Mechanisms of Acoustic Trama

Mechanisms of Acoustic Trauma

Temporary and Permanent Threshold Shifts

Noise trauma is a well-investigated phenomenon for air-adapted ears (see Lehnhardt, 1986;

Lipscomb, 1978; and Richardson, et al., 1991 for reviews). For the sake of completeness in the

following discussion, noise trauma has been divided into lethal and sublethal impacts. Lethal

impacts are those that result in the immediate death or serious debilitation of the majority of

animals in or near an intense source; i.e., profound injuries related to shock wave or blast effects

which are not, technically, pure acoustic trauma. Lethal impacts are discussed briefly at the end

of this section. Sublethal impacts are those in which a hearing loss is caused by exposures to

sounds that exceed the ear's tolerance to some acoustic parameter; i.e., auditory damage occurs

from metabolic exhaustion or over-extension of one or more inner ear components. Of course,

sublethal impacts may ultimately be as devastating as lethal impacts, causing death indirectly

through behavioural reactions, such as panic, as well as impaired foraging or predator detection,

but the potential for this type of extended or delayed impact from any sound source is not well

understood for any mammal.

To determine whether any one animal or species is subject to a sublethal noise impact from a

particular sound requires understanding how its hearing abilities interact with that sound.

Basically, any noise at some level has the ability to damage hearing by causing decreased

sensitivity. The loss of sensitivity is called a threshold shift. Not all noises will produce

equivalent damage at some constant exposure level. The extent and duration of a threshold shift

depends upon the synergistic effect of several acoustic features, including how sensitive the

subject is to the sound. Most recent research efforts have been directed at understanding the

basics of how frequency, intensity, and duration of exposures interact to produce damage rather

than interspecific differences: that is, what sounds, at what levels, for how long, or how often will

commonly produce recoverable (TTS - Temporary Threshold Shift) vs permanently (PTS) hearing

loss.

Three fundamental effects are known at this time:

1) the severity of the loss from any one signal may differ among species.

2) for pure tones, the loss centers around the incident frequency.

3) for all tones, at some balance of noise level and time, the loss is irreversible.

Hearing losses are recoverable (TTS - temporary threshold Shift) or permanent (PTS)

primarily based on extent of inner ear damage the received sound causes (see Lipscomb 1978,

Lehnhardt 1986, Richardson et al. 1991 for reviews). Temporary threshold shifts (TTS) will be

broad or punctate, according to source characteristics. The majority of studies have been

conducted with cats and rodents, using relatively long duration stimuli (> 1 hr.) and mid to low

frequencies (1-4 kHz) (see Lehnhardt, 1986, for summary). Inner ear damage location and

severity are correlated with the power spectrum of the signal in relation to the sensitivity of the

animal. Virtually all studies show that losses are centered around the peak spectra of the source

and are highly dependent upon the frequency sensitivity of the subject. For narrow band, high

frequency signals, losses typically occur in or near the signal band, but intensity and duration can

act synergistically to broaden the loss.

It has also been established that repeated exposures to TTS level stimuli without adequate

recovery periods can induce permanent, acute threshold shifts. Liberman (1987) showed that

losses were directly correlated with graded damage to the outer and inner hair cells, and that the

majority of cells recover. With short duration, narrow band stimuli, recovery periods can vary

from hours to days. In effect, the duration of a threshold shift, is correlated with both the length

of time and the intensity of exposure. In general, if the duration to intense noise is short and the

noise is narrow, the loss is limited and recoverable. Based on both the available experimental data

and on human data from occupational hearing loss, moderate to protracted exposures to a signal

intensity of 80 dB or more over the individual threshold at each frequency for land species is

required for significant threshold shifts (see NIH./CDC, 1990; Yost, 1994 for overview). These

findings led to the current allowable limit of 80-90 dB re 20 µPa for human workplace exposures

for broad spectrum signals, as well as an allowance of the 3-5 dB increase in exposure as a tradeoff

for halving of exposure times (Lehnhardt, 1986). While the commonality of 80 dB suggests

that TTS is a dynamic range dependent phenomenon which is probably related to fundamental

mammalian inner ear mechanisms, this specific dB criterion for exposure limits cannot be

supported nor refuted with current data for marine mammals, particularly since some marine

species have inner ear adaptations that could alter these responses (see Marine Mammal Issues

section).

Given the complex nature of the interaction of species-specific hearing parameters with each

signal feature a simplistic rule for species dependent impacts based on any one acoustic feature or

hearing characteristic is not possible, as is shown in a quick review of Table 3. Some broad

trends do emerge, however, from inter-species comparisons of sources that induce TTS in air.

At the grossest level, TTS effects from approximately equivalent exposures appear to be

inversely related to weight or mass; i.e., effects were less pronounced in humans than in cat or in

chinchilla, but this may be a secondary effect of frequency sensitivities differing also with animal

size. The majority of effects appear to be species independent, suggesting that basic cochlear

mechanisms may be the dominating factor. Effects that were common to all species were the

following:

1. Shifts were strongly dependent on interactions of timing, level, and frequency.

2. Cumulative or compound effects are common.

2. Asymptotic shifts appear to depend on similar metabolic and mechanical fatigue

phenomena.

3. Hearing impaired individuals have approximately the same absolute exposure limit for TTS

as unimpaired individuals, which is manifested in an apparently smaller exposure window

prior to TTS.

4. Effects spread primarily upward in frequency, which is a reflection of the basilar

membrane's tonotopic organization and the asymmetric distribution of the traveling wave

envelope (Fig. 4).

5. Frequency discrimination is unaffected.

6. Temporal integration is reduced.

Effects that showed strong species dependence were:

1. Loss at a particular frequency are correlated with species sensitivity.

2. Losses at all frequencies are correlated with metabolic, hair cell, and neural differences

throughout the cochlea.

The majority of PTS effects are minimally species dependent, but nevertheless equally

complex. One important aspect of PTS is that signal rise-time and duration of peak pressure are

significant factors. If the exposure is short, hearing is recoverable; if long, or has a sudden,

intense onset and is broadband, hearing, particularly in the higher frequencies, can be permanently

lost (PTS). Experimentally, PTS is induced with multi-hour exposures to narrow band noise. In

humans, PTS results most often from protracted, repeat intense exposures (e.g., occupational

auditory hazards from background noise) or sudden onset of intense sounds (e.g., rapid, repeat

gun fire). Sharp rise-time signals have been shown also to produce broad spectrum PTS at lower

intensities than slow onset signals both in air and in water (Lipscomb, 1978; Lehnhardt, 1986;

Liberman, 1987). Hearing loss with aging (presbycusis) is the accumulation of PTS and TTS

insults to the ear. Typically, high frequencies are lost first with the loss gradually spreading to

lower frequencies over time.

In experiments, multi-hour exposures to narrow band noise are used to induce PTS. As noted

above, most mammals with air-adapted ears incur losses when the signal is 80 dB over threshold.

TTS has been produced in humans for frequencies between 0.7 and 5.6 kHz (our most sensitive

range) from underwater sound sources when received levels were 150-180 dB re 1 µPa (Smith

and Wojtowicz 1985, Smith et al. 1988). Taking into account differences in measurements of

sound pressure in air vs. water (equations 4 and 5), these underwater levels are consistent with

the 80-90 dB exposure levels that induce TTS in humans at similar frequencies in air. Sharp risetime

signals produce broad spectrum PTS at lower intensities than slow onset signals both in air

and in water (Lipscomb 1978, Lehnhardt 1986).

Blast Effects

Simple intensity related loss is not synonymous with blast injury. Acoustic trauma induced by

sudden onset, loud noise ( a "blast" of sound) is not synonymous with blast trauma, nor are noise

and blast effects of the same magnitude. Blast injuries generally result from a single exposure to

an explosive shock wave which has a compressive phase with a few microseconds initial rise time

to a massive pressure increase over ambient followed by a rarefactive wave in which pressure

drops well below ambient.

Blast injuries may be reparable or permanent according to the severity of the exposure and are

conventionally divided into three groups based on severity of symptoms, which parallel those of

barotrauma:

MILD - Recovery

Pain

Vertigo

Tinnitus

Hearing Loss

Tympanic tear

MODERATE - Partial loss

Otitis media

Tympanic membrane rupture

Tympanic membrane hematoma

Serum-blood in middle ear

Dissection of mucosa

SEVERE - Permanent loss - death

Ossicular Fracture/Dislocation

Round/Oval window rupture

CSF leakage into middle ear

Cochlear and saccular damage

Moderate to severe stages result most often from blasts, extreme intensity shifts, and trauma;

i.e., explosions or blunt cranial impacts that cause sudden, massive systemic pressure increases

and surges of circulatory or spinal fluid pressures (Schuknecht, 1993). Hearing loss in these cases

results from an eruptive injury to the inner ear; i.e., with the rarefactive wave of a nearby

explosion, cerebrospinal fluid pressures increase and the inner ear window membranes blow out

due to pressure increases in the inner ear fluids. Inner ear damage frequently coincides with

fractures to the bony capsule of the ear or middle ear bones and with rupture of the eardrum.

Although technically a pressure induced injury, hearing loss and the accompanying gross

structural damage to the ear from blasts are more appropriately thought of as the result of the

inability of the ear to accommodate the sudden, extreme pressure differentials and over-pressures

from the shock wave.

At increasing distance from the blast, the effects of the shock wave lessen and even though

there is no overt tissue damage, mild damage with some permanent hearing loss occurs (Burdick,

1981, in Lehnhardt, 1986). This type of loss is generally called an asymptotic threshold shift

(ATS) because, as was found with protracted exposures in TTS experiments, ATS derives from a

saturation effect. Like TTS, the hair cells are damaged, but as in PTS, recovery is unlikely to take

place. Because ATS depends upon complex interactions of rise time and wave form, not simply

intensity at peak frequency, hearing losses are typically broader and more profound than simple

PTS losses.

There is no well defined single criterion for sublethal ATS from blasts (Roberto, et al., 1989),

but eardrum rupture, which is common to all stages of blast injury, has been moderately well

investigated. Although rupture per se is not synonymous with permanent loss (eardrum ruptures

have occurred at as little as 2.5 kPa overpressure and are strongly influenced by the health of the

ear), the incidence of tympanic membrane rupture is strongly correlated with distance from the

blast (Kerr, & Byrne, 1975). As frequency of rupture increases so does the incidence of

permanent hearing loss. In zones where >50% tympanic membrane rupture occurred, 30% of the

victims had long term or permanent loss.

Recent experimental work has shown that weighted sound exposure level is a more robust

predictor of permanent loss than peak pressure (Patterson, 1991). Data with weighted levels are

rare; overpressure data are more common and have been shown to be highly correlated with

received levels (Roberto et al., 1989). In general, complex and fast-rise time sounds cause

ruptures at lower overpressures than slow-rise time waveforms, and smaller mammals will be

injured by lower pressures larger animals. Of the animals tested to date, sheep and pig have ears

anatomically closest to those of whales and seals. The air-based data for pigs and sheep imply

that overpressures <70 kPa are needed to induce 100% tympanic membrane rupture. However,

cross-study/cross-species comparisons and extrapolations are risky because of radically different

experimental conditions as well as differences in acoustic energy transmission in the air and

water. The data available for submerged and aquatic animals imply that lower pressures in water

than in air induce serious trauma (Myrick et al., 1989; see also summary in Richardson, et al.

1991). For submerged terrestrial mammals, lethal injuries have occurred at overpressures >55

kPa (Yelverton, 1973, in Myrick, et al., 1989; Richmond, et al., 1989). In a study of Hydromex

blasts in Lake Erie the overpressure limit for 100% mortality for fish was 30 kPa (Chamberlain,

1976). The aquatic studies imply therefore that overpressures between 30 and 50 kPa are

sufficient for a high incidence of severe blast injury. Minimal injury limits in both land and fish

studies coincided with overpressures of 0.5 to 1 kPa.

Marine Mammal Issues

Major impacts from noise can be divided into direct physiologic effects, such as permanent

vs. temporary hearing loss, and those that are largely behavioral, such as masking, aversion, or

attraction. Although there is no substantial research accomplished in any of these areas in marine

mammals, behavioral effects have been at least preliminarily investigated through playback and

audiometric experiments, while marine mammal susceptibility to physiologic hearing loss is

virtually unexplored. Despite increasing concern over the effects on marine mammals of manmade

sound in the oceans, we still have little direct information about what sound frequencyintensity

combinations damage marine mammal ears, and at present there are insufficient data to

accurately determine acoustic exposure guidelines for any marine mammal.

Is acoustic trauma even moderately debatable in marine mammals? Recalling the paradox

mentioned earlier, there are a variety of reasons to hypothesize that marine mammals may have

evolved useful adaptations related to noise trauma. Vocalizations levels in marine mammals are

frequently cited as indicating high tolerance for intense sounds. Some whales and dolphins have

been documented to produce sounds with source levels as high as 180 to 220 dB re 1 µPa

(Richardson et al., 1991; Au, 1993). Vocalizations are accepted indicators for perceptible

frequencies because peak spectra of vocalizations are near best frequency of hearing in most

species, but it is important to recall that the two are not normally precisely coincident.

It must be borne in mind also that animals, including humans, commonly produce sounds

which would produce discomfort if they were received at the ear at levels equal to levels at the

production site, and arguments that marine mammals, simply by nature of their size and tissue

densities, can tolerate higher intensities are not persuasive. First, mammal ears are protected

from self-generated sounds not only by intervening tissues (head shadow and impedance

mismatches) but also by active mechanisms (eardrum and ossicular tensors). These mechanisms

do not necessarily provide equal protection from externally generated sounds largely because

the impact is not anticipated as it is in self-generated sounds. Our active mechanisms are

initiated in coordination and in anticipation of our own sound production. Just as the level of a

shout is not indicative of normal or tolerable human hearing thresholds, source level calculations

for vocalizations recorded in the wild should not be viewed as reliable sensitivity measures. As

was indicated earlier, while there is little question of anomalous dysfunction of the middle ear in

pinnipeds, middle ear function continues to debated for cetaceans. However, it is very

important to recall also that cetaceans do have very well developed middle ear anatomies,

including stapedial ligaments (Ketten, 1984; 1992) which argues that they have the capability for

middle ear attenuation responses. Further, the large head size of a whale is not acoustically

exceptional when the differences in pressure and sound speed in water vs. air are taken into

account. As noted earlier, ear separation in a bottle-nosed dolphin is acoustically equivalent to

that of a rat when the distances are corrected for the speed of sound in water. Exactly how

head size in water affects attenuation of incident sound at the inner ear has not been investigated

and remains an important open question.

Data from several pilot studies may, however, provide some useful insights into both facets

of the paradox. In one investigation (detailed below, Ketten et al, 1993; Lien et al. 1993), ears

from humpbacks that died following underwater explosions had extensive mechanical trauma

while animals that were several kilometers distant from the blasts and at the surface showed no

significant behavioral effects. These findings indicate adaptations that prevent barotrauma do not

provide special protection from severe auditory blast trauma, but it remains unclear whether lower

intensity purely acoustic stimuli induce temporary and/or acute threshold shifts in marine

mammals.

A second study compared inner ears from one long-term captive dolphin with a documented

hearing loss with the ears of one juvenile and two young adult dolphins (Ketten et al., 1995). CT,

MRI, and histologic studies of the oldest dolphin ears showed cell loss and laminar

demineralization like that found in humans with presbycusis, the progressive sensorineural hearing

loss that accompanies old age. The location and degree of neural degeneration in this animal

implied a substantial, progressive, hearing loss beginning in the high frequency regions. This too

is consistent with the pattern commonly observed in humans. Frequency-position estimates of the

elder animal's hearing loss done blind; i.e., without prior knowledge of its audiogram, predicted a

profound loss for all frequencies >58 kHz. A review of the animal's behavioral audiogram

subsequently showed that over a 12 year period this dolphin's hearing curve shifted from normal

threshold responses for all frequencies up to 165 kHz to no functional hearing over 60 kHz prior

to his death at age 28. For this animal at least, the conclusion was that significant hearing loss had

occurred attributable only to age-related changes in the ear. Similar significant differences in the

hearing thresholds of two Zalophus have also been reported by Kastak and Schusterman (1995)

that are consistent with age-related hearing differences between the animals but which are also

consistent with protracted exposures to construction noise.

Micrographs from young adult dolphin ears show several important cochlear duct cellular

adaptations that are markedly different from those of conventional land mammals and seals.

Transmission electron micrographic studies revealed dolphins have active fibrocytes in the spiral

ligament and four times as many cell layers in stria vascularis as any other mammal. The stria is

considered to be the principal dictator of mammalian cochlear metabolism. If these results are

confirmed in other dolphin ears, these structural differences could mean dolphins have faster hair

cell recovery times than air adapted ears and may therefore be less subject to temporary

threshold shifts than most land animals or pinnipeds.

Unfortunately, these data only beg the question. The problem of hearing loss has not been

realistically considered prior to this point in any systematic way in any marine mammal. In fact,

the most studied group, odontocetes, have generally been thought of as ideal underwater

receivers. A captive animal's age or history is not normally considered in analyzing its auditory

responses, and, in the absence of overt data (e.g., antibiotic therapy), we assume a test animal

has a normal ear with representative responses for that species. It is not clear that this is both

reasonable and realistic. Particularly when data are obtained from one animal, it is important to

question whether that hearing curve is representative of the normal ear for that species. The

pilot studies noted clearly suggest age and/or exposure to noise can significantly alter hearing in

marine mammals. In fact, in some cases (compare the two curves shown in figure 3a for

Tursiops), "individual differences" that are seen in "normal" audiograms for two animals from the

same species may be the result of undetected hearing loss in on of the animals. The fact that

some studies show losses in marine mammals consistent with age-related hearing changes and

disease considerably complicates the diagnosis and assessment of hearing loss from

anthropogenic sources based on small samplings of populations. Natural loss should be

considered in any animal for which there is little or no history, therefore the finding of a single

animal with some hearing decrement in the vicinity of a loud source cannot be taken as a clear

indicator of a population level hazard from that source. On the other hand, because of the

importance of hearing to these animals, it is also unlikely that a high incidence of loss will be

normally found in any wild population, and a finding of substantial hearing loss from, for

instance, a mass-stranding or fishery coincident with a long-term exposure to an intense source

would be appropriate cause for significant concern.

Given the minimal state of marine mammal data, the only comprehensive database that can be

brought to bear at this time for predicting physiologic impacts is from acoustic trauma studies of

land mammals and fish.

Few reports exist that detail injuries in marine mammals from blast induced trauma. Bohne et

al. (1985) reported on inner ear damage in Weddell seals that survived blasts, but they were

unable to determine exposure levels or number of exposures for each animal. There are scattered

reports of opportunistic examinations of animals exposed to large blasts, including one on otters

with extensive trauma from nuclear explosions (Richardson et al., 1991) that concluded that peak

pressures of 100-300 psi were invariably lethal. Recently, several humpbacks exposed to

TOVEX blasts were shown to have severe blast injuries (Ketten et al., 1993). TOVEX, like

Hydromex, is a TNT clone explosive similar to HBX-1 with a detonation velocity of ~7500 m/sec

(Ketten, 1994). Received levels in the humpbacks could not be calculated with confidence;

however, the charge weights associated with the injuries ranged from 1700 to 5000 kg. The

animals died within three days of the blasts, and the extent of the injuries found implied they were

close to the blast site. Mechanical trauma in these ears included round window rupture, ossicular

chain disruption, bloody effusion of the peribullar spaces, dissection of the middle ear mucosa

with pooled sera, and bilateral periotic fractures. These observations are consistent with classic

blast injuries reported in humans, particularly with victims near the source who had massive,

precipitous increases in cerebrospinal fluid pressure and brain trauma. There was no evidence of

ship collision or prior concussive injury in these humpbacks, and no similar abnormalities were

found in ears from humpbacks not exposed to blasts. These findings imply that despite

adaptations in whales and seals that minimize barotrauma, marine mammals are not immune to

blast trauma. Given the similarities of seal and whale ears to land mammal ears, it is clear that

explosions and the shock wave and intense transient sound field that result can produce both blast

injury and acoustic trauma in marine animals. More important, even though the whale ear is

ostensibly a fluid-to-fluid coupler, marine mammals, having retained an air-filled middle ear

(Ketten, 1994), are subject to all ranges of compressive-rarefactive/blast injury.

The level of impact from blast will depend on both an animal's location and, at outer zones, on

its sensitivity to the residual noise. Factors that are most important for trauma from explosive

sources are the following:

1. Topography

2. Proximity of ear

3. Anatomy and health of ear

4. Charge weight and type

5. Rise time

6. Overpressure

7. Pressure and duration of positive pressure phase

Topographic effects for open ocean are minimal for most boat deployed sources. Surface

reflections will have a significant effect on the blast and acoustic wave spread patterns at some

depth that is largely dependent on detonation depth. This effect also complicates predictions of

received levels for animals at surface or within the air-sea boundary layer.

The health of individual ears that may be impacted cannot be estimated in advance. It is

reasonable to assume an average distribution. Many explosives (TNT clones and water-gel

explosives;e.g., HBX, Tovex, etc.) currently in use have high detonation velocities and are

therefore effectively an instantaneous onset, high peak pressure, broad spectrum blast.

Consequently, effects of the acoustic signature and certainly of the blast wave from these charges

are likely to be similar in all species in the target area; i.e., individual hearing ranges are largely

irrelevant in assessing TTS/PTS and blast effects in the near field, except for those species that

have no discrete air pockets.

Although multiple parameters are associated with both lethal and sublethal effects, virtually all

studies agree fairly closely on baseline criteria for lethal or compulsory injury zones for fast-rise

time, complex waveforms: ~ 30-50 kPa peak overpressure in water and > 180 dB re 20 µPa in air

(~240 dB re 1 µPa in water), (Chamberlain, 1976; Yelverton and Richmond, 1981; Phillips et al.,

1989: Richmond, et al.,, 1989; Myrick, et al.,, 1989). If, for comparison, the lowest otter impact

estimate were chosen (100 psi), the impact range is substantially greater. Depending upon this

range of criteria, a lethal impact zone limit for a 1200 lb source could be placed at 40 m. (absolute

minimum, land mammal) or 300 m (conservative estimate of 100 psi based on otter observations).

For a 10,000 lb. charge, the equivalent min-max limits for a killing ground are 70 m to 800 m. If

a conservative average overpressure of ~30 kPa is used as the criterion, the lethality limit for both

large charges is approximately 100 m. in comparison to approximately 10 m. and 50 m. for the 9

and 50 lb. charges.

Criteria for differentiating PTS or ATS zones from TTS are less clear. For this discussion,

peak pressures of ~150 psi, which are consistent with 50% incidence of eardrum rupture (30%

hearing loss) in larger mammals were chosen to define PTS/ATS limits. For a 9 lb. charge,

pressures that result in significant auditory damage can be expected along a long axis radius of

nearly 50 m. from the source. For a 50 lb charge, the equivalent PTS/ATS radius is nearly 100

m. For the 1200 and 10,000 lb charges, the transitional lethal zones in which serious sublethal

injury will predominate are estimated as 300 m and 750 m, respectively Beyond these zones,

the relative incidence of PTS to TTS will largely depend on individual susceptibility. That is,

the variables that will determine TTS vs PTS are highly dependent on both species-specific and

individual ear factors.

There is consensus in the literature on the criteria for an outer limit for mild TTS zones. 5-15

psi is accepted as the frontier at which TTS and detectable injury become rare (Yelverton and

Richmond, 1981; Smith et al., 1985, 1988; Myrick et al., 1989; Roberto et al, 1989). This is also

the zone in which the greatest differences are found in effects among charge weights. For 9 lb.

charges, moderate incidence of TTS may be expected up to 700 m from the epicenter; the 50 lb

TTS zone could extend to 1600 m in contrast to a 5 and 10 km radius from the heavier charges

before the acoustic impact could be expected to drop precipitously.

Acoustic Devices, Fisheries, and Mitigation Measures

Potential impacts

Although the remainder of this discussion is concerned with purely physiologic elements of

the effects of sound, it is important first to note that acoustic trauma per se is only one side of a

significant effect coin.

Acoustic trauma is a very real and appropriate physiologic concern. It is also one for which

we can obtain a metric that will allow us to provide a usable limit. That is, given that we know

sound level X induces TTS while Y induces PTS, for frequency Z in a specific species, we can

apply these data to the estimated exposure curve for that species and determine its risk of

hearing loss. As discussed earlier, this is the basic principle behind both the 80 dB/5 dB rule

currently in use for workplace exposures. Because of the importance of hearing to marine

mammals, understanding how man-made sources may impact that sense is an important and

reasonable step towards minimizing adverse impacts from man-made sound sources in the

oceans, but it is imperative that we employ a scientifically valid, marine specific meter-stick for

underwater exposures.

Above all, it is equally important to consider that sub-trauma levels of sound can have

profound effects on individual fitness that propogate to the population level. These effects can

take the form of masking of important signals, including echolocation signals, intra-species

communication, and predator-prey cues; of disrupting important behaviors through startle and

repellence, or of acting as attractive nuisances, all of which may alter migration patterns or result

in abandonment of important habitats. Unfortunately, these issues are beyond the scope of this

document as well as the expertise of the author and therefore cannot be productively and

responsibly discussed here. Nevertheless, it is important to at least note the concern, and above

all to suggest that there is a substantial need for field monitoring of behaviors in wild populations

in tandem with controlled studies directed at expanding our audiometric data and understanding

of acoustic trauma mechanisms.

As indicated earlier, there are no discrete data at this time that provide a direct measure of

acoustic impact from a calibrated, underwater sound source for any marine mammal.

Preliminary data from work underway on captive cetaceans and pinnipeds (Ridgway, pers.

comm.; Schusterman, pers. comm.) suggest that odontocetes may have asymptotic responses

while pinnipeds are more similar to land mammals in their dynamic range for threshold shift

effects. This response difference as well as the difference in hearing ranges - if these data are

shown to be robust - suggest that pinnipeds are the more acoustically fragile group from most

anthropogenic sound sources and that odontocetes are relatively immune or require substantially

higher sound levels to incur TTS.

In terms of the specifics of tuna-marine mammal-echo-ranging device interactions, the principal

acoustic concern is to determine a balance of frequencies vs. level vs. duty cycle that will

effectively detect and census commercially viable schools at long ranges but will not repel the

target species nor harm marine mammals within that sound field. To accomplish these goals it is

necessary to determine and balance the following components:

1. What are the effective frequencies for longer range detection? Presumably this will

require a moderately low frequency for maximizing distance of detection balanced against

a need to detect relatively small targets.

2. What is the hearing curve of the target species for capture? This feature must be

considered in order to avoid startle or repellent effects in the fish schools that are to be

detected by the source.

3. What are the hearing curves for non-target species within the sound field? This has the

same concern as the second component, with a different end objective; i.e., to avoid

impact or harassment but is driven also by an additional desire to prevent long-term,

multiple exposure effects that can compound the probability of hearing loss.

Put simply, the device must be able to detect fish without cueing them but at the same time

avoid frequency-intensity-sensitivity combinations likely to impact non-targeted, acoustically

fragile species. Detection devices proposed recently (see Nero, 1996; Rees, 1996; Denny et al.

1997) commonly employ frequencies in the low to mid-sonic ranges (50-5000 Hz) with a wide set

of emission algorithms, including repeat pulsed signals, and, in at least one scenario,

explosive/high intensity impulsive source. Source levels proposed vary widely but can range as

high as 235 dB re 1 µPa at 1 m. These spectra are coincident with virtually all marine mammal

hearing ranges, and ironically may be well perceived by at least some fish species. In fact, for

clupeids, recent data show a coincident high frequency sensitivity that suggests convergence of

predator and prey auditory systems at both mid-sonic (2-4 kHz) and ultrasonic (20-40 kHz)

ranges (Popper, 1997). Rather than complicating the issue, this coincidence may prove beneficial

by driving the frequency choice in the same direction; i.e., avoiding these frequencies may

maximize the utility of the device for finding fish without disturbance of the school while

minimizing the probability of its impact on marine mammals.