[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