Aquatic Mammals Aquatic Mammals, Volume 33, Number 4, 2007 ISSN 0167-5427

Marine Mammal Noise Exposure Criteria: Initial Scientific Recommendations Brandon L. Southall, Ann E. Bowles, William T. Ellison, James J. Finneran, Roger L. Gentry, Aquatic Charles R. Greene Jr., David Kastak, Darlene R. Ketten, James H. Miller, Paul E. Nachtigall, W. John Richardson, Jeanette A. Thomas, & Peter L. Tyack Contents

Overview ...... 411 Chapter 1. Introduction ...... 415 Mammals Objectives...... 415 Historical Perspective ...... 416 Acoustic Measures and Terminology...... 417 Sound Production and Use in Marine Mammals ...... 419 Responses to Sound ...... 420 Chapter 2. Structure of the Noise Exposure Criteria ...... 427 Sound Types...... 427 Marine Mammal Functional Hearing Groups ...... 430 Exposure Criteria Metrics ...... 434 Levels of Noise Effect: Injury and Behavioral Disturbance ...... 436 Chapter 3. Criteria for Injury: TTS and PTS ...... 437 Effects of Noise on Hearing in Marine Mammals: TTS Data ...... 437 Injury from Noise Exposure: PTS-Onset Calculation ...... 441 Criteria for Injury from a Single Pulse ...... 442 Criteria for Injury from Multiple Pulses ...... 444 Criteria for Injury from Nonpulses ...... 444 Chapter 4. Criteria for Behavioral Disturbance ...... 446 Behavioral Response Data Analysis Procedures: Disturbance Criteria and Severity Scaling...... 448 Criteria for Behavioral Disturbance: Single Pulse...... 451 Behavioral Response Severity Scaling: Multiple Pulses ...... 452 Behavioral Response Severity Scaling: Nonpulses ...... 456 Chapter 5. Research Recommendations ...... 474 Measurements of Anthropogenic Sound Sources and Ambient Noise ...... 474 Marine Mammal Auditory Processes...... 474 Behavioral Responses of Marine Mammals to Sound...... 477 Effects of Noise Exposure on Marine Mammal Hearing and Other Systems ...... 478 Particularly Sensitive Species ...... 480 Necessary Progressions of Marine Mammal Noise Exposure Criteria ...... 481 Acknowledgments ...... 482 Literature Cited ...... 482 Marine Mammal Noise Exposure Criteria: Appendix A. Acoustic Measures and Terminology ...... 498 Initial Scientiﬁ c Recommendations Appendix B. Studies Involving Marine Mammal Behavioral Responses to Multiple Pulses ...... 502 Appendix C. Studies Involving Marine Mammal Behavioral Responses to Nonpulses ...... 509 Supported through Joint Sponsorship by the European Association for

Volume 33, Number 4, 2007 Aquatic Mammals, the Alliance of Marine Mammal Parks and Aquariums, Aquatic Mammals and the International Marine Animal Trainer’s Association Founded by EAAM in 1974 Papers dealing with all aspects of the care, conservation, medicine, and science of aquatic mammals

Editor and Copyright: Jeanette A. Thomas Department of Biological Sciences, Western Illinois University–Quad Cities, Moline, IL 61265, USA Co-Editor: Kathleen Dudzinski Dolphin Communication Project, Stonington, Connecticut 06378, USA

Prepared and published by Document and Publication Services, Western Illinois University, Macomb, IL 61455, USA Established by the EAAM in 1974 AQUATIC MAMMALS AQUATIC MAMMALS European Association for Aquatic Mammals The European Association for Aquatic Mammals (EAAM), the Alliance of Marine Board of the European Association for Aquatic Mammals Mammal Parks and Aquariums (AMMPA), and the International Marine Animal Trainer’s President: Niels van Elk, Dolfinarium Harderwijk, Strandboulevard Post 1, Association jointly sponsor Aquatic Mammals (ISSN 0167-5427), printed four times per year. 3841 AB, Harderwijk, The Netherlands; E-mail: [email protected] President Elect: Birgitta Mercera, Parc Asterix, BP 8, 60128 Plailly, France; Managing Editor E-mail: [email protected] Jeanette A. Thomas, Ph.D., Professor Past President (2004-2006): Pedro Roberto Lavia, Mundo Aquatico, Zoomarine, Estrada Nacional Department of Biological Sciences, Western Illinois University–Quad Cities, 125 KM 65, Guia, 8200 Albufeira, Portugal; E-mail: [email protected] 3561 60th Street, Moline, Illinois 61265, USA Secretary/Treasurer: Sabrina Brando, Animal Concepts, Sea Mammal Research Unit, Gatty Marine Laboratory, E-mail: [email protected], Tel: (309) 762-9481, ext. 311, Fax: (309) 762-6989 University of St. Andrews, St. Andrews, Fife, UK; E-mail: [email protected] Co-Editor Information Committee: Emily M. Walter, Western Illinois University–Quad Cities, Kathleen M. Dudzinski, Ph.D., Director Moline, Illinois 61265, USA; E-mail: [email protected] Dolphin Communication Project, 222 Wolf Neck Road, Stonington, Connecticut 06378, USA Alliance of Marine Mammal Parks and Aquariums E-mail: [email protected], Tel: (860) 514-4704, Manual Fax: (860) 536-1740 Executive Director: Marilee Menard Book Review Editor 418 North Pitt Street, Alexandria, Virginia 22314, USA; E-mail: [email protected] Justin Gregg 2007 Board of Directors School of Psychology, Áras an Phiarsaigh, Trinity College, Dublin 2, Ireland President: Dave Merritt, Indianapolis Zoo E-mail: [email protected], Tel: +31 (0)475-330006 (Netherlands), Immediate Past President: Jeff Jouett, Dolphin Quest Fax: +1-860-572-5973 (USA) Vice President: Clint Wright, Vancouver Aquarium Graduate Assistants to the Editors Treasurer: Bill Hughes, SeaWorld Orlando Emily M. Walter and Sara Crowell Secretary: John Rupp, Point Defiance Zoo and Aquarium Western Illinois University–Quad Cities, 3561 60th Street, Moline, Illinois 61265, USA Director-at-Large: Mark Swingle, Virginia Aquarium & Marine Science Center E-mail: [email protected] and [email protected] Director-at-Large: Dave Blasko, Six Flags Marine World Tel: (309) 762-9481, ext. 289, Fax: (309) 762-6989 International Marine Animal Trainers’ Association Editors Emeriti c/o John G. Shedd Aquarium, 1200 S. Lake Shore Drive, Chicago, Illinois, 60605, USA Willem H. Dudok van Heel, Victor J. A. Manton, and Paul E. Nachtigall 2007 Board of Directors Editorial Board President: Billy Hurley Arne Bjørge, Institute of Marine Research, Gaustadalleen 21, 03439 Oslo, Norway President Elect: Shelly Ballman Manuel E. dos Santos, Instituto Superior de Psicologia Aplicada, Lisboa, Portugal Past President: Al Kordowski Manuel García Hartmann, Duisburg Zoo, Duisburg, Germany Secretary: Michael Hunt Philip Hammond, Sea Mammal Research Unit, University of St Andrews, St Andrews, Fife, UK Treasurer: Traci Belting Heidi Harley, Division of Social Sciences, New College of Florida, Sarasota, Florida, USA First Vice President: Dave Roberts A. Rus Hoelzel, School of Biological and Biomedical Sciences, University of Durham, Durham, UK Second Vice President: Andrew Scullion Christina Lockyer, North Atlantic Marine Mammal Commission, Polar Environmental Centre, Tromsø, Norway Third Vice President: Mike Osborne Lee A. Miller, Institute of Biology, University of Southern Denmark, Odense, Denmark Paul E. Nachtigall, Hawaii Institute of Marine Biology, Kailua, Hawaii, USA For instructions to authors, abstracts of previous issues, and publication fees, see the journal website: Giuseppe Notarbartolo di Sciara, Tethys Research Institute, Milano, Italy www.aquaticmammalsjournal.org Dan Odell, Hubbs-Sea World Research Institute, Orlando, Florida, USA For information on EAAM, see website: Grey Stafford, Wildlife World Zoo, Litchfield Park, Arizona, USA http://eaam.org Publication and Administrative Services Document and Publication Services (DPS), Western Illinois University, Macomb, Illinois 61455, USA For information on the Alliance of Marine Mammal Parks and Aquariums, see website: www.ammpa.org To obtain hard copies, back issues, or CD versions, contact Gina Colley, DPS, Western Illinois University, Macomb, Illinois 61455, USA; E-mail: [email protected] For information on IMATA, see website: www.imata.org Cover Photo Credits C. Richter, Sperm Whale Seismic Study, Gulf of Mexico, U.S. Minerals Management Service (2002); For details about the online version of Aquatic Mammals, see the Ingenta website: J. Thomas, Antarctica (YEAR); J. Thomas, Antarctica (YEAR); J. Anderson, Patagonia (1999); J. Anderson, www.ingentaconnect.com Southern Ocean (2002); J. Anderson, Patagonia (1999); A. S. Friedlaender, Bay of Fundy (2002); J. Anderson, San Simeon, CA (1999); P. Scheifele, St. Lawrence Seaway (mid-1990s) Marine Mammal Noise Exposure Criteria i

Marine Mammal Noise Exposure Criteria: Initial Scientific Recommendations Brandon L. Southall,1, 2 Ann E. Bowles,3 William T. Ellison,4 James J. Finneran,5 Roger L. Gentry,6 Charles R. Greene Jr.,7 David Kastak,2 Darlene R. Ketten,8, 9 James H. Miller,10 Paul E. Nachtigall,11 W. John Richardson,12 Jeanette A. Thomas,13 and Peter L. Tyack8

1National Oceanic and Atmospheric Administration (NOAA), National Marine Fisheries Service, Office of Science and Technology, Marine Ecosystems Division, NOAA’s Ocean Acoustics Program, 1315 East-West Highway #12539, Silver Spring, MD 20910-6233, USA; E-mail: [email protected] 2Long Marine Laboratory, University of California at Santa Cruz, 100 Shaffer Road, Santa Cruz, CA 95060, USA 3Hubbs-Sea World Research Institute, 2595 Ingraham Street, San Diego, CA 92109, USA 4Marine Acoustics, Inc., 809 Aquidneck Avenue, Middletown, RI 02842, USA 5U.S. Navy Marine Mammal Program, Space and Naval Warfare Systems Center–San Diego, 53560 Hull Street, San Diego, CA 92152-5000, USA 6ProScience Consulting, LLC, P.O. Box 177, Dickerson, MD 20842-0177, USA 7Greeneridge Sciences, Inc., 4512 Via Huerto, Santa Barbara, CA 93110, USA 8Woods Hole Oceanographic Institution, Woods Hole, MA 02543, USA 9Harvard Medical School, Department of Otology and Laryngology, Boston, MA 02114, USA 10University of Rhode Island, Department of Ocean Engineering, South Ferry Road, Narragansett, RI 02882, USA 11Hawai’i Institute of Marine Biology, P.O. Box 1346, Kane’ohe, HI 96744, USA 12LGL Ltd., environmental research associates, P.O. Box 280, 22 Fisher Street, King City, ON L7B 1A6, Canada 13Western Illinois University–Quad Cities, Department of Biological Sciences, 3561 60th Street, Moline, IL 61265, USA

Contents Overview ...... 411

Chapter 1. Introduction ...... 415 Objectives ...... 415 Historical Perspective ...... 416 Acoustic Measures and Terminology ...... 417 Sound Production and Use in Marine Mammals...... 419 Responses to Sound ...... 420

Chapter 2. Structure of the Noise Exposure Criteria ...... 427 Sound Types...... 427 Marine Mammal Functional Hearing Groups ...... 430 Exposure Criteria Metrics ...... 434 Levels of Noise Effect: Injury and Behavioral Disturbance ...... 436

Chapter 3. Criteria for Injury: TTS and PTS ...... 437 Effects of Noise on Hearing in Marine Mammals: TTS Data ...... 437 Injury from Noise Exposure: PTS-Onset Calculation ...... 441 Criteria for Injury from a Single Pulse ...... 442 Criteria for Injury from Multiple Pulses ...... 444 Criteria for Injury from Nonpulses ...... 444

Chapter 4. Criteria for Behavioral Disturbance ...... 446 Behavioral Response Data Analysis Procedures: Disturbance Criteria and Severity Scaling...... 448 Criteria for Behavioral Disturbance: Single Pulse...... 451 Behavioral Response Severity Scaling: Multiple Pulses ...... 452 Behavioral Response Severity Scaling: Nonpulses ...... 456 ii Southall et al.

Chapter 5. Research Recommendations ...... 474 Measurements of Anthropogenic Sound Sources and Ambient Noise ...... 474 Marine Mammal Auditory Processes...... 474 Behavioral Responses of Marine Mammals to Sound...... 477 Effects of Noise Exposure on Marine Mammal Hearing and Other Systems ...... 478 Particularly Sensitive Species ...... 480 Necessary Progressions of Marine Mammal Noise Exposure Criteria ...... 481

Acknowledgments ...... 482

Literature Cited ...... 482

Appendix A. Acoustic Measures and Terminology ...... 498

Appendix B. Studies Involving Marine Mammal Behavioral Responses to Multiple Pulses ...... 502

Appendix C. Studies Involving Marine Mammal Behavioral Responses to Nonpulses ...... 509 Marine Mammal Noise Exposure Criteria iii

Acronyms

Acronym Definition

A-weighting Frequency-selective weighting for aerial hearing in humans derived from the inverse of the idealized 40-phon equal loudness hearing function across frequencies ABR Auditory brainstem response ADD Acoustic deterrent device AEP Auditory evoked potentials AHD Acoustic harassment device ANSI American National Standards Institute ASSR Auditory steady-state response ATOC Acoustic Thermometry of Ocean Climate program CF Center frequency C-weighting Frequency-selective weighting for aerial hearing in humans derived from the inverse of the idealized 100-phon equal loudness hearing function across frequencies EFR Envelope following response EPA U.S. Environmental Protection Agency ES Explosion simulator fhigh Estimated upper functional hearing limit flow Estimated lower functional hearing limit HESS High Energy Seismic Survey HPA Hypothalamic-pituitary-adrenal axis IMAPS Integrated Marine Mammal Monitoring and Protection System ISO International Standards Organization JNCC U.K. Joint Nature Conservation Committee

LeqT Equivalent-continuous sound level over period T

LIeqT Impulse equivalent-continuous sound level over period T LFA Low Frequency Active (sonar) M-weighting Generalized frequency weightings for various groups of marine mammals, allowing for their functional bandwidths and appropriate in characterizing auditory effects of strong sounds

Mlf Frequency weighting for low-frequency cetaceans (mysticetes)

Mmf Frequency weighting for mid-frequency cetaceans (most odontocetes)

Mhf Frequency weighting for high-frequency cetaceans (odontocetes specialized for use of very high frequencies)

Mpw Frequency weighting for pinnipeds, listening in water

Mpa Frequency weighting for pinnipeds, listening in air MMPA U.S. Marine Mammal Protection Act NIHL Noise-induced hearing loss NIPTS Noise-induced permanent threshold shift NIOSH U.S. National Institute for Occupational Safety and Health NMFS U.S. National Marine Fisheries Service iv Southall et al.

Acronym Definition

NOAA U.S. National Oceanic and Atmospheric Administration NRC U.S. National Research Council NRL U.S. Naval Research Laboratory

Pmax Maximum sound pressure OBN Octave-band noise PCAD National Research Council’s Population Consequences of Acoustic Disturbance Model PICE Porpoise incidental catch elimination PTS Permanent threshold shift REFMS A computer program for predicting shock-wave propagation from underwater explosions RL Received level RMS Root-mean-square SEL Sound exposure level SL Source level (received level measured or estimated 1 m from the source) SLM Sound level meter SPL Sound pressure level TS Threshold shift TTS Temporary threshold shift USC United States Code VAFB Vandenberg Air Force Base Aquatic Mammals 2007, 33(4), 411-414, DOI 10.1578/AM.33.4.2007.411

Overview

A group of experts in acoustic research from and quality. In many respects, data gaps severely behavioral, physiological, and physical disciplines restrict the derivation of scientifically-based noise was convened over a several year period. The pur- exposure criteria and, in some cases, explicit pose of this panel was to review the expanding lit- threshold criteria for certain effects are not appro- erature on marine mammal hearing and on physi- priate given the amount and type of data available. ological and behavioral responses to anthropogenic Scientific inquiry into acoustic communication sound, and to propose exposure criteria for certain among marine mammals extends back more than effects. The group employed all available relevant half a century, but most of the specific data rel- data to predict noise exposure levels above which evant to the proposed criteria have been published adverse effects on various groups of marine mam- within the last two decades. Owing to the mount- mals are expected. Recent advances in these fields ing public, scientific, and regulatory interest in and the pressing need for a science-based para- conservation issues related to acoustics, the avail- digm to assess the effects of sound exposure were able science is progressing rapidly (e.g., see NRC, the primary motivations for this effort. Two cat- 2003, 2005). egories of effects were considered: (1) injury and This paper proposes, for various marine mammal (2) behavioral disturbance. The proposed criteria groups and sound types, levels above which there for the onset of these effects were further segre- is a scientific basis for expecting that exposure gated according to the functional hearing capa- would cause auditory injury to occur. Controlled bilities of different marine mammal groups, and measurements of hearing and of the effects of according to the different categories and metrics underwater and aerial sound in laboratory settings of typical anthropogenic sounds in the ocean. The have greatly expanded the ability to assess audi- group achieved many of its objectives but acknowl- tory effects. While understanding of the hearing edges certain limitations in the proposed criteria capacities among all marine mammals remains because of scarcity or complete absence of infor- admittedly rudimentary, there is a fairly detailed mation about some key topics. A major component understanding of some key aspects of underwater of these recommendations is a call for specific and aerial hearing in a few representative species research on critical topics to reduce uncertainty of odontocetes, pinnipeds, and sirenians, although and improve future exposure criteria for marine hearing in mysticetes remains untested. Available mammals. This publication marks the culmination data, along with the compelling evidence of similar of a long and challenging initial effort, but it also auditory processes among all mammals, enables initiates a necessary, iterative process to apply and some reasonable extrapolations across species for refine noise exposure criteria for different species estimating auditory effects, including the exposure of marine mammals. levels of probable onset of injury. Recent evidence The process of establishing policy guidelines suggests that exposure of beaked whales to under- or regulations for anthropogenic sound exposure water noise may, under certain (generally unknown) (i.e., the application of these exposure criteria) will conditions, result in non-auditory injury as well vary among nations, jurisdictions, and legal/policy (e.g., Fernández et al., 2005). At present, however, settings. Such processes should carefully consider there are insufficient data to allow formulation of the limitations and caveats given with these pro- quantitative criteria for non-auditory injuries. posed criteria in deciding whether sufficient data There are many more published accounts of currently exist to establish simplistic, broad crite- behavioral responses to noise by marine mammals ria based solely on exposure levels. In many cases, than of direct auditory or physiological effects. especially for behavioral disturbance, context- Nevertheless, the available data on behavioral specific analyses considering previous studies on responses do not converge on specific exposure species and conditions similar to those in question conditions resulting in particular reactions, nor do might, at least for the foreseeable future, be more they point to a common behavioral mechanism. appropriate than general guidelines. Even data obtained with substantial controls, precision, and standardized metrics indicate high State of Current Knowledge variance both in behavioral responses and in expo- The available data on the effects of noise on sure conditions required to elicit a given response. marine mammals are quite variable in quantity It is clear that behavioral responses are strongly 412 Southall et al.

affected by the context of exposure and by the ani- and (5) pinnipeds listening in air (Mpa). These cri- mal’s experience, motivation, and conditioning. teria do not specifically address sirenians, the sea This reality, which is generally consistent with otter, or the polar bear, in part because of the lack patterns of behavior in other mammals (includ- of key data in these species. ing humans), hampered our efforts to formulate The M-weighting functions were defined based broadly applicable behavioral response criteria for on known or estimated auditory sensitivity at dif- marine mammals based on exposure level alone. ferent frequencies rather than vocal characteristics per se. Owing to the paucity of relevant data, these Frequency-Weighting Functions auditory functions are intentionally precaution- In humans, hearing processes in a large number ary (wide) and likely overestimate the functional of male and female subjects of different ages bandwidth for most or all species. Their primary have been tested to determine a basic audiomet- application is in predicting auditory damage rather ric curve, equal-loudness curve, and the levels and than levels of detection or behavioral response. exposure durations needed to induce either recov- Consequently, it is more appropriate to use “flat- erable hearing loss (called temporary threshold ter” functions than would be obtained by employ- shift or TTS) or permanent threshold shift (PTS). ing a simple inverse-audiogram function. In addition, the manner in which successive exposures to noise contribute to TTS growth has been Exposure Criteria Metrics well-documented in humans (e.g., Kryter, 1994; To further complicate the derivation of noise expo- Ward, 1997). In assessing the effects of noise sure criteria, sounds can be described with various on humans, either an A- or C-weighted curve acoustic metrics, including sound pressure levels is applied to correct the sound-level measure- and sound exposure levels. The latter is a measure ment for the frequency-dependent hearing func- of received sound energy. Available literature pro- tion of humans. Early on, the panel recognized vides a mixture of both measures, but many sound that similar, frequency-weighted hearing curves sources have primarily been described in pressure were needed for marine mammals; otherwise, level units. To accommodate these two measures, extremely low- and high-frequency sound sources and to account for all relevant acoustic features that are detected poorly, if at all, might be subject that may affect marine mammals, we developed to unrealistic criteria. dual criteria for noise exposures in each of the five One of the major accomplishments in this functional hearing groups, using both sound pres- effort was the derivation of recommended fre- sure and sound exposure levels. quency-weighting functions for use in assessing the effects of relatively intense sounds on hearing Exposure Criteria for Injury in some marine mammal groups. It is abundantly Another area in which we provide substantive clear from measurements of marine mammal hear- conclusions is in the determination of sound ing in the laboratory, call characteristics, and audi- exposures believed to cause direct auditory injury tory morphology that there are major differences to marine mammals. By all accounts, the inner in auditory capabilities across marine mammal ear is the organ system most directly sensitive to species (e.g., Wartzok & Ketten, 1999). Most pre- sound exposure and, thus, the most susceptible to vious assessments of acoustic effects either failed sound-derived damage. We define the minimum to account for differences in functional hearing exposure criterion for injury as the level at which bandwidth among marine mammal groups or did a single exposure is estimated to cause onset of not recognize that the “nominal” audiogram might permanent hearing loss (PTS). Data on TTS in be a relatively poor predictor of how the auditory marine mammals, and on patterns of TTS growth system responds to relatively strong exposures. and its relation to PTS in other mammals, were The authors delineated five groups of marine used to estimate thresholds for injury. Owing to mammals based on similarities in their hearing, and the limited availability of relevant data on TTS they developed a generalized frequency-weight- and PTS, the extrapolation procedures underlying ing (called “M-weighting”) function for each. these estimations are necessarily precautionary. The five groups and the associated designators are To account for all of the potentially injurious (1) mysticetes (baleen whales), designated aspects of exposure, dual criteria for injury were as “low-frequency” cetaceans (Mlf); (2) some established for each functional marine mammal odontocetes (toothed whales), designated as hearing group based on instantaneous peak pres- “mid-frequency” cetaceans (Mmf); (3) odontocetes sure (unweighted) and total energy (M-weighted). specialized for using high frequencies (i.e., por- Exposure criteria for injury are given for two types poises, river dolphins, and the genera Kogia and of sounds, pulse and nonpulse, and for single and Cephalorhynchus) (Mhf); (4) pinnipeds (i.e., seals, multiple exposures. The term pulse is used here to sea lions, and walruses) listening in water (Mpw); describe brief, broadband, atonal, transients (ANSI, Marine Mammal Noise Exposure Criteria 413

1986; Harris, 1998, Chapter 12), which are charac- to derive explicit and broadly applicable numeri- terized by a relatively rapid rise-time to maximum cal threshold values for delineating behavioral pressure followed by a decay that may include a disturbance. We did develop a quantitative scor- period of diminishing and oscillating maximal and ing paradigm that numerically ranks, as a severity minimal pressures. Examples of pulses are sounds scaling, behavioral responses observed in either from explosions, gunshots, sonic booms, seismic field or laboratory conditions. We applied this airgun pulses, and pile driving strikes. Nonpulse approach to the appropriate behavioral data for (intermittent or continuous) sounds can be tonal, multiple pulses and nonpulses. Some of these data broadband, or both. They may be of short dura- suffer from poor statistical power, limited infor- tion but without the essential properties of pulses mation on received sound levels and background (e.g., rapid rise-time). Examples of anthropogenic, noise, insufficient measurements of all potentially oceanic sources producing such sounds include important contextual variables, and/or insufficient vessels, aircraft, machinery operations such as controls. Some such data are analyzed here solely drilling or wind turbines, and many active sonar for illustrative purposes. Most behavioral studies systems. As a result of propagation, sounds with suffered from at least some of these problems. the characteristics of a pulse at the source may lose Therefore, we do not intend to give uniform sci- pulsatile characteristics at some (variable) distance entific credence to all of the cited data, and we and can be characterized as a nonpulse by certain expect future studies to give greater attention and receivers. rigor to these critical requirements. Regardless of the anthropogenic sound, if a This review and scoring process, while not a marine mammal’s received exposures exceed the formal meta-analysis for normalizing and pool- relevant (pulse or nonpulse) criterion, auditory ing disparate observations, corroborated certain injury (PTS) is assumed to be likely. Chapter 3, interesting aspects of marine mammal behavioral “Criteria for Injury,” provides details regarding responses to sound exposure. Foremost was that the exposure levels required to cause TTS-onset a behavioral response is determined not only by and the extrapolation of those results to estimate simple acoustic metrics, such as received level levels above which PTS-onset may occur. For all (RL), but also by contextual variables (e.g., labo- five functional hearing groups, we propose dual ratory vs field conditions, animal activity at the exposure criteria above which auditory injury is time of exposure, habituation/sensitization to likely. the sound, etc.). Also important is the presence or absence of acoustic similarities between the Exposure Criteria for Behavior anthropogenic sound and biologically relevant One challenge in developing behavioral criteria natural signals in the animal’s environment (e.g., is to distinguish a significant behavioral response calls of conspecifics, predators, prey). Within from an insignificant, momentary alteration in certain similar conditions, there appears to be behavior. For example, the startle response to a some relationship between the exposure RL and brief, transient event is unlikely to persist long the magnitude of behavioral response. However, enough to constitute significant disturbance. Even in many cases, such relationships clearly do not strong behavioral responses to single pulses, other exist, at least when response data are pooled than those that may secondarily result in injury across multiple species and contexts. This argues or death (e.g., stampeding), are expected to dis- for a context-based approach to deriving noise sipate rapidly enough as to have limited long-term exposure criteria for behavioral responses. That consequence. Consequently, upon exposure to a concept, along with our review and scaling of the single pulse, the onset of significant behavioral available observational data, provides a founda- disturbance is proposed to occur at the lowest tion for establishing dose-response relationships level of noise exposure that has a measurable for some specific circumstances and a starting transient effect on hearing (i.e., TTS-onset). We point for future analyses when additional data are recognize that this is not a behavioral effect per available. se, but we use this auditory effect as a de facto behavioral threshold until better measures are Conclusions and Research Recommendations identified. Lesser exposures to a single pulse are This process has resulted in several significant not expected to cause significant disturbance, advances. These include a review and interpre- whereas any compromise, even temporarily, to tation of the available literature on injury and hearing functions has the potential to affect vital behavioral data using precautionary extrapola- rates through altered behavior. tion procedures, derivation of marine mammal For other anthropogenic sound types (multiple frequency-weighting functions, specification of pulses, nonpulses), we conducted an extensive quantitative criteria for auditory injury, and deriva- review of the available literature but were unable tion of a “severity scale” for behavioral responses. 414 Southall et al.

The inability to identify broadly applicable, constraining important human activities in the quantitative criteria for behavioral disturbance in oceans will continue to be challenging for the response to multiple-pulse and nonpulse sounds is foreseeable future. With sustained and focused an acknowledged limitation. research in key areas, future scientists will be Our efforts to derive marine mammal noise equipped to make informed improvements to the exposure criteria clearly illustrate the fact that, initial scientific recommendations presented here. at present, research in this field remains limited These improvements should ideally be integrated in many areas. The need for extrapolation pro- into science-based risk assessment models that cedures and precautionary assumptions points consider all aspects of sound exposure and other directly to research needs in a variety of areas on a potential stressors on individual marine mammals, variety of species. In certain conditions, proposed populations, and marine ecosystems. criteria for an entire marine mammal group are based on the most precautionary measurement or observation for a species within that group, despite the fact that, for other species within that group, there are empirical data indicating that higher exposures are required to induce the same effect. We believe it is appropriate to use the most precautionary data in proposing group-wide criteria applicable for species where there are no direct measurements. We also feel it is appropriate on a case-by-case basis to apply the most relevant empirical data (i.e., from the species or genus of concern) in setting the exposure thresholds specified in policy guidelines. Finally, we emphasize that exposure criteria for single individuals and relatively short-term (not chronic) exposure events, as discussed here, are insufficient to describe the cumulative and ecosystem-level effects likely to result from repeated and/or sustained human input of sound into the marine environment and from potential interactions with other stressors. Also, the injury criteria proposed here do not appear to predict what may have been indirect injury from acoustic exposure in several cases where cetaceans of several species mass-stranded following exposure to military sonar. The extensive research recommendations given here (see Chapter 5) represent our collective view of the concerted effort that will be required over the coming decades. High priority categories of research include (1) continued expansion of knowledge on basic marine mammal hearing capabilities, including sound localization, the detection of realistic sound signals, communication masking, and auditory “scene analysis”; (2) continued expansion of knowledge on baseline marine mammal behavioral patterns; (3) well- controlled, direct measurements (using appropriate, standardized acoustic metrics) of the effects of sound exposure on marine mammal hearing, behavior, and physiology; and (4) risk-assessment studies of the cumulative and synergistic effects of noise and other exposure(s) on individuals and populations. Understanding and managing the effects of noise on marine life without unjustifiably Aquatic Mammals 2007, 33(4), 415-426, DOI 10.1578/AM.33.4.2007.415

1. Introduction

Objectives legal considerations limit the level of scientific information that is available for deriving criteria Recent interest and concern about the effects applicable to either humans or marine mammals. of anthropogenic noise on marine mammals Consequently, certain assumptions and criteria has triggered considerable new research (e.g., proposed here were based on information from Costa et al., 2003; Fristrup et al., 2003; Finneran other mammalian groups, where justified. Where et al., 2005a), summaries of available information such data present a variety of options, we made (Richardson et al., 1995; Wartzok & Ketten, 1999), intentionally precautionary decisions (i.e., lower and recommendations for specific action (NRC, proposed exposure levels) to reduce the risk of 1994, 2000, 2003, 2005). Systematic, objective, assuming no effect when one was actually present. science-based interpretation of the available data is The term “precautionary” is used here without ref- critically needed to inform management agencies erence to any regulatory or policy implication of charged with mitigating adverse effects of anthro- this word. Scientists would more conventionally pogenic noise on protected species. In response to use the term “conservative” in this regard rather this need, we use here the full body of scientific than the more bureaucratic “precautionary,” but in data on marine mammal hearing and the effects of certain complex instances here, the term “conser- noise on hearing and behavior, augmented where vative” would be potentially ambiguous, depend- appropriate by interpretations of terrestrial mammal ing on the perspective of the reader. When infor- (including human) data, to develop proposed expo- mation was limited, extrapolations were made sure criteria that are as comprehensive, defensible, cautiously to minimize the risk of failing to recog- and precise as is currently possible. The scope of nize an effect when one actually occurs (Type-II these criteria includes injurious and behavioral statistical error) as can occur with small sample effects of a single noise exposure event on an indi- sizes or imprecise measurements. vidual cetacean (whales, dolphins, and porpoises) Each generalization/extrapolation was identi- or pinniped (seals, sea lions, and walruses). fied, all precautionary decisions were noted, and The recommended noise exposure criteria are the logic leading to each proposed criterion was science-based, developed without addressing the specified. Thus, when new data become available, commercial, societal, or practical ramifications appropriate modifications can be made readily. of implementing the conclusions reached here. Studies that are needed to resolve the uncertain- We intend to mirror the process used in the devel- ties encountered in developing the current criteria opment of damage risk criteria for humans (see are discussed in detail (see Chapter 5, “Research Crocker, 1997). Policy “guidelines” developed for Recommendations”). Realistically, however, the regulatory and societal purposes are based both on generalization of information between related scientific evidence (as summarized in this paper species will remain essential in many cases for the for marine mammals) and on other considerations foreseeable future. (e.g., economic, practical, social, and ethical) not Our intent was to derive recommended noise dealt with here. Thus, on certain points, policy exposure criteria using the best information cur- guidelines that are developed separately for the rently available, identify weaknesses in the present purposes of various jurisdictions, nations, or users approach, call for relevant research, and structure of these criteria may differ from the science-based the criteria such that future improvements can be criteria recommended here. incorporated easily. Lack of data limited the pro- All forms of anthropogenic noise received by posed noise exposure criteria to individual marine marine mammals were considered, whether pro- mammals exposed to acute exposure events (such duced under water or in air, and we adopted a as the passage of one vessel or a series of active comparative approach, which we regard as essen- sonar transmissions). Also, the proposed criteria tial to any criteria-setting process for nonhuman are limited to cetaceans and pinnipeds. We expect animals. For most of the ~128 marine mammal that noise exposure criteria for other marine species and subspecies (Rice, 1998) considered mammals (manatees, dugongs, polar bears, and here, no empirical data were available on nominal sea otters), as well as other marine taxa, will be hearing characteristics or on the effects of noise developed as additional data become available and on hearing or behavior. Practical, ethical, and are evaluated. In fact, a separate expert panel (S3/ 416 Southall et al.

WG92: “Effects of Sound on Fish and Turtles”) has et al., 1988). In contrast, early observations of been established under the Standards Committee bowhead and gray whales exposed to continu- (S3) of the Acoustical Society of America to con- ous industrial sounds, such as those associated sider noise exposure criteria for fish and turtles. with drilling operations, suggested 120 dB re: Additionally, criteria are clearly needed for cumu- 1 µPa as the approximate threshold for behavioral lative effects and for effects at species or even eco- disturbance of these baleen whales (Malme et al., system levels, but data to support those types of 1984; Richardson et al., 1990a, 1995 [pp. 286- criteria do not currently exist. 287]). Significant individual variability was noted The present recommended criteria represent a in “typical” behavioral responses, however, with major step in initiating a lengthy, systematic pro- some individual whales responding only when cess to predict and identify acoustic exposure con- very close to sound sources and others reacting ditions (natural or anthropogenic) associated with at much longer distances (and to lower received various effects on marine mammals. This paper is sound levels). This variability raises questions as deliberately structured in a somewhat formulaic to whether behavioral responses are most appro- and report-like manner so that the logic underly- priately described by the exposure received level ing certain assumptions and extrapolations (as (RL) of the stimulus at the animal, the signal-to- well as the data needed to test and/or strengthen ambient noise differential, the rate of change of them) is self-evident. We expect there will be an the signal, or simply to the presence of the human iterative process of improving and expanding the activity as indicated by acoustic cues and/or visual complexity of the exposure criteria, similar to the stimuli. decades-long development of human noise expo- Concern about the effects of acoustic pulses sure criteria (see Crocker, 1997). Because of the from seismic exploration and continuous sound matrix structure of the proposed criteria, thresh- from other industrial activities resulted in the olds in specific cells can be updated independently imposition of mitigation requirements on some as new information becomes available. industrial activities in certain jurisdictions by There is an extensive history and diversity of the early- to mid-1980s. Subsequent events, exposure criteria for humans with various kinds of such as the Heard Island Feasibility Test in acoustic exposure. A full discussion of these crite- 1991 (Baggeroer & Munk, 1992), the Acoustic ria is beyond the scope of this paper, but examples Thermometry of Ocean Climate (ATOC) pro- include workplace noise standards (e.g., NIOSH, gram in the late-1990s (see NRC, 1994, 2000; Au 1998), standards for the protection of military et al., 1997; Costa et al., 2003), and the U.S. personnel (U.S. DoD, 1997), and national policy Navy’s low-frequency active sonar program (e.g., guidelines (e.g., EPA, 1974; BG PPG, 1994). Croll et al., 2001) resulted in popular and govern- Several additional examples were also considered, mental interest in setting criteria for safe levels of whether received under water or in air, in various sound for marine mammal exposure (NRC, 1994, decisions underlying the marine mammal criteria 2000, 2003; Richardson et al., 1995). This interest proposed here. The process of establishing human has expanded with the finding that tactical, mid- noise exposure criteria has been difficult and con- frequency, military sonar transmissions are some- tentious, but establishing noise exposure criteria times correlated, in specific conditions, with mass for marine mammals is considerably more daunt- stranding events of (predominantly) several beaked ing given the diversity of marine mammal species whale species, including Cuvier’s (Ziphius caviro- across three orders, the complexity of aerial and stris), Blainville’s (Mesoplodon densirostris), and underwater acoustic exposures, and profound data Gervais’ (Mesoplodon europaeus) beaked whales limitations. (see Evans & England, 2001; Fernández et al., 2005; Cox et al., 2006). Historical Perspective In 1995, the U.S. National Marine Fisheries Service (NMFS) set underwater “do not exceed” Concerns about potential adverse effects of anthro- criteria for exposure of marine mammals to pogenic noise on marine life began in the 1970s underwater pulses from seismic airguns. These (e.g., Payne & Webb, 1971) and expanded in the criteria were 190 dB re: 1 µPa for pinnipeds and 1980s. Experiments during the 1980s with seismic most odontocete cetaceans and 180 dB re: 1 µPa airguns indicated that bowhead whales (Balaena for mysticetes and sperm whales (Physeter mac- mysticetus) and gray whales (Eschrichtius robus- rocephalus) (and, by inference, for pygmy and tus) exhibited clear, sustained avoidance of opera- dwarf sperm whales [Kogia spp.]). These exposure tional areas at distances where pulse root-mean- limits were intended as precautionary estimates of square (RMS) sound pressure levels (SPLs) were exposures below which physical injury would not 160 to 170 dB re: 1 µPa (Malme et al., 1983, 1984, occur in these taxa. There was no empirical evi- 1986, 1988; Richardson et al., 1986; Ljungblad dence as to whether exposure to higher levels of Marine Mammal Noise Exposure Criteria 417 pulsed sounds would or would not cause auditory of our severity scaling paradigm would almost or other injuries. Given the limited data then avail- certainly not constitute biologically significant able, however, it could not be guaranteed that disturbance (or consequently level B harassment marine mammals exposed to higher levels would under the MMPA). However, our exposure criteria not be injured. Further, it was recognized that were derived without regard for policy decisions behavioral disturbance could, and in some cases of the U.S. or any nation and should therefore not likely would, occur at lower RLs. be assumed to correspond with regulatory catego- In June 1997, the High Energy Seismic Survey ries or definitions of effects. Since harassment (HESS) team (1999, Appendix 5) convened a definitions under the MMPA are not uniform for panel of experts to assess noise exposure criteria all human activities and are subject to change, for marine mammals exposed to seismic pulses. additional interpretation of the information pre- The consensus was that, given the best available sented would be required to evaluate effects with data at that time, exposure to airgun pulses with regard to this (or any other) statute. RLs above 180 dB re: 1 µPa (averaged over the pulse duration) was “likely to have the potential to Acoustic Measures and Terminology cause serious behavioral, physiological, and hearing effects.” The panel noted the potential for ± 10 This section briefly considers those acoustic mea- dB variability around the 180 dB re: 1 µPa level, sures and terminology that are directly relevant depending on species, and that more information to these marine mammal exposure criteria. More was needed. detailed descriptions of some of the terms given The NMFS has continued to use a “do not in this and other sections, including equations exceed” exposure criterion of 180 dB re: 1 µPa for relevant to many of the definitions, are given in mysticetes and (recently) all odontocetes exposed Appendix A. Basic acoustic terminology is pre- to sequences of pulsed sounds, and a 190 dB re: sented in numerous other sources (e.g., Kinsler 1 µPa criterion for pinnipeds exposed to such et al., 1982; ANSI, 1986, 1994; Richardson et al., sounds. Higher thresholds have been used in the 1995; Harris, 1998; NRC, 2003). U.S. for single pulses such as explosions used in Sound is appropriately described as having two naval vessel-shock trials. Behavioral disturbance components: (1) a pressure component and (2) a criteria for pulsed sounds have typically been set particle motion component. Particle motion—the at an SPL value of 160 dB re: 1 µPa, based mainly oscillatory displacement, velocity, or acceleration on the earlier observations of mysticetes reacting of the actual “particles” of the medium at a par- to airgun pulses (e.g., Malme et al., 1983, 1984; ticular location—is directional and best described Richardson et al., 1986). The relevance of the 160 by a 3-dimensional vector. Marine mammal sen- dB re: 1 µPa disturbance criterion for odontocetes sitivity to particle motion is poorly understood, and pinnipeds exposed to pulsed sounds is not but it appears to be functionally limited (Finneran at all well-established, however. Although these et al., 2002a) in contrast to the sensory capabili- criteria have been applied in various regulatory ties of most or all fish (see Popper et al., 2003). actions (principally in the U.S.) for more than a Conversely, as compared to fish, marine mammals decade, they remain controversial, have not been generally have greater sensitivity to sound pres- applied consistently in the U.S., and have not been sure (lower detection thresholds) and much wider widely accepted elsewhere. functional hearing bandwidths (see Fay, 1988; More recently, a considerable body of data has Richardson et al., 1995; Popper et al., 2003). accumulated on the levels at which transient and Consequently, in considering the potential effects more prolonged sounds cause the onset of tempo- of sound on marine mammals, particle motion is rary threshold shift (TTS) and various behavioral rarely discussed. Except for special circumstances reactions. Some of these data are not consistent (e.g., plane and spherical waves), there is no with the aforementioned de facto criteria used in simple relationship between pressure and particle recent years in the United States. velocity. The vast majority of studies of hearing One main purpose of this paper is to synthe- in captive marine mammals have been conducted size and apply all available information to derive in relatively small enclosed volumes of water, proposed objective noise exposure criteria for a making the plane wave assumption (and a priori large subset of marine mammals. The effect levels knowledge of the relationship between pressure considered (injury and significant behavioral and velocity) invalid. disturbance) were generally consistent with the It is important to distinguish between the source definitions of levels A and B harassment, respec- level (SL), or level measured 1 m from the source, tively, of the U.S. Marine Mammal Protection Act vs the received level (RL), which is the level mea- (MMPA) of 1972 (16 USC, § 1361); however, sured at the receiver (usually a marine mammal many of the behaviors considered at the lower end herein). 418 Southall et al.

The term “intensity” is often used generally time integral of the squared-instantaneous sound with respect to subjective acoustic parameters pressure normalized to a 1-s period. It can be an (i.e., loudness), but it is used here in a strict sense. extremely useful metric for assessing cumulative Sound intensity is normally defined as the time- exposure because it enables sounds of differing averaged active intensity (Kinsler et al., 1982; duration, sometimes involving multiple expo- Fahy, 1995); this quantity corresponds to local sures, to be compared in terms of total energy. net transport of sound energy and is related to Several methods exist for summing energy over the product of the sound pressure and the particle multiple exposures to generate a single exposure velocity component in-phase with the sound pres- “equivalent” value. The relatively straightforward sure. In the majority of laboratory studies, complex approach used here is described in Appendix A (eq. sound fields typically create complex, spatially 5). This summation procedure essentially generates varying relationships between pressure and veloc- a single exposure “equivalent” value that assumes ity. In these circumstances, sound intensity cannot no recovery of hearing between repeated expo- be estimated from pressure measurements alone sures. As discussed below, recovery functions for (which assume that pressure and particle velocity marine mammal TTS during and following multi- are in-phase), and specific measurements of the ple exposures are still unknown; however, consid- sound particle velocity (or pressure gradient) are ering nominal TTS recovery functions in terrestrial required in order to characterize intensity. mammals when exposures occur minutes to hours We distinguished two basic sound types: apart (see Kryter, 1994; Ward, 1997), the above (1) pulse and (2) nonpulse. Our operational defi- summation procedure would likely overestimate nitions of sound types are given in Chapter 2, the effect of multiple exposures in many condi- “Structure of the Noise Exposure Criteria,” and tions. This summation procedure was intentionally are discussed at greater length in Appendix A. The selected as a precautionary measure in the absence pulse/nonpulse distinction is important because of empirical information, although note the tem- pulses generally have a different potential to cause poral conditions given in the “Sound Types” sec- physical effects, particularly on hearing (e.g., tion of Chapter 2. The appropriate units are dB re: Ward, 1997). 1 µPa2-s for underwater SEL and dB re: (20 µPa)2-s Peak sound pressure (Pmax) is the maximum for aerial SEL. absolute value of the instantaneous sound pressure Frequency-selective weighting is often employed during a specified time interval and is denoted in to measure (as a single number) sound pressure or units of Pascals (Pa). It is in no sense an averaged energy in a specific frequency band of sound, with pressure. Peak pressure is a useful metric for either emphasis or de-emphasis on particular frequencies pulse or nonpulse sounds, but it is particularly as a function of the relative sensitivity of a receiver. important for characterizing pulses (ANSI, 1986; For aerial hearing in humans, A-weighting is derived Harris, 1998, Chapter 12). Peak-to-peak sound from the inverse of the idealized 40-phon equal pressure is the algebraic difference between the loudness hearing function across frequencies, stan- maximum positive and maximum negative instan- dardized to 0 dB at 1 kHz (Harris, 1998). This pro- taneous peak pressure. The mean-squared pres- vides level measures denoted as dB(A). C-weight- sure is the average of the squared pressure over ing is determined from the inverse of the idealized some duration. Sound pressure levels are given as 100-phon equal loudness hearing function (which the decibel (dB) measures of the pressure metrics differs in several regards from the 40-phon func- defined above. The RMS SPL is given as dB re: tion), standardized to 0 dB at 1 kHz (Harris, 1998). 1 µPa for underwater sound and dB re: 20 µPa This provides level measures denoted as dB(C). In for aerial sound. Peak sound pressure levels are the absence of equal-loudness contours for marine denoted hereafter as dB re: 1 µPa (peak) in water mammals, special frequency-weighting functions and dB re: 20 µPa (peak) in air. Peak-to-peak based loosely on human C-weighting and general sound pressure levels are dB re: 1 µPa (peak-to- knowledge of functional hearing bandwidth were peak) in water and dB re: 20 µPa (peak-to-peak) developed here for functional marine mammal hear- in air. ing groups (see the “Marine Mammal Functional Duration is the length of a sound in seconds. Hearing Groups” section of Chapter 2). Duration is important because it affects other Other measures of noise interference with sound measures, specifically mean-square and/or critical functions in humans, including the RMS sound pressure (Madsen, 2005). Because of Articulation Index (French & Steinberg, 1947) background noise and reverberation, duration can and the more recent Speech Interference Level be difficult to specify precisely, but a functional (see Beranek & Ver, 1992), focused on the percep- definition (see Appendix A) is used here. tion of speech and effects of noise. Consequently, Sound exposure level (SEL) is a measure of exposure criteria geared toward speech percep- energy. Specifically, it is the dB level of the tion (e.g., Beranek, 1989) focus on a frequency Marine Mammal Noise Exposure Criteria 419 bandwidth narrower than the audible bandwidth. and behavioral responses to natural (including For a detailed discussion of speech intelligibility conspecific) and anthropogenic sounds, a more and noise impacts, see Chapter 6 in Kryter (1994). refined means of frequency-weighting than the It is clear that the perception of conspecific vocal intentionally precautionary (broad) M-weighting signals in marine mammals is critically important functions may be recommended. in various life history functions (discussed below; Kurtosis is a statistical measure of a probability see Wartzok & Ketten, 1999) and that interference distribution often applied to describe the shape of with these functions may have particularly nega- the amplitude distribution (Hamernik & Hsueh, tive consequences. 1991; Lei et al., 1994; Hamernik et al., 2003). In The hypothesis that vocalizations coincide some regards, it appears to be a highly relevant with the range of hearing is based on an adaptive metric in that impulsive sound with high nega- argument that vocal energy should be selected to tive kurtosis, rapid onset, and high instantaneous lie within the range of hearing for maximum effi- peak-pressure may be particularly injurious to ciency of communication. However, several lines some mammals (Hamernik et al., 2003). of evidence suggest that other adaptive pressures may shape the vocal range. First, vocal anatomy Sound Production and Use in Marine Mammals may produce energy at other frequencies as a byproduct of producing sound within the hearing As a general statement, all studied marine mam- range. If there is no pressure to eliminate these mals can produce sounds in various important con- frequencies, they can be expected to persist. An texts. They use sound in social interactions as well example is the ultrasonic components of humming- as to forage, to orient, and to respond to predators. bird song, which lie well outside the range of bird Interference with these functions, through the var- hearing (Pytte et al., 2004). Second, to promote ious effects of noise on hearing and/or behavior long-range transmission, the vocal range may be identified below, thus has the potential to interfere adapted to produce greater energy at the low end with vital rates identified by the NRC (2005) as of the range than would be expected based on the particularly significant effects of exposure. auditory threshold function (Larom et al., 1997). The noise exposure criteria given here are Greater relative energy at low frequencies is also focused on current knowledge of hearing and seen in a number of primate species as a byprod- the effects of noise on hearing and/or behavior in uct of producing the formant structure of their calls marine mammals. Thus, a detailed discussion and (Fitch & Hauser, 1995). Finally, animals may pro- review of the expansive literature on the produc- duce sounds with disproportionate low-frequency tion and the uses of sound is beyond the scope of information to signal greater size, potentially tar- this paper; interested readers are referred to the geting predators rather than conspecifics (Fitch, many reviews of marine mammal acoustic signals 1999; Matrosova et al., 2007). Thus, a number of (e.g., Schusterman, 1981; Watkins & Wartzok, selective forces can drive the development of an 1985; Au, 1993; Richardson et al., 1995; Wartzok emphasis on low-frequency energy in vocalizations & Ketten, 1999; Clark & Ellison, 2004). Because not matched by the shape of the auditory threshold of the extreme importance of detecting conspecific function. While vocal range can be expected to cor- social signals in marine mammal life history func- relate with hearing range to some degree, giving a tions, however, a brief and very general discus- rough indication of the frequency range of hearing, sion of sound output characteristics in the major it cannot be used to estimate either the shape of the marine mammal groups is given here. auditory threshold function or to assign upper and The large whales (mysticete cetaceans, as lower frequency limits. described below) generally produce low-fre- We lack sufficient empirical data on whether quency sounds in the tens of Hz to the several kHz vocal frequency range sufficiently predicts all band, with a few signals extending above 10 kHz. frequencies that are biologically significant, These sounds appear to serve predominantly social however. functions, including reproduction and maintaining Certain marine mammal responses to anthropo- contact, but they may also play some role in spa- genic sounds, such as the sometimes strong reac- tial orientation. tions by beaked whales to mid-frequency sonar, The dolphins and porpoises (odontocete ceta- would not be expected if only sounds within the ceans, also described below) produce sounds bandwidth of vocal output were important in pre- across some of the widest frequency bands that dicting a behavioral response. Hence, our precau- have been observed in animals. Their social sounds tionary frequency-weighting approach assumes are generally in the range audible to humans, from that the full audible band is relevant. As additional a few hundreds of Hz to several tens of kHz, but data become available on both hearing capa- specialized clicks used in biosonar (echolocation) bilities (specifically, equal-loudness contours) 420 Southall et al. systems for prey detection and navigation extend Audibility well above 100 kHz. When a sound can be perceived amidst background Pinnipeds (seals, sea lions, and walruses) also noise, it is considered to be audible. Audibility can produce a diversity of sounds, though generally differ from detectability in that a receiving system over a lower and more restricted bandwidth (gen- may detect a signal at some level even when it is erally from 100 Hz to several tens of kHz). Their incapable of meaningful perception. Audibility sounds are used primarily in critical social and is determined by the characteristics of received reproductive interactions. Pinnipeds spend time sound, characteristics of the receiving system, and both at sea and on land, however, and thus pro- background noise conditions (either external or duce sounds in both water and air. internal). Audition (hearing) is a well-developed Because sound production in marine mam- and primary sensory modality for most, if not all, mals is integral to so many important behaviors, marine vertebrates (Schusterman, 1981; Tyack, interference with these communicative functions 1998; Fay & Popper, 2000). It involves coding, is considered to be particularly adverse (see sever- processing, integrating, and responding to sound ity scaling described in Chapter 4, “Criteria for in a variety of ways, some not outwardly evident Behavioral Disturbance”). As discussed in Chapter (Yost, 2000). Like other animals, marine mam- 5, considerable additional research is needed to mals have multiple sound-reception pathways and identify conditions in which anthropogenic noise rely on signal processing at multiple levels inte- exposure interferes with acoustic communication grated within the cochlea and nervous system to as well as ways in which marine mammals cope optimize perception. with masking noise to overcome interference in Marine mammal hearing capabilities are detecting real-world signals in complex, 3-dimen- quantified in live subjects using behavioral audi- sional marine environments. ometry and/or electrophysiological techniques (e.g., Schusterman, 1981; Au, 1993; Kastak & Responses to Sound Schusterman, 1998; Wartzok & Ketten, 1999; Nachtigall et al., 2000, 2007; Finneran & Houser, Animals exposed to either natural or anthropo- 2006; André & Nachtigall, 2007; Supin & Popov, genic sound may experience physical and psycho- 2007). For species not studied with in vivo audi- logical effects, ranging in magnitude from none to ometry, some auditory characteristics can be esti- severe. This brief discussion considers the range mated based on sound production frequencies; on of potential impacts, which depend on spatial rela- observations of sound characteristics that either do tionships between a sound source and the animal or do not elicit behavioral responses in untrained receiver; sensitivity of the receiver; received expo- animals (e.g., Richardson et al., 1995; Erbe, 2002); sure level, duration, and duty cycle; and many or on auditory morphology, including biomechan- other factors (see also Richardson et al., 1995). ical properties of the basilar membrane and other The same acoustic source may have radically characteristics (Wartzok & Ketten, 1999). different effects depending on operational and Behavioral audiograms are obtained from cap- environmental variables, and on the physiological, tive, trained animals using standard psychometric sensory, and psychological characteristics of testing procedures. With appropriate controls and exposed animals. It is important to note that these sufficient training, behavioral data are presently animal variables may differ (greatly in some cases) considered to most accurately represent hearing among individuals of a species and even within capabilities of a test subject. Behavioral audio- individuals depending on various factors (e.g., metric studies are time-consuming, however, and sex, age, previous history of exposure, season, and the results depend on the training and attention of animal activity). Responses elicited can depend subjects as well as the background noise condi- both on the context (feeding, mating, migrating, tions in captive settings. Because marine mam- etc.) in which an individual is ensonified and mals are large and difficult to maintain, behav- on a host of experiential variables (see Wartzok ioral audiograms representing an entire species et al., 2004). Consequently, certain effects may are typically based on a few individuals (often be poorly described with simple measures such one animal). Additionally, subjects are generally as SPL alone, and may only be predictable when obtained opportunistically (e.g., individuals reha- additional variables are considered. We consid- bilitated after stranding) rather than by random ered all known factors in developing the noise sampling of individuals from wild populations. exposure criteria proposed here, but data limita- This may provide a somewhat biased representa- tions precluded the derivation of explicit exposure tion of “normal” hearing for the species if reha- criteria for all of the effects discussed below. bilitated animals have compromised hearing capabilities (see André et al., 2007). Individual differences in hearing sensitivity among subjects, Marine Mammal Noise Exposure Criteria 421 and methodological differences among investiga- (lowest hearing thresholds) and frequencies both tors, can lead to improper conclusions when nom- below and above this range where sensitivity is inal species audiograms are based on data from relatively poor (higher threshold values). Species a single animal (e.g., compare Hall & Johnson, differ in absolute sensitivity and functional fre- 1972, with Szymanski et al., 1999). Hearing sen- quency bandwidth (see Fay, 1988; Richardson sitivity has been measured using behavioral meth- et al., 1995), such that identical sounds may be ods in fewer than 20 of the ~128 cetacean and perceived radically differently by individuals of pinniped species (based on the taxonomy of Rice, different species. Individual differences within 1998). species have also been demonstrated in some ter- Electrophysiological audiometry involves mea- restrial species (see Fay, 1988) and, to a lesser suring small electrical voltages (auditory evoked extent, in marine mammals as well (see Houser & potentials [AEPs]) produced by neural activity Finneran, 2006b, for the most definitive example when the auditory system is stimulated by sound. of this). Sounds whose levels barely exceed back- With this technique, neural responses are typi- ground noise levels may be detectable but may cally averaged while many relatively short dura- or may not elicit changes in individual behavior. tion signals are presented. This technique is com- Ideally, “absolute” or unmasked hearing thresh- paratively fast and less sensitive to factors such as olds should be measured in low background noise subject experience and reproductive, behavioral, conditions such as anechoic testing enclosures. or motivational states that affect behavioral audi- While this is standard practice in human audiometry. Whereas behavioral audiograms can only ometry, very few of the marine mammal hearing be made with trained, captive animals, AEP mea- data obtained to date have been measured in such sures of sound detection can also be made with conditions. Limited recent data obtained with pin- untrained individuals that are stranded, tempo- nipeds tested in a hemi-anechoic testing chamber rarily restrained, or in rehabilitation (see Cook in air (described in Kastak et al., 2005) suggest et al., 2006; André et al., 2007; Delory et al., 2007; that masking from environmental noise in testing Taylor et al., 2007). enclosures may have significantly affected mea- AEP and behavioral techniques measure differ- surements of “absolute” hearing; thresholds in a ent features of the auditory system and may gener- harbor seal (Phoca vitulina) were in fact ≥ 30 dB ate somewhat different measured results. Relevant lower in very low background noise conditions comparisons of AEP and behavioral audiograms (Holt et al., 2001). are limited and are the subject of ongoing scien- While the above concepts and studies are essen- tific investigation. Besides the need to obtain both tial in understanding general hearing capabilities types of data on the same individuals, there are (e.g., functional bandwidth, range of best hearing complications due to differences in the types of sensitivity) of marine mammals, animals in the test stimuli used by different researchers, prob- “real world” rarely listen for simple acoustic sig- lems in estimating the true RL at the relevant nals from point sources and do not live in a noise- sensory organ(s), and the difficulty of determin- controlled environment. Rather, they are presented ing absolute signal amplitudes that barely elicit with spatially complex and time-varying streams neural responses. Even so, Yuen et al. (2005), of acoustic information in often noisy environ- Finneran et al. (2007b), and Schlundt et al. (2007) ments. Measurements using simple sound stimuli demonstrated that, with carefully calibrated and have indicated that marine mammals are generally repeated measurements, the two procedures can quite adept at localizing acoustic sources in labo- produce comparable detection thresholds in at ratory conditions (Møhl, 1964; Gentry, 1967; least a few cetacean species. Terhune, 1974; Moore & Au, 1975; Renaud & An auditory threshold, estimated by either Popper, 1975; Holt et al., 2004, 2005). Many of behavioral or electrophysiological responses, is the behavioral observations discussed in Chapter the level of the quietest sound audible in a speci- 4 (and in Appendices B & C) indicated relatively fied percent of trials. An auditory threshold is not precise orientation behaviors to sound sources (or an invariant critical value above which a sound is sound localization) in the field as well. Limited always heard and below which it is never heard. laboratory data are also available regarding how Instead, it is a sound level at which there is an marine mammals detect relatively simple stim- explicit signal detection probability (often 50%; uli over background masking noise (discussed determined a priori). This probability depends below). A more complex perceptual matter related on a number of intrinsic factors (Green & Swets, to localization and detection over masking noise 1974; Egan, 1975; McMillan & Creelman, 1991). is the manner in which vertebrates process com- In all species tested thus far, the hearing response plex information to perceive the acoustic (or audi- in relation to frequency is a generally U-shaped tory) scene—that is, gain useful information from curve with a frequency range of best sensitivity 422 Southall et al. the suite of sounds around them in the real world above those of the masking stimulus is more sig- (e.g., Fay & Popper, 2000). nificantly impeded (see Buus, 1997; Yost, 2000). Bregman (1990) considered how the human Because of common biomechanical cochlear auditory system constructs a perceptual acoustic properties across taxa (Echteler et al., 1994), image of the surrounding environment and events masking is expected to follow similar principles in occurring in that environment. He posits that, as in other mammals (including marine mammals). The visual perception, hearing systems are organized structure and function of the outer and middle ear in such a manner that related acoustic events (such differ profoundly between terrestrial and marine as the frequency structure of a harmonic signal or mammals (Wartzok & Ketten, 1999); however, a repeated signal from the same source in a 3- the characteristics of auditory masking are strik- dimensional space) are grouped perceptually in ingly similar among nonspecialized mammals in a meaningful way. According to the process of general (Fay, 1988; Echteler et al., 1994), includ- auditory scene analysis, the auditory system sorts- ing marine mammals tested in air and in water out related elements of a complex natural acous- (Turnbull & Terhune, 1990; Southall et al., 2000, tic environment into those arising from different 2003). Similarities in morphology and mamma- sound sources. Furthermore, previous experience lian cochlear functional dynamics (as revealed by can have powerful effects on the processing and masking studies) suggest that auditory data from interpretation of sounds. This too is similar to psy- terrestrial mammals may be reliably used in some chological processes underlying visual perception situations where marine mammal data are lacking. in which the range to an object may be inferred Data on auditory masking in marine mammals are from knowledge of an object’s general size and not presented in detail here because they are not physical appearance. directly used in formulating the recommended Presuming such capabilities occur in marine noise exposure criteria (but see Southall et al., vertebrates, which is logical given the importance 2000, 2003, for reviews). of sound to marine mammals, it seems likely that they could perceive range and the general nature Auditory Threshold Shift (e.g., movement) of sound sources. Acoustic Animals exposed to sufficiently intense sound stream segregation, the identification of relatively exhibit an increased hearing threshold (i.e., poorer simple stimuli from different, overlapping patterns, sensitivity) for some period of time following has been demonstrated in several bird and bat spe- exposure; this is called a noise-induced thresh- cies (MacDougall-Shackleton et al., 1998; Moss old shift (TS). Factors that influence the amount & Surlykke, 2001). Neither acoustic stream seg- of TS include the amplitude, duration, frequency regation nor auditory scene analysis has yet been content, temporal pattern, and energy distribution investigated in marine mammals (but see Madsen of noise exposure. The magnitude of TS normally et al., 2005a). Each of these processes, along with decreases over time following cessation of the more data on sound localization, may be relevant in noise exposure. The amount of TS just after expo- the continued development of appropriate marine sure is called the initial TS. mammal noise exposure criteria (see the “Marine If TS eventually returns to zero (i.e., the thresh- Mammal Functional Hearing Groups” section of old returns to the pre-exposure value), it is called Chapter 5, for research recommendations). TTS. The following physiological mechanisms are thought to play some role in inducing TTS, Auditory Masking also referred to as auditory fatigue: effects on sen- Noise may partially or entirely reduce the audi- sory hair cells in the inner ear that reduce their bility of signals, a process known as auditory sensitivity, modification of the chemical environ- masking. The extent of interference depends on ment within sensory cells, residual middle-ear the spectral, temporal, and spatial relationships muscular activity, displacement of certain inner between signals and masking noise, in addition ear membranes, increased blood flow, and post- to other factors. Human auditory systems per- stimulatory reduction in both efferent and sensory form frequency-based assessment (similar to neural output (Kryter, 1994; Ward, 1997). Where Fourier analysis) on incoming signals such that, these effects result in TTS rather than a permanent for most exposure levels, significant masking of change in hearing sensitivity, they are within the tonal signals is almost exclusively by noise in a nominal bounds of physiological variability and narrow band (called the critical band) of similar tolerance and do not represent physical injury frequencies (Wegel & Lane, 1924; Fletcher, 1940; (Ward, 1997). Recovery of nominal hearing func- Greenwood, 1961). With increasing masker level, tion may occur quickly, and the amount of TTS however, there is an asymmetrical spread in the measured depends on the time elapsed since the masking effect such that detection of frequencies cessation of noise exposure; subscripts are used to indicate the time in minutes after exposure. For Marine Mammal Noise Exposure Criteria 423

example, TTS2 means TTS measured 2 min after exposure (see the “Criteria for Injury: TTS and exposure cessation. PTS” section of Chapter 3). If TS does not return to zero after a relatively long interval (on the order of weeks), the residual Behavioral Reactions to Sound TS is called a noise-induced permanent threshold Behavioral responses to sound are highly variable shift (PTS). The distinction between PTS and TTS and context-specific (see Wartzok et al., 2004, for depends on whether there is a complete recovery a discussion). Some sounds that are audible to ani- of TS following noise exposure. PTS is considered mals may elicit no overt behavioral response. This to be auditory injury. Some of the apparent causes is most common when the sound does not greatly of PTS in mammals are severe extensions of exceed the minimum detectable level and is not effects underlying TTS (e.g., irreparable damage increasing or fluctuating (Richardson et al., 1995). to the sensory hair cells). Others involve different Inability to detect an overt response does not nec- mechanisms, such as exceeding the elastic limits essarily mean that there is no subtle behavioral (or of certain tissues and membranes in the middle other) effect, however. and inner ears and resultant changes in the chemi- When observable reactions do occur, they may cal composition of inner ear fluids (Ward, 1997; include orientation or attraction to a sound source; Yost, 2000). The relationship between TTS and increased alertness; modification of characteristics PTS depends on a highly complex suite of vari- of their own sounds; cessation of feeding or social ables concerning the study subject and the expo- interaction; alteration of movement/diving behav- sure. This relationship remains poorly understood, ior; temporary or permanent habitat abandonment; even for humans and small terrestrial mammals in and, in severe cases, panic, flight, stampede, or which this topic has been investigated intensively stranding, sometimes resulting in injury or death (see Kryter, 1994; Yost, 2000). (e.g., Richardson et al., 1995; Evans & England, In addition to the potential for discrete, intense 2001; Gordon et al., 2004; Scheifele et al., 2005; sounds to result in TTS or PTS, chronic sound Cox et al., 2006; Nowacek et al., 2007). Minor exposure, common in industrialized societies, can or temporary behavioral effects are often simply result in noise-induced PTS in humans as they age evidence that an animal has heard a sound and (see Kryter, 1994). Reduced hearing sensitivity may not indicate lasting consequence for exposed as a simple function of development and aging individuals. For the purposes of setting crite- (presbycusis) has been demonstrated in both chil- ria, the effects of greatest concern are those that dren (Roche et al., 1978) and adults (e.g., Brant may negatively impact reproduction or survival. & Fozard, 1990). In the long-term, noise-induced Ultimately, it is the biological relevance of the hearing loss and presbycusis appear to result in reaction in terms of vital parameters that must be a progressive PTS that is a complex, nonlinear determined. In proposing noise exposure criteria, process and particularly affects high-frequency one must clearly and explicitly differentiate triv- hearing. Limited research in cetaceans and pin- ial effects from those with the potential to affect nipeds has revealed patterns of presbycusis that vital rates. However, it has proven to be exceed- are similar to those observed in humans (Ridgway ingly challenging to distinguish among and rank & Carder, 1997; Brill et al., 2001; Schusterman the various effects and to establish a generally et al., 2002; Houser & Finneran, 2006b; Reichmuth accepted definition of biologically meaningful et al., 2007), further underscoring certain gen- behavioral disturbance (see NRC, 2005). eral similarities in auditory processes across Except for naïve individuals, behavioral mammals. responses depend critically on the principles PTS and TTS data from humans and non- of habituation and sensitization. An animal’s human terrestrial mammals were used to develop exposure history with a particular sound affects safe exposure guidelines for human work environ- whether it is subsequently less likely (habitua- ments (e.g., NIOSH, 1998). For marine mammals, tion) or more likely (sensitization) to respond to recent data are available regarding sounds that a stimulus such as sound exposure. The processes cause modest TTS (generally < 20 dB decrease of habituation and sensitization do not necessar- in sensitivity) in a few species of odontocetes and ily require an association with a particular adverse pinnipeds. No data exist on exposures that would or benign outcome. Rather, individuals may be cause PTS in these taxa, however (see Chapter 2 innately predisposed to respond to certain stimuli for detailed discussions). Consequently, the only in certain ways. These responses may interact current option for estimating exposure conditions with the processes of habituation and sensitiza- that would cause PTS-onset in marine mammals tion for subsequent exposure. Where associative is to use the available marine mammal TTS data learning occurs, individuals link a particular expo- combined with data from terrestrial mammals sure with a known outcome (positive, negative, on TTS growth rates with increasing acoustic or neutral) and use that information in guiding 424 Southall et al. future decisions on whether and how to respond recurrence of exposure) may be of considerably to similar stimuli. The relationship between these greater relevance to the behavioral response than two categories of learning (non-associative and simple acoustic variables such as exposure RL. associative) can be highly complex, particularly For example, if an anthropogenic sound is per- for experienced individuals (see Deecke et al., ceived as indicating the presence of a predator, 2002). it is likely to trigger a strong defensive reaction Many contextual variables may be power- at relatively low RLs. On the other hand, sounds ful contributors to an animal’s perception of and that resemble conspecific signals may be ignored reaction to the acoustic scene. These include the or induce approach or avoidance, depending upon perception of source proximity (nearness), relative the context. Further, typically neutral sounds may movement (encroachment or retreat), and general cause increasing annoyance reactions (such as novelty or familiarity, all of which may affect the avoidance) as a function of exposure level. This type and magnitude of the resulting behavioral makes it difficult or impossible to justify basing response(s). In terms of proximity, the presence broad, objective determinations of impact thresh- of high-frequency components in a sound and the olds on RL alone. This is the primary reason why lack of reverberation, both of which are indica- this paper does not propose explicit behavioral tive of proximity, may be more relevant acoustic disturbance criteria levels for certain sound types. cues of spatial relationship than simply exposure Rather, we collated available data relating acous- level alone (see P. Miller, 2002). If a source is per- tic exposure to the severity of observed behav- ceived to be approaching, the response is often ioral response in a form that allows a variety of stronger. In addition, the activity of the individual relationships to be estimated (Chapter 4). When and its fidelity to a current location often affect research allows the separation of annoyance from the response. cases where an animal interprets sounds as sig- Thus, in addition to source characteristics, nals from predators, prey, or conspecifics, it may other factors that may be critical in determining become possible to classify signals and predict behavioral effects include past experience, situ- responses more precisely. ational variables, receiver auditory systems, and the extent to which the sound resembles familiar Non-Auditory Effects benign or noxious stimuli (e.g., Irvine et al., 1981; The auditory system appears to include the organs NRC, 2005). Animals that fail to exhibit general most susceptible to noise exposure, at least in avoidance when exposed to a certain sound source humans (e.g., Ward, 1997). The limited data on may still detect the sound but are either habituated captive marine mammals exposed to various to exposure or may display less dramatic behav- kinds of noise support a similar conclusion, sug- ioral responses (e.g., altering vocal behavior, gesting that TTS-onset occurs at levels which may modifying orientation/movement patterns). be below those required for direct non-auditory The magnitude of a given behavioral response physiological trauma (but see discussion of deep- may not be a direct function of exposure levels diving species below). Noise exposure does have or even of the animal’s experiential history. If the potential to induce a range of direct or indi- the sound triggers an anti-predator response rect physiological effects on non-auditory struc- in the subject (e.g., Irvine et al., 1981; Finley tures. These may interact with or cause certain et al., 1990), the response magnitude may reflect behavioral or auditory effects, or they may occur the individual’s underlying physiological con- entirely in the absence of those effects. dition, the relative costs in fitness of failing to Noise exposure may affect the vestibular and respond, the availability of alternative refuges, neurosensory systems. For instance, in humans, and other factors specific to predator defense (Gill dizziness and vertigo can result from exposure & Sutherland, 2000; Frid & Dill, 2002; Beale & to high levels of noise, a condition known as nys- Monaghan, 2004). tagmus (see Oosterveld et al., 1982; Ward, 1997; For all these reasons, behavioral responses Halmagyi et al., 2005). Little is known about ves- to anthropogenic sounds are highly variable. tibular functions in marine mammals. There are Meaningful interpretation of behavioral response significant differences in vestibular structures in data (and biologically relevant conservation deci- some marine mammal species compared to most sions) must consider not only the relative mag- land mammals (Wartzok & Ketten, 1998; Ketten, nitude and apparent severity of behavioral reac- 2000). In cetaceans in particular, the vestibular tions to human disturbance but also the relevant components are sufficiently reduced and have acoustic, contextual, and ecological variables. such low neural representation that the princi- In many cases, specific acoustic features of the pal function may be essentially to provide lim- sound and contextual variables (e.g., proximity, ited gravitational and linear acceleration cues. subject experience and motivation, duration, or Pinnipeds by contrast have a well-developed, Marine Mammal Noise Exposure Criteria 425 more conventional vestibular apparatus that Sound at certain frequencies can cause an air- likely provides multiple sensory cues similar to filled space to vibrate at its resonant frequency those of most land mammals. Both pinnipeds and (acoustic resonance), which may increase the like- cetaceans retain the direct coupling through the lihood of mechanical trauma in the adjacent or sur- vestibule of the vestibular and auditory systems; rounding tissue. The resonant frequencies of most therefore, it is possible, albeit not known, that marine mammal lungs are below the operating marine mammals may be subject to noise-induced frequencies of many anthropogenic sound sources effects on vestibular function as has been shown (Finneran, 2003). Further, biological tissues are in land mammals and humans. Responses to heavily damped, estimated tissue displacement underwater sound exposures in human divers and at resonant frequencies is predicted to be exceed- other immersed land mammals suggest that ves- ingly small, and lung tissue damage is generally tibular effects are produced from intense under- uncommon in acoustic-related marine mammal water sound at some lower frequencies (Steevens stranding events. For these reasons, specialists do et al., 1997). Theoretical effects on the human ves- not regard lung resonance as a likely significant tibular system as well as other organs (e.g., lungs) non-auditory effect for marine mammals exposed from underwater sound exposures also have been to anthropogenic noise sources that operate above explored through models (Cudahy & Ellison, 100 Hz (U.S. Department of Commerce, 2002). 2002); however, there are no comparable mea- This conclusion might not apply to lower-fre- surements or models for marine mammals at this quency sources that operate at a particular fre- point from which to estimate such effects. Data quency for a significant duration. are clearly needed for all major marine mammal The non-auditory effect now being most taxa to more fully assess potential impacts on non- actively discussed in marine mammalogy is nitro- auditory systems. gen gas bubble growth, resulting in effects similar Relatively low-level physiological responses to decompression sickness in humans. Jepson et al. include changes in cardiac rate (bradycardia or (2003) and Fernández et al. (2004, 2005) hypoth- tachycardia) and respiratory patterns, which may esized that lesions (gas and fat emboli) observed lead to changes in metabolism. Stress reactions in individual beaked whales found stranded after in humans and other vertebrates include various military sonar exercises were somehow caused by physiological changes to pulmonary, cardiac, in vivo nitrogen bubble formation. Osteonecrosis metabolic, neuro-endocrine, immune, and repro- in sperm whales has further been suggested as ductive functions (e.g., Hales, 1973; Lee, 1992; a chronic result of nitrogen bubble formation Vrijkotte et al., 2000). Studies of noise-induced (Moore & Early, 2004). stress in marine mammals are very limited, but To date, the gas bubble hypothesis remains endocrine secretions of glucocorticoids and altered untested, and the acoustic causative mechanism cardiovascular function have been documented in for formation of emboli, if any, is unknown. odontocetes exposed to high-level sound (Romano Theoretically, bubble precursors in supersaturated, et al., 2004; cf. Thomas et al., 1990c). Noise expo- homogenized tissue may incrementally enlarge sure also often leads to changes in surfacing-res- during the successive passage of compression and piration-dive cycles of cetaceans (e.g., Richardson rarefaction portions of acoustic waves that exceed & Malme, 1993), which may have various physi- static pressure (rectified diffusion; Crum & Mao, ological effects. Assuming that effects in marine 1996). Alternatively, a single acoustic exposure and terrestrial mammals are similar, intermediate could activate bubble precursors, allowing them physiological responses to stressors (including to grow by gradual expansion into bubbles in noise) may accompany avoidance or aggres- nitrogen-supersaturated tissue (static diffusion; sive behaviors and include single auditory star- see Potter, 2004). The diving patterns of some tle responses, the initiation and sustenance of marine mammals increase gas-tissue saturation the catecholamine response, and physiological and potentially could increase the susceptibility of preparation for fight or flight. The most severe noise-exposed animals to bubble growth via either physiological responses would include multiple mechanism (Ridgway & Howard, 1979; Houser or repeated auditory startle responses, trigger- et al., 2001b). Nitrogen supersaturation levels for ing of the hypothalamic-pituitary-adrenal (HPA) deep-diving species of interest, including beaked axis and associated elevated blood glucocorticoid whales, are based on theoretical models, however level, substantially altered metabolism or energy (Houser et al., 2001b). No unequivocal support for reserves, lowered immune response, diminished either pathway presently exists. reproductive effort, and potential tissue trauma The evidence for bubble formation as a causal (e.g., Sapolsky et al., 2000). [The issue of stress mechanism between certain types of acoustic responses to noise exposure has been discussed exposure and stranding events remains equivo- recently by Wright et al. (in press).] cal. At a minimum, scientific disagreement and/or 426 Southall et al. complete lack of information exists regarding the following important points: (1) received acoustic exposure conditions for animals involved in stranding events; (2) pathological interpretation of observed lesions in stranded marine mammals (Fernández et al., 2004; Piantadosi & Thalmann, 2004); (3) acoustic exposure conditions required to induce such physiological trauma directly; (4) whether noise exposure may cause behavioral reactions (e.g., atypical diving behavior) that secondarily induce bubble formation and tissue damage (Jepson et al., 2003; Fernández et al., 2005; Zimmer & Tyack, 2007); and (5) the extent that post mortem artifacts introduced by decompo- sition before sampling, handling, freezing, or nec- ropsy procedures affect interpretation of observed lesions. Tests of the gas bubble hypothesis may yield data pertinent to future marine mammal noise exposure criteria, but too little is currently known to establish explicit exposure criteria for this proposed mechanism.

Courtesy: A. Friedlander Aquatic Mammals 2007, 33(4), 427-436, DOI 10.1578/AM.33.4.2007.427

2. Structure of the Noise Exposure Criteria

When de facto noise exposure guidelines are The format of the recommended marine used by management agencies, they generally are mammal noise exposure criteria is thus a matrix of based on a small number of categories of marine 15 “cells” that systematically considers three sound mammals and sound types. Though it would be types (see next section) and five functional marine convenient to have a single exposure criterion mammal hearing groups (see the “Marine Mammal for all species and sound sources, such a simpli- Functional Hearing Groups” section of this chap- fied approach is not supported by available sci- ter). Within each of those 15 cells, we consider ence. However, some categorization of species two general acoustic metrics (see the “Exposure and sources is warranted based on current infor- Criteria Metrics” section) and two levels of expo- mation. The many anthropogenic sound sources sure effect (“Levels of Noise Effect: Injury and used in marine environments can be categorized Behavorial Disturbance” section of this chapter). based on certain acoustic and operational features. Sixty possible criteria result (i.e., 3 sound types × Similarly, there is great diversity in hearing and 5 marine mammal groups × 2 metrics × 2 impact in the biological effects of noise among marine levels), although fewer than 60 are reported due to mammals, but current knowledge supports some data limitations. Whereas sound types are defined functional and/or phylogenetic groupings. by source features, criteria values represent levels It is also neither possible nor desirable to derive received by individual marine mammals. distinct exposure criteria for every species and sound source. Important generalizations across Sound Types taxa would be missed even if resources and time were adequate to study each species and expo- Three sound types are used: (1) a single pulse, sure condition. Further, it is impractical to apply (2) multiple pulses, and (3) nonpulses. The separa- numerous, species-specific criteria when predict- tion between pulses and nonpulses is supported by ing and/or attempting to mitigate effects. data on auditory fatigue and acoustic trauma in ter- A standard scientific approach in such situa- restrial mammals (e.g., Dunn et al., 1991; Hamernik tions is to categorize animals based on functional et al., 1993) and is generally consistent with the characteristics and sound sources based on physi- sound types distinguished for damage risk criteria cal similarities, and to summarize the information in humans (e.g., U.S. DoD, 1997; NIOSH, 1998). in a matrix format. We subdivide cetaceans and pin- Pulses and nonpulses are distinguished by nipeds into five functional hearing categories based numerous definitions and mathematical distinc- on the frequencies they hear. Other methods of cat- tions (e.g., Burdic, 1984). The empirical distinc- egorization are, of course, possible. For instance, tion used here is based on a measurement proce- Verboom (2002) relied heavily on direct measure- dure using several temporal weightings. Various ments of noise impacts on hearing to quantify the exponential time-weighting functions applied in effects of noise exposure on marine mammals. Some measuring pulse and nonpulse sounds may yield of his proposed criteria are comparable with those different measured received levels (RLs) (see presented here. The present effort makes broader Harris, 1998). Most sound level meters (SLM) pro- use of laboratory and field behavioral and audiomet- vide options for applying either a “slow” or “fast” ric data, additional recent data, and extrapolations time constant (1,000 or 125 ms, respectively) for from terrestrial mammals not used by Verboom. We measuring nonpulses or an impulse time con- divide sound sources into three types according to stant (35 ms) appropriate for measuring pulses. acoustic characteristics defined at the source. Note For a sound pulse, the slow or fast SLM settings that at a distance, a sound may have significantly result in lower sound pressure level (SPL) mea- different features; categorizing sounds based on surements than those obtained using the impulse source characteristics is a precautionary and prag- setting. Each of these time constants is selected matic approach (as is described in the next section). based on properties of the human auditory system. The justifications for and assumptions underlying These may be at least generally relevant for other our categorization of functional hearing groups and mammalian auditory systems, although further sound types are described here. The number of sub- empirical data on temporal resolution in marine divisions in future noise exposure criteria will likely mammals are needed (see Chapter 5, “Research increase as more supporting data are acquired. Recommendations”). 428 Southall et al.

Harris (1998) proposed a measurement-based Treating pulses and nonpulses as discrete sound distinction of pulses and nonpulses that is adopted types is justified by data on mammals in general and here in defining sound types. Specifically, a ≥ 3-dB several cetacean species in particular (Dunn et al., difference in measurements between continuous 1991; Hamernik et al., 1993; also see the “Effects and impulse SLM settings indicates that a sound is of Noise on Hearing in Marine Mammals TTS a pulse; a < 3-dB difference indicates that a sound Data” section in Chapter 3). Mammalian hearing is a nonpulse. We note the interim nature of this is most readily damaged by transient sounds with distinction for underwater signals and the need for rapid rise-time, high peak pressures, and sustained an explicit distinction and measurement standard duration relative to rise-time (for humans: Thiery such as exists for aerial signals (ANSI, 1986). & Meyer-Bisch, 1988; for chinchillas [Chinchilla Harris’s (1998) definitions assumed use of A- lanigera], Dunn et al., 1991). Consistent with these weighting as do most human-oriented definitions results, those odontocetes tested thus far have been of acoustical measurements; however, different shown to experience TTS-onset at lower respective frequency-weighting functions should be used for exposure levels if the sound is a pulse rather than a various animal taxa (as discussed below). Leaving nonpulse (Finneran et al., 2002b, 2005a). that question aside temporarily, it is instructive to Mammals are also apparently at greater risk compare the impulse equivalent-continuous sound from rapidly repeated transients and those with level (LIeqT) for a sound that increases in level with high impulse amplitude kurtosis (Erdreich, 1986). the corresponding equivalent-continuous level Hamernik et al. (1993, 2003) argued that the dis- (LeqT). Here, LIeqT has an impulse integration time of tinction between exposures with relatively high 35 ms and LeqT, defined as sound exposure divided and low “peakedness” is to some extent an over- by T, is expressed as a level. As an example, sup- simplification. Highly variable threshold shifts pose that a source is examined over a 2-s period can result from exposures of variable peaked- (T = 2 s). The highest LAIeq2s (“A” here denotes A- ness but comparable overall levels, depending on weighting) during this period is 75.2 dB, and the a host of factors. Hamernik et al. (1993, 2003) highest LALeq2s is 65.1 dB. The difference of 10.1 also noted that peak pressure levels sufficient to dB is greater than the 3-dB criterion given by exceed mechanical limits of the cochlea, and thus Harris (1998); therefore, the sound is considered more likely to induce acoustic trauma, tend to be to be a pulse. more typical of pulses than nonpulses. The distinction between pulses and nonpulses is The present criteria also categorize sound types not always clear in practice. For instance, certain based on repetition. For mammals, single and mul- signals (e.g., acoustic deterrent and harassment tiple noise exposures at various levels and dura- devices) have characteristics of both pulses and tions generally differ in their potential to induce nonpulses. Also, certain sound sources (e.g., seis- auditory fatigue or trauma. This results principally mic airguns and pile driving) may produce pulses from the temporal interaction between exposure at the source but, through various propagation and recovery periods (e.g., Kryter, 1994) and dif- effects, may meet the nonpulse definition at greater ferences in received total acoustic energy. Further, distances (e.g., Greene & Richardson, 1988). This multiple exposures may increase the likelihood of means that a given sound source might be subject behavioral responses because of increased prob- to different exposure criteria, depending on the ability of detection and the (generally) greater distance to the receiver and intervening propaga- biological significance of continued exposure as tion variables. While this is certainly realistic for opposed to a single, transient event (although see many real-world exposures, measurements at the discussion of habituation in the “Responses to animal are often not practical. Changes in sound Sound” section of Chapter 1). characteristics with distance generally result in Single exposures are considered here as dis- exposures becoming less physiologically damag- crete acoustic events in which received sound ing with increasing distance because sharp tran- levels exceed ambient noise in at least some por- sient peaks become less prominent. Therefore, tion of the frequency band of functional marine these criteria use a precautionary approach and mammal hearing once in a 24-h period; multi-path classify sound types based on acoustic character- receptions of a single exposure are not consid- istics at the source. Additional empirical measure- ered multiple exposures. Multiple exposures are ments are needed to advance our understanding of considered to be acoustic events causing RLs to sound type classification as a function of source, exceed ambient noise within the functional band- range, and environmental variables. We empha- width more than once, with an intervening quiet size that the use of source parameters to classify period not exceeding 24 h. If the exposure event sound types does not negate our decision to rec- is interrupted, even briefly (other than as a result ommend exposure criterion levels relative to RLs of the animal’s own action—e.g., breaching), it is at the animal. considered a multiple exposure. Marine Mammal Noise Exposure Criteria 429

Exposures should be categorized as either pulsed negates the need for numerically different injury cri- or nonpulsed sounds as described above. Single and teria for single and multiple exposures at the expense multiple exposures to either pulse or nonpulse sounds of neglecting assumed, but as-yet poorly understood (or both) are possible. Examples of single pulses and recovery phenomena during intervals between expo- single nonpulses are sounds from a single firing of sures. This is a precautionary approach, pending an airgun or a single vessel passage, respectively. availability of data on acoustic recovery by marine Multiple pulse or multiple nonpulse sounds are mammals during intervals between exposures. more difficult to delineate, given the diversity and When considering behavioral responses, single complexity of sound sources. A series exclusively and multiple nonpulse exposures are considered as consisting of two or more nonpulses would clearly be a single category. Insufficient information exists to a multiple nonpulse exposure (e.g., multiple vessel assess the use of SEL as a relevant metric in the con- passages). A multiple pulse exposure would simi- text of marine mammal behavioral disturbance for larly be described as a series exclusively containing anything other than a single pulse exposure. Future pulses (e.g., repeated pile strikes) or a combination noise exposure criteria for behavioral disturbance of pulses and nonpulses (e.g., the combined vessel may distinguish SPL and SEL exposure criteria for noise and airgun transmissions of a seismic vessel). additional conditions, but for most sound types (the One justification for treating combined pulses and exception being single pulses), the available data nonpulses as pulses is that the proposed exposure are best assessed in relation to SPL (discussed in criteria for injury are more precautionary (lower) detail in Chapter 4). Consequently, the structure of in the case of pulses than for nonpulses. Specific the exposure criteria matrix includes a categorical consideration should be given, on a case-by-case distinction between single and multiple pulses given basis, as to whether such a distinction would neces- that numerical SEL thresholds are recommended for sarily be the more precautionary. For instance, if a a single pulse, but not for multiple pulses. No such compound exposure included relatively high-level distinction is made for nonpulses where the available nonpulses as well as relatively low-level pulses, the data do not (at least currently) support differential more appropriate and protective distinction might behavioral criteria for single vs multiple exposures. be to classify it as a nonpulse exposure. Thus, the current state of scientific knowledge The proposed exposure criteria for injury from regarding mammalian hearing and various noise single and multiple exposures to both sound types impacts supports three distinct sound types as are numerically identical (Chapter 3). This is another relevant for marine mammal noise exposure cri- precautionary decision, arising from the fact that no teria: (1) single pulse, (2) multiple pulses, and marine mammal data were available regarding the (3) nonpulses. Examples of sound sources belong- effects of inter-exposure interval on recovery from ing in each of these categories (based on character- auditory effects (e.g., TTS). A summation procedure istics of the sound emitted at the source) are given is applied to quantify the fatiguing effects of multi- in Table 1. A simplistic measurement procedure ple exposures with an equivalent SEL value (Chapter using source characteristics (the 3-dB distinction 1; also Appendix A, eq. 5). The SEL metric takes based on Harris, 1998, described above) is used account of the pressure waveform and duration of here to distinguish a pulse from a nonpulse, while either single or multiple sound events; it represents the simple definitions above distinguish single cumulative received energy. This approach effectively and multiple exposures.

Table 1. Sound types, acoustic characteristics, and selected examples of anthropogenic sound sources; note sound types are based on characteristics measured at the source. In certain conditions, sounds classified as pulses at the source may lack these characteristics for distant receivers.

Sound type Acoustic characteristics (at source) Examples

Single pulse Single acoustic event; > 3-dB difference between Single explosion; sonic boom; single airgun, received level using impulse vs equivalent watergun, pile strike, or sparker pulse; single ping continuous time constant of certain sonars, depth sounders, and pingers Multiple pulses Multiple discrete acoustic events within 24 h; Serial explosions; sequential airgun, watergun, > 3-dB difference between received level using pile strikes, or sparker pulses; certain active sonar impulse vs equivalent continuous time constant (IMAPS); some depth sounder signals Nonpulses Single or multiple discrete acoustic events within Vessel/aircraft passes; drilling; many construc- 24 h; < 3-dB difference between received level tion or other industrial operations; certain sonar using impulse vs equivalent continuous time systems (LFA, tactical mid-frequency); acoustic constant harassment/deterrent devices; acoustic tomography sources (ATOC); some depth sounder signals 430 Southall et al.

Marine Mammal Functional Hearing Groups (i.e., external auditory meatus, air-filled middle ear, and spiral-shaped cochlea). Most of the mech- Species of cetaceans and pinnipeds were assigned anisms of mammalian hearing are also conserved to one of five functional hearing groups based on such as the basic lever structure of the ossicles and behavioral psychophysics, evoked potential audi- the tonotopic organization of the hair cells along ometry, auditory morphology, and (for pinnipeds) the inner ear’s basilar membrane. the medium in which they listen. Cetaceans and However, marine mammal auditory systems differ pinnipeds are broadly separable based on phylo- in having some adaptations that seem to be related genetic and functional differences (Reynolds & to pressure, hydrodynamics, and sound reception in Rommel, 1999). Cetaceans were further subdi- water (see Wartzok & Ketten, 1999). For instance, vided according to differences in their measured the pinna has been reduced or eliminated in most or estimated hearing characteristics and not neces- species, owing to hydrodynamic adaptations. sarily according to their phylogeny (as in Wartzok Tissue modifications may enable the reduction or & Ketten, 1999). Pinnipeds are considered a single elimination of gas spaces in the middle ear of some group, but as amphibious mammals, their hearing marine mammals. Consequently, bone conduction, differs in air and in water (Kastak & Schusterman, rather than the conventional ossicular chain, may be 1998); separate criteria were required for each an additional (or primary) sound transmission path medium. The taxa in each functional hearing to the cochlea (e.g., Repenning, 1972; Au, 1993). group (based on Rice, 1998) are given in Table 2. There are important differences in these adaptations within and between marine mammal taxa. Marine Mammal Hearing Knowledge of marine mammal hearing varies All marine mammals evolved from terrestrial, air- widely among groups, but for most species it is adapted ancestors (Domning et al., 1982; Barnes quite limited compared to knowledge of terrestrial et al., 1985) and, at least in part, retain the nominal mammal hearing. Because of the sheer size, lim- mammalian tripartite peripheral auditory system ited and disproportionate availability in captive

Table 2. Functional marine mammal hearing groups, auditory bandwidth (estimated lower to upper frequency hearing cut-off), genera represented in each group, and group-specific (M) frequency-weightings

Functional hearing Estimated auditory Genera represented Frequency-weighting group bandwidth (Number species/subspecies) network

Low-frequency 7 Hz to 22 kHz Balaena, Caperea, Eschrichtius, Mlf cetaceans Megaptera, Balaenoptera (lf: low-frequency cetacean) (13 species/subspecies)

Mid-frequency 150 Hz to 160 kHz Steno, Sousa, Sotalia, Tursiops, Stenella, Mmf cetaceans Delphinus, Lagenodelphis, Lagenorhynchus, (mf: mid-frequency Lissodelphis, Grampus, Peponocephala, cetaceans) Feresa, Pseudorca, Orcinus, Globicephala, Orcaella, Physeter, Delphinapterus, Monodon, Ziphius, Berardius, Tasmacetus, Hyperoodon, Mesoplodon (57 species/subspecies)

High-frequency 200 Hz to 180 kHz Phocoena, Neophocaena, Mhf cetaceans Phocoenoides, Platanista, Inia, Kogia, (hf: high-frequency Lipotes, Pontoporia, Cephalorhynchus cetaceans) (20 species/subspecies)

Pinnipeds in water 75 Hz to 75 kHz Arctocephalus, Callorhinus, Mpw Zalophus, Eumetopias, Neophoca, (pw: pinnipeds in water) Phocarctos, Otaria, Erignathus, Phoca, Pusa, Halichoerus, Histriophoca, Pagophilus, Cystophora, Monachus, Mirounga, Leptonychotes, Ommatophoca, Lobodon, Hydrurga, and Odobenus (41 species/subspecies)

Pinnipeds in air 75 Hz to 30 kHz Same species as pinnipeds in water Mpa (41 species/subspecies) (pa: pinnipeds in air) Marine Mammal Noise Exposure Criteria 431 settings, and, for many species and jurisdictions, for mysticetes may need to be revisited based the protected status of marine mammals, there on emerging knowledge. At present, we esti- are limitations in obtaining hearing data for many mate the lower and upper frequencies for func- species. Behavioral or electrophysiological audio- tional hearing in mysticetes, collectively, to be grams exist for fewer than 20 marine mammal 7 Hz and 22 kHz (Ketten et al., 2007). species (of ~128 species and subspecies; Rice, Mid- and high-frequency cetaceans are all 1998). By combining these data with comparative odontocetes (toothed whales). Unlike the mystice- anatomy, modeling, and response measured in tes, all odontocete cetaceans appear to have highly ear tissues from species that are difficult to study, advanced echolocation (biosonar) systems that however, it is possible to describe the frequency use intermediate to very high frequencies (tens of sensitivity and critical adaptations for underwa- kHz to 100+ kHz: see Au, 1993; Richardson et al., ter hearing in each of the five functional hearing 1995; Wartzok & Ketten, 1999). They also produce groups of marine mammals considered here. social sounds in a lower-frequency band, including Low-frequency cetaceans consist of 13 species generally low to intermediate frequencies (1 kHz to and subspecies of mysticete (baleen) whales in tens of kHz). Consequently, their functional hear- five genera (based on Rice, 1998; see Table 2). No ing would be expected to cover a wider absolute direct measurements of hearing exist for these ani- frequency range than is assumed for mysticetes or mals, and theories regarding their sensory capabil- has been demonstrated for pinnipeds (discussed ities are consequently speculative (for a detailed below). This has been experimentally confirmed assessment by species using the limited available in the odontocete species whose hearing has been information, see Erbe, 2002). They are too large to measured (discussed below); however, their best maintain in the laboratory for psychophysical test- hearing sensitivity typically occurs at or near the ing. The limited evoked potential measurements frequency where echolocation signals are stron- on animals of this size have not yet yielded hearing gest. Based on the differential characteristics of thresholds (Ridgway & Carder, 2001), but techno- echolocation signals in two groups of odontocetes logical advances may soon enable evoked poten- (see Au, 1993) and on the hearing data described tial audiometry on relatively small and/or young below, odontocetes were divided into mid- and mysticetes. In these species, hearing sensitivity high-frequency functional groups (as seen gener- has been estimated from behavioral responses ally in Wartzok & Ketten, 1999). (or lack thereof) to sounds at various frequencies, Mid-frequency cetaceans include 32 species vocalization frequencies they use most, body size, and subspecies of “dolphins,” six species of larger ambient noise levels at the frequencies they use toothed whales, and 19 species of beaked and bot- most, and cochlear morphometry (Richardson tlenose whales (see Table 2). “Functional” hear- et al., 1995; Wartzok & Ketten, 1999; Houser ing in this group was estimated to occur over a et al., 2001a; Erbe, 2002; Clark & Ellison, 2004). wide range of low to very high frequencies. Based Until better information is available regard- on the combined available data, mid-frequency ing the relationship between auditory sensitivity species are estimated to have lower and upper and marine environmental noise, the sensitivity frequency “limits” of nominal hearing at approxi- of mysticetes cannot be easily inferred from the mately 150 Hz and 160 kHz, respectively. As for acoustic environment. the other hearing groups, there is variability within The combined information strongly suggests and among species, intense signals below and that mysticetes are likely most sensitive to sound above the stated bounds may be weakly detect- from perhaps tens of Hz to ~10 kHz. However, able, and there is a progressive rather than instan- recent data indicated that humpback whales taneous reduction in hearing sensitivity near these (Megaptera novaeangliae) produce some signals limits. Mid-frequency cetaceans generally do not with harmonics extending above 24 kHz (Au appear well-adapted to detect or to discriminate et al., 2006). These harmonics have considerably signals outside this frequency band, however. The lower levels than occur at lower frequencies, and scarcity (and variability) of empirical data pre- their presence does not necessarily indicate they cludes a finer subdivision of this relatively diverse are audible to the whales. Nonetheless, some and large group of marine mammals, though it is high-frequency energy is present. [Additionally, acknowledged that some mid-frequency species some recent anatomical modeling work by likely have a narrower functional hearing band Ketten et al. (2007) suggested that some mysti- than the range given above. cetes may have functional hearing capabilities at Behavioral hearing data are available for the frequencies as high as 30 kHz.] While we do not following mid-frequency cetacean species: bot- include these recent results at this time, we note tlenose dolphin (Tursiops truncatus: Johnson, their presence and the possibility that the upper 1967; Ljungblad et al., 1982; Finneran et al., frequency limit of the M-weighting function 2005a), beluga (Delphinapterus leucas: White 432 Southall et al. et al., 1978; Awbrey et al., 1988; Johnson, 1992; et al., 2006); and Amazon river dolphin (Popov & Ridgway et al., 2001; Finneran et al., 2005b), Supin, 1990). killer whale (Orcinus orca: Hall & Johnson, The pinnipeds include 16 species and subspecies 1972; Szymanski et al., 1999), false killer whale of sea lions and fur seals (otariids), 23 species and (Pseudorca crassidens: Thomas et al., 1988, subspecies of true seals (phocids), and two sub- 1990a; Au et al., 1997), Risso’s dolphin (Grampus species of walrus (odobenids). Pinnipeds produce griseus: Nachtigall et al., 1995; Au et al., 1997); a wide range of social signals, most occurring at and Pacific white-sided dolphin (Lagenorhynchus relatively low frequencies. They lack the highly- obliquidens: Tremel et al., 1998). specialized active biosonar systems of odontocete Audiograms derived using auditory evoked cetaceans, possibly as a result of their amphibious potential (AEP) methodology (Supin et al., 2001) lifestyle (see Schusterman et al., 2000). Because have been obtained for a number of cetacean spe- of this aspect of their life history, pinnipeds com- cies. Specific AEP techniques, which involve municate acoustically in air and water, have sig- measuring electrophysiological responses to nificantly different hearing capabilities in the sound, include those measuring transient evoked two media, and may be subject to both aerial and responses, such as the auditory brainstem response underwater noise exposure (Schusterman, 1981; (ABR) or mid-latency response, and those mea- Kastak & Schusterman, 1998, 1999). These dif- suring steady-state evoked responses such as the ferences necessitate separate noise exposure crite- envelope following response (EFR) or auditory ria for pinnipeds in each medium. steady-state response (ASSR). Mid-frequency For pinnipeds in water, behavioral measures cetacean species tested include the bottlenose of hearing are available for the northern fur seal dolphin (Bullock et al., 1968; Seeley et al., 1976; (Callorhinus ursinus: Moore & Schusterman, Popov & Supin, 1990; Houser & Finneran, 2006b; 1987; Babushina et al., 1991), California sea Finneran et al., 2007a; Hernandez et al., 2007; lion (Zalophus californianus: Schusterman et al., Popov et al., 2007), killer whale (Szymanski 1972; Moore & Schusterman, 1987; Kastak & et al., 1999), beluga (Popov & Supin, 1990; Schusterman, 1998, 2002; Southall et al., 2004), Klishin et al., 2000), common dolphin (Delphinus northern elephant seal (Mirounga angustirostris: delphis: Popov & Klishin, 1998), Risso’s dolphin Kastak & Schusterman, 1998, 1999; Southall (Dolphin, 2000; Nachtigall et al., 2005, 2007), et al., 2004), Hawaiian monk seal (Monachus tucuxi dolphin (Sotalia fluviatilis: Popov & Supin, schauinslandi: Thomas et al., 1990b), harp seal 1990), striped dolphin (Stenella coeruleoalba: (Pagophilus groenlandicus: Terhune & Ronald, Kastelein et al., 2003), Pacific white-sided dol- 1972), ringed seal (Phoca hispida: Terhune phin (Au et al., 2007), false killer whale (Supin & Ronald, 1975), harbor seal (Møhl, 1967, et al., 2003), and Gervais’ beaked whale (Cook 1968; Terhune & Turnbull, 1995; Kastak & et al., 2006). Additionally, Yuen et al. (2005) con- Schusterman, 1995, 1998; Southall et al., 2004), ducted a comparative study of behavioral and AEP and walrus (Odobenus rosmarus: Kastelein thresholds for the false killer whale, and Finneran et al., 2002b). Ridgway & Joyce (1975) measured & Houser (2006), Houser & Finneran (2006a), the gray seal’s (Halichoerus grypus) underwater and Finneran et al. (2007b) have compared behav- hearing using evoked potential audiometry. ioral and AEP thresholds in multiple bottlenose For pinnipeds in air, behavioral measures of hear- dolphins. ing are available for the northern fur seal (Moore The high-frequency cetaceans include eight & Schusterman, 1987; Babushina et al., 1991), species and subspecies of true porpoises, six spe- California sea lion (Schusterman, 1974; Kastak & cies and subspecies of river dolphins plus the fran- Schusterman, 1998; Kastak et al., 2004b), north- ciscana, Kogia, and four species of cephalorhyn- ern elephant seal (Kastak & Schusterman, 1998, chids (see Table 2). “Functional” hearing in this 1999; Kastak et al., 2004b), harp seal (Terhune group was estimated to occur between 200 Hz and & Ronald, 1971), and harbor seal (Møhl, 1968; 180 kHz. Behavioral audiograms are available for Kastak & Schusterman, 1998; Kastak et al., the following high-frequency cetacean species: 2004b). Aerial hearing in pinnipeds has also been harbor porpoise (Phocoena phocoena: Andersen, measured using evoked potential audiometry in 1970; Kastelein et al., 2002a), Chinese river dol- the gray seal (Ridgway & Joyce, 1975), California phin (Lipotes vexillifer: Wang et al., 1992), and sea lion (Bullock et al., 1971; Ridgway & Joyce, Amazon river dolphin (Inia geoffrensis: Jacobs & 1975; Mulsow & Reichmuth, 2007; Reichmuth Hall, 1972). Audiograms using AEP methodol- et al., 2007), harbor seal (Thorson et al., 1998; ogy have been obtained for three species: harbor Wolski et al., 2003; Mulsow & Reichmuth, 2007; porpoise (Popov et al., 1986, 2006; Beedholm Reichmuth et al., 2007), and northern elephant & Miller, 2007; Lucke et al., 2007b); finless seal (Houser et al., 2007; Mulsow & Reichmuth, porpoise (Neophocaena phocaenoides: Popov 2007; Reichmuth et al., 2007). Marine Mammal Noise Exposure Criteria 433

The combined results of these studies indi- These frequency-weighting functions take into cate that pinnipeds are sensitive to a broader account both the frequency bandwidth of human range of sound frequencies in water than in air. hearing and loudness perception. For use as fre- The data further suggest differences in the func- quency filters, the functions are inverted; normal- tional hearing range among otariids, phocids, ized to 0 dB in the frequency range of best hearing and odobenids, especially under water (Kastak & (specifically at 1,000 Hz for humans); and ideal- Schusterman, 1998; Kastelein et al., 2002b). For ized for implementation in hearing aids, sound these proposed noise exposure criteria, however, level meters, and other measurement devices. pinnipeds are considered a single functional hear- At minimum, metrics used for animals should ing group because the data are too limited, both in eliminate inaudible frequencies both below and terms of absolute hearing data and TTS measure- above the range of functional hearing. The “abso- ments (see “The Effects of Noise on Hearing in lute” auditory threshold function (audiogram) has Marine Mammals: TTS Data” section in Chapter been suggested as a frequency-weighting func- 3), to support finer subdivisions. We estimate that tion for marine species exposed to underwater pinnipeds have “functional” underwater hearing sound (e.g., Malme et al., 1989; Thorson et al., between 75 Hz and 75 kHz and “functional” aerial 1998; Heathershaw et al. 2001; Nedwell et al., hearing between 75 Hz and 30 kHz. These ranges 2007) as well as for terrestrial animals (Delaney are essentially based on data for phocid seals, et al., 1999; Bjork et al., 2000). However, the which have the broadest auditory bandwidths of auditory threshold function does not characterize the pinnipeds. This approach results in a precau- the flattening of equal-loudness perception with tionary functional bandwidth for estimating fre- the increasing stimulus level that has been dem- quency-weighting functions (below) and noise onstrated in humans (Fletcher & Munson, 1933). impacts on pinnipeds. Acoustic injury would only be expected to occur In summary, based on current knowledge at levels far above the detection threshold—that of functional hearing in marine mammals, five is, levels for which the flattening effect would be distinct, functional hearing categories were expected. Consequently, it is unclear how useful defined: (1) low-frequency cetaceans (i.e., mys- or appropriate the auditory threshold function is ticetes), (2) mid-frequency cetaceans (i.e., most in deriving frequency-weighting filters in marine odontocetes), (3) high-frequency cetaceans (i.e., mammals for which psychophysical equal-loud- porpoises, river dolphins, pygmy sperm whale, ness measurements are generally unavailable and Cephalorhynchus), (4) pinnipeds in water, (although see preliminary measurements by and (5) pinnipeds in air. The genera in each group, Ridgway & Carder, 2000). Further, the limited and the estimated lower and upper frequency TTS data for cetaceans exposed to tones at differ- hearing “limits,” are shown in Table 2. Because ent frequencies (discussed below) suggest that an the five functional hearing groups of marine mam- audiogram-based frequency-weighting function mals differ in hearing bandwidth, each may be would produce too much filtering at lower fre- affected differently by identical noise exposures. quencies (i.e., the weighting function for hearing Therefore, frequency-weighting functions are effects should be flatter than the inverted audio- required to develop marine mammal noise expo- gram procedure would indicate). sure criteria. Therefore, a precautionary procedure was used to derive frequency-specific, marine mammal Frequency-Weighting Functions weighting functions. Each was based on an algo- As a general statement, animals do not hear rithm that requires only the estimated (as ~80 dB equally well at all frequencies within their func- above best hearing sensitivity) lower and upper tional hearing range. Frequency weighting is a frequencies of functional hearing as given in the method of quantitatively compensating for the above description of each marine mammal group differential frequency response of sensory sys- and in Table 2. The resulting functions were tems. Generalized frequency-weighting functions designed to reasonably represent the bandwidth were derived for each functional hearing group of where acoustic exposures can have auditory effects marine mammals using principles from human and were designed to be most accurate for describ- frequency-weighting paradigms, with adjustments ing the adverse effects of high-amplitude noise for the different hearing bandwidths of the various where loudness functions are expected to flatten marine mammal groups. significantly. The weighting functions (designated For humans, substantial improvement in dose- “M” for marine mammal) are analogous to the response models is obtained by filtering noise C-weighting function for humans, which is com- through equal-loudness functions, particularly monly used in measuring high-amplitude sounds. the 40-phon, equal-loudness function (“A-weight- In the general absence of empirical data, however, ing”) and the 100-phon function (“C-weighting”). the upper and lower frequency roll-offs of the 434 Southall et al.

M-weighting functions are symmetrical, whereas frequency-weighting functions includes, within its C-weighting admits more energy at the lower than relatively flat portion, the full audible range for at the upper frequency limits (ANSI, 2001). each species for which auditory data are available. The M-weighting functions assume a logarith- In other words, none of the species included within mic reduction in auditory sensitivity outside of the each functional hearing group has been shown or range of best hearing sensitivity, with the function is expected to have any portion of its best hearing being 6 dB down from peak sensitivity at the lower sensitivity outside the flat portion of the relevant and upper frequency “limits.” Auditory detection frequency-weighting function. Thus, the functions thresholds at these “limits” (see above discussion are quite precautionary, which is appropriate given of lower and upper frequency “cut-offs”) can be that data are limited or lacking for most species. ≥ 80 dB higher (less sensitive) than those at the frequencies of best hearing sensitivity. Consequently, Exposure Criteria Metrics these frequency filters are much “flatter” than audiograms and probably quite precautionary Many acoustic metrics (e.g., RMS or peak SPL, even considering the expected flattening of equal- SEL, kurtosis) could be considered in relation loudness contours at high exposure levels. The to noise impacts on animals. It is impossible to M-weighting functions are also precautionary in predict unequivocally which one is best associ- that regions of best hearing sensitivity for most ated with the likelihood of injury or significant species are likely considerably narrower than the behavioral disturbance across all taxa because of M-weighting functions (designed for the overall species differences and the fact that real-world marine mammal group) would suggest. The gen- sound exposures contain many widely differing eral expression for M-weighting (M[f]), using temporal patterns and pressure signatures. To the estimated lower and upper “functional” hear- account for such differences and to allow for cur- ing limits (flow and fhigh) for each of the five func- rent scientific understanding of tissue injury from tional marine mammal hearing groups, is given in noise exposure, the proposed injury criteria incor- Appendix A (eq. 7 & 8). These frequency-weight- porate a dual-criteria approach based on both peak ing functions are identified in Table 2, and each is pressure and energy. For an exposed individual, depicted graphically in Figure 1. whichever criterion is exceeded first (i.e., the more The M-weighting functions de-emphasize fre- precautionary of the two measures) is used as the quencies that are near the lower and upper fre- operative injury criterion. Similarly, a dual-crite- quency ends of the estimated hearing range as rion approach (peak sound pressure and energy) indicated by negative relative values (Figure 1). is also proposed for behavioral disturbance from This de-emphasis is appropriate because, to have a single pulse. a given auditory effect, sound at these frequencies The pressure criteria for injury are defined as must have higher absolute amplitude than sound in those peak SPLs above which tissue injury is pre- the region of best hearing sensitivity. As a corol- dicted to occur, irrespective of exposure duration. lary, sound at a given level will have less effect if Any single exposure at or above this peak pressure it is near or (especially) beyond the lower or upper is considered to cause tissue injury, regardless of bounds of the functional hearing range than if it is the SPL or SEL of the entire exposure. For each well within that frequency range. It is important marine mammal group, the recommended pres- to note the incremental nature of the frequency- sure-based injury criteria are the same for all sound weighting functions, which approximate the types and are based on the criterion for a single gradual reduction in auditory effect at frequencies pulse. This is a precautionary procedure; pressure outside the range of greatest sensitivity. criteria based on TTS data for nonpulses would Use of such M-frequency-weighting functions yield much higher estimates of the exposure nec- is superior to flat weighting across all frequencies essary for PTS-onset. By proposing, for all cases, because it accounts for known or estimated differ- pressure criteria appropriate to a single pulse, we ences in the frequency response characteristics for protect against the possibility that, for some sound each functional hearing group. At least in the context sources, one or more intense pulses may occasion- of injury criteria, it is superior to frequency-weight- ally be embedded in nonpulse sounds. ing via the inverse-audiogram method as it takes For exposures lacking intense peak pressure into account the expected “flattening” of equal- components, available data indicate that measure- loudness curves at the high exposure levels where ments integrating instantaneous pressure squared TTS and PTS are expected. It is also superior to a over the duration of sound exposure are well corre- “boxcar-type” step function because it more closely lated with the probability of TTS-onset and tissue approximates the gradual roll-off of sensitivity injury. Consequently, for exposures other than below and above the range of optimum sensitivity. those containing intense peak pressure transients, Furthermore, each of the recommended “shallow” SEL is the (or at least one of the) appropriate Marine Mammal Noise Exposure Criteria (A) 435 M-weighting (A) (A) M-weighting Weighting (dB) Weighting Weighting (dB) Weighting high-frequency cetaceans, M hf mid-frequency cetaceans, M high-frequency cetaceans, Mmf hf

mid-frequency cetaceans, Mmf

low-frequency cetaceans, M lf

low-frequency cetaceans, M lf Frequency (Hz) Frequency (Hz) (B) (B) M-weighting (B) M-weighting Weighting (dB) Weighting Weighting (dB) Weighting

pinnipeds in water, Mpw

pinnipeds in air, Mpa pinnipeds in water, Mpw

pinnipeds in air, Mpa

Frequency (Hz)

Figure 1. The M-weighting functions for (A) low-, mid-,Frequency and high-frequency (Hz) cetaceans, as well as for (B) pinnipeds in water and air. 436 Southall et al. metric(s) for estimating TTS-onset and predicting phenomena of concern (TTS, PTS, etc.) will and PTS-onset in humans (ISO, 1990). will not likely occur. This use of SEL is based on the assumption that sounds of equivalent energy will have generally Levels of Noise Effect: similar effects on the auditory systems of exposed Injury and Behavioral Disturbance human subjects, even if they differ in SPL, duration, and/or temporal exposure pattern (Kryter, Direct auditory tissue effects (injury) and behav- 1970; Nielsen et al., 1986; Yost, 1994; NIOSH, ioral disruption are the two categories of noise 1998). Under the equal-energy assumption, at effect that are considered in these marine mammal exposure levels above TTS-onset, each doubling exposure criteria. Chapter 3 summarizes all of sound duration is associated with a 3 dB reduc- available data on the effects of noise on marine tion in the SPL theoretically required to cause the mammal hearing. It also describes how these data same amount of TTS. This relationship has been are applied and extrapolated using precautionary used in the derivation of exposure guidelines for measures to predict auditory injury and to derive humans (e.g., NIOSH, 1998). Numerous authors thresholds and proposed criteria for injury. have questioned the predictive power of using In Chapter 4 and Appendices B & C, we summa- a simplistic total energy approach in all condi- rize the current understanding and available data tions. It fails to account for varying levels and regarding marine mammal behavioral responses temporal patterns of exposure/recovery, among to noise. Chapter 4 includes a quantitative sever- other factors, and will thus likely overestimate ity scale based generally on the NRC’s (2005) the TTS resulting from a complex noise exposure Population Consequences of Acoustic Disturbance (Hamernik & Hsueh, 1991; Hamernik et al., 1993, (PCAD) Model. Chapter 4 also includes a limited 2002; Ahroon et al., 1993; Ward, 1997; Strasser and cautious entry of behavioral-response data et al., 2003). A comparative assessment of TTS as into a matrix of severity scaling as a function of a function of exposure level in mammals, fish, and RL. Currently available data, pooled by functional birds suggests that there are direct relationships hearing group, do not support specific numerical but that the slopes vary among taxa (Smith et al., criteria for the onset of disturbance. Rather, they 2004). The debate over the validity of the equal indicate the context-specificity of behavioral reac- energy “rule” of noise exposure remains unre- tions to noise exposure and point to some general solved, even for humans. conclusions about response severity in certain, Some limited evidence favoring an SEL specific conditions. approach exists for marine mammals, however. Specifically, an equal-energy relationship for TTS-onset appears to hold reasonably well for certain noise exposure types within several mid-frequency cetacean species (Finneran et al., 2002b, 2005a; see “Effects of Noise on Hearing in Marine Animals: TTS Data” section in Chapter 3). A recent study of in- air TTS in a California sea lion (Kastak et al., 2007) illustrates some conditions in which exposures with identical SEL result in considerably different levels of TTS. Nevertheless, because the very limited marine mammal data agree reasonably well (at least as a first-order approximation) with equal-energy predictions, and predictions based on SEL will be precautionary for intermittent exposures, we regard it as appropriate to apply the SEL metric for certain noise exposure criteria until future research indicates an alternate and more specific course. In certain applications, there is much more scientific justification for use of SEL-based criteria than for previous ad hoc SPL criteria (discussed in the “Historical Perspective” section in Chapter 1). In applications involving auditory effects, SEL-based criteria will likely more reliably distinguish cases where Aquatic Mammals 2007, 33(4), 437-445, DOI 10.1578/AM.33.4.2007.437

3. Criteria for Injury: TTS and PTS

The criteria for injury for all marine mammal To date, TTSs measured in marine mammals groups and sound types are received levels (fre- have generally been of small magnitude (mostly quency-weighted where appropriate) that meet the < 10 dB). The onset of TTS has been defined as definition of PTS-onset used here (40 dB-TTS, being a temporary elevation of a hearing thresh- described below). Criteria were derived from mea- old by 6 dB (e.g., Schlundt et al., 2000), although sured or assumed TTS-onset thresholds for each smaller threshold shifts have been demonstrated to marine mammal group plus TTS growth rate esti- be statistically significant with a sufficient number mates (given below). Available TTS data for two of samples (e.g., Kastak et al., 1999; Finneran mid-frequency cetacean species and three species et al., 2005a). Normal threshold variability within of pinnipeds are used as the basis for estimating and between both experimental and control ses- PTS-onset thresholds in all cetaceans (“cetacean sions (no noise) does warrant a TTS-onset crite- procedure” described below; see “PTS-Onset for rion at a level that is always clearly distinguish- Pulses”) and in all pinnipeds (see “PTS-Onset able from that of no effect. We considered a 6 dB for Nonpulse Sounds”), respectively. The pro- TTS sufficient to be recognized as an unequivo- posed injury criteria are presented by sound type cal deviation and thus a sufficient definition of because, for a given sound type, many of the same TTS-onset. extrapolation and summation procedures apply Most of the frequencies used in TTS experi- across marine mammal hearing groups. ments to date are within the flat portions of the A dual-criterion approach was used for the rec- M-weighting functions given here, but not nec- ommended injury criteria. That is, any received essarily within the regions of greatest hearing noise exposure that exceeds either a peak pressure sensitivity. Within the range of best hearing sen- or a SEL criterion for injury is assumed to cause sitivity for a given individual, detection thresholds tissue injury in an exposed marine mammal. Of are generally similar. Within this band, exposures the two measures of sound exposure, peak pres- with the same absolute level but different fre- sures are to be unweighted (i.e., “flat-weighted”), quency are thus similar in terms of their effective whereas SEL metrics are to be M-weighted for the sensation level. Sensation level is the amount (in relevant marine mammal group (Figure 1). In prac- dB) by which an RL exceeds the threshold RL tice, the received noise conditions should be com- for that signal type within a prescribed frequency pared to the two exposure criteria for that sound band (Yost, 2000). If two exposures with identical type and functional hearing group, and the more absolute level are both audible, but one is outside precautionary of the two outcomes accepted. the frequency range of best hearing sensitivity, sensation level will be less for the latter exposure, Effects of Noise on Hearing in and its potential effects will be diminished. By Marine Mammals: TTS Data creating frequency-weighted functions that are flat across virtually the entire functional hearing Noise exposure criteria for auditory injury ideally band, rather than just the region of best sensitivity, should be based on exposures empirically shown to we have made another precautionary decision in induce PTS-onset; however, no such data presently the absence of underlying data on equal-loudness exist for marine mammals. Instead, PTS-onset must functions. be estimated from TTS-onset measurements and Auditory fatigue (i.e., TTS) in mid-frequency from the rate of TTS growth with increasing expo- cetaceans has been measured after exposure to sure levels above the level eliciting TTS-onset. PTS tones, impulsive sounds, and octave-band noise is presumed to be likely if the threshold is reduced (OBN). In pinnipeds, it has been measured upon by ≥ 40 dB (i.e., 40 dB of TTS). We used available exposure to construction noise and OBN in both marine mammal TTS data and precautionary extrap- air and water. olation procedures based on terrestrial mammal data (see “Level of Noise Effect” in Chapter 2) Cetacean TTS to estimate exposures associated with PTS-onset. The sound exposures that elicit TTS in cetaceans Existing TTS measurements for marine mammals have been measured in two mid-frequency spe- are reviewed in detail here since they serve as the cies—bottlenose dolphin and beluga (specific ref- quantitative foundation for the injury criteria. erences given below)—with at least limited data 438 Southall et al. being available for exposures to a single pulse and were inconclusive: one dolphin exhibited a TTS to nonpulsed sounds ranging from 1-s to ~50-min after exposure at 182 dB SPL re: 1 µPa (182 dB duration. There are no published TTS data for any re: 1 µPa2-s) but not at higher exposure levels. The other odontocete cetaceans (either mid- or high- other dolphin experienced no threshold shift after frequency) or for any mysticete cetaceans (low- exposure to maximum SPL levels of 193 dB re: frequency). This review is organized according to 1 µPa (193 dB re: 1 µPa2-s). The shifts occurred the duration of the fatiguing stimulus, with short- most often at frequencies above the fatiguing est exposures discussed first. stimulus. Finneran et al. (2000) exposed two bottlenose Finneran et al. (2005a) measured TTS in bot- dolphins and one beluga to single pulses from an tlenose dolphins exposed to 3 kHz tones with “explosion simulator” (ES). The ES consisted of durations of 1, 2, 4, and 8 s and at various SPL an array of piezoelectric sound projectors that values. Tests were conducted in a quiet pool in generated a pressure waveform resembling that contrast to previous studies in San Diego Bay, from a distant underwater explosion. The pressure where thresholds were masked by broadband waveform was generally similar to waveforms noise. Small amounts of TTS (3 to 6 dB) occurred predicted by the Navy REFMS model (Britt et al., in one dolphin following exposures with SELs of 1991). The ES failed to produce realistic energy 190 to 204 dB re: 1 µPa2-s. These results are con- at frequencies below 1 kHz, however. No substan- sistent with those of Schlundt et al. (2000), indi- tial (i.e., ≥ 6 dB) threshold shifts were observed cating that their results had not been significantly in any of the subjects exposed to a single pulse at affected by the use of masked hearing thresholds the highest received exposure levels (peak: 70 kPa in quantifying TTS. In general, the SEL necessary [10 psi]; peak-to-peak: 221 dB re: 1 µPa (peak-to- for TTS-onset was relatively consistent across the peak); SEL: 179 dB re: 1 µPa2-s)]. range of exposure durations, whereas exposure Finneran et al. (2002b) repeated this experiment SPL values causing TTS-onset tended to decrease using a seismic watergun that produced a single with increasing exposure duration. These results acoustic pulse. Experimental subjects consisted of confirmed that, for these testing conditions (bot- one beluga and one bottlenose dolphin. Measured tlenose dolphins exposed to £ 8-s tones of variable TTS2 was 7 and 6 dB in the beluga at 0.4 and 30 SPL), TTS magnitude was best correlated with kHz, respectively, after exposure to intense single exposure SEL rather than SPL. pulses (peak: 160 kPa [23 psi]; peak-to-peak: 226 Schlundt et al. (2006) reported on the growth dB re: 1 µPa (peak-to-peak); SEL: 186 dB re: 1 and recovery of TTS in a bottlenose dolphin µPa2-s). Thresholds returned to within ± 2 dB of exposed to 3 kHz tones with SPLs up to 200 dB the pre-exposure value within 4 min of exposure. re 1 µPa and durations up to 128 s. The maximum No TTS was observed in the bottlenose dolphin exposure SEL was 217 dB re 1 µPa2-s, which pro- at the highest exposure condition (peak: 207 kPa duced a TTS4 of ~23 dB. All thresholds recovered [30 psi]; peak-to-peak: 228 dB re: 1 µPa (peak-to- to baseline values within 24 h, most within 30 2 peak); SEL: 188 dB re: 1 µPa -s). These studies min. The growth of TTS4 with increasing expo- demonstrated that, for very brief pulses, higher sure SEL was ~1 dB TTS per dB SEL for TTS4 of sound pressures were required to induce TTS ~15 to 18 dB. than had been found for longer tones (discussed Finneran et al. (2007b) measured TTS in a below). bottlenose dolphin after single and multiple expo- Schlundt et al. (2000) reported TTS in five bot- sures to 20 kHz tones. Hearing thresholds were tlenose dolphins and two belugas exposed to 1-s estimated at multiple frequencies (10 to 70 kHz) pure tones (nonpulses). This paper also included both behaviorally and electrophysiologically (by a re-analysis of TTS data from a technical report measurement of multiple auditory steady-state by Ridgway et al. (1997). At frequencies of 3 kHz, responses). Three experiments were performed. 10 kHz, and 20 kHz, SPLs necessary to induce The first two featured single exposures (20 kHz, TTS-onset were 192 to 201 dB re: 1 µPa (SEL: 64-s tones at 185 and 186 dB re 1 µPa). The third 192 to 201 dB re: 1 µPa2-s). The mean exposure featured three 20 kHz, 16-s exposures separated SPL for TTS-onset was 195 dB re: 1 µPa (195 by 11 and 12 min, with a mean SPL of 193 dB re dB re: 1 µPa2-s). Note the appropriately differ- 1 µPa (SD = 0.8 dB). Hearing loss was frequency- ent metrics for the nonpulse sources used in this dependent, with the largest TTS occurring at 30 study and those involving pulses. Also note that kHz, less at 40, and then 20 kHz, and little or no the SPL and SEL values are identical in this spe- TTS at other measured frequencies. AEP thresh- cial case because of the 1-s duration fatiguing old shifts reached 40 to 45 dB and were always stimuli. At 0.4 kHz, no subjects exhibited shifts larger than behavioral shifts, which were 19 to 33 after exposures up to SPL exposures of 193 dB dB. Complete recovery required up to 5 d, with re: 1 µPa (193 dB re: 1 µPa2-s). Data at 75 kHz the recovery rate at 20 kHz being ~2 dB/doubling Marine Mammal Noise Exposure Criteria 439 of time and the rate at 30 and 40 kHz ~5 to 6 dB/ lower SPLs (but similar SEL values) are required doubling of time. to induce TTS when exposure duration is longer. Nachtigall et al. (2003) measured TTS (ca. 20 These findings are generally consistent with mea- min after noise cessation) in a bottlenose dolphin surements in humans and terrestrial mammals and found an average 11 dB shift following a 30- (Kryter, 1970; Harris, 1998; NIOSH, 1998) and min net exposure to OBN with a 7.5 kHz center support the use of SEL to approximate the audi- frequency (CF) (max SPL: 179 dB re: 1 µPa; SEL: tory effects of variable exposure level/duration ~212 to 214 dB re: 1 µPa2-s). The net exposure conditions. Although there are certain (possibly time was calculated as the total experimental time many) conditions under which an explicit “equal- minus the time required for the subject to surface energy rule” may fail to adequately describe the to breathe. Exposure during breathing periods was auditory effects of variable and/or intermittent measured and factored into the SEL measurement. noise exposure, the combined cetacean TTS data No TTS was observed after exposure to the same presented above generally support the use of SEL OBN at maximum SPL values of 165 and 171 dB as a first-order approximation, at least until addi- re: 1 µPa (SEL: ~198 to 200 dB re: 1 µPa2-s and tional data are available. 204 to 206 dB re: 1 µPa2-s, respectively). For cetaceans, published TTS data are limited Using AEP methods, Nachtigall et al. (2004) to the bottlenose dolphin and beluga (Finneran found TTS5 of ca. 4 to 8 dB following nearly 50- et al., 2000, 2002b, 2005a; Schlundt et al., 2000; min exposures to OBN with a CF of 7.5 kHz (max Nachtigall et al., 2003, 2004). Where data exist for SPL: 160 dB re: 1 µPa; SEL: ~193 to 195 dB re: 1 both species, we use the more precautionary result µPa2-s). The difference in results between the two (usually for beluga) to represent TTS-onset for all Nachtigall et al. studies (slightly lower TTS after mid-frequency cetaceans. No published data exist exposure to much lower exposure energy) was on auditory effects of noise in either low- or high- attributed to measuring TTS at a shorter interval frequency cetaceans (an area of needed research after the exposure ended (5 vs ~20 min), and thus as discussed in Chapter 5); therefore, data from allowing less opportunity for hearing recovery. mid-frequency cetaceans are used as surrogates Further, Nachtigall et al. (2004) repeatedly mea- for these two other groups (cetacean proce- sured hearing until recovery had occurred. TTS dure). [We are aware of some very recent TTS recovery was shown to occur within minutes or measurements for an individual harbor porpoise tens of minutes, depending on the amount of the exposed to single pulses (Lucke et al., 2007a) threshold shift. Generally, the recovery rate was but lack sufficient details regarding methodology 1.5 dB of recovery per doubling of time and was and data analysis to directly consider those data consistent in both studies (Nachtigall et al., 2003, quantitatively.] 2004). Low-frequency cetaceans (mysticetes), based The National Research Council (NRC) (1994) on their auditory anatomy (Wartzok & Ketten, identified the need to know whether marine mam- 1999) and ambient noise levels in the frequency 1 mals experience greatest TTS at a frequency ⁄2- ranges they use (Clark & Ellison, 2004), almost octave above the frequency of exposure when certainly have poorer absolute sensitivity (i.e., exposed to loud tones as has been shown in terres- higher thresholds) across much of their hearing trial mammals. Nachtigall et al. (2004) observed range than do the mid-frequency species (but an average threshold shift of 4 dB at 8 kHz but 8 see earlier discussion). Mid-frequency cetaceans dB shift at 16 kHz following the exposure to OBN experience TTS-onset at relatively high levels centered at 7.5 kHz as described above. A similar compared with their absolute hearing sensitivity upward frequency shift also has been observed by at similar frequencies (i.e., high sensation levels), Schlundt et al. (2000) and Finneran et al. (2007b) although it is not known that this is similarly char- for mid-frequency cetaceans. These findings pro- acteristic of low-frequency cetaceans. Our use of vide “strong evidence for fundamental similarities TTS data from mid-frequency cetaceans as a sur- in cochlear micromechanics in marine and land rogate for low-frequency cetaceans presumes that mammals” (NRC, 1994, p. 51) and further justify the two groups have similar auditory mechanisms the judicious extrapolation of TTS data within and are not radically different in relative sensitiv- marine mammal functional hearing groups and ity to fatiguing noise, and that relative differences from terrestrial to marine mammals. in absolute sensitivity between the two groups are The above results provide empirical measures of generally as expected. exposure conditions associated with TTS-onset in For high-frequency species, data from mid- mid-frequency cetaceans exposed to single pulses frequency cetaceans are currently used as a sur- and nonpulses. Combined, these data demonstrate rogate in the absence of available group-specific that, as compared with the exposure levels neces- data. Aside from their extended upper-frequency sary to elicit TTS when exposure duration is short, hearing, high-frequency cetaceans appear to be 440 Southall et al. generally similar in auditory anatomy and hear- kHz OBN, higher sensation levels (up to 95 dB), ing capabilities to mid-frequency species, though and longer exposure durations (up to 50-min net there are some general differences between the exposure). These data largely corroborate previ- groups in sound production. Based on available ous findings concerning TTS-onset in these pin- information and our extrapolation procedures, nipeds. They also support sensation level as a rele- slightly lower estimates of TTS-onset may be war- vant metric for normalizing exposures with similar ranted for high-frequency cetaceans exposed to durations across species having different absolute very high-frequency sounds (≥ 100 kHz). [Also, hearing capabilities. Comparative analyses of preliminary measurements of TTS in a harbor the combined underwater pinniped data (Kastak porpoise exposed to a single airgun pulse (Lucke et al., 2005) indicated that, in the harbor seal, a et al., 2007a) suggest that this species may experi- TTS of ca. 6 dB occurred with 25-min exposure to ence TTS-onset at levels lower than would be sug- 2.5 kHz OBN with SPL of 152 dB re: 1 µPa (SEL: gested by extrapolating from mid-frequency ceta- 183 dB re: 1 µPa2-s). Under the same test condi- ceans. Those results, if confirmed, may provide a tions, a California sea lion showed TTS-onset at more empirical basis for estimating TTS-onset in 174 dB re: 1 µPa (SEL: 206 dB re: 1 µPa2-s), and high-frequency cetaceans and deriving group-spe- a northern elephant seal experienced TTS-onset at cific injury criteria.] 172 dB re: 1 µPa (SEL: 204 dB re: 1 µPa2-s). Data on underwater TTS-onset in pinnipeds Pinniped TTS (Under Water) exposed to pulses are limited to a single study. Sound exposures that elicit TTS in pinnipeds under Finneran et al. (2003) exposed two California water have been measured in individual subjects of sea lions to single underwater pulses from an three pinniped species (harbor seal, California sea arc-gap transducer. They found no measurable lion, and northern elephant seal). Available data TTS following exposures up to 183 dB re: 1 µPa involved exposures to either broadband or octave- (peak-to-peak) (SEL: 163 dB re: 1 µPa2-s). Based band nonpulse noise over durations ranging from on the Kastak et al. (2005) measurements using ~12 min to several hours, plus limited data on nonpulse sounds, the absence of TTS for the sea exposure to underwater pulses. Interestingly, lions following such exposures is generally not there were consistent among-species differences surprising. in the exposure conditions that elicited TTS under water. For the conditions tested, the harbor seal Pinniped TTS (In Air) experienced TTS at lower exposure levels than did Auditory fatigue has been measured following the California sea lion or northern elephant seal. exposure of pinnipeds to single pulses of in-air There are no underwater TTS data for any other sound and to nonpulse noise. pinniped species. Bowles et al. (unpub. data) measured TTS- The following review first considers expo- onset for harbor seals exposed to simulated sonic sure to nonpulses, organized chronologically, booms at peak SPLs of 143 dB re: 20 µPa (peak) followed by a brief discussion of the lone study (SEL: 129 dB re: [20 µPa]2-s). Higher exposure on exposure to pulses. All but one of the studies levels were required to induce TTS-onset in both (Finneran et al., 2003) came from one laboratory California sea lions and northern elephant seals in and from the same individual test subjects. Kastak the same test setting, consistent with the results & Schusterman (1996) reported a TTS of ~8 dB for nonpulse sound both under water and in air. (measured under water at 100 Hz) in a harbor seal Auditory fatigue to airborne sound has also following exposure to broadband airborne, non- been measured in the same three species of pinni- pulse noise from nearby construction. Under con- peds after exposure to nonpulse noise, specifically trolled conditions, Kastak et al. (1999) measured 2.5 kHz CF OBN for 25 min (Kastak et al., 2004a). TTS of ca. 4 to 5 dB in a harbor seal, California The harbor seal experienced ca. 6 dB of TTS at sea lion, and northern elephant seal following 20- 99 dB re: 20 µPa (SEL: 131 dB re: [20 µPa]2-s). to 22-min exposure to underwater OBN centered Onset of TTS was identified in the California at frequencies from 100 Hz to 2 kHz. Exposures sea lion at 122 dB re: 20 µPa (SEL: 154 dB re: were normalized to octave-band levels 60 to 75 [20 µPa]2-s). The northern elephant seal experi- dB above each subject’s hearing threshold (i.e., 60 enced TTS-onset at 121 dB re: 20 µPa (SEL: 163 to 75 dB sensation level) to present similar effec- dB re: [20 µPa]2-s). The subjects in these tests tive exposure conditions to each of the three sub- were the same individuals tested in water (Southall jects. Because of this approach, absolute exposure et al., 2001; Kastak et al., 2005). values (in terms of both SPL and SEL) were quite Kastak et al. (2007) measured TTS-onset and variable depending on subject and test frequency. growth functions for the same California sea lion Subsequently, Kastak et al. (2005) made TTS exposed to a wider range of noise conditions. A measurements on the same subjects using 2.5 total of 192 exposure sequences were conducted Marine Mammal Noise Exposure Criteria 441 with OBN (centered at 2.5 kHz) at levels 94 to Data that would be needed to support alternate 133 dB re: 20 µPa and durations 1.5 to 50 min criteria allowing for presbycusis are lacking. Our net exposure duration. In these more intense noise approach, which uses TTS data from subjects pre- exposures, TTS magnitudes up to 30 dB were sumed to have “normal” hearing as the starting measured at the 2.5 KHz test frequency. Full point for estimating PTS-onset, is precautionary. recovery was observed following all exposures; It is expected to overestimate damaging effects for this occurred rapidly (likely within tens of min- those individuals with diminished absolute hear- utes) for small shifts but took as long as 3 d in the ing sensitivity and/or functional bandwidth prior case of the largest TTS. The estimated SEL value to the exposure. coinciding with TTS-onset across these varied Data on the effects of noise on terrestrial mam- exposure conditions was 159 dB re: (20 µPa)2-s mals can be useful in considering the effects on with a TTS growth function of ~2.5 dB TTS/dB marine mammals in certain conditions (as dis- noise. For TTS exceeding 20 dB, a recovery rate cussed in Chapter 1) because of similarities in of ~2.6 dB/doubling of time was calculated. These morphology and functional dynamics among results generally agree with those of Kastak et al. mammalian cochleae. Under that premise, it is (2004a) but provide a larger data set, across a assumed that a noise exposure capable of induc- wider range of exposure conditions with which ing 40 dB of TTS will cause PTS-onset in marine to derive an empirical TTS-growth function. They mammals. Based on available data for terrestrial also support the conclusion that patterns of TTS mammals, this assumption is likely somewhat growth and recovery are generally similar to those precautionary as there is often complete recov- of terrestrial mammals and that sensation level for ery from TTS of this magnitude or greater. Such the particular species and medium (water or air) is precaution is appropriate, however, because the the appropriate metric for comparing the effects of precise relationship between TTS and PTS is not underwater and aerial noise exposure. fully understood, even for humans and small terrestrial mammals despite hundreds of studies (see Injury from Noise Exposure: Kryter, 1994; Ward, 1997). For marine mammals, PTS-Onset Calculation this presumably complex relationship is unknown, and likely will remain so. The available marine As discussed in Chapter 1, PTS is an irreversible mammal TTS data provide a basis for establish- elevation of the hearing threshold (i.e., a reduction ing a maximum allowable amount of TTS up to in sensitivity) at a specific frequency (Yost, 2000). which PTS is unlikely, however, and for conclud- This permanent change following intense noise ing that PTS is increasingly likely to occur above exposure results from damage or death of inner this point. In using TTS data to estimate the expo- or outer cochlear hair cells. It is often followed by sure that will cause PTS-onset, our approach is to retrograde neuronal losses and persistent chemical acknowledge scientific uncertainty and to err on and metabolic cochlear abnormalities (Saunders the side of overestimating the possibility of PTS et al., 1991; Ward, 1997; Yost, 2000). (i.e., on the side of underestimating the exposure Noise-induced PTS represents tissue injury, but required to cause PTS-onset). TTS does not. Although TTS involves reduced In humans, when TTS2 magnitude for a single hearing sensitivity following exposure, it results exposure exceeds ca. 40 dB, the likelihood of primarily from the fatigue (as opposed to loss) PTS begins to increase substantially (Kryter of cochlear hair cells and supporting structures et al., 1966; Kryter, 1994). Threshold shifts and is, by definition, reversible (Nordmann et al., greater than 40 dB have been demonstrated to 2000). Many mammals, including some pinnipeds be fully recoverable after some period of time in (Kastak et al., 1999, 2005) and cetaceans (e.g., some terrestrial mammal species (human: Ward, Schlundt et al., 2000; Nachtigall et al., 2004), 1959; Ahroon et al., 1996; chinchilla: Miller demonstrate full recovery even after repeated et al., 1971; Mongolian gerbil [Meriones unguicu- TTS. Since TTS represents a temporary change in latus]: Boettcher, 1993). Generally, however, TTS sensitivity without permanent damage to sensory exceeding 40 dB requires a longer recovery time cells or support structures, it is not considered to than smaller shifts, suggesting a higher probability represent tissue injury (Ward, 1997). Instead, the of irreversible damage (Ward, 1970) and possibly onset of tissue injury from noise exposure is con- different underlying mechanisms (Kryter, 1994; sidered here as PTS-onset. Nordman et al., 2000). PTS as a function of age (presbycusis; discussed Our derivation of proposed injury criteria for in Chapter 1) generally appears to be a normal pro- marine mammals begins with measured or esti- cess of aging in mammals (including humans and mated noise exposure conditions associated with marine mammals), but no specific allowance for TTS-onset in cetaceans and pinnipeds. Procedures this is included in our proposed exposure criteria. for estimating PTS-onset, assumed to occur in 442 Southall et al. conditions causing 40 dB of TTS, were derived by PTS-Onset for Nonpulse Sounds combining (1) measured or estimated TTS-onset The peak pressure values assumed to be associated levels in marine mammals and (2) the estimated with onset of injury (PTS-onset) are numerically “growth” of TTS in certain terrestrial mammals equivalent for nonpulse and pulse sounds. Among exposed to increasing noise levels. The general other considerations, this allows for the possibility PTS-onset procedures differ according to sound that isolated pulses could be embedded within the type (pulses and nonpulses), the extent of available predominantly nonpulse sound. information, and required extrapolation. To esti- To estimate the SEL value that would cause mate exposure conditions that will result in PTS- PTS-onset for nonpulse sounds, we used the fol- onset, SEL and SPL were considered separately. lowing procedure. In humans, each added dB of nonpulse noise exposure above TTS-onset PTS-Onset for Pulses results in up to 1.6 dB of additional TTS (Ward Henderson & Hamernik (1986) reported that in et al., 1958, 1959). Assuming this relationship chinchillas exposed to pulses up to a certain level, applies to marine mammals, ~20 dB of additional for each dB of added exposure above that which noise exposure above that causing TTS-onset is caused TTS-onset, a further TTS of about 0.5 dB required to induce PTS-onset (i.e., [40 dB TTS – resulted. For the highest exposure levels, as much 6 dB TTS] / [1.6 dB TTS/dB noise exposure] = as 3 dB of additional TTS was found per additional 21.3 dB of additional noise exposure). We rounded dB of noise. Thus, in extrapolating TTS growth this down to a slightly more precautionary value functions from terrestrial to marine mammals, a of 20 dB of additional noise exposure above TTS- precautionary approach is justified such as using onset. Consequently, to estimate PTS-onset and a slope nearer the upper extreme of this range to derive the SEL injury criteria for nonpulses, we estimate the growth of TTS with exposure level. add 20 dB to the M-weighted SEL values esti- When dealing with pulsed sound, to estimate SEL mated to cause TTS-onset. The lone exception exposures coincident with PTS-onset, we assume to this approach is for pinnipeds in air (discussed a slope of 2.3 dB TTS/dB noise. This is relatively below) where a more precautionary TTS growth precautionary in relation to the data by Henderson rate was used based on a relatively large empirical & Hamernik (1986) on chinchillas. This slope trans- data set (Kastak et al., 2007). lates to an injury criterion (for pulses) that is 15 dB above the SEL of exposures causing TTS-onset Criteria for Injury from a Single Pulse (defined above as 6 dB TTS). That is, PTS-onset (40 dB TTS) is expected to occur on exposure to an As per the “PTS-Onset Calculation” section of this M-weighted SEL 15 dB above that associated with chapter, the recommended criteria for injury from TTS-onset ([40 dB TTS – 6 dB TTS] / [2.3 dB TTS/ exposure to a single pulse, expressed in terms of dB noise exposure] ª 15 dB noise exposure above peak pressure, are TTS-onset levels plus 6 dB of TTS-onset). additional exposure. In terms of SEL, the recom- In terms of sound pressure, TTS-onset thresh- mended criteria are TTS-onset levels plus 15 dB olds in marine mammals, particularly cetaceans, of additional exposure. are quite high (see above). The predicted PTS- For all cetaceans exposed to pulses, the data onset values would be very high (perhaps unreal- of Finneran et al. (2002b) were used as the basis istically so as they would approach the cavitation for estimating exposures that would lead to TTS- limit of water) if the aforementioned 15 dB dif- onset (and, consequently, PTS-onset). They esti- ference between TTS-onset and PTS-onset were mated that, in a beluga exposed to a single pulse, assumed. Consequently, an additional precaution- TTS-onset occurred with unweighted peak levels ary measure was applied by arbitrarily assuming of 224 dB re: 1 µPa (peak) and 186 dB re: 1 µPa2-s. that the pressure difference between TTS-onset The latter is equivalent to a weighted (Mmf) SEL and PTS-onset for pulses might be just 6 dB. This exposure of 183 dB re: 1 µPa2-s as some of the results in a TTS “growth” relationship of 6 dB energy in the pulse was at low frequencies to TTS/dB noise (i.e., [40 dB TTS – 6 dB TTS] / [6 which the beluga is less sensitive. Adding 6 dB dB TTS/dB noise exposure] ª 6 dB noise expo- to the former (224 dB) values, the pressure cri- sure above TTS-onset). That is an extremely con- terion for injury for mid-frequency cetaceans is servative slope function given that it is double the therefore 230 dB re: 1 µPa (peak) (Table 3, Cell highest rate found in chinchillas by Henderson & 4). Adding 15 dB to the latter (183 dB) value, Hamernik (1986). This 6 dB of added exposure, the M-weighted SEL injury criterion is 198 dB above the exposure eliciting TTS-onset, essen- re: 1 µPa2-s (Table 3, Cell 4). These results are tially establishes a proposed (unweighted) peak- assumed to apply (see cetacean procedure, p. 439) pressure ceiling value for all sound types. to low- and perhaps high-frequency cetaceans (Table 3, Cells 1 & 7, respectively) as well as to Marine Mammal Noise Exposure Criteria 443 mid-frequency cetaceans. These injury criteria, 2005; Southall et al., 2001; Schusterman expressed in SEL, are slightly more precautionary et al., 2003 vs Finneran et al., 2000, 2005a; than, but generally consistent with, Ketten’s 1998 Schlundt et al., 2000; Nachtigall et al., 2003, prediction (pers. comm.) that 30% of individual 2004). Assuming that this difference for cetaceans exposed to pulses with an SEL of 205 nonpulse sounds exists for pulses as well, dB re: 1 µPa2-s would experience PTS. TTS-onset in pinnipeds exposed to single For pinnipeds in water, there are no empirical underwater pulses is estimated to occur at data concerning the levels of single pulses that a peak pressure of 212 dB re: 1 µPa (peak) would lead to TTS-onset. At least for the California and/or an SEL exposure of 171 dB re: 1 sea lion, the required exposure is expected to be µPa2-s. Each of these metrics is 12 dB less greater than 183 dB re: 1 µPa (peak) and 163 dB than the comparable value for mid-frequency re: 1 µPa2-s) because Finneran et al. (2003) found cetaceans (see Finneran et al., 2002b, and no TTS in two California sea lions following such above). exposures. In the absence of specific data on the (3) As per the “PTS-onset Procedure” (discussed level of a sound pulse that would cause TTS-onset earlier), we added 6 dB to the former (212 for pinnipeds in water, we used a three-step pro- dB) value to derive the recommended injury cess to estimate this value: pressure criterion of 218 dB re: 1 µPa (peak) (1) We began with the Finneran et al. (2002b) (unweighted) for pinnipeds in water exposed data on TTS-onset from single pulse expo- to a single pulse. Similarly, we added 15 dB sures in a mid-frequency cetacean. TTS- to the latter value (171 dB) to derive the rec- onset occurred with a peak pressure of 224 ommended M-weighted SEL injury criterion 2 dB re: 1 µPa (peak) and Mmf-weighted SEL of 186 dB re: 1 µPa -s (Table 3, Cell 10). of 183 dB re: 1 µPa2-s. These proposed criteria are likely precaution- (2) We assumed that the known pinniped-to- ary because the harbor seal is the most sen- cetacean difference in TTS-onset upon sitive pinniped species tested to date, based exposure to nonpulse sounds would also on results from a single individual (Kastak apply (in a relative sense) to pulses. et al., 1999, 2005). Specifically, with nonpulse sounds, harbor For pinnipeds in air exposed to a single sound seals experience TTS-onset at ca. 12 dB pulse, the proposed criteria for injury were lower RLs than do belugas (i.e., 183 vs based on measurements by Bowles et al. (unpub. 195 dB re: 1 µPa2-s; Kastak et al., 1999, data), which indicated that TTS-onset in harbor

Table 3. Proposed injury criteria for individual marine mammals exposed to “discrete” noise events (either single or multiple exposures within a 24-h period; see Chapter 2)

Sound type Marine mammal group Single pulses Multiple pulses Nonpulses

Low-frequency cetaceans Cell 1 Cell 2 Cell 3 Sound pressure level 230 dB re: 1 µPa (peak) (flat) 230 dB re: 1 µPa (peak) (flat) 230 dB re: 1 µPa (peak) (flat) 2 2 2 Sound exposure level 198 dB re: 1 µPa -s (Mlf) 198 dB re: 1 µPa -s (Mlf) 215 dB re: 1 µPa -s (Mlf) Mid-frequency cetaceans Cell 4 Cell 5 Cell 6 Sound pressure level 230 dB re: 1 µPa (peak) (flat) 230 dB re: 1 µPa (peak) (flat) 230 dB re: 1 µPa (peak) (flat) 2 2 2 Sound exposure level 198 dB re: 1 µPa -s (Mmf) 198 dB re: 1 µPa -s (Mmf) 215 dB re: 1 µPa -s (Mmf) High-frequency cetaceans Cell 7 Cell 8 Cell 9 Sound pressure level 230 dB re: 1 µPa (peak) (flat) 230 dB re: 1 µPa (peak) (flat) 230 dB re: 1 µPa (peak) (flat) 2 2 2 Sound exposure level 198 dB re: 1 µPa -s (Mhf) 198 dB re: 1 µPa -s (Mhf) 215 dB re: 1 µPa -s (Mhf) Pinnipeds (in water) Cell 10 Cell 11 Cell 12 Sound pressure level 218 dB re: 1 µPa (peak) (flat) 218 dB re: 1 µPa (peak) (flat) 218 dB re: 1 µPa (peak) (flat) 2 2 2 Sound exposure level 186 dB re: 1 µPa -s (Mpw) 186 dB re: 1 µPa -s (Mpw) 203 dB re: 1 µPa -s (Mpw) Pinnipeds (in air) Cell 13 Cell 14 Cell 15 Sound pressure level 149 dB re: 20 µPa (peak) (flat) 149 dB re: 20 µPa (peak) (flat) 149 dB re: 20 µPa (peak) (flat) 2 2 2 Sound exposure level 144 dB re: (20 µPa) -s (Mpa) 144 dB re: (20 µPa) -s (Mpa) 144.5 dB re: (20 µPa) -s (Mpa) Note: All criteria in the “Sound pressure level” lines are based on the peak pressure known or assumed to elicit TTS-onset, plus 6 dB. Criteria in the “Sound exposure level” lines are based on the SEL eliciting TTS-onset plus (1) 15 dB for any type of marine mammal exposed to single or multiple pulses, (2) 20 dB for cetaceans or pinnipeds in water exposed to nonpulses, or (3) 13.5 dB for pinnipeds in air exposed to nonpulses. See text for details and derivation. 444 Southall et al. seals occurs following exposure to 143 dB re: high peak pressure transients embedded within 20 µPa (peak) and 129 dB re: (20 µPa)2-s. As for exposures of longer duration but lower-pressure underwater exposures to nonpulse sounds (Kastak magnitude. There are related limitations with SEL et al., 1999, 2005), higher exposure levels were in that temporal integration is involved. required to induce TTS in California sea lions and To account for the potentially damaging aspects northern elephant seals. Consequently, using harbor of high-pressure transients embedded within seal TTS data to establish injury criteria for expo- nonpulse exposures, a precautionary approach sure to a single aerial pulse in pinnipeds is likely a was taken, and the same peak pressure criterion precautionary approximation. Based on these esti- for injury proposed for single pulses is also rec- mates of peak pressure and SEL associated with ommended as the criterion for multiple pulses in TTS-onset, plus 6 dB and 15 dB, respectively, to all functional hearing groups. Thus, if any compo- estimate PTS-onset, the injury criteria for pinni- nent of a nonpulse exposure (unweighted) exceeds peds exposed to a single aerial pulse are 149 dB re: the peak pressure criterion, injury is assumed to 20 µPa (peak) (unweighted) and 144 dB re: (20 µPa)2-s, occur. We expect that only rarely will the injury M-weighted (Table 3, Cell 13). pressure criterion for nonpulse sound be exceeded if the injury SEL criterion is not exceeded (i.e., Criteria for Injury from Multiple Pulses the SEL criterion will be the effective criterion in most exposure conditions). For all marine mammal groups, the recommended For nonpulsed sounds, the recommended SEL criteria for exposure to multiple pulses, expressed criteria for injury (PTS-onset) are M-weighted in both SPL and SEL units, were numerically exposures 20 dB higher than those required identical to the criteria for a single pulse. Any for TTS-onset (see “PTS-Onset Calculation: exposure in a series that exceeds the peak pressure Nonpulses”). Injury SEL criteria for multiple non- criterion would be considered potentially injuri- pulses are numerically identical to those for single ous. In addition, the cumulative SEL for multiple nonpulses for all hearing groups. We make no exposures should be calculated using the summa- distinction between single and multiple nonpulses tion technique described in Chapter 1 (Appendix except that the cumulative SEL for multiple expo- A, eq. 5). The resulting SEL value for multiple sures is calculated as described in Chapter 1 and pulses is then compared to the SEL injury crite- Appendix A, eq. 5. rion for a single pulse in the same functional hear- For all cetaceans exposed to nonpulses, the rec- ing group. As for the single pulse criteria, peak ommended pressure criterion for injury is 230 dB pressures are unweighted (i.e., “flat-weighted”), re: 1 µPa (peak) (Table 3, Cells 3, 6, & 9), the same but SEL should be weighted by the appropriate criterion as for single pulses in these functional M-weighting function (Figure 1). hearing groups. Injury SEL criteria are based on For cetaceans, the proposed criteria for injury TTS data for mid-frequency species and extrapo- by multiple pulses are therefore 230 dB re: lated to the other cetacean groups (see cetacean 1 µPa (peak) and, following summation, 198 procedure, p. 439). The SEL criterion for non- dB re: 1 µPa2-s in terms of SEL (Table 3, Cells pulse injury in cetaceans is calculated to be an M- 2, 5 & 8). As for single pulses, this approach is weighted exposure of 215 dB re: 1 µPa2-s (Table 3, considered precautionary for mid- and low-fre- Cells 3, 6 & 9). This is based on 195 dB re: 1 µPa2-s quency species, but some caution is warranted in as an estimate of TTS-onset in mid-frequency ceta- applying it to high-frequency species (cf. Lucke ceans (Finneran et al., 2002b, 2005a; Schlundt et al., 2007a). et al., 2000; Nachtigall et al., 2003, 2004) plus 20 Following the same logic, the proposed injury dB to estimate PTS-onset. Applying this approach pressure criterion for pinnipeds in water exposed to low-frequency cetaceans is considered pre- to multiple pulses is 218 dB re: 1 µPa (peak) and cautionary, but some caution may be warranted the injury SEL criterion is 186 dB re: 1 µPa2-s in extrapolating to high-frequency cetaceans (cf. (Table 3, Cell 11). For pinnipeds in air, the pro- single-pulse data of Lucke et al., 2007a). posed injury pressure criterion for multiple pulses We note that special injury criteria, different is 149 dB re: 20 µPa (peak) and the injury SEL cri- from those shown in Cell 6 of Table 3, are likely terion is 144 dB re: (20 µPa)2-s (Table 3, Cell 14). needed for exposure of beaked whale species to nonpulses. Under certain conditions, beaked Criteria for Injury from Nonpulses whales of several species (primarily Cuvier’s, Blainville’s, and Gervais’ beaked whales) have SPL and SEL appear to be appropriate metrics stranded in the presence of sound signals from for quantifying exposure to nonpulse sounds. But tactical mid-frequency military sonars (Frantzis, because SPL measures involve averaging over 1998; Evans & England, 2001; Fernández et al., some duration, they may not adequately quantify 2005; Cox et al., 2006). There have been other Marine Mammal Noise Exposure Criteria 445 incidents (e.g., NMFS, 2005; Hohn et al., 2006) For pinnipeds in air exposed to nonpulse sound, where marine mammal strandings or other anom- the injury pressure criterion is a flat-weighted value alous events involving other marine mammal of 149 dB re: 20 µPa (peak) (Table 3, Cell 15), con- species have occurred in association with sistent with that for pulses. The SEL criterion is mid-frequency sonar operations. They are, how- based on occurrence of TTS-onset in a harbor seal ever, much more ambiguous, difficult to interpret, exposed in air to 131 dB re: (20 µPa)2-s (Kastak and appear fundamentally different than the spe- et al., 2004a). In estimating the exposure that cific beaked whale events. Little is known about would cause PTS-onset, we use empirical mea- the exposure levels, or about the positions or reac- surements of TTS growth as a function of expo- tions of other marine mammals in the areas during sure SEL in a California sea lion. Kastak et al. mid-frequency sonar training operations. The (2007) found a TTS growth rate of 2.5 dB TTS/dB most extreme, ultimate response of some beaked noise based on nearly 200 exposure sequences whales in specific conditions (stranding and sub- involving variable exposure level and duration sequent death) does not appear to be typical of conditions. This growth rate implies a 13.5 dB dif- other marine mammals. ference between TTS- and PTS-onset as opposed Sound fields resulting from sonar operations to the 20 dB value used for marine mammals in have been modeled in several of the above cases water. When the 13.5 dB figure is added to the (e.g., the 1996 event in Greece and the 2000 TTS-onset value for harbor seals (131 dB re: [20 2 event in the Bahamas), and it is possible to at µPa] -s), we obtain a proposed Mpa-weighted SEL least roughly bound the estimated exposures for criterion of 144.5 dB re: (20 µPa)2-s for pinnipeds some of the individuals that stranded (D’Spain in air (Table 3, Cell 15). et al., 2006). While the specific exposure levels The use for all pinnipeds of harbor seal TTS will never be quantitatively known, it does appear data combined with the sea lion growth function likely that the exposures for some of the beaked would be an exceedingly precautionary procedure. whales that stranded were below the criteria for This PTS-onset estimate is considerably below tissue injury proposed above. the TTS-onset estimates for both the northern ele- Consequently, the general injury criteria do not phant seal (163 dB re: [20 µPa]2-s; Kastak et al., seem sufficiently precautionary for beaked whales 2004a) and the California sea lion (159 dB re: [20 exposed to some nonpulse sounds under certain µPa]2-s; Kastak et al., 2007). Applying the TTS conditions. Empirical data to support discrete, growth function of 2.5 dB TTS/dB noise from science-based injury criteria specific to beaked Kastak et al. (2007) to these TTS-onset estimates whales exposed to tactical, mid-frequency, mili- would yield PTS-onset values of 172.5 and 176.5 tary sonar are lacking, however. Regulatory agen- dB re: (20 µPa)2-s for the California sea lion and cies should consider adopting provisional injury northern elephant seal, respectively. As noted in criteria for beaked whales exposed to active, mid- the “Overview,” where specific data are available frequency, military sonars that are lower (in terms for the species or genus of concern, it is appropri- of RL) than the criteria used for mid-frequency ate for criteria to be based on those data rather than cetaceans and nonpulse sources generally. Of the generalized criteria that are recommended for foremost importance, specific studies are needed the overall group of marine mammals. to better define the mechanism of injury in these apparently sensitive species (see Chapter 5). For pinnipeds in water, the recommended pressure criterion for injury from exposure to nonpulse sounds is the same value as applied to pulses: an unweighted value of 218 dB re: 1 µPa (peak) (Table 3, Cell 12). To derive the associated SEL criterion, we began with the measured nonpulse exposure eliciting TTS-onset in a harbor seal, 183 dB re: 1 µPa2-s (Kastak et al., 1999, 2005). This is likely a precautionary choice because SEL values ~10 to 20 dB higher were required to induce TTS- onset in a California sea lion and a northern elephant seal. We assume that 20 dB of additional noise exposure will elicit PTS-onset (see “Effects of Noise on Hearing” section of this chapter), resulting in an Mpw-weighted SEL criterion of 203 dB re: 1 µPa2-s for pinnipeds exposed to nonpulse sound in water (Table 3, Cell 12). Aquatic Mammals 2007, 33(4), 446-473, DOI 10.1578/AM.33.4.2007.446

4. Criteria for Behavioral Disturbance

Behavioral reactions to acoustic exposure are been quantified (e.g., Schultz, 1978; Angerer generally more variable, context-dependent, and et al., 1991) and used to develop dose-response less predictable than effects of noise exposure on relationships for noise exposure in human com- hearing or physiology. Animals detecting one kind munity noise applications (see Kryter, 1994, of signal may simply orient to hear it, whereas Chapter 10). Practical issues (e.g., difficulties in they might panic and flee for many hours upon training nonverbal species to provide interpretable hearing a different sound, potentially even one responses and questions about the applicabil- that is quieter, but with some particular signifi- ity of captive data to free-ranging animals) have cance to the animal. The conservation of cochlear prevented this or similar approaches from being properties across mammals justifies judicious applied to marine mammals. Instead, most efforts application of auditory data from terrestrial mam- have focused on analyses of observable reactions mals where data on marine mammals are missing. to known noise exposure. However, the context-specificity of behavioral For most free-ranging marine mammals, behav- responses in animals generally makes extrapola- ioral responses are often difficult to observe. Also, tion of behavioral data inappropriate. Assessing precise measurements of received noise exposure the severity of behavioral disturbance must conse- and other relevant variables (e.g., movement of quently rely more on empirical studies with care- source, presence of high-frequency harmonics fully controlled acoustic, contextual, and response indicating relative proximity, and prior experience variables than on extrapolations based on shared of exposed individuals) can be difficult to obtain. phylogeny or morphology. Only a subset of disturbance studies have esti- Considerable research has been conducted mated received sound levels, and only a very small to describe the behavioral responses of marine number have actually measured RLs at the subject. mammals to various sound sources. Fortunately, Further, exposures are often complicated by mul- at least limited data are available on behavioral tiple contextual covariants such as the presence of responses by each of the five functional marine vessels and/or humans close to subjects either for mammal groups to each sound type considered observation or to deploy playback sources (e.g., here. As evident in the extensive literature review Frankel & Clark, 1998). Interpretation of the summarized below and described in detail in observed results is highly limited by uncertainty Appendices B & C, however, very few studies as to what does and does not constitute a mean- involving sufficient controls and measurements ingful response. Also, most behavioral-response exist. In addition, the influence of experience with studies have concentrated on short-term and local- the experimental stimulus or similar sounds has ized behavioral changes whose relevance to indi- usually been unknown. vidual well-being and fitness, let alone population To assess and quantify adverse behavioral parameters, is likely to be low. effects of noise exposure, a metric for the impact A further complication is that observations from such changes might have on critical biological laboratory and field settings cannot be directly parameters such as growth, survival, and reproduc- equated. Laboratory studies are usually precise in tion is needed. Behavioral disturbances that affect quantifying exposures and responses. The expo- these vital rates have been identified as particularly sure conditions very rarely approximate those in important in assessing the significance of noise the field, however, and measured behavior may exposure (NRC, 2005). Unfortunately, as Wartzok have little or no relevance to the ways in which et al. (2004) pointed out, no such metric is cur- unconstrained, untrained wild animals respond. rently available, and it is likely to take decades of Conversely, field measurements may address research to provide the analytical framework and responses of free-ranging mammals to a specific empirical results needed to create such a metric, if sound source but often lack adequate controls and one in fact is ultimately even viable. precision in quantifying acoustic exposures and In humans, a common and useful means of esti- responses. Clearly, there is a need for a framework mating behavioral disturbance from noise expo- to integrate laboratory and field data, despite the sure is to ask individuals to rate or describe the challenges in constructing that framework. degree to which various sounds are bothersome. Another difficult issue concerns the appropri- Subjective perception of noise “annoyance” has ate noise exposure metric for assessing behavioral Marine Mammal Noise Exposure Criteria 447 reactions. Most bioacousticians recommend was done only for studies that provided sufficient reporting several different measures of acoustic information on source and environmental charac- exposure, such as SPL and SEL (as in Blackwell teristics. Our approach does not presume that SPL et al., 2004a, 2004b). Of the many studies that is necessarily the acoustic metric best correlated report source SPL, relatively few specify whether with behavioral changes (significant or otherwise). RMS, peak, peak-to-peak, or other sound pressure In particular, SPL fails to account for the dura- measurements were made. Additionally, relatively tion of exposure whereas this is captured using few papers provide sufficient relevant informa- SEL. SPL is the metric that has most often been tion about sound transmission loss in the study measured or estimated during disturbance studies, area. A small number of papers report estimates or however. Thus, it is currently the best metric with direct measurements of received SPL, but very few which to assess the available behavioral response report SEL. The appropriate measure for predict- data. Future studies should report the full range of ing probability of a behavioral response is likely standard acoustic measurements appropriate to the to vary depending upon the behavioral context. For sound source in question and should also include example, if an animal interprets a sound as indicat- measurements of background noise levels in order ing the presence of a predator, a short faint signal to assess signal-to-noise ratios. These additional may evoke as strong a response as a longer, strong data should eventually clarify which exposure sound. But if an animal is responding to a context- metrics best predict different kinds of behavioral neutral stimulus that is merely annoying, the prob- responses and which are most appropriate for use ability of response may well scale with duration and in policy guidelines applicable to different types level of exposure. of noise exposures. It is difficult to define the SEL for individual Beyond the discussion of which metric is most animals in the wild exposed to a specific sound appropriate to quantify the exposure level of a source. Ideally, received SEL over the animal’s sound, it is recognized that many other variables full duration of exposure would be measured affect the nature and extent of responses to a par- (Madsen et al., 2005a). We expect that the prob- ticular stimulus. Wartzok et al. (2004) discussed ability and severity of some kinds of response will in detail the highly variable response of belugas vary with duration as well as level of exposure; exposed to similar sounds in different locations— for those situations, an SEL metric may be most for example, Frost et al. (1984) vs Finley et al. appropriate. However, the most practical way to (1990). In those cases, it appears that the context look for consistent patterns of response as a func- (recent experience of the belugas with the sound tion of RL and duration, given the current state stimulus, their current activity, and their motiva- of science, is to evaluate how different animals tion to remain or leave) was much more significant respond to similar sound sources used in similar in governing their behavioral responses. Similarly, contexts. For example, the relationship between reactions of bowhead whales to seismic airgun acoustic exposure and animal responses is likely sounds depend on whether the whales are feeding to be quite different for mammals exposed to (Richardson et al., 1986; Miller et al., 2005) vs sounds from a slow-moving seismic survey vessel migrating (Richardson et al., 1999). Reactions of operating in a given habitat for many weeks as bowheads and other cetaceans to boats depend on compared with a torpedo transmitting directional whether the boats are moving or stationary, and on high-frequency sonar pings as it transits an area the relative movement of the boat and the whale once at many tens of knots. Similarly, an acous- (see Richardson et al., 1995; Wartzok et al., 2004). tic harassment device placed in a habitat for years In these and some other cases, simple metrics of is likely to evoke a different severity of response exposure (without considering context) will not than would several short pulses at a comparable reliably predict the type and severity of behav- SPL. Until more controlled studies become avail- ioral response(s). Our analyses here, which use able with calibrated measurements of RLs and exposure SPL alone, are admittedly rudimentary ambient noise measurements (including signal-to- and limited by the fact that—for most species and noise ratio), the best way to predict likely effects situations—current data do not support a more will be a common-sense approach that assesses sophisticated approach. available data from situations similar to the situ- Another key consideration involves differ- ation of concern. entiating brief, minor, biologically unimportant Considering all of these limitations and the reactions from profound, sustained, and/or bio- nature of the available data, as a practical matter, logically meaningful responses related to growth, we use SPL as the acoustic metric for the behav- survival, and reproduction. The biological rel- ioral analyses given below. Where necessary and evance of a behavioral response to noise expo- appropriate, simple assumptions regarding trans- sure may depend in part on how long it persists. mission loss were applied to predict RLs. This Many mammals perform vital functions (e.g., 448 Southall et al. feeding, resting, traveling, socializing) on a diel pulses and summarize the literature considered in cycle. Repeated or sustained disruption of these the severity scaling analyses for multiple pulses functions is more likely to have a demonstrable and nonpulse sources. More detailed discus- effect on vital rates than a single, brief disturbance sions of this literature are given in Appendix B episode. The NRC (2005) argued that, although for multiple pulses and Appendix C for nonpulse the duration of behaviors likely to affect vital rates sources. is believed to be particularly significant, current scientific knowledge is insufficient to support an Behavioral Response Data Analysis Procedures: analytical treatment of biological significance and Disturbance Criteria and Severity Scaling ad hoc criteria are needed in the interim. Here, substantive behavioral reactions to noise expo- Single Pulse sure (such as disruption of critical life functions, Due to the transient nature of a single pulse, the displacement, or avoidance of important habitat) most severe behavioral reactions will usually be are considered more likely to be significant if they temporary responses, such as startle, rather than last more than one diel period, or recur on subse- prolonged effects, such as modified habitat utili- quent days. Consequently, a reaction lasting less zation. A transient behavioral response to a single than 24 h and not recurring on subsequent days is pulse is unlikely to result in demonstrable effects not regarded as particularly severe unless it could on individual growth, survival, or reproduction. directly affect survival or reproduction. Consequently, for the unique condition of a single In the absence of an overarching means of quan- pulse, an auditory effect is used as a de facto dis- tifying the biological significance of an effect, we turbance criterion. It is assumed that significant had to adopt a more descriptive method of assess- behavioral disturbance might occur if noise expo- ing the range of possible responses and the sever- sure is sufficient to have a measurable transient ity of behavioral response. To do this, we took effect on hearing (i.e., TTS-onset). Although TTS two different approaches. For the unusual case of is not a behavioral effect per se, this approach is exposure to a single pulse, where the exposure is used because any compromise, even temporar- very brief and responses are usually brief as well, ily, to hearing functions has the potential to affect a procedure for determining recommended criteria vital rates by interfering with essential communi- is identified and applied. For all other conditions, cation and/or detection capabilities. This approach an ordinal and subjective response severity scal- is expected to be precautionary because TTS at ing was developed and applied to those data on onset levels is unlikely to last a full diel cycle or to marine mammal behavioral responses for which have serious biological consequences during the estimates of received SPL were available. These time TTS persists. Because this approach is based analyses were limited to peer-reviewed literature on an auditory phenomenon, the exposure criteria (published or in press) and peer-reviewed techni- can reasonably be developed for entire functional cal reports, with some exceptions on a case-by- hearing groups (as in the injury criteria) rather case basis. than on a species-by-species basis. The extrapo- The severity scale was designed to provide lation procedures used to estimate TTS-onset for some analytical basis for assessing biological single pulse exposures for each hearing group are significance, but it had to be rooted in the kinds described in Chapter 3 (see the “Injury from Noise of descriptions provided in the available scien- Exposure: PTS-Onset Calculation” section). tific literature. Our current understanding of the A dual-criterion approach (using both SPL influences of contextual variables on behavioral [peak] and SEL) was used to determine behavioral responses in free-ranging marine mammals is criteria for a single pulse exposure. Consistent very limited. The analyses presented here should with the injury criteria, which also were based on be considered with these cautions and caveats in auditory fatigue data, RLs that exceed the criterion mind. Our goal was to review the relevant scien- for either metric are considered to have greater tific literature, tally behavioral effects by the type potential to elicit a biologically significant behav- of acoustic exposure for each category of marine ioral response. Proposed criteria for exposure to mammal and sound type, and draw what conclu- a single pulse for each functional hearing group sions were appropriate based on the information are given in the next section. These criteria are the available. TTS-onset thresholds discussed in Chapter 3. The general procedures for determining behav- An exception was made in any case where ioral response exposure criteria for a single pulse, behavioral data indicate that a single pulse expo- and for conducting the severity analyses of indi- sure may elicit a sustained and potentially signifi- vidual behavioral responses vs received SPL, are cant response when the RL is below that required discussed in the next section. Subsequent sections for TTS-onset. This can apply to hauled-out pin- discuss the exposure criterion levels for single nipeds, which sometimes stampede from a beach Marine Mammal Noise Exposure Criteria 449 upon exposure to a sonic boom and may not return rudimentary framework for assessing the relative for many hours (e.g., Holst et al., 2005a, 2005b). biological importance of behavioral responses and In cases where such behavioral responses may is likely a closer approximation of reality than pre- result in the injury or death of pups or other indi- vious step-function thresholds (as discussed in the viduals, exposure levels should be considered in “Historical Perspectives” section of Chapter 1). the context of injury criteria. Conversely, if avail- This approach emphasizes that “disturbance” is a able behavioral data indicate that the response graduated, rather than a “yes-or-no,” phenomenon threshold for exposure to a single pulse is above and that some noise-induced changes in behavior the level required for TTS-onset, then the TTS- are more significant than others. We expect that onset level is retained as the behavioral criterion future studies involving multivariate analysis of as a further precautionary procedure. multiple behavioral response variables, multiple measures of acoustic exposure, and multiple con- Multiple Pulses and Nonpulses textual variables will provide a foundation for For all other sound types than single pulses, we more sophisticated interpretations. expect that significant behavioral effects will occur Second, we reviewed available research and more commonly at levels below those involved in observations for each of the five marine mammal temporary or permanent losses of hearing sensi- functional hearing groups exposed to either mul- tivity. This argues against basing threshold criteria tiple pulse or nonpulse sounds (i.e., Cells 2, 3, 5, exclusively on TTS and indicates the need for a 6, 8, 9, 11, 12, 14 & 15 in our matrix of sound paradigm to predict the probability of significant type by animal group). We considered measure- behavioral response as a function of noise expo- ments of behavioral response both in the field sure. However, because of the extreme degree and in the laboratory according to the behavioral of group, species, and individual variability in severity scale. Studies with insufficient informa- behavioral responses in various contexts and con- tion on exposures and/or responses were conditions, it is less appropriate to extrapolate behav- sidered but not included in the severity analysis. ioral effects as opposed to auditory responses. Where individual (and/or groups considered as an The available data on marine mammal behavioral “individual”; see below) behavioral responses and responses to multiple pulse and nonpulse sounds associated received sound levels were reported, are simply too variable and context-specific to jus- the observations were assigned the appropriate tify proposing single disturbance criteria for broad behavioral “score” from Table 4 and the case was categories of taxa and of sounds. included in a severity scoring table for the relevant This should not, however, lead to the conclusion matrix cell. One dimension in this type of table that there are insufficient data to conduct a system- was the behavioral score (defined in Table 4); atic assessment of the likelihood that certain sound the other dimension was the received SPL within exposures will induce behavioral effects of variable 10-dB ranges. Where multiple responses were seriousness in marine mammals. On the contrary, reported for the same individual and/or group in a this field has seen many and accelerating strides study (or where it was possible that this had been in characterizing how certain kinds of sounds can done—pseudoreplication), appropriate fractions affect marine mammal behavior. Quantification of a single observation were assigned to relevant of the severity or significance of these effects will cells in the scoring table. As a result, there are frac- continue to be challenging. However, based on tional responses for some individual and/or group the NRC (2005) model described above in which responses in the tabular severity-scaling forms. behavioral reactions with a greater potential to For example, a single behavioral observation for affect vital rates are of particular concern, a sim- one individual was weighted as equivalent to ten plistic scaling paradigm in which to consider the observations for another individual by assigning available data appears to provide the most justifi- each observation (some potentially in different able way forward at present. RL/severity score bins) of the second individual a First, we developed an ordinal ranking of relative weight of 0.1. behavioral response severity (see Table 4). The Many observations of marine mammals involve intent of this scaling was to delineate those behav- multiple individuals because many species occur iors that are relatively minor and/or brief (scores in large social groups and are followed as a group. 0-3); those with higher potential to affect forag- In this case, if one individual responds to a sound, ing, reproduction, or survival (scores 4-6); and the other group members may respond to the those considered likely to affect these vital rates response as opposed to the sound. In such obser- (scores 7-9). This is an admittedly simplistic vations, the full group was considered to repre- way of scaling the strikingly complex and poorly sent an “individual” (i.e., the group became the understood behavioral patterns of marine mam- unit of analysis). As a precautionary approach, the mals in real-world conditions. It does provide a most severe response by any individual observed 450 Southall et al.

Table 4. Severity scale for ranking observed behavioral responses of free-ranging marine mammals and laboratory subjects to various types of anthropogenic sound

Response Corresponding behaviors Corresponding behaviors score1 (Free-ranging subjects)2 (Laboratory subjects)2

0 - No observable response - No observable response 1 - Brief orientation response (investigation/visual orientation) - No observable response 2 - Moderate or multiple orientation behaviors - No observable negative response; may - Brief or minor cessation/modification of vocal behavior approach sounds as a novel object - Brief or minor change in respiration rates 3 - Prolonged orientation behavior - Minor changes in response to trained - Individual alert behavior behaviors (e.g., delay in stationing, - Minor changes in locomotion speed, direction, and/or dive extended inter-trial intervals) profile but no avoidance of sound source - Moderate change in respiration rate - Minor cessation or modification of vocal behavior (duration < duration of source operation), including the Lombard Effect 4 - Moderate changes in locomotion speed, direction, and/or dive - Moderate changes in response to profile but no avoidance of sound source trained behaviors (e.g., reluctance to - Brief, minor shift in group distribution return to station, long inter-trial - Moderate cessation or modification of vocal behavior (duration intervals) ª duration of source operation) 5 - Extensive or prolonged changes in locomotion speed, direction, - Severe and sustained changes in and/or dive profile but no avoidance of sound source trained behaviors (e.g., breaking away - Moderate shift in group distribution from station during experimental - Change in inter-animal distance and/or group size (aggregation sessions) or separation) - Prolonged cessation or modification of vocal behavior (duration > duration of source operation) 6 - Minor or moderate individual and/or group avoidance of sound - Refusal to initiate trained tasks source - Brief or minor separation of females and dependent offspring - Aggressive behavior related to noise exposure (e.g., tail/flipper slapping, fluke display, jaw clapping/gnashing teeth, abrupt directed movement, bubble clouds) - Extended cessation or modification of vocal behavior - Visible startle response - Brief cessation of reproductive behavior 7 - Extensive or prolonged aggressive behavior - Avoidance of experimental situation - Moderate separation of females and dependent offspring or retreat to refuge area (£ duration of - Clear anti-predator response experiment) - Severe and/or sustained avoidance of sound source - Threatening or attacking the sound - Moderate cessation of reproductive behavior source 8 - Obvious aversion and/or progressive sensitization - Avoidance of or sensitization to exper- - Prolonged or significant separation of females and dependent imental situation or retreat to refuge offspring with disruption of acoustic reunion mechanisms area (> duration of experiment) - Long-term avoidance of area (> source operation) - Prolonged cessation of reproductive behavior 9 - Outright panic, flight, stampede, attack of conspecifics, or - Total avoidance of sound exposure stranding events area and refusal to perform trained - Avoidance behavior related to predator detection behaviors for greater than a day

1Ordinal scores of behavioral response severity are not necessarily equivalent for free-ranging vs laboratory conditions. 2Any single response results in the corresponding score (i.e., all group members and behavioral responses need not be observed). If multiple responses are observed, the one with the highest score is used for analysis. Marine Mammal Noise Exposure Criteria 451 within a group was used as the ranking for the shown in the studies reviewed here. The results whole group. from prior behavioral studies in which these vari- A specific category of behavioral studies was ables are fairly similar to those in the anticipated one in which marine mammal distributions were exposure situation will very likely be the most rel- measured around a sound source during quiet and evant. Information from those studies should be active periods. The available data typically involve most strongly weighted in assessing the likelihood comparisons of the distribution of animals before of significant behavioral disturbance. exposure (“control” or “reference”) vs during exposure (“experimental”); the difference in distribution Criteria for Behavioral Disturbance: Single Pulse of the group was the behavioral response. Using this method, and given equivalent range measure- For all cetaceans exposed to single pulses, the ments for control and experimental observations, criteria were based on the Finneran et al. (2002b) “phantom” RLs for mammals detected during results for TTS-onset in a beluga exposed to a control periods (RLs that would have existed if in single pulse. The unweighted peak sound pressure fact the source was active) can be calculated and values of 224 dB re: 1 µPa (peak) and weighted 2 compared to actual RLs during experimental con- (Mmf) SEL values of 183 dB re: 1 µPa -s are rec- ditions. In this way, the percentage of avoidance ommended as “behavioral” disturbance criteria responses by individuals during the exposure was for mid-frequency cetaceans (Table 5, Cell 4). By then calculated. extrapolation (see cetacean procedure, Chapter For the studies used in this analysis, noise 3, p. 439), the same values were also proposed exposure (including source and RL, frequency, for low- and high-frequency cetaceans (Table 5, duration, duty cycle, and other factors) was either Cells 1 & 7, respectively). The only difference in directly reported or was reasonably estimated the application of these criteria to the three ceta- using simple sound propagation models deemed cean groups is the influence of the respective fre- appropriate for the sources and operational envi- quency-weighting functions for SEL criteria (Mlf ronment. Because of the general lack of precision and Mhf vs Mmf). in many studies and the difficulties in pooling the For pinnipeds exposed to single pulses in water, results from disparate studies here, we pooled the proposed “behavioral” disturbance criteria are individual exposure SPL into 10-dB bins. also the estimated TTS-onset values. For pinni- Our analysis of the available behavioral peds as a whole, these are 212 dB re: 1 µPa (peak) 2 response studies presents raw, individual obser- and weighted (Mpw) SEL of 171 dB re: 1 µPa -s vations of reactions to multiple pulses and non- (Table 5, Cell 10) as discussed in Chapter 3. pulses as a function of exposure RL. The basic For pinnipeds in air, the proposed behavioral output of this procedure is a series of tables, one criteria are based on the strong responses (stam- for each combination of the five marine mammal peding behavior that could injure some indi- functional hearing groups and these two sound viduals or separate mothers from pups) of some types (multiple pulses and nonpulses). The over- species, especially harbor seals, to sonic booms all tally within each cell represents the number of from aircraft and missile launches in certain individuals and/or independent group behavioral conditions (Berg et al., 2001, 2002; Holst et al., responses with estimated and/or measured RL in 2005a, 2005b). No responses resulting in injury the specified 10-dB category. were observed in these specific studies, but the This analysis is intended to provide some behavioral responses were, in some cases, among foundation for judging the degree to which avail- those that would be considered relatively severe able data suggest the existence of dose-response in regards to vital rates. It was therefore deter- relationships between noise exposure and marine mined appropriate to use results from these stud- mammal behavior. An example of such a dose- ies rather than TTS-based thresholds for behav- response function is the Schultz (1978) curve ioral response criteria. The proposed criteria are used to predict growth of human annoyance with 109 dB re: 20 µPa (peak) and frequency-weighted 2 increasing noise level. The reader should note, (Mpa) SEL of 100 dB re: (20 µPa) -s (Table 5, however, that the substantial, acknowledged cave- Cell 13). These levels are substantially below ats and limitations of the current approach, partic- TTS-onset values. They are also probably quite ularly those related to contextual variables other precautionary as behavioral response criteria for than simply exposure level. Any application of the group as a whole, especially for species other the severity analyses given below should carefully than harbor seals where higher exposures were not consider the nature of the available information observed to induce strong (or in some cases any) regarding sound source, species, sex/age class, responses. sound-propagation environment, and especially the overall context of exposure relative to that 452 Southall et al.

Table 5. Proposed behavioral response criteria for individual marine mammals exposed to various sound types; specific threshold levels are proposed for single pulses. See the referenced text sections and tables for severity scale analyses of behavioral responses to multiple pulses and nonpulses.

Sound type Marine mammal group Single pulses Multiple pulses Nonpulses

Low-frequency cetaceans Cell 1 Cell 21 Cell 36 Sound pressure level 224 dB re: 1 µPa (peak) (flat) Tables 6 & 7 Tables 14 & 15 2 Sound exposure level 183 dB re: 1 µPa -s (Mlf) Not applicable Not applicable Mid-frequency cetaceans Cell 4 Cell 52 Cell 67 Sound pressure level 224 dB re: 1 µPa (peak) (flat) Tables 8 & 9 Tables 16 & 17 2 Sound exposure level 183 dB re: 1 µPa -s (Mmf) Not applicable Not applicable High-frequency cetaceans Cell 7 Cell 83 Cell 98 Sound pressure level 224 dB re: 1 µPa (peak) (flat) [Tables 18 & 19] Tables 18 & 19 2 Sound exposure level 183 dB re: 1 µPa -s (Mhf) Not applicable Not applicable Pinnipeds (in water) Cell 10 Cell 114 Cell 129 Sound pressure level 212 dB re: 1 µPa (peak) (flat) Tables 10 & 11 Tables 20 & 21 2 Sound exposure level 171 dB re: 1 µPa -s (Mpw) Not applicable Not applicable Pinnipeds (in air) Cell 13 Cell 145 Cell 1510 Sound pressure level 109 dB re: 20 µPa (peak) (flat) Tables 12 & 13 Tables 22 & 23 2 Sound exposure level 100 dB re: (20 µPa) -s (Mpa) Not applicable Not applicable

1 “Low-Frequency Cetaceans/Multiple Pulses (Cell 2)” section 2 “Mid-Frequency Cetaceans/Multiple Pulses (Cell 5)” section 3 “High-Frequency Cetaceans/Multiple Pulses (Cell 8)” section 4 “Pinnipeds in Water/Multiple Pulses (Cell 11)” section 5 “Pinnipeds in Air/Multiple Pulses (Cell 14)” section 6 “Low-Frequency Cetaceans/Nonpulses (Cell 3)” section 7 “Mid-Frequency Cetaceans/Nonpulses (Cell 6)” section 8 “High-Frequency Cetaceans/Nonpulses (Cell 9)” section 9 “Pinnipeds in Water/Nonpulses (Cell 12)” section 10 “Pinnipeds in Air/Nonpulses (Cell 15)” section

Behavioral Response Severity Scaling: contextual variables. A few of these “excluded” Multiple Pulses studies are listed at the bottom of Table 6. Table 7 shows the results of the severity scaling analy- Low-Frequency Cetaceans/Multiple Pulses (Cell 2) ses of individual and/or group responses, con- Numerous field observations have been made sidering the studies deemed to contain sufficient of low-frequency cetaceans reacting to multiple data on exposure conditions and behavioral pulses either incidentally during ongoing opera- responses. For migrating bowhead whales, the tions or intentionally during experiments. A mod- onset of significant behavioral disturbance from erate number of species and experimental condi- multiple pulses occurred at RLs (RMS over pulse tions have been considered, but the sources have duration) around 120 dB re: 1 µPa (Richardson usually been seismic airgun arrays. Some of the et al., 1999). For all other low-frequency cetaceans studies focused on migrating whales seen from (including bowhead whales not engaged in migra- fixed observation platforms or in/near migratory tion), this onset was at RLs around 140 to 160 dB corridors. This approach minimizes pseudorepli- re: 1 µPa (Malme et al., 1983, 1984; Richardson cation without the need for identifying individuals et al., 1986; Ljungblad et al., 1988; Todd et al., because individuals are unlikely to pass observers 1996; McCauley et al., 1998, 2000) or perhaps more than once. higher (Miller et al., 2005). There is essentially no Table 6 summarizes the methods used to obtain overlap in the RLs associated with onset of behav- acoustic measurements and observations of behavioral responses by members of these two groups ioral or distributional responses (see Appendix B based on the information currently available. for more details). As in most cells, a number of reported observations were not scored or reported Mid-Frequency Cetaceans/Multiple Pulses (Cell 5) here due to lack of some key information and, in A limited number of behavioral observations have some cases, difficulties in accounting for various been made of mid-frequency cetaceans exposed to Marine Mammal Noise Exposure Criteria 453 N/A N/A N/A N/A ; response sever- ; response sever- ; response sever- ; response sever- ; response sever- ; response sever- Summary of (see Table 7) Table (see SPL SPL SPL SPL SPL SPL severity scale analysis severity Exposure RLs 140-180 dB ity scores: 0, 1, 3, 5 & 6 Exposure RLs 140-180 dB ity scores: 0, 1, 3, 5 & 6 Exposure RLs 140-180 dB ity scores: 0, 1 & 6 Exposure RLs 140-180 dB ity score: 6 Exposure RLs 150-160 dB ity score: 3 Exposure RLs 150-170 dB SPL; response severity scores: 6 & 7 Exposure RLs 110-140 dB ity scores: 0, 1, 5 & 6 Exposure RLs 140-180 dB SPL; response severity score: 6 Exposure RLs 140-180 dB SPL; response severity scores: 0 & 6 No No No No Yes Yes Yes Yes Yes Yes Yes Yes Yes Study included in severity scale severity Type of individual and/or of individual Type group behavioral responses group behavioral Land-based observations of individuals/ Land-based observations and respiration patterns groups; movement during and without airguns of individuals/ Land-based observations and respiration patterns groups; movement during and without airguns of individuals/groups; Aerial observations and respiration patterns during movement and without airguns of individuals/groups; Aerial observations patterns and behavioral movement responses during and without airguns of whale behavior observations Visual before and during use of explosives of indi- Aerial and boat-based observations and behavioral viduals/groups; movement patterns during and without airguns of individu- of distribution Aerial surveys als/groups of individuals/ Boat-based observations patterns and behavioral groups; movement during and without airguns of individuals; observations Vessel-based movement/ of distribution; aerial surveys responses patterns and behavioral diving during and without airguns data for this analysis Insufficient data for this analysis Insufficient data for this analysis Insufficient data for this analysis Insufficient near near near near near near near near in situ in situ in situ in situ in situ in situ in situ in situ measurements Type of acoustic Type Calibrated RL measurements made areas of exposure Calibrated RL measurements made areas of exposure Calibrated RL measurements made areas of exposure Calibrated RL measurements made areas of exposure Limited to nominal measurements of explosives used (not measured on or near subjects) Calibrated RL measurements made areas of exposure Calibrated RL measurements made areas of exposure Calibrated RL measurements made areas of exposure Calibrated RL measurements made areas of exposure data for this Insufficient analysis data for this Insufficient analysis data for this Insufficient analysis data for this Insufficient analysis Pa) over the duration of a pulse. over Pa) m Sound source Single airgun (1.64 Single airgun L) & 20-gun 65.5-L array airgun and Single airgun array airgun (0.66 Single airgun L) and 30-gun 47-L array (1.3 Single airgun L) or 18- to 20-gun array Explosions (0.33 Single airgun L) and several arrays array (6 to Airgun 16 guns; 9 to 25 L) Single airgun (0.33 L) array (24 Airgun guns; 36.9 L) Seismic airgun array Seismic airgun array and Single airgun array airgun Seismic airgun . Specific severity scores for each study are given in Table 7, and more details are given in Appendix B. Exposure RLs are given in dB SPL, dB in given are RLs Exposure B. Appendix in given are details more and 7, Table in given are study each for scores severity Specific . 3 species Subject Gray whales Gray whales Bowhead whales (feeding) Bowhead whales (feeding) Humpback whales Humpback whales (migrating) Bowhead whales (migrating) Humpback whales (socializing) Bowhead whales (feeding) Bowhead whales (migrating) Humpback whales Gray whales Bowhead whales (migrating) 1 2 3 4 5 6 7 8 9 Table 7) Table Reference number (for Not included Not included Not included Not included Not Summary of behavioral responses by different species of low-frequency cetaceans exposed to multiple pulses (Cell 2) by type of sound source, acoustic available metrics, Study Malme et al. (1983) Malme et al. (1984) Richardson et al. (1986) Ljungblad et al. (1988) et al. Todd (1996) McCauley et al. (1998) Richardson et al. (1999) McCauley et al. (2000) Miller et al. (2005) et al. Reeves (1984) Malme et al. (1985) Malme et al. (1986, 1988) & Koski Johnson (1987) Table 6. Table all for volume total the is this stated, is volume array airgun where score(s); severity corresponding of summary a and group), and/or individual (by response behavioral of description in 61 = L 1 array; the in airguns operating (dB re: 1 which are RMS sound pressure levels 454 Southall et al.

Table 7. Number (in bold) of low-frequency cetaceans (individuals and/or groups) reported as having behavioral responses to multiple pulse noise; responses were categorized into 10-dB RL bins, ranked by severity of the behavioral response (see Table 4 for severity scaling), and combined with other observations having the same RL/severity score. A summary of the individual studies included in this table is given in the “Low-Frequency Cetaceans/Multiple Pulses (Cell 2)” section of this chapter. Parenthetical subscripts indicate the reference reporting the observations as listed in Table 6.

Received RMS sound pressure level (dB re: 1 µPa) Response 80 to 90 to 100 to 110 to 120 to 130 to 140 to 150 to 160 to 170 to 180 to 190 to score < 90 < 100 < 110 < 120 < 130 < 140 < 150 < 160 < 170 < 180 < 190 < 200

9 8 7 1.0 (6) 6 9.5 47.4 2.2 3.4 5.8 4.5 8.3 (7) (7) (7) (4, 6, 8) (1, 2, 3, 6) (1, 2, 3, 4, 6) (1, 2, 4, 8, 9)

5 1.0 1.0 1.0 (7) (4) (1, 2) 4 3 1.0 1.0 (1, 2) (1, 2) 2 1 5.0 6.0 1.0 2.5 3.0 (7) (7) (7) (1, 2, 3) (5) 0 59.8 17.7 1.1 0.1 0.6 6.8 6.3 (7) (7) (7, 9) (9) (3, 9) (1, 2, 3, 9) (1, 2, 9) multiple pulses. Field observations have involved High-Frequency Cetaceans/Multiple Pulses (Cell 8) sperm whales and a few other odontocete spe- Based on our source type distinction (see Chapter cies exposed to seismic airguns and explosives. 2), virtually all sources of transient sound used in Laboratory investigations have considered behav- quantitative behavioral studies of high-frequency ioral responses to various kinds of multiple pulse cetaceans—for example, acoustic harassment sources. Again, some observations were excluded devices (AHDs) and acoustic deterrent devices due to lack of relevant information. Four studies (ADDs)—would be characterized as nonpulse of individual mid-frequency cetacean responses sounds. While individual elements produced by to multiple pulse exposures contained sufficient some of these sources could be characterized as acoustic and behavioral information for inclusion pulses, and sequences of them as multiple pulses, in this analysis. These include field observations of they are generally emitted in such rapid fashion free-ranging sperm whales and belugas studied by that some mammalian auditory systems likely Madsen & Møhl (2000), Madsen et al. (2002), and perceive them as nonpulses. Further, some AHDs Miller et al. (2005), as well as laboratory observa- and ADDs, and most other sources used in behav- tions of captive false killer whales by Akamatsu ioral studies with high-frequency cetaceans, lack et al. (1993). The information from these studies the characteristics of pulses such as extremely fast is summarized in Table 8 and discussed in detail rise-time, correspondingly broad frequency band- in Appendix B; the companion severity scaling width, and high kurtosis. Due to uncertainty over analysis is shown in Table 9. the extent to which some of these signals may be The combined data for mid-frequency ceta- perceived and the overarching paucity of data, it ceans exposed to multiple pulses do not indicate is not possible to present any data on behavioral a clear tendency for increasing probability and responses of high-frequency cetaceans as a func- severity of response with increasing RL. In cer- tion of received levels of multiple pulses. Available tain conditions, multiple pulses at relatively low data for nonpulse sounds are considered below RLs (~80 to 90 dB re: 1 µPa) temporarily silence (see the “High-Frequency Cetaceans/Nonpulses individual vocal behavior for one species (sperm [Cell 9]” section). We note the need for empirical whales). In other cases with slightly different behavioral research in these animals using sound stimuli, RLs in the 120 to 180 dB re: 1 µPa range sources (such as airgun or pile-driving stimuli) failed to elicit observable reaction from a signifi- unequivocally classified as multiple pulses (see cant percentage of individuals either in the field or Chapter 5). in the laboratory. Marine Mammal Noise Exposure Criteria 455 N/A N/A Table 9) Table scale analysis (see Summary of severity Summary of severity Exposure RLs 170- 180 dB SPL; response score: 0 severity Exposure RLs 120- 140 dB SPL; response score: 0 severity Exposure RLs 100- 150 dB SPL; response scores: 0 & 6 severity Exposure RLs 170- 180 dB SPL; response scores: 0 & 6 severity No No Yes Yes Yes Yes scale Study included in severity in severity Pa) over the duration of a pulse. over Pa) m Type of individual and/or of individual Type responses group behavioral of individu- observations Vessel-based als, including visual detection and passive acoustic monitoring of vocalizations of individu- observations Vessel-based als, including visual detection and passive acoustic monitoring of vocalizations of individuals; observations Vessel-based movement/ of distribution; aerial surveys responses patterns and behavioral diving during and without airguns responses of behavioral observations Visual in laboratory context within experimental conditions of individuals; observations Vessel-based behavior patterns and vocal diving of individuals; observations Vessel-based behavior sightings data and avoidance near near near within in situ in situ in situ in situ measurements Type of acoustic Type Calibrated RL measurements made areas of exposure Calibrated RL measurements made areas of exposure Calibrated RL measurements made areas of exposure Calibrated RL measurements made enclosure experimental data for this Insufficient analysis data for this Insufficient analysis Sound source Small explosives Small explosives per day) (several array (distant) Airgun array (24 guns; Airgun 36.9 L) Numerous sounds, including pulse sequences Natural and artificial pulses (repeated) arrays Seismic airgun (various) species Subject Sperm whales Sperm whales Beluga killer False whales (captive) Sperm whales mid- Several freq. cetacean species 1 2 3 4 Table 9) Table Reference number (for Not included Not included Summary of behavioral responses by different species of mid-frequency cetaceans exposed to multiple pulses (Cell 5) by type of sound source, acoustic available metrics, Study Madsen & Møhl (2000) Madsen et al. (2002) Miller et al. (2005) Akamatsu et al. (1993) André et al. (1997) Stone (2003) Table more 8. and Table 9 Table in given are study each for scores severity specific score(s); severity corresponding of summary a and group), and/or individual (by response behavioral of description (dB re: 1 in dB SPL, which are RMS sound pressure levels Appendix B. Exposure RLs are given in details are given 456 Southall et al.

Table 9. Number (in bold) of mid-frequency cetaceans (individuals and/or groups) reported as having behavioral responses to multiple pulse noise; responses were categorized into 10-dB RL bins, ranked by severity of the behavioral response (see Table 4 for severity scaling), and combined with other observations having the same RL/severity score. A summary of the individual studies included in this table is given in the “Mid-Frequency Cetaceans/Multiple Pulses (Cell 5)” section of this chapter. Parenthetical subscripts indicate the reference reporting the observations as listed in Table 8.

9 8 7 6 0.17 0.17 0.17 1.3 (3) (3) (3) (4) 5 4 3 2 1 0 0.25 0.25 3.0 4.0 6.7 (3) (3) (2) (2) (1, 4)

Pinnipeds in Water/Multiple Pulses (Cell 11) data on responses to pulses were from single pulse Information on behavioral reactions of pinnipeds events (e.g., rocket launches) over populations of in water to multiple pulses involves exposures to pinnipeds exposed to such signals repeatedly (e.g., small explosives used in fisheries interactions, Thorson et al., 1998, 1999, 2000a, 2000b; Berg impact pile driving, and seismic surveys. Several et al., 2001, 2002, 2004). These events do not occur studies lacked matched data on acoustic expo- frequently enough for the exposures to be consid- sures and behavioral responses by individuals. As ered multiple pulses, and many of them contained a result, the quantitative information on reactions nonpulse as well as pulse exposures. They are of pinnipeds in water to multiple pulses is very discussed in some detail in Appendix B (as well limited (see Table 10). The severity scaling analy- as in Appendix C when nonpulses are involved). sis for individual behavioral responses for Cell 11 Appendix B also discusses several other studies is given in Table 11. potentially relevant to Cell 14 but ultimately not Our general finding is that, based on the limited used in this analysis. Consequently, the quantita- data on pinnipeds in water exposed to multiple tive information analyzed for reactions of pinni- pulses, exposures in the ~150 to 180 dB re: 1 µPa peds in air exposed to multiple pulses (see Tables range (RMS values over the pulse duration) gen- 12 & 13) focused on the aerial data by Blackwell erally have limited potential to induce avoidance et al. (2004b). These extremely limited data sug- behavior in pinnipeds. RLs exceeding 190 dB gest very minor, if any, observable behavioral re: 1 µPa are likely to elicit responses, at least in responses by pinnipeds exposed to airborne pulses some ringed seals (Harris et al., 2001; Blackwell with RLs 60 to 80 dB re: 20 µPa. et al., 2004b; Miller et al., 2005). Note that the SEL associated with a single 190 dB re: 1 µPa Behavioral Response Severity Scaling: (RMS) pulse from an airgun is typically ca. 175 Nonpulses dB re: 1 µPa2-s. That exceeds the estimated TTS threshold for the closely related harbor seal (171 Low-Frequency Cetaceans/Nonpulses (Cell 3) dB re: 1 µPa2-s; see Chapter 3). Thus, in the case While there are clearly major areas of uncertainty of ringed seals exposed to sequences of airgun remaining, there has been relatively extensive pulses from an approaching seismic vessel, most behavioral observation of low-frequency ceta- animals may show little avoidance unless the RL ceans exposed to nonpulse sources. As summa- is high enough for mild TTS to be likely. rized in Table 14 (and discussed in greater detail in Appendix C), these field observations involve Pinnipeds in Air/Multiple Pulses (Cell 14) the majority of low-frequency cetacean species How multiple pulses produced in air affect pinni- exposed to a wide range of industrial, active sonar, peds was among the least well-documented of the and tomographic research active sources (Baker conditions we considered. Most of the available et al., 1982; Malme et al., 1983, 1984, 1986; Marine Mammal Noise Exposure Criteria 457 N/A N/A N/A Table 11) Table scale analysis (see Summary of severity Summary of severity Exposure RLs 160- 200 dB SPL; response scores: 0 & 6 severity Exposure RLs 150- 160 dB SPL; response scores: 0 & 1 severity Exposure RLs 170- 200 dB SPL; response score: 0 severity No No No Yes Yes Yes severity scale severity Study included in Pa) over the duration of a pulse. over Pa) m Type of individual and/or of individual Type group behavioral responses group behavioral Vessel-based observations of observations Vessel-based within specified individuals a limited range zones over of Land-based observations and movement individuals; response patterns during pipe- (note that construction driving had been underway activities for a considerable period before observations) of observations Vessel-based pat- movement individuals; responses terns and behavioral during and without airguns of indi- observations Visual multiple vidual responses over exposures of indi- observations Visual multiple vidual responses over exposures Complicated by simultaneous to pulse and nonpulse exposure sources near in situ in situ in situ measurements Type of acoustic Type RLs measured observed near individuals spatial zones in defined RLs measured observed near individuals (detailed measurements, including peak pressure, RMS, SEL, and duration) Calibrated RL measurements made areas of exposure data for this Insufficient analysis data for this Insufficient analysis Calibrated measurements made in the area of exposure Sound source Single airgun and Single airgun 11-gun, 21.6-L array sounds Pipe-driving (construction) array (24 Airgun 36.9 L) airguns; Seal bombs (small explosives) Seal bombs (small explosives) Subject species Ringed (mainly), bearded, and spotted seals Ringed seals Ringed and bearded seals California sea lions California sea lions Ringed seals 1 2 3 Table 11) Table Reference number (for Not included Not included Not included Summary of behavioral responses by different species of pinnipeds in water exposed to multiple pulses (Cell 11) by type of sound source, available acoustic metrics, Study Harris et al. (2001) Blackwell et al. (2004b) Miller et al. (2005) Shaughnessy et al. (1981) Mate & Harvey (1987) Moulton et al. (2003, 2005) Table Table 10. description of behavioral response (by individual and/or group), and a summary of corresponding score(s); severity specific scores severity for each study 11 are and in given Table (dB re: 1 in dB SPL, which are RMS sound pressure levels Appendix B. Exposure RLs are given in more details are given 458 Southall et al.

Table 11. Number (in bold) of pinnipeds in water (individuals and/or groups) reported as having behavioral responses to multiple pulse noise. Responses were categorized into 10-dB RL bins, ranked by severity of the behavioral response (see Table 4 for severity scaling), and combined with other observations having the same RL/severity score; a summary of the individual studies included in this table is given in the “Pinnipeds in Water/Multiple Pulses (Cell 11)” section of this chapter. Parenthetical subscripts indicate the reference reporting the observations as listed in Table 10.

9 8 7 6 1.7 2.1 45.4 (1) (1) (1) 5 4 3 2 1 0.3 (2) 0 0.7 5.3 30.3 0.3 9.9 (2) (1) (1, 3) (3) (1, 3)

Richardson et al., 1990b; McCauley et al., 1996; An additional challenge in interpreting many Biassoni et al., 2000; Croll et al., 2001; Palka & of the field data for this condition is isolating Hammond, 2001; Nowacek et al., 2004). the effect of RL from the effects of mere source The combined information generally indicates presence (as possibly indicated by visual stimuli no (or very limited) responses at RLs 90 to 120 or other aspects of acoustic exposure such as the dB re: 1 µPa and an increasing probability of presence of high-frequency components) and avoidance and other behavioral effects in the 120 other contextual variables. For this reason, several to 160 dB re: 1 µPa range (severity scaling: Table studies were considered but not integrated into the 15). However, these data also indicated consid- analysis. The laboratory observations are of cap- erable variability in RLs associated with behav- tive cetaceans exposed to precisely controlled and ioral responses. Contextual variables (e.g., source known noise exposures in the context of hearing proximity, novelty, operational features) appear to and TTS experiments. However, the relevance of have been at least as important as exposure level behavioral reactions of trained, food-reinforced in predicting response type and magnitude. captive animals exposed to noise to the reactions of free-ranging marine mammals is debatable. Mid-Frequency Cetaceans/Nonpulses (Cell 6) This is discussed in greater detail in Appendix C. A relatively large number of mid-frequency cetaceans The combined field and laboratory data for mid- have been observed in the field and in the laboratory frequency cetaceans exposed to nonpulse sounds responding to nonpulse sounds, including vessels do not lead to a clear conclusion about RLs coinci- and watercraft (LGL & Greeneridge, 1986; Gordon dent with various behavioral responses (see sever- et al., 1992; Palka & Hammond, 2001; Buckstaff, ity scaling, Table 17). In some settings, individuals 2004; Morisaka et al., 2005), pulsed pingers and in the field showed behavioral responses with high AHDs/ADDs (Watkins & Schevill, 1975; Morton & severity scores to exposures from 90 to 120 dB re: Symonds, 2002; Monteiro-Neto et al., 2004), indus- 1 µPa, while others failed to exhibit such responses trial activities (Awbrey & Stewart, 1983; Richardson for exposure RLs from 120 to 150 dB re: 1 µPa. et al., 1990b), mid-frequency active sonar (NRL, Contextual variables other than exposure RL, and 2004a, 2004b; NMFS, 2005), and tones or bands probable species differences, are the likely rea- of noise in laboratory conditions (Nachtigall et al., sons for this variability in response. Context may 2003; Finneran & Schlundt, 2004). Summary infor- also explain why there is great disparity in results mation on these studies is given in Table 16; detailed from field and laboratory conditions—exposures descriptions are given in Appendix C. As in other in captive settings generally exceeded 170 dB re: conditions, a number of potentially relevant field 1 µPa before inducing behavioral responses. studies are not included in the severity scaling analysis due to lack of sufficiently detailed information. Marine Mammal Noise Exposure Criteria 459 N/A re: 20 µPa; re: 20 µPa; Summary of (see Table 13) Table (see severity scale analysis severity Exposure RLs 60-80 dB SPL response severity scores: 0 & 1 No Yes severity scale severity Study included in Pa) over the duration of a pulse. over Pa) m Type of individual and/or of individual Type group behavioral responses group behavioral Land-based observations of indi- Land-based observations and response viduals; movement (note patterns during pipe-driving had been that construction activities for a considerable period underway before observations) of animal Land-based observations and heart rate presence, behavior, (note long history of sonic booms in the area) near in situ measurements Type of acoustic Type RLs measured (detailed observed individuals measurements, including peak pressure, RMS, SEL, and duration) Measured sound overpressure Measured sound overpressure on breeding beaches, levels not RLs at positions of but animals exposed Sound source Pipe-driving sounds Pipe-driving (construction) Repeated sonic booms species Subject Ringed seals Harbor and gray seals 1 Table 13) Table Reference number (for Not included Summary of behavioral responses by different species of pinnipeds in air exposed to multiple pulses (Cell 14) by type of sound source, available acoustic metrics, description metrics, acoustic available source, sound of type by 14) (Cell pulses multiple to exposed air in pinnipeds of species different by responses behavioral of Summary Study Blackwell et al. (2004b) Perry et al. (2002) Table 12. Table 13 and of response more and/or behavioral (by details group), individual scores and for a Table each summary in score(s); severity study of specific are corresponding given severity (dB re: 20 in dB SPL, which are RMS sound pressure levels Appendix B. Exposure RLs are given in are given 460 Southall et al.

Table 13. Number (in bold) of pinnipeds in air (individuals and/or groups) reported as having behavioral responses to multiple pulse noise; responses were categorized into 10-dB RL bins, ranked by severity of the behavioral response (see Table 4 for severity scaling), and combined with other observations having the same RL/severity score. A summary of the individual studies included in this table is given in the “Pinnipeds in Air/Multiple Pulses (Cell 14)” section of this chapter. Parenthetical subscripts indicate the reference reporting the observations as listed in Table 12.

Received RMS sound pressure level (dB re: 20 µPa) Response score 50 to < 60 60 to < 70 70 to < 80 80 to < 90 90 to < 100 100 to < 110 110 to < 120

9 8 7 6 5 4 3 2 1 0.125 (1) 0 0.625 0.25 (1) (1)

High-Frequency Cetaceans/Nonpulses (Cell 9) Habituation to sound exposure was noted in some Numerous controlled studies have been conducted but not all studies. Strong initial reactions of high- on the behavioral reactions of high-frequency frequency cetaceans at relatively low levels may in cetaceans to various nonpulse sound sources some conditions wane with repeated exposure and both in the field (Culik et al., 2001; Olesiuk subject experience. et al., 2002; Johnston, 2002) and in laboratory settings (Kastelein et al., 1997, 2000, 2005, Pinnipeds in Water/Nonpulses (Cell 12) 2006a). However, only one high-frequency spe- The effects of nonpulse exposures on pinni- cies (harbor porpoise) has been extensively peds in water are poorly understood. Studies studied and that species provided all the avail- for which enough information was available for able data on behavioral response magnitude vs analysis include field exposures of harbor seals received exposure conditions. The original stud- to AHDs (Jacobs & Terhune, 2002) and exposure ies were attempts to reduce harbor porpoise by- of translocated freely diving northern elephant catch by attaching warning pingers to fishing gear. seals to a research tomography source (Costa More recent studies consider whether AHDs and et al., 2003), as well as responses of captive harbor ADDs also exclude harbor porpoises from criti- seals to underwater data communication sources cal habitat areas, and whether these devices affect (Kastelein et al., 2006b). These limited available harbor porpoise behavior in controlled laboratory data (see Table 20 & Appendix C) suggested that conditions. exposures between ~90 and 140 dB re: 1 µPa The combined wild and captive animal data generally do not appear to induce strong behav- (summarized in Table 18 and discussed in detail in ioral responses in pinnipeds exposed to nonpulse Appendix C) clearly support the observation that sounds in water; no data exist regarding exposures harbor porpoises are quite sensitive to a wide range at higher levels. The severity scale results for Cell of human sounds at very low exposure RLs (~90 to 12 are given in Table 21. 120 dB re: 1 µPa), at least for initial exposures. This It is important to note that among these stud- observation is also evident in the severity scaling ies of pinnipeds responding to nonpulse exposures analysis for Cell 9 (Table 19). All recorded expo- in water, there are some apparent differences in sures exceeding 140 dB re: 1 µPa induced profound responses between field and laboratory condi- and sustained avoidance behavior in wild harbor tions. Specifically, in this case, captive subjects porpoises. Whether this apparently high degree of responded more strongly at lower levels than did behavioral sensitivity to anthropogenic acoustic animals in the field. Again, contextual issues are sources extends to other high-frequency cetacean the likely cause of this difference. Captive subspecies (or nonpulse sources other than AHDs and jects in the Kastelein et al. (2006b) study were not ADDs) is unknown. Given the lack of informa- reinforced with food for remaining in noise fields, tion to the contrary, however, such a relationship in contrast to the laboratory studies for mid-fre- should be assumed as a precautionary measure. quency cetaceans described above. Subjects in the Marine Mammal Noise Exposure Criteria 461 details more Table 15) Table scale analysis (see Summary of severity Summary of severity Exposure RLs 100- 140 dB SPL; severity scores: 0 & 6 Exposure RLs 90- 150 dB SPL; severity scores: 0 & 6 Exposure RLs 100- 120 dB SPL; severity scores: 0 & 6 Exposure RLs 100- 140 dB SPL; severity scores: 0 & 6 Exposure RLs 110- 130 dB SPL; severity score: 6 Exposure RLs 120- 130 dB SPL; severity score: 6 Exposure RLs 110- 160 dB SPL; severity scores: 2 & 4 Exposure RLs 140- 150 dB SPL; severity score: 0 scale Study included in severity in severity Yes Yes Yes Yes Yes Yes Yes Yes Type of individual and/or of individual Type Pa). group behavioral responses group behavioral m Vessel-based observations of individual of individual observations Vessel-based patterns around and behavioral movement vessels of individual Shore-based observations patterns around and behavioral movement simulated drilling operations/platforms of individual observations Vessel-based patterns before and and behavioral movement during playbacks movement of individual observations Visual patterns before, during, and and behavioral to drilling sounds after exposure of individual observations Visual patterns during and behavioral movement approaches vessel movement of individual observations Visual patterns before, during, and and behavioral after playbacks movement of individual observations Visual patterns before, during, and and behavioral after playbacks a general responses not reported but Individual with/without of feeding behavior observation sonar near near near indi- in situ in situ in situ near areas of near areas of measurements Type of acoustic Type in situ in situ Individual RLs not reported but RLs not reported but Individual identical to previous vessels measurements RLs measured observed individuals RLs measured observed individuals Detailed and calibrated source transmission loss measurements good RL estimates allowed RLs measured viduals observed Calibrated RL measurements made exposure Calibrated RL measurements made exposure Calibrated RL measurements and modeling for area of exposure Sound source Vessel noise Vessel and presence Playbacks of drilling and machinery noise Playbacks of noise drilling Drilling noise playbacks noise Vessel and presence Low- frequency M-sequence playback Low- frequency sonar playback Low- frequency sonar playback Subject species Humpback whales Gray whales (migrating) Gray whales (feeding) Bowhead whales (migrating) Humpback whales Humpback whales Humpback whales Blue and fin whales (feeding) 1 2 3 4 5 6 7 8 number Reference (for Table 15) Table (for Summary of behavioral responses by different species of low-frequency cetaceans exposed to nonpulses (Cell 3) by type of sound source, available acoustic metrics, descrip- metrics, acoustic available source, sound of type by 3) (Cell nonpulses to exposed cetaceans low-frequency of species different by responses behavioral of Summary Study et al. Baker (1982) Malme et al. (1983, 1984) Malme et al. (1986) Richardson et al. (1990b) McCauley et al. (1996) & Frankel Clark (1998) Biassoni et al. (2000); Miller et al. (2000) Croll et al. (2001) Table 14. Table and 15 Table in given are study each for scores severity specific score(s); severity corresponding of summary a and group), and/or individual (by response behavioral of tion (dB re: 1 in dB SPL, which are RMS sound pressure levels Appendix C. Exposure RLs are given in are given 462 Southall et al. N/A N/A N/A N/A N/A N/A N/A Table 15) Table scale analysis (see Summary of severity Summary of severity Exposure RLs 110- 120 dB SPL; severity score: 3 Exposure RLs 120- 150 dB SPL; severity scores: 0 & 7 scale Study included in severity in severity Yes Yes No No No No No No No Type of individual and/or of individual Type group behavioral responses group behavioral Visual observations of individual and group of individual observations Visual patterns during and behavioral movements approaches vessel and physical Detailed measurements of vocal reactions of animals before, during, and after playbacks of individuals and acoustic observations Visual responses for data on individual Insufficient this analysis responses for data on individual Insufficient this analysis around rigs of individuals observations Visual responses for data on individual Insufficient this analysis during of individuals observations Visual approaches responses for data on individual Insufficient this analysis measurements Type of acoustic Type RL estimates based on source and characteristics environmental calibrated tags that Subjects wore measured RL and behavior/ movement data for this analysis Insufficient Some RL measurements and modeling in area Some RL measurements and modeling in area data for this analysis Insufficient Some RL measurements and modeling in area data for this analysis Insufficient Some RL measurements and modeling in area Sound source Vessel noise Vessel and presence Playbacks of several nonpulses Playbacks of nonpulses Industrial noise ATOC source Drillships ATOC source noise Vessel and presence ATOC source Subject species whales Minke Right whales Gray whales Various cetaceans Humpback whales Bowhead whales Humpback whales Fin whales Humpback whales 9 10 number Reference (for Table 15) Table (for Not included Not included Not included Not included Not included Not included Not included Study & Palka Hammond (2001) Nowacek et al. (2004) Dahlheim (1987) Borggaard et al. (1999) & Frankel Clark (2000) Schick & Urban (2000) & Frankel Clark (2002) Jahoda et al. (2003) Mobley (2005) Table 14 (continued) Table Marine Mammal Noise Exposure Criteria 463

Table 15. Number (in bold) of low-frequency cetaceans (individuals and/or groups) reported as having behavioral responses to nonpulses; responses were categorized into 10-dB RL bins, ranked by severity of the behavioral response (see Table 4 for severity scaling), and combined with other observations having the same RL/severity score. A summary of the individual studies included in this table is given in the “Low-Frequency Cetaceans/Nonpulses (Cell 3)” section of this chapter. Parenthetical subscripts indicate the reference reporting the observations as listed in Table 14.

9 8 7 2.5 1.5 (10) (10) 6 4.9 7.4 16.2 13.6 4.2 0.8 (2) (1, 2, 4) (1, 2, 3, 5) (2, 5) (1, 2) (2) 5 4 3.0 1.0 1.0 (5, 7) (7) (7) 3 1,117 0.27 (9) (6) 2 0.5 4.0 5.0 2.0 1.0 (7) (7) (7) (7) (7) 1 0 1.1 82.6 33.9 7.08 7.2 1.45 (2) (2, 3, 4) (1, 2, 3, 4) (2, 4, 6, 10) (4, 10) (2, 8, 10) field may have been more tolerant of exposures those observations (Thorson et al., 1999, 2000b; because of motivation to return to a safe location Berg et al., 2002) for which there was clearly just (Costa et al., 2003) or motivation to approach nonpulse exposure were considered in the severity enclosures holding prey items (Jacobs & Terhune, scaling analyses for this condition. 2002). The limitations of these and other potentially applicable studies resulted in a very limited data Pinnipeds in Air/Nonpulses (Cell 15) set for use in this analysis (see summary in Table There has been considerable effort to study the 22 and severity scaling analysis in Table 23). As a effects of aerial nonpulse sounds on pinniped general statement from the available information, behavior, primarily involving rocket launches, pinnipeds exposed to intense (~110 to 120 dB re: aircraft overflights, powerboat approaches, and 20 µPa) nonpulse sounds tended to leave haulout construction noise. Unfortunately, as discussed in areas and seek refuge temporarily (minutes to a few Appendix C, many of the studies are difficult to hours) in the water, whereas pinnipeds exposed to interpret in terms of exposure RL and individual distant launches at RLs ~60 to 70 dB re: 20 µPa or group behavioral responses. In many cases, tended to ignore the noise. It is difficult to assess it was difficult or impossible to discern whether the relevance of either of these observations to the reported behavioral response was induced by naïve individuals, however, given the repeated the noise from a specific operation or some cor- exposure of study colonies to such noise events and related variable such as its visual presence. For the potential that observed individuals were habitu- these reasons, most of the observational studies ated. Due to the limitations of available data, it is of behavioral disturbance were not appropriate for not currently possible to make any further general scoring behavioral responses relative to exposure characterizations regarding this condition. RL. However, a number of the technical reports and analyses of rocket launches are relevant for this cell and contain sufficiently detailed information regarding estimated RLs. These observations are, however, complicated by the fact that all studies were conducted in the same general area with subjects likely habituated to the presence of launch noise. Further, in many cases, exposures contained both a nonpulse component and a pulse component (described below). Only 464 Southall et al. Table 17) Table scale analysis (see Summary of severity Summary of severity Exposure RLs 80-90 dB SPL; severity score: 3 Exposure RLs 110-150 dB SPL; scores: 0, 1, severity 2 & 6 Exposure RLs 90- 120 dB SPL; severity scores: 0, 1, 2, 3 & 8 Exposure RLs 100-130 dB SPL; scores: 0, 1, severity 3 & 4 Exposure RLs 110- 120 dB SPL; severity score: 3 Exposure RLs 110- 120 dB SPL; severity score: 3 Exposure RLs 140- 150 dB SPL; severity score: 8 Exposure RLs 110- 120 dB SPL; severity score: 2 Exposure RLs 160- 170 dB SPL; severity score: 6 Yes Yes Yes Yes Yes Yes Yes Yes Yes scale Study included in severity in severity Pa). m Type of individual and/or of individual Type group behavioral responses group behavioral Passive acoustic monitoring of vocal output of acoustic monitoring of vocal Passive during exposure individuals and group of individual observations Visual patterns during and behavioral movements and control trials exposure of groups Ice-based and aerial observations patterns and behavioral animals; movement before, during, and after ice-breaking of individual Ice-based and aerial observations during and behavior and group movements and control trials exposure acoustic and passive observations Vessel-based patterns movement monitoring of individuals; responses and behavioral and group of individual observations Visual patterns during and behavioral movements approaches vessel and group sightings Census data for individual zones used to estimate “exclusion” vocal acoustic monitoring of individual Passive approaches output during vessel and group of individual observations Visual patterns before, and behavioral movements during, and after incidental exposure near indi- in situ near areas of near areas of near areas of measurements Type of acoustic Type in situ in situ in situ RLs measured viduals observed RL estimates based on source and characteristics environmental Calibrated RL measurements made exposure RL estimates based on source and characteristics plus environmental data sonobuoy Calibrated RL measurements made exposure RL estimates based on source and characteristics environmental RL estimates based on source and characteristics environmental Calibrated RL measurements made exposure Some calibrated RL measurements and RL estimates from modeling source and characteristics environmental Sound source Pingers Playbacks of drilling sounds Ship and ice-breaking noise Playbacks of drilling sounds noise Vessel and presence noise Vessel and presence Various AHDs noise Vessel and presence (approaches) Mid- frequency mili- active tary sonar Exposure RLs are given in dB SPL, which are RMS sound pressure levels (dB re: 1 in dB SPL, which are RMS sound pressure levels Exposure RLs are given Subject species Sperm whales Belugas and Belugas narwhals Belugas Sperm whales White-sided and white- beaked dolphins Killer whales Bottlenose dolphins Killer whales 1 2 3 4 5 6 7 8 9 Table 17) Table Reference number (for Summary of behavioral responses by different species of mid-frequency cetaceans exposed to nonpulses (Cell 6) by type of sound source, available acoustic metrics, descrip- metrics, acoustic available source, sound of type by 6) (Cell nonpulses to exposed cetaceans mid-frequency of species different by responses behavioral of Summary Study Watkins & Schevill (1975) Awbrey & Stewart (1983) LGL & Greeneridge (1986) Richardson et al. (1990b) Gordon et al. (1992) & Palka Hammond (2001) Morton & Symonds (2002) Buckstaff (2004) NRL (2004a, 2004b); NMFS (2005) Table 16. Table tion of behavioral response (by individual and/or group), and a summary of corresponding severity score(s); specific severity scores for each study are given in 17 Table and more Appendix C. in details are given Marine Mammal Noise Exposure Criteria 465 N/A N/A N/A N/A N/A N/A N/A N/A Table 17) Table scale analysis (see Summary of severity Summary of severity Exposure RLs 110- 120 dB SPL; severity score: 6 Exposure RLs 120- 130 dB SPL; severity score: 5 Exposure RLs 170- 180 dB SPL; severity score: 6 Exposure RLs 180- 200 dB SPL; severity scores: 0 & 8 No No No No No No No No Yes Yes Yes Yes scale Study included in severity in severity Type of individual and/or of individual Type group behavioral responses group behavioral Visual observations of individual and group of individual observations Visual patterns during and behavioral movements and control trials exposure acoustic monitoring of individual Passive approaches output during vessel vocal responses of behavioral observations Visual lab context within experimental responses of behavioral observations Visual in laboratory context within experimental conditions acoustic measurements of whistle rates Passive foraging of individual observations Visual behavior data for this analysis Insufficient and diving of movement observations Visual behavior and diving of movement observations Visual behavior and diving of movement observations Visual behavior behavior of movement observations Visual exposures/ data on individual Insufficient responses for this analysis in near areas of within test enclosure measurements Type of acoustic Type in situ in situ within test enclosure RL estimates based on source and characteristics environmental Calibrated RL measurements made exposure Calibrated RL measurements situ Calibrated RL measurements made data for this analysis Insufficient data for this analysis Insufficient data for this analysis Insufficient Acoustic measurements of source no estimates of RL but levels data for this analysis Insufficient data for this analysis Insufficient data for this analysis Insufficient data for this analysis Insufficient ® Sound source Dukane Netmark ADDs noise Vessel and presence Nonpulse noise (bands) Nonpulse noise (tones) mili- Active tary sonar noise Vessel and presence ADDs noise Vessel and presence ADDs noise Vessel and presence noise Vessel and presence General increase in vessels Subject species (river Tucuxi dolphins) Indo-Pacific dolphins Bottlenose dolphins (captive) Bottlenose dolphins (captive) Long-finned pilot whales Bottlenose dolphins Franciscana dolphins Killer whales Bottlenose dolphins Bottlenose dolphins Bottlenose dolphins Killer whales 10 11 12 13 Table 17) Table Reference number (for Not included Not included Not included Not included Not included Not included Not included Not included Study Monteiro- Neto et al. (2004) Morisaka et al. (2005) Nachtigall et al. (2003) Finneran & Schlundt (2004) Rendell & Gordon (1999) & Chilvers Corkeron (2001) Bordino et al. (2002) et al. Williams (2002) Cox et al. (2003) Hastie et al. (2003) Lusseau (2003) et al. Foote (2004) Table 16 (continued) Table 466 Southall et al.

Table 17. Number (in bold) of mid-frequency cetaceans (individuals and/or groups) reported as having behavioral responses to nonpulses; responses were categorized into 10-dB RL bins, ranked by severity of the behavioral response (see Table 4 for severity scaling), and combined with other observations having the same RL/severity score. A summary of the individual studies included in this table is given in the “Mid-Frequency Cetaceans/Nonpulses (Cell 6)” section of this chapter. Parenthetical subscripts indicate the reference reporting the observations as listed in Table 16.

9 8 1.0 7.0 5.0 1.0 5.0 1.5 (3) (3) (2) (7) (13) (13) 7 6 3.0 1.0 1.0 6.0 (2, 10) (2) (9) (12) 5 1.0 (11) 4 1.0 2.0 (4) (4) 3 5.0 4.0 134 1.0 (1) (3, 5) (4, 6) (4)

2 15.0 (2, 3, 8) 1 1.0 1.0 1.0 (4) (2, 3) (2, 4) 0 8.0 2.0 1.0 1.0 3.0 1.5 (3, 4) (2, 4) (2, 4) (2) (13) (13)

Courtesy: A. Friedlander Marine Mammal Noise Exposure Criteria 467 scale analysis (see Table 19) Table (see Summary of severity Summary of severity Exposure RLs 80-120 dB SPL; response severity scores: 0 & 6 Exposure RLs 140-160 dB SPL; response severity score: 6 Exposure RLs 120-130 dB SPL; response severity scores: 0 & 6 Exposure RLs 80-120 dB SPL; response severity scores: 0, 4 & 6 Exposure RLs 90-120 dB SPL; response severity scores: 0 & 6 Exposure RLs 90-120 dB SPL; response severity scores: 0 & 6 Exposure RLs 100-120 dB SPL; response severity scores: 0 & 6 Yes Yes Yes Yes Yes Yes Yes scale Study included in severity in severity Type of individual and/or of individual Type group behavioral responses group behavioral Visual observations of individual and group of individual observations Visual patterns before and behavioral movements deployment and following and group of individual observations Visual patterns before and behavioral movements deployment and following and group of individual observations Visual patterns before and behavioral movements deployment and following respiration, of movement, observations Visual in laboratory conditions and behavior respiration, of movement, observations Visual in laboratory conditions and behavior respiration, of movement, observations Visual in laboratory conditions and behavior respiration, of movement, observations Visual in laboratory conditions and behavior Pa). m within test enclosure within test enclosure within test enclosure within test enclosure measurements Type of acoustic Type in situ in situ in situ in situ RL estimates based on source and characteristics environmental RL estimates based on source and characteristics environmental RL estimates based on source and characteristics environmental Calibrated RL measurements made Calibrated RL measurements made Calibrated RL measurements made Calibrated RL measurements made ® ® Sound source PICE pinger Airmar AHDs Airmar AHDs non- Various pulse sounds (laboratory) non- Various pulse sounds (laboratory) Various nonpulse sounds (laboratory) non- Various pulse sounds (laboratory) Subject species Harbor porpoises (wild) Harbor porpoises (wild) Harbor porpoises (wild) Harbor porpoises (captive) Harbor porpoises (captive) Harbor porpoises (captive) Harbor porpoises (captive) 1 2 3 4 5 6 7 Table 19) Table Reference number (for Exposure RLs are given in dB SPL, which are RMS sound pressure levels (dB re: 1 in dB SPL, which are RMS sound pressure levels Exposure RLs are given Summary of behavioral responses to Summary cetaceans of nonpulses of exposed high-frequency (Cell behavioral acoustic 9) metrics, by description type of of behavioral sound source, available Study Culik et al. (2001) Olesiuk et al. (2002) Johnston (2002) Kastelein et al. (1997) Kastelein et al. (2000) Kastelein et al. (2005) Kastelein et al. (2006a) Table 18. Table response (by individual and/or group), and a summary of corresponding score(s); severity specific scores severity for each study 19 are and in given Table more details are in given Appendix C. 468 Southall et al. N/A N/A N/A N/A N/A N/A N/A scale analysis (see Table 19) Table (see Summary of severity Summary of severity No No No No No No No scale Study included in severity in severity Type of individual and/or of individual Type group behavioral responses group behavioral Measurements of by-catch rates in commercial fisheries analysis Review zones of “exclusion” observations Visual zones of “exclusion” observations Visual movement of individuals; Aerial observations and respiration patterns during without airguns Measurements of by-catch rates in commercial fisheries monitoring of general distribution Visual patterns near areas of measurements Type of acoustic Type in situ Insufficient data for this analysis Insufficient analysis Review data for this analysis Insufficient data for this analysis Insufficient Calibrated RL measurements made exposure data for this analysis Insufficient measure- Calibrated source-level data insufficient ments made but on RL ® Sound source Dukane pingers General nonpulse sounds Various AHDs Various ADDs non- Various pulse sounds (laboratory) Various ADDs Simulated wind turbine noise Subject species Harbor porpoises (wild) Harbor porpoises (wild) Harbor porpoises (wild) Harbor porpoises (wild) Harbor porpoises (captive) Harbor porpoises (wild) Harbor porpoises (wild) Table 19) Table Reference number (for Not included Not included Not included Not included Not included (same subjects as 2000 study) Not included Not included Study Kraus et al. (1997) et al. Taylor (1997) Johnston & Woodley (1998) Cox et al. (2001) Kastelein et al. (2001) & Barlow Cameron (2003) Koschinski et al. (2003) Table 18 (continued) Table Marine Mammal Noise Exposure Criteria 469

Table 19. Number (in bold) of high-frequency cetaceans (individuals and/or groups) reported as having behavioral responses to nonpulses; responses were categorized into 10-dB RL bins, ranked by severity of the behavioral response (see Table 4 for severity scaling), and combined with other observations having the same RL/severity score. A summary of the individual studies included in this table is given in the “High-Frequency Cetaceans/Nonpulses (Cell 9)” section of this chapter. Parenthetical subscripts indicate the reference reporting the observations as listed in Table 18.

9 8 7 6 0.3 0.3 0.9 3.3 1.0 52.1 9.3 4.6 (4) (4) (1, 2, 4, 5, 6, 7) (1, 2, 4, 5, 6, 7) (3, 7) (2) (2) (2) 5 4 0.1 0.1 (4) (4) 3 2 1 0 12.8 23.1 0.4 0.1 0.3 (1, 5) (1, 2, 5, 6) (4, 7) (7) (3)

Courtesy: A. Friedlander 470 Southall et al. N/A N/A N/A N/A N/A N/A N/A scale analysis (see Table 21) Table (see Summary of severity Summary of severity Exposure RLs 120-130 dB SPL; response severity score: 0 Exposure RLs 110-140 dB SPL; response severity scores: 0, 3 & 4 Exposure RLs 80-110 dB SPL; response severity scores: 0 & 6 scale Study included in severity in severity Yes Yes Yes No No No No No No No Type of individual and/or of individual Type group behavioral responses group behavioral Pa). m Visual observations of individuals and of individuals observations Visual and behavioral groups of seals; movement AHDs patterns during and without tags placed on animals resulted in Archival measures of individual detailed quantitative responses, and exposure behavior, diving RLs in well-characterized contexts subject positions and the mean Individual during con- behaviors number of surfacing intervals trol and exposure data for this analysis Insufficient data for this analysis Insufficient responses data on individual Insufficient for this analysis measurements but Some behavioral responses data on individual insufficient as a function of RL data for this analysis Insufficient responses data on individual Insufficient for this analysis responses data on individual Insufficient for this analysis in areas in areas on individuals on individuals in situ in situ in situ within experimental within experimental measurements Type of acoustic Type in situ RLs measured observed where individuals RLs measured using calibrated tags archival during exposure Calibrated RL measurements made enclosure data for this analysis Insufficient data for this analysis Insufficient Calibrated acoustic measurements around arrays of the devices taken data for this analysis Insufficient data for this analysis Insufficient RLs measured observed where individuals data for this analysis Insufficient AHDs ® dB plus dB plus ® ® Sound source Airmar AHD II (see ATOC Appendix B) non- Various pulse sounds used in under- data water communications Underwater drilling sounds Underwater drilling sounds Applied Cascade Sciences Airmar AHD II AHD Simulated wind turbine noise Construction noise Subject species Harbor seals Elephant seals Harbor seals Ringed seals Ringed and bearded seals California sea lions California sea lions Harbor seals Harbor seals Ringed seals Exposure RLs are given in dB SPL, which are RMS sound pressure levels (dB re: 1 in dB SPL, which are RMS sound pressure levels Exposure RLs are given 1 2 3 number Reference (for Table 21) Table (for Not included Not included Not included Not included Not included Not included Not included Summary of behavioral responses by different species of pinnipeds in water exposed to nonpulses (Cell 12) by type of sound source, available acoustic metrics, description metrics, acoustic available source, sound of type by 12) (Cell nonpulses to exposed water in pinnipeds of species different by responses behavioral of Summary Study Jacobs & Terhune (2002) Costa et al. (2003) Kastelein et al. (2006b) Frost & (1988) Lowry Richardson et al. (1990b, 1991) & Norberg Bain (1994) Norberg (2000) (2000) Yurk Koschinski et al. (2003) Moulton et al. (2003) Table 20. Table 21 and of response more and/or behavioral (by details group), individual scores and for a Table each summary in score(s); severity study of specific are corresponding given severity Appendix C. in are given Marine Mammal Noise Exposure Criteria 471

Table 21. Number (in bold) of pinnipeds in water (individuals and/or groups) reported as having behavioral responses to nonpulses; responses were categorized into 10-dB RL bins, ranked by severity of the behavioral response (see Table 4 for severity scaling), and combined with other observations having the same RL/severity score. A summary of the individual studies included in this table is given in the “Pinnipeds in Water/Nonpulses (Cell 12)” section of this chapter. Parenthetical subscripts indicate the reference reporting the observations as listed in Table 20.

9 8 7 6 1.0 (3) 5 4 1.0 5.0 (2) (2) 3 1.0 2.0 (2) (2) 2 1 0 1.0 1.0 1.0 5.0 (3) (3) (2) (1, 2)

Courtesy: A. Friedlander 472 Southall et al. N/A N/A N/A N/A N/A N/A N/A N/A N/A ; response Table 23) Table Summary of analysis (see severity scale severity SPL Exposure RLs 110-120 dB SPL; response severity score: 6 Exposure RLs 60- 70 and 110-120 dB scores: 0 severity & 6 Exposure RLs 110-120 dB SPL; response severity score: 6 No No No No No No No No No Yes Yes Yes scale Study included in severity in severity Type of individual and/or of individual Type group behavioral responses group behavioral Visual observations of movement and of movement observations Visual in breeding of individuals behavior before, during, and after rookeries launches rocket and of movement observations Visual in breeding of individuals behavior before, during, and after rookeries launches rocket and of movement observations Visual in breeding of individuals behavior before, during, and after rookeries launches rocket data for this analysis Insufficient data for this analysis Insufficient data for this analysis Insufficient data for this analysis Insufficient data for this analysis Insufficient data for this analysis Insufficient data for this analysis Insufficient data for this analysis Insufficient of animal presence observations Visual before launches and and distribution launches during and following behavior Pa). m in and in and in and in and in and in and in situ in situ in situ in situ in situ in situ measurements Type of acoustic Type RLs measured around breeding rookeries RLs measured around breeding rookeries RLs measured around breeding rookeries data for this Insufficient analysis data for this Insufficient analysis data for this Insufficient analysis RLs measured around breeding rookeries data for this Insufficient analysis RLs measured around breeding rookeries RLs measured around breeding rookeries data for this Insufficient analysis RLs measured near observed pinnipeds, including peak, RMS, SEL, and duration Sound source Athena 2 missile IKONOS-1 launch IV B-28 Titan missile launch IV B-34 Titan missile launch noise Aerial vessel and presence Underground explosions and quarrying operations noise Aerial vessel and presence A-18 IV Titan missile launch Aircraft noise and presence II G-13 Titan missile launch Delta II EO-1 missile launch Industrial equipment noise & presence Small- and mid- sized missile launches species Subject Harbor seals, northern elephant seals, California sea lions, and northern fur seals Harbor seals, northern elephant seals, and California sea lions Harbor seals Harbor seals Northern fur seals Harbor seals Harbor seals Ringed seals Harbor seals California sea lions and northern elephant seals Ringed seals Harbor seals, California sea lions, and northern elephant seals 1 2 3 Exposure RLs are given in dB SPL, which are RMS sound pressure levels (dB re: 20 in dB SPL, which are RMS sound pressure levels Exposure RLs are given number Reference (for Table 23) Table (for Not included Not included Not included Not included Not included Not included Not included Not included Not included Summary of behavioral responses by different species of pinnipeds in air exposed to nonpulses (Cell 15) by type of sound source, available acoustic metrics, description of description metrics, acoustic available source, sound of type by 15) (Cell nonpulses to exposed air in pinnipeds of species different by responses behavioral of Summary Study Thorson et al. (1999) Thorson et al. (2000b) et al. Berg (2002) Allen et al. (1984) Gentry et al. (1990) Suryan & (1998) Harvey Thorson et al. (1998) Born et al. (1999) Thorson et al. (2000a) et al. Berg (2001) Moulton et al. (2002) Holst et al. (2005a, 2005b) Table 22. Table are details more and 23 Table in given are study each for scores severity specific score(s); severity corresponding of summary a and group), and/or individual (by response behavioral Appendix C. in given Marine Mammal Noise Exposure Criteria 473

Table 23. Number (in bold) of pinnipeds in air (individuals and/or groups) reported as having behavioral responses to nonpulses; responses were categorized into 10-dB RL bins, ranked by severity of the behavioral response (see Table 4 for severity scaling), and combined with other observations having the same RL/severity score. A summary of the individual studies included in this table is given in the “Pinnipeds in Air/Nonpulses (Cell 15)” section in this chapter. Parenthetical subscripts indicate the reference reporting the observations as listed in Table 22.

Received RMS sound pressure level (dB re: 20 µPa) Response 50 to 60 to 70 to 80 to 90 to 100 to 110 to score < 60 < 70 < 80 < 90 < 100 < 110 < 120

9 8 7 6 1.0 (1, 2, 3) 5 4 3 2 1 0 1.0 (2)

Courtesy: Peter M. Scheifele Aquatic Mammals 2007, 33(4), 474-497, DOI 10.1578/AM.33.4.2007.474

5. Research Recommendations

The marine mammal noise exposure criteria tags should be deployed on free-ranging marine proposed here represent a synthesis and precau- mammals and/or platforms near the animals in tionary application of current scientific informa- controlled exposure conditions. tion. Clearly, the reliance on extrapolation proce- If future noise exposure criteria are to consider dures, extreme data gaps and limitations in many the important matters of auditory masking, cumu- areas, and precautionary assumptions throughout lative exposure effects on individuals, and ecosys- point to the need for targeted research to fill spe- tem effects (discussed below), additional data are cific gaps in support of subsequent criteria. We needed concerning ambient ocean noise on vari- consider the current noise exposure criteria to be ous spatial and temporal scales. These data should merely an initial step in an iterative process to be used to determine how ambient noise “budgets” understand and better predict the effects of noise vary as a function of natural and human activi- on marine mammal hearing and behavior. ties. These data will need to be integrated with Research recommendations are given below expanded information on marine mammal abun- in several broad categories relevant to improving dance and distribution. The NRC (2003) recom- marine mammal noise exposure criteria. No pri- mended that a systematic effort be made to obtain oritization is implied in the ordering of these areas passive acoustic data, including average (steady- or research topics within them, however, and this state) ambient noise from 1 Hz to 200 kHz, and is by no means an exhaustive list. We present, in including transient human sources not identified in abbreviated form, what we regard as critical, tar- classical ambient noise measurements. We concur geted research needs to improve future iterations and call for wide-ranging acoustic measurements of these exposure criteria. Some of the most impor- designed to test explicit hypotheses about spatial tant research recommendations are summarized in and temporal variability in marine ambient noise. Table 24; each is discussed in more detail in the relevant section below. Many of these research Marine Mammal Auditory Processes recommendations are similar to recommendations made previously (NRC, 1994, 2000, 2003, 2005; “Absolute” Hearing Data Richardson et al., 1995). Although there has been Future iterations of these criteria will be sig- progress in the last decade, much important work nificantly improved by increased knowledge of remains to be done. hearing sensitivity derived from behavioral and electrophysiological measurements and anatomi- Measurements of Anthropogenic cal models. The most pressing needs are for data Sound Sources and Ambient Noise on deep-diving cetaceans such as beaked whales and on low-frequency specialists (mysticetes). Comprehensive and systematic measurements Better information on inter-species differences are needed of all relevant anthropogenic sound is also needed to validate the functional hearing sources that have a reasonable likelihood of groups used here or alternatively to identify other adversely affecting marine mammal hearing or relevant subdivisions (e.g., phocid vs otariid pin- behavior. Empirical measures of sound fields nipeds or potential partitioning of mid-frequency enable more accurate estimation of RLs using cetaceans). The number of individuals tested propagation models and validate the selection of should be increased in all species, with the pos- different propagation models as appropriate. Such sible exception of the bottlenose dolphin, in order studies must report the full range of relevant stan- to better understand individual differences within dard acoustic measurements and should include species. Hearing sensitivity across the full func- detailed information about equipment calibration tional hearing range should be measured, where and/or propagation modeling methods used (e.g., possible, rather than just those frequencies con- Goold & Fish, 1998; Wales & Heitmeyer, 2002; tained within the communication signals of spe- Blackwell et al., 2004a). Measurements are also cies being investigated. needed to describe conditions where sounds clas- Improvements are needed in both electro- sified as pulses at the source transition to non- physiological and behavioral testing methods to pulse exposures. To measure in situ exposures increase the number of individuals of each spe- from specific sound sources, archival acoustic cies that can be tested, and to distinguish absolute Marine Mammal Noise Exposure Criteria 475

Table 24. Research recommendations in various subject areas needed to enhance future marine mammal noise exposure criteria (as discussed in Chapter 5)

Research topic General description Critical information needs

Acoustic Detailed measurements needed of Comprehensive, calibrated measurements of the properties measurements of source levels, frequency content, and of human-generated sound sources, including frequency- relevant sound radiated sound fields around intense dependent propagation and received characteristics in sources and/or chronic noise sources. different environments. Comprehensive, calibrated measurements of ambi- Systematic measurements of underwa- ent noise, including spectral, temporal, and directional Ambient noise ter ambient noise are needed to quan- aspects, in different oceanic environments; ambient noise measurements tify how human activities are affecting “budgets” indicating relative contribution of natural and the acoustic environment. anthropogenic sources and trends over time. Audiometric data are needed to deter- Carefully controlled behavioral and electrophysiological mine functional bandwidth, species and measurements of hearing sensitivity vs frequency for more “Absolute” hearing individual differences, dynamic hearing individuals and species, particularly for high-priority spe- measurements ranges, and detection thresholds for cies, such as beaked whales and mysticetes. Also, detec- realistic biological stimuli. tion thresholds for complex biological signals. Measurements to determine the sophis- Measurements of stream segregation, spatial perception, ticated perceptual and processing capa- multidimensional source localization, frequency discrimi- Auditory scene bilities of marine mammals that enable nation, temporal resolution, and feedback mechanisms analysis them to detect and localize sources in between sound production and hearing systems. complex, 3-D environments. Carefully constructed observational and exposure experi- Measurements of behavioral reactions Marine mammal ments that consider not only RL but also source range, to various sound types are needed, behavioral responses motion, signal-to-noise ratio, and detailed information on including all relevant acoustic, contex- to sound exposure receivers, including baseline behavior, prior experience tual, and response variables. with the sound, and responses during exposure. Masked hearing thresholds for simple stimuli in more spe- Effects of sound cies and individuals, as well as complex biological signals Continued effort is needed on the exposure on marine and realistic maskers; allowance for directional effects; simultaneous and residual physiologi- mammal hearing: comparative data on TTS-onset and growth in a greater cal effects of noise exposure on marine masking, TTS, and number of species and individuals for nonpulse and pulsed mammal hearing. PTS anthropogenic sources; recovery functions after exposures and between repeated exposures. Effects of sound Various baseline and exposure-condition measurements, Physiological measurements are needed exposure on marine including nitrogen saturation levels; bubble nuclei; the for both acute and chronic sound expo- mammal formation of hemorrhages, emboli, and/or lesions; stress sure conditions to investigate effects on non-auditory hormones; and cardiovascular responses to acute and non-auditory systems. systems chronic noise exposure. Various studies, including measurements and modeling related to (1) hearing sensitivity, (2) diving and vocaliza- Baseline and exposure data on these Particularly tion parameters, (3) tissue properties, (4) gas/fat emboli poorly understood taxa to assess their sensitive species: formation and significance, (5) advanced detection capa- apparent sensitivity to certain anthropo- beaked whales bilities for localizing and tracking them, and (6) behav- genic sound sources. ioral reactions to various anthropogenic and natural sound sources. 476 Southall et al. from masked thresholds. Auditory evoked poten- Auditory Scene Analysis tial (AEP) techniques should continue to be While baseline hearing information is clearly improved and standardized for pinnipeds and needed, urgently in some cases, more advanced, small cetaceans. Researchers should continue to comprehensive, and innovative measurements are develop procedures applicable to stranded indi- also needed that provide insight into the ways in viduals of species generally not represented in which animals use their auditory sense to derive captive settings, particularly for species that may detailed information about their surrounding envi- be especially sensitive to certain types of acous- ronment. For future iterations of noise exposure tic exposure. The massive body size of mysticetes criteria to consider multiple stimuli and cumula- may require that AEP studies begin using smaller tive effects, additional data will be needed on species (e.g., minke whale) that may be stranded, sound localization in three-dimensional auditory trapped in tidal fishing enclosures (weirs), or tem- space, frequency discrimination, temporal resolu- porarily available in a holding facility. Behavioral tion, and, specifically, detection of biological sig- audiometric methods, which investigate the effect nals in complex sound fields. of the overall perceptual and cognitive system on Several studies of terrestrial animals detection, should also continue to be employed (MacDougall-Shackleton et al., 1998; Moss & and improved, particularly those that increase the Surlykke, 2001) have investigated how subjects speed with which results are obtained without sac- process multiple acoustic stimuli that are simul- rificing precision of measurements. taneously present but differ in acoustic signature Additionally, behavioral methods should be either temporally or spatially. The acoustic scene developed to measure hearing characteristics concept, owing largely to the work of Bregman that require a subjective judgment of perception (1990), has the potential to play a major role in such as evaluation of equal loudness between two the development and progression of acoustic expo- acoustic stimuli. Equal-loudness hearing contours sure criteria. Bregman draws powerful analogies for marine mammals are needed to refine the broad between modalities of perception, including the frequency-weighting networks derived here. fundamental ways in which higher processing sys- A final consideration is that behavioral audio- tems associate common elements of complex stim- metric research should eventually move beyond uli in highly cluttered perceptual environments. the use of relatively simple artificial stimuli (e.g., One analogy that Bregman (1990) makes with pure tones, noise bands, broadband clicks, tone regard to the innate power of visual scene analysis pips). Such stimuli can be precisely controlled is the ability of the visual processing portion of the and can be used to clearly indicate which acoustic human brain to estimate object size without regard feature triggers the response in the whole animal to distance. The implication is that the reverse is or its auditory system. Animals in nature, how- true as well—if the size of something is known, ever, rarely encounter such sounds. While some its distance can be inferred from visual appear- biological signals consist of combinations of tonal ance. Extending this ability to animals that rely on elements, most are exceedingly complex. Marine underwater hearing to determine distance, similar mammal detection thresholds for complex, bio- perceptual processes may occur. If so, mammals logically relevant stimuli may be poorly predicted may determine range by using various effects of by experiments using simple artificial stimuli. the propagation medium on sound transmission Humans, for example, are particularly adept at (e.g., presence of structured multi-path signal identifying speech-like sounds in noise (Yost, spreading, frequency dependent multi-path losses, 2000). Animals are expected to be similarly sensi- and absorption effects in particular; Ellison & tive to important natural sounds. To base future Weixel, 1994). Further, both loudness modulation noise criteria on more relevant audiometric data, and source movement relative to the receiver pro- research is needed on detection thresholds for bio- vide significant clues as to the distance and general logically meaningful sounds, such as vocalizations nature of the sound source. If one considers sound of conspecifics, prey, and predators, and sounds to play a role in the life of marine wildlife similar needed for active or passive acoustic navigation. to that of sight in terrestrial animals, then context Such measurements will further be useful in inves- clues such as tempo, encroachment, and proximity tigating the potential active space (detection range must take on a powerful role in determining an ani- in three dimensions) for acoustic communication mal’s response to any given sound. These hypoth- (e.g., Brenowitz, 1982; Janik, 2000; Au et al., eses need to be studied in marine mammals. 2004) and the effects of anthropogenic sound on the active space. Field studies using biologically relevant sounds would be more relevant to real- world communication and masking than studies involving simple, artificial test stimuli. Marine Mammal Noise Exposure Criteria 477

Behavioral Responses of specific behavioral criteria. Duty cycle (the pro- Marine Mammals to Sound portion of time when the noise is present) is also likely to be important. Magnitude and duration of There is an urgent need for better and more exten- response are the most readily quantified param- sive data on behavioral responses to sound, includ- eters that may be useful in determining whether ing measurement of the specific acoustic features a behavioral response is likely to have a biologi- of exposures and consideration of previous expe- cally meaningful outcome. Noise exposure criteria rience with the sound and all relevant contextual should attempt to distinguish between minor, tem- variables. The current behavioral exposure criteria porary behavioral changes and those with greater are quite limited in several ways. Insufficient data significance. This is necessary in order to focus exist to support criteria other than those based on on biologically significant behavioral responses SPL alone, and this metric fails to account for the (see NRC, 2005) and the exposure conditions that duration of exposure beyond the separation of elicit them. pulse from nonpulse sounds. Also, there is much Considering the many contextual cues that free- variability in responses among species of the same ranging animals use to perceive and characterize functional hearing group and also within species. sound sources and to determine a response, it is Because of the poorly understood modify- not surprising that our analysis revealed a high ing influences of numerous variables, behavioral degree of variability in behavioral responses as a responses usually cannot be predicted a priori function of RL. Consequently, the logic of relying with much confidence given present information. solely on exposure RL as the metric for behav- In addition, the biological significance of any ioral responses is substantially diminished. A host observed behavioral response is even more diffi- of variables additional to RL may be important cult to assess (NRC, 2005). to marine mammals in assessing a sound and Research is needed to quantify behavioral reac- determining how to react. This argues for care- tions of a greater number of free-ranging marine ful design and execution of controlled exposure mammal species to specifically controlled or well- experiments to replicate the signal of interest in characterized exposures from different human as many dimensions as possible. Serious con- sound sources. The most direct way to obtain these sideration should be given to developing a broad kinds of extremely detailed data would be to attach multi-variable approach to behavioral research acoustic dosimeter tags to individuals and directly that takes into account not only source type and measure noise exposure, behavioral response, exposure level but also distance, motion, and rela- and physiological changes, if any. It is essential tive signal-to-noise ratio. Some studies are already that future research investigates responses in con- developing data of the scale and quality needed texts as similar as possible to those of interest. for such an approach. This includes studies pro- Responses of both naïve and previously exposed viding broad, long-term measurements of ambi- individuals should be studied and distinguished to ent sounds in areas cohabited by anthropogenic the greatest extent possible. sources and marine wildlife. Where these studies Further, such experiments must ensure that all include remotely deployed passive acoustic sen- relevant acoustic measurements of sound exposure sors and tagged animals, they approach what may be reported more systematically than in many pre- become the new standard. As additional infor- vious studies. Specifically, behavioral responses mation becomes available, future noise exposure need to be directly correlated with the physical criteria may assess behavioral reactions not only parameters (e.g., SPL, SEL) of the stimuli most according to RL measured with multiple acoustic likely to evoke the responses. Such research parameters, range (near and far), relative motion clearly requires greater knowledge of exposure (towards, parallel, etc.), and rate of change, but parameters (including SPL over some duration) also in relation to the animal’s activity or per- than currently exists for most studies. ceptual situation (e.g., neutral; threatened, as by The relationship between exposure SPL and/or a predator; or positive, related to food, mating, SEL and behavioral reaction should be determined etc.). for representative species within each functional The role of habituation and sensitization in hearing group. Whether the relationship follows behavioral reactions to noise exposure is a criti- a dose-response-like function for various sound cal subject for future research. These processes types, and under what conditions, is a significant can only be studied under controlled or well- and pressing open question. defined conditions (as in Deecke et al., 2002). A We need more data on the magnitude and time key question is how habituation and sensitization course of behavioral responses to known noise develop with repeated exposure in specific eco- exposures to test the validity of concepts outlined logically relevant circumstances. For example, here, and to make progress toward identifying the pattern of habituation to a neutral stimulus 478 Southall et al. is likely to follow quite a different pattern from reasonably well-known from laboratory studies in selective habituation to a harmless stimulus that mammals, including marine mammals. To enable is initially perceived as a threat (Deecke et al., masking to be included in subsequent noise expo- 2002). Furthermore, it would be desirable to know sure criteria, however, data are needed on mask- if there are common acoustic features in sounds ing and its effects in real-world conditions for to which marine mammals become sensitized. For all functional hearing groups. Data are needed example, to which acoustic features of a threat, on the masking effects of natural and anthropo- such as a vessel used to hunt animals, does an genic noise sources; on detection of simple, artifi- animal become sensitized? cial stimuli; and, increasingly, on more complex, Analyses of the behavior of various animal spe- biologically meaningful signals. Directionality in cies in the presence of predators suggest that they the masking sound and/or the signal of interest is have evolved anti-predator responses that mirror very likely to affect the severity of masking and their responses to human disturbance. According needs to be considered. Baseline measurements to predation risk theory, various ecological con- are needed on functional communication ranges siderations beyond simply disturbance magnitude for different acoustic signals and on the reduction are very likely involved in determining and pre- of those ranges caused by either natural or anthro- dicting behavioral response (Frid & Dill, 2002). pogenic maskers. Also needed are additional field The biological relevance of behavioral changes measurements of the behavioral adjustments that can only be determined in natural populations in marine mammals make to offset masking effects which vital life history parameters (e.g., reproduc- (e.g., Lesage et al., 1999; Serrano & Terhune, tion, growth, and survival rates) can be measured 2002; Foote et al., 2004; Scheifele et al., 2005). before and after noise exposure and in conditions where other potential stressors have been controlled Temporary Threshold Shift (TTS) (NRC, 2005). One important question is whether TTS studies in marine mammals remain limited these life history parameters are the same in popu- to a very few species and individuals, limiting the lations that have apparently habituated to expo- certainty with which they may be extrapolated sure and remain in relatively noisy environments within and among groups. A number of specific as they are in populations living in quieter condi- TTS studies are needed to improve future crite- tions. Because of the apparently major influence ria. For instance, it is critical to future iterations of experience and the strong context-specificity of of these noise exposure criteria that research on behavioral responses to noise, field measurements TTS-onset, TTS growth with noise exposure, and must be made for long periods following repeated recovery rates expands to larger numbers of indi- or continual exposure. Longitudinal studies should viduals and species, and to species in the low- and be conducted to assess the time course of expo- high-frequency cetacean groups. Presently, extrap- sure to various existing sound sources known or olation procedures must be used because TTS suspected to cause relatively long-term (seasonal) data are unavailable for certain functional hearing habitat abandonment. Where possible, parallel groups. Additionally, certain highly precautionary studies should be done in neighboring areas with procedures are used here in the estimation of PTS- different levels of noise exposure. Such studies onset because the growth rate of TTS with increas- should allow for other non-acoustic factors likely ing exposure level is generally poorly understood, to affect distribution such as predators, prey, and even for the few marine mammal species in which other important environmental covariates. These TTS has been studied. The relationship between studies will often need to extend over long peri- auditory sensitivity and susceptibility to TTS/PTS ods (many years) in order to be effective, and they should be determined by group. should be planned and funded recognizing that. To the extent possible, electrophysiological Ideally, such a study should start collecting data techniques should be used to obtain these TTS well in advance of the introduction of anthropo- data to increase sample size and knowledge of genic noise, and continue throughout the period of recovery functions. anticipated impact and for long enough thereafter More data for pinnipeds also are needed, par- to observe return to baseline. ticularly for pulse exposures where extrapolations of cetacean data currently must be used. Particular Effects of Noise Exposure on Marine Mammal emphasis should be placed on determining Hearing and Other Systems whether harbor seals have increased sensitivity to noise exposure relative to other pinniped species, Auditory Masking as current information suggests, and if so, whether Auditory masking is likely the most widespread species closely related to the harbor seal also are effect of anthropogenic noise on populations of more sensitive than are other pinnipeds. marine mammals. The principles of masking are Marine Mammal Noise Exposure Criteria 479

To minimize the need for such extrapolation to their own intense outgoing clicks (Henson, and reduce the assumptions required to predict 1965). The middle ears of marine mammals have PTS-onset, empirical data are needed on TTS some specialized adaptations relative to terrestrial growth rates up to greater shift magnitudes (10 to mammals (see Wartzok & Ketten, 1999). In water, 30 dB). These data are needed for both pulse and if bone conduction (rather than ossicular conduc- nonpulse sound types, at a variety of exposure fre- tion) is the predominant transmission path, it is quencies, in both single and multiple exposures. possible that a stapedial reflex, if present, may These results should further elucidate whether, have limited or no protective function for intense and in what conditions, the “equal energy hypoth- acoustic exposures. Research is also needed on the esis” may be appropriate for comparing the effects role of meatal closure in pinnipeds during noise of variable noise exposures in marine mammals. exposure. Such closures could be an alternative For pulse exposures, particular attention should or additional way of reducing auditory sensitivity. be paid to whether TTS growth is directly related Either mechanism could also affect the interpreta- to overall noise energy, and whether the kurtosis tion of threshold if performed during audiometric of exposure is also a factor (see Erdreich, 1986; measurements. Thiery & Meyer-Bisch, 1988; Dunn et al., 1991; Hamernik et al., 1993, 2003). Permanent Threshold Shift (PTS) A further topic for future research is deter- Sound exposures causing PTS-onset, used here mining whether using 40 dB of TTS as a proxy to define injury from acoustic exposure, have for PTS-onset is a precautionary approach, and not been measured in marine mammals. Instead, whether TTSs on the order of 25 to 35 dB are fully exposures that would cause PTS-onset are esti- recoverable in marine mammals as expected from mated from measured TTS-onset using assump- terrestrial mammal data. To avoid any possibility tions about the growth of TTS with noise expo- of injury, such studies should continue to take a sure level. Direct measurements of PTS in marine precautionary approach, using gradual increases mammals are highly desirable for establishing in exposure level and duration. future injury criteria, but they are unlikely to be A related question is how TTS recovers fol- obtained due to ethical, legal, and/or practical lowing noise cessation in variable conditions. considerations. Data from modeling and exposure Data on recovery functions and TTS magnitude of cadavers to very intense acoustic stimuli give are needed for representative species from each some indication of conditions causing PTS but do functional hearing group. Electrophysiological not reveal the exposure conditions that produce techniques may be particularly useful in this PTS in vivo, nor active processes that affect basi- regard. These data may be useful in comparing lar membrane displacement. Consequently, our basic auditory system responses to noise exposure research recommendations for improving PTS- and determining how summation procedures for onset predictions for marine mammals involve multiple exposures should be modified to more more indirect measures. precisely consider exposure intermittence. Levels One recommended type of indirect measure is of relatively long duration noise exposure causing to compare age-related hearing changes in captive asymptotic TTS, in which TTS values do not con- individuals that have been involved in TTS experi- tinue to increase in magnitude with exposure but ments with those that have not. This comparison may have longer-lasting effects, should be deter- may provide some insight into the complex rela- mined. Recovery functions from asymptotic TTS tionship between repeated TTS and PTS, which of various levels should be compared with recov- remains poorly understood for terrestrial mam- ery functions from non-asymptotic TTS. mals, including humans. One main impediment, Finally, the existence of a stapedial reflex in however, is that confounding variables likely exist marine mammals and its possible role in mitigat- other than controlled noise exposure. For captive ing TTS and other effects of intense noise expo- individuals used in TTS studies, absolute hear- sure are areas of needed research. For certain ing should be tested both during and following noise exposures, particularly those with relatively sequences of noise exposure experiments. For low frequencies and long duration, the middle ear captive individuals not used in TTS experiments, muscles (tensor tympani and stapedial) of terres- absolute hearing should be measured at regular trial mammals may contract and reduce the ampli- intervals over extended periods. The latter group fying function of the ossicular chain (Yost, 2000). may more readily display natural age-related hear- This muscular contraction reduces the amount of ing loss (presbycusis) than the former, as well as acoustic energy transmitted into the cochlea via potential sex differences. For both groups, efforts the stapes. This stapedial reflex has been demon- should be made to characterize long-term ambient strated in humans exposed to intense sound (Davis noise conditions experienced by test animals. et al., 1955) as well as echolocating bats exposed 480 Southall et al.

Non-Auditory Effects of Noise Exposure deep-diving marine mammals to determine behav- Lack of specific data on acoustic exposures caus- ioral responses to sound sources, including sonar. ing non-auditory effects in marine mammals cur- These experiments should use realistic source and rently prevents deriving explicit exposure criteria received levels. If responses are identified, this for such effects. Research is underway, however, may identify situations where it would be useful that may make this possible in future versions to conduct observational studies of responses of the criteria. Non-auditory effects of noise are during uncontrolled use of anthropogenic sound potentially significant but remain generally poorly sources. Research should characterize the changes understood. in diving behavior and should determine what they A current hypothesis regarding non-audi- mean in terms of bubble formation or growth with tory effects is that acoustic exposure may pro- continued exposure. duce nitrogen bubbles in blood or other tissues. Other possible non-auditory effects of acoustic Hemorrhages, gas and fat emboli, and other exposure should be investigated as well. Stress lesions have been reported in some marine mam- hormone levels associated with noise exposure mals exposed to mid-frequency military sonar should be more fully investigated. As of now, (Jepson et al., 2003; Fernández et al., 2004, 2005). they have been investigated following exposure of Substantial empirical questions remain, however. captive odontocetes to high-level sound (Thomas First among these is whether nitrogen bubbles are et al., 1990c; Romano et al., 2004). The ability of in fact responsible for the hemorrhages, emboli, animals to recruit effective stress responses should and other lesions reported. Conversely, are enough also be studied during chronic exposures—for nitrogen bubbles produced to pose a risk of related example, in captive animals that live permanently tissue injuries, under any set of circumstances, in noisy vs quiet environments. Effects of noise arising from high nitrogen supersaturation levels, exposure on marine mammal vestibular and car- acoustic exposure, and/or drastic changes in behav- diovascular systems should also be studied. ior? Do high levels of nitrogen supersaturation or gas or fat emboli occur in diving mammals that Particularly Sensitive Species have not been exposed to intense anthropogenic sound? Do these or related phenomena occur in In rare circumstances, marine mammals (primar- species other than beaked whales? If bubble for- ily beaked whales) have been known to strand mation is acoustically mediated, does it occur and ultimately die following exposure to tactical, as a direct result of acoustic exposure of bubble mid-frequency active sonar (see Cox et al., 2006; precursors (nuclei) in tissue, or indirectly through Nowacek et al., 2007). Our knowledge of these changes in diving behavior? If the pathway is kinds of extreme reactions to acute exposures direct, how does bubble formation and/or growth remains poor. However, the available informa- occur? A more thorough understanding is needed tion suggests that at least some species of beaked of lipid biochemistry in tissues that may be par- whales are particularly sensitive to this one spe- ticularly sensitive to acoustically mediated bubble cific category of sound sources. formation (e.g., acoustic fats). Modeling studies Gas bubble formation is a hypothesized path- are needed on tissue properties and their relevance way of this effect (e.g., Fernández et al., 2005), to nitrogen bubble formation at specific frequen- but it remains poorly understood and the precise cies of interest. These studies should consider the mechanism underlying these strandings remains growth of discrete bubbles from precursors in var- unknown (e.g., Cox et al., 2006). The controlled ious tissues, and the interaction among coalesced exposure experiments outlined above are essential aggregations of acoustically activated bubbles. to revealing the conditions and responses underly- If the pathway is indirect and mediated by behav- ing this effect. Until such research is conducted, ior, is rapid surfacing more risky than remaining deriving science-based exposure criteria specifi- submerged too long and exceeding physiological cally for beaked whales or other deep-diving ceta- limits? How does the dive profile affect the limits of ceans exposed to active sonar will prove difficult nitrogen supersaturation in normal diving? Do high or impossible. levels of nitrogen supersaturation and gas emboli One current hypothesis is that behavioral reac- occur in marine mammals that have voluntary tions influence beaked whale diving patterns in a control over depth, diving profile, and inter-dive way that induces physically debilitating or disori- interval? Resolution of these questions is likely to enting injuries (Cox et al., 2006). Both the specif- require interplay between modeling and empirical ics of this potential mechanism and whether it is measurements (Zimmer & Tyack, 2007). specific to beaked whales remains unknown, how- In conjunction with the above physiological ever. Mammals, including some marine mammals, modeling and measurements, controlled expo- show strong avoidance responses when evading sure experiments should be conducted with predators. Sounds from tactical mid-frequency Marine Mammal Noise Exposure Criteria 481

sonars somewhat resemble, in frequency band Necessary Progressions of Marine Mammal and modulation, the social signals of one of the Noise Exposure Criteria only predators of large marine mammals, the killer whale. If beaked whales inherit a broad tem- The currently proposed noise exposure criteria plate for acoustic detection of these predators, as are for individual sound exposures and individual waterfowl do for visual detection of aerial preda- marine mammals. The research recommended tors (Lorenz, 1939; Tinbergen, 1948), they might above is needed to substantiate and improve future respond to sonar as if it were a predator. Learning iterations of these types of criteria. Future itera- is required for selective habituation to safe stim- tions of behavioral disturbance criteria may derive uli that resemble those from predators (Deecke dose-response functions based on an ordinal scor- et al., 2002). Many of the strandings that coincide ing paradigm similar to that provided. This may with sonar exercises have occurred in sites where occur for subcategories of sound sources within killer whales are rare. Possibly these stranded the general categories here (e.g., seismic signals animals have not had enough experience with as a subset of multiple pulses, vessel noise as a either sonar or killer whales to learn the differ- subset of nonpulses). It may also occur for sub- ence. Propagation of sound in the ocean may also groups of species within the broad categories rec- degrade acoustic features that help differentiate ognized here (e.g., phocid vs otariid pinnipeds) and the two classes of stimuli at a distance. It is plau- for other types of marine mammals not addressed sible that this type of reaction could occur at rela- here (e.g., sirenians, sea otters, polar bears). tively long distances from the source if the sound Future iterations of these noise exposure criteria is alarming based on properties other than high should also perhaps distinguish several different RL. categories of response that are expected, for both Whether beaked whales in certain conditions theoretical and empirical reasons, to vary with mistake tactical mid-frequency sonar signals RL in different ways. For example, if an animal for killer whales and consequently change their responds to a sound as if it were from a predator behavior in a way that injures them is an empiri- (Frid & Dill, 2002), one would expect the dose- cal question. This should be carefully investigated response function to have a very different shape using controlled experiments that take into account as compared to that if the animal responds based the relevant contextual variables discussed above. on interference with the animal’s ability to com- Additional baseline data on beaked whale physi- municate acoustically or echolocate. Predicting ology, life history, and behavior are also needed whether a sound might trigger an anti-predator to appropriately address questions regarding the response would require more detailed analyses of apparent sensitivity of these animals to certain acoustic parameters of the anthropogenic sound kinds of anthropogenic sound. Finally, in some compared to signals of predators. Further, in some specific conditions, such as sonar training ranges, non-marine taxa, different anti-predator responses where sophisticated listening arrays make it pos- may be triggered depending on levels and other sible to detect marine mammals over large ranges characteristics of acoustic stimuli (Spangler, 1988; before and during active sonar operations, active Hoy, 1989) and may be modulated by the cost of or passive detection of marine mammal behav- the response as well as the perceived risk (Frid ioral patterns may become increasingly possible. & Dill, 2002). Behavioral ecologists hypothesize While these observations have limitations, given that anti-predator behavior should balance risk of that they may be able to detect more individuals predation against cost of response, including cost without requiring tagging efforts, they may be an of foregone benefits from alternative activities important complement to directed experiments. (Frid & Dill, 2002). These non-acoustic param- Some other species of marine mammals are eters must be taken into account in order to under- unusually responsive to certain anthropogenic stand disturbance responses. The acoustic param- sounds, either generally or under particular condi- eters affecting anti-predator behavior may involve tions, and this can result in strong and sometimes detection thresholds, ambient noise conditions, large-scale avoidance. Examples include harbor source distance and source movement, as well porpoises and, in some but not all situations, as the more direct measures of received sound. beluga and bowhead whales (Finley et al., 1990; In future studies, most or all of these parameters Richardson et al., 1999; Olesiuk et al., 2002; should be measured. Miller et al., 2005). There is a need for additional Additionally, further exposure criteria are behavioral and acoustic information to better char- needed to fully consider the effects of anthropo- acterize these extreme responses, the situations in genic noise on other types of marine life, includ- which they occur, and whether similar responses ing the effects of single and multiple exposures can occur in other related species or in response to on individual invertebrates, fish, and sea turtles as other similar stimuli. well as sirenians, sea otters, and polar bears. There 482 Southall et al. are fewer data to support criteria for marine biota Solow, Sofie Van Parijs, Willem Verboom, Donna other than cetaceans and pinnipeds, and criteria Wieting, and five anonymous reviewers. are perhaps as urgently (or more urgently) needed for some other groups. Some fish and most sea Literature Cited turtle species are considered threatened or endan- gered. The effects of anthropogenic noise on Ahroon, W. A., Hamernik, R. P., & Davis, R. I. (1993). fish are also of particular importance given their Complex noise exposures: An energy analysis. Journal central role as both predators and prey in many of the Acoustical Society of America, 93, 997-1006. marine ecosystems and because of human depen- Ahroon, W. A., Hamernik, R. P., & Lei, S-F. (1996). The dence on fisheries. effects of reverberant blast waves on the auditory Additional criteria are also needed for the cumu- system. Journal of the Acoustical Society of America, lative effects of repetitive or long-term noise expo- 100, 2247-2257. sure on marine mammals. Ideally, spatiotemporal Akamatsu, T., Hatakeyama, Y., & Takatsu, N. (1993). data on marine ambient noise and long-term expo- Effects of pulse sounds on escape behavior of false killer sure histories of individuals should be integrated whales. Bulletin of the Japanese Society of Scientific with vital rate data for marine mammal popula- Fisheries, 59, 1297-1303. tions to address this question. Considerably more Allen, S. G., Ainley, D. G., Page, G. W., & Ribic, C. A. data are needed on how noise impacts in single (1984). The effect of disturbance on harbor seal haul-out animals can be extended to the population level. patterns at Bolinas Lagoon, California. Fishery Bulletin, Such measurements will likely require extensive 82, 493-500. measurements on a few representative species and American National Standards Institute (ANSI). (1986). conservative extrapolations within and between Methods for measurement of impulse noise (ANSI S12.7- functional hearing groups. 1986). New York: Acoustical Society of America. Noise exposure criteria that consider ecosystem- ANSI. (1994). Acoustical terminology (ANSI S1.1-1994). level effects are needed as well. It is possible that New York: Acoustical Society of America. the effects of noise exposure on some elements of ANSI. (2001). Acoustical terminology (ANSI S1.42-2001). local food webs may have a cascade effect to other New York: Acoustical Society of America. elements within the web. No data are available on Andersen, S. (1970). Auditory sensitivity of the harbor the ecological effects of underwater noise, even porpoise, Phocoena phocoena. In G. Pilleri (Ed.), at a local scale. However, given the upward trend Investigations on Cetacea, Volume II (pp. 255-259). in human activities in many nearshore areas, such Berne, Switzerland: Berne-Bümpliz. ecological effects should be anticipated. André, M., & Nachtigall, P. E. (2007). Electrophysiological Progress in each of these research areas will measurements of hearing in marine mammals. Aquatic involve iterative processes that depend on the Mammals, 33, 1-5. availability of relevant scientific data. Like the André, M., Terada, M., & Watanabe, Y. (1997). Sperm process of improving and expanding future noise whale (Physeter macrocephalus) behavioural response exposure criteria, our ability to understand and after the playback of artificial sounds. Report of the predict the effects of anthropogenic noise expo- International Whaling Commission, 47, 499-504. sure on marine ecosystems will continue to evolve André, M., Delory, E., Degollada, E., Alonso, J., del Rio, over a period of many decades. J., van der Schaar, M., et al. (2007). Identifying cetacean hearing impairment at stranding sites. Aquatic Mammals, Acknowledgments 33, 100-109. Angerer, J. R., McCurdy, D. A., & Erickson, R. A. (1991). The authors acknowledge three former mem- Development of an annoyance model based upon audi- bers of the marine mammal noise exposure cri- tory sensations for steady-state aircraft inter-noise teria group: Whitlow Au, Sam Ridgway, and containing tonal components (NASA TM-104147). Ronald Schusterman. Partial financial support Hampton, VA: NASA, Langley Research Center. for meetings of the noise exposure criteria group Au, D., & Perryman, W. (1982). Movement and speed of was provided by the U.S. National Oceanic and dolphin schools responding to an approaching ship. Atmospheric Administration, National Marine Fishery Bulletin, 80, 371-379. Fisheries Service, Ocean Acoustics Program. Au, W. W. L. (1993). The sonar of dolphins. New York: The authors also appreciate the expertise, Springer-Verlag. review, and assistance of the following people: Au, W. W. L., & Green, M. (2000). Acoustic interaction of Laurie Allen, Robyn Angliss, Dan Costa, Tara humpback whales and whale-watching boats. Marine Cox, Ned Cyr, Adam Frankel, Ron Kastelein, Environmental Research, 49, 469-481. Steve Leathery, Rob McCauley, David Mellinger, Au, W. W. L., Nachtigall, P. E., & Pawloski, J. L. (1997). Richard Merrick, Douglas Nowacek, Deborah Acoustic effects of the ATOC signal (75 Hz, 195 dB) on Palka, Michael Payne, Arthur Popper, Andrew Marine Mammal Noise Exposure Criteria 483

dolphins and whales. Journal of the Acoustical Society Berg, E. A., Nieto, M. P., Thorson, P. H., Francine, J. K., of America, 101, 2973-2977. & Oliver, G. (2001). Acoustic measurements of the 21 Au, W. W. L., Ford, J. K. B., Horne, J. K., & Newman November 2000 Delta II EO-1 launch and quantitative Allman, K. A. (2004). Echolocation signals of free- analysis of behavioral responses of Pacific harbor seals, ranging killer whales (Orcinus orca) and modeling of brown pelicans, and southern sea otters on Vandenberg foraging for chinook salmon (Oncorhynchus tshawyts- Air Force Base and selected pinnipeds on San Miguel cha). Journal of the Acoustical Society of America, 115, Island, CA. Los Angeles Air Force Base: Report by 901-909. SRS Technologies, Systems Development Division, Au, W. W. L., Pack, A. A., Lammers, M. O., Herman, L. M., for Space and Missile Systems Center, U.S. Air Force Deakos, M. H., & Andrews, K. (2006). Acoustic proper- Materiel Command. ties of humpback whale songs. Journal of the Acoustical Berg, E. A., Nieto, M. P., Thorson, P. H., Francine, J. K., Society of America, 120, 1103-1110. & Oliver, G. (2002). Acoustic measurements of the 5 Au, W. W. L., Thomas, J. A., & Ramirez, K. T. (2007). October 2001 Titan IV B-34 launch and quantitative Characteristics of the auditory brainstem evoked poten- analysis of behavioral responses of Pacific harbor seals tial of a Pacific white-sided dolphin (Lagenorhynchus on Vandenberg Air Force Base, California. Los Angeles obliquidens). Aquatic Mammals, 33, 76-84. Air Force Base: Report by SRS Technologies, Systems Awbrey, F. T., & Stewart, B. S. (1983). Behavioral responses Development Division, for Space and Missile Systems of wild beluga whales (Delphinapterus leucas) to noise Center, U.S. Air Force Materiel Command. from oil drilling. Journal of the Acoustical Society of Berg, E. A., Cordova, D. A., Fillmore, L. E., Oliver, America, 74, S54. G. W., & Thorson, P. H. (2004). Acoustic measurements Awbrey, F. T., Thomas, J. A., & Kastelein, R. A. (1988). of the 2 December 2003 Atlas IIAS MLV-14 launch and Low frequency underwater hearing sensitivity in belu- quantitative analysis of behavioral responses of pin- gas, Delphinapterus leucas. Journal of the Acoustical nipeds on San Miguel Island, California. Los Angeles Society of America, 84, 2273-2275. Air Force Base: Report by SRS Technologies, Babushina, Ye. S., Zaslavskii, G. L., & Yurkevich, L. I. Information Systems Division, for U.S. Air Force (1991). Air and underwater hearing characteristics of Materiel Command. the northern fur seal: Audiograms, frequency and differ- Biassoni, N., Miller, P. J., & Tyack, P. L. (2000). Preliminary ential thresholds. Biophysics, 36, 900-913. results of the effects of SURTASS-LFA sonar on singing Baggeroer, A., & Munk, W. (1992). The Heard Island feasi- humpback whales (Technical Report #2000-06). Woods bility test. Physics Today, 45(9), 22-30. Hole, MA: Woods Hole Oceanographic Institute. 23 pp. Baker, C. S., Herman, L. M., Bays, B. G., & Stifel, Bjork, E., Nevalainen, T., Hakumaki, M., & Voipio, H-M. W. F. (1982). The impact of vessel traffic on the behav- (2000). R-weighting provides better estimation for rat ior of humpback whales in southeast Alaska. Honolulu: hearing sensitivity. Laboratory Animals, 34, 136-144. Research from Kewalo Basin Marine Mammal Blackwell, S. B., Greene, C. R., Jr., & Richardson, W. J. Laboratory for U.S. National Marine Fisheries Service, (2004a). Drilling and operational sounds from an oil Seattle, WA. 78 pp. production island in the ice-covered Beaufort Sea. Barlow, J., & Cameron, G. A. (2003). Field experiments Journal of the Acoustical Society of America, 116, 3199- show that acoustic pingers reduce marine mammal 3211. by-catch in the California drift gill net fishery. Marine Blackwell, S. B., Lawson, J. W., & Williams, M. T. (2004b). Mammal Science, 19, 265-283. Tolerance by ringed seals (Phoca hispida) to impact Barnes, L. G., Domning, D. P., & Ray, C. E. (1985). Status pile-driving and construction sounds at an oil production of studies on fossil marine mammals. Marine Mammal island. Journal of the Acoustical Society of America, Science, 1, 15-53. 115, 2346-2357. Beale, C. M., & Monaghan, P. (2004). Behavioral responses Boettcher, F. A. (1993). Auditory brain-stem response cor- to human disturbance: A matter of choice? Animal relates of resistance to noise-induced hearing loss in Behavior, 68, 1065-1069. Mongolian gerbils. Journal of the Acoustical Society of Beedholm, K., & Miller, L. A. (2007). Automatic gain con- America, 94, 3207-3214. trol in harbor porpoises (Phocoena phocoena)? Central Bordino, P., Kraus, S., Albareda, D., Fazio, A., Palmerio, A., versus peripheral mechanisms. Aquatic Mammals, 33, Mendez, M., et al. (2002). Reducing incidental mortality 69-75. of Franciscana dolphins with acoustic warning devices Begall, S., Burda, H., & Schneider, B. (2004). Hearing in attached to fishing nets. Marine Mammal Science, 18, coruros (Spalacopus cyanus): Special audiogram fea- 833-842. tures of a subterranean rodent. Journal of Comparative Borggaard, D., Lien, J., & Stevick, P. (1999). Assessing the Physiology A, 190, 963-969. effects of industrial activity on large cetaceans in Trinity Beranek, L. L. (1989). Applications of NCB noise criterion Bay, Newfoundland (1992-1995). Aquatic Mammals, curves. Noise Control Engineering Journal, 33, 64-73. 25, 149-161. Beranek, L. L., & Ver, I. L. (1992). Noise & vibration con- Born, E. W., Riget, F. F., Dietz, R., & Andriashek, D. trol engineering. New York: Wiley. (1999). Escape responses of hauled out ringed seals 484 Southall et al.

(Phoca hispida) to aircraft disturbance. Polar Biology, of juvenile northern elephant seals, Mirounga angustiro- 21, 171-178. stris. Journal of the Acoustical Society of America, 113, Brant, L. J., & Fozard, J. L. (1990). Age changes in pure- 1155-1165. tone hearing thresholds in a longitudinal study of normal Cox, T. M., Read, A. J., Solow, A., & Tregenza, N. (2001). human aging. Journal of the Acoustical Society of Will harbor porpoises (Phocoena phocoena) habitu- America, 88, 813-820. ate to pingers? Journal of Cetacean Research and Bregman, A. S. (1990). Auditory scene analysis. Cambridge: Management, 3, 81-86. MIT Press. Cox, T. M., Read, A. J., Swanner, D., Urian, K., & Waples, Brenowitz, E. A. (1982). The active space of red-winged D. (2003). Behavioral responses of bottlenose dolphins blackbird song. Journal of Comparative Physiology, (Tursiops truncatus) to gillnets and acoustic alarms. 147, 511-522. Biological Conservation, 115, 203-212. Brill, R. L., Moore, P. W. B., & Dankiewicz, L. A. (2001). Cox, T. M., Ragen, T. J., Read, A. J., Vos, E., Baird, R. W., Assessment of dolphin (Tursiops truncatus) auditory Balcomb, K., et al. (2006). Understanding the impacts sensitivity and hearing loss using jawphones. Journal of of anthropogenic sound on beaked whales. Journal of the Acoustical Society of America, 109, 1717-1722. Cetacean Research and Management, 7, 177-187. British Government Planning Policy Guidance (BG PPG). Crocker, M. J. (1997). Rating measures, descriptors, criteria, (1994). Planning and noise (Meeting summary). and procedures for determining human response to noise. Britt, J. R., Eubanks, R. J., & Lumsden, M. G. (1991). In M. J. Crocker (Ed.), Encyclopedia of acoustics, Volume Underwater shock wave reflection and refraction in deep III (pp. 943-965). New York: John Wiley & Sons. and shallow water. In A user’s manual for the REFMS Croll, D. A., Clark, C. W., Calambokidis, J., Ellison, W. T., code (Version 4.0) (Volume 1). St. Joseph, LA: Science & Tershy, B. R. (2001). Effects of anthropogenic low- Applications International Corporation. frequency noise on the foraging ecology of Balaenoptera Buckstaff, K. C. (2004). Effects of watercraft noise on whales. Animal Conservation, 4, 13-27. the acoustic behavior of bottlenose dolphins, Tursiops Crum, L. A., & Mao, Y. (1996). Acoustically enhanced truncatus, in Sarasota Bay, Florida. Marine Mammal bubble growth at low frequencies and its implications Science, 20, 709-725. for human diver and marine mammal safety. Journal of Bullock, T. H., Grinnell, A. D., Ikezono, F., Kameda, the Acoustical Society of America, 99, 2898-2907. K., Katsuki, Y., Nomoto, M., et al. (1968). Electro- Cudahy, E., & Ellison, W. (2002). A review of the poten- physiological studies of the central auditory mechanisms tial for in vivo tissue damage by exposure to underwater in cetaceans. Zeitschrift für Vergleichende Physiologie sound. Groton, CT: Naval Submarine Medical Research (Springer-Verlag), 59, 117-156. Library. Bullock, T. H., Ridgway, S. H., & Suga, N. (1971). Culik, B. M., Koschinski, S., Tregenza, M., & Ellis, G. Acoustically evoked potentials in midbrain auditory (2001). Reactions of harbor porpoise (Phocoena pho- structures in sea lions (Pinnipedia). Zeitschrift für coena) and herring (Clupea harengus) to acoustic Vergleichende Physiologie (Springer-Verlag), 74, 372- alarms. Marine Ecology Progress Series, 211, 255-260. 387. D’Spain, G. L., D’Amico, A., & Fromm, D. M. (2006). Burdic, W. S. (1984). Underwater acoustic system analysis. Properties of the underwater sound fields during some well Englewood Cliffs, NJ: Prentice Hall. documented beaked whale mass stranding events. Journal Buus, S. (1997). Auditory masking. In M. J. Crocker (Ed.), of Cetacean Research and Management, 7, 223-238. Encyclopedia of acoustics, Volume III (pp. 1427-1446). Dahlheim, M. E. (1987). Bio-acoustics of the gray whale New York: John Wiley & Sons. (Eschrichtius robustus). Ph.D. dissertation, University Chilvers, B. L., & Corkeron, P. J. (2001). Trawling and of British Columbia, Vancouver. 315 pp. bottlenose dolphins’ social structure. Proceedings of the Dahlheim, M. E., Fisher, H. D., & Schempp, J. D. (1984). Royal Society of London B, 268, 1901-1905. Sound production by the gray whale and ambient noise Clark, C. W., & Ellison, W. T. (2004). Potential use of levels in Laguna San Ignacio, Baja California Sur, low-frequency sound by baleen whales for probing Mexico. In M. L. Jones, S. L. Swartz, & S. Leatherwood the environment: Evidence from models and empirical (Eds.), The gray whale Eschrichtius robustus (pp. 511- measurements. In J. A. Thomas, C. F. Moss, & M. Vater 541). New York: Academic Press. (Eds.), Echolocation in bats and dolphins (pp. 564-581). Davis, R. C., Buchwald, A. M., & Frankman, R. W. (1955). Chicago: University of Chicago Press. Autonomic and muscular stimuli and their relation to Cook, M. L. H., Varela, R. A., Goldstein, J. D., McCulloch, simple stimuli. Psychological Monographs, 69(405). S. D., Bossart, G. D., Finneran, J. J., et al. (2006). Beaked Deecke, V. B., Slater, P. J. B., & Ford, J. K. B. (2002). whale auditory evoked potential hearing measurements. Selective habituation shapes acoustic predator recogni- Journal of Comparative Physiology A, 192, 489-495. tion in harbour seals. Nature, 420, 171-173. Costa, D. P., Crocker, D. E., Gedamke, J., Webb, P. M., Delaney, D. K., Grubb, T. G., Beier, P., Pater, L. L., & Houser, D. S., Blackwell, S. B., et al. (2003). The Reiser, M. H. (1999). Effects of helicopter noise on effect of a low-frequency sound source (Acoustic Mexican spotted owls. Journal of Wildlife Management, Thermometry of Ocean Climate) on the diving behavior 63, 60-76. Marine Mammal Noise Exposure Criteria 485

Delory, E., del Rio, J., Castell, J., van der Schaar, M., & Fernández, A., Edwards, J. F., Rodríguez, F., Espinosa André, M. (2007). OdiSEA: An autonomous portable de los Monteros, A., Herráez, P., Castro, P., et al. (2005). auditory screening unit for rapid assessment of hearing Gas and fat embolic syndrome involving a mass strand- in cetaceans. Aquatic Mammals, 33, 85-92. ing of beaked whales (Family Ziphiidae) exposed to Dolphin, W. F. (2000). Electrophysiological measures of anthropogenic sonar signals. Veterinary Pathology, 42, auditory processing in odontocetes. In W. W. L. Au, 446-457. A. N. Popper, & R. R. Fay (Eds.), Hearing by whales and Finley, K. J., Miller, G. W., Davis, R. A., & Greene, C. R., dolphins (pp. 294-329). New York: Springer-Verlag. Jr. (1990). Reactions of belugas, Delphinapterus leucas, Dolphin, W. F., Au, W. W. L., Nachtigall, P. E., & Pawloski, and narwhals, Monodon monoceros, to ice-breaking J. (1995). Modulation rate transfer functions to low- ships in the Canadian high arctic. Canadian Bulletin of frequency carriers in three species of cetaceans. Journal Fisheries and Aquatic Sciences, 224, 97-117. of Comparative Physiology, 177, 235-245. Finneran, J. J. (2003). Whole-lung resonance in a bottle- Domning, D. P., Morgan G. S., & Ray, C. E. (1982). North nose dolphin (Tursiops truncatus) and a white whale American Eocene sea cows (Mammalia: Sirenia). (Delphinapterus leucas). Journal of the Acoustical Smithsonian Contributions to Paleobiology, 52, 1-69. Society of America, 114, 529-535. Dunn, D. E., Davis, R. R., Mercy, C. J., & Franks, J. R. Finneran, J. J., & Houser, D. S. (2006). Comparison of in- (1991). Hearing loss in the chinchilla from impact and air evoked potential and underwater behavioral hearing continuous noise exposure. Journal of the Acoustical thresholds in four bottlenose dolphins (Tursiops trunca- Society of America, 90, 1979-1985. tus). Journal of the Acoustical Society of America, 119, Echteler, S. M., Fay, R. R., & Popper, A. N. (1994). 3181-3192. Structure of the mammalian cochlea. In R. R. Fay & Finneran, J. J., & Schlundt, C. E. (2004). Effects of intense A. N. Popper (Eds.), Comparative hearing: Mammals pure tones on the behavior of trained odontocetes (pp. 134-171). New York: Springer-Verlag. (SPAWAR Systems Command Technical Report #1913). Egan, J. P. (1975). Signal detection theory and ROC analy- San Diego: U.S. Navy. sis. New York: Academic Press. Finneran, J. J., Schlundt, C. E., Carder, D. A., Clark, J. A., Ellison, W. T., & Weixel, K. (1994). Considerations for Young, J. A., Gaspin, J. B., et al. (2000). Auditory and designing underwater acoustical playback experiments. behavioral responses of bottlenose dolphins (Tursiops Journal of the Acoustical Society of America, 96(5, pt. truncatus) and white whales (Delphinapterus leucas) A), 3316-3317. to impulsive sounds resembling distant signatures of Environmental Protection Agency (EPA). (1974). underwater explosions. Journal of the Acoustical Society Information on levels of environmental noise requisite of America, 108, 417-431. to protect public health and welfare with an adequate Finneran, J. J., Carder, D. A., & Ridgway, S. H. (2002a). margin of safety (Report #550/9-74-004). Washington, Low frequency acoustic pressure, velocity and intensity DC. thresholds in a bottlenose dolphin (Tursiops truncatus) Erbe, C. (2002). Hearing abilities of baleen whales and white whale (Delphinapterus leucas). Journal of the (Contractor report #DRDC Atlantic CR 2002-065). Acoustical Society of America, 111, 447-456. Dartmouth, NS: Defence R&D Canada – Atlantic. 30 Finneran, J. J., Schlundt, C. E., Dear, R., Carder, D. A., pp. & Ridgway, S. H. (2002b). Temporary shift in masked Erdreich, J. (1986). A distribution based definition of hearing thresholds (MTTS) in odontocetes after expo- impulse noise. Journal of the Acoustical Society of sure to single underwater impulses from a seismic America, 79, 990-998. watergun. Journal of the Acoustical Society of America, Evans, D. L., & England, G. R. (2001). Joint interim 111, 2929-2940. report/Bahamas marine mammal stranding event of Finneran, J. J., Dear, R., Carder, D. A., & Ridgway, S. H. 14-16 March 2000. Washington, DC: National Oceanic (2003). Auditory and behavioral responses of California and Atmospheric Administration, U.S. Department of sea lions (Zalophus californianus) to single underwa- Commerce, and U.S. Navy. 61 pp. ter impulses from an arc-gap transducer. Journal of the Fahy, F. J. (1995). Sound intensity (2nd ed.). London: E and Acoustical Society of America, 114, 1667-1677. FN Spon. 295 pp. Finneran, J. J., Carder, D. A., Schlundt, C. E., & Ridgway, Fay, R. R. (1988). Hearing in vertebrates: A psychophysics S. H. (2005a). Temporary threshold shift (TTS) in bot- databook. Winnetka, IL: Hill-Fay Associates. tlenose dolphins (Tursiops truncatus) exposed to mid- Fay, R. R., & Popper, A. N. (2000). Evolution of hearing frequency tones. Journal of the Acoustical Society of in vertebrates: The inner ears and processing. Hearing America, 118, 2696-2705. Research, 149, 1-10. Finneran, J. J., Carder, D. A., Dear, R., Belting, T., McBain, Fernández, A., Arbelo, M., Deaville, R., Patterson, I. A. P., J., Dalton, L., et al. (2005b). Pure tone audiograms and Castro, P., Baker, J. R., et al. (2004). Pathology: Whales, possible aminoglycoside-induced hearing loss in belugas sonar and decompression sickness (reply). [Brief (Delphinapterus leucas). Journal of the Acoustical Communications]. Nature, 428(6984), U1-2. Society of America, 117, 3936-3943. 486 Southall et al.

Finneran, J. J., Houser, D. S., & Schlundt, C. E. (2007a). In B. R. Melteff & D. H. Rosenberg (Eds.), Workshop Objective detection of bottlenose dolphin (Tursiops on biological interactions among marine mammals and truncatus) steady-state auditory evoked potentials in commercial fisheries in the southeastern Bering Sea (pp. response to AM/FM tones. Aquatic Mammals, 33, 43- 187-200). Anchorage: University of Alaska Press. 54. Gentry, R. L. (1967). Underwater auditory localization in Finneran, J. J., Schlundt, C. E., Branstetter, B., & Dear, the California sea lion (Zalophus californianus). Journal R. L. (2007b). Assessing temporary threshold shift in a of Auditory Research, 7, 187-193. bottlenose dolphin (Tursiops truncatus) using multiple Gentry, R. L., Gentry, E. C., & Gilman, J. F. (1990). simultaneous auditory evoked potentials. Journal of the Responses of northern fur seals to quarrying operations. Acoustical Society of America, 122, 1249-1264. Marine Mammal Science, 6, 151-155. Fitch, W. T. (1999). Acoustic exaggeration of size in birds Gill, J. A., & Sutherland, W. J. (2000). Predicting the via tracheal elongation: Comparative and theoretical consequences of human disturbance from behavioral analyses. Journal of Zoology London, 248, 31-48. decisions. In L. M. Gosling & W. J. Sutherland (Eds.), Fitch, W. T., & Hauser, M. D. (1995). Vocal production in Behavior and conservation (pp. 51-64). Cambridge, nonhuman primates: Acoustics, physiology and func- UK: Cambridge Press. tional constraints on “honest” advertisement. American Gill, J. A., Norris, K., & Sutherland, W. J. (2001). Why Journal of Primatology, 37, 191-219. behavioral responses may not reflect the population Fletcher, H. F. (1940). Auditory patterns. Review of Modern consequences of human disturbance. Biological Physics, 12, 47-65. Conservation, 97, 265-268. Fletcher, H. F., & Munson, W. A. (1933). Loudness, its Goold, J. C., & Fish, P. J. (1998). Broadband spectra of seis- definition, measurement, and calculation. Journal of the mic survey air-gun emissions, with reference to dolphin Acoustical Society of America, 5, 82-108. auditory thresholds. Journal of the Acoustical Society of Foote, A. D., Osborne, R. W., & Hoelzel, A. R. (2004). America, 103, 2177-2184. Whale-call response to masking boat noise. Nature, 428, Gordon, J., Leaper, R., Hartley, F. G., & Chappell, O. 910. (1992). Effects of whale-watching vessels on the surface Frankel, A. S., & Clark, C. W. (1998). Results of low-fre- and underwater acoustic behaviour of sperm whales quency playback of M-sequence noise to humpback off Kaikoura, New Zealand. In Science and research whales, Megaptera novaeangliae, in Hawai’i. Canadian series (p. 64). Wellington: New Zealand Department of Journal of Zoology, 76, 521-535. Conservation. Frankel, A. S., & Clark, C. W. (2000). Behavioral responses Gordon, J., Gillespie, D., Potter, J., Frantzis, A., Simmonds, of humpback whales (Megaptera novaeangliae) to full- M. P., Swift, R., et al. (2004). A review of the effects of scale ATOC signals. Journal of the Acoustical Society of seismic surveys on marine mammals. Marine Technology America, 108, 1930-1937. Society Journal, 37, 16-34. Frankel, A. S., & Clark, C. W. (2002). ATOC and other fac- Green, D. M. (1976). An introduction to hearing. tors affecting the distribution and abundance of hump- New York: John Wiley & Sons. back whales (Megaptera novaengliae) off the north Green, D. M., & Swets, J. (1974). Signal detection theory shore of Kaua’i. Marine Mammal Science, 18, 644-662. and psychophysics. New York: Robert E. Krieger, Frantzis, A. (1998). Does acoustic testing strand whales? Huntington. Nature, 392, 29. Greene, C. R., Jr. (1997). Physical acoustics measurements. French, N. R., & Steinberg, J. C. (1947). Factors govern- In W. J. Richardson (Ed.), Northstar marine mammal ing the intelligibility of speech sounds. Journal of the monitoring program, 1996 (LGL Report 2121-2, Section Acoustical Society of America, 19, 90-119. 3). LGL Ltd. report for BP Exploration (Alaska) Inc., Frid, A., & Dill, L. (2002). Human-caused disturbance stim- Anchorage, AK, and National Marine Fisheries Service, uli as a form of predation risk. Conservation Ecology, 6, Silver Spring, MD. 245 pp. 11. Greene, C. R., Jr., & Richardson, W. J. (1988). Characteristics Fristrup, K. M., Hatch, L. T., & Clark, C. W. (2003). of marine seismic survey sounds in the Beaufort Sea. Variation in humpback whale (Megaptera novaeangliae) Journal of the Acoustical Society of America, 83, 2246- song length in relation to low-frequency sound broad- 2254. casts. Journal of the Acoustical Society of America, 113, Greenwood, D. D. (1961). Auditory masking and the criti- 3411-3424. cal band. Journal of the Acoustical Society of America, Frost, K. J., & Lowry, L. F. (1988). Effects of industrial 33, 484-502. activities on ringed seals in Alaska, as indicated by aerial Hales, J. R. S. (1973). Effects of heat stress on blood flow surveys. In W. M. Sackinger, M. O. Jeffries, J. L. Imm, in respiratory and non-respiratory muscles in the sheep. & S. D. Treacy (Eds.), Port and ocean engineering under Pflüg. Archive of the European Journal of Physiology Arctic conditions, Volume II (pp. 15-25). Fairbanks: (Historical Archive), 345, 123-130. Geophysical Institute, University of Alaska–Fairbanks. Hall, J. D., & Johnson, C. S. (1972). Auditory thresholds Frost, K. J., Lowry, L. F., Gilbert, J. R., & Burns, J. J. of a killer whale, Orcinus orca, Linnaeus. Journal of the (1984). Beluga whale studies in Bristol Bay, Alaska. Acoustical Society of America, 51, 515-517. Marine Mammal Noise Exposure Criteria 487

Halmagyi, G. M., Curthoys, I. S., Colebatch, J. G., & Aw, Management Service. 39 pp. Retrieved 22 October 2007 S. T. (2005). Vestibular responses to sound. Clinical and from www.mms.gov/omm/pacific/lease/fullhessrept.pdf. basic oculomotor research: In honor of David S. Zee. Holst, M., Lawson, J. W., Richardson, W. J., Schwartz, S. J., Annals of the New York Academy of Sciences, 1039, 54- & Smith, G. (2005a). Pinniped responses during Navy 67. missile launches at San Nicolas Island, California. In Hamernik, R. P., & Hsueh, K. D. (1991). Impulse noise: D. K. Garcelon & C. A. Schwemm (Eds.), Proceedings Some definitions, physical acoustics and other consid- of the 6th California Islands Symposium (Ventura, CA) erations. Journal of the Acoustical Society of America, (National Park Service Technical Publication CHIS- 90, 189-196. 05-01) (pp. 477-484). Arcata, CA: Institute of Wildlife Hamernik, R. P., Ahroon, W. A., Hsueh, K. D., Lei, S. F., & Studies. Davis, R. I. (1993). Audiometric and histological differ- Holst, M., Greene, C. R., Jr., Richardson, W. J., McDonald, ences between the effects of continuous and impulsive T. L., Bay, K., Elliott, R. E., et al. (2005b). Marine noise exposures. Journal of the Acoustical Society of mammal and acoustical monitoring of missile launches America, 93, 2088-2095. on San Nicolas Island, California, August 2001-May Hamernik, R. P., Ahroon, W. A., Patterson, J. H., Jr., & Qiu, 2005 (LGL Report TA2665-5). Report from LGL W. (2002). Relations among early postexposure noise- Ltd., King City, Ontario, for Naval Air Warfare Center induced threshold shifts and permanent threshold shifts Weapons Division, Point Mugu, CA, and National in the chinchilla. Journal of the Acoustical Society of Marine Fisheries Service, Silver Spring, MD. 165 pp. America, 111, 320-326. Holt, M. M., Southall, B. L., Kastak, D., Reichmuth Kastak, Hamernik, R. P., Qiu, W., & Davis, B. (2003). The effects of C. J., & Schusterman, R. J. (2001). Aerial hearing sen- the amplitude distribution of equal energy exposures on sitivity in pinnipeds: A comparison of free-field and noise-induced hearing loss: The kurtosis metric. Journal headphone thresholds. 14th Biennial Conference on the of the Acoustical Society of America, 114, 386-395. Biology of Marine Mammals. Vancouver, BC, Canada. Harris, C. M. (1998). Handbook of acoustical measure- 102 pp. ments and noise control (3rd ed.). Huntington, NY: Holt, M. M., Schusterman, R. J., Southall, B. L., & Kastak, Acoustical Society of America. D. (2004). Localization of aerial broadband noise by Harris, R. E., Miller, G. W., & Richardson, W. J. (2001). pinnipeds. Journal of the Acoustical Society of America, Seal responses to airgun sounds during summer seismic 115, 2339-2345. surveys in the Alaskan Beaufort Sea. Marine Mammal Holt, M. M., Schusterman, R. J., Kastak, D., & Southall, Science, 17, 795-812. B. L. (2005). Localization of aerial pure tones by pinni- Hastie, G. D., Wilson, B., Tufft, L. H., & Thompson, peds. Journal of the Acoustical Society of America, 118, P. M. (2003). Bottlenose dolphins increase breathing 3921-3926. synchrony in response to boat traffic. Marine Mammal Houser, D. S., & Finneran, J. J. (2006a). A comparison of Science, 19, 74-84. underwater hearing sensitivity in bottlenose dolphins Heathershaw, A. D., Ward, P. D., & David, A. M. (2001). (Tursiops truncatus) determined by electrophysiologi- The environmental impact of underwater sound. In cal and behavioral methods. Journal of the Acoustical Second Symposium on Underwater Bio-sonar and Society of America, 120, 1713-1722. Bioacoustic Systems (pp. 1-12). St Albans, Hertfordshire, Houser, D. S., & Finneran, J. J. (2006b). Variation in the UK: Loughborough University, Institute of Acoustics. hearing sensitivity of a dolphin population obtained 202 pp. through the use of evoked potential audiometry. Journal Henderson, D., & Hamernik, R. P. (1986). Impulse noise: of the Acoustical Society of America, 120, 4090-4099. Critical review. Journal of the Acoustical Society of Houser, D. S., Helweg, D. A., & Moore, P. W. B. (2001a). A America, 80, 569-584. bandpass filter-bank model of auditory sensitivity in the Henson, O. W., Jr. (1965). The activity and function of the humpback whale. Aquatic Mammals, 27, 82-91. middle ear muscles in echolocating bats. Journal of Houser, D. S., Howard, R., & Ridgway, S. (2001b). Can Physiology, 180, 871-887. diving-induced tissue nitrogen supersaturation increase Hernandez, E. N., Kuczaj, S., Houser, D. S., & Finneran, the chance of acoustically driven bubble growth in J. J. (2007). Middle- and long-latency auditory evoked marine mammals? Journal of Theoretical Biology, 213, potentials in bottlenose dolphins (Tursiops truncatus) 183-195. resulting from frequent and oddball stimuli. Aquatic Houser, D. S., Crocker, D. E., Reichmuth, C., Mulsow, J., Mammals, 33, 34-42. & Finneran, J. J. (2007). Auditory evoked potentials Hewitt, R. P. (1985). Reaction of dolphins to a survey vessel: in northern elephant seals (Mirounga angustirostris). Effects on census data. Fishery Bulletin, 83, 187-193. Aquatic Mammals, 33, 110-121. High Energy Seismic Survey (HESS). (1999). High Energy Hoy, R. R. (1989). Startle, categorical response, and atten- Seismic Survey review process and interim operation in the acoustic behavior of insects. Annual Review tional guidelines for marine surveys offshore Southern of Neuroscience, 12, 355-375. California. Camarillo: Report from HESS Team for International Standardization Organization (ISO). (1990). California State Lands Commission and U.S. Minerals Acoustics – Determination of occupational noise 488 Southall et al.

exposure and estimation of noise-induced hearing Kastak, D., & Schusterman, R. J. (2002). Changes in audi- impairment (ISO 1990-01-15). Geneva, Switzerland. tory sensitivity with depth in a free-ranging California Irvine, A. B., Scott, M. D., Wells, R. S., & Kaufmann, J. H. sea lion (Zalophus californianus). Journal of the (1981). Movements and activities of the Atlantic bottle- Acoustical Society of America, 112, 329-333. nose dolphin, Tursiops truncatus, near Sarasota, Florida. Kastak, D., Schusterman, R. J., Southall, B. L., & Fishery Bulletin, 79, 671-688. Reichmuth, C. J. (1999). Underwater temporary thresh- Jacobs, D. W., & Hall, J. D. (1972). Auditory thresholds of a old shift induced by octave-band noise in three species of fresh water dolphin, Inia geoffrensis Blainville. Journal pinniped. Journal of the Acoustical Society of America, of the Acoustical Society of America, 51, 530-533. 106, 1142-1148. Jacobs, S. R., & Terhune, J. M. (2002). The effectiveness Kastak, D., Southall, B. L., Holt, M. M., Kastak, C. R., & of acoustic harassment devices in the Bay of Fundy, Schusterman, R. J. (2004a). Noise-induced temporary Canada: Seal reactions and a noise exposure model. threshold shift in pinnipeds: Effects of noise energy. Aquatic Mammals, 28, 147-158. Journal of the Acoustical Society of America, 116 (4, pt. Jahoda, M., Lafortuna, C. L., Biassoni, N., Almirante, C., 2), 2531. Azzellino, A., Panigada, S., et al. (2003). Mediterranean Kastak, D., Kastak, C. R., Holt, M. M., Schusterman, fin whale’s (Balaenoptera physalus) response to small R. J., & Southall, B. L. (2004b). Aerial hearing sensitiv- vessels and biopsy sampling assessed through passive ity in some pinnipeds is comparable to that of humans. tracking and timing of respiration. Marine Mammal Journal of the Acoustical Society of America, 115 (5, pt. Science, 19, 96-110. 2), 2406. Janik, V. M. (2000). Source levels and the estimated active Kastak, D., Southall, B. L., Schusterman, R. J., & Reichmuth space of bottlenose dolphin (Tursiops truncatus) whistles Kastak, C. (2005). Underwater temporary threshold shift in the Moray Firth, Scotland. Journal of Comparative in pinnipeds: Effects of noise level and duration. Journal Physiology A, 186, 673-680. of the Acoustical Society of America, 118, 3154-3163. Jepson, P. D., Arbelo, M., Deaville, R., Patterson, I. A. P., Kastak, D., Reichmuth, C., Holt, M. M., Mulsow, J., Castro, P., Baker, J. R., et al. (2003). Gas-bubble lesions Southall, B. L., & Schusterman, R. J. (2007). Onset, in stranded cetaceans. Nature, 425, 575-576. growth, and recovery of in-air temporary threshold Johnson, C. S. (1967). Sound detection thresholds in marine shift in a California sea lion (Zalophus californianus). mammals. In W. N. Tavolga (Ed.), Marine bioacoustics Journal of the Acoustical Society of America, 122, 2916- (pp. 247-260). New York: Pergamon. 2924. Johnson, C. S. (1992). Detection of tone glides by the Kastelein, R. A., de Haan, D., Goodson, A. D., Staal, C., beluga. In J. A. Thomas, R. A. Kastelein, & A. Ya. Supin & Vaughan, N. (1997). The effects of various sounds (Eds.), Marine mammal sensory systems (pp. 241-248). on harbor porpoise. In A. J. Read, P. R. Wiepkema, & New York: Plenum Press. P. E. Nachtigall (Eds.), The biology of the harbor por- Johnston, D. W. (2002). The effect of acoustic harassment poise (pp. 367-383). Woerden, The Netherlands: De Spil devices on harbor porpoises (Phocoena phocoena) in the Publishers. Bay of Fundy, Canada. Biological Conservation, 108, Kastelein, R. A., Rippe, H. T., Vaughan, N., Schooneman, 113-118. N. M., Verboom, W. C., & de Haan, D. (2000). The Johnston, D. W., & Woodley, T. H. (1998). A survey of effects of acoustic alarms on the behavior of harbor por- acoustic harassment device (AHD) use in the Bay of poises in a floating pen. Marine Mammal Science, 16, Fundy, NB, Canada. Aquatic Mammals, 24, 51-61. 46-64. Kastak, D., & Schusterman, R. J. (1995). Aerial and under- Kastelein, R. A., Rippe, H. T., Vaughan, N., Staal, C., & water hearing thresholds for 100 Hz pure tones in two Schooneman, N. M. (2001). The influence of three species of pinnipeds. In R. A. Kastelein, J. A. Thomas, acoustic alarms on the behavior of harbor porpoises & P. E. Nachtigall (Eds.), Sensory systems of aquatic (Phocoena phocoena) in a floating pen. Marine mammals (pp. 71-79). Woerden, The Netherlands: De Environmental Research, 52, 351-371. Spil Publishers. Kastelein, R. A., Bunskoek, P., Hagedoorn, M., Au, Kastak, D., & Schusterman, R. J. (1996). Temporary thresh- W. W. L., & de Haan, D. (2002a). Audiogram of a old shift in a harbor seal (Phoca vitulina). Journal of the harbor porpoise (Phocoena phocoena) measured with Acoustical Society of America, 100, 1905-1908. narrow-band frequency modulated signals. Journal of Kastak, D., & Schusterman, R. J. (1998). Low-frequency the Acoustical Society of America, 112, 334-344. amphibious hearing in pinnipeds: Methods, measure- Kastelein, R. A., Mosterd, P., van Santen, B., Hagedoorn, ments, noise, and ecology. Journal of the Acoustical M., & deHaan, D. (2002b). Underwater audiogram of Society of America, 103, 2216-2228. a Pacific walrus (Odobenus rosmarus divergens) mea- Kastak, D., & Schusterman, R. J. (1999). In-air and sured with narrow-band frequency-modulated signals. underwater hearing sensitivity of a northern elephant Journal of the Acoustical Society of America, 112, 2173- seal (Mirounga angustirostris). Canadian Journal of 2182. Zoology, 77, 1751-1758. Kastelein, R. A., Hagedoorn, M., Au, W. W. L., & de Haan, D. (2003). Audiogram of a striped dolphin (Stenella Marine Mammal Noise Exposure Criteria 489

coeruleoalba). Journal of the Acoustical Society of Norris (Eds.), Dolphin societies: Discoveries and puz- America, 113, 1130-1137. zles (pp. 149-159). Berkeley: University of California Kastelein, R. A., Verboom, W. C., Muijsers, M., Jennings, Press. N. V., & van der Heul, S. (2005). The influence of acous- Kraus, S., Read, A., Anderson, E., Baldwin, K., Solow, tic emissions for underwater data transmission on the A., Spradlin, T., et al. (1997). A field test of the use of behaviour of harbor porpoises (Phocoena phocoena) acoustic alarms to reduce incidental mortality of harbor in a floating pen. Marine Environmental Research, 59, porpoise in gill nets. Nature, 388, 525. 287-307. Kryter, K. D. (1970). Effects of noise on man. New York: Kastelein, R. A., Jennings, N. V., Verboom, W. C., de Haan, Academic Press. 633 pp. D., & Schooneman, N. M. (2006a). Differences in the Kryter, K. D. (1994). The handbook of hearing and the response of a striped dolphin (Stenella coeruleoalba) and effects of noise. New York: Academic Press. 673 pp. a harbor porpoise (Phocoena phocoena) to an acoustic Kryter, K. D., Ward, W. D., Miller, J. D., & Eldredge, D. H. alarm. Marine Environmental Research, 61, 363-378. (1966). Hazardous exposure to intermittent and steady- Kastelein, R. A., van der Heul, S., Verboom, W. C., state noise. Journal of the Acoustical Society of America, Triesscheijn, R. V. J., & Jennings, N. V. (2006b). The 39, 451-464. influence of underwater data transmission sounds on the Larom, D., Garstang, M., Payne, K., Raspet, R., & Lindeque, displacement behaviour of captive harbor seals (Phoca M. (1997). The influence of surface atmospheric condi- vitulina). Marine Environmental Research, 61, 19-39. tions on the range and area reached by animal vocaliza- Ketten, D. R. (1997). Structure and function in whale ears. tions. Journal of Experimental Biology, 200, 421-431. Bioacoustics, 8, 103-135. Lee, A. S. (1992). Mammalian stress response: Induction of Ketten, D. R. (2000). Cetacean ears. In W. W. L. Au, the glucose-regulated protein family. Current Opinions R. R. Fay, & A. N. Popper (Eds.), Hearing by whales in Cell Biology, 4, 267-273. and dolphins (SHAR Series for Auditory Research) (pp. Lei, S. F., Ahroon, W. A., & Hamernik, R. P. (1994). The 43-108). New York: Springer-Verlag. application of frequency and time domain kurtosis to the Ketten, D. R., Arruda, J, Cramer, S., Yamato, M., Zosuls, assessment of hazardous noise exposures. Journal of the M., Mountain, D., et al. (2007). How low can they go: Acoustical Society of America, 96, 1435-1444. Functional analysis of the largest land and marine Lesage, V., Barrette, C., Kingsley, M. C. S., & Sjare, B. mammal ears. [Presentation abstract.] 17th Biennial (1999). The effects of vessel noise on the vocal behavior Conference on the Biology of Marine Mammals, Cape of belugas in the St. Lawrence River Estuary, Canada. Town, South Africa. Marine Mammal Science, 15, 65-84. Kinsler, L. E., Frey, A. R., Coppens, A. B., & Sanders, J. V. LGL Ltd. & Greeneridge Sciences. (1986). Reactions (1982). Fundamentals of acoustics (3rd ed.). New York: of beluga whales and narwhals to ship traffic and ice- John Wiley & Sons. 480 pp. breaking along ice edges in the eastern Canadian High Kjellberg, A., Tesarz, M., Holmberg, K., & Landström, U. Arctic: 1982-1984. In Environmental studies (No. 37). (1997). Evaluation of frequency-weighted sound level Ottawa, ON, Canada: Indian and Northern Affairs measurements for prediction of low-frequency noise Canada. 301 pp. annoyance. Environment International, 23(4), 519-527. Ljungblad, D. K., Scoggins, P. D., & Gilmartin, W. G. Klishin, V. O., Popov, V. V., & Supin, A. Ya. (2000). Hearing (1982). Auditory thresholds of a captive Eastern Pacific capabilities of a beluga whale, Delphinapterus leucas. bottle-nosed dolphin, Tursiops spp. Journal of the Aquatic Mammals, 26, 212-228. Acoustical Society of America, 72, 1726-1729. Koschinski, S., & Culik, B. (1997). Deterring harbor por- Ljungblad, D. K., Würsig, B., Swartz, S. L., & Keene, poises (Phocoena phocoena) from gillnets: Observed J. M. (1988). Observations on the behavioral responses reactions to passive reflectors and pingers. Report of the of bowhead whales (Balaena mysticetus) to active geo- International Whaling Commission, 47, 659-668. physical vessels in the Alaskan Beaufort Sea. Arctic, 41, Koschinski, S., Culik, B. M., Henriksen, O. D., Tregenza, 183-194. N., Ellis, G., Jansen, C., et al. (2003). Behavioral reac- Lorenz, K. (1939). Vergleichende Verhaltensforschung. tions of free-ranging porpoises and seals to the noise of a Verhandlungen der Deutschen Zoologischen simulated 2 MW windpower generator. Marine Ecology Gesellschaft, 1939, 69-102. Progress Series, 265, 263-273. Lucke, K., Lepper, P. A., Blanchet, M.-A., & Siebert, U. Koski, W. R., & Johnson, S. R. (1987). Behavioral studies (2007a). Testing the auditory tolerance of harbour and aerial photogrammetry. In Responses of bowhead porpoise hearing for impulsive sounds (Presentation whales to an offshore drilling operation in the Alaskan abstract). Effects of Noise on Marine Life Conference, Beaufort Sea, autumn 1986 (Section 4). Report from Nyborg, Denmark. LGL Ltd., King City, ON, and Greeneridge Sciences Lucke, K., Lepper, P. A., Hoeve, B., Everaarts, E., van Elk, Inc., Santa Barbara, CA, for Shell Western Exploration N., & Siebert, U. (2007b). Perception of low-frequency and Production Inc., Anchorage, AK. 371 pp. acoustic signals by a harbour porpoise (Phocoena Kraus, S. (1991). The interactions between killer whales phocoena) in the presence of simulated offshore wind and boats in Johnstone Strait, B.C. In K. Pryor & K. S. turbine noise. Aquatic Mammals, 33, 55-68. 490 Southall et al.

Lusseau, D. (2003). Effects of tour boats on the behavior Malme, C. I., Würsig, B., Bird, J. E., & Tyack, P. L. (1988). of bottlenose dolphins: Using Markov chains to model Observations of feeding gray whale responses to con- anthropogenic impacts. Conservation Biology, 17, trolled industrial noise exposure. In W. M. Sackinger, 1785-1793. M. O. Jeffries, J. L. Imm, & S. D. Treacy (Eds.), MacDougall-Shackleton, S. A., Hulse, S. H., Gentner, Port and ocean engineering under Arctic conditions, T. Q., & White, W. (1998). Auditory scene analysis Volume II (pp. 55-73). Fairbanks: University of Alaska, by European starlings: Perceptual segregation of tone Geophysical Institute. sequences. Journal of the Acoustical Society of America, Malme, C. I., Miles, P. R., Miller, G. W., Richardson, 103, 3581-3587. W. J., Roseneau, D. G., Thomson, D. H., et al. (1989). Madsen, P. T. (2005). Marine mammals and noise: Problems Analysis and ranking of the acoustic disturbance poten- with root-mean-square sound pressure level for tran- tial of petroleum industry activities and other sources of sients. Journal of the Acoustical Society of America, noise in the environment of marine mammals in Alaska 117, 3952-3957. (BBN Report No. 6945; OCS Study MMS 89-0006; Madsen, P. T., & Møhl, B. (2000). Sperm whales (Physeter NTIS PB90-188673). Report from BBN Systems and catodon L. 1758) do not react to sounds from detona- Technologies Corp. for U.S. Minerals Management tors. Journal of the Acoustical Society of America, 107, Services, Anchorage, AK. Var. pag. 668-671. Mansfield, A. W. (1983). The effects of vessel traffic in the Madsen, P. T., Møhl, B., Nielsen, B. K., & Wahlberg, M. Arctic on marine mammals and recommendations for (2002). Male sperm whale behavior during exposures future research. Canadian Technical Report of Fisheries to distant seismic survey pulses. Aquatic Mammals, 28, and Aquatic Sciences, 1186. 97 pp. 231-240. Mate, B. R., & Harvey, J. T. (Eds.). (1987). Acoustical Madsen, P. T., Johnson, M., Aguilar de Soto, N., Zimmer, deterrents in marine mammal conflicts with fisher- W. M. X., & Tyack, P. (2005a). Biosonar performance ies (ORESU-W-86-001). Corvallis: Oregon State in foraging beaked whales (Mesoplodon densirostris). University, Sea Grant College Program. 116 pp. Journal of Experimental Biology, 208, 181-194. Matrosova, V. A., Volodin, I. A., Volodina, E. V., & Babitsky, Madsen, P. T., Carder, D. A., Bedholm, K., & Ridgway, A. F. (2007). Pups crying bass: Vocal adaptation for S. H. (2005b). Porpoise clicks from a sperm whale avoidance of age-dependent predation risk in ground nose—Convergent evolution of 130 kHz pulses in squirrels? Behavioral Ecology and Sociobiology, 62, toothed whale sonars? Bioacoustics, 15, 195-206. 181-191. Malme, C. I., Miles, P. R., Clark, C. W., Tyack, P., & Bird, McCauley, R. D., Cato, D. H., & Jeffery, A. F. (1996). A J. E. (1983). Investigations of the potential effects of study of the impacts of vessel noise on humpback whales underwater noise from petroleum industry activities on in Hervey Bay. Queensland, Australia: Report for the migrating gray whale behavior (BBN Report No. 5366; Queensland Department of Environment and Heritage, NTIS PB86-174174). Report from Bolt Beranek and Maryborough Office, from the Department of Marine Newman Inc. for U.S. Minerals Management Service, Biology, James Cook University, Townsville. 137 pp. Anchorage, AK. McCauley, R. D., Jenner, M-N., Jenner, C., McCabe, Malme, C. I., Miles, P. R., Clark, C. W., Tyack, P., & Bird, K. A., & Murdoch, J. (1998). The response of humpback J. E. (1984). Investigations of the potential effects of whales (Megaptera novaeangliae) to offshore seismic underwater noise from petroleum industry activities on survey noise: Preliminary results of observations about migrating gray whale behavior. Phase II: January 1984 a working seismic vessel and experimental exposures. migration (BBN Report No. 5586; NTIS PB86-218377). Australian Petroleum Production and Exploration Report from Bolt Beranek and Newman Inc. for U.S. Association Journal, 38, 692-707. Minerals Management Service, Anchorage, AK. McCauley, R. D., Fewtrell, J., Duncan, A. J., Jenner, Malme, C. I., Miles, P. R., Tyack, P. L., Clark, C. W., & C., Jenner, M.-N., Penrose, J. D., et al. (2000). Marine Bird, J. E. (1985). Investigations of the potential effects seismic surveys: A study of environmental implica- of underwater noise from petroleum industry activities tions. Australian Petroleum Production and Exploration on feeding humpback whale behavior (Bolt Beranek Association Journal, 40, 692-708. & Newman Report No. 5851, NTIS PB86-218385). McMillan, N. A., & Creelman, C. D. (1991). Detection Submitted to Minerals Management Service, Anchorage, theory: A user’s guide. Cambridge, UK: Cambridge AK. University Press. Malme, C. I., Würsig, B., Bird, J. E., & Tyack, P. L. (1986). Miles, P. R., & Malme, C. I. (1983). The acoustic environ- Behavioral responses of gray whales to industrial ment and noise exposure of humpback whales in Glacier noise: Feeding observations and predictive modeling Bay, Alaska (BBN Technical Memorandum 734). Report (BBN Report No. 6265, OCS Study MMS 88-0048; from Bolt Beranek & Newman Inc. for National Marine NTIS PB88-249008). NOAA Outer Continental Shelf Mammal Laboratory, Seattle, WA. 81 pp. Environmental Assessment Program, Final Reports of Miller, G. W., Moulton, V. D., Davis, R. A., Holst, M., Principal Investigators, 56, 393-600. Millman, P., MacGillivray, A., et al. (2005). Monitoring seismic effects on marine mammals – southeastern Marine Mammal Noise Exposure Criteria 491

Beaufort Sea, 2001-2002. In S. L. Armsworthy, P. J. in British Columbia, Canada. ICES Journal of Marine Cranford, & K. Lee (Eds.), Offshore oil and gas environ- Science, 59, 71-80. mental effects monitoring: Approaches and technologies Moss, C. F., & Surlykke, A. (2001). Auditory scene analy- (pp. 511-542). Columbus, OH: Battelle Press. sis by echolocation in bats. Journal of the Acoustical Miller, J. D., Rothenberg, S. J., & Eldredge, D. H. (1971). Society of America, 110, 2207-2226. Preliminary observations on the effects of exposure to Moulton, V. D., Richardson, W. J., McDonald, T. L., Elliott, noise for seven days on the hearing and inner ear of the R. E., & Williams, M. T. (2002). Factors influencing chinchilla. Journal of the Acoustical Society of America, local abundance and haulout behavior of ringed seals 50, 1199-1203. (Phoca hispida) on landfast ice of the Alaskan Beaufort Miller, P. J. O. (2002). Mixed-directionality of killer Sea. Canadian Journal of Zoology, 80, 1900-1917. whale stereotyped calls: A direction of movement cue? Moulton, V. D., Richardson, W. J., Williams, M. T., & Behavioral & Ecology Sociobiology, 52, 262-270. Blackwell, S. B. (2003). Ringed seal densities and noise Miller, P. J. O., Biassoni, N., Samuels, A., & Tyack, P. L. near an icebound artificial island with construction and (2000). Whale songs lengthen in response to sonar. drilling. Acoustics Research Letters Online, 4, 112-117. Nature, 405, 903. Moulton, V. D., Richardson, W. J., Elliott, R. E., McDonald, Mobley, J. R. (2005). Assessing responses of humpback T. L., Nations, C., & Williams, M. T. (2005). Effects of whales to North Pacific Acoustic Laboratory (NPAL) an offshore oil development on local abundance and dis- transmissions: Results of 2001-2003 aerial surveys north tribution of ringed seals (Phoca hispida) of the Alaskan of Kauai. Journal of the Acoustical Society of America, Beaufort Sea. Marine Mammal Science, 21, 217-242. 117, 1666-1673. Mulsow, J., & Reichmuth, C. (2007). Electrophysiological Mobley, J. R., Herman, L. M., & Frankel, A. S. (1988). assessment of temporal resolution in pinnipeds. Aquatic Responses of wintering humpback whales (Megaptera Mammals, 33, 122-131. novaeangliae) to playback of recordings of winter Nachtigall, P. E., Au, W. W. L., Pawloski, J. L., & Moore, and summer vocalizations and of synthetic sounds. P. W. B. (1995). Risso’s dolphin (Grampus griseus) Behavioral Ecology and Sociobiology, 23, 211-223. hearing thresholds in Kaneohe Bay, Hawaii. In R. A. Møhl, B. (1964). Preliminary studies on hearing in seals. Kastelein, J. A. Thomas, & P. E. Nachtigall (Eds.), Videnskabelige Meddelelser fra Dansk Naturhistorisk Sensory systems of aquatic mammals (pp. 49-53). Forening i Kjobenhavn, 127, 283-294. Woerden, The Netherlands: De Spil Publishers. Møhl, B. (1967). Frequency discrimination in the common Nachtigall, P. E., Lemonds, D. W., & Roitblat, H. L. (2000). seal and a discussion of the concept of upper hearing Psychoacoustic studies of dolphin and whale hearing. In limit. In V. M. Albers (Ed.), Underwater acoustics, W. W. L. Au, A. N. Popper, & R. R. Fay (Eds.), Hearing Volume 2 (pp. 43-54). New York: Plenum Press. by whales and dolphins (pp. 330-363). New York: Møhl, B. (1968). Auditory sensitivity of the common seal in Springer. air and water. Journal of Auditory Research, 8, 27-38. Nachtigall, P. E., Pawloski, J. L., & Au, W. W. L. (2003). Monteiro-Neto, C., Ávila, F. J. C., Alves-Jr., T. T., Araújo, Temporary threshold shifts and recovery following noise D. S., Campos, A. A., Martins, A. M. A., et al. (2004). exposure in the Atlantic bottlenosed dolphin (Tursiops Behavioral responses of Sotalia fluviatilis (Cetacea, truncatus). Journal of the Acoustical Society of America, Delphinidae) to acoustic pingers, Fortaleza, Brazil. 113, 3425-3429. Marine Mammal Science, 20, 141-151. Nachtigall, P. E., Supin, A. Ya., Pawloski, J. L., & Au, Moore, M. M., & Early, G. A. (2004). Cumulative sperm W. W. L. (2004). Temporary threshold shifts after noise whale bone damage and the bends. Science, 306, 2215. exposure in the bottlenose dolphin (Tursiops truncatus) Moore, P. W. B., & Au, W. W. L. (1975). Underwater local- measured using auditory evoked potentials. Marine ization of pulsed pure tones by the California sea lion Mammal Science, 20, 673-687. (Zalophus californianus). Journal of the Acoustical Nachtigall, P. E., Yuen, M. M. L., Mooney, T. A., & Taylor, Society of America, 58, 721-727. K. A. (2005). Hearing measurements from a stranded Moore, P. W. B., & Schusterman, R. J. (1987). Audiometric infant Risso’s dolphin, Grampus griseus. Journal of assessment of northern fur seals (Callorhinus ursinus). Experimental Biology, 108, 4181-4188. Marine Mammal Science, 3, 31-53. Nachtigall, P. E., Mooney, T. A., Taylor, K. A., & Yuen, Moore, S. E., & Clarke, J. T. (2002). Potential impact of M. M. L. (2007). Hearing and auditory evoked poten- offshore human activities on gray whales (Eschrichtius tial methods applied to odontocete cetaceans. Aquatic robustus). Journal of Cetacean Research and Mammals, 33, 6-13. Management, 4, 19-25. National Institute for Occupational Safety and Health Morisaka, T., Shinohara, M., Nakahara, F., & Akamatsu, (NIOSH). (1998). Criteria for a recommended stan- T. (2005). Effects of ambient noise on the whistles of dard: Occupational noise exposure (Publication #98- Indo-Pacific bottlenose dolphin populations. Journal of 126). Washington, DC: U.S. Department of Health and Mammalogy, 86, 541-546. Human Services, NIOSH. Morton, A. B., & Symonds, H. K. (2002). Displacement National Marine Fisheries Service (NMFS). (1995). of Orcinus orca (Linnaeus) by high amplitude sound Small takes of marine mammals incidental to specified 492 Southall et al.

activities; offshore seismic activities in southern Norris, K. S. (1968). The evolution of acoustic mechanisms California: Notice of issuance of an incidental harass- in odontocete cetaceans. In E. T. Drake (Ed.), Evolution ment authorization. Federal Register, 60(200, 17 and environment (pp. 297-324). New Haven, CT: Yale October), 53753-53760. University Press. NMFS. (2005). Assessment of acoustic exposures on marine Norris, T. F. (1994). Effects of boat noise on the acoustic mammals in conjunction with USS Shoup active sonar behavior of humpback whales. Journal of the Acoustical transmissions in Haro Strait, Washington, 5 May 2003 Society of America, 96, 3251. (NMFS Office of Protected Resources report). Nowacek, D. P., Johnson, M. P., & Tyack, P. L. (2004). National Research Council (NRC). (1994). Low-frequency North Atlantic right whales (Eubalaena glacialis) ignore sound and marine mammals: Current knowledge ships but respond to alerting stimuli. Proceedings of the and research needs. Washington, DC: The National Royal Society of London Series B: Biological Sciences, Academies Press. 75 pp. 271, 227-231. NRC. (2000). Marine mammals and low-frequency sound. Nowacek, D. P., Thorne, L. H., Johnston, D. W., & Tyack, Washington, DC: The National Academies Press. P. L. (2007). Responses of cetaceans to anthropogenic 146 pp. noise. Mammal Review, 37, 81-115. NRC. (2003). Ocean noise and marine mammals. Nowacek, S. M., Wells, R. S., & Solow, A. R. (2001). Washington, DC: The National Academies Press. Short-term effects of boat traffic on bottlenose dolphins, 192 pp. Tursiops truncatus, in Sarasota Bay, Florida. Marine NRC. (2005). Marine mammal populations and ocean Mammal Science, 17, 673-688. noise: Determining when noise causes biologically Olesiuk, P. F., Nichol, L. M., Sowden, M. J., & Ford, J. K. B. significant effects. Washington, DC: The National (1996). Effect of sounds generated by an acoustic deter- Academies Press. 126 pp. rent device on abundance and distribution of harbor Naval Research Laboratory (NRL). (2004a). Acoustic mod- porpoise (Phocoena phocoena) in Retreat Passage, eling results of the Haro Strait for 5 May 2003 (Naval British Columbia (Draft report for the Department of Research Laboratory Report). Arlington, VA: Office of Fisheries and Oceans). Nanaimo, BC: Pacific Biological Naval Research. Station. 47 pp. NRL. (2004b). EEEL analysis of Shoup transmissions in the Olesiuk, P. F., Nichol, L. M., Sowden, M. J., & Ford, Haro Strait on 5 May 2003 (Naval Research Laboratory J. K. B. (2002). Effect of the sound generated by an briefing of 2 September 2004). acoustic harassment device on the relative abundance Nedwell, J. R., Turnpenny, A. W. H., Lovell, J., Parvin, and distribution of harbor porpoises, Phocoena pho- S. J., Workman, R., Spinks, J. A. L., et al. (2007). coena, in Retreat Passage, British Columbia. Marine

A validation of the dBht as a measure of the behav- Mammal Science, 18, 843-862. ioural and auditory effects of underwater noise Oosterveld, W., Polman, A., & Schoonheyt, J. (1982). (Subacoustech Report 534R1231). Report to Chevron Vestibular implications of noise-induced hearing loss. Ltd, TotalFinalElf Exploration UK PLC, Department British Journal of Audition, 16, 227-232. of Business, Enterprise, and Regulatory Reform, Shell Palka, D., & Hammond, P. S. (2001). Accounting for UK, ITF, JNCC. Retrieved 4 January 2008 from www. responsive movement in line transect estimates of subacoustech.com/information/downloads/reports/ abundance. Canadian Journal of Fisheries and Aquatic 534R1231.pdf. Sciences, 58, 777-787. Nielsen, D. W., Bauman, M. J., & Brandt, D. K. (1986). Parks, S. E., Clark, C. W., & Tyack, P. L. (2007). Short- Changes in auditory threshold during and after long- and long-term changes in right whale calling behavior: duration noise exposure: Species differences. In The potential effects of noise on acoustic communica- R. J. Salvi, D. Henderson, R. P. Hamernik, & V. Colletti tion. Journal of the Acoustical Society of America, 122, (Eds.), Basic and applied aspects of noise-induced hear- 3725-3731. ing loss (pp. 281-293). New York: Plenum Press. Payne, R., & Webb, D. (1971). Orientation by means of Norberg, B. (2000). Looking at the effects of acoustic deter- long range acoustic signaling in baleen whales. Annals rent devices on California sea lion predation patterns of the New York Academy of Science, 188, 110-141. at a commercial salmon farm. Seattle, WA: National Perry, E. A., Boness, D. J., & Insley, S. J. (2002). Effects of Marine Fisheries Service. 17 pp. sonic booms on breeding gray seals and harbor seals on Norberg, B., & Bain, D. (1994). Implementation and assess- Sable Island, Canada. Journal of the Acoustical Society ment of the acoustic barrier at the Hiram M. Chittenden of America, 111, 599-609. Locks using calibrated measurements of the sound field. Piantadosi, C. A., & Thalmann, E. D. (2004). Pathology: Seattle, WA: National Marine Fisheries Service. 67 pp. Whales, sonar and decompression sickness [Brief Nordmann, A. S., Bohne, B. A., & Harding, G. W. (2000). Communications]. Nature, 428(6984), U1. Histopathological differences between temporary and Popov, V. V., & Klishin, V. O. (1998). EEG study of hear- permanent threshold shift. Hearing Research, 139, 31- ing in the common dolphin Delphinus delphis. Aquatic 41. Mammals, 24, 13-21. Marine Mammal Noise Exposure Criteria 493

Popov, V. V., & Supin, A. Ya. (1990). Electrophysiological in the eastern Beaufort Sea, 1980-84 (OCS Study MMS studies of hearing in some cetaceans and a manatee. In 85-0034; NTIS PB87-124376). Report from LGL J. A. Thomas & R. A. Kastelein (Eds.), Sensory abilities Ecological Research Associates, Inc., for U.S. Minerals of cetaceans: Laboratory and field evidence (pp. 405- Management Service, Reston, VA. 306 pp. 415). New York: Plenum Press. 710 pp. Richardson, W. J., & Malme, C. I. (1993). Man-made noise Popov, V. V., Ladygina, T. F., & Supin, A. Ya. (1986). and behavioral responses. In J. J. Burns, J. J. Montague, Evoked potentials of the auditory cortex of the por- & C. J. Cowles (Eds.), The bowhead whale (Special poise, Phocoena phocoena. Journal of Comparative Publication 2) (pp. 631-700). Lawrence, KS: Society for Physiology A, 158, 705-711. Marine Mammalogy. 787 pp. Popov, V. V., Supin, A. Ya., Wang, D., & Wang, K. (2006). Richardson, W. J., Fraker, M. A., Würsig, B., & Wells, Nonconstant quality of auditory filters in the porpoises, R. S. (1985). Behaviour of bowhead whales, Balaena Phocoena phocoena and Neophocaena phocaenoides mysticetus, summering in the Beaufort Sea: Reactions to (Cetacea, Phocoenidae). Journal of the Acoustical industrial activities. Biological Conservation, 32, 195- Society of America, 119, 3173-3180. 230. Popov, V. V., Supin, A. Ya., Pletenko, M. G., Tarakanov, Richardson, W. J., Würsig, B., & Greene, C. R., Jr. (1986). M. B., Klishin, V. O., Bulgakova, T. N., et al. (2007). Reactions of bowhead whales, Balaena mysticetus, Audiogram variability in normal bottlenose dolphins to seismic exploration in the Canadian Beaufort Sea. (Tursiops truncatus). Aquatic Mammals, 33, 24-33. Journal of the Acoustical Society of America, 79, 1117- Popper, A. N., Fay, R. R., Platt, C., & Sand, O. (2003). 1128. Sound detection mechanisms and capabilities of tele- Richardson, W. J., Würsig, B., & Greene, C. R., Jr. (1990a). ost fishes. In S. P. Collin & N. J. Marshall (Eds.), Reactions of bowhead whales, Balaena mysticetus, to Sensory processing in aquatic environments (pp. 3-38). drilling and dredging noise in the Canadian Beaufort New York: Springer-Verlag. Sea. Marine Environmental Research, 29, 135-160. Potter, J. R. (2004). A possible mechanism for acous- Richardson, W. J., Greene, C. R., Jr., Koski, W. R., Malme, tic triggering of decompression sickness symptoms in C. I., Miller, G. W., Smultea, M. A., et al. (1990b). deep-diving marine mammals. Paper presented to the Acoustic effects of oil production activities on bowhead 2004 IEEE International Symposium on Underwater and white whales visible during spring migration near Technology, Taipei, Taiwan. Pt. Barrow, Alaska—1989 phase (OCS Study MMS 90- Pytte, C. L., Ficken, M. S., & Moiseff, A. (2004). Ultrasonic 0017; NTIS PB91-105486). LGL Ltd. report for U.S. singing by the blue-throated hummingbird: A com- Minerals Management Service, Herndon, VA. 284 pp. parison between production and perception. Journal of Richardson, W. J., Greene, C. R., Jr., Koski, W. R., Smultea, Comparative Physiology A, 190, 665-673. M. A., Cameron, G., Holdsworth, C., et al. (1991). Reeves, R. R., Ljungblad, D. K., & Clarke, J. T. (1984). Acoustic effects of oil production activities on bowhead Bowhead whales and acoustic seismic surveys in the and white whales visible during spring migration near Beaufort Sea. Polar Record, 22, 270-280. Pt. Barrow, Alaska—1990 phase (OCS Study MMS 91- Reichmuth, C., Mulsow, J., Finneran, J. J., Houser, D. S., & 0037; NTIS PB92-170430). LGL Ltd. report for U.S. Supin, A. Ya. (2007). Measurement and response char- Minerals Management Service, Herndon, VA. 311 pp. acteristics of auditory brainstem responses in pinnipeds. Richardson, W. J., Greene, C. R., Jr., Malme, C. I., & Aquatic Mammals, 33, 132-150. Thomson, D. H. (1995). Marine mammals and noise. Renaud, D. L., & Popper, A. N. (1975). Sound localization New York: Academic Press. 576 pp. by the bottlenose porpoise, Tursiops truncatus. Journal Richardson, W. J., Miller, G. W., & Greene, C. R., Jr. of Experimental Biology, 63, 569-585. (1999). Displacement of migrating bowhead whales Rendell, L. E., & Gordon, J. C. D. (1999). Vocal response of by sounds from seismic surveys in shallow waters of long-finned pilot whales (Globicephala melas) to mili- the Beaufort Sea. Journal of the Acoustical Society of tary sonar in the Ligurian Sea. Marine Mammal Science, America, 106, 2281. 15, 198-204. Ridgway, S. H., & Carder, D. A. (1997). Hearing deficits Repenning, C. A. (1972). Underwater hearing in seals: measured in some Tursiops truncatus and the discovery Functional morphology. In R. J. Harrison (Ed.), of a deaf/mute dolphin. Journal of the Acoustical Society Functional anatomy of marine mammals, Volume 1 (pp. of America, 101, 590-594. 307-331). New York: Academic Press. Ridgway, S. H., & Carder, D. A. (2000). A preliminary Reynolds III, J. E., & Rommel, S. A. (Eds.). (1999). Biology study of loudness at frequencies of 5 to 120 kHz based of marine mammals. Washington, DC: Smithsonian on whistle response time (RT) in a dolphin. Journal of Institution Press. the Acoustical Society of America, 108, 2515. Rice, D. W. (1998). Marine mammals of the world: Ridgway, S. H., & Carder, D. A. (2001). Assessing hear- Systematics and distribution (Special Publication 4). ing and sound production in cetaceans not available for Lawrence, KS: Society for Marine Mammalogy. 231 pp. behavioral audiograms: Experiences with sperm, pygmy Richardson, W. J. (1985). Behavior, disturbance responses sperm, and gray whales. Aquatic Mammals, 27, 267- and distribution of bowhead whales (Balaena mysticetus) 276. 494 Southall et al.

Ridgway, S. H., & Howard, R. (1979). Dolphin lung col- in a bottlenose dolphin (Tursiops truncatus). Journal of lapse and intramuscular circulation during free diving: the Acoustical Society of America, 122, 615-622. Evidence from nitrogen washout. Science, 206, 1182- Schultz, T. (1978). Synthesis of social surveys on noise 1183. annoyance. Journal of the Acoustical Society of America, Ridgway, S. H., & Joyce, P. L. (1975). Studies on seal brain 64, 377-405. by radiotelemetry. Rapports et Procès-Verbeaux des Schusterman, R. J. (1974). Auditory sensitivity of a Réunions, Conseil International pour l’Exploration de California sea lion to airborne sound. Journal of the la Mer, 169, 81-91. Acoustical Society of America, 56, 1248-1251. Ridgway, S. H., Carder, D. A., Smith, R. R., Kamolnick, T., Schusterman, R. J. (1981). Behavioral capabilities of seals Schlundt, C. E., & Elsberry, W. R. (1997). Behavioral and sea lions: A review of their hearing, visual, learning, responses and temporary threshold shift in masked hear- and diving skills. Psychological Record, 31, 125-143. ing thresholds of bottlenose dolphins, Tursiops trunca- Schusterman, R. J., Balliet, R. F., & Nixon, J. (1972). tus, to 1-second tones of 141-201 dB re: 1mPa (Technical Underwater audiogram of the California sea lion by Report #1751). San Diego: Naval Command, Control, the conditioned vocalization technique. Journal of the and Ocean Surveillance Center, RDT&E Division. 27 Experimental Analysis of Behavior, 17, 339-350. pp. Schusterman, R. J., Barrett, R., & Moore, P. W. B. (1975). Roche, A. F., Siervogel, R. M., Himes, J. H., & Johnson, Detection of underwater signals by a California sea D. L. (1978). Longitudinal study of hearing in children: lion and a bottlenose porpoise: Variation in the payoff Baseline data concerning auditory thresholds, noise matrix. Journal of the Acoustical Society of America, 57, exposure, and biological factors. Journal of the 1526-1532. Acoustical Society of America, 64, 1593-1601. Schusterman, R. J., Kastak, D., Levenson, D. H., Romano, T. A., Keogh, M. J., Kelly, C., Feng, P., Berk, Reichmuth, C. J., & Southall, B. L. (2000). Why pinni- L., Schlundt, C. E., et al. (2004). Anthropogenic sound peds don’t echolocate. Journal of the Acoustical Society and marine mammal health: Measures of the nervous of America, 107, 2256-2264. and immune systems before and after intense sound Schusterman, R. J., Southall, B. L., Kastak, D., & Reichmuth exposure. Canadian Journal of Fisheries and Aquatic Kastak, C. (2002). Age-related hearing loss in sea lions Sciences, 61, 1124-1134. and their scientists. Journal of the Acoustical Society of Sapolsky, R. M., Romero, L. M., & Munck, A. U. (2000). America, 111, 2342-2343. How do glucocorticoids influence stress responses? Schusterman, R. J., Kastak, D., Southall, B. L., Reichmuth, Integrating permissive, suppressive, stimulatory, and C. J., & Holt, M. M. (2003). Noise-induced tempo- preparative actions. Endocrine Reviews, 21, 55-89. rary threshold shift in pinnipeds: Effects of exposure Saunders, J. C., Cohen, Y. E., & Szymko, Y. M. (1991). The medium, intermittence, duration and intensity. Abstract structural and functional consequences of acoustic injury & Talk at Symposium on Environmental Consequences in the cochlea and peripheral auditory systems. Journal of Underwater Sound (ECOUS). San Antonio, TX. of the Acoustical Society of America, 90, 136-146. Seeley, R. L., Flanigan, W. F., Jr., & Ridgway, S. H. (1976). Scheifele, P. M., Andrews, S., Cooper, R. A., Darre, A technique for rapidly assessing the hearing of the M., Musick, F. E., & Max, L. (2005). Indication of a bottlenosed porpoise, Tursiops truncatus (NUC TP 522; Lombard vocal response in the St. Lawrence River NTIS AD-A 029178). San Diego, CA: Naval Undersea beluga. Journal of the Acoustical Society of America, Center. 15 pp. 117, 1486-1492. Serrano, A., & Terhune, J. M. (2002). Anti-masking aspects Schick, R. S., & Urban, D. L. (2000). Spatial components of of harp seal (Pagophilus groenlandicus) underwater bowhead whale (Balaena mysticetus) distribution in the vocalizations. Canadian Journal of Zoology, 79, 1410- Alaskan Beaufort Sea. Canadian Journal of Fisheries 1413. and Aquatic Sciences, 57, 2193-2200. Shaughnessy, P. D., Semmelink, A., Cooper, J., & Frost, P. Schlundt, C. E., Finneran, J. J., Carder, D. A., & Ridgway, G. H. (1981). Attempts to develop acoustic methods of S. H. (2000). Temporary shift in masked hearing thresh- keeping Cape fur seals (Arctocephalus pusillus) from olds (MTTS) of bottlenose dolphins and white whales fishing nets. Biological Conservation, 21, 141-158. after exposure to intense tones. Journal of the Acoustical Smith, M. E., Kane, A. S., & Popper, A. N. (2004). Society of America, 107, 3496-3508. Acoustical stress and hearing sensitivity in fishes: Does Schlundt, C. E., Dear, R. L., Carder, D. A., & Finneran, the linear threshold shift hypothesis hold water? Journal J. J. (2006). Growth and recovery of temporary threshold of Experimental Biology, 207, 3591-3602. shifts in a dolphin exposed to midfrequency tones with Southall, B. L., Schusterman, R. J., & Kastak, D. (2000). durations up to 128 s. Journal of the Acoustical Society Masking in three pinnipeds: Underwater, low-frequency of America, 120, 3227. critical ratios. Journal of the Acoustical Society of Schlundt, C. E., Dear, R. L., Green, L., Houser, D. S., America, 108, 1322-1326. & Finneran, J. J. (2007). Simultaneously measured Southall, B. L., Schusterman, R. J., Kastak, D., & Kastak, behavioral and electrophysiological hearing thresholds C. R. (2001). Pinniped hearing and anthropogenic noise. Marine Mammal Noise Exposure Criteria 495

Journal of the Acoustical Society of America, 110 (5, pt. Proceedings of the Institute of Acoustics, 19(9), 267- 2), 2722. 275. Southall, B. L., Schusterman, R. J., & Kastak, D. (2003). Terhune, J. M. (1974). Directional hearing of a harbor seal Auditory masking in three pinnipeds: Aerial criti- in air and water. Journal of the Acoustical Society of cal ratios and direct critical bandwidth measurements. America, 56, 1862-1865. Journal of the Acoustical Society of America, 114, 1660- Terhune, J. M., & Ronald, K. (1971). The harp seal, 1666. Pagophilus groenlandicus (Erxleben, 1777). X. The air Southall, B. L., Schusterman, R. J., Kastak, D., & Kastak, audiogram. Canadian Journal of Zoology, 49, 385-390. C. R. (2004). Underwater hearing thresholds in pin- Terhune, J. M., & Ronald, K. (1972). The harp seal, nipeds measured over a 6-year period. Journal of the Pagophilus groenlandicus (Erxleben, 1777). III. The Acoustical Society of America, 116 (4, pt. 2), 2504. underwater audiogram. Canadian Journal of Zoology, Spangler, H. G. (1988). Moth hearing, defense, and com- 50, 565-569. munication. Annual Review of Entomology, 33, 59-81. Terhune, J. M., & Ronald, K. (1975). Underwater hearing Steevens, C. C., Sylvester, R., & Clark, J. (1997). Effects of sensitivity of two ringed seals. Canadian Journal of low-frequency water-borne sound on divers: Open water Zoology, 53, 227-231. trial (Naval Submarine Medical Research Laboratory Terhune, J. M., & Turnbull, S. (1995). Variation in the psy- Groton, CT, Interim Report A727133). chometric functions and hearing thresholds of a harbor Stone, C. J. (2003). The effects of seismic activity on seal. In R. A. Kastelein, J. A. Thomas, & P. E. Nachtigall marine mammals in UK waters, 1998-2000. Joint Nature (Eds.), Sensory systems of aquatic mammals (pp. 81- Conservation Committee Report, 323. 78 pp. 93). Woerden: The Netherlands: DeSpil Publishers. Stone, C. J., & Tasker, M. L. (2006). The effects of seismic Thiery, L., & Meyer-Bisch, C. (1988). Hearing loss due airguns on cetaceans in UK waters. Journal of Cetacean to partly impulsive industrial noise exposure at levels Research and Management, 8, 255-263. between 87 and 90 dB (A). Journal of the Acoustical Strasser, H., Irle, H., & Legler, R. (2003). Temporary hear- Society of America, 84, 651-659. ing threshold shifts and restitution after energy-equiva- Thomas, J. A., Chun, N., Au, W. W. L., & Pugh, K. lent exposures to industrial noise and classical music. (1988). Underwater audiogram of a false killer whale Noise Health, 5, 75-84. (Pseudorca crassidens). Journal of the Acoustical Supin, A. Ya., & Popov, V. V. (2007). Improved techniques Society of America, 84, 936-940. of evoked-potential audiometry in odontocetes. Aquatic Thomas, J. A., Pawloski, J. L., & Au, W. W. L. (1990a). Mammals, 33, 14-23. Masked hearing abilities in a false killer whale Supin, A. Ya., Popov, V. V., & Mass, A. M. (2001). The sen- (Pseudorca crassidens). In J. A. Thomas & R. A. sory physiology of aquatic mammals. Boston: Kluwer Kastelein (Eds.), Sensory abilities of cetaceans (pp. 395- Academic Publishers. 332 pp. 404). New York: Plenum Press. Supin, A. Ya., Nachtigall, P. E., Pawloski, J. L., & Au, Thomas, J. A., Moore, P., Withrow, R., & Stoermer, M. W. W. L. (2003). Evoked potential recording during echo- (1990b). Underwater audiogram of a Hawaiian monk location in a false killer whale (Pseudorca crassidens). seal (Monachus schauinslandi). Low-frequency amphib- Journal of the Acoustical Society of America, 113, 2408- ious hearing in pinnipeds: Methods, measurements, 2411. noise, and ecology. Journal of the Acoustical Society of Suryan, R. M., & Harvey, J. T. (1998). Variability in reac- America, 87, 417-420. tions of Pacific harbor seals, Phoca vitulina richardsi, to Thomas, J. A., Kastelein, R. A., & Awbrey, F. T. (1990c). disturbance. Fishery Bulletin, 97, 332-339. Behavior and blood catecholamines of captive belugas Szymanski, M. D., Bain, D. E., Kiehl, K., Pennington, during playbacks of noise from an oil drilling platform. S., Wong, S., & Henry, K. R. (1999). Killer whale Zoo Biology, 9, 393-402. (Orcinus orca) hearing: Auditory brainstem response Thompson, D., Sjöberg, M., Bryant, E. B., Lovell, P., and behavioral audiograms. Journal of the Acoustical & Bjørge, A. (1998). Behavioural and physiologi- Society of America, 106, 1134-1141. cal responses of harbour (Phoca vitulina) and grey Taylor, B. L., & Dawson, P. K. (1984). Seasonal changes (Halichoerus grypus) seals to seismic surveys. Abstracts in density and behavior of harbor porpoise (Phocoena of the World Marine Mammal Science Conference, phocoena) affecting census methodology in Glacier Monaco, p. 134. Bay National Park, Alaska. Report of the International Thorson, P. H., Francine, J. K., & Eidson, D. A. (1998). Whaling Commission, 34, 479-483. Acoustic measurement of the Titan IV A-18 launch and Taylor, K. A., Nachtigall, P. E., Mooney, T. A., Supin, quantitative analysis of harbor seal (Phoca vitulina rich- A. Ya., & Yuen, M. M. L. (2007). A portable system ardsi) behavioral and auditory responses. Los Angeles for the evaluation of the auditory capabilities of marine Air Force Base: Report by SRS Technologies, Systems mammals. Aquatic Mammals, 33, 93-99. Development Division, for Space and Missile Systems Taylor, V. J., Johnston, D. W., & Verboom, W. C. (1997). Center, U.S. Air Force Materiel Command. Acoustic harassment device (AHD) use in the aquacul- Thorson, P. H., Francine, J. K., Berg, E. A., Meyers, ture industry and implications for marine mammals. L. E., Oliver, G. W., & Eidson, D. A. (1999). Quantitative 496 Southall et al.

analysis of behavioral responses for selected pinni- Active (SURTASS LFA) Sonar (Vol. 1). Arlington, VA: peds on Vandenberg Air Force Base and San Miguel Chief of Naval Operations. Island, California and acoustic measurement of the 24 Verboom, W. C. (2002). Noise criteria for marine mammals September 1999 Athena 2 IKONOS-II Launch. Report (TNO Technical Report #HAG-RPT-010120). Delft, by SRS Technologies, Systems Development Division, Netherlands: TNO TPD. for Lockheed Martin Environmental Services. Vrijkotte, T. G. M., van Doornen, E. J. P., & de Geus, E. J. C. Thorson, P. H., Francine, J. K., Berg, E. A., Fillmore, (2000). Effects of work stress on ambulatory blood pres- L. E., & Eidson, D. A. (2000a). Acoustic measurement of sure, heart rate, and heart rate variability. Hypertension, the 17 August 2000 Titan IV B-28 launch and quantita- 35, 800-886. tive analysis of auditory and behavioral responses for Wales, S. C., & Heitmeyer, R. M. (2002). An ensemble selected pinnipeds on Vandenberg Air Force Base and source spectra model for merchant ship-radiated noise. San Miguel Island, CA. Los Angeles Air Force Base: Journal of the Acoustical Society of America, 111, 1211- Report by SRS Technologies, Systems Development 1231. Division, for Space and Missile Systems Center, U.S. Wang, D., Wang, K., Ziao, Y., & Sheng, G. (1992). Auditory Air Force Materiel Command. sensitivity of a Chinese river dolphin (Lipotes vexilli- Thorson, P. H., Francine, J. K., Berg, E. A., Fillmore, L. E., fer). In J. A. Thomas, R. A. Kastelein, & A. Ya. Supin & Eidson, D. A. (2000b). Acoustic measurement of the (Eds.), Marine mammal sensory systems (pp. 213-221). 21 September 2000 Titan II G-13 launch and quantita- New York: Plenum Press. tive analysis of behavioral responses for selected pin- Ward, W. D. (1970). Temporary threshold shift and damage- nipeds on Vandenberg Air Force Base, CA. Los Angeles risk criteria for intermittent noise exposure. Journal of Air Force Base: Report by SRS Technologies, Systems the Acoustical Society of America, 48, 561-574. Development Division, for Space and Missile Systems Ward, W. D. (1997). Effects of high-intensity sound. In Center, U.S. Air Force Materiel Command. M. J. Crocker (Ed.), Encyclopedia of acoustics, Volume Tinbergen, N. (1948). Social releasers and the experimen- III (pp. 1497-1507). New York: John Wiley & Sons. tal method required for their study. Wilson Bulletin, 60, Ward, W. D., Glorig, A., & Sklar, D. L. (1958). Dependence 6-51. of temporary threshold shift at 4 kc on intensity and Todd, S., Stevick, P., Lien, J., Marques, F., & Ketten, D. time. Journal of the Acoustical Society of America, 30, (1996). Behavioral effects of exposure to underwater 944-954. explosions in humpback whales (Megaptera novaean- Ward, W. D., Glorig, A., & Sklar, D. L. (1959). Temporary gliae). Canadian Journal of Zoology, 74, 1661-1672. threshold shift from octave-band noise: Applications to Tremel, D. P., Thomas, J. A., Ramirez, K., Dye, G. S., damage-risk criteria. Journal of the Acoustical Society Bachman, W. A., Orban, A. N., et al. (1998). Underwater of America, 31, 522-528. hearing sensitivity of a Pacific white-sided dolphin, Wartzok, D., & Ketten, D. R. (1999). Marine mammal Lagenorhynchus obliquidens. Aquatic Mammals, 24, sensory systems. In J. E. Reynolds II & S. A. Rommel 63-69. (Eds.), Biology of marine mammals (pp. 117-175). Turnbull, S. D., & Terhune, J. (1990). White noise and pure Washington, DC: Smithsonian Institution Press. tone masking of pure tone thresholds of a harbor seal Wartzok, D., Popper, A. N., Gordon, J., & Merrill, J. listening in air and under water. Canadian Journal of (2004). Factors affecting the responses of marine mam- Zoology, 68, 2090-2097. mals to acoustic disturbance. Marine Technology Society Tyack, P. L. (1981). Interactions between singing Hawaiian Journal, 37, 6-15. humpback whales and conspecifics nearby. Behavioral Watkins, W. A., & Schevill, W. E. (1975). Sperm whales Ecology and Sociobiology, 8, 105-116. (Physeter catodon) react to pingers. Deep Sea Research Tyack, P. L. (1998). Acoustic communication under the I, 22, 123-129. sea. In S. L. Hopp, M. J. Owren, & C. S. Evans (Eds.), Watkins, W. A., & Wartzok, D. (1985). Sensory biophysics Animal acoustic communication (pp. 163-220). Berlin: of marine mammals. Marine Mammal Science, 1, 219- Springer-Verlag. 260. U.S. Department of Commerce. (2002). Report of the Wegel, R. L., & Lane, C. E. (1924). The auditory masking Workshop on Acoustic Resonance as a Source of Tissue of one pure tone by another and its probable relation to Trauma in Cetaceans. Retrieved 19 September 2007, from the dynamics of the inner ear. Physics Review, 23, 266- www.nmfs.noaa.gov/pr/pdfs/acoustics/cetaceans.pdf. 285. U.S. Department of Defense (DoD). (1997). Department White, M. J., Jr., Norris, J., Ljungblad, D., Baron, K., & of Defense design criteria standard: Noise limits (MIL- di Sciara, G. (1978). Auditory thresholds of two beluga STD-1474D, AMSC-A7245). Washington, DC. whales (Delphinapterus leucas) (Technical Report No. U.S. Marine Mammal Protection Act of 1972 (16 USC 78-109). San Diego: Hubbs Marine Research Institute. §1361). Williams, R., Bain, D. E., Ford, J. K. B., & Trites, A. W. U.S. Navy. (2001). Final overseas environmental impact (2002). Behavioral responses of male killer whales to statement and environmental impact statement for a “leapfrogging” vessel. Journal of Cetacean Research Surveillance Towed Array Sensor System Low Frequency and Management, 4, 305-310. Marine Mammal Noise Exposure Criteria 497

Wolski, L. F., Anderson, R. C., Bowles, A. E., & Yochem, P. K. (2003). Measuring hearing in the harbor seal (Phoca vitulina): Comparison of behavioral and auditory brainstem response techniques. Journal of the Acoustical Society of America, 113, 629-637. Wright, A. J., Aguilar Soto, N., Baldwin, A. L., Bateson, M., Beale, C. M., Clark, C., et al. (In press). Anthropogenic noise and physiological stress: Behaviour, context and animal welfare. International Journal of Comparative Psychology. Yost, W. A. (1994). Fundamentals of hearing (3rd ed.). New York: Academic Press. 326 pp. Yost, W. A. (2000). Fundamentals of hearing: An introduction (4th ed.). New York: Academic Press. 349 pp. Yuen, M. M. L., Nachtigall, P. E., Breese, M., & Supin, A. Ya. (2005). Behavioral and auditory evoked potential audiograms of a false killer whale (Pseudorca crassidens). Journal of the Acoustical Society of America, 118, 2688-2695. Yurk, H. (2000). Experimental attempts to reduce predation by harbor seals on out-migrating juvenile salmonids. Transactions of the American Fisheries Society, 129, 1360-1366. Zimmer, W. M. X., & Tyack, P. L. (2007). Repetitive shallow dives pose decompression risk in deep-diving beaked whales. Marine Mammal Science, 23, 888-925. Aquatic Mammals 2007, 33(4), 498-501, DOI 10.1578/AM.33.4.2007.498

Appendix A. Acoustic Measures and Terminology

This appendix provides a more detailed descrip- Pulses and nonpulses are distinguished here by tion of many key acoustic measurements and an empirical approach based on several temporal terms used throughout the noise exposure crite- weightings. Various exponential time-weighting ria. It is not intended as an exhaustive or instruc- functions applied in measuring pulse and nonpulse tive text on these exceedingly complex issues sounds may yield different measured received (for more detailed treatments, see Kinsler et al., levels (RLs) (see Harris, 1998). By way of illus- 1982; ANSI, 1986, 1994; Richardson et al., 1995; tration, most sound level meters (SLM) provide Harris, 1998; NRC, 2003). Rather, it is intended options for applying either a slow or fast time to provide fairly straightforward definitions and constant (1,000 or 125 ms, respectively) for mea- equations related to the marine mammal noise suring nonpulses, or an impulse time constant (35 exposure criteria. ms) appropriate for measuring pulses. If applied to a sound pulse, the slow or fast SLM settings Pulses and Nonpulse Sounds result in lower sound pressure level (SPL) measurements than those obtained using the impulse The distinction between these two general sound setting. Each of these time constants was selected types is important because they have differing based on the physical properties of the human potential to cause physical effects, particularly auditory system. It is clear that further empiri- with regard to hearing (e.g., Ward, 1997). cal measures of temporal resolution in marine Pulses, as used in the context of this paper, mammals are needed, particularly for animal taxa are defined as brief, broadband, atonal, tran- whose hearing extends to significantly higher or sients (ANSI, 1986; Harris, 1998, Chapter 12). lower frequencies than in humans (see Chapter 5, Examples of pulses (at least at the source) are “Research Recommendations”). Future noise cri- explosions, gunshots, sonic booms, seismic airgun teria are expected to include distinctions between pulses, and pile driving strikes. These sounds are pulse and nonpulse sounds that may be more spe- all characterized by a relatively rapid rise from cifically appropriate for marine mammals than ambient pressure to a maximal pressure value fol- is this current simple approach. We note also the lowed by a decay period that may include a period need for an explicit distinction and measurement of diminishing, oscillating maximal and minimal standard, such as exists for aerial signals (ANSI, pressures. The rapid rise-time characteristic of 1986). these sounds ensures that they are also broad- Peak sound pressure is the maximum absolute band in nature, with the higher-frequency com- value of the instantaneous sound pressure during ponents being related to the rapidity of the rise- a specified time interval and is denoted as Pmax in time. Pulses, either as isolated events or repeated units of Pascals (Pa). It is not an averaged pressure. in some succession, generally have an increased Peak pressure is a useful metric for either pulse or capacity to induce physical injury as compared nonpulse sounds, but it is particularly important with sounds that lack these features. for characterizing pulses (ANSI, 1986; Harris, Nonpulse (intermittent or continuous) sounds 1998, Chapter 12). Because of the rapid rise-time can be tonal, broadband, or both. Some of these of such sounds, it is imperative to use an adequate nonpulse sounds can be transient signals of short sampling rate, especially when measuring peak duration but without the essential properties of pressure levels (Harris, 1998, Chapter 18). pulses (e.g., rapid rise-time). Examples of sources Peak-to-peak sound pressure is the algebraic producing nonpulse sounds include vessels; air- difference between the maximum positive and craft; machinery operations, such as drilling or maximum negative instantaneous peak pressure. wind turbines; and many active sonar systems. The The mean-squared pressure is the average duration of such sounds, as received at a distance, of the squared pressure over some duration. For can be greatly extended in highly reverberant envi- nonpulse sounds, the averaging time is any con- ronments. It is critical to note that a sound that has venient period sufficiently long enough to permit characteristics of a pulse at the source may, as a averaging the variability inherent in the type of result of propagation effects, lose those charac- sound. Note that some of the variability of the teristics at some (variable) distance and could be received sound typically arises simply from the characterized as a nonpulse for certain receivers. Marine Mammal Noise Exposure Criteria 499

t2 2 1 2 Pamb )( dttP (1) relative movement of a free-ranging animal and a ³ t2 tt 12 1 t source, whether the latter is moving or stationary. P 2 1 2 )( dttP (1) Sound pressure levels (SPLs) are given as the amb ³ tt 12 decibel (dB) measures of the pressure metrics t1 (1) eq. defined above. The root-mean-square (RMS) The temporal (or event) sound exposure 2 SPL is given as dB re: 1 µPa for underwater sound Etemp (t) (in Pa -s) is then calculated as and dB re: 20 µPa for aerial sound. Peak SPLs are t t2 given as dB re: 1 µPa (peak) in water and dB re: 2 1 2 2 2 Pamb )()( )( dttP dtPtPtE (2) (1) 20 µPa (peak) in air. Peak-to-peak SPLs are dB temp ³ t ³ amb t tt 12 t re: 1 µPa (peak-to-peak) in water and dB re: 20 2 1 2 t2 2 (2) eq. temp2 1 )()( 2 amb dtPtPtE (2) µPa (peak-to-peak) in air. Source level (SL) is the The 0% sound exposure³ point (ta) signifies the Pamb )( dttP (1) received level measured or estimated 1 m from the “start” of the acoustict2 event tt ³and the 100% sound 12 t1 source. exposure point (tb) signifies the “end” of the event. Duration is the length of a sound, generally in These two points are where the E(t) curve begins seconds. Duration is important because it affects to rise and wheret it tlevels off, respectively. Their various acoustic metrics, including mean-square selection can be difficult2 2due 2to variation2 in ambi- temp )()( amb)()( dtPtPtE amb dtPtPtE (3) (2) and/or RMS sound pressure (Madsen, 2005). ent noise preceding³ t (and³ overlapping) the acoustic t Because of background noise and reverbera- event, as well tasa reverberation2 2 t )()( plus2 ambientdtPtPtE noise (3) tion, duration can be difficult to define precisely. following the event.³ Consequently,2 amb many2 investi- temp )()( amb dtPtPtE (2) Various definitions of duration exist in the litera- gators identify theseta points³ subjectively. ture such as the time between the points on the The sound exposure tE(t)2 (in Pa2-s), where t £ pressure-time waveform P(t) determined to be tb, is then calculated as either 10 dB (0.316 times) or 20 dB (0.1 times) t t below the instantaneous peak pressure (Hamernik 2 2 2 )()( )()( dttPtE amb dtPtPtE (4) (3) & Hsueh, 1991). Malme et al. (1983, 1984) used a ³ ³ t t similar approach. Harris (1998, Chapter 11) sug- ta a t 2 (3) eq. 2)()( dttPtE 2 (4) gested alternative constructs, including exponen- where E100 = E(tb³) is 100% of the sound expo- t )()( amb dtPtPtE (3) tial time weighting. This topic is discussed below sure. For the 5% apoint,³ E(t) is determined as E5 ta with regard to updating measurement standards = 0.05•E100 = E(t5), while E(t) for the 95% energy for impulse sounds. Greene (1997) described a point is determined as E95 = 0.95•E100 = E(t95). practical definition of pulse duration based on Thus, E90 = E95 –t E5 and duration (Td) = t95 – t5 (s) N T ½ the interval over which 90% of the sound energy where the received pressure2 2)()( dttPtE level greatly exceeds (4) ° n )( dttp ° arrived at the receiver. This interval could also be the ambient level,³¦ eq.³N 3 Tcan be reduced to n 1 used as the averaging time for mean-square pres- °ta 0t 2 ° ½ SEL log10 10 ® )( ¾dttp (5) sure (Madsen, 2005). This approach has been °¦³ 2 2n ° ° ° n p 1ref )( )()( dttPtE ° ° (4) widely used in measuring exposure duration and SEL log10 ³ 0 (5) °10 ® ta 2 ° ¾ SPL values for seismic airgun and pile driving sig- ¯ p )( ¿ (4) eq. nals (e.g., McCauley et al., 1998; Blackwell et al., Sound exposure° level ref (SEL) is° the decibel ° ° 2004b). Defined as such, duration is the interval level of the cumulativeN T sum-of-square pressures ¯ ½¿ 2 between the 5% and 95% bounds of the time-inte- over the duration of a sound2 (e.g.,)( dttp dB re: 1 µPa -s) gral of the instantaneous sound-pressure squared for sustained nonpulse°¦³ soundsn where° the exposure ° n 1 0 ° (sound exposure [E(t)] as defined below) while is of a constant natureN (i.e.,T source and animal SEL log10 10 ® ¾ (5) accounting for background noise and low-level positions are held roughly constant).22 4 ½ However, > XO pref)( P )( @ dttp reverberation (assumed to be continuous). That is, this measureXkurt is)( ° also° extremely¦³ n useful° ° for pulses (6) the background noise is measured over a period of and transient nonpulse° ° n sounds 1V04 because°4 ° it enables SEL log10 ¯10 ® > XO P ¿ @¾ (5) time before the pulse occurs and then is subtracted sounds of differingXkurt )( durationp to )( be2 characterized (6) from the cumulative sum-of-square pressures to in terms of total energy° for purposesVref4 of° assessing ° ° determine the sum-of-square pressures from the exposure risk. ¯ ¿ impulsive sound alone. This is done by manually The SEL metric also enables integrating sound identifying a period of time (t1, t2) preceding the energy across multiple exposures from sources fR )( 4 event, deemed to be representative of ambient suchfM as seismiclog20)( airguns,> XO pile P driving, @ and most (7) noise. The mean-square pressure (in Pa2) of the sonar signals. Xkurt 10Several)( methodsfR })(max{ exist for summing (6) 2 fR 4)( ambient (Pamb) is determined with the following energyfM over multiplelog20)( exposures.V We use a rela- (7) relationship: tively straightforward10 approach> XO fR here,})(max{ P specifically4 @ Xkurt )( (6) V 4

22 high ff fR )( fR )( fM log20)( (7) 10 22 ffff 2222 ))(( (8) high high ff lowfR })(max{ fR )( fR )( 2222 fM log20)( ffff ))(( (8)(7) high10 lowfR })(max{

ff 22 fR )( high 2222 (8) high ffff low ))(( ff 22 fR )( high 2222 (8) high ffff low ))(( 1 t2 P 2 2 )( dttP (1) amb ³ tt 12 tt 1 t2 1 12 P 2 2 )( dttP P 2 (1) 2 )( dttP (1) amb ³ amb ³ tt tt 12 12 t1 t1

t 2 )()( 2 dtPtPtE (2) temp ³ amb t t2t 2 )()( 2 dtPtPtE (2) 2 )()( 2 dtPtPtE (2) temp ³ amb temp ³ amb t2 t2

t 2 )()( 2 dtPtPtE (3) ³ amb t tat t 1 22 2 2 2 2 2)()( amb dtPtPtE (3) )()( amb dtPtPtE (3) Pamb ³ )( dttP (1)³ ³ t tatt a 12 t1

t ³ 2 )()( dttPtE (4) t tat t 2 2 2 )()( dttPtE 2 (4) )()( dttPtE (4) temp ³ )()( amb dtPtPtE (2)³ ³ t ta a t2 500 Southall et al. N T ½ 2 )( dttp °¦³ n ° ° n 1 ° T marine mammal hearing groupsN 0T (see the “Marine t N SEL log10 (5) 2 ½ Mammal Functional10 ® Hearing Groups”2 2 ¾½ section 2 2 p )( ° n )( dttp ° in Chapter 2). M-weighting°° hasrefn a )( mathematicaldttp °° ¦)()( ³ amb dtPtPtE (3) ¦³ ³ ° n 1 0 ° structure similar to the °C-weighting° n 1 0 used in°° human SEL log10 ta 10 ® ¾ SEL log10 10¯® (5) ¿¾ (5) p )( 2 hearing, which reflects the factp that)( 2 sounds must ° ref ° be more intense at high° and lowref frequencies° for ° ° ° ° them to have equal auditory¯ effect. C-weighting ¯ ¿ (5) eq. is most appropriate for determining the effects¿ of where instantaneous sound-pressure (p) is mea- intense sounds—that is, those with loudness equal sured in µPa fort n exposures and the reference to that of a tone 100 dB above threshold4 at 1,000 2 > XO P @ pressure (pref) is 1 µPa under)()( waterdttPtE and 20 µPa in air. Hz. The M-weighting (4)Xkurt )( was designed to do much (6) This summation ³procedure essentially generates a the same for the different marine mammal4 groups ta V single exposure “equivalent”> XO value P that 4 @ assumes with the only difference being> theirXO varyingP 4 @ low- no recovery of hearingXkurt )( between repeated expo- and high-frequency cutoffs.Xkurt )( (6) The M-weighting for (6) sures. The appropriate units forV underwater4 SEL marine mammals, like the C-weightingV 4 used in are dB re: 1 µPa²-s, and the appropriate units for humans, rolls off at a rate of 12 dB per octave. aerial SEL are dB re: (20 µPa)2-s. The general expression for M-weighting (M[f]), Kurtosis is a statisticalT measure of the prob- using estimated frequency cutoffsfR )( for each func- N ½ fM log20)( (7) ability distribution of sound2 pressure)( dttp amplitudes tional marine mammal10 hearing group, is given as (Hamernik & Hsueh,°¦ ³ 1991;n Lei ° et al., 1994; fR })(max{ ° n 1 ° Hamernik et al., 2003)0 that describesfR )( the shape fR )( SEL log10 10 ® 2 ¾ fM log20)( (5) (7) of the fM amplitude log20)( distribution.10 In some regards, 10(7) ° pref )( fR ° })(max{ fR })(max{ it appears to be ° a highly relevant ° metric in that (7) eq. impulsive sound¯ with high kurtosis¿ and high where instantaneous peak pressure may be particularly ff 22 injurious to some mammals (Hamernik et al., fR )( high 2003). Kurtosis is related to the fourth central- ffff 2222 ))(( (8) moment and is defined for random22 variable X as high ff 22 low (8) eq. high ff The estimated lower and highupper “functional” hear- fR )( 4 fR )( 2222 XO P 2222 ing limits (flow and fhigh) (8)for each of the five func- (8) high> ffff low@ ))(( high ffff low ))(( Xkurt )( 4 tional marine mammal(6) hearing groups and the V (6) eq. names of the frequency-weighting functions are where O is the expectation operator, µ is the mean, given in Table 2. The weighting functions de- and S is the standard deviation. When kurtosis is emphasize frequencies that are near the lower and high, amplitude distribution is generally more cen- upper frequency ends of the estimated hearing trally peaked and may have broader tails. Normal range as indicated by the negative relative values distributions have a kurtosisfR )( value of 3 indepen- in Figure 1. dent of the mean or standard deviation. Audition (hearing) is a well-developed and pri- fM log20)( 10 (7) Frequency-selective weightingfR is})(max{ often employed mary sensory modality for most, if not all, marine to measure (as a single number) sound pressure or vertebrates (Schusterman, 1981; Tyack, 1998; Fay energy in a specific frequency band, with emphasis & Popper, 2000). The vertebrate hearing system or de-emphasis on particular frequencies as a func- involves coding, processing, integrating, and tion of the sensitivity to those frequencies. For aerial responding to sound in a variety of ways, some hearing in humans, A-weighting is derived from the not outwardly evident (Yost, 2000). inverse of the idealized 40-phonff 22 equal loudness hear- Hearing (auditory) threshold is most com- ingfR function)( across frequencieshigh standardized to 0 dB monly measured by behavioral or electrophysi- 2222 (8) at 1 kHz (Harris, high 1998), providing ffff low ))(( level measures ological responses and is defined as the SPL of denoted as dB(A). C-weighting is determined from the quietest sound audible in some percentage the inverse of the idealized 100-phon equal loudness of experimental trials. In air, measurements are hearing function (which differs in several regards often conducted in specially constructed sound from the 40-phon function) standardized to 0 dB at chambers. When that is not possible, tests must be 1 kHz (Harris, 1998); level measures are denoted as conducted in low background noise conditions to dB(C). yield meaningful threshold data. Absent equal-loudness contours for marine Sensation level represents the difference (in mammals, special weighting functions based dB) between the overall level of a sound and the loosely on human weighting functions and receiver’s auditory threshold at similar sound fre- general knowledge of functional hearing band- quencies. It is particularly useful as a means of width, were developed here for the five functional comparing the relative exposure level of a sound Marine Mammal Noise Exposure Criteria 501 for individuals that may have different hearing Temporary Threshold Shift (TTS) is a revers- capabilities. Sensation level is sometimes abbre- ible elevation in hearing threshold (i.e., a non- viated SL, but this is not done in this document to permanent reduction in hearing sensitivity) most avoid confusion with the very different concept of commonly resulting from noise exposure. source level. Permanent Threshold Shift (PTS) is a perma- Auditory masking is the partial or complete nent elevation in hearing threshold (i.e., an unre- reduction in the audibility of signals due to the coverable reduction in hearing sensitivity). PTS presence of interfering noise (see Buus, 1997). can occur from a variety of causes, but it is most The degree of masking depends on the spectral often the result of intense and/or repeated noise and temporal relationships between signals and exposures. In that case it is also referred to as noise masking noise as well as their respective RLs induced hearing loss (NIHL) or noise induced per- (e.g., Fletcher, 1940). manent threshold shift (NIPTS). Sound localization is the determination of source location based on features of received sounds. This critical, complex process of the auditory system can involve the detection of sounds produced directly by a source (passive listening) or the detection of echoes reflected off a target (as in the case of biosonar). Auditory scene analysis is the process by which the auditory system sorts out related elements of a complex acoustic environment into those arising from discrete sound sources. This process is similar to psychological processes underlying visual perception whereby many different visual images are perceived as discrete elements within a visual scene.

Courtesy: A. Friedlander Aquatic Mammals 2007, 33(4), 502-508, DOI 10.1578/AM.33.4.2007.502

Appendix B. Studies Involving Marine Mammal Behavioral Responses to Multiple Pulses

Low-Frequency Cetaceans/Multiple Pulses Beaufort Sea. Whales in their westward autumn (Cell 2) migration over three seasons (1996 to 1998) were detected with aerial surveys on days with and Numerous field observations have been made without seismic survey activity. Using the obser- of low-frequency cetaceans reacting to multiple vations of dozens of migrating whales during peri- pulses, either opportunistically exposed to ongo- ods when airguns were not active, we were able ing operations or by intentional exposures. A mod- to calculate the percentage of observed whales erate number of species and experimental condi- during seismic surveys that demonstrated avoid- tions have been considered, but the source was ance behavior at various RLs (see Table 7). These usually a seismic airgun or arrays of these intense results indicate that migrating bowhead whales in sources. Some studies focused on migrating ani- the Richardson et al. (1999) study often avoided mals observed from fixed observation platforms or areas where RLs exceeded 120 to 130 dB re: 1 µPa in/near migration corridors. (RMS over pulse duration). The general results of the severity scaling In contrast, Richardson et al. (1986) observed analysis for this condition suggest the onset of quite different movement patterns of bowhead more significant behavioral disturbances from whales exposed to seismic airgun sounds on multiple pulses for migrating bowhead whales at their summer feeding grounds in the Canadian RLs (RMS over pulse duration) around 120 dB Beaufort Sea. Received levels from a single seis- re: 1 µPa (Richardson et al., 1999). For all other mic airgun (0.66-L) were measured in situ near low-frequency cetaceans (including feeding bow- individual whales being observed 3 to 5 km from head whales), this onset was at RLs around 150 the sound source, and ranged from 118 to 133 dB to 160 dB re: 1 µPa (Malme et al., 1983, 1984; re: 1 µPa. Visual orientation by groups of whales Richardson et al., 1986; Ljungblad et al., 1988; during airgun exposure was observed on two of Todd et al., 1996; McCauley et al., 1998). There is five occasions; only minor changes in swim- essentially no overlap in the RLs associated with ming and respiration patterns were observed. the onset of behavioral responses by members of Richardson et al. (1986) also made opportunistic these two groups based on the information cur- observations of groups of bowhead whales near rently available. a seismic vessel operating an airgun array. At the Seismic airguns operated near bowhead whales highest RLs, some measurements exceeded the generally initiate avoidance reactions as well as dynamic range of the recording equipment and are changes in locomotion and respiration (Reeves considered exposure minima, although this was et al., 1984; Richardson et al., 1985, 1986, 1999; not the case for most measurements relevant to Ljungblad et al., 1988). During the autumn migra- the behavioral observations. From these observation, avoidance behavior has been observed at rel- tions and the controlled exposure to sounds from a atively great (20+ km) ranges from source opera- single airgun, Richardson et al. (1986) concluded tions (Koski & Johnson, 1987; Richardson et al., that some whales responded subtly by changing 1999). Ljungblad et al. (1988) did not investigate diving and breathing patterns at relatively low behavioral reactions over such ranges. During RLs (ca. 120 to 140 dB re: 1 µPa), but that avoid- the summer, feeding bowheads exhibited subtle ance and other more profound behavioral changes behavioral responses but not active avoidance were generally not observed unless the RL was at distances beyond 6 km from airgun sources ≥ 160 dB re: 1 µPa. (Richardson et al., 1986; see also Miller et al., Ljungblad et al. (1988) conducted a series of 2005). acoustic experiments on behavioral reactions of Richardson et al. (1999) studied autumn- bowhead whales exposed to sounds from ships migrating bowhead whale and found avoidance with operating airgun(s). Experiment 1 was con- by most individual whales at distances up to 20 ducted on a group of eight whales. When a seis- km and some avoidance at 20 to 30 km. Seismic mic vessel approached to within 3.5 km (max. surveys using airgun arrays with 6 to 16 guns and RL near observed individuals was 142 dB re: 1 total volumes of 560 to 1,500 in3 were conducted µPa), the bowhead whales coalesced and moved in shallow (generally < 20 m) water of the Alaskan in a tight group away from the approaching vessel. Marine Mammal Noise Exposure Criteria 503

Experiment 2 involved a group of three bowhead around active airguns was very limited. The whales that demonstrated startle responses at the approximate difference in mean sighting distance onset of sounds from an airgun 7 km away (max. was ~600 m. Similarly, the aerial surveyors did measured RL was 165 dB re: 1 µPa). Behavior not detect any large-scale avoidance of the airgun returned to pre-exposure values shortly after the operations by bowheads. These observations were operation was terminated. Experiment 3 involved a generally consistent for both years in which mea- group of seven bowhead whales that demonstrated surements were made and are generally consistent avoidance behavior at ranges of ~3.5 km (max. with the observations of Richardson et al. (1986) measured RL of 178 dB re: 1 µPa). Experiment in the same region and season (summer). Animals 4 involved a group of 50 bowhead whales. not exhibiting observable behavioral reactions Behavioral reactions were first observed at ranges (response score: 0) were consistently sighted in of about 8 km (max. measured RLs of 157 dB re: areas where RLs very likely ranged from 130 to 1 µPa) and avoidance behavior was noted at ~3 180 dB re: 1 µPa. The general lack of sightings km (RLs ~165 dB re: 1 µPa). Avoidance behavior within a small area around the seismic vessel sug- in this instance similarly abated shortly following gests behavioral avoidance (response score: 6) at cessation of exposure (and was thus assigned a RLs exceeding 180 dB re: 1 µPa. Exposures were behavioral score of 6). not estimated to exceed 190 dB re: 1 µPa. The Recent work on summering bowhead whales entire study was treated as a single observation for by Miller et al. (2005) also found that avoidance the purposes of the behavioral analysis. Half of responses were limited to distances of at most a the “observation” was scored as avoidance behav- few kilometers and RLs exceeding 160 dB re: 1 ior and half as no response, with exposure RL bins µPa. Miller et al. conducted a monitoring program from 130 to 190 dB re: 1 µPa (Table 6). over two summers for various marine mammals The combined data for bowhead whale avoid- offshore of the Mackenzie Delta in the Southeast ance of airgun sounds (Richardson et al., 1986, Beaufort Sea before and during seismic surveys. 1999; Ljungblad et al., 1988; Miller et al., 2005) They presented observational data from both indicated that, when migrating, these animals can vessel-based and aerial observations of bowhead be particularly prone to behavioral disturbance, whales, belugas, and several pinniped species. with the onset of significant responses occur- The general methodology is briefly discussed ring at approximately 120 dB re: 1 µPa (RMS here as well as data on behavioral responses by over pulse duration) (Table 6). In contrast, when low-frequency cetaceans (bowhead whales) and feeding, they may show subtle effects at low RLs the corresponding rank on the severity scale. but only tend to display active avoidance at RLs The airgun operations involved 3-D seismic exceeding 160 dB re: 1 µPa. profiling from a 67-m vessel using two identi- Low-frequency cetaceans, other than migrating cal 2,250 in3 sleevegun arrays, each with 24 air- bowhead whales, appear to be much more toler- guns. Shots were at 8-s intervals and at a depth of ant of exposure to multiple pulses, although data 5 m below the surface of the water. Surveys were are limited to a few species and (primarily) airgun conducted in very shallow water (13 m average). sources. Available data for species other than bow- Acoustic monitoring with calibrated hydrophones heads include reactions to opportunistic and inten- across the 10 Hz to 24 kHz bandwidth was con- tional exposures of humpback whales (Malme ducted while seismic operations were underway. et al., 1985; Todd et al., 1996; McCauley et al., Physical properties of the operational environ- 1998, 2000) and gray whales (Malme et al., 1983, ment, and hence sound propagation in the shal- 1984, 1986, 1988; also see review by Moore & low water environments, were highly variable, Clarke, 2002). Todd et al. (1996), Malme et al. but RLs as a function of range from active airgun (1983, 1984), and McCauley et al. (1998) are arrays were measured. Vessel-based observers included in the behavioral scoring analysis here and aerial surveyors used line-transect methods because they contain sufficient information on to monitor marine mammals in and adjacent to exposures and individual responses of low-fre- seismic operational areas, both before and during quency cetaceans other than bowhead whales. shooting. Bowhead whales observed during the Todd et al. (1996) analyzed the impact of con- periods coincident with seismic operations were struction activity (explosions and drilling) on the presumed to be feeding (i.e., not migrating). Many entanglement of three foraging humpback whales bowheads (355 individuals in 232 groups) were off Newfoundland. They conducted observations seen by marine mammal observers aboard the seis- of whale behavior during and following explo- mic vessel. Sighting rates were lower and mean sions and obtained acoustic measurements of sighting distances were somewhat larger during underwater sound signatures. The data suggest seismic operations than at times when the airguns few short-term changes in movement and behav- were not operating, but the zone of avoidance ior patterns in response to discrete exposures; 504 Southall et al. however, repeated exposures to high levels may and social behaviors. Five trials were excluded have resulted in sensory impairment in whales and from consideration in our analysis, but behavioral perhaps greater susceptibility to entanglement in observations were reported for the remaining 11. fishing gear. Of these, ten included cow pods of various sizes, Malme et al. (1983, 1984) documented behav- and one was a lone male. Since the cow pods were ioral reactions of migrating gray whales to seismic not migrating and were not individually identi- pulses from both single airguns and an array. Only fied, a single behavioral observation is included land-based observers were used, which meant that in Table 7 for the ten observations. The results the observers could not have affected the whales’ for the cow pods were very consistent, indicating behavior. Both phases of the investigation yielded clear avoidance (severity score = 6) of the airgun the general conclusion that RLs exceeding 160 dB at exposures in the 140 to 150 dB re: 1 µPa range re: 1 µPa (on an approximate RMS basis) were (RMS over pulse duration). The lone male essen- required to cause migrating gray whales to avoid tially ignored the airgun until within ca. 100 m, airgun sounds, although statistically significant when the received level approached 180 dB re: reactions that were less profound occurred at 1 µPa (RMS); this response may have had as much much larger ranges and lower levels. From their or more to do with the presence of the vessel than empirical phase II results, Malme et al. (1984) exposure to the airgun sound. Noting this con- calculated 10, 50, and 90% probabilities of gray textual complexity here, a single observation for whale avoidance reactions in these conditions to this individual is reported in the 170 to 180 dB re: be 164, 170, and 180 dB re: 1 µPa, respectively. 1 µPa exposure bin in Table 7 as general avoid- McCauley et al. (1998) made behavioral obser- ance (severity score = 6). vations of migrating humpback whales off western Australia during seismic operations with a single Mid-Frequency Cetaceans/Multiple Pulses airgun and several airgun array configurations. (Cell 5) Seismic track lines were oriented perpendicular to the migration paths of humpback whales moving A limited number of behavioral observations have through the area. Aerial surveys were conducted been made of mid-frequency cetaceans exposed to to determine the presence of humpback whales multiple pulses. Field observations have involved moving through the survey area. Detailed obser- sperm whales and a few other odontocete species vational data were presented for individuals and exposed to seismic airguns and small explosives groups of whales; RLs were measured at vari- (Madsen & Møhl, 2000; Madsen et al., 2002; able ranges. The seismic survey did not appear to Miller et al., 2005). Laboratory investigations have grossly affect the migration of humpback whales considered behavioral responses to various kinds through the area; however, avoidance behav- of multiple pulse sources (Akamatsu et al., 1993). ior was observed to begin at ranges from 5 to 8 As in most criteria cells, a number of reported km and to be almost universal at ranges of 1 to observations were not scored and reported here 4 km. Exposures to a single airgun (20 in3) were due to lack of relevant information and difficulties extrapolated to equivalent ranges for exposure to in accounting for various contextual variables. A full arrays based on empirical measurements. The summary of those studies used and others consid- data indicated an onset of behavioral avoidance at ered is given in Table 8; the severity scaling analy- ~159 dB re: 1 µPa (peak-to-peak), roughly equiva- sis for Cell 5 is shown in Table 9. lent to the full array at 5 km. General behavioral The combined data for mid-frequency ceta- avoidance (most individuals) occurred at a range ceans exposed to multiple pulses do not indicate of about 1 km for the single gun (~168 dB re: 1 a clear pattern of increasing probability and sever- µPa [peak-to-peak]), equivalent to the full array at ity of response with increasing RLs. In certain about 3 km. Some individual whales did approach conditions, multiple pulses at relatively low RLs closer than the typical 3-km stand-off range; these (~80 to 90 dB re: 1 µPa) temporarily silence indi- may have been males investigating the presence of vidual acoustic behavior for one species (sperm the low-frequency source. whales). In other cases with slightly different In addition to presenting again the results given stimuli, RLs in the 120 to 180 dB re: 1 µPa range in the McCauley et al. (1998) paper, McCauley failed to elicit observable reactions from a signifi- et al. (2000) provide additional behavioral observa- cant percentage of individuals of the same species, tions of 16 humpback whale pods that approached both in the field and in the laboratory. as a single airgun (Bolt PAR 600b 20-in3) was operated. These whales were also observed after Field Observations (Cell 5) termination of airgun operations. These trials were Madsen & Møhl (2000) investigated sperm whale conducted in a large embayment (Exmouth Gulf) responses to small underwater detonators that as the animals were engaged in a variety of resting included 1-g TNT charges, producing a 1-ms Marine Mammal Noise Exposure Criteria 505 broadband (300 Hz to 15 kHz) pulse; several purposes of our behavioral analyses, the combined charges were triggered per day. Echolocation click beluga results were treated as a single observa- behavior was monitored, and one whale was local- tion that was subdivided equally into either avoid- ized acoustically. This individual demonstrated no ance behavior or no observable response. Belugas modulation of vocal behavior when exposed to an exposed to RLs of 100 to 120 dB re: 1 µPa (RMS RMS-equivalent RL of ~173 dB re: 1 µPa. There over pulse duration) were determined to have had was also one observation of a whale exposed to no observable reaction (response score: 0) to seis- 179 dB re: 1 µPa; it continued breathing normally mic exposures. RLs between 120 and 150 dB re: 1 with no visible response. µPa were determined to have induced temporary Madsen et al. (2002) studied responses of sperm avoidance behavior (response score: 6) in belugas, whales in Norway to sounds associated with dis- based on the vessel-based and aerial observations. tant seismic survey operations. Calibrated RLs Based on both the vessel-based and aerial surveys, for individuals and correlated acoustic behavior exposures apparently did not exceed 150 dB re: are reported for three discrete sightings over a 5-d 1 µPa. Weighted behavioral response scores for period. The first observation involved three sperm each of these five exposure RL bins are given in whales tracked by acoustic localization within a Table 7. dispersed array of calibrated hydrophones, which Several studies involved behavioral reactions also recorded airgun sounds from an array operat- of free-ranging, mid-frequency cetaceans but ing 40 km away. RL at the position of the whale lacked specific measures to be included directly was estimated to be 123 dB re: 1 µPa. The second in our analyses. André et al. (1997) exposed sperm observation (3 d later) involved a single sperm whales to various stimuli, including two pulse whale recorded before, during, and after airgun sounds (recorded coda playbacks and a 10-kHz exposure at a range of 86 km; measured RL was pulse). A significant number of exposed whales 130 dB re: 1 µPa. The third observation (2 d later) exhibited vocal modulations and modified diving involved three individuals; the survey vessel was behavior, but insufficient information is available 94 km away and measured RL was 130 dB re: on received exposures of individual whales. Stone 1 µPa. No change in sperm whale acoustic behav- (2003) compiled a large database of sighting ior was noted in any of these cases. The authors data of several mid-frequency cetacean species also played artificial codas and noticed that two observed from seismic survey vessels. Sighting whales directed their sonar beams at the speaker, rates of small odontocetes were significantly lower but insufficient information is given to associate when airguns were firing, and they were sighted at this response with a particular RL. greater distances from vessels, indicating avoid- Miller et al. (2005) documented behavioral reac- ance behavior. The study sponsors (JNCC) kindly tions of various marine mammal species, includ- provided raw data for use in our quantitative ing belugas, to airgun operations. The general avoidance analyses, but they are not included due methodology is detailed above (see the “Cell 2” to difficulties in estimating exposure RL for indi- section). Owing to their normal seasonal patterns vidual sightings. (See also Stone & Tasker, 2006, in the Beaufort Sea, belugas were most abundant for a recently published account.) in the Miller et al. (2005) study area prior to the start of seismic operations. There were relatively Laboratory Observations (Cell 5) few vessel-based sightings, most of which were Akamatsu et al. (1993) investigated avoid- made when airguns were not active. Many belugas ance behavior in two captive false killer whales were observed during aerial surveys, however, and exposed to 15 different kinds of sounds, including these data were used to compare beluga sightings pulse sequences (manual strikes on a metal pipe within concentric 10-km bands around the active once every 2 s) in the 24 to 115 kHz range. For seismic source with sighting rates in non-airgun this stimulus, no avoidance was seen following conditions. During airgun operations, Miller et al. the first exposure (174 dB re: 1 µPa), but tempo- detected significantly fewer animals 10 to 20 km rary avoidance behavior (response score: 6) was from seismic operations and an unexpectedly high observed for successive exposures at 174 and 178 number of sightings in the 20- to 30-km zone. This dB re: 1 µPa. was suggestive of behavioral avoidance of seismic Finneran et al. (2000) observed behavioral operations at distances up to 20 km. These observa- responses of two captive bottlenose dolphins and tions may in part explain why so few animals were a beluga whale during TTS experiments involving observed by shipboard marine mammal observers. a series of impulsive exposures designed to rep- Miller et al. noted that the apparent avoidance of licate distant explosions. Each animal exhibited seismic operations was much greater than expected alterations of nominal trained behaviors (reluc- if the whales were responding to non-airgun tance to return to experimental stations) during the sounds associated with vessel operation. For the experiment; the onset of behavioral disturbance 506 Southall et al. occurred in the beluga at 220 dB re: 1 µPa (peak- potential to induce avoidance behavior in pinni- to-peak) and in the two bottlenose dolphins at 196 peds, whereas RLs exceeding 190 dB re: 1 µPa are and 209 dB re: 1 µPa (peak-to-peak), respectively. likely to elicit responses, at least in some ringed In a related study, Finneran et al. (2002b) observed seals (Harris et al., 2001; Blackwell et al., 2004b; behavioral responses of a bottlenose dolphin and a Miller et al., 2005). beluga whale after exposure to impulsive sounds Harris et al. (2001) documented responses produced by a water gun. Both individuals showed of pinnipeds (primarily ringed seals, but a few a similar reluctance to return to experimental sta- bearded and spotted seals) and obtained calibrated tions (beluga at 202 dB re: 1 µPa (peak-to-peak); measures of RLs within defined spatial zones bottlenose dolphin at 229 dB re: 1 µPa [peak-to- during operation of a single airgun, an 11-airgun peak]). Romano et al. (2004) studied physiologi- array totaling 1,320 in3, and during control peri- cal responses to these exposures in these same ods. Visual observations from the seismic vessel animals. They observed clear neuro-immune were limited to the area within a few hundred responses in the beluga at exposures above 222 meters, and 79% of the seals observed were within dB re: 1 µPa (peak-to-peak) and significant differ- 250 m of the vessel. During daylight, seals were ences in aldosterone and monocyte counts in the observed at nearly identical rates with no airguns, dolphin for exposures exceeding 225 dB re: 1 µPa one airgun, or when a full airgun array was firing. (peak-to-peak). Seals were significantly further away during full array operations compared to the other two con- High-Frequency Cetaceans/Multiple Pulses ditions. Also, there was some avoidance within (Cell 8) 150 m of the vessel in these conditions (0.37 seals seen per hour in control periods compared to 0.21 Based on our source type distinction (see Chapter seals/h during full array operations). Seismic 2), virtually all sound sources used in behavioral operations were not believed to cause many, if studies of high-frequency cetaceans (e.g., acoustic any, seals to desert the operational area. harassment devices [AHDs] and acoustic deterrent Blackwell et al. (2004b) investigated behav- devices [ADDs]) would be characterized as non- ioral reactions of ringed seals to impact sounds pulses. While individual elements produced by associated with the driving of steel pipes in the some of these sources would be characterized as construction of an oil production facility. Multiple pulses, and sequences of them as multiple pulses, strikes were recorded under water at distances up they are generally emitted in such rapid fashion to 3 km from the source. Unweighted peak pres- that mammalian auditory systems are likely to sure level, SPL, and SEL measurements were perceive them as nonpulses. Further, some AHDs, made at various distances. At the closest point (63 ADDs, and all other sources used in behavioral m), RLs were 151 dB re: 1 µPa (RMS), 157 dB re: studies with high-frequency cetaceans lack the 1 µPa (peak), and 145 dB re: 1 µPa2-s (SEL). Pulses characteristics of pulses. Due to the lack of data, it had measurable components extending to over 10 is not possible to present any behavioral response kHz, although more than 95% of the energy in the data on multiple pulses for high-frequency ceta- signals was below 225 Hz. A frequency-weight- ceans; available data for nonpulse sounds are ing metric somewhat similar to that proposed here considered elsewhere (see the “High-Frequency was applied to the recorded signals in estimating Cetaceans/Nonpulses [Cell 9]” sections of Chapter 4 audibility ranges. Individuals demonstrated no and Appendix C). We note the need for behavioral or low-level behavioral responses to pile-driving research on these animals using sound sources sounds, but were somewhat responsive to helicop- unequivocally classified as pulses. ter overflights. Blackwell et al. noted, however, that their data were collected after a prolonged Pinnipeds in Water/Multiple Pulses (Cell 11) period of intensive construction activity and may reflect the least responsive part of the original Information on behavioral reactions of pinnipeds population of seals that may have already habitu- in water to multiple pulses is derived from studies ated to the noise source. Individual observations using small explosives similar to those used in fish- in which helicopters were not present are consid- eries interactions, construction activity, and seis- ered in our behavioral analysis, weighted by the mic surveys. Several studies lacked matched data total number of relevant observations (Table 11). on acoustic exposures and behavioral responses by Aerial measurements of multiple pulse exposures individuals. As a result, the quantitative informa- were also obtained in this study and are consid- tion on reactions of pinnipeds in water to multiple ered in the relevant condition below. pulses is very limited. Our general finding is that Miller et al. (2005) documented behavioral exposures in the ~150 to 180 dB re: 1 µPa range reactions of various marine mammal species, (RMS over pulse duration) generally have limited including pinnipeds in water, to seismic airgun Marine Mammal Noise Exposure Criteria 507 operations. The general methodology is detailed on visual observations. Thus, future studies may above (see the “Cell 2” section). The vast major- show some seals to be more responsive to multiple ity (> 90%) of the seals were ringed seals and pulses than Table 11 would suggest. the remainder were bearded seals. Vessel-based observers saw seals around the vessel, and often Pinnipeds in Air/Multiple Pulses (Cell 11) quite close to it, throughout the period of seismic operations. Seals were observed significantly The effects of multiple aerial pulses on pinnipeds further away during airgun operations in the first are among the least well-documented of the condi- summer, whereas the reverse pattern was actu- tions we considered. Most of the available data on ally the case in the second season. Combined, the responses to pulses are from single-pulse events results suggest essentially no observable behav- (e.g., rocket launches) over populations of pin- ioral response in pinnipeds exposed to seismic nipeds exposed to such signals repeatedly (e.g., signals in these specific conditions. Based on the Thorson et al., 1998, 1999, 2000a, 2000b; Berg acoustic measurements that were conducted and et al., 2001, 2002, 2004). These launches are not the areas in which these pinnipeds were observed, repeated so frequently that the exposure can be RLs were likely 170 to 200 dB re: 1 µPa (RMS considered as involving multiple pulses, and many over pulse duration). A single observation of no of the exposures include nonpulse components. reaction (response score: 0) for pinnipeds in water However, they are discussed in some detail in this is reported for this study and is weighted equally appendix (as well as in Appendix C for nonpulses across these exposure RL bins (Table 8). within these studies) along with several other stud- Several other studies were deleted from our ies potentially relevant to Cell 14 but ultimately analysis due to a lack of certain information. not used in the analysis here. Consequently, the Two studies investigated small firecracker-like quantitative information analyzed for reactions of explosives (called “seal bombs”) and their effect pinnipeds in air exposed to multiple pulses (see on the underwater behavior of pinnipeds around Table 12 for summary and Table 13 for severity fishing gear (Shaughnessy et al., 1981; Mate & scaling analysis) focuses on the aerial data of Harvey, 1987). Initially, these explosives tend to Blackwell et al. (2004b). These extremely limited induce the desired avoidance behavior, but this data suggest very minor, if any, observable behav- response fades quickly due to habituation (see ioral responses for exposures ranging from 60 to Richardson et al., 1995). Mate & Harvey (1987) 80 dB re: 20 µPa. reported fairly extensive descriptions of startle Blackwell et al. (2004b) reported behavioral and temporary avoidance data as well as some reactions of ringed seals to aerial impact sounds information on exposure conditions. Besides the from pile-driving (described above). Multiple challenging matter of interpreting the apparently strikes were recorded in air at distances up to rapid habituation to this sound source, however, 500 m from the source. Unweighted SPL, peak data are lacking that relate discrete exposures with sound pressure levels, and SEL measurements defined behavioral responses of specific individual were made at various distances. At the closest point pinnipeds. For these reasons, we excluded data (63 m) average RLs were 93 dB re: 20 µPa (RMS), on responses to seal bombs from our analysis. 111 dB re: 20 µPa (peak), and 87 dB re: (20 µPa)2- Moulton et al. (2003, 2005) conducted surveys s (SEL). Mean pulse durations were between of ringed seal distribution before and during the 0.17 and 0.63 s, with measurable energy to over construction and operation of the same oil produc- 10 kHz, but with 95% of the energy occurring tion facility described by Blackwell et al. (2004a, between 89 and 3,534 Hz. A frequency-weight- 2004b). Sound sources included nonpulse as well ing metric somewhat similar to that proposed here as multiple pulse sources (including impact pile- was applied to the recorded signals in estimat- driving). Their observations across multiple sea- ing audibility ranges. Individuals demonstrated sons indicated little or no behavioral avoidance of very limited behavioral responses to pile-driving the area in response to various industrial activi- sounds in some conditions (most appeared either ties. Due to difficulties with control observations “indifferent or curious”) but were more respon- across seasons and the lack of information about sive to helicopter overflights. Data were collected discrete exposures and individual reactions, how- after prolonged construction activities, and some ever, we excluded the Moulton et al. (2003, 2005) habituation probably had taken place already. data from our analysis. A final study for which Individual observations for which helicopters available data were insufficient for inclusion here were not present are considered in the behavioral is Thompson et al. (1998). That telemetry study analysis here and weighted by the total number of seemed to show much higher responsiveness of relevant observations (Table 13) to equal a single gray and harbor seals to airgun sounds than has observation for the study. been demonstrated in other studies, which relied 508 Southall et al.

Perry et al. (2002) measured the effects of repeated (0 to 5/d) sonic booms from Concorde aircraft on harbor and gray seals on Sable Island, Nova Scotia. They measured the number of animals on shore before and after booms as well as the frequency of various behaviors. Additionally, they compared heart rates in exposure and control conditions using recording devices deployed on the animals. They reported received sound overpressure of booms on the breeding beaches of both pinniped species. Observed effects on animal presence, behavior, and heart rate were generally minor and not statistically significant; animals were largely tolerant of the sounds but became somewhat more alert following them. However, Perry et al. (2002) note that there is a long history of sonic booms from aircraft in the area and the animals are likely habituated to their presence. Due to this complication and the lack of explicit received SPL measures at exposed individuals, we did not score the results of Perry et al. (2002) here. Aquatic Mammals 2007, 33(4), 509-521, DOI 10.1578/AM.33.4.2007.509

Appendix C. Studies Involving Marine Mammal Behavioral Responses to Nonpulses

Low-Frequency Cetaceans/Nonpulses investigation yielded the general conclusion that (Cell 3) RLs exceeding 120 dB re: 1 µPa induced demonstrable behavioral reactions (avoidance). Malme While there are clearly major areas of uncertainty et al. (1984) calculated 10%, 50%, and 90% remaining, there has been relatively extensive probabilities of gray whale avoidance reactions behavioral observation of low-frequency ceta- in these conditions as 110, 120, and 130 dB re: ceans exposed to nonpulse sources. As sum- 1 µPa. Malme et al. (1986) observed the behavior marized in Table 14, these field observations of feeding gray whales during four experimental involve the majority of low-frequency cetacean playbacks of drilling sounds (50 to 315 Hz; 21- species exposed to a wide range of industrial, min overall duration and 10% duty cycle; source active sonar, and tomographic research active levels 156 to 162 dB re: 1 µPa-m). In two cases sources (Baker et al., 1982; Malme et al., 1983, for RLs 100 to 110 dB re: 1 µPa, there was no 1984, 1986; Richardson et al., 1990b; McCauley observed behavioral reaction. Avoidance behavior et al., 1996; Frankel & Clark, 1998; Borggaard was observed in two cases where RLs were 110 to et al., 1999; Biassoni et al., 2000; Croll et al., 120 dB re: 1 µPa. 2001; Palka & Hammond, 2001; Nowacek et al., Richardson et al. (1990b) performed 12 play- 2004). Observations from several related studies back experiments in which bowhead whales in the (Dahlheim, 1987; Frankel & Clark, 2000, 2002; Alaskan Arctic were exposed to drilling sounds. Schick & Urban, 2000; Moore & Clarke, 2002; Low-frequency source characteristics and trans- Jahoda et al., 2003; Mobley, 2005) were reviewed mission loss were well-characterized, enabling briefly but not analyzed here because key infor- RL estimates to be made for individual cases. mation was lacking. Whales generally did not respond to exposures in These papers generally indicate no (or very the 100 to 130 dB re: 1 µPa range, although there limited) responses at RLs 90 to 120 dB re: 1 µPa was some indication of minor behavioral changes and an increasing probability of avoidance and in several instances. other behavioral effects in the 120 to 160 dB re: Using different detection and sampling tech- 1 µPa range (Table 14). However, the data also niques, McCauley et al. (1996) reported several indicate considerable variability in RLs associ- cases of humpback whales responding to vessels ated with behavioral responses. Contextual vari- in Hervey Bay, Australia, along with measure- ables (e.g., source proximity, novelty, operational ments of noise RL. Not all cases reported provided features) appear to have been at least as important sufficient information to associate a response or as exposure level in predicting response type and lack of response with exposure, but in three cases, magnitude. individual responses and noise RL were reported. Baker et al. (1982) investigated behavioral Results indicated clear avoidance at RLs between responses of individual humpback whales to vessel 118 to 124 dB re: 1 µPa. traffic in southeast Alaska. Individual RLs were Palka & Hammond (2001) analyzed line transect not reported, but sufficient information regarding census data in which the orientation and distance individual ranges was obtained to approximate off transect line were reported for large numbers of exposures given that the acoustic characteristics minke whales. General additive models were used of identical classes of vessel classes involved to estimate the range at which cetaceans respond were measured in similar conditions by Miles & to the noise of the research vessel by approach or Malme (1983). Results indicate some behavioral avoidance. The typical avoidance distance for 272 avoidance when RL was in the 110 to 120 dB re: minke whales in the Gulf of Maine was 717 m; 1 µPa range and clear avoidance at 120 to 140 dB for 352 minke whales in the North Sea, it was 563 re: 1 µPa. m; and for 493 minke whales in the Northeastern Malme et al. (1983, 1984) used playback meth- Atlantic, it was 695 m. Received levels were esti- ods to document behavioral reactions of migrat- mated based on a nominal source level for that ing gray whales to intermittent sounds of heli- class of research vessel (ca. 170 to 175 dB re: copter overflights and continuous sounds from 1 µPa-m) and an assumption of spherical (20 log drilling rigs and platforms. Both phases of the R) spreading loss (54 dB loss @ 500 m; likely 510 Southall et al. reasonable for these conditions). These data are for which 12 to 30% of exposures exceeded 140 represented in Table 14 by the 110 to 120 dB re: dB re: 1 µPa. A single observation was scored for 1 µPa exposures and a relatively low (less severe) this study because individual responses were not behavioral response score of three (i.e., minor reported. changes in locomotion speed, direction, and/or Frankel & Clark (1998) conducted playback diving profile). experiments with wintering humpback whales Several additional studies have used playback around the Big Island of Hawai’i. The sound source experiments with active sound sources to investi- was a single speaker producing a low-frequency gate the behavioral reactions of low-frequency ceta- “M-sequence” (sine wave with multiple-phase ceans to nonpulse sources. Biassoni et al. (2000) and reversals) signal in the 60 to 90 Hz band. This was Miller et al. (2000) report behavioral observations similar in bandwidth to the ATOC source, but had for humpback whales exposed to a low-frequency a much lower output level (172 dB re: 1 µPa @ sonar stimulus (160- to 330-Hz frequency band; 1 m). A vertical line array of calibrated hydrophones 42-s tonal signal repeated every 6 min; source levels was deployed from a spar buoy to measure received 170 to 200 dB re: 1 µPa-m). Measured RLs ranged signals in situ. Detailed observations of many from 120 to 150 dB re: 1 µPa. In nine cases, indi- behavioral patterns (including respiration, diving, vidual whales continued singing throughout expo- and general movements) were recorded before, sures, while in four instances, individuals ceased during, and after playback (n = 50) and control calling when they joined another whale. The cessa- (n = 34) sequences. A single trial also involved tion of song and joining another individual is typi- playback of humpback foraging sounds. Most of cal of normal mysticete social interactions (Tyack, the playback sequences involved very low-level 1981). Consequently, these events were not scored RLs, ca. 90 to 120 dB re: 1 µPa, though not speci- as a vocal response to the playback but as a mod- fied in sufficient detail to include in the analysis erate orienting behavior (severity score = 2). For here. For 11 playbacks, exposures were between the remaining five playbacks, individual whales 120 and 130 dB re: 1 µPa and included suffi- stopped singing during exposure without joining cient information regarding individual responses. other whales (severity scale = 4). Although singers During eight of the trials, there were no measur- also stop spontaneously under control conditions, able differences in tracks or bearings relative to the latter five experimental trials were considered control conditions, whereas on three occasions, vocal cessation resulting from sound exposure whales either moved slightly away from (n = 1) or (Biassoni et al., 2000). However, there are insuf- towards (n = 2) the playback speaker during expo- ficient data to compare control and experimental sure. Because it was not possible to determine cases for spontaneous rates of cessation. Analysis whether the same individual whales were repre- of all singers indicated an increase in song dura- sented more than once in the playback sequences, tion during exposure due to increased repetition a single observation was recorded for Frankel & of elements of the song. Since it was possible that Clark (1998), with 0.73 of this observation (8/11) some individual whales were represented multiple scored as a 0 (no response) and 0.27 (3/11) scored times within the playbacks, the Biassoni et al. as a 3 (minor changes in locomotion speed, direc- (2000) and Miller et al. (2000) data were scored as tion, and/or diving). A final important observation a single behavioral observation. The 18 individual from the detailed statistical analysis by Frankel & observations were weighed inversely by the total Clark was that the presence of the source vessel number (1/18) in Table 15. itself had a greater effect than did the M-sequence Croll et al. (2001) investigated responses of for- playback. aging fin and blue whales to the same LFA sonar Finally, Nowacek et al. (2004) used controlled stimulus off southern California. Unlike the pre- exposures to demonstrate behavioral reactions of vious two studies, where individual experimental northern right whales to various nonpulse sounds. subjects were tracked on a behavioral scale, this Playback stimuli included ship noise; social study used sighting data on an ecological scale. sounds of conspecifics; and a complex, 18-min Playbacks and control intervals with no trans- “alert” sound consisting of repetitions of three mission were used to investigate behavior and different artificial signals (alternating 1-s pure distribution on time scales of several weeks and tones [500 and 850 Hz]; a 2-s, tonal, frequency spatial scales of tens of kilometers. Sightings and downsweep [4,500 to 500 Hz]; and a pair of 1- whale diving behavior were not random but were s pure tones [1,500 Hz and 2,000 Hz] amplitude related to environmental features such as the con- modulated at 120 Hz). A total of ten whales were tinental shelf break and its effects on prey abun- tagged with calibrated instruments that measured dance rather than operation and location of the received sound characteristics and concurrent nonpulse sonar source. The general conclusion animal movements in three dimensions. Five out was that whales remained feeding within a region of six exposed whales reacted strongly to alert Marine Mammal Noise Exposure Criteria 511 signals at measured RLs between 130 and 150 dB & Clark (2000) observed whales from a land station re: 1 µPa (i.e., ceased foraging and swam rapidly and determined that the average distance between to the surface; severity scale = 7). Two of these the sound source and the whale groups sighted was individuals were not exposed to ship noise and significantly greater during source operation. These are given as a discrete observation in Table 15, and other data were also considered in the context whereas the other four were exposed to both stim- of other factors affecting humpback whale distribu- uli and thus weighted as 0.5 (1/2) observations for tion off the island of Kaua’i. Mobley (2005) con- the respective RL and severity score. These whales ducted aerial surveys in each of three years (2001, reacted mildly to conspecific signals (not scored source off; 2002 & 2003, source on) during the here because of biological signals). Seven whales, peak season of humpback residency. Abundance including the four exposed to the alert stimulus, and distribution of whales were very similar in the had no measurable response to either ship sounds area surrounding the source over all three years; or actual vessel noise. This study by Nowacek small differences in sighting rates, sighting loca- et al. included the careful experimental design, tion depth, and distances from the source and shore controls, and detailed information on exposure and were not statistically significant. Frankel & Clark individual behavioral response that were required (2002) and Mobley (2005) lack explicit data on for behavioral analysis. More studies of this type RLs associated with individual behavioral observa- and rigor are urgently needed (see Chapter 5). tions, which precludes their inclusion here. We reviewed additional studies concerning low- frequency cetaceans and nonpulse sounds but did Mid-Frequency Cetaceans/Nonpulses (Cell 6) not include them in the analysis here, generally due A relatively large number of mid-frequency to the absence of key information. Dahlheim (1987) cetaceans have been observed in the field and in exposed gray whales to playbacks of outboard the laboratory responding to nonpulse sounds, noise, gray whale calls, and tonal sounds. Whales including vessels and watercraft (LGL & significantly increased calling rate and modified Greeneridge, 1986; Gordon et al., 1992; Palka call structure for sources other than the test tone & Hammond, 2001; Buckstaff, 2004; Morisaka (the latter caused all vocalization to cease). During et al., 2005), pulsed pingers and ADD/AHDs and following longer duration playbacks of oil (Watkins & Schevill, 1975; Morton & Symonds, drilling and killer whale sounds with more precise 2002; Monteiro-Neto et al., 2004), industrial tracking of gray whale locations, individuals spent activities (Awbrey & Stewart, 1983; Richardson more time milling, and whales remained farther off- et al., 1990b), mid-frequency active military shore during killer whale playbacks. Unfortunately, sonar (NRL, 2004a, 2004b; NMFS, 2005), and insufficient information is presented to associate tones or bands of noise in laboratory conditions changes with specific RLs. Borggaard et al. (1999) (Nachtigall et al., 2003; Finneran & Schlundt, measured the effects of industrial activity on several 2004). Summary information on these stud- mysticete species in Newfoundland, but insufficient ies is given in Table 16. As in other conditions, information is reported on individually discernible a number of potentially relevant field studies are responses. Schick & Urban (2000) applied statisti- not included in the severity scaling analysis due to cal methods to assess spatial avoidance of active lack of sufficiently detailed information. drilling rigs by bowhead whales, but no acoustic An additional challenge in interpreting many of data are reported. Moore & Clarke (2002) synthe- the field data for this condition is isolating the effect sized previously published data (all considered of RL from the effects of mere source presence (as separately above) on numerous nonpulse sources, possibly indicated by visual stimuli or other aspects in order to assess the avoidance probability of gray of acoustic exposure such as the presence of high- whales for various exposure RLs. Jahoda et al. frequency components) and other contextual vari- (2003) studied individual responses of fin whales ables. For this reason, several studies were consid- (n = 25) to close rapid approaches of small vessels; ered but not integrated into the analysis. 18 observations included control and experimental The laboratory observations are of captive ceta- data. Clear behavioral responses were observed, ceans exposed to precisely controlled and known but neither RL nor range from source to individu- noise exposures in the context of hearing and TTS als were given. Results are further complicated by experiments. The relevance of behavioral reac- whale tagging attempts from the vessel. Frankel tions of trained, food-reinforced captive animals & Clark (2000) and Mobley (2005) investigated exposed to noise in assessing reactions of free-rang- the distribution of humpback whales in Hawai’i in ing marine mammals is not well-known, however relation to the operation of a low-frequency tomo- (discussed below). graphic source (ca. 75 Hz; 37.5-Hz nominal band- The combined field and laboratory data for width; 20-min duration every 2 h during daylight mid-frequency cetaceans exposed to nonpulse hours; source level: 195 dB re: 1 µPa-m). Frankel sounds do not lead us to a clear conclusion about 512 Southall et al.

RLs coincident with various behavioral responses measurements on 28 June 1982. Four groups of (see severity scaling, Table 17). In some settings, narwhals (n = 9 to 10, 7, 7, and 6) responded when individuals in the field showed profound (and the ship was 6.4 km away (exposure RLs of ~100 what we regard here as significant) behavioral dB re: 1 µPa in the 150- to 1,150-Hz band). At responses to exposures from 90 to 120 dB re: a later point, observers sighted belugas moving 1 µPa, while others failed to exhibit such responses away from the source at > 20 km (exposure RLs for exposure RLs from 120 to 150 dB re: 1 µPa. of ~90 dB re: 1 µPa in the 150- to 1,150-Hz band). Contextual variables other than exposure RL, and The total number of animals observed fleeing was probable species differences, are the likely reasons about 300, suggesting approximately 100 inde- for this variability. Context, including the fact that pendent groups (of three individuals each), which captive subjects were often directly reinforced is the sample size used here. No whales were with food for tolerating noise exposure, may also sighted the following day, but some were sighted explain why there is great disparity in results on 30 June, with ship noise audible at spectrum from field and laboratory conditions—exposures levels of approximately 55 dB re 1 µPa/Hz (up to in captive settings generally exceeded 170 dB re: 4 kHz). 1 µPa before inducing behavioral responses. Observations during 1983 (LGL & Greeneridge, 1986) involved two ice-breaking ships with aerial Field Observations (Cell 6) survey and ice-based observations during seven The most extensive series of observations regard- sampling periods. As the first vessel approached ing vessels and watercraft is from LGL and at a distance of about 65 km, ice-based observ- Greeneridge (1986) and Finley et al. (1990), who ers noted reactions from both narwhals (seven documented belugas and narwhals (Monodon groups) and belugas (eight groups) (exposure monoceros) congregated near ice edges reacting RLs of ~101 to 105 dB re: 1 µPa in the 20- to to the approach and passage of ice-breaking ships. 1,000-Hz band). After 22 h without operation, the Over a 3-y period (1982 to 1984), they used both vessel commenced ice-breaking, and a second ice-based local observations of whales and aerial icebreaker approached (exposure RLs of ~120 surveys, and also made detailed acoustic measure- dB re: 1 µPa in the 20- to 1,000-Hz band). This ments. The survey method made it difficult to resulted in the rapid movement of > 225 belugas assess independent groups of animals. Some large- (estimated as a sample size of 75 for this analy- scale groupings could be identified for several dif- sis); belugas were neither seen nor heard for ferent “disturbance” periods, however. Pre-dis- the remainder of the second observation period. turbance group size was ~3; we divided reported Behavioral responses were also observed for numbers of disturbed “herds” by three to estimate 10 groups of narwhals. A total of 73 narwhals the number of independent groups. Aerial surveys were seen and/or heard, but their reactions are in 1984 lumped sightings by minute, which cor- not clearly reported and are thus excluded from responded to about 3.4 km in distance. We consid- analysis here. At the onset of the third sampling ered this distance sufficient to treat each minute as period, following a 4.5-h silent interval, four an independent unit for avoidance analysis. The narwhal groups were observed in nominal social responses of both species over a 3-y period were behavior (diving and vocalizing). An ice-breaking generally similar to responses they make to preda- vessel operated intermittently, but no change was tors as described by Inuit hunters. observed in narwhal behavior. Belugas in the area Beluga whales responded to oncoming ves- did modify vocalization parameters during opera- sels by (1) fleeing at speeds of up 20 km/h from tions (exposure RLs of ~116 dB re: 1 µPa in the distances of 20 to 80 km, (2) abandoning normal 20- to 1,000-Hz band). A 6-h quiet interval was pod structure, and (3) modifying vocal behavior followed by 10.5 h of ice-breaking operation, but and/or emitting alarm calls. Narwhals, in contrast, bad weather precluded animal observations. After generally demonstrated a “freeze” response, lying an additional 9-h hiatus, ice-breaking commenced motionless or swimming slowly away (as far as again by both vessels (exposure RLs of ~121 dB 37 km down the ice edge), huddling in groups, re: 1 µPa in the 20- to 1,000-Hz band). Ice-based and ceasing sound production. There was some observers documented 14 narwhals and 11 belu- evidence of habituation and reduced avoidance 2 gas leaving the area, and aerial surveys indicated to 3 d after onset. Due to the detailed and exten- 80% of 673 belugas moving away from sound sive nature of these observations, data from each sources (estimated number of groups calculated season, and how they are interpreted here, are as [.8]*[673/3] = 179.5). As noise levels from ice- given in detail. breaking operations diminished, a total of 45 nar- The 1982 season observations by LGL & whals returned to the area and engaged in diving Greeneridge (1986) involved a single passage and foraging behavior. The sixth observation of an icebreaker with both ice-based and aerial period followed 6.5 h without a vessel in the area, Marine Mammal Noise Exposure Criteria 513 during which 30 belugas (estimated as 10 groups) The white-sided dolphins exhibited simple avoid- and 15 narwhals (estimated as five groups) were ance behavior (as indicated by their orientations) observed diving in the area (exposure RLs of out to an estimated range of 592 m based on 85 ~105 dB re: 1 µPa in the 20- to 1,000-Hz band). group sightings (n > 1). White-beaked dolphins A single beluga vocal response was noted at RL actually approached vessels between 150 and 300 = 116 dB re: 1 µPa in the 20- to 1,000-Hz band. m away, but demonstrated avoidance at distances Aerial surveys indicated dense concentrations of of 300 to 700 m. Typical avoidance distance was narwhals (n = 50) and belugas (n = 400) appar- estimated as 716 m based on 48 groups sighted. ently foraging well away from the disturbance site. Buckstaff (2004), using repeated samples During the final sampling period, following an of the behavior of 14 individual bottlenose dol- 8-h quiet interval, no reactions were seen from 28 phins, observed 1,233 vessel approaches (within narwhals and 17 belugas (exposure RLs ranging 400 m) near Sarasota, Florida. Dolphin whistle up to 115 dB re: 1 µPa). rates became elevated before vessel noise was The final season (1984) reported in LGL & detectable to the researcher listening via towed Greeneridge (1986) involved aerial surveys before, hydrophones. Vessel RLs measured near dolphin during, and after the passage of two ice-breaking subjects ranged from 113 to 138 dB re: 1 µPa. ships. The lack of ice camps precluded acoustic Dolphin vocal responses were observed before measurements as well as behavioral observations. vessel sounds were audible, and apparently A preliminary survey was conducted the day occurred with RLs in the 110 to < 120 dB re: 1 before operations, and an additional aerial survey µPa category. was conducted as both ships commenced operat- Morisaka et al. (2005) compared whistles from ing. During operations, no belugas and few nar- three populations of Indo-Pacific bottlenose dol- whals were observed in an area approximately 27 phins (Tursiops aduncus). One population was km ahead of the vessels, and all whales sighted exposed to vessel noise with spectrum levels of over 20 to 80 km from the ships were swimming ~85 dB re: 1 µPa/Hz in the 1- to 22-kHz band strongly away. Additional observations confirm (broadband RL ~128 dB re: 1 µPa) as opposed to the remarkable spatial extent of avoidance reac- ~65 dB re: 1 µPa/Hz in the same band (broadband tions to this sound source in this context. In the RL ~108 dB re: 1 µPa) for the other two sites. absence of acoustic measurements, however, it Dolphin whistles in the noisier environment had was necessary to estimate RLs from the detailed lower fundamental frequencies and less frequency data from the same ice-breaking vessel during the modulation, suggesting a shift in sound param- previous season. eters as a result of increased ambient noise. Behavioral responses at fairly low exposure Morton & Symonds (2002) used census data RLs are suggested by studies of some other mid- on killer whales in British Columbia to evaluate frequency cetaceans as well. Gordon et al. (1992) avoidance of nonpulse AHD sources. They con- conducted opportunistic visual and acoustic moni- sidered unusually long time scales, comparing toring of sperm whales in New Zealand exposed pre-exposure data from 1985 to 1992, exposure to nearby whale-watching boats (within 450 m). from 1993 to 1998, and post-exposure from 1999 Individuals could not be used as the units of to 2000. The response data were simply pres- analysis because it was difficult to re-sight spe- ence or absence, making it difficult to assess RLs. cific individuals during both exposure and control Using some monitoring and reasonable assump- conditions. Sperm whales respired significantly tions, however, they estimated audibility ranges less frequently, had shorter surface intervals, and throughout the complex study area. Avoidance took longer to start clicking at the start of a dive ranges were ca. 4 km. Also, there was a dramatic descent when boats were nearby than when they reduction in the number of days “resident” killer were absent. Noise spectrum levels of whale- whales were sighted during AHD-active periods watching boats ranged from 109 to 129 dB re: compared to pre- and post-exposure periods and 1 µPa/Hz. Over a bandwidth of 100 to 6,000 Hz, a nearby control site. Morton & Symonds did not equivalent broadband source levels are ~157 dB indicate how many pods were involved in their re: 1 µPa-m; RLs at a range of 450 m are ~104 dB analysis. Consequently, we assume a single inde- re: 1 µPa. pendent group in our analysis. Palka & Hammond (2001) applied a General Monteiro-Neto et al. (2004) studied avoidance Additive Model to line transect data to estimate responses of tucuxi (Sotalia fluviatilis) to Dukane® the range at which mid-frequency cetaceans typi- Netmark ADDs. Source characteristics are not cally responded to the noise of research vessels. given, but identical devices were used by Culik et The subjects were Atlantic white-sided dolphins al. (2001), and acoustic parameters are reported in in the Gulf of Maine and white-beaked dolphins detail there (and in the “Cell 9” section). In a total (Lagenorhynchus albirostris) in the North Sea. of 30 exposure trials, ~5 groups each demonstrated 514 Southall et al. significant avoidance compared to 20 pinger off al., 2003; Hastie et al., 2003; Lusseau, 2003; Foote and 55 no-pinger control trials over two quad- et al., 2004; Scheifele et al., 2005) were reviewed rats of about 0.5 km2. Neither avoidance range in detail. The results were excluded from Table 17 nor RLs are given, but based upon a central dis- due to limited or no information on animal num- tance from the quadrat of 10 m, and assuming 15 bers and/or location relative to the source, acous- log R transmission loss in this shallow environ- tic properties of sources, propagation variables, or ment (water depth 1 to 5 m), estimated exposure received exposure conditions. The general obser- RLs were ~115 dB re: 1 µPa. vations of each study are given here briefly. Hastie The only specific situation involving exposure et al. (2003) documented increased swimming of wild marine mammals to active mid-frequency and diving synchrony of bottlenose dolphins off military sonar for which exposure conditions are northern Scotland in the presence of vessel traf- known with any degree of specificity involved fic. Lusseau (2003) observed effects on behavior incidental exposure of killer whales to sounds of New Zealand bottlenose dolphins within 400 m from the naval vessel USS Shoup (NRL, 2004a, of boats. Chilvers & Corkeron (2001) considered 2004b; NMFS, 2005). A group (J-pod) of south- differences in behavior of bottlenose dolphins that ern resident killer whales in the eastern Strait of do and do not forage around trawlers. Williams Juan de Fuca and Haro Strait, Washington, was et al. (2002) observed that some killer whales adopt observed by researchers before, during, and after erratic movement patterns, suggestive of avoid- the approaching USS Shoup transmitted sonar sig- ance, when whale-watching vessels accelerate to nals from its 53C sonar at a source level of ca. 235 intersect the whale’s course. RLs of vessel sound dB re: 1 µPa-m once every ca. 28 s for several increased approximately 14 dB with increased hours. At its point of closest approach, the mean speed associated with leapfrogging. Bordino et al. direct-path RL within a specified area around the (2002) determined that ADDs were initially effec- animals was ca. 169 dB re: 1 µPa (NRL, 2004a, tive at reducing by-catch of Franciscana dolphins 2004b). As indicated by NMFS (2005), there is in Argentine subsistence gillnet fisheries. Cox some discrepancy in interpretation of the behav- et al. (2003) investigated reactions of bottle- ioral responses among researchers who were nose dolphins to Dukane® NetMark 1000 ADDs either on the water or who observed video record- attached to commercial gillnets and found very ings of behavioral responses. The lead researcher limited to no behavioral avoidance. A group of following and observing the animals during the long-finned pilot whales (Globicephala melas) event indicated that individuals in the group dem- demonstrated significant elevations of whistle onstrated abnormal avoidance behavior, most rates following each exposure to mid-frequency dramatically at the point of closest approach. military sonar reported to be at a “high” level However, the behavior of the whales apparently (Rendell & Gordon, 1999). returned to normal within a short period following Finally, two recent papers deal with important cessation of sonar transmissions. A severity score issues relating to changes in marine mammal vocal of 6 (mild/moderate avoidance) is subsequently behavior as a function of variable background reported in the 160 to 170 dB re: 1 µPa bin for this noise levels. Foote et al. (2004) found increases in single observation of the group. the duration of killer whale calls over the period Awbrey & Stewart (1983) played back semi-sub- 1977 to 2003, during which time vessel traffic in mersible drillship sounds (source level: 163 dB re: Puget Sound, and particularly whale-watching 1 µPa-m) to belugas in Alaska. They reported avoid- boats around the animals, increased dramatically. ance reactions at 300 and 1,500 m and approach Scheifele et al. (2005) demonstrated that belugas by groups at a distance of 3,500 m (RLs ~110 to in the St. Lawrence River increased the levels of 145 dB re: 1 µPa over these ranges assuming a their vocalizations as a function of the background 15 log R transmission loss). Similarly, Richardson noise level (the “Lombard Effect”). (See also et al. (1990b) played back drilling platform Parks et al., 2007, for a related new paper on mys- sounds (source level: 163 dB re: 1 µPa-m) to belu- ticetes.) These papers demonstrate some degree of gas in Alaska. They conducted aerial observations plasticity in the vocal signal parameters of marine of eight individuals among ~100 spread over an mammals in response to the ambient condition area several hundred meters to several kilometers (likely affected by the presence of human sound from the sound source and found no obvious reac- sources). These studies were not particularly tions. Moderate changes in movement were noted amenable to the kind of analysis conducted in the for three groups swimming within 200 m of the severity scaling. We note the particular impor- sound projector. tance of direct measurements of noise impacts on A number of additional studies (Rendell & marine mammal vocalization and communication Gordon, 1999; Chilvers & Corkeron, 2001; systems. Bordino et al., 2002; Williams et al., 2002; Cox et Marine Mammal Noise Exposure Criteria 515

Laboratory Observations (Cell 6) Dall’s porpoises do not (Richardson et al., 1995). Several researchers conducting laboratory experi- Whether this apparently high degree of behavioral ments on hearing and the effects of nonpulse sensitivity by harbor porpoises to anthropogenic sounds on hearing in mid-frequency cetaceans sounds extends to other high-frequency cetacean have reported concurrent behavioral responses. species (or to nonpulse sources other than ADDs, Nachtigall et al. (2003) reported that noise expo- AHDs, and boats) is unknown. However, given sures up to 179 dB re: 1 µPa and 55-min duration the lack of information to the contrary, such a affected the trained behaviors of a bottlenose dol- relationship should be assumed as a precautionary phin participating in a TTS experiment. Finneran measure. & Schlundt (2004) provided a detailed, compre- Habituation to sound exposure was noted in hensive analysis of the behavioral responses of some but not all studies. In certain field condi- belugas and bottlenose dolphins to 1-s tones (RLs tions, strong initial reactions of high-frequency 160 to 202 dB re: 1 µPa) in the context of TTS cetaceans at relatively low RLs appeared to wane experiments. Romano et al. (2004) investigated the rather rapidly with repeated exposure (Cox et al., physiological responses of a bottlenose dolphin 2001). In contrast, several laboratory observations and a beluga exposed to these tonal exposures and showed little or no indication of reduced behav- demonstrated a decrease in blood cortisol levels ioral sensitivity as a function of exposure experi- during a series of exposures between 130 and 201 ence (Kastelein et al., 1997, 2005). dB re: 1 µPa. Collectively, the laboratory observations suggested the onset of behavioral response Field Observations (Cell 9) at higher RLs than did field studies (see Table 16). Kraus et al. (1997) found (and Barlow & Cameron, The differences were likely related to the very dif- 2003, later confirmed) that ADDs can affect by- ferent conditions and contextual variables between catch rates of harbor porpoises in commercial untrained, free-ranging individuals vs laboratory fishing applications. Kraus et al. (1997) found that subjects that were rewarded with food for tolerat- nets with Dukane® pingers (10-kHz fundamental ing noise exposure. frequency, 300-ms duration, 132 dB re: 1 µPa source level) were sufficiently avoided that sig- High-Frequency Cetaceans/Nonpulses (Cell 9) nificantly fewer porpoises were entangled than in nets lacking pingers. Their observations suggest Numerous controlled studies have been con- an ADD avoidance range of at least 10 m (expo- ducted recently on the behavioral reac- sure RL ~110 dB re: 1 µPa) but are not explicit tions of high-frequency cetaceans to vari- enough in documenting exposure conditions or ous nonpulse sound sources both in the field individual responses to include in the behavioral (Culik et al., 2001; Johnston, 2002; Olesiuk scoring analysis here. et al., 2002) and in laboratory settings (Kastelein Culik et al. (2001) conducted behavioral et al., 1997, 2000, 2005, 2006a). However, only observations of groups of harbor porpoises near one high-frequency species (harbor porpoise) has Vancouver Island before, during, and after the been extensively studied. For that species, suf- removal of a PICE pinger (eight different wide- ficient data are available to estimate behavioral band swept frequency signals between 20 and response magnitude vs received exposure condi- 160 kHz; 300-ms duration at random intervals tions. The original studies were attempts to reduce [5 to 30 s]; max. broadband SL = 145 dB re: harbor porpoise by-catch by attaching warning 1 µPa @ 1 m). Source characteristics of the alarm pingers to fishing gear. More recent studies con- were known, but propagation measurements were sider whether ADDs and AHDs also exclude not made in situ. Exposure RLs are estimated here harbor porpoises from critical habitat areas and based on source characteristics and assumptions whether these devices affect harbor porpoise regarding propagation, allowing for measures behavior in controlled laboratory conditions. of similar sources in similar conditions. A large The combined wild and captive animal data exclusion zone of approximately 530-m radius (summarized in Table 18) clearly support the surrounding active acoustic alarms was observed observation that harbor porpoises are quite sensi- (corresponding to exposure RLs of ~90 to 100 dB tive to a wide range of human sounds at very low re: 1 µPa). Individual sighting and avoidance data exposure RLs (~90 to 120 dB re: 1 µPa), at least for during CEE active and control conditions were initial exposures. This observation is also evident scored for individuals within and outside this in the severity scaling analysis for Cell 9 (Table exclusion zone (see Table 18). 19). All recorded exposures exceeding 140 dB Johnston & Woodley (1998) conducted an exten- re: 1 µPa induced profound and sustained avoid- sive survey of AHDs used in the Bay of Fundy to ance behavior in wild harbor porpoises. Harbor exclude pinnipeds from salmon aquaculture sites. porpoises also tend to avoid boats, although Based on the behavioral observations of Olesiuk 516 Southall et al. et al. (1996), Johnston & Woodley (1998) deter- limited available information about noise exposure mined that harbor porpoises were likely being conditions and individual behavioral responses. excluded from extensive areas of important habitat Finally, while their study was not considered in as a result of overlapping AHD deployments. This the severity scaling here, we note the importance study lacked the discrete observational data neces- of the Cox et al. (2001) observations regarding sary for analysis here, but two subsequent studies harbor porpoise habituation. They found that wild contained such measurements for harbor porpoises porpoises were initially displaced by approxi- exposed to AHDs. mately 208 m from active ADDs, but this displace- Olesiuk et al. (2002) conducted a controlled ment decreased by 50% in 4 d, and reached control exposure in which they deactivated an array of levels in 10 to 11 d. Because of the potential for four Airmar® AHDs for 3 wk and then activated habituation, it should be noted that many or most the array for three 3-wk intervals over an 18-wk of the field observations reported here, other than period. Source characteristics are known (10-kHz those that involve long-duration deployments, are fundamental frequency; 194 dB re: 1 µPa-m (peak- likely most relevant for naïve individuals. to-peak) source level; series of 1.8-ms pulses, repeated at 40-ms intervals grouped into 2.3-s Laboratory Observations (Cell 9) trains separated by 2.1-s quiet periods). However, Relatively extensive laboratory data are available no empirical acoustic measurements were obtained. on captive, individual high-frequency cetaceans Exposure RLs were estimated here based on source exposed to some of the same acoustic alarms characteristics and simple assumptions about the (ADDs and AHDs) and scaring devices deployed propagation of high-frequency sounds in shallow- in field applications. We applied our behavioral water environments. Active AHDs resulted in clear scoring paradigm to data from each of the captive avoidance behavior by individuals and groups of studies conducted by Kastelein and colleagues, harbor porpoises. The sighting rate during active which included relatively detailed information on periods as a percent of that during control periods individual responses and directly measured expo- was only 1.4% at ranges from 400 to 599 m, 2.5 sure RLs. to 3.3% at 600 to 2,499 m, and 8.1% at 2.5 to 3.5 Kastelein et al. (1997) recorded behavioral km. RLs at 3.5 km were estimated to be ~123 dB responses (location, swimming speed, and re: 1 µPa. Avoidance data during active and control respiration patterns) of a naïve, captive harbor periods were scored here for individuals within the porpoise exposed to a variety of clicks, tones, and various exposure zones (Table 18). frequency sweeps. All of the relatively low expo- Johnston (2002) observed similar harbor por- sure RLs (~90 to 115 dB re: 1 µPa) resulted in poise behavioral avoidance of active AHDs. They strong behavioral avoidance (subjects basically used an Airmar® dB II Plus AHD System (10-kHz swam rapidly as far from the devices as pos- fundamental frequency; 180 dB re: 1 µPa-m source sible within the enclosure) as well as changes in level, producing a short train of 2.5-ms signals swimming speed and breathing patterns. Although repeated every 17 s) deployed from a small boat. this response quickly abated following noise cessa- They sighted fewer animals when the AHD was tion, no habituation was observed across multiple active, and these animals were significantly fur- exposure events. Data from individual exposure ther away than during control phases. Approximate trials were presented by Kastelein et al. and are exposure RLs at the point of closest approach were analyzed here. To avoid pseudoreplication, these estimated here as ~128 dB re: 1 µPa; mean clos- data are inversely weighted by the total number est approach distance was consistent with exposure of trials to approximate a single exposure for the RLs of ~125 dB re: 1 µPa. individual. Based on harbor porpoise hearing mea- Additional field observations of harbor por- surements (Andersen, 1970) and the Kastelein et poises suggest that their apparently high degree al. (1997) data on behavioral reactions, Taylor of behavioral sensitivity extends to sources other et al. (1997) estimated zones of noise influence than ADDs and AHDs. Koschinski et al. (2003) (audibility, behavioral disturbance, and hearing observed behavioral responses of harbor porpoises damage) for free-ranging harbor porpoises. to simulated wind turbine noise (max. energy Subsequently, Kastelein et al. (2000) exposed between 30 and 800 Hz; spectral density source two naïve subjects to three different nonpulse levels of 128 dB re: 1 µPa/Hz at 80 and 160 Hz). sources and observed generally similar behavioral They sighted harbor porpoises at greater ranges avoidance in all conditions. Pooled data for each during playbacks of simulated wind turbine noise subject were scored and reported here; pooled and observed that animals more frequently used data for each alarm in the dose-response analysis echolocation signals during industrial activity. were weighted to equate with a single exposure These data are not scored here, however, due to event for each individual. Kastelein et al. (2001) later measured similar behavioral responses of Marine Mammal Noise Exposure Criteria 517 the same two individual harbor porpoises to three Field Observations (Cell 12) different acoustic alarms, but these data were not Jacobs & Terhune (2002) observed harbor seal included in this analysis because subjects were no reactions to Airmar® dB plus II AHDs (general longer naïve to controlled noise exposures. source characteristics given in the “Cell 9” section Kastelein et al. (2005) exposed two additional above; source level in this study was 172 dB re: naïve harbor porpoises to various sounds associ- 1 µPa-m) deployed around aquaculture sites. From ated with underwater data transmission systems 1 to 10 AHDs were deployed around nine differ- (clicks, tones, sweeps, and impulsive distance ent sites. Jacobs & Terhune measured received sensors with a range of source characteristics). SPLs around the AHDs and measured the behav- They directly measured source levels of each ior of seals in the surrounding area. Seals in this sound type and RLs at numerous positions within study were generally unresponsive to sounds from the experimental pool. Observed behavioral the AHDs. During two specific events, individu- responses (avoidance and changes in swimming als came within 43 and 44 m of active AHDs and and respiration patterns) were very similar to failed to demonstrate any measurable behavioral those during the previous Kastelein et al. (1997, response; estimated exposure RLs based on the 2000, 2001) studies. Pooled data for each indi- measures given were ~120 to 130 dB re: 1 µPa. vidual response and source type were analyzed These individual observations are weighted to rep- here in the same manner as we applied to the resent a single observation for this study, scored Kastelein et al. (2000) measurements. Kastelein (as 0), and reported in Table 21. et al. (2006a) exposed yet another naïve individual Costa et al. (2003) measured received noise harbor porpoise and reported very similar find- levels from an ATOC sound source off north- ings, which we incorporated as a single pooled ern California using acoustic data loggers placed result, with all exposures equally weighted. on translocated elephant seals. Subjects were captured on land, transported to sea, instrumented Pinnipeds in Water/Nonpulses (Cell 12) with archival acoustic tags, and released such that their transit would lead them near an active ATOC The effects of nonpulse exposures on pinni- source (at 939-m depth; 75-Hz signal with 37.5- peds in water are poorly understood. Studies for Hz bandwidth; 195 dB re: 1 µPa-m max. source which enough information was available for our level, ramped up from 165 dB re: 1 µPa-m over analysis include field exposures of harbor seals to 20 min) on their return to a haulout site. Costa et AHDs (Jacobs & Terhune, 2002) and of translo- al. provided a wide range of detailed quantitative cated diving northern elephant seals to a research measures of individual diving behavior, responses, tomography source (Costa et al., 2003), as well and exposure RLs in well-characterized contexts; as responses of captive harbor seals to underwa- this kind of information was ideal for the present ter data communication sources (Kastelein et al., purposes. Dive depth and duration, descent/ascent 2006b). These limited data (see Table 20) suggest velocity, surface interval, and exposure RL were that exposures between ~90 and 140 dB re: 1 µPa recorded from a total of 14 seals. An additional generally do not appear to induce strong behav- three seals were exposed to the ATOC source ioral responses in pinnipeds exposed to nonpulse during translocations and behavioral observations sounds in water; no data exist regarding exposures were made, but exposure RLs were unavailable. at higher levels. The severity scaling for Cell 12 is Seven control seals were instrumented similarly given in Table 21. and released when the ATOC source was not active. It is important to note that among these stud- Received exposure levels of the ATOC source for ies of pinnipeds responding to nonpulse exposures experimental subjects averaged 128 dB re: 1 µPa in water, there are some apparent differences in (range 118 to 137) in the 60- to 90-Hz band. None responses between field and laboratory conditions. of the instrumented animals terminated dives or In contrast to the mid-frequency odontocetes, cap- radically altered behavior upon exposure, but tive pinnipeds responded more strongly at lower some statistically significant changes in diving levels than did animals in the field. Again, contex- parameters were documented in nine individuals. tual issues are the likely cause of this difference. The behavioral scores assigned here for statisti- Captive subjects in the Kastelein et al. (2006b) cally significant responses were either three or four study were not reinforced with food for remaining depending on the change in diving behavior during in noise fields, whereas free-ranging subjects may exposure relative to mean values for the same indi- have been more tolerant of exposures because viduals before and after exposure (< 50% change of motivation to return to a safe location (Costa scored 3; > 50% change scored 4). Translocated et al., 2003) or to approach enclosures holding northern elephant seals exposed to this particular prey items (Jacobs & Terhune, 2002). nonpulse source (ATOC) began to demonstrate 518 Southall et al. subtle behavioral changes at ~120 to 140 dB re: 1 no avoidance of the area around the various indus- µPa exposure RLs (Table 21). trial sources, most of which emitted nonpulses. Several other field studies (discussed briefly Because of the continuous exposure to multiple below) were considered but not included in the sound sources at varying distances, this study behavioral analyses due to limited information did not produce data on discrete exposures and about source and/or propagation characteristics, responses. individual responses during and/or in the absence of exposure, or both. While studying cetaceans, Laboratory Observations (Cell 12) Richardson et al. (1990b, 1991) made some Kastelein et al. (2006b) exposed nine captive observation of ringed and bearded seal responses harbor seals in a ~25 × 30 m enclosure to non- to playbacks of underwater drilling sounds. Their pulse sounds used in underwater data communi- findings generally suggested a fairly high degree cation systems (similar to acoustic modems). Test of tolerance by exposed pinnipeds to these sounds. signals were identical to those used by Kastelein This contrasts to some extent with the results of et al. (2005) in harbor porpoise exposure studies Frost & Lowry (1988) who found some reduction (frequency modulated tones, sweeps, and bands in ringed seal densities around islands on which of noise with fundamental frequencies between drilling was occurring. Norberg & Bain (1994) 8 and 16 kHz; 128 to 130 [± 3] dB re: 1 µPa-m made detailed acoustic measurements of several source levels; 1- to 2-s duration [60-80% duty arrays of Cascade Applied Sciences® AHDs (11.9- cycle]; or 100% duty cycle). They recorded seal to 14.7-kHz frequency sweeps; 195 dB re: 1 µPa- positions and the mean number of individual m source level; 1-ms pulse produced in 57 to 58 surfacing behaviors during control periods (no discrete pulse chirps of 2.3-s total duration). These exposure), before exposure, and in 15-min experi- devices were placed on the Chittenden Locks in mental sessions (n = 7 exposures for each sound Puget Sound, Washington, in an effort to dissuade type). Background noise and exposure RLs (in predation of wild steelhead trout by California terms of Leq; 32-s total time) were measured at sea lions. Behavioral responses of individual ani- numerous positions around the enclosure for mals, however, were not reported. Norberg (2000) each acoustic source. Acoustic discomfort was evaluated the behavioral responses of California recognized based on movement out of areas that sea lions to Airmar® AHDs (10-kHz fundamen- animals used during control periods. An acoustic tal frequency; 195 dB re: 1 µPa-m source level; discomfort threshold was calculated for the group short train of 2.5-ms signals repeated every 17 of seals for each source type, and for each sound s) intended to reduce predation on salmonids in source this was ca. 107 dB re: 1 µPa. Seals gener- aquaculture facilities. Behavioral observations ally swam away from each source, avoiding it by suggested limited behavioral deterrence by the ~5 m, although they did not haul out of the water devices (predation rates were similar in experi- or change surfacing behavior. Seal reactions did mental and control conditions), but measures of not appear to wane over repeated exposure (i.e., RLs and individual response behavior are absent. there was no obvious habituation), and the colony Yurk (2000) also observed pinnipeds exposed to of seals generally returned to baseline conditions AHDs in the context of fisheries interactions. He following exposure. determined that active AHDs were more effec- For the behavioral analysis conducted here, the tive than a mechanical barrier or altered lighting Kastelein et al. (2006b) results were interpreted as conditions in dissuading harbor seals from prey- follows. Because the behavior of individuals within ing on fish under bridges. Again, however, insuf- the same pool at the same time cannot be consid- ficient information regarding received sounds ered independent, the group of nine harbor seals and individual responses is available to consider was considered a single observation. Because of these observations explicitly here. Koschinski similarity of sources and exposure conditions and et al. (2003) observed harbor seals during under- the close temporal timing of exposures, we com- water playbacks of simulated wind turbine noise bined observations across the four sound types and (maximum energy between 30 and 800 Hz; spec- include a single observation within each appropri- tral density source levels of 128 dB re: 1 µPa/Hz ate 10-dB bin. Exposures between ~80 and 107 dB at 80 and 160 Hz). Harbor seals were sighted at re: 1 µPa seemed insufficient to induce behavioral greater distances during playbacks than during avoidance in the colony of seals, but higher expo- control conditions. However, limited information sures were considered sufficient. Consequently, on received exposures and individual behavioral single observations indicating no response (0) responses precluded inclusion in our analysis. appear in the 80 to 90 and in the 90 to 100 dB re: Moulton et al. (2003, 2005) studied ringed seals 1 µPa exposure bins, and a single observation before and during the construction and operation indicating avoidance behavior (6) is shown in the of an oil production facility. They found little or 100 to 110 dB re: 1 µPa condition (Table 21). Marine Mammal Noise Exposure Criteria 519

Pinnipeds in Air/Nonpulses (Cell 15) as the rocket goes supersonic. Extensive research has been conducted on the effects of both sound There has been considerable effort to study the types on pinnipeds. Nonpulse exposures are effects of aerial nonpulse sounds on pinniped considered in this section, whereas behavioral behavior, primarily involving rocket launches, air- responses to the pulse component of some of the craft overflights, power-boat approaches, and con- same launches are considered in Appendix B. struction noise. Unfortunately, many of the studies Because many measurements were made on the are difficult to interpret in terms of exposure RL same few colonies of pinnipeds that were exposed and individual or group behavioral responses. In to multiple launches, it is likely that some of the many cases, it was difficult or impossible to dis- same individuals were resampled. Therefore, we cern whether the reported behavioral response was weighted the combined results across studies for induced by the noise from a specific operation or each species and breeding location into a single some correlated variable such as its visual presence. observation for the behavioral analysis here. That For these reasons, most of the observational studies is, we considered each species in an individual of behavioral disturbance are not appropriate for breeding colony a single unit of observation quantitative analyses relating exposure level and across studies. The results were pooled accord- scored behavioral response. However, a number ingly in Table 22, but the studies are discussed of the technical reports and analyses of rocket longitudinally below. The studies discussed below launches are relevant for this cell and contain suf- reported exposure conditions on or near pinniped ficiently detailed information regarding estimated breeding rookeries during launches of different RLs. These observations are complicated, how- types of rockets using a variety of metrics, includ- ever, by the fact that all studies were conducted in ing A-weighted values and a frequency-weighting the same general area with subjects likely habitu- function derived from the harbor seal audiogram; ated to the presence of launch noise. Further, in we used unweighted SPL values for the analysis many cases, exposures contained both a nonpulse here. component and a pulse component (described Thorson et al. (1998) measured harbor seal below). Only those observations for which there responses and conducted AEP measurements on was clearly just nonpulse exposure were con- seals exposed to a Titan IV A-18 launch from sidered in the severity scaling analysis (Thorson Vandenberg Air Force Base (VAFB), California. et al., 1999, 2000b; Berg et al., 2002). They studied colonies both on the mainland at VAFB The limitations of these and other potentially and on nearby Santa Cruz Island. Unfortunately, applicable studies resulted in a very limited data the launch occurred at night and during a period set for use in this analysis (see summary in Table of relatively high tide, limiting both the number of 22 and severity scaling analysis in Table 23). As a seals present on the rookeries and the observation general statement from the available information, of individuals. However, behavioral monitoring pinnipeds exposed to intense (~110 to 120 dB re: over several days after the launch did not indicate 20 µPa) nonpulse sounds often leave haulout areas any abandonment of the breeding rookeries at and seek refuge temporarily (minutes to a few either site. Hearing measures (AEP) on individuals hours) in the water. In contrast, pinnipeds exposed tested before and several hours after the launch did to distant launches at RLs ~60 to 70 dB re: 20 µPa not indicate any loss of sensitivity. tend to ignore the noise. It is difficult to assess the Thorson et al. (1999) conducted similar observa- relevance of either of these observations to naïve tions of harbor seals at VAFB and also observed individuals, however, given the repeated exposure northern elephant seals, California sea lions, and of colonies studied to such noise events. Also, northern fur seals at nearby San Miguel Island. there are strong species differences, with harbor Following the launch (of an Athena 2 IKONOS-1 seals being much more responsive than northern missile), 33 harbor seals (including six pups) at the elephant seals (e.g., Holst et al., 2005a, 2005b). VAFB rookery entered the water. They began to Due to the limitations of the available data, it is return to the beach beginning 16 min after the launch, not currently possible to make any further general and no pups were observed to have died as a result of characterizations regarding this condition. the event. This behavior was considered to represent A series of highly detailed, quantitative analy- both minor avoidance and a brief/minor potential ses on the behavior of pinnipeds exposed to the or actual separation of females and dependent off- sounds of various large missile launches were spring (scored 6 here). The maximum unweighted reviewed. These sources generally produce sus- SPL value was 119 dB re: 20 µPa. Individuals tained, generally low-frequency (little energy of the three pinniped species monitored on above 1,000 Hz) “rumbling” sounds lasting tens San Miguel Island reacted similarly. However, their of seconds (nonpulse) associated with launch responses were to the sonic boom generated by the boosters, as well as a sonic boom (pulse) in flight rocket once airborne rather than to the nonpulse 520 Southall et al. sounds associated with the launch per se, and thus entered the water immediately following the onset are not scored here. of launch noise. More seals (n = 56) were present Thorson et al. (2000a) conducted observations at the locations 90 min after the launch event, indi- of harbor seal abundance, distribution, and haulout cating the temporary and minor nature of the dis- patterns at VAFB for several days before and after turbance, and no injured animals were located. The the launch of a Titan II G-13 missile from VAFB. avoidance behavior was coincident with a maxi- This launch occurred during the middle of the night, mum unweighted SPL value near the rookeries of precluding direct observation of seal reactions (and 119 dB re: 20 µPa (unweighted aerial SEL value behavioral scoring here), although observations was 130 dB re: [20 µPa]2-s). on subsequent days indicated generally nominal Finally, Berg et al. (2004) observed behavioral harbor seal presence and distribution in the area. responses of California sea lions, northern elephant Thorson et al. (2000b) measured behavioral and seals, and northern fur seals on San Miguel Island auditory responses of harbor seals at VAFB and to the launch of an Atlas IIAS MLV-14 missile behavioral responses of northern elephant seals from VAFB. Received signals were sonic booms and California sea lions on San Miguel Island to which had little to no effect on the behavior of the the launch of a Titan IV B-28 missile from VAFB. pinnipeds, other than minor orienting behaviors They observed all 54 harbor seals at the VAFB site and movements in some of the California sea lions. moving from the breeding rookery into the water These results are not scored here, in part because within 2 min of the onset of the launch (47 entered the sounds included pulses. the water immediately). The maximum unweighted Other researchers have investigated the effects SPL value near the rookery was 116 dB re: 20 µPa; of other kinds of human activities (e.g., aircraft, this exposure was considered here to be consistent motorboats, general human presence) as well as with a behavioral score of 6 for this group of seals. rocket launches on the haulout behavior, including The sound persisted for several minutes, and the avoidance, of pinnipeds (Allen et al., 1984; Suryan unweighted SEL value was 127 dB re: (20 µPa)2- & Harvey, 1998; Born et al., 1999; Moulton et al., s. There was no difference in the hearing capabili- 2002). The combined results indicated that hauled- ties of three young seals tested using AEP meth- out pinnipeds in certain conditions can be disturbed, ods before and after the missile launch. Neither significantly in some cases, by the presence of vari- the California sea lions nor elephant seals on San ous human activities. However, these studies lack Miguel Island were observed to respond at all to either specific estimates of received noise expo- the “faint” noise associated with the launch, cor- sure conditions or individual-specific behavioral responding to a severity scaling score of 0 (Table responses or both. Additionally, multiple stimuli 23). These sounds were from the launch boosters were generally simultaneously present, including (nonpulses) rather than sonic booms and were esti- the visual presence of sources, which preclude their mated here as ~60 to 70 dB re: 20 µPa based on the inclusion here. Gentry et al. (1990) determined that measurements and descriptions given. northern fur seals were generally tolerant of under- Berg et al. (2001) obtained similar measure- ground explosions and other quarrying operations ments of behavioral responses of harbor seals at in relatively close proximity; only a few orienting VAFB and California sea lions and northern ele- behaviors were observed in response to the largest phant seals at San Miguel Island to a Delta II EO-1 blasts. Some acoustic measurements were made, missile launch from VAFB. Observations were also but individual behaviors or group responses and made of southern sea otter (Enhydra lutris nereis) received exposure levels were not reported and were and California brown pelican (Pelecanus occiden- thus not scored here. talis californicus) responses. No harbor seals were Holst et al. (2005a, 2005b) observed behavioral hauled out on the VAFB rookery during this launch. responses in three species of pinnipeds—harbor seal, Berg et al. note that subsequent harbor seal abun- California sea lion, and northern elephant seal—on dance and distribution in the days after the launch San Nicolas Island to 47 small- and mid-sized mis- were within normal variability, and there appeared sile launches over a 4-y period. They observed to be no lasting behavioral reactions. Elephant seals animal presence and distribution before launches and California sea lions at San Miguel Island did and behavior during and following launches. Some not noticeably respond to sounds associated with of the missiles generated sonic booms, but the the launch, which in this case were predominantly majority of the exposures were relatively low-fre- the sonic boom (pulse) component. quency, long-duration rumbling sounds that would Berg et al. (2002) measured behavioral responses be categorized as nonpulses. During many launches, of harbor seals on VAFB rookeries to the launch acoustic measurements were made near the animals of a Titan IV B-34 missile from a launch pad at whose behavior was videotaped. Peak, SPL, and SEL VAFB ~8.6 km away. At the time of the launch, 38 exposures were reported. This dataset has not been seals were present at two haulout sites, all of which incorporated into the present analysis. However, Marine Mammal Noise Exposure Criteria 521 results indicated that California sea lions had mixed considered as they relate to injury criteria. In reactions to rocket launches, with some individuals California sea lions and northern elephant seals, exhibiting startle responses and increased vigilance there were significant correlations between behav- and others showing virtually no reaction. Northern ioral responses and both the missile’s closest dis- elephant seal reactions were minimal, consisting tance and the RL of the launch sound near the only of minor movements and orienting responses pinnipeds (SEL). Corresponding relationships for that rapidly subsided. Conversely, harbor seals were harbor seals were weaker. Holst et al. (2005b) con- by far the most responsive of the pinnipeds observed, cluded that the temporary behavioral responses, with many individuals entering the water from even the relatively severe ones observed in harbor haulout sites following rocket launches and failing seals, do not appear to have substantial adverse to return for periods of hours. No cases of long- effects on pinniped populations. This conclusion term pup separation or of injury were documented. is based on the decades-long occurrence of missile If those phenomena had occurred, they would be launches and the presence of increasing numbers considered relatively severe in terms of the behav- of pinnipeds of all three species in the area. ioral scoring paradigm given here and should also be

(Left to Right): David Kastak, Roger Gentry, Peter Tyack, Brandon Southall, Darlene Ketten, W. John Richardson, Jeanette Thomas, James Miller, Ann Bowles, James Finneran, Charles Greene, Jr., Paul Nachtigall, and William Ellison This manuscript is respectfully dedicated to our co-author, David Kastak.

Dave was a brilliant scientist, but even more importantly, he was a man with great sincerity, integrity, and a sharp wit. He was an inspiration and mentor to many, and his significant, incisive research on marine mammal cognition and sensory systems over the past two decades provided advances that shaped and will continue to guide the future of these fields. Dave was a valued colleague and treasured friend to all of us. He will be missed but never forgotten. Did you know that you *OURNALCOVER can subscribe to Aquatic Mammals online?

Now available at www.ingentaconnect.com

The most comprehensive collection of academic and professional publications online with more than 17 million articles available

For online access to Aquatic Mammals . . .

Register free at Plus, take advantage of extra free www.ingentaconnect.com. services available on IngentaConnect, including up to five New Issue Alerts, Select personal subscriptions. RSS feeds, marked lists, and more. Select Aquatic Mammals and enter your subscription number. If you are a library or institution not currently registered at You will be notified by e-mail once www.ingentaconnect.com and you want your online access has been activated. to know more, please contact our Customer Services department at [email protected] for further information.

2008 Subscription Rates for Aquatic Mammals Effective October 2007 – October 2008 Circle type of subscription desired.

Printed version, 4 issues per year Printed version per year, plus online versions: (approximately 375 pages): Library: $446 USD Library: $263 USD Agency: $415 USD Agency: $244 USD Individual: $168 USD Individual: $131 USD Student or retired colleague: $158 USD Student or retired colleague: $105 USD CD Version, 4 CDs per year: Online access through Ingenta’s Library: $184 USD Web page, 4 issues per year Agency: $171 USD (Articles are searchable and printable Individual: $79 USD from the web page.): Student or retired colleague: $53 USD Library: $210 USD Agency: $195 USD Individual: $63 USD Student or retired colleague: $47 USD (Students: Include a copy of current student ID with order.)

Sold to: Ship to:

______

Submit payment as check in U.S. dollars or by VISA or MasterCard to . . . Aquatic Mammals Subscription cost: $_____USD c/o Gina Colley Shipment outside N. Am., add $_____50 USD Document and Publication Services Total Charge: $_____USD Western Illinois University Macomb, Illinois 61455, USA E-mail: [email protected] Telephone: (309) 298-1917 Fax: (309) 298-2869

Credit card number: ______-______-______-______Expiration date: ____/____

Name on credit card: ______

Receipt available upon request.