Click for next page ( 42


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 41
Assessment of Continuing Research Needs In its 1994 report the NRC recommended future research that would provide a better understanding of the effects of low-frequency sounds on marine mammals and their prey. This chapter assesses the progress made in addressing the targeted issues since publication of that report. The major aims are presented as described in that 1994 report; specific goals are described in boldface type. Since 1994 the Office of Naval Research (ONR) has invested significant funds into attempts to address a number of the issues raised in the 1994 report. These studies were described to the Committee by Robert Gisiner at the April 1999 meeting, and their results are cited below as appropriate. Although an excellent start has been made in addressing the many questions, there is still a dearth of data on marine mammal bioacoustics. Some of these research needs fall within the purview of ONR, but it is a mission agency and its goals are highly oriented toward the missions of the Navy. Thus, it cannot be expected to deal with all of the important issues raised in the 1994 NRC report, implying that additional sources of funding for marine mammal bioacoustic research are required if better knowledge of marine mammal hearing is deemed by policy makers to be desirable. Since publication of the 1994 NRC report, the anatomy of the inner ear of several additional marine mammal species has been studied. Computational models based on the anatomical parameters of marine mammal cochleae have been developed, and predictions from such models have correlated well with behaviorally determined audiograms in several species of toothed whales (Ketten, 1997~. This kind of modeling provides an important new tool for assessing the auditory sensitivity and frequency range in whales that are not amenable to experimental measurements. There have been significant advances in our knowl- 41

OCR for page 41
42 MARINE MAMMALS AND LOW-FREQUENCY SOUND edge of low-frequency hearing capabilities of several toothed whale species as well as in techniques required to acquire such data. Significant data have been obtained on temporary threshold shift (TTS)1 in several marine mammal species (Ridgway et al., 1997; Kastak and Schusterman, 1998; Kastak et al., 1999; Schlundt et al., 2000~. Additional work needed includes (1) anatomical studies of species with known audiograms to validate the use of anatomical features to predict auditory capabilities and (2) studies using both behavioral and auditory evoked potentials techniques to determine auditory capabilities of marine mam- mals (especially of baleen whales). Additional studies are needed to determine the abilities of marine mammals to detect natural sound in the presence of human- generated background noise (Erbe and Farmer, 1998~. Several papers have appeared since 1994 reporting on use of auditory evoked potentials to study dolphins (Szymanski et al., 1995, 1998; Popov and Klishin, 1998; Popov and Supin, 1998; Popov et al., 1998~. Dolphin (1997) reviewed auditory processing in several toothed whales (Grampus, Orcinus, Tursiops, Delphinapterus) that were studied using evoked potential methods. He pointed out that auditory evoked potentials can be used to study a wide range of questions about hearing and the auditory system. These include determination of auditory filter shapes that may provide clues about the potential masking effects of some human-generated sounds, determination of ITS resulting from sound exposure, and studies of the masking effects of specific sounds. Dolphin (1995, 1996) and Dolphin et al. (1995) examined temporal processing in several whale species in response to amplitude-modulated stimuli using evoked potential techniques. Ridgway and Au (1999) reviewed earlier work on processing by the auditory central nervous system and approaches to sound conduction to the dolphin ear. Popov and Supin (1998) studied dolphin auditory evoked responses to rhythmic sound pulses. Popov and Klishin (1998) reported on a study of common dolphin hearing using the electroencephalogram. Popov et al. (1998) reported on fre- quency tuning of the dolphin auditory system using evoked potential methods. 1Temporary threshold shift is a temporary increase in the threshold audible sound level presumed to be caused by temporary inactivation of the outer hair cells at a given frequency. 2Auditory evoked potentials (AEPs) are electrophysiological recordings of minute voltages gener- ated by neural activity in the brain in response to acoustic stimuli. AEPs can be noninvasively recorded from the scalp skin surface and have been broadly applied with great success in humans. The response following the presentation of a brief acoustic stimulus is a series of peaks or waves that arise as a consequence of more or less simultaneous firings in sets of neurons located in successively higher auditory nuclei. Evoked potentials are quite weak, meaning that multiple presentations of the acoustic stimulus and averaging techniques must be used to measure them. Among the advantages of AEP measurements are that (1) they require no or only minimal cooperation from the subject, (2) responses are rapidly obtained and are highly robust, (3) response detection can be fully auto- mated and based on totally objective acceptance criteria, and (4) tests are noninvasive and therefore amenable to examination of protected species. AEPs have now been obtained from a wide range of species. The availability of such data greatly facilitates comparative studies of hearing and auditory function.

OCR for page 41
ASSESSMENT OF CONTINUING RESEARCH NEEDS 43 Employing the auditory brainstem response (ABR), a type of evoked poten- tial technique, Szymanski et al. (1999) reported that the ABR of killer whales provided a means of suprathreshold hearing measurement. The ABR thresholds of individual killer whales were, at most frequencies, within 12 dB of thresholds measurable by behavioral responses. At the most sensitive frequency (20 kHz), mean thresholds determined by behavioral and physiological methods differed by only 3 dB, thus showing the usefulness of this technique as a proxy for measure- ments of behavioral thresholds. The two killer whales studied are the largest animals (2,000 to 3,000 kg) ever successfully investigated with evoked potential methods (see also Szymanski et al., 1995, 1998~. Although the ABRs were lower in amplitude than those for dolphins less than one-tenth as heavy, ABRs from the killer whale (Szymanski et al., 1999) were adequate for determining values near the behavioral threshold (ABRs averaged 5 dB higher than behavioral thresholds for the most sensitive range of 18 to 42 kHz). Because ABR amplitudes appear to correlate with the relative size of auditory structures in the brain, Szymanski et al. (1999) suggested that successful use of this method in the larger baleen whales may be more difficult. Experience with a young gray whale at Sea World (San Diego) demonstrated that working with larger whales requires more sensitive techniques, quieter conditions, and more time to use ABR techniques on whales than is needed for smaller marine mammals. However, the promise of this technique for validating behavioral observations on large whales indicates that the NRC (1994) recommendation about testing hearing in beached, stranded, or entrapped larger whales should continue to be a goal. These investigations should build on those mentioned above using teams experienced in electrophysiological techniques. A comprehensive set of species groups, signal types, and biological param- eters that should be measured for marine mammals is presented in Chapter 5. Priority for acoustic studies should also be given to species that are (1) endan- gered or threatened (e.g., the northern right whale, Eubalaena glacialis); and (2) known or suspected to hear and communicate using low-frequency sound (e.g., baleen whales, sperm whales, elephant seals). BEHAVIOR OF MARINE MAMMALS IN THE WILD Aim: To determine the normal behaviors of marine mammals in the wild and their behavioral responses to human-generated acoustic signals (NRC, 1994, pp. 41-47~. This aim can be conceptualized as a number of specific topics, which are shown here in bold print. . Determine how marine mammals utilize natural sound for communi- cation and for maintaining their normal behavioral repertoires. Although much is unknown about communication among baleen whales, and

OCR for page 41
44 MARINE MAMMALS AND LOW-FREQUENCY SOUND little research is currently directed at this topic, some work has been published since 1994 on species that may be affected by the Acoustic Thermometry of Ocean Climate (ATOC) signal or low-frequency active (LFA) sonar. Gray whale (Eschrichtius robustus) vocalizations were documented at two locations along their southward migration by Crane and Lashkari (1996), who found that vocal behavior varied with location. More vocalizations occurred in shallow water than in deep water, and the type of vocalization (call structure) produced most frequently during the southward migration differed from vocalizations in the lagoons of Baja California where the whales calve and spend the winter. Crane and Lashkari concluded that vocal activity is an important component of migratory behavior in gray whales, probably for communication rather than navi- gation. These data are important as a reminder that the potential for acoustic interference by human-generated noise can be site-dependent and seasonally vari- able. Clearly, there is a greater potential for disruption of normal behavior in areas where vocalization rates are high. Recordings of blue whale vocalizations in the Pacific Ocean have indicated that the occurrence of a two-part call is characteristic of populations in both the Gulf of California and the waters off California (Thompson et al., 1996; Rivers, 1997; Stafford et al., 1999~. Whether these two populations are part of a single Pacific Ocean population or whether there is limited exchange is yet to be deter- mined. The presence of a subspecies (pygmy blue whale, Balaenoptera musculus brevicauda) off Australia (Ljungblad et al., 1997) adds further incentive for documenting the occurrence and types of vocalizations of blue whales in the Pacific Ocean (Stafford et al., 1998~. The recent availability of the Navy's Integrated Undersea Surveillance Sys- tem (IUSS) and other hydrophore arrays to scientific researchers has permitted the use of passive acoustic tracking as a method for documenting migration routes and critical habitats where baleen whales are seasonal visitors or residents (e.g., Stafford and Fox, 1996; Stafford et al., 1998, 1999; Watkins et al., 2000~. Although direct visual observations of vocalizing whales provide more details of behavior compared to the use of remote monitoring (where animals are not observed as they produce sounds), the detection of well-described species-specific calls to track baleen whale migrations and activity using the IUSS and other hydrophore arrays has great potential. Several studies (listed above) have provided descriptions of vocal behavior, and such data are critical for interpreting the biological significance of any changes in vocal behavior induced by human-generated noise. A shift in the focus of regulations from harassment to the biological significance of behavioral disruption (recommended later in this report) will require a much better under- standing of the functions of vocal behavior. Therefore, a high priority should be given to basic research on the behavioral ecology of how marine mammals use the sounds they produce; the results of such research would have immediate regulatory implications.

OCR for page 41
ASSESSMENT OF CONTINUING RESEARCH NEEDS 45 Determine the responses of free-ranging marine mammals to human- generated acoustic stimuli, including repeated exposure of the same individuals. How is the use of natural sounds altered by the presence of human-generated sounds? ATOC transmissions from a fixed source off Hawaii theoretically could provide some data on repeated exposure to this signal; however, few individuals were seen close enough to the Hawaii source site either in experimental (ATOC signal on) or control (ATOC signal off) conditions to provide evidence of any response to source levels greater than 130 dB. Individual identification of hump- back whales was part of the MMRP, but the activities of specific individuals exposed to the ATOC stimulus repeatedly have not been presented to date. In addition, humpback whale song was recorded prior to and during exposure to the ATOC signal, but these data also have not been examined in sufficient detail to determine if the presence of the ATOC sound has a concurrent and/or long- lasting effect on the songs of individual whales. The Quick-Look reports did indicate that there was no change in average acoustic energy in the 200- to 800- Hz band, a dominant frequency band for humpback song. Detailed examination of the actual songs for changes in frequency content or amplitude of individual components that may overlap with the broadband components of the ATOC signal was not completed at the time of the Committee's April 1999 meeting. Clearly, such data are unique and warrant a detailed examination. If subsequent analyses of data support the finding of no major change in the vocalizations of humpback whales, that would support the contention that ATOC is unlikely to have long-term effects on this species at this important breeding site. Two studies on white whales (Delphinaptera leucas) have provided perti- nent data since the publication of NRC (1994~. Erbe and Farmer (1998) trained captive animals to detect vocalizations from other white whales in the presence of background noise to assess the maskings effects of icebreaker activities. Erbe and Farmer found that the aeration system used to clear ice debris had the greatest masking effect, followed by propeller noise. This innovative protocol may be useful for determining masking effects in other toothed whales using a variety of human-generated noises, including the ATOC signal and LEA sonar. A logisti- cally challenging field study was conducted in the St. Lawrence estuary with free-ranging white whales (Lesage et al., 1999~. The vocal activity of the animals was monitored prior to, during, and after exposure to two vessels with different Masking is the reduction in the audibility of one sound due to the presence of a second sound. of greatest interest here is simultaneous masking, in which both sounds overlap in time, but masking can also occur when the signal is a brief sound presented immediately before or immediately after the masker (backward and forward temporal masking, respectively). In humans the amount of simulta- neous spectral masking observed is related to the width of the person s auditory filter located at the signal frequency.

OCR for page 41
46 MARINE MAMMALS AND LOW-FREQUENCY SOUND noise signatures: a motorboat and a ferry. More than 70 recording sessions were conducted, but only six met the criteria for a successful experimental sequence with acceptable signal-noise ratio for analysis. Despite the small sample size (N= three for each vessel), differences were noted in the vocalizations. Calling rates declined in five of the six sessions as noise levels increased, the occurrence of certain call types increased, repetition rates of certain calls increased when a vessel was within 1 km of the white whale, and the mean frequencies produced were higher, presumably to move the frequency of the call outside the frequency band of the masking noise produced by the vessels. These two studies provide evidence that specific human-generated noises can affect the vocal activity of white whales in the short term. Although generali- zations to other species are not without risk, baleen whales exposed to low- frequency noise (from ATOC, LEA, ships) may respond similarly. In addition, the field study by Lesage et al. (1999) indicates that such data are difficult but not impossible to obtain given a sustained field effort. Tyack (1998) and Tyack and Clark (1998) concluded that 10 of 17 singing humpback whales exposed to low-frequency sounds from the SURTASS-LFA4 sonar system stopped singing during playback with a source level that ranged from 155 to 205 dB, resulting in maximum received levels of 120 to 150 dB. Four of these 10 stopped when they were joined by other whales, a behavior that has been observed in previous studies with no human-generated sound (Tyack, 1983), and their responses cannot be definitively associated with the playback. There was no difference in the received levels for the six whales that stopped singing, apparently in response to playback, compared to the seven that did not stop, suggesting individual differences in perception and/or response. In order to evaluate the significance of disruption induced by cessation of singing, it will be necessary to make a decision about what proportion of the population of singers can be disrupted before the disruption will have a population-level effect. An important task is to determine how different sound types and levels affect migration and other movement patterns of marine mammals. The 1994 NRC report specifically recommended shore-based studies similar to the Malme et al. investigation of migrating gray whales. Tyack and Clark (1998) replicated the Malme et al. (1983) study using a sound projector from the SURTASS-LFA. Two important characteristics of the Tyack and Clark study were that (1) the intensity of the source was adjustable and (2) its location could be changed. The results differed considerably depending on both of those variables. When the source level was 170 dB and was located in the migratory path of the whales, animals deflected around the source by a maximum of only several hundred 4SURTASS-LFA is the acronym for a low-frequency active sonar system developed by the U.S. Navy to detect submarines. This sonar uses a vertical array of sound projectors deployed from a surface ship to broadcast sounds in the 100- to 500-Hz range. The Navy made this system available for three marine mammal studies conducted over a period of 1 year.

OCR for page 41
ASSESSMENT OF CONTINUING RESEARCH NEEDS 47 meters. When the source level was increased to 185 dB and was located in the migratory path of the whales, the animals changed course so as to avoid passing within a kilometer of the source on both the onshore and the offshore sides. The evidence suggested that most whales avoided exposure to received levels of 140 dB or more. These findings agreed with earlier findings of Malme et al. (1983, 1984) using noises associated with oil industry activities. This study confirmed that whales change their response as source level is changed, which demonstrates that they are responding to received level, not just distance or sound gradient. When the source was located on the offshore side of the migratory path, there was little evidence of any diversion in the individual migratory paths for source levels of both 185 and 200 dB (Tyack and Clark, 1998~. This finding is espe- cially interesting because calculated received levels were higher in these cases than in the condition producing the strong avoidance effect when the source was located in the migratory path at a source level of 185 dB. For example, at the 200-dB source level for offshore playbacks, the received levels measured at ranges of 2 to 2.5 km were >140 dB. Since the offshore source was placed about 2 km offshore from the inshore source location, this means that during the off- shore playbacks most of the whales were exposed to received levels that would almost certainly have elicited an avoidance reaction had the source been placed in the inshore location. Thus, for these whales there clearly was something disturb- ing about a strange sound source located in the migratory path that was not disturbing when that same source was on the offshore side of the migratory path. This study illustrates that behavioral responses to noise sources may not be solely dependent on the acoustic nature of the noise, but on the location of the noise as well. Apparently, the 120-dB avoidance model, which seemed correct for non- impulse noises in the migratory path, is not valid for offshore sources. High-frequency pingers and submarine sonar pings are known to affect sperm whale vocalization rates and behavior (Watkins and Schevill, 1975; Watkins et al., 1985, 1993~. Low-frequency sound also may affect sperm whales because their wide-band clicks contain energy between 100 and 2,000 Hz (Watkins et al., 1985; Moore et al., 1993), which is suggestive of low-frequency hearing. Gordon et al. (1996) were funded by the MMRP for two three-week cruises off the Azores to study how sperm whales responded to experimental playbacks of M-code sequences similar to the ATOC stimulus as well as similar stimuli of higher frequency. Most standard playbacks were conducted at 75 Hz (the ATOC frequency), and one each was conducted at 2, 3, and 4 kHz. There were 16 control trials. Gordon et al. observed no significant difference between playback and control groups in blow rates of whales before diving or the relative bearing of sperm whales with respect to the source vessel. Sperm whales produce a regular series of clicks starting about three minutes after the onset of a dive. Gordon et al. (1996) compared six different measures from these clicks and found no differ- ence for the 75-Hz playbacks, although the initial click rate was higher for the three higher frequency playbacks than during controls.

OCR for page 41
48 MARINE MAMMALS AND LOW-FREQUENCY SOUND Many earlier reports suggest that sperm whales may silence or move out of an area in response to manmade noise (Watkins et al., 1985; Bowles et al., 1994; Mate et al., 1994~. The contrasting lack of response in Gordon et al. (1996) may reflect different responsiveness to different stimuli or perhaps that different groups of sperm whales have differing responsiveness depending on their prior exposure history. As with the Frankel et al. (1995) study, this six-week pilot study confirmed the utility of observations for which the study site and experimental protocol are optimized for marine mammal studies. One of the most important limitations of the Gordon et al. study was that the limited source level of the sound source meant that few whales were exposed to received levels above 120 dB. Gordon et al. (1996) advocate further studies using more powerful sources and more sensitive methods for measuring or estimating received levels and for monitoring responses of the whales. In another recent study of sperm whales in the Atlantic Ocean, Andre et al. (1997) presented individuals off the Canary Islands with various noise sources to determine if acoustic deterrence could reduce whale collisions with ferry boats. Although there is no documentation of the received levels for the sound stimuli used, Andre et al. observed approaches to the source, which were interpreted as curiosity, when an artificial click "coda" was presented. The authors concluded that sperm whales exposed to high levels of shipping noise have a high tolerance for noise. Alternatively, those animals may have permanent threshold shifts. There is no way to distinguish between those possibilities with the data available at present, although postmortem examinations on the cochleae of a few animals could resolve this uncertainty. A careful study of the response of sperm whales to low-frequency sound seems warranted. Several papers have suggested that beaked whales tend to strand when there are naval operations offshore. Simmonds and Lopez-Jurado (1991) reported on four mass strandings between 1985-1989 of Cuvier's beaked whale (Ziphius cavirostris) on the coast of Fuerteventura in the Canary Islands that may have been related to naval maneuvers. Frantzis (1998) reported on another mass stranding of 12 or more beaked whales sighted over 38 km of coastline during two days (May 12 and 13, 1996) in the Kyparissiakos Gulf in Greece. There was no external sign of injury or disease in any of these animals. Frantzis (1998) concluded that the mass stranding was associated with a concurrent NATO sonar exercise. The Frantzis paper stimulated the NATO research center that con- ducted the sonar tests to convene panels to review the data (D'Amico,1998~. The NATO sonar transmitted two simultaneous signals, one at 450-700 Hz and one at 2.8-3.3 kHz at source levels of just under 230 dB. This combined signal lasted four seconds and was repeated once every minute. The NATO analysis suggested close timing between the onset of sonar transmissions and the first strandings. Unfortunately, it was not possible to determine the received levels experienced by the stranded whales. D'Amico (1998) states that received levels as high as 150-160 dB were estimated to occur at ranges of 50 km. Sperm whales were

OCR for page 41
ASSESSMENT OF CONTINUING RESEARCH NEEDS 49 heard within 10-25 km of the sound source, but demonstrated no obvious changes in their clicking patterns before, during, and after sonar transmissions. Although these papers raise concern about the effects of noise on beaked whales, they provide little guidance regarding what exposures may be dangerous and which are safe. There is a clear need for experimental studies of the responses of beaked whales to carefully controlled exposures of noise. NATO sonars have been tested in the Mediterranean Sea on many occasions without strandings. Both Simmonds and Lopez-Jurado (1991) and Frantzis (1998) started with rare strandings and then looked for some other rare event that might correlate, but neither paper makes a strong case for having performed a thorough systematic survey of when naval or sonar exercises might have occurred in these areas in the absence of strandings. There is a clear need for studies designed to test this association more systematically. In areas where beaked whales are common or have historically stranded, it would be good to set up a prospective study monitoring noise exposure, systematically logging mass strandings and sources of loud noise such as naval exercises. Careful necropsy of stranded animals would help test for any noise-induced injuries. . Determine the response of deep-diving marine mammals to low- frequency sounds whose characteristics (source level, frequency bandwidth, duty cycle) duplicate or approximate those produced by acoustic oceanographers. The only research on this issue known to the Committee was the previously mentioned study on elephant seal deep diving conducted as part of the California ATOC observations. This should be an area of priority for future studies since it is directly related to the issue of the effects of the ATOC source on the animals that may approach closest to the source. Several other studies have reported on the effects of ATOC-like sounds on various marine mammals (e.g., Mattlin, 1995; Aburto et al., 1997; Harvey and Eguchi,1997~. However, none of these studies has appeared in the peer-reviewed literature, and weaknesses in data acquisition, analysis, or interpretation limit their usefulness. Although these studies were examined by the Committee, their results will not be considered further. Relevant results should be published in the peer-reviewed scientific literature, and funding agencies should provide support for this critical component of research. STRUCTURE AND FUNCTION OF MARINE MAMMAL AUDITORY SYSTEMS Aim: To determine the structure and capabilities of the auditory system in marine mammals (NRC, 1994, pp. 47-53~.

OCR for page 41
so MARINE MAMMALS AND LOW-FREQUENCY SOUND This aim can also be conceptualized as a number of specific topics, which are shown below in bold print. . Determine basic hearing capabilities of various species of marine mammals. Low-frequency audiograms have now been obtained from multiple individu- als of several species of toothed whales and seals by using a variety of behavioral and electrophysiological techniques. Audiograms extending below 100 Hz have been obtained for white whales, the bottlenose dolphin (Tursiops truncates), the false killer whale (Pseudorca crassidens), and Risso's dolphin (Grampus griseus), as well as the harbor seal (Phoca vitulina) (Au et al., 1997; Kastak et al., 1999~. Additional species of toothed whales must be tested because auditory differences among species can be substantial (e.g., porpoises versus killer whales). There is still no audiogram for any baleen whale species, something that is understandable considering the difficulties of working with these giant species. Since the publication of NRC (1994), several studies have been published that reveal new information on auditory capabilities of seals and toothed whales. Underwater thresholds for three seal species (one eared seal and two true seals) were obtained with lower frequencies than had been tested previously (Kastak and Schusterman, 1998~. All three species had relatively high thresholds to sounds with a frequency of 100 Hz (89.9 to 116.3 dB, depending on the species). In addition, Kastak and Schusterman calculated ranges at which the sound from the ATOC source would be just detectable, not necessarily annoying, for all three species: 9 to 34 km for the California sea lion, 160 km for the harbor seal, and 279 km for the elephant seal.5 Using a different experimental design, Kastak and Schusterman compared low-frequency hearing sensitivity in air and in water for the same three pinniped species. They presented the argument that in water sound pressure is a more biologically appropriate measure than sound intensity.6 Using sound pressure sensitivity as the critical parameter, they found consistent correlations between hearing sensitivities and the environment in which the ani- mal spends the greater portion of its time. Thus, the elephant seal, which spends a greater proportion of time in water, has better underwater hearing sensitivity than the sea lion, which hears better in air than in water, and the harbor seal, which has almost equal sensitivity in air and in water. 5Kastak and Schusterman performed these calculations assuming a simplified propagation model of spherical spreading to a distance of 1 km (20 log R), followed by propagation described by 51Og R at greater distances. The authors did not give a reason for using this combination. 6Sound intensity is the power of the sound or pressure squared. This raises questions about the biological relevance of comparing in-air and underwater sounds by ``correcting for,, intensity rather than by simply comparing pressure levels.

OCR for page 41
ASSESSMENT OF CONTINUING RESEARCH NEEDS 51 Ridgway et al. (1997) tested the hypothesis that the sensitivity of white whale hearing might diminish with depth. To test the effect of depth, two trained white whales made dives to a platform at 5, 100, 200, and 300 m in the Pacific Ocean. During dives to the platform for up to 12 minutes, the whales whistled in response to 500-ms tones projected at random intervals to assess hearing threshold at each of the four depths. Analysis of response whistle spectra, whistle latency to tones, and hearing thresholds showed that the increased hydrostatic pressure at depth changed each whale's whistle response at depth but did not attenuate hearing. Hearing is attenuated in the aerial ear of humans and other land mam- mals when tested in pressure chambers due to changes in middle-ear impedance that result from increased air density (Fluur and Adolfson, 1966; Pantev and Pantev, 1979; Levendag et al., 1981~. The finding that whale hearing is not attenuated at depth suggests that sound is conducted through whale head tissues to the ear without requiring the usual eardrum/ossicular chain amplification of the aerial middle ear. These first-ever hearing tests in the open ocean demonstrate that zones of audibility for human-generated sounds are as great at 300 m, and potentially much deeper, as in shallow water. (These tests could not have been performed without trained whales.) Au et al. (1997) tested two members of the dolphin family (false killer whale and Risso's dolphin) for their responses to the ATOC signal using a behavioral paradigm. They concluded that neither species could be negatively affected by the ATOC experiments due to their high auditory thresholds (139 and 141 dB, respectively) for the ATOC signal.7 Based on calculated transmission signal loss, Au et al. concluded that at a horizontal range of 0.5 km these whales would not hear the ATOC signal. They also concluded that the ATOC signal is unlikely to harm baleen whales or to mask their vocalizations, but these conclusions are based on calculations and interpretations of source levels of whale vocalizations without using actual behavioral data. Source levels for baleen whale vocaliza- tions have been calculated from measurements for single animals at considerable distances from the hydrophore. Such vocalization levels may be more appropri- ately compared to humans shouting, rather than to conversational speech levels. Members of the same species probably are rarely, if ever, exposed to vocaliza- tions at levels as high as the maxima reported by some researchers (cited in Richardson et al., 1995; Au et al., 1997~. The reported maxima also may be outliers or erroneous because available data are relatively scarce; thus, assump- tions about noise levels that may or may not disturb whales should not be based on such measurements. 7According to the authors, small toothed whales swimming directly above the ATOC source will not hear the signal unless they dive to 400 m. The authors used a propagation loss model calculated for the source off Kauai. They then stated the horizontal distance without a depth, so it is unclear whether the animals would have to be at the surface for these measurements to be valid.

OCR for page 41
54 MARINE MAMMALS AND LOW-FREQUENCY SOUND animal should be viewed as only a temporary substitute for average hearing capabilities across members of wild populations (such as the bottlenose dolphins that have been studied for several years in Sarasota Bay, Florida). In the absence of large datasets on the hearing capabilities of multiple animals, Terhune and Turnbull suggested a "broad-brush approach" in using available data that assumes average auditory capabilities, for example, rather than using the lowest thresholds measured as standard values. This approach also may be useful when comparing auditory data for various types of whales. Knowledge of the individual differ- ences in hearing sensitivity within species may help explain the large differences observed in the behavior of individual animals (e.g., gray whales) when con- fronted with a noise source in their path of travel (Tyack and Clark, 1998~. Audiograms from many individuals, preferably from the wild, are critically important as a baseline to understanding the hearing capabilities of populations. . Determine sound pressure levels that produce temporary and per- manent hearing loss in marine mammals. In humans, exposure to intense sounds has the potential to produce a number of temporary or permanent aftereffects, depending on the level and duration of the exposure. These aftereffects include reductions in perceived loudness, ring- ing in the ears (tinnitus), and changes in perceived pitch (Kryter, 1985; Ward, 1997~. The most studied aftereffect is loss of hearing sensitivity, commonly known as temporary and permanent threshold shift (TTS and PTS, respectively). Repeated exposures that produce TTS eventually produce a PTS. The fact that some hearing losses last only minutes, hours, or days suggests that some cochlear structures have the ability to recover from whatever damage is inflicted by the exposure stimulus. In terrestrial mammals the receptor cells known as outer hair cells are known to be far more susceptible to acoustic damage than are the less numerous inner hair cells (Saunders et al., 1991), and temporary inactivation of the outer hair cells is presumed to be the primary factor in exposure-induced hearing losses of about 30 dB (in air) or less that last only a few hours. Tremen- dous variability exists in the susceptibility of individual ears to both TTS and PTS; in some human survey studies the subjects are even partitioned into differ- ent groups on the basis of the apparent "toughness" of their ears. For a typical human ear exposed daily to essentially continuous noise in the workplace, 90 dB (in air) has been widely adopted as the point at which precautions need to be taken (see Appendix D). Under U.S. regulations, for every additional 5 dB of exposure, the allowable duration of exposure is halved. However, the 90-dB value (in air) in not ubiquitous across animals and cannot be compared directly with in-water values. Across mammalian species there are known to be quite large differences in the sound levels required to damage the auditory system. For example, 12 min- utes of exposure to a 1,000-Hz tone of 120 dB (in air) sound pressure level can

OCR for page 41
ASSESSMENT OF CONTINUING RESEARCH NEEDS 55 produce more hearing loss in a chinchilla than 12 hours of the same exposure in a squirrel monkey (Hunter-Duvar and Bredberg, 1974~. Decory et al. (1992) made direct comparisons of hearing loss and cochlear damage in cat, guinea pig, and chinchilla using several exposure frequencies and a fixed duration of 20 minutes. Across exposure frequencies there was between 10 and 25 dB more hearing loss in the chinchilla than in the cat, with the guinea pig being intermedi- ate between the two. At the highest exposure frequency (8 kHz), there were approximately 10 times the number of damaged hair cells in the chinchilla than in the cat. Because of certain controls implemented by the experimenters, it is possible to rule out any contribution of differences in the outer-ear system to these differences in susceptibility, but it is not yet possible to say with certainty what the relative contributions were from differences in the middle-ear system and differences in cochlear mechanisms. Decory et al. (1992) presented an interesting argument about the possible contribution to damage made by the angular displacement to which a stereocilium is exposed in each of these species, a suggestion that warrants further study and attention by comparative anatomists studying marine mammals. The work of Luz and Lipscomb (1973) suggests that large discrepancies exist across species for susceptibility to impulse noise as well as continuous noise. There is some evidence to indicate that similar differences in susceptibility to exposure-induced hearing loss exist in marine mammals as well. Kastak and Schusterman (1996) suggested that exposure (in air) to a band level only about 10 to 25 dB above absolute sensitivity was adequate to produce 8 dB of TTS in a harbor seal. By comparison, the work of Schlundt et al. (2000) suggests that for the bottlenose dolphin the sound levels necessary to produce a small masked TTS (in water) must be between about 115 and 150 dB above absolute sensitivity. The existence of such large differences among species in terms of suscepti- bility to exposure-induced hearing loss creates two problems for regulators of human-generated sound in the ocean. First, critical exposure levels cannot be extrapolated from a few species, although it may be possible to identify a set of representative species for the initial studies (see Box 5.1~. Second, it almost certainly will never be possible to specify one single value of sound level at which damage to the auditory system will begin for all, or even most, marine mammals. An interesting characteristic of TTS and PTS is that for terrestrial mammals the maximum hearing loss typically occurs in a frequency region above the exposure frequency (McFadden, 1986~. Investigators studying TTS in marine mammals should design their experiments to obtain information about any upward shifts in maximum hearing loss that exist in marine mammal ears. In terrestrial mammals, cochlear mechanics are known to be somewhat different in the apical (low-frequency) regions of the cochlea than in the middle and basal regions. Further, some species of bats have cochleae that are highly specialized to process stimuli from certain narrow frequency bands that are important to their survival.

OCR for page 41
56 MARINE MAMMALS AND LOW-FREQUENCY SOUND The extent to which these differences affect the growth and spread of TTS and PTS is not completely understood, but because there may be corresponding regional differences in the cochlear mechanics of those marine mammals that depend on low-frequency sounds, it will be important to study TTS following exposures to low-frequency sounds as well as to mid- and high-frequency sounds. A serendipitous occurrence of TTS was documented in a captive harbor seal exposed to construction noise (Kastak and Schusterman, 1996~. Although the exposure was in air, the TTS (also measured in air) nonetheless reveals the levels above threshold at which the seal ear could be affected and the duration of the effect after prolonged repeated exposure. The noise was present over 6 days with 6 to 7 hours of intermittent daily exposure. The longest continuous exposure was believed to be about 1.5 hours. The third octave sound pressure level, centered at 100 Hz, varied from 75 to 90 dB (in air), which was only 10 to 25 dB above threshold for this seal. On day 6 of daily exposures, an 8-dB temporary threshold shift was measured at 100 Hz (the only frequency tested). Also, the seal's false alarm error rate in responding to sounds increased by 23 percent, which the authors suggested may have been the result of exposure-induced tinnitus. Hear- ing sensitivity recovered completely in about 1 week. Kastak et al. (1999) went on to test underwater TTS in three species of seals. They used test frequencies ranging from 100 to 2,000 Hz and "octave-band noise exposure levels that were approximately 60-75 dB SL (sensation level at center frequency)." They found TTS averaging 4.8 dB for one harbor seal, 4.9 dB for two California sea lions, and 4.6 dB for one northern elephant seal. The animals' hearing returned to baseline levels when tested within 24 hours of noise exposure. Schlundt et al. (2000) reported measures of TTS in four bottlenose dolphins (three females and one male) following exposures of differing levels at differing frequencies. The exposures were always 1 second in duration because the experi- ment was designed to determine the effects of a sonar sound commonly used by the U.S. Navy. The behavioral task used was imaginative and requires some description. A platform contained two stations, each having a prescribed place for the animal to position itself in front of a speaker. Under instructions from its trainer, a dolphin would position itself at the first station where it would first be exposed to a sound having a duration of 1 second. Immediately after receiving the first sound, the animal swam to the second station where it was presented with a series of 250-ms tones of fixed frequency and varying level. The animal was trained to whistle whenever it heard a tone from this second speaker, and the level of these test tones was adjusted up and down in accordance with the animal's responses to the series of test tones. After extensive training with this procedure, the level of the 1-second first sound was increased, with the objective of producing 6 dB of hearing loss as detected using the series of second tones. Different animals were tested with the exposure and test tones set to 3, 20, and 75 kHz. In part because of the background level at the test location, the series of second tones was actually masked by a broadband noise whose level could be varied.

OCR for page 41
ASSESSMENT OF CONTINUING RESEARCH NEEDS 57 Thus, the measure extracted from these data was in fact the difference in the level of the second tone necessary for a fixed level of detection performance in the presence of the background masker. Accordingly, it is probably better character- ized as masked temporary threshold shift (MTTS) in order to distinguish it from standard (unmasked) measures of TTS. This paradigm, while unusual, may provide a more realistic estimate of threshold shifts for animals tested in a noisy environment. The results were that small amounts of MTTS were observed when the exposure sounds were in the range of 192 to 201 dB, and this was true for all three exposure frequencies. (As already noted, this corresponds to exposures approximately 1 15 to 150 dB above absolute sensitivity, values that are higher than the nominal 90 dB commonly cited for humans.) Further, the animals exhibited varying degrees of behavioral disturbance at exposure levels about 12 to 15 dB below the values necessary to produce MTTS, a fact that Schlundt et al. interpreted to mean that dolphins, and perhaps other whales, may naturally avoid sounds that pose a threat to their hearing. To test this possibility, whales could be monitored closely during the ramp-up phase of experimental presentations of the ATOC signal near the test subjects. There is a particular priority for obtaining TTS data for endangered baleen whales. Many environmental impact statements (EISs) suggest levels of 150 to 160 dB as safe exposure levels for marine mammals. This is particularly difficult to justify with animals for which no audiometric data exist. Even if there is not enough time to conduct a complete TTS study on a stranded whale, it would still be particularly useful to test for TTS after several tens of minutes of exposure to 160-dB noise. . Determine condition of important cochlear structures in wild marine mammals using postmortem examinations (this topic did not appear in earlier reports). For many locations around the world the ocean is an extremely noisy place due to shipping, petroleum exploration or drilling, underwater explosions (Ketten et al., 1993), and other human activities. Accordingly, large numbers of marine mammals are already being exposed to anthropogenic sounds on a regular basis, and many of these sounds are of high intensity. Unfortunately, little is known about the auditory consequences of these exposures even though that knowledge could be extremely informative about the possible consequences of exposure to other intense sources such as those used for ATOC and LEA sonar. One obvious source of information would be behavioral or physiological measures of hearing from animals living at these especially noisy sites. In the absence of such difficult- to-obtain information, there is a less direct, but still potentially informative, approach that deserves attention. Specifically, postmortem examinations of the cochleae of marine mammals have the potential to correlate noise levels and

OCR for page 41
58 MARINE MAMMALS AND LOW-FREQUENCY SOUND hearing damage. Many marine mammals of various species die each year, in strandings, fish nets, and accidents. If the cochleae of some of these animals could be collected, preserved correctly, and sent to scientists capable of histologi- cal examination of the cells known to be damaged by exposure to intense sounds, some idea about the current (ambient) level of cochlear damage and, by impli- cation, hearing loss gradually would emerge. Eventually, this information could i' provide a set of baselines for use in establishing new regulations and against which new exposures could be measured. Much is currently known about the progression of cellular damage that occurs in terrestrial mammals following noise exposures of various types and durations (see Saunders et al., 1991), meaning that the cochlear physiologists have good expectations about what to look for, and where, in the cochleae of marine mam- mals. Damage of particular types and extents will be definitive evidence of permanent extensive hearing loss, and less severe damage will be evidence of less extensive hearing loss. Of special interest should be the cochleae of young marine mammals because they are the least likely to have cochlear deterioration attributable to factors other than noise exposure (such as the presbyacusis attrib- utable to aging). In terrestrial mammals, at least, hearing loss induced by a number of agents s known to progress from the basal (high-frequency) part of the cochlea toward the apical (low-frequency) part. Assuming that this pattern of progression exists in marine mammals as well, discovering damage to apical regions of the cochlea that is not accompanied by correspondingly greater damage to basal regions will be strong evidence that the damage was induced by an intense low-frequency sound source. Once frequency maps are established for the cochleae of various representative mammals (Ketten, 1994, 1997), localized damage at a particular region of the cochlea can be traced back to sound sources with maximum energy in a particular frequency band. It is important to undertake a parallel effort to monitor noise exposure in areas where strandings are well monitored. Regular monitoring of noise, strandings, and cochlear damage is required to move research from an anecdotal correlation toward causal links between noise and strandings. When implementing this procedure, the so-called half-octave shift in PTS needs to be taken into account; see McFadden (1986~. Because this idea about gathering information concerning the current state of hearing loss in marine mammals arose late in its deliberations, the Committee was not able to determine how best to implement a program of collecting cochleae from dead marine mammals. Clearly, the SWAT (Standard Whale Auditory Test) teams described elsewhere in this report should include personnel capable of extracting and preserving the peripheral auditory systems of marine mammals. In addition, it might be possible to train other personnel that are likely to come into contact with potential specimens. All interested governmental agencies should be advised about the importance of this collecting work and should be encouraged to implement policies that minimize any existing impediments to its

OCR for page 41
ASSESSMENT OF CONTINUING RESEARCH NEEDS 59 efficient progression. Because of the impossibility of predicting the location of specimens that might become available for this study, and because specimens can deteriorate relatively rapidly, transportation of auditory specimens across inter- national borders for purposes of histological examination should be made as simple as possible. This is especially important for vulnerable endangered species, where the Convention on International Trade in Endangered Species of Wild Fauna and Flora (CITES) permit processes inhibit timely delivery of samples to appropriate experts. . Determine morphology and sound conduction paths of the auditory system in various marine mammals. No studies have been published since 1994 that reveal new information about the structural details of the auditory system in marine mammals, although ONR has funded a project in the laboratory of Darlene Ketten (Woods Hole Oceanographic Institution) to examine potential auditory pathways in the lower jaw and skull of dolphins as part of a project to model sound conduction in these animals. The existence of a second pathway to the ear may be confirmed by these studies (Ketten, 1994, 1997~. Determine temporal-resolving power for various marine mammals (this topic did not appear in earlier reports). In humans and other terrestrial mammals, hearing sensitivity varies with signal duration. Specifically, the longer the duration, the less strong the signal needs to be for equal detectability. This relationship holds out to a limiting duration of about 300 to 500 ms, beyond which the signal level for equal detect- ability remains constant. What is most interesting about this trade-off between duration and level is that level declines (or increases) about 3 dB for every doubling (or halving) of duration below the limiting duration. That is, power is traded almost perfectly for time (e.g., Plomp and Bouman, 1959), suggesting that energy is the feature of the signal that determines detectability. This temporal integration function rises about 15 dB as the signal duration decreases from 300 ms to 10 ms in any animal using signal energy as the basis for its performance. The implications of temporal-resolving power differ for marine mammals with low-frequency hearing versus those with high-frequency hearing. Malme et al. (1983, 1984) reported that the levels required to elicit a response in gray whales need to be as much as 50 dB greater for short impulses than for continuous sounds. These data suggest that this species may be disproportionately insensitive to very short sounds; that is, their temporal integration function appears to be much steeper or displaced from that of typical terrestrial mammals. Data on the hearing capabilities of other baleen whales may suggest that long temporal inte- gration functions, with a corresponding relative insensitivity to transient sounds,

OCR for page 41
60 i' MARINE MAMMALS AND LOW-FREQUENCY SOUND is to be expected in other baleen whales. One implication is that these species should have generally poor temporal resolution; for example, they should have difficulty detecting a brief silent gap in an ongoing waveform. In contrast to baleen whales, characterized by their sensitivity to low- frequency sounds, the toothed whales, characterized by high-frequency hearing, do appear to exhibit short temporal integration functions. Measures of temporal resolution capability and temporal integration times were obtained in several species of toothed whales by Dolphin (1995, 1996) using evoked potential tech- niques. The species examined demonstrated temporal resolution capabilities exceeding those observed in humans and other terrestrial mammal species. Thus, one would expect these species to exhibit high sensitivity to brief, transient sounds. Moreover, the possession by dolphins of the capability for both very high temporal resolution and extremely sharp frequency resolution is enigmatic and points out our lack of understanding of the auditory processing of these animals. Accordingly, it would be extremely valuable to obtain information about temporal processing in a wider range of species from each of the groups identified in Box 5.1. Although the evidence suggesting differences in temporal processing capa- bilities between certain marine mammals and terrestrial mammals is unquestion- ably scarce, a logical argument can be made for such a difference. Species that are highly dependent on low frequencies may have acquired long time constants for temporal integration, making them highly insensitive to very short sounds (the rise time of their auditory systems may be so long that the response to impulsive sounds is weak), whereas those species that depend on high frequen- cies and/or echolocation have evolved auditory systems with short time constants and high temporal resolution capabilities. If this is true, it would have major impact on what kinds of sounds are regarded as dangerous for different species; namely, it may be that brief sounds, even quite intense ones, are not unduly dangerous to hearing in some species of baleen whale. These facts emphasize that temporal resolution can vary substantially across species having hearing that is specialized for operation in different frequency regions and they lend plausibil- ity to the speculation above about temporal resolution possibly being poor in those marine mammals specialized to communicate with low-frequency sound. EFFECTS OF LOW-FREQUENCY SOUNDS ON THE FOOD CHAIN Aim: To determine whether low-frequency sounds affect the behavior and physi- ology of organisms that serve as part of the food chain for marine mammals (NRC, 1994, pp. 53-54~. The most serious effects of noise on potential prey species are those that involve growth and reproduction. Increases in noise (above ambient levels) have been implicated in reduced growth and reproduction in a variety of marine organ-

OCR for page 41
ASSESSMENT OF CONTINUING RESEARCH NEEDS i' 61 Ems in tanks, although these data are still very limited and in need of replication. A single study of the effects of intense sound on the auditory systems of fresh- water fish has been published (Hastings et al., 1996~. In this study, freshwater fish were exposed to varying levels of a sinusoidal sound, and some evidence of auditory damage was reported for sounds above 180 dB when fish were exposed to continuous signals for 4 hours and then allowed to survive for several days before damage was assessed. It should be noted that these sounds were very different in intensity, duty cycle, and duration from those used in the MMRP studies and the fish used in this experiment could not escape the sound source. Additional studies are needed, particularly of fish species that are endangered, important commercially, or are a component of the food chain of marine mammal species. Growth rates in two fish species, sheepshead minnows (Cyprinodon variegates) and killifish (Fundulus similis), were significantly lower in aquariums exposed to noise 20 to 30 dB above ambient levels in the natural habitat (Banner and Hyatt, 1973~. Tanks with noise levels 20 dB higher than ambient levels reduced the viability of eggs in sheepshead minnows. Evidence reviewed by Corwin and Oberholtzer (1997) suggests that fish and perhaps some shark and amphibian species have the capacity to regenerate damaged hair cells in their auditory and balance organs (see also Lombarte et al., 1993~. To the extent that this ability is widespread in those species, they may be at lesser long-term risk from exposure to intense sounds than are marine mammals, which are presum- ably like terrestrial mammals in being incapable of regenerating new receptor cells to replace damaged ones (e.g., Hastings et al., 1996; Corwin and Oberholtzer, 1997~. No assessments of pre-ATOC shark abundances were made, nor has the potential attraction of sharks by low-frequency ATOC sound been investigated, despite extensive data in the literature showing that low-frequency sounds, such as those used by ATOC, attract sharks (e.g., Myrberg, 1972, 1978; Myrberg et al., 1976~. The potential for redistribution of sharks cannot be ignored, and some effort should be made in the future to monitor any ATOC source with appropriate methods (methods that would not alter the behavior of sharks and other organ- isms in significant ways) to determine if sharks are attracted to the site. Effects of intense sound have been observed in a shrimp species (Crangon crangon) (Lagardere, 1982~. Shrimp exposed to noise levels 20 to 30 dB higher than normal ambient levels exhibited reduced growth and reproduction and increased aggression and mortality relative to a control group. Although population surveys have indicated the presence of endangered turtle species at both ATOC sites (the Kauai EIS mentions green sea turtles [Chelonia mydas], leatherback turtles [Dermochelys coriacea], olive Ridley turtles [Lepidochelys olivacea], and hawksbill turtles [Eretmochelys imbricataj) no specific results for studies of hearing or behavioral observations on any shark or turtle species were presented by the MMRP scientists.

OCR for page 41
62 MARINE MAMMALS AND LOW-FREQUENCY SOUND POTENTIAL NONAUDITORY ACOUSTIC EFFECTS ON MARINE ANIMAL HEALTH Extremely intense sound can result in injury to many bodily organs and physiological processes. Aside from the direct effects of nearby blasting from explosives, intense sound can cause injury to lungs and other air-containing spaces. In addition, there may be direct effects on the nervous system. Extremely intense sound can also affect the vestibular system and can cause disorientation. In humans these vestibular impacts result in readily observable nystagmus, an uncontrolled movement of the eyes (Stephens and Ballam, 1974~. Studies of these effects in humans and other terrestrial species are being funded by ONR. Research by Crum and Mao (1996) suggests the possibility of potential effects that only occur in certain physiological states that would be very difficult to study in nonhumans. For example, human divers are susceptible to decom- pression sickness ("bends"), a disabling and sometimes fatal condition in which bubbles of nitrogen gas form in the blood, joints, and other tissues. Low-frequency sound might induce bends episodes in human divers whose tissues are saturated with gas due to breathing pressurized gas at depth, that would not otherwise occur. Crum and Mao (1996) showed experimental data suggesting that intense (160 to 220 dB) low-frequency sound may induce bubble growth in tissues (see also Lettvin et al., 1982), and therefore divers ensonified with low-frequency pulsed sound when they are near decompression limits could be severely injured (Crum and Mao, 1996~. This is unlikely to be a problem with the ATOC sources because they are so far offshore from dive sites and in deep water, but it may be a problem with more powerful shallow-water sources, such as SURTASS-LFA. Although marine mammals do not carry a tank of pressurized breathing gas as human divers do, they make frequent dives to depths greater than 100 m, which may produce over 200 percent supersaturation of nitrogen in muscle tissue after repetitive dives (Ridgway and Howard, 1979, 1982~. In Ridgway and Howard's 1979 study, dolphins made 23 to 25 dives to 100 m (lo atmospheres) in 1 hour. Dolphins did not suffer from decompression sickness even with muscle nitrogen at supersaturated levels that would produce bends in humans (Ridgway and Howard, 1979~. However, Lettvin et al. (1982) and Crum and Mao (1996) suggested that sound exposure could induce bubble growth in blood. Crum and Mao (1996) suggested that this might be an issue for both humans and marine mammals. Therefore, it should be considered whether intense low-frequency sound might cause bubbles in the circulatory systems of whales returning to the surface after a long series of deep, but rapid, dives similar to those studied by Ridgway and Howard (1979~. At the present time, bubble formation in diving marine mammals must be considered as conjecture based on findings in unrelated studies. However, research on marine mammals should probably consider the issue after ONR studies on human subjects have been completed. If results from research on humans suggest that marine mammals could experience such effects,

OCR for page 41
ASSESSMENT OF CONTINUING RESEARCH NEEDS 63 sophisticated electronic tags that record depth of dive and other relevant data might reveal the extent to which such a nonauditory threat may be realistic in wild populations. As part of the SURTASS-LFA program, ONR funded research on the effects of low-frequency sound on human divers. These studies show that the likelihood of scuba divers aborting a dive is lowest at 250 Hz but rises for source frequen- cies both above and below 250 Hz. Sound levels were relatively nonaversive until reaching a received sound pressure level of 148 dB, at which point 15 percent of dives were aborted. Aversion at lower frequencies resulted from a sensation of vibration in the air-filled cavities in the head, chest, and abdomen. No vestibular effects were observed for sound up to 157 dB. Research on marine mammals should also be devoted to evaluating other physiological measures of general health, both short-term and long-term, that can be monitored non-invasively over time with tags. For now, these measurements are limited to heart rate and respiration. Data on stress and stress indicators in marine mammals is sorely lacking, and there is not even a baseline from which to determine normal values. At the same time, the significance of such reactions to stress in terms of reproduction and survival should be assessed. LONG-TERM ACOUSTIC MONITORING OF CRITICAL HABITATS The issue of cumulative impacts from human-generated noise is best dealt with as a habitat degradation issue. Undersea noise should be monitored in important marine mammal habitats (after these have been identified). This moni- toring effort should be designed in parallel with surveys of marine mammal distribution, abundance, and strandings using methods that allow temporal and spatial analysis of how noise may lead to changes in these population character- istics. These data are particularly important for populations that are either not recovering or are declining due to habitat degradation and other causes. Monitor- ing should include the ambient noise field, marine mammal vocalizations, and transient noises, particularly in locations and times of the year in which marine mammals are known to be common. This monitoring optimally should also include or be coordinated with other assessments of habitat quality such as prey fields and chemical pollutants. Coordination of noise monitoring with stranding networks would enable more systematic and controlled evaluation of whether noise influences strandings and whether cochlear damage in stranded animals is associated with acute noise exposure. NMFS, the Navy, and other agencies with responsibilities for marine mam- mals or that conduct or permit activities that introduce significant levels of sound to the ocean should evaluate the costs and benefits of an array of acoustic receivers designed to monitor both human-generated sound in the ocean and the vocaliza- tions of whales in acoustic hotspots (NRDC, 1999~. One possibility is to use existing arrays such as the IUSS (JOI, 1994; Clark, 1995; Gisiner, 1998) devel-

OCR for page 41
64 MARINE MAMMALS AND LOW-FREQUENCY SOUND oped by the U.S. Navy to detect submarines. Evaluation of the appropriateness of the IUSS should determine whether the bandwidth and geographic coverage of the IUSS is adequate for the task of monitoring ambient noise and marine mam- mals or whether it is necessary to design and build an array of sensors specifically to monitor marine mammals. Such a system could be automated to activate when significant sounds are detected. Whales could be located and tracked in real time and in three-dimensional space, thus identifying natural paths and avoidance paths. This capability was demonstrated in the Whales '93 program in which the IUSS was used to routinely detect, locate, and track blue, finback, and humpback whales in the North Atlantic Ocean (JOI, 1994~. Hundreds of thousands of whale vocalizations were documented, allowing the description of seasonal movements of the whales. Autonomous underwater recorders, sonobuoys, or towed arrays of hydrophores can be used in areas where (or at times when) more intensive moni- toring is desired (Richardson et al., 1986; Thomas et al., 1986; Moore et al., 1999~.