the 100- and 400-m isobaths, from 87.5[deg] W to
27.5[deg] N (buffered by 6 km), LOA-holders conducting airgun surveys
must monitor a moored passive listening array and may not use airguns
when a confirmed or potential Bryde's whale is detected within 6 km of
the acoustic source.''
The proposal of a three-month seasonal restriction on airgun
surveys in Area 3 with no additional monitoring requirement is included
in the regulatory text at the end of this document, following the
preamble.
As mentioned in the ``Proposed Mitigation'' section, we are
interested in public comment on these proposals, including any data
that may support the necessary findings regarding potential impacts to
the GOM Bryde's whale for these proposals, as well as any additional
alternative proposals that could vary the time period or length of
seasonal closure from what NMFS has proposed.
Shutdowns
For the proposal requiring shutdown upon a confirmed acoustic
detection of sperm whales within 1 km or upon a confirmed visual or
acoustic detection of Bryde's whales, large whales with calf, beaked
whales, or Kogia spp. within 1 km, the regulatory text at Sec.
217.184(b)(6) would read: ``Buffer Zone and Exclusion Zone--The PSOs
shall establish and monitor a 500-m exclusion zone and additional 500-m
buffer to the exclusion zone. For all confirmed detections of baleen
whales, beaked whales, and Kogia spp., and for confirmed acoustic
detections of sperm whales, the full 1,000-m zone shall function as an
exclusion zone. These zones shall be based upon radial distance from
any element of the airgun array (rather than being based on the center
of the array or around the vessel itself). During use of the acoustic
source, occurrence of marine mammals within the buffer zone (but
outside the exclusion zone) shall be communicated to the operator to
prepare for the potential shutdown of the acoustic source. PSOs must
monitor the 1,000-m zone for a minimum of 30 minutes prior to ramp-up
(i.e., pre-clearance).'' Regulatory text at Sec. 217.184(b)(8)(ii)
would read: ``Upon completion of ramp-up, if a marine mammal appears
within, enters, or appears on a course to enter the exclusion zone, the
acoustic source must be shut down (i.e., power to the acoustic source
must be immediately turned off). If a marine mammal (excluding
delphinids) is detected acoustically and is determined to be within 1
km of the acoustic source, the acoustic source must be shut down.''
Regulatory text at Sec. 217.184(b)(8)(iv) would read: ``Shutdown of
the acoustic source is required upon detection (visual or acoustic) of
a baleen whale, beaked whale, or Kogia spp. within 1 km.''
For the proposal waiving the shutdown or power-down requirement
upon detection of small dolphins within a 500-m exclusion zone,
regulatory text at Sec. 217.184(b)(8)(iii) would read: ``This shutdown
requirement is waived for dolphins of the following genera: Tursiops,
Stenella, Steno, and Lagenodelphis. If there is uncertainty regarding
identification (i.e., whether the observed animal(s) belongs to the
group described above), shutdown must be implemented.''
The other proposals discussed in the ``Proposed Mitigation''
section for detection of Bryde's whales, beaked whales, sperm whales,
Kogia spp., and small dolphins are included in the regulatory text
following the preamble. As mentioned in the ``Proposed Mitigation''
section, we are interested in public comment on these proposals.
Scope of the Rule
NMFS requests comment on the issuance of incidental take
regulations that do not apply to BOEM's Eastern Planning Area. In the
regulatory text, 217.180(b) would be replaced with the following text:
``The taking of marine mammals by oil and gas industry operators may be
authorized in a Letter of Authorization (LOA) only if it occurs within
the Bureau of Ocean Energy Management's Western or Central Planning
Areas in the Gulf of Mexico.'' Under this alternative scope, NMFS would
continue working on a programmatic approach to the authorization of
take incidental to geophysical survey operations in the Eastern
Planning Area, but applicants could apply for individual permits (IHAs)
until that process is completed.
This revision of scope, if it occurred, would result in less
impacts to affected species or stocks of marine mammals relative to
what was considered in the analyses presented previously in this
preamble. Based on the analysis included in the preceding sections, if
no other changes are made to the scope of the rule or the required
mitigation measures analyzed in the preceding sections (i.e., the
measures are not modified as considered above in this Alternatives for
Consideration section), we preliminarily find that the total marine
mammal take from the proposed activity (reflecting the revised scope
considered here) will have a negligible impact on all affected marine
mammal species or stocks and the mitigation
[[Page 29303]]
measures included would effect the least practicable adverse impact on
the affected species and stocks and their habitat.
Request for Information
NMFS requests interested persons to submit comments, information,
and suggestions concerning the proposed rule and regulations, including
the variations of the proposed rule, two economic baselines, and other
information provided in the Regulatory Impact Analysis and associated
appendices (www.fisheries.noaa.gov/national/marine-mammal-protection/incidental-take-authorizations-oil-and-gas) (see ADDRESSES). All
comments will be reviewed and evaluated as we prepare the final rule.
This proposed rule and referenced documents provide all environmental
information relating to our proposed action for public review.
Classification
Pursuant to the procedures established to implement Executive Order
12866, the Office of Management and Budget has determined that this
proposed rule is significant. Accordingly, a regulatory impact analysis
(RIA) has been prepared and is available for review online at:
www.fisheries.noaa.gov/national/marine-mammal-protection/incidental-take-authorizations-oil-and-gas. The RIA evaluates the potential costs
and benefits of these proposed incidental take regulations, as well as
a more stringent alternative, against two baselines. The baselines
correspond with regulatory requirements associated with management of
geophysical survey activity in the GOM prior to 2013 (pursuant to
BOEM's authorities under the Outer Continental Shelf Lands Act) and
conditions in place since 2013 pursuant to a settlement agreement, as
amended through stipulated agreement, involving a stay of litigation
(NRDC et al. v. Zinke et al., Civil Action No. 2:10 cv-01882 (E.D.
La.)). Under the settlement agreement that is in effect, industry trade
groups representing operators agreed to include certain mitigation
requirements for geophysical surveys in the GOM. As described
previously in this preamble (``Economic Baseline''), NMFS is seeking
comment on the most appropriate baseline against which to measure the
costs and benefits of the proposed regulatory action.
The proposed rule would require new mitigation measures relative to
the baseline and, thus, new costs for survey operators. However, the
proposed rule would also alleviate the regulatory burden of
implementing minimum separation distance requirements for deep
penetration airgun surveys. The proposed rule also would result in
indirect (but non-monetized) costs as a result of the proposed time-
area restrictions. However, we do not believe that these would be
significant, as described in the RIA and in the ``Proposed Mitigation''
section. Moreover, as described in the RIA, total costs related to
compliance for survey activities are small compared with expenditures
on other aspects of oil and gas industry operations, and direct
compliance costs of the regulatory requirements are unlikely to result
in materially reduced oil and gas activities in the GOM.
The proposed rule would also result in certain non-monetized
benefits. The protection of marine mammals afforded by this rule
(pursuant to the requirements of the MMPA) would benefit the regional
economic value of marine mammals via tourism and recreation to some
extent, as mitigation measures applied to geophysical survey activities
in the GOM region are expected to benefit the marine mammal populations
that support this economic activity in the GOM. In addition, some
degree of benefits can be expected to accrue solely via ecological
benefits to marine mammals and other wildlife as a result of the
proposed regulatory requirements. The published literature (described
in the RIA) is clear that healthy populations of marine mammals and
other co-existing species benefit regional economies and provide social
welfare benefits to people; however, it does not provide a basis for
quantitatively valuing the cost of anticipated incremental changes in
environmental disturbance and marine mammal harassment associated with
the proposed rule.
Notably, the proposed rule would also afford significant benefit to
the regulated industry by providing an efficient framework within which
to achieve compliance with the MMPA, and the attendant regulatory
certainty. In particular, cost savings may be generated by the reduced
administrative effort required to obtain an LOA under the framework
established by a rule compared to what would be required to obtain an
incidental harassment authorization (IHA) under section 101(a)(5)(D).
Absent the rule, survey operators in the GOM would likely be required
to apply for an IHA. Although not monetized in the RIA, NMFS's analysis
indicates that the upfront work associated with the rule (e.g.,
analyses, modeling, process for obtaining LOA) would likely save
significant time and money for operators. A conservative cost savings
calculation, based on estimates of the costs for IHA applications
(provided by a contractor providing such services) relative to LOA
application costs and an assumption of the number of likely
authorizations based on total annual survey days and survey estimates
included in the RIA, ranges from $500,000 to $1.5 million annually. In
terms of timing, NMFS recommends that IHA applicants contact the agency
six to nine months in advance of the planned activity, whereas NMFS
anticipates a timeframe of just three months for LOA applications under
a rule.
We prepared an initial regulatory flexibility analysis (IRFA), as
required by Section 603 of the Regulatory Flexibility Act (RFA), for
this proposed rule. The IRFA describes the economic effects this
proposed rule, if adopted, would have on small entities. A description
of this action, why it is being considered, the objectives of, and
legal basis for this proposed rule are contained in the preamble of
this proposed rule. A copy of the full analysis is available online at
www.fisheries.noaa.gov/national/marine-mammal-protection/incidental-take-authorizations-oil-and-gas. The MMPA provides the statutory basis
for this proposed rule. No duplicative, overlapping, or conflicting
Federal rules have been identified. A summary of the IRFA follows.
This proposed rule is expected to directly regulate businesses that
conduct geophysical surveys in the GOM with the potential to
incidentally take marine mammals. Some of these businesses may be
defined as small entities. The IRFA identifies these businesses as well
as potential indirect impacts to small business boat owners and
operators, who would not be directly regulated by the rule, but who may
be involved in the implementation of the survey activities. The IRFA
found that, for ten years of relevant permit data (2006-2015), 62 U.S.
based-companies applied for 284 permits for relevant surveys, in 15
different industry NAICS codes. The IRFA also found that, for the
period 2012-2014, 33 U.S.-flagged vessels operated under contract to
permit applicants; the parent companies and primary NAICS codes under
which those vessels operated were also identified where possible.
Of the total number of relevant survey applications from 2006-2015,
12 percent (75 applications) were put forth by small entities. In
total, 34 U.S.-based small businesses applied for relevant permits in
the GOM between 2006-2015, representing only 12 percent of permit
applications during this period.
[[Page 29304]]
Foreign businesses and U.S.-based large businesses applied for more
permits per business than did small businesses. Companies involved in
crude petroleum and natural gas extraction (NAICS 211111) and support
activities for oil and gas (NAICS 213112) conducted the majority of the
surveys by small companies (87 percent of companies). Historically,
small entities undertook a larger percentage of HRG surveys (airgun and
non-airgun) than did businesses as a whole (85 percent of surveys
conducted by small businesses were HRG, compared to 57 percent of
surveys by all entities). Small businesses did not undertake larger
surveys (e.g., 3D WAZ), according to the permit database reviewed.
Using this information, the IRFA finds that small entities would
participate in approximately 33 to 57 surveys over the five years, or
approximately 7 to 11 surveys annually, and that approximately 15 to 26
small companies will likely apply for relevant permits over the five
years (approximately 3 to 5 small companies each year). The future
distribution of small companies by industry is not known, but the
historical pattern suggests that companies involved in crude petroleum
and natural gas extraction (NAICS 211111) and support activities for
oil and gas (213112) will conduct the majority of the surveys by small
companies.
Annual median revenues for small entities who applied for relevant
permits were $12.26 million. Incremental costs of the proposed rule for
non-airgun surveys, which comprised most of the HRG surveys (95 percent
are forecast to be non-airgun, as opposed to airgun, surveys), are
anticipated to range from $5,700 to $12,300 per survey. Airgun HRG
survey costs are anticipated to range from $25,800 to $37,500 per
survey. Approximately four small entities are anticipated to be
involved in survey activities annually over the five years. As such,
impacts would not be universally experienced by all small entities, and
would depend on the specific survey types the companies engaged in.
Incremental impacts for HRG surveys, which historically comprised most
small business surveys, are anticipated to increase costs to small
entities by one percent or less of annual revenues. For those entities
engaged in other types of surveys, costs could comprise a larger
portion of annual revenues.
In summary, the IRFA finds: (1) In the majority of cases (88
percent), survey permit applicants are large businesses; (2) When the
permit applicants are small businesses, the majority of the time (63
percent) they are oil and gas extractors (NAICS 211111); (3) Together
these permits (for large businesses and small businesses with high
annual revenues for which rule costs are a small fraction) account for
96 percent of the survey permits; (4) While small entities in other
industries occasionally apply for permits (four percent historically),
these businesses are quite small, with average annual revenues in the
millions or even less. Given their size, it is unlikely that these
permit applicants bear survey costs; otherwise it would be reflected in
their annual revenues (i.e., their revenues on average would reflect
that they recover their costs). Accordingly, we expect it is most
likely the survey costs are passed on to oil and gas extraction
companies who commission the surveys or purchase the data; and (5)
Overall, up to five small businesses (NAICS 211111) per year may
experience increased costs of between 0.1 and 1.1 percent of average
annual revenues.
NMFS's RIA evaluates the incremental regulatory impact of the
proposed rule, as well as the incremental regulatory impact of a more
stringent alternative to the mitigation, monitoring, and reporting
requirements of the proposed rule. NMFS is requesting comment on the
costs of these proposed incidental take regulations on small entities,
with the goal of ensuring a thorough consideration and discussion at
the final rule stage. We request comments on the analysis of entities
affected, as well as information on regulatory alternatives that would
simultaneously reduce the burden on small entities and afford
appropriate protections to affected marine mammal species and stocks.
This proposed rule contains a collection-of-information requirement
subject to the provisions of the Paperwork Reduction Act (PRA).
Notwithstanding any other provision of law, no person is required to
respond to nor shall a person be subject to a penalty for failure to
comply with a collection of information subject to the requirements of
the PRA unless that collection of information displays a currently
valid OMB control number. These requirements have been approved by OMB
under control number 0648-0151, currently under application for
renewal, and include applications for regulations, subsequent LOAs, and
reports. Send comments regarding any aspect of this data collection,
including suggestions for reducing the burden, to NMFS and the OMB Desk
Officer (see Addresses).
List of Subjects in 50 CFR Part 217
Exports, Fish, Imports, Indians, Labeling, Marine mammals,
Penalties, Reporting and recordkeeping requirements, Seafood,
Transportation.
Dated: June 12, 2018.
Donna S. Wieting,
Acting Deputy Assistant Administrator for Regulatory Programs, National
Marine Fisheries Service.
For reasons set forth in the preamble, 50 CFR part 217 is proposed
to be amended as follows:
PART 217--REGULATIONS GOVERNING THE TAKING AND IMPORTING OF MARINE
MAMMALS
0
1. The authority citation for part 217 continues to read as follows:
Authority: 16 U.S.C. 1361 et seq.
0
2. The heading of part 217 is revised to read as set forth above.
0
3. Add Subpart S to read as follows:
Subpart S--Taking Marine Mammals Incidental to Geophysical Survey
Activities in the Gulf of Mexico
Sec.
217.180 Specified activity and specified geographical region.
217.181 Effective dates.
217.182 Permissible methods of taking.
217.183 Prohibitions.
217.184 Mitigation requirements.
217.185 Requirements for monitoring and reporting.
217.186 Letters of Authorization (LOA).
217.187 Renewals and modifications of Letters of Authorization.
217.188 [Reserved]
217.189 [Reserved]
Subpart S--Taking Marine Mammals Incidental to Geophysical Survey
Activities in the Gulf of Mexico
Sec. 217.180 Specified activity and specified geographical region.
(a) Regulations in this subpart apply only to oil and gas industry
operators (LOA-holders), and those persons authorized to conduct
activities on their behalf, for the taking of marine mammals that
occurs in the area outlined in paragraph (b) of this section and that
occurs incidental to geophysical survey activities.
(b) The taking of marine mammals by oil and gas industry operators
may be authorized in a Letter of Authorization (LOA) only if it occurs
within the Gulf of Mexico.
Sec. 217.181 Effective dates.
Regulations in this subpart are effective from [EFFECTIVE DATE OF
FINAL RULE] through [DATE 5 YEARS AFTER EFFECTIVE DATE OF FINAL RULE].
[[Page 29305]]
Sec. 217.182 Permissible methods of taking.
Under LOAs issued pursuant to Sec. 216.106 of this chapter and
Sec. 217.186, LOA-holders may incidentally, but not intentionally,
take marine mammals within the area described in Sec. 217.180(b) by
Level A and Level B harassment associated with geophysical survey
activities, provided the activity is in compliance with all terms,
conditions, and requirements of the regulations in this subpart and the
appropriate LOA.
Sec. 217.183 Prohibitions.
Notwithstanding takings contemplated in Sec. 217.180 and Sec.
217.182, and authorized by a LOA issued under Sec. 216.106 of this
chapter and Sec. 217.186, no person in connection with the activities
described in Sec. 217.180 may:
(a) Violate, or fail to comply with, the terms, conditions, and
requirements of this subpart or a LOA issued under Sec. 216.106 of
this chapter and Sec. 217.186;
(b) Take any marine mammal not specified in such LOAs;
(c) Take any marine mammal specified in such LOAs in any manner
other than as specified;
(d) Take a marine mammal specified in such LOAs if NMFS determines
such taking results in more than a negligible impact on the species or
stocks of such marine mammal; or
(e) Take a marine mammal specified in such LOAs if NMFS determines
such taking results in an unmitigable adverse impact on the species or
stock of such marine mammal for taking for subsistence uses.
Sec. 217.184 Mitigation requirements.
When conducting the activities identified in Sec. 217.180, the
mitigation measures contained in any LOA issued under Sec. 216.106 of
this chapter and Sec. 217.186 must be implemented. These mitigation
measures shall include but are not limited to:
(a) General conditions:
(1) A copy of any issued LOA must be in the possession of the LOA-
holder, the vessel operator and other relevant personnel, the lead
protected species observer (PSO), and any other relevant designees of
the LOA-holder operating under the authority of the LOA.
(2) The LOA-holder shall ensure that the vessel operator and other
relevant vessel personnel are briefed on all responsibilities,
communication procedures, marine mammal monitoring protocol,
operational procedures, and LOA requirements prior to the start of
survey activity, and when relevant new personnel join the survey
operations. The LOA-holder shall instruct relevant vessel personnel
with regard to the authority of the protected species monitoring team,
and shall ensure that relevant vessel personnel and protected species
monitoring team participate in a joint onboard briefing led by the
vessel operator and lead PSO to ensure that responsibilities,
communication procedures, marine mammal monitoring protocol,
operational procedures, and LOA requirements are clearly understood.
This briefing must be repeated when relevant new personnel join the
survey operations.
(b) Deep penetration airgun surveys:
(1) Deep penetration airgun surveys are defined as surveys using
airgun arrays with total volume greater than 400 in\3\.
(2) The LOA-holder must use independent, dedicated, trained PSOs,
meaning that the PSOs must be employed by a third-party observer
provider, may have no tasks other than to conduct observational effort,
record observational data, and communicate with and instruct relevant
vessel crew with regard to the presence of marine mammals and
mitigation requirements (including brief alerts regarding maritime
hazards), and must have successfully completed an approved PSO training
course. NMFS will maintain a list of approved PSOs and, for PSOs not on
the list, NMFS must review and approve PSO resumes accompanied by a
relevant training course information packet that includes the name and
qualifications (i.e., experience, training completed, and educational
background) of the instructor(s), the course outline or syllabus, and
course reference material as well as a document stating the PSO's
successful completion of the course. NMFS shall have one week to
approve PSOs from the time that the necessary information is submitted,
after which PSOs meeting the minimum requirements shall automatically
be considered approved.
(3) At least one visual PSO and two acoustic PSOs must have a
minimum of 90 days at-sea experience working in those roles,
respectively, during a deep penetration seismic survey, with no more
than eighteen months elapsed since the conclusion of the at-sea
experience. One visual PSO with such experience shall be designated as
the lead for the entire protected species observation team. The lead
shall coordinate duty schedules and roles for the PSO team and serve as
primary point of contact for the vessel operator. To the maximum extent
practicable, the lead PSO shall devise the duty schedule such that
experienced PSOs are on duty with those PSOs with appropriate training
but who have not yet gained relevant experience.
(4) Visual observation:
(i) During survey operations (e.g., any day on which use of the
acoustic source is planned to occur, and whenever the acoustic source
is in the water, whether activated or not), a minimum of two PSOs must
be on duty and conducting visual observations at all times during
daylight hours (i.e., from 30 minutes prior to sunrise through 30
minutes following sunset) and 30 minutes prior to and during nighttime
ramp-ups of the airgun array.
(ii) Visual monitoring must begin not less than 30 minutes prior to
ramp-up and must continue until one hour after use of the acoustic
source ceases or until 30 minutes past sunset.
(iii) Visual PSOs shall coordinate to ensure 360[deg] visual
coverage around the vessel from the most appropriate observation posts,
and shall conduct visual observations using binoculars and the naked
eye while free from distractions and in a consistent, systematic, and
diligent manner.
(iv) Visual PSOs shall immediately communicate all observations to
acoustic PSOs, including any determination by the PSO regarding species
identification, distance, and bearing and the degree of confidence in
the determination.
(v) Visual PSOs may be on watch for a maximum of two consecutive
hours followed by a break of at least one hour between watches and may
conduct a maximum of 12 hours of observation per 24-hour period.
(vi) Any observations of marine mammals by crew members aboard any
vessel associated with the survey shall be relayed to the PSO team.
(vii) During good conditions (e.g., daylight hours; Beaufort sea
state (BSS) 3 or less), visual PSOs shall conduct observations when the
acoustic source is not operating for comparison of sighting rates and
behavior with and without use of the acoustic source and between
acquisition periods, to the maximum extent practicable.
(5) Acoustic observation:
(i) All surveys must use a towed passive acoustic monitoring (PAM)
system at all times when operating in waters deeper than 100 m, which
must be monitored beginning at least 30 minutes prior to ramp-up and at
all times during use of the acoustic source.
(ii) Acoustic PSOs shall immediately communicate all detections to
visual PSOs, when visual PSOs are on duty, including any determination
by the PSO regarding species identification, distance, and bearing and
the degree of confidence in the determination.
[[Page 29306]]
(iii) Acoustic PSOs may be on watch for a maximum of four
consecutive hours followed by a break of at least two hours between
watches and may conduct a maximum of 12 hours of observation per 24-
hour period.
(iv) Survey activity may continue for brief periods of time when
the PAM system malfunctions or is damaged. Activity may continue for 30
minutes without PAM while the PAM operator diagnoses the issue. If the
diagnosis indicates that the PAM system must be repaired to solve the
problem, operations may continue for an additional two hours without
acoustic monitoring under the following conditions:
(A) Daylight hours and sea state is less than or equal to BSS 4;
(B) No marine mammals (excluding delphinids) detected solely by PAM
in the exclusion zone in the previous two hours;
(C) NMFS is notified via email as soon as practicable with the time
and location in which operations began without an active PAM system;
and
(D) Operations with an active acoustic source, but without an
operating PAM system, do not exceed a cumulative total of four hours in
any 24-hour period.
(6) Exclusion Zone and Buffer Zone--The PSOs shall establish and
monitor a 500-m exclusion zone and additional 500-m buffer zone. These
zones shall be based upon radial distance from any element of the
airgun array (rather than being based on the center of the array or
around the vessel itself). During use of the acoustic source,
occurrence of marine mammals within the buffer zone (but outside the
exclusion zone) shall be communicated to the operator to prepare for
the potential shutdown of the acoustic source. PSOs must monitor the
1,000-m zone for a minimum of 30 minutes prior to ramp-up (i.e., pre-
clearance).
(7) Ramp-up--A ramp-up procedure, involving a step-wise increase in
the number of airguns firing and total array volume until all
operational airguns are activated and the full volume is achieved, is
required at all times as part of the activation of the acoustic source.
Ramp-up may not be initiated if any marine mammal is within the
designated exclusion zone or buffer zone. If a marine mammal is
observed within these zones during the pre-clearance period, ramp-up
may not begin until the animal(s) has been observed exiting the 1,000-m
zone or until an additional time period has elapsed with no further
sightings (i.e., 15 minutes for small odontocetes and 30 minutes for
all other species). PSOs shall monitor the exclusion zone during ramp-
up, and ramp-up must cease and the source shut down upon observation of
marine mammals within the zones. Ramp-up may occur at times of poor
visibility if appropriate acoustic monitoring has occurred with no
detections in the 30 minutes prior to beginning ramp-up. Acoustic
source activation may only occur at times of poor visibility where
operational planning cannot reasonably avoid such circumstances. The
operator must notify a designated PSO of the planned start of ramp-up
as agreed-upon with the lead PSO; the notification time should not be
less than 60 minutes prior to the planned ramp-up. A designated PSO
must be notified again immediately prior to initiating ramp-up
procedures and the operator must receive confirmation from the PSO to
proceed. Ramp-up shall begin by activating a single airgun of the
smallest volume in the array and shall continue in stages by doubling
the number of active elements at the commencement of each stage, with
each stage of approximately the same duration. Duration should not be
less than 20 minutes. The operator must provide information to the PSO
documenting that appropriate procedures were followed. Ramp-ups shall
be scheduled so as to minimize the time spent with source activated
prior to reaching the designated run-in.
(8) Shutdown requirements:
(i) Any PSO on duty has the authority to delay the start of survey
operations or to call for shutdown of the acoustic source pursuant to
the requirements of this subpart. When shutdown is called for by a PSO,
the acoustic source must be immediately deactivated and any dispute
resolved only following deactivation. The operator must establish and
maintain clear lines of communication directly between PSOs on duty and
crew controlling the acoustic source to ensure that shutdown commands
are conveyed swiftly while allowing PSOs to maintain watch. When there
is certainty regarding the need for mitigation action on the basis of
either visual or acoustic detection alone, the relevant PSO(s) must
call for such action immediately. When there is uncertainty regarding
the nature of the observation, all on duty PSOs must agree upon the
mitigation action. When only the acoustic PSO is on duty and there is
uncertainty regarding the need for mitigation action on the basis of a
detection, the PSO may request that the acoustic source be shut down as
a precaution.
(ii) Upon completion of ramp-up, if a marine mammal appears within,
enters, or is clearly on a course to enter the exclusion zone, the
acoustic source must be shut down (i.e., power to the acoustic source
must be immediately turned off). If a marine mammal (excluding
delphinids) is detected acoustically, the acoustic source must be shut
down.
(iii) This shutdown requirement is waived for dolphins of the
following genera: Tursiops, Stenella, Steno, and Lagenodelphis. Instead
of shutdown, the acoustic source must be powered down to the smallest
single element of the array if a dolphin of the indicated genera
appears within or enters the 500-m exclusion zone, or is acoustically
detected and localized within the zone. Power-down conditions shall be
maintained until the animal(s) is observed exiting the exclusion zone
or for 15 minutes beyond the last observation of the animal, following
which full-power operations may be resumed without ramp-up.
(iv) Shutdown of the acoustic source is required upon detection
(visual or acoustic) of a baleen whale, beaked whale, or Kogia spp. at
any distance.
(v) Shutdown of the acoustic source is required upon observation of
a whale (i.e., sperm whale or any baleen whale) with calf at any
distance, with ``calf'' defined as an animal less than two-thirds the
body size of an adult observed to be in close association with the
calf.
(vi) Upon implementation of shutdown, the source may be reactivated
after the animal(s) has been observed exiting the exclusion zone or
following a 30-minute clearance period with no further observation of
the animal(s). Where there is no relevant zone (e.g., shutdown due to
observation of a baleen whale), a 30-minute clearance period must be
observed following the last observation of the animal(s).
(vii) If the acoustic source is shut down for reasons other than
mitigation (e.g., mechanical difficulty) for brief periods (i.e., less
than 30 minutes), it may be activated again without ramp-up if PSOs
have maintained constant visual and acoustic observation and no visual
detections of any marine mammal have occurred within the exclusion zone
and no acoustic detections (excluding delphinids) have occurred. For
any longer shutdown, pre-clearance watch and ramp-up are required. For
any shutdown at night or in periods of poor visibility (e.g., BSS 4 or
greater), ramp-up is required but if the shutdown period was brief and
constant observation maintained, pre-clearance watch is not required.
(9) Miscellaneous protocols:
(i) The acoustic source must be deactivated when not acquiring data
or preparing to acquire data, except as
[[Page 29307]]
necessary for testing. Unnecessary use of the acoustic source shall be
avoided. Notified operational capacity (not including redundant backup
airguns) must not be exceeded during the survey, except where
unavoidable for source testing and calibration purposes. All occasions
where activated source volume exceeds notified operational capacity
must be noticed to the PSO(s) on duty and fully documented. The lead
PSO must be granted access to relevant instrumentation documenting
acoustic source power and/or operational volume.
(ii) Testing of the acoustic source involving all elements requires
normal mitigation protocols (e.g., ramp-up). Testing limited to
individual source elements or strings does not require ramp-up but does
require pre-clearance.
(c) Shallow penetration surveys:
(1) Shallow penetration surveys are defined as surveys using airgun
arrays with total volume equal to or less than 400 in\3\ or boomers.
(2) LOA-holders shall follow the requirements defined for deep
penetration airgun surveys at Sec. 217.184(b), with the following
exceptions:
(i) Use of a towed PAM system is not required except to begin use
of the airgun(s) at night in waters deeper than 100 m. Use of a PAM
system is required for nighttime start-up, with monitoring by a trained
and experienced acoustic PSO during a 30-minute pre-clearance period
and during the ramp-up period (if applicable). The required acoustic
PSO may be a crew member.
(ii) Ramp-up is not required for shallow penetration surveys using
only a single airgun or boomer.
(iii) The exclusion zone shall be established at a distance of 200
m, with an additional 200-m buffer monitored during pre-clearance.
(iv) No shutdown or power-down action is required upon detection of
the dolphin genera described at Sec. 217.184(b)(8)(iii) for surveys
using a single airgun or boomer.
(v) Shutdowns are not required for observations beyond the
exclusion zone under any circumstance.
(d) Non-airgun surveys:
(1) Non-airgun surveys are defined as surveys using an acoustic
source other than an airgun(s) or boomer that operates at frequencies
less than 200 kHz (i.e., side-scan sonar, multibeam echosounder, or
subbottom profiler).
(2) LOA-holders conducting non-airgun surveys shall follow the
requirements defined for shallow penetration surveys at Sec.
217.184(c), with the following exceptions:
(i) Use of a towed PAM system is not required under any
circumstances;
(ii) Ramp-up is not required under any circumstances;
(iii) Non-airgun surveys shall employ a minimum of one trained and
experienced independent visual PSO during all daylight operations (as
described at Sec. 217.184(b)) when operating in waters deeper than 200
m. In waters shallower than 200 m, non-airgun surveys shall employ one
trained visual PSO, who may be a crew member, to monitor the exclusion
zone and buffer during the pre-clearance period; and
(iv) No shutdown or power-down action is required upon detection of
the dolphin genera described at Sec. 217.184(b)(8)(iii).
(e) Restriction areas:
(1) From February 1 through May 31, no use of airguns may occur
shoreward of the 20-m isobath (buffered by 13 km).
(2) No use of airguns may occur within the area bounded by the 100-
and 400-m isobaths, from 87.5[deg] W to 27.5[deg] N (buffered by 6 km),
during June through August.
(3) No use of airguns may occur within the area bounded by the 200-
and 2,000-m isobaths from the northern border of BOEM's Howell Hook
leasing area to 81.5[deg] W (buffered by 9 km).
(f) To avoid the risk of entanglement, LOA-holders conducting
surveys using ocean-bottom nodes or similar gear must:
(1) Use negatively buoyant coated wire-core tether cable;
(2) Retrieve all lines immediately following completion of the
survey;
(3) Attach acoustic pingers directly to the coated tether cable;
acoustic releases should not be used; and
(4) Employ a third-party PSO aboard the node retrieval vessel in
order to document any unexpected marine mammal entanglement.
(g) To avoid the risk of vessel strike, LOA-holders must adhere to
the following requirements:
(1) Vessel operators and crews must maintain a vigilant watch for
all marine mammals and slow down or stop their vessel or alter course,
as appropriate and regardless of vessel size, to avoid striking any
marine mammal. A visual observer aboard the vessel must monitor a
vessel strike avoidance zone around the vessel, which shall be defined
according to the parameters stated in this subsection, to ensure the
potential for strike is minimized. Visual observers monitoring the
vessel strike avoidance zone can be either third-party observers or
crew members, but crew members responsible for these duties must be
provided sufficient training to distinguish marine mammals from other
phenomena and broadly to identify a marine mammal as a baleen whale,
sperm whale, or other marine mammal;
(2) All vessels, regardless of size, must observe a 10 kn speed
restriction within the restriction area described previously at Sec.
217.184(e)(2);
(3) Vessel speeds must also be reduced to 10 kn or less when
mother/calf pairs, pods, or large assemblages of cetaceans are observed
near a vessel;
(4) All vessels must maintain a minimum separation distance of 500
yd (457 m) from baleen whales;
(5) All vessels must maintain a minimum separation distance of 100
yd (91 m) from sperm whales;
(6) All vessels must attempt to maintain a minimum separation
distance of 50 yd (46 m) from all other marine mammals, with an
exception made for those animals that approach the vessel;
(7) When cetaceans are sighted while a vessel is underway, vessels
shall attempt to remain parallel to the animal's course, and shall
avoid excessive speed or abrupt changes in direction until the animal
has left the area; and
(8) If cetaceans are sighted in a vessel's path or in close
proximity to a moving vessel, the vessel shall reduce speed and shift
the engine to neutral, not engaging the engines until animals are clear
of the area. This does not apply to any vessel towing gear.
Sec. 217.185 Requirements for monitoring and reporting.
(a) LOA-holders must provide bigeye binoculars (e.g., 25 x 150; 2.7
view angle; individual ocular focus; height control) of appropriate
quality (i.e., Fujinon or equivalent) solely for PSO use. These shall
be pedestal-mounted on the deck at the most appropriate vantage point
that provides for optimal sea surface observation, PSO safety, and safe
operation of the vessel. The operator must also provide a night-vision
device suited for the marine environment for use during nighttime ramp-
up pre-clearance, at the discretion of the PSOs. At minimum, the device
should feature automatic brightness and gain control, bright light
protection, infrared illumination, and optics suited for low-light
situations.
(b) PSOs must also be equipped with reticle binoculars (e.g., 7 x
50) of appropriate quality (i.e., Fujinon or equivalent), GPS, a
digital single-lens reflex camera of appropriate quality (i.e., Canon
or equivalent), a compass, and any other tools necessary to adequately
perform necessary tasks, including accurate determination of
[[Page 29308]]
distance and bearing to observed marine mammals.
(c) PSO qualifications:
(1) PSOs must successfully complete relevant training, including
completion of all required coursework and passing (80 percent or
greater) a written and/or oral examination developed for the training
program.
(2) PSOs must have successfully attained a bachelor's degree from
an accredited college or university with a major in one of the natural
sciences and a minimum of 30 semester hours or equivalent in the
biological sciences and at least one undergraduate course in math or
statistics. The educational requirements may be waived by NMFS if the
PSO has acquired the relevant skills through alternate experience.
Requests for such a waiver shall be submitted to NMFS and must include
written justification. Requests shall be granted or denied (with
justification) by NMFS within one week of receipt of submitted
information. Alternate experience that may be considered includes, but
is not limited to:
(i) Secondary education and/or experience comparable to PSO duties;
(ii) Previous work experience conducting academic, commercial, or
government-sponsored marine mammal surveys; or
(iii) Previous work experience as a PSO; the PSO should demonstrate
good standing and consistently good performance of PSO duties.
(d) Data collection--PSOs must use standardized data forms, whether
hard copy or electronic. PSOs shall record detailed information about
any implementation of mitigation requirements, including the distance
of animals to the acoustic source and description of specific actions
that ensued, the behavior of the animal(s), any observed changes in
behavior before and after implementation of mitigation, and if shutdown
was implemented, the length of time before any subsequent ramp-up of
the acoustic source to resume survey. If required mitigation was not
implemented, PSOs should record a description of the circumstances. We
require that, at a minimum, the following information be recorded:
(1) Vessel names (source vessel and other vessels associated with
survey) and call signs;
(2) PSO names and affiliations;
(3) Dates of departures and returns to port with port name;
(4) Dates and times (Greenwich Mean Time) of survey effort and
times corresponding with PSO effort;
(5) Vessel location (latitude/longitude) when survey effort begins
and ends; vessel location at beginning and end of visual PSO duty
shifts;
(6) Vessel heading and speed at beginning and end of visual PSO
duty shifts and upon any line change;
(7) Environmental conditions while on visual survey (at beginning
and end of PSO shift and whenever conditions change significantly),
including wind speed and direction, Beaufort sea state, Beaufort wind
force, swell height, weather conditions, cloud cover, sun glare, and
overall visibility to the horizon;
(8) Factors that may be contributing to impaired observations
during each PSO shift change or as needed as environmental conditions
change (e.g., vessel traffic, equipment malfunctions);
(9) Survey activity information, such as acoustic source power
output while in operation, number and volume of airguns operating in
the array, tow depth of the array, and any other notes of significance
(i.e., pre-ramp-up survey, ramp-up, shutdown, testing, shooting, ramp-
up completion, end of operations, streamers, etc.); and
(10) If a marine mammal is sighted, the following information
should be recorded:
(i) Watch status (sighting made by PSO on/off effort,
opportunistic, crew, alternate vessel/platform);
(ii) PSO who sighted the animal;
(iii) Time of sighting;
(iv) Vessel location at time of sighting;
(v) Water depth;
(vi) Direction of vessel's travel (compass direction);
(vii) Direction of animal's travel relative to the vessel;
(viii) Pace of the animal;
(ix) Estimated distance to the animal and its heading relative to
vessel at initial sighting;
(x) Identification of the animal (e.g., genus/species, lowest
possible taxonomic level, or unidentified), also note the composition
of the group if there is a mix of species;
(xi) Estimated number of animals (high/low/best);
(xii) Estimated number of animals by cohort (adults, yearlings,
juveniles, calves, group composition, etc.);
(xiii) Description (as many distinguishing features as possible of
each individual seen, including length, shape, color, pattern, scars or
markings, shape and size of dorsal fin, shape of head, and blow
characteristics);
(xiv) Detailed behavior observations (e.g., number of blows, number
of surfaces, breaching, spyhopping, diving, feeding, traveling; as
explicit and detailed as possible; note any observed changes in
behavior);
(xv) Animal's closest point of approach (CPA) and/or closest
distance from the center point of the acoustic source;
(xvi) Platform activity at time of sighting (e.g., deploying,
recovering, testing, shooting, data acquisition, other); and
(xvii) Description of any actions implemented in response to the
sighting (e.g., delays, shutdown, ramp-up, speed or course alteration,
etc.); time and location of the action should also be recorded.
(11) If a marine mammal is detected while using the PAM system, the
following information should be recorded:
(i) An acoustic encounter identification number, and whether the
detection was linked with a visual sighting;
(ii) Time when first and last heard;
(iii) Types and nature of sounds heard (e.g., clicks, whistles,
creaks, burst pulses, continuous, sporadic, strength of signal, etc.);
and
(iv) Any additional information recorded such as water depth of the
hydrophone array, bearing of the animal to the vessel (if
determinable), species or taxonomic group (if determinable),
spectrogram screenshot, and any other notable information.
(e) LOA-holders shall provide to NMFS within 90 days of survey
conclusion geo-referenced time-stamped vessel tracklines for all time
periods in which airguns were operating. Tracklines should include
points recording any change in airgun status (e.g., when the airguns
began operating, when they were turned off, or when they changed from
full array to single gun or vice versa). GIS files shall be provided in
ESRI shapefile format and include the UTC date and time, latitude in
decimal degrees, and longitude in decimal degrees. All coordinates
shall be referenced to the WGS84 geographic coordinate system.
(f) Reporting:
(1) Annual reporting: LOA-holders shall submit an annual summary
report to NMFS on all activities and monitoring results within 90 days
of the completion of the survey or expiration of the LOA, whichever
comes sooner. The report must describe all activities conducted and
sightings of marine mammals near the activities, must provide full
documentation of methods, results, and interpretation pertaining to all
monitoring, and must summarize the dates and locations of survey
operations and all marine mammal sightings (dates, times, locations,
activities, associated survey activities). Geospatial data regarding
locations where the acoustic source was used, provided to NMFS under
subparagraph Sec. 217.185(e), must
[[Page 29309]]
be provided as an ESRI shapefile with all necessary files and
appropriate metadata. The report must summarize the data collected as
required under Sec. 217.185(d). In addition to the report, all raw
observational data shall be made available to NMFS. The draft report
must be accompanied by a certification from the lead PSO as to the
accuracy of the report, and the lead PSO may submit directly to NMFS a
statement concerning implementation and effectiveness of the required
mitigation and monitoring. A final report must be submitted within 30
days following resolution of any comments on the draft report.
(2) Comprehensive reporting: LOA-holders shall contribute to the
compilation and analysis of data for inclusion in an annual synthesis
report addressing all data collected and reported through annual
reporting in each calendar year. The synthesis period shall include all
annual reports deemed to be final by NMFS from July 1 of one year
through June 30 of the subsequent year. The report must be submitted to
NMFS by October 1 of each year.
(g) Reporting of injured or dead marine mammals:
(1) In the unanticipated event that the activity defined in Sec.
217.180 clearly causes the take of a marine mammal in a prohibited
manner, the LOA-holder shall immediately cease such activity and report
the incident to the Office of Protected Resources (OPR), NMFS, and to
the Southeast Regional Stranding Coordinator, NMFS. Activities shall
not resume until NMFS is able to review the circumstances of the
prohibited take. NMFS will work with the LOA-holder to determine what
measures are necessary to minimize the likelihood of further prohibited
take and ensure MMPA compliance. The LOA-holder may not resume their
activities until notified by NMFS. The report must include the
following information:
(i) Time, date, and location (latitude/longitude) of the incident;
(ii) Name and type of vessel involved;
(iii) Vessel's speed during and leading up to the incident;
(iv) Description of the incident;
(v) Status of all sound source use in the 24 hours preceding the
incident;
(vi) Water depth;
(vii) Environmental conditions (e.g., wind speed and direction,
Beaufort sea state, cloud cover, visibility);
(viii) Description of all marine mammal observations in the 24
hours preceding the incident;
(ix) Species identification or description of the animal(s)
involved;
(x) Fate of the animal(s); and
(xii) Photographs or video footage of the animal(s).
(2) In the event that the LOA-holder discovers an injured or dead
marine mammal and determines that the cause of the injury or death is
unknown and the death is relatively recent (e.g., in less than a
moderate state of decomposition), the LOA-holder shall immediately
report the incident to OPR and the Southeast Regional Stranding
Coordinator, NMFS. The report must include the information identified
in paragraph (f)(1) of this section. Activities may continue while NMFS
reviews the circumstances of the incident. NMFS will work with the LOA-
holder to determine whether additional mitigation measures or
modifications to the activities are appropriate.
(3) In the event that the LOA-holder discovers an injured or dead
marine mammal and determines that the injury or death is not associated
with or related to the activities defined in Sec. 217.180 (e.g.,
previously wounded animal, carcass with moderate to advanced
decomposition, scavenger damage), the LOA-holder shall report the
incident to OPR and the Southeast Regional Stranding Coordinator, NMFS,
within 24 hours of the discovery. The LOA-holder shall provide
photographs or video footage or other documentation of the stranded
animal sighting to NMFS.
Sec. 217.186 Letters of Authorization (LOA).
(a) To incidentally take marine mammals pursuant to these
regulations, prospective LOA-holders must apply for and obtain a LOA.
(b) A LOA, unless suspended or revoked, may be effective for a
period not to exceed the expiration date of these regulations.
(c) In the event of projected changes to the activity or to
mitigation and monitoring measures required by a LOA, the LOA-holder
must apply for and obtain a modification of the LOA as described in
Sec. 217.187.
(d) The LOA shall set forth:
(1) Permissible methods of incidental taking;
(2) Means of effecting the least practicable adverse impact (i.e.,
mitigation) on the species or stock and its habitat; and
(3) Requirements for monitoring and reporting.
(e) Issuance of the LOA shall be based on a determination that the
level of taking will be consistent with the findings made for the total
taking allowable under these regulations and a determination that the
amount of take authorized under the LOA is of no more than small
numbers.
(f) For LOA issuance, where either:
(1) The conclusions put forth in an application (e.g., take
estimates) are based on analytical methods that differ substantively
from those used in the development of the rule; or
(2) The proposed activity or anticipated impacts vary substantively
in scope or nature from those analyzed in the preamble to the rule,
NMFS may publish a notice of proposed LOA in the Federal Register,
including the associated analysis of the differences, and solicit
public comment before making a decision regarding issuance of the LOA.
(g) Notice of issuance or denial of a LOA shall be published in the
Federal Register within thirty days of a determination.
Sec. 217.187 Renewals and modifications of Letters of Authorization.
(a) A LOA issued under Sec. 216.106 of this chapter and Sec.
217.186 for the activity identified in Sec. 217.180 shall be modified
upon request by the applicant, provided that:
(1) The proposed specified activity and mitigation, monitoring, and
reporting measures, as well as the anticipated impacts, are the same as
those described and analyzed for these regulations (excluding changes
made pursuant to the adaptive management provision in paragraph (c)(1)
of this section); and
(2) NMFS determines that the mitigation, monitoring, and reporting
measures required by the previous LOA under these regulations were
implemented.
(b) For LOA modification requests by the applicant that include
changes to the activity or the mitigation, monitoring, or reporting
(excluding changes made pursuant to the adaptive management provision
in paragraph (c)(1) of this section) that result in more than a minor
change in the total estimated number of takes (or distribution by
species or years), NMFS may publish a notice of proposed LOA in the
Federal Register, including the associated analysis of the change, and
solicit public comment before issuing the LOA.
(c) A LOA issued under Sec. 216.106 of this chapter and Sec.
217.186 for the activity identified in Sec. 217.180 may be modified by
NMFS under the following circumstances:
(1) Adaptive Management--NMFS may modify (including augment) the
existing mitigation, monitoring, or reporting measures (after
consulting with the LOA-holder regarding the practicability of the
modifications) if doing so is practicable and creates a
[[Page 29310]]
reasonable likelihood of more effectively accomplishing the goals of
the mitigation and monitoring set forth in the preamble for these
regulations;
(i) Possible sources of data that could contribute to the decision
to modify the mitigation, monitoring, or reporting measures in a LOA:
(A) Results from monitoring from previous years;
(B) Results from other marine mammal and/or sound research or
studies; and
(C) Any information that reveals marine mammals may have been taken
in a manner, extent or number not authorized by these regulations or
subsequent LOAs.
(ii) If, through adaptive management, the modifications to the
mitigation, monitoring, or reporting measures are substantial, NMFS
will publish a notice of proposed LOA in the Federal Register and
solicit public comment.
(2) Emergencies--If NMFS determines that an emergency exists that
poses a significant risk to the well-being of the species or stocks of
marine mammals specified in a LOA issued pursuant to Sec. 216.106 of
this chapter and Sec. 217.186, a LOA may be modified without prior
notice or opportunity for public comment. Notice would be published in
the Federal Register within thirty days of the action.
Sec. 217.188 [Reserved]
Sec. 217.189 [Reserved]
[FR Doc. 2018-12906 Filed 6-21-18; 8:45 am]
BILLING CODE 3510-22-P