Effects of Sound
on Fish
by
Mardi C. Hastings, 1 Ph. D.
&
Arthur N. Popper, 1 Ph. D.
Subconsultants to Jones & Stokes Under
California Department of Transportation Contract No. 43A0139, Task Order 1
Funding Provided by the California Department of Transportation
Prime Contractor:
Jones & Stokes
2600 V Street
Sacramento, CA 95818
January 28, 2005
August 23, 2005 ( Revised Appendix B)
1 Any opinions or positions expressed in this report are those of the authors' and do not necessarily represent the opinions or positions of their employers, the State of California, the State of Maryland, or the United States Government
1
Table of Contents
Summary____________________________________________________________________ 4
A. Effects of Pile- Driving Sound on Fish____________________________________________ 4
B. Areas of Uncertainty and Studies Needed________________________________________ 5
Table 1: Outline of studies to investigate pile driving and its effects on fishes._______________________ 6
C. Terminology________________________________________________________________ 7
I. Introduction_______________________________________________________________ 8
II. Characterization of Pile Driving Sound and Its Effect on Fishes___________________ 10
A. Overview of Pile Driving Sound_______________________________________________ 10
B. Comparison of Pile Driving Sound Waveforms with an Ideal Impulse Wave__________ 12
C. Overview of Results from Recent Pile Driving Studies_____________________________ 13
1. Caltrans ( 2001)______________________________________________________________________ 13
2. Abbott and Bing- Sawyer ( 2002)_________________________________________________________ 14
3. Nedwell et al. ( 2003)__________________________________________________________________ 15
4. Caltrans ( 2004)______________________________________________________________________ 15
5. Port of Oakland Prelininary Study ( Abbott 2004; Marty 2004)_________________________________ 16
III. Biology of Fishes_________________________________________________________ 16
A. Fishes of the Pacific Coast and River Systems____________________________________ 16
B. Fish Hearing and Its Importance______________________________________________ 17
1. Sound Production and Communication___________________________________________________ 19
2. Hearing Capabilities of Fishes__________________________________________________________ 20
3. Auditory Structures___________________________________________________________________ 23
4. Hearing Mechanisms_________________________________________________________________ 24
IV. Effects on Fish from Exposure to Sound______________________________________ 25
A. Behavioral Responses and Masking of Biologically Relevant Sounds_________________ 26
Table 3: Citations of selected studies examining the effects of exposure to sound on fishes that have most relevance to pile driving. Note that there are no data for non- teleosts ( e. g., sturgeon, elasmobranchs).___ 26
B. Stress – Physiological Responses_______________________________________________ 28
C. Temporary and Permanent Hearing Loss_______________________________________ 29
D. Structural and Cellular Damage of Auditory and Non- Auditory Tissues_____________ 33
1. Juvenile and Adult Fish_______________________________________________________________ 37
2. Eggs and Larvae_____________________________________________________________________ 38
3. Lateral Line________________________________________________________________________ 40
E. Effects of Pile Driving on Fishes_______________________________________________ 40
V. Areas of Uncertainty and Studies Needed______________________________________ 42
A. Guidance for Protection of Fish from Exposure to Pile Driving Sound_______________ 42
Table 4: Recommendations for Preliminary Guidance for Teleost Fishes_________________________ 46
B. Recommended Studies_______________________________________________________ 47
VI. Literature Cited__________________________________________________________ 53
Glossary___________________________________________________________________ 61 2
Figures____________________________________________________________________ 66
Appendix A_________________________________________________________________ 74
Appendix B_________________________________________________________________ 77 3
Summary
A. Effects of Pile- Driving Sound on Fish
The purpose of this report is to describe what is known about the effects of sound ( including those from pile driving activities) on fishes and to identify studies needed to address areas of uncertainty relative to measurement of sound and the response of fishes. Exposure to sound is defined to include both the received level and duration of the signal.
The emphasis of this report is on the known effects of sound received by fishes. The effects are known if both the received sound and its elicited effect are well defined. Detailed source characteristics of various types of piles and detailed analyses of the effectiveness of various sound attenuation technologies ( e. g., bubble curtains) are beyond the scope of this report.
The results in the peer- reviewed and gray literature on the effects of sound on fishes are variable and, as yet, give no clear- cut “ rules” as to what sounds will affect fish and how they will be affected. A limited number of quantitative and qualitative studies and observations show mortality related to pile driving and also provide some data pertaining to the effects of sound on fishes. Results based on sound signals other than pile driving indicate that some exposures to sound will cause a change in the hearing capabilities of some test fish species or actually damage the sensory structures of the inner ear. There is also a very limited body of evidence that leads to the suggestion that exposure to sound has the potential for affecting other aspects of the physiology of fish, and that these effects may range from the macro ( destruction of the swim bladder) to the cellular and molecular.
Data from explosive blast studies, while not directly comparable to pile driving, indicate that very fast, high- level acoustic exposures can cause physical damage and/ or mortally wound fishes. There is also reason to believe that lesser effects might also occur, but these have not been well documented. Just as in investigations testing the effects of sound, however, the number of species studied in tests of the effects of explosives is very limited, and there have been no investigations to determine whether blasts that do not kill fish have had any impact on short- or long- term hearing loss, or on other aspects of physiology ( e. g., cell membrane permeability, metabolic rate, stress), and/ or behavior ( e. g., feeding or reproductive behavior, movement from preferred home sites).
While these earlier studies provide a preliminary indication of the potential impact of pile driving on fishes, there are no peer- reviewed studies that examine the effects of pile driving on fish hearing, and there are only a few non- peer- reviewed reports about effects on non- sensory structures. 2 While we are able to use available data as a very preliminary indication of the kinds of effects that might be encountered as a result of pile driving, only additional well- controlled studies of behavioral and physiological responses to pile driving or to signals specifically
2 In the course of conducting this review, the authors reviewed numerous reports and publications provided by the Fisheries Hydroacoustics Working Group and other colleagues. We are grateful for having many important reports and papers brought to our attention and many of these have been incorporated into this report. At the same time, we have elected to not include a number of reports and papers provided to us because we did not see them as being relevant to the subject and/ or scope of this report.
4
designed to have the same acoustic characteristics as pile driving sounds, will provide clear scientific support of any criteria that are to be established.
B. Areas of Uncertainty and Studies Needed
To date, there are few data for fish on the effects of exposure to sound from pile driving, and these only appear in the gray literature ( e. g., Anderson 1990; Feist 1992; Bonar 1995; Shin 1995; Caltrans 2001, 2004; Abbott and Bing- Sawyer 2002; Nedwell et al. 2003; Abbott 2004). Although these studies provide some information about exposures to pile- driving sounds, there is little that can be definitively concluded from them. By way of example, there are data and general observations of mortality and some injury to fishes that are close to the source where the level of sound is very high. Additionally, there are observations based on the numbers of fish that come to the surface dead after pile driving that suggest that there is less ( or no) mortality at greater distances from the source ( where the received level of sound would be lower than close to the pile). Finally, experimental cage studies also suggest that fishes further from the pile have little or no mortality and/ or damage ( e. g., Caltrans 2001, 2004; Abbott and Bing- Sawyer 2002; Nedwell et al. 2003; Abbott 2004; Marty 2004).
It does appear, however, that the degree of damage is not related directly to the distance of the fish from the pile, but to the received level and duration of the sound exposure. Because monitoring data show that sound pressure levels do not necessarily decrease monotonically with increasing distance from the pile, it is imperative that received sound levels be measured in future studies in order to develop exposure metrics that correlate with mortality and different types of damage observed in fish exposed to pile driving. The only study we are aware of to date ( Caltrans 2004) that was intended to measure the differential in survival between fish exposed to pile driving with a bubble curtain attenuation device turned on and those exposed with the bubbles turned off, was not able to show a statistical difference in survival between the two conditions because the sample sizes were too small. Though in a study using an explosive sound souce, Keevin et al. ( 1997) showed that use of a bubble curtain significantly reduced mortality of caged bluegill ( Lepomis macrochirus) during demolition of a dam and locks on the Mississippi River.
It is also very difficult to extrapolate to pile driving from studies using other types of signals ( e. g., pure tones, air guns) because such signals are not analyzed or described in a format that can be interpreted in terms of a pile- driving signal ( e. g., acoustic energy flux or acoustic intensity over time). Moreover, signals used in other studies often differ markedly from those emitted by pile driving in terms of duration, rise and fall times, and frequency content ( e. g., Yelverton et al. 1975; Hastings et al. 1996; McCauley et al. 2003). Thus, specific signal components that affect fish may be very different in, for example, a study that uses continuous white noise vs. a study that uses impact sound exposures such as generated during pile driving.
The authors of this report conclude that it is imperative to initiate studies that start with very basic questions about the effects on fishes from exposure to pile driving sound. Table 1 ( page 6) gives an overview of the types of studies that need to be accomplished to better understand the issues of pile driving and the biological effects caused by such signals. Note that
5
this table is presented in much greater detail in Section V of this report ( Table 5, page 49), and summarized in Figure 9 ( page 73).
It is important to note, as discussed in detail in Section V ( page 42), that the body of scientific and commercial data currently available is inadequate for the purpose of developing more than the most preliminary scientifically supportable criteria that will protect fish from exposure to pile driving sound. As a consequence, such criteria are not proposed in this report. Instead, the information from earlier blast and pure tone studies has been used to develop recommendations for interim guidance to address physical injury and mortality and damage to auditory sensory cells, while recognizing the need for well- controlled studies to provide clear direction for development of scientifically supported criteria. It is critical to note, however, that the interim guidance developed must be used with the utmost caution, and that such guidance should not be used for any other signal than pile driving. The interim guidance recommended for pile driving is only applicable to that source and not for other sources such as air guns or sonars because it is based on results of effects studies that had received signals with temporal and spectral characteristics similar to those of pile driving signals.
Table 1: Outline of studies to investigate pile driving and its effects on fishes.
Characteristics of pile driving
Define acoustic dose for exposure to pile driving sound – Develop ways to express exposure to pile driving sounds in terms of total energy received over time and degree of temporal variation, and to define the acoustic particle velocity within the sound field.
Structural acoustic analysis of piles – Develop structural acoustics models of piles to investigate how their physical characteristics affect the radiated sound and how modifications to piles could alter the sounds received by fish. Acoustic analysis could also indicate how best to describe the waveform and how it is affected by pile material, geometry and size, hammer type, and environmental factors such as water temperature, depth, and substrate. These studies could lead to a better ability to develop attenuation of sounds produced during pile driving by modifying structural material, attenuation technologies, etc., especially if they are linked to modeling of the underwater sound propagation as described below.
Characteristics of underwater sound field – Develop underwater sound propagation models for locations of interest and integrate with pile structural acoustics models to estimate received levels of sound pressure and particle velocity in the vicinity of pile driving operations and define zones of impact on fishes. Verify with field measurements of underwater sound pressure measurements.
Effects on fishes
Hearing capabilities of Pacific Coast fishes – Determine hearing capabilities ( using Auditory Brainstem Response [ see Glossary, page 61]) of representative species. 3
3 All studies involve what are called in this report “ representative species.” These are defined as species that serve as models for fishes in the region of question – in this case, the Pacific Coast. Species for study need to be selected to represent differences in: ( a) habitat; ( b) presumed hearing capabilities ( e. g., hearing specialists vs. non- specialists; ( c) ear structure and connections of the ear to peripheral structures such as an air bubble; ( d) swim bladder presence/ absence, biomechanical properties, and connection to the gut; ( e) bony fish vs. non- bony fish ( including elasmobranchs); ( f) fish size/ mass; and ( g) other comparable factors. A minimum set of fishes should be defined so as to have the fewest possible studies and yet represent as many of the parameters for the fishes of the area of question as possible.
6
Table 1: Outline of studies to investigate pile driving and its effects on fishes.
Mortality of fishes exposed to pile driving – Determine mortality immediately and at longer intervals after exposure of representative species as a result of exposure to pile driving sounds. Measure pathology ( using accepted necropsy studies) of the effects of sounds on fishes at different levels and durations of exposure.
Effects of exposure to pile driving sound on non- auditory tissues – Using precisely the same paradigm as used to study the effects on the ear, examine other tissues using standard fish necropsy and histopathological techniques to assess gross, cellular, and molecular damage to fish. Furthermore, determine stress effects on fish using appropriate stress measures ( e. g., hormone levels).
Effects of pile driving sound exposure on hearing capabilities – Determine permanent hearing loss ( PTS) and temporary hearing loss ( TTS) on representative species.
Effects of pile driving sound exposure on fish eggs and larvae – Determine mortality, growth rates, and pathological changes in developing fishes of representative species with exposure at different times during the development cycle
Behavioral responses of fish to pile driving – Observe, in large cages, the short- term behavioral responses of representative species to pile driving sounds. ( e. g., Do fish attempt to swim from the source? Do they react to the sounds? Do they “ freeze” in place?)
More general behavioral responses of fish to pile driving – Determine if there are longer- term effects that might alter movement patterns of fish schools, preferences for breeding sites, feeding behavior, mating and reproductive behavior, etc.
Effects of exposure to pile driving sound on the ear and lateral line – Determine morphological changes over time for representative species on sensory cells of the ear and lateral line, and whether such changes are reversible
Effects of multiple exposures to sound from pile driving on fish – For the appropriate experiments cited above, determine effects of multiple exposures, over time, of pile driving
At the same time, it should be noted that the interim guidance provided in this report ( Table 4, page 46) is conservative and protective of fish. While we do not propose a single set of numbers, the guidelines are values that are well below signal levels that are known to have damaged fishes in other experimental situations. Doing this is particularly important because most of the data providing the basis for interim guidance involved experiments that did not use pile- driving signals, and so extrapolation from such signals to pile driving is difficult, at best.
C. Terminology
This report contains a wide range of acoustic and biological terms. To facilitate understanding of terminology, most of the terms are defined in a Glossary that appears at the end of the report ( page 61).
In addition, it is important to define what is meant by “ behavior” in this report because the word is used for a wide range of activities, and usage varies between different investigators. For example, behavior may be used to refer to the complex interaction of signals and rituals that animals use during mating, or the movements of animals from one feeding ground to another. In the context of this report, “ behavior” is used in its broadest possible sense unless otherwise qualified. In terms of pile driving, the behavioral effects may include small startle movements when the sound is heard or, at the other extreme, behavior may include changes in the reproductive rituals of fish caused by the presence of long- term pile driving at some distance which alters hormone levels.
7
I. Introduction
Over the past decade it has become increasingly apparent that human- generated ( often called“ anthropogenic”) sound has the potential to impact the health and well- being of animals as well as humans. There has been, in this same time frame, an increasing awareness of the presence of human- generated sounds in the aquatic environment, and concern has arisen that these sounds could impact aquatic mammals, diving birds, fishes, amphibians, reptiles, and perhaps even invertebrates ( e. g., NRC 1994, 2000, 2003; Richardson et al. 1995; Popper 2003; Popper et al. 2004).
Despite the concerns raised by increased human- generated sound in the aquatic environment, very little is known about the effects of exposure to such sounds on marine mammals, and far less is known about the effects on fishes ( see reviews in NRC 1994, 2000, 2003; Popper 2003; Popper et al. 2004). And, even in cases where data are available for fishes, they are so few that one must be extremely cautious in attempting to extrapolate between species, even for identical stimuli. Moreover, one must also be extremely cautious with any attempts to extrapolate results between stimuli because the characteristics of the sources ( e. g., air guns, sonars, ship noise, pile driving) differ significantly from one another.
The purpose of this report is to describe what is known about the effects of sound on fish and to identify needed studies to address areas of uncertainty relative to measurement of exposure to sound and the response of fishes. Exposure to sound is defined to include a measure of both the received level and the duration of the signal. For example, the received level can be expressed in terms of acoustic pressure, particle velocity, or intensity ( energy flux), which all vary with time over the duration of the signal. Sound exposure metrics usually convey an integration of the received level over the time duration of a single acoustic event. The effects of multiple events primarily depend on the degree of damage caused by a single event, recovery of the damaged region during the time interval between events, and the total number of events. If a single sound exposure causes absolutely no damage whatsoever, then no damage will occur for repeated multiple exposures. 4
The focus of this review is the evaluation of all known literature related to the effects of pile driving on fishes, with particular emphasis on fishes of the Pacific Coast region, including fishes in bay, estuarine, lake, river, and stream habitats. Pile driving commonly occurs in shallow water and is related to construction and repair of bridges, docks, and other over water infrastructure. Very little, if anything, is known about ( 1) particular characteristics of a pile- driving signal that are responsible for different observed effects, and ( 2) the differential in effects that can be attributed to differences in signals generated by different types of piles ( e. g., concrete vs. steel). Indeed, structural acoustics analysis of different piles, including how variations in material and structure could alter sounds received by fish, is a recommended study ( Tables 1 and 5). Moreover, only the experimental studies recommended to determine effects on fishes will reveal the relationships between characteristics of pile driving signals and observed effects. Thus the emphasis of this report is on what is known about effects related to sound received by fish, and not on the sounds emitted by any particular type of pile or pile driving operation.
4 Although it should be noted that some effects may not be easily observed, such as damage at the cellular levels, but as a result of repeated exposures the damage may show up on a larger, and more easily determined, scale.
8
This report describes the potential for effects on fish that is supported or inferred from available information and sets the stage for future studies by outlining what is known about detection of acoustic signals by fishes, sound detection by Pacific Coast fishes, effects of exposure to sound on various species of fishes, and characteristics of the sounds produced by pile driving. Far too little is known about the effects on fish from exposure to sound to allow for definitive conclusions to be drawn from the literature. A series of well- defined research programs, with suitable and appropriate experimental design and controls, would help garner needed information ( see Tables 1 [ page 6] and 5 [ page 49]).
The material presented in this report is based on a careful evaluation of technical reports ( gray literature) and peer- reviewed articles ( see footnote 1, page 4). The approach and analysis in each study differs, and so extrapolation between studies, especially those done in different locations or by different groups of investigators, is difficult. Moreover, we have been particularly careful in our use of the gray literature because we have no way of knowing if these studies have undergone rigorous scientific review.
To help resolve the problems in using the gray literature reports, we have attempted to review the potentially useful gray literature reports ourselves, and have used this material based on our views about the quality of the science and the validity of the conclusions in these studies. We have, for the most part, avoided use of material that is presented only as pages on the Internet because we have no basis for knowing if that material has received any review whatsoever.
In addition to primary peer- reviewed literature and gray literature reports, we also include citations to a number of reviews and overviews of various aspects of the material presented here. It must be recognized that the reviews, even if they have gone through appropriate peer review, are often the opinions of the authors and may be based on analysis of material from peer- reviewed articles and/ or the gray literature.
Based on the literature review, this report provides the rationale to establish interim guidance for impact thresholds for the purpose of protecting listed and commercially important species, and identifies future research needs. Once research provides sufficient data to address identified critical information gaps such as effects on physiology and behavior, results can be used to establish formal criteria to protect fish.
The report is divided into several sections. Section I ( page 8) is an overall introduction. Section II ( page 10) describes the signals produced by pile driving and includes an overview of a number of the experimental studies that have explicitly looked at the effect of pile driving sounds on fishes. Section III ( page 16) provides a background on sound and fish, and includes an overview of sound detection and the use of sound by fishes. This is followed in Section IV ( page 25) with a discussion of what is known about the effects of exposure to sound on fishes. As appropriate, Section IV provides some analysis of the little we know about the effects of pile driving, and provides considerably more background on the effects of exposure to sounds generated by other means, including underwater sound projectors used in laboratory experiments and others more representative of human- generated sources ( e. g., air guns, explosives, sonar). The discussion includes effects that range from fish mortality to the movement of fishes from areas of ensonification. Finally, Section V ( page 42) describes gaps in our knowledge on the
9
effects of pile driving on fishes. It provides interim guidelines that might be used to help protect fishes before firmer criteria can be established based on new data that need to be obtained. Section V ends with recommendations for studies needed to develop realistic and usable criteria.
II. Characterization of Pile Driving Sound and Its Effect on Fishes
A. Overview of Pile Driving Sound
Impact sounds result from a rapid release of energy when two objects hit one another. The physical characteristics of impact sounds primarily depend on the mechanical properties of the impacting objects. When a pile- driving hammer strikes a pile, impact sound propagates in the air and a transient ( stress) wave, or pulse, propagates down the length of the pile. The impact will also create flexural ( or transverse) stress waves in the wall of the pile which couple with the surrounding fluids ( air and water) to radiate sound into the water and additional sound into the air.
Moreover, the pulse propagating down the length of the pile couples to the substrate at the water bottom, and this causes waves to propagate outward through the bottom sediment. These transient sound waves ( or pulses) in the substrate can be transmitted from the bottom into the water at some distance away from the pile to create localized areas of very low and/ or very high sound pressure and acoustic particle motion because of destructive or constructive ( respectfully) interference with the sound pulse traveling outward through the water directly from the pile. Thus it is possible that at certain locations received levels of sound could be higher further from the pile than at locations closer to it and this has been observed in some monitoring data ( Caltrans 2001).
Sound pulses as a function of time are referred to as waveforms. The sound pressure associated with passage of a waveform at some point away from the pile can be measured at a selected location in the water column using a hydrophone ( an underwater microphone) or sound level meter with an underwater probe. Typically, pile driving sounds underwater are characterized by multiple rapid increases and decreases in sound pressure over time as shown in the measured waveform displayed in Figure 1( a) ( page 66). The peak pressure is the highest absolute value of the measured waveform, and can be a negative or positive pressure peak.
The root- mean- square or “ rms” level is determined by analyzing the waveform and computing the square root of the average of the squared pressures over the time period that comprises that portion of the waveform containing 90 percent of the sound ( pressure squared) energy ( Richardson et al. 1995; Illingworth & Rodkin 2001). This calculated rms sound pressure level ( SPL) is described as “ RMS ( impulse)” and is used to report an overall average SPL for a single pile driving pulse ( Caltrans 2001; Illingworth & Rodkin 2001). The frequency content of the sound pressure level shown in Figure 1( b) provides some indication of the bandwidth of the pile driving pulse. The frequency bandwidth for pile driving sounds is typically below 1,000 Hz, the same bandwidth as hearing in many species of fish ( see Fig. 2, page 67).
Another measure of the pressure waveform that can be used to describe the pile driving pulse is the sound energy. Typically, the effects of short- duration or transient sounds are not
10
only characterized by rise time, duration, and peak pressure, but also total energy received over time ( or dose). While the effects are described most often in terms of humans, all indications are that the same effects are likely to occur in all vertebrates. The energy contained in a sound wave is a measure of the amount of work it does pushing on the fluid ( or substrate material) as it travels. The sound wave “ pushes” with pressure, or force acting over a unit area, and this force causes the fluid to move locally. This fluid motion is called “ acoustic particle velocity.” If the sound impinges on an aquatic animal, the energy will create forces and motions inside its body just as it does in the fluid.
For a plane wave traveling in open space without any interaction with objects or boundaries, the relationship between sound pressure ( p) and particle velocity ( v) is p = ( ρc) v, where ρ ( kg/ m3) is the density of the fluid and c ( m/ s) is the speed of sound in the fluid ( or other material). The acoustic energy flux or intensity ( I) of a sound wave is the product of the pressure multiplied by the particle velocity, and has the units of Joule per square meter per second ( J/ m2- s) or watts per square meter ( W/ m2). For a plane wave the intensity ( or energy flux) is given by I = p2/( ρc). It is equivalent to the amount of energy in Joules passing through a unit area per unit time as the sound wave travels unbounded in the fluid.
How rapidly the energy accumulates over time may be significant in assessing the potential effects of exposure to transient sounds on fish and other aquatic animals ( e. g., Johnson and Robinson 1969; Hamernik and Hsueh 1991). Indeed, Finneran et al. ( 2002) found that cumulative sound energy could be used as a criterion for temporary threshold shift in hearing of a few species of marine mammals ( odontocetes) for a single exposure to several different types of transient signals. In addition, Rasmussen ( 1967) reported that fish could withstand higher peak pressure from a gunpowder explosion than from a dynamite explosion because of the “ more rounded front” of the pressure waveform. A more rounded front would result in a slower rate of sound energy accumulation than would occur for a waveform typical of a dynamite explosion.
Because sound is a form of energy, the damage potential of a given sound environment will depend not only on its level, but also its duration. For constant sound levels this is a straightforward analysis, but if sound level varies it must be sampled repeatedly over a well- defined time window ( or sampling period). In human studies, these samples have been averaged together to form a single value known as the Equivalent Continuous Sound Level or Leq, which has the same energy content as a varying sound level.
A common alternative energy metric to the Leq is the sound exposure level ( SEL), which is defined as the constant sound level acting for one second, which has the same amount of acoustic energy as the original sound. An SEL measurement is often used as an energy metric for a single acoustic event. Because all SEL measurements are normalized to a one second time interval, it may be used to compare the energy content of different exposures to sound. SEL is calculated by summing the cumulative pressure squared ( p2) over time and is often used as an indication of the energy dose. The unit for SEL is dB re 1μPa2- s. This metric accounts for both negative and positive pressure excursions because p2 is positive for both and therefore negative and positive pressures are treated equivalently in the cumulative sum of p2.
The cumulative sound pressure squared ( also commonly referred to as accumulated sound energy) plotted in Figure 1( c) ( page 66) provides a comparison of the differences in
11
estimated energy between transient waveforms because it contains aspects related to the effects of both peak pressure and rise time. If a sound pulse contains higher pressure peak amplitudes and/ or faster rise and fall times, then the cumulative pressure squared will increase at a higher rate than for a pulse with lower peak amplitudes and longer rise and fall times. As previously stated, this could be significant in assessing the potential effects of exposure to transient sounds, on fish.
The caveat in using SEL is that its calculation inherently assumes a plane wave in which the acoustic energy flux ( or intensity as defined above) is directly proportional to p2. Thus in many underwater environments where the relationship between acoustic pressure and particle velocity is more complex, the true “ total energy flux,” will not be equivalent to SEL ( Weston 1960; Hamernik and Hsueh 1991).
In the case of pile driving, there is rarely a plane wave because the sounds are produced in shallow water near shore with numerous boundaries and may interact with sound traveling in the substrate. These conditions produce a very complex sound field that does not have a simple relationship between sound pressure and particle velocity. Moreover, it is necessary to know the sound particle velocity because it is the stimulus that is detected by the ears of fishes ( e. g., Popper et al. 2003). Because of the complexity of the sound field produced in pile driving environments, relatively simple models, such as the one developed by Dzwilewski and Fenton ( 2003), are not very useful in predicting the impact zones for aquatic animals. More detailed models of the acoustic environment are needed to understand where sound energy is concentrated and adequately predict impact zones for aquatic animals ( Ward et al. 1998).
B. Comparison of Pile Driving Sound Waveforms with an Ideal Impulse Wave
Impulse sound is a transient sound that also arises from a rapid release of energy, usually electrical or chemical, such as from circuit breakers or explosives. Although impact and impulse sounds result from different processes, they share many characteristics: high peak overpressure, rapid rise and fall times, and relatively short durations. Thus “ impulsive” and “ impact” are often used interchangeably to describe many high level, short duration sounds.
The ideal impulse is mathematically described by the Friedlander wave ( Hamernik and Hsueh 1991). If transient sounds, such as those produced by pile driving, could be characterized using a waveform similar to this type, then effects of pile driving on aquatic animals could potentially be extrapolated from data based on effects observed from exposure to other transient signals ( e. g., explosives, air guns, sonic booms) or other transient waveforms that could be described by the Friedlander wave model. These estimates could provide a basis for developing interim guidance for exposure to sound from pile driving until more research is completed.
Figure 2 ( page 67) shows an approximation of a pile driving sound using a Friedlander wave. Figures 2( a), ( b), and ( c) compare the temporal characteristics, sound exposure spectral density, and cumulative pressure squared over time, respectively, for the idealized and actual pile driving sound characterized in Figure 1 ( page 66). These waves are very close in exposure characteristics, which indicate that the key characteristics for pile driving may be the peak
12
positive and negative pressures and their time durations, which are combined to calculate the cumulative pressure squared and sound exposure level ( SEL).
SEL is based on the cumulative sum of the square of the pressure, so positive and negative pressures have equivalent contributions because the pressure squared is always positive. Even though a Friedlander wave does not have a large negative pressure, its SEL will be approximately the same as that of a pile- driving signal if both waveforms have the same peak pressure magnitude, and if the time interval for the rise and fall of positive pressure in the Friedlander waveform is the same as the sum of the time intervals for the rise and fall of the largest positive and negative pressures in the pile- driving waveform. Thus a systematic approach to approximate pile- driving signals using mathematically modeled Friedlander type waves could provide a way to determine how data, which have been obtained in effects studies using blasts or other transient sources, relate to different pile driving scenarios. Appendix A ( page 74) provides a detailed explanation of the derivation of Figure 2 that could be applied to other sets of data.
A mathematical model that captures the essential characteristics of pile driving signals could also be used to investigate the effects of changes in the waveform that could be created by modifications in the structural acoustics design of the pile. Such an approach was used to investigate the reshaping of sonic booms to achieve both reduced loudness and sound exposure level ( Leatherwood and Sullivan 1994).
C. Overview of Results from Recent Pile Driving Studies
There are five recent experimental studies that have examined the effects of pile driving on fish ( Caltrans 2001, 2004; Abbott and Bing- Sawyer 2002; Nedwell et al. 2003; Abbott 2004). In this section we provide a very brief overview of these studies as background to the discussions that follow. An analysis of the results is provided in Section IV. E. ( page 40) and Appendix B.
1. Caltrans ( 2001)
Caltrans ( 2001) did an assessment of impacts to fish in San Francisco Bay during a Pile Installation Demonstration Project ( PIDP) leading up to the work on the San Francisco- Oakland Bay Bridge ( SFOBB) East Span. The Caltrans study involved examining fish that died during exposure to pile driving operations, as well as determining the effects on shiner surfperch ( Cymatogaster aggregata) held in cages at different distances from the pile driving source, although these results were listed as being very preliminary.
Results indicate that there was mortality caused by exposure to pile driving sounds, with dead fish of several different species found to at least 50 meters from the pile being driven. There was also an increase in catch by over flying gulls during pile driving, further indicating fish mortality. The report suggests that the use of a bubble curtain cut down on mortality, but there was no statistical demonstration of this result. Dead/ dying fish showed a number of forms of damage including bleeding and damage to the swim bladder. Numbers were relatively low, reflecting difficulty in retrieving dead/ dying fish and the possibility that fish did not come to the surface at all, or not until they were away from the collecting operation.
13
The extent of damage and mortality of caged fish was greater when the cages were closer to the source than further away. There was significant variability between experiments with the hammer used, the duration of exposure, and distance of the fish cages from the source. Insufficient numbers of exposures or animals make it difficult to reach firm conclusions, but there may be a distance effect that is possibly related to decreasing levels of the signal, though the signal levels at the test cages were not measured. Moreover, because of complexities in the underwater sound field, especially when there is also a potential contribution from energy that has traveled through the substrate interacting with the water- borne signal, it is not clear that the signal at some distance from a pile- driving source will be lower than a signal closer to the source ( see Section II. A., page 10).
2. Abbott and Bing- Sawyer ( 2002)
Abbott and Bing- Sawyer ( 2002) investigated the effects of pile driving construction on Sacramento blackfish ( Orthodon microlepidotus) as a surrogate for the threatened Sacramento splittail ( Pogonichthys mactolepidotus). The authors reported significant problems with the study in that there were many uncontrolled factors as well as inadequate sample size to allow any definitive conclusions from the work. Significantly, sound levels at the test cages were not calibrated, and there was potential acoustic interference in the water that may have affected sound levels so that they could not be predicted based on any predictable attenuation from a site of known source levels.
Fish were placed in cages near a pile that was struck 43 times with the presence of a bubble curtain, and then 45 additional times after the bubble curtain was removed. The same fish were used both with and without the bubble curtain. Sound levels at the location of the fish ( received sound levels) were not made. Cages were placed at five different distances from the pile, from 45 m to 850 m. At the end of the exposure, fish were removed from the cages, observed for injuries and abnormal behavior for five hours, and then placed in plastic bags and frozen on ice. Subsequently, non- microscopic necropsy was done on each fish to determine any external or internal damage resulting from exposure to the pile driving.
Results of the necropsy suggested that there was more damage to fish closer to the source than further away. At the same time, there was substantial inter- animal variation, even within the same cage, in damage revealed by gross necropsy. While the authors concluded that the damage was only found in fishes subject to 193 dB ( peak) re 1 μPa sounds and that there was no damage to fishes exposed to sounds below 183 dB ( peak) re 1 μPa, it must be emphasized that sound levels were extrapolated and were not actually measured at the cages. Accordingly, these conclusions are not well supported and the lack of actual measured sound levels and problems with controls and necropsy methods confound interpretation of the injury results.
No behavioral effects were seen prior to sacrifice of the fish, although the authors indicate that they did not have suitable facilities in which to observe behavior. There were no deaths of any animals at the time of removal from the cages or in the five- hour period prior to sacrifice for necropsy.
14
3. Nedwell et al. ( 2003)
Nedwell et al. ( 2003) examined the effects on caged brown trout ( Salmo trutta) in response to pile driving at the Red Funnell’s Southampton Terminal, England. Caged fish were placed at distances of 25 to 400 m from the piles being driven, with a control cage 10 km away. Animals were observed by closed circuit TV as they were exposed to pile- driving sounds. The sound levels at the various cages were not given, though the authors estimated that the level of the pile- driving signal ( presumably, but not stated, as an impact operation) was 134 dB re 1 μPa ( peak) at 400 m from the pile.
Behavioral results reportedly showed no reaction to vibropiling for fish as close as 25 m to the source. The behavior of animals in response to hammer pile operations were only reported for fish at 400 m from the source. These fish showed no apparent response to the pile driving. Fish injuries were only monitored for fish at 400 m. No criteria for injuries were given other than gross descriptions ( e. g., hemorrhage of the eyes, rupture of the swim bladder), nor was there any analysis of any other possible injuries. The authors concluded that no injuries were observed. Data were not, however, provided for fish closer to the source than 400 m to the pile driving operation.
4. Caltrans ( 2004)
Caged fish monitoring was conducted during the San Francisco – Oakland Bay Bridge East Span seismic safety project using caged shiner surfperch ( Cymatogaster aggregata) and steelhead ( Oncorhynchus mykiss). Caged fish were exposed to pile- driving sounds at distances from 23 to 314 meters from the pile driving operation, with exposure durations of 1 to 20 minutes. Controls included fish that were placed in test cages and in the same locations as the test animals for 3 to 10 minutes but without exposure to pile driving ( this was a shorter time period of being underwater than for fish exposed to pile driving) and fish that were never placed in cages and either transported to the test site ( transport controls) or kept in the laboratory.
Fish were observed behaviorally following exposure and then held for 48 hours to monitor survival. Fish were then sacrificed by placing them in a plastic bag and freezing. Fish were thawed for necropsy, which was done rapidly ( 3- 5 minutes). A number of fish died before the end of the 48- hour holding period. The authors report that all control animals had the same low level of trauma. They also reported that there was more trauma in animals exposed to pile driving, and lesser levels of trauma in animals exposed to pile driving in the presence of an air bubble curtain. However, no statistical difference between any effects seen in fishes exposed to pile driving with and without the presence of the air bubble curtain could be provided because the number of fish exposed while air bubble curtains were used did not provide a sufficient sample size. The authors also reported no mortalities that could be associated with exposure to pile driving.
15
5. Port of Oakland Prelininary Study ( Abbott 2004; Marty 2004)
This study ( Abbott 2004; Marty 2004) investigated the effects of pile driving on caged fish of three species: shiner perch, Chinook salmon, and northern anchovy. The fish were caged and lowered to about 25 feet at a location that was about 32 feet from the pile being driven. The fish were exposed to four minutes of pile driving ( something over 200 impulses). The 2- foot diameter jetted concrete piles were driven using using a diesel hammer at the construction site of Berth 22 in the Port of Oakland. Following exposure, fish were returned to the surface, mortality determined, behavior observed for one- minute, and then later sacrificed using excellent pathology methodology ( Marty 2004). Controls were treated in the same way as experimental animals and pathology was analyzed without the investigator knowing the treatment group of each animal examined.
The results showed no differences in mortality between sound- exposed and control animals ( Abbott 2004). The investigators suggested that there were also no behavioral differences between sound- exposed and control animals, but this was based on behavioral analysis after the fish had been removed from the test cages and it is impossible to ascertain whether there were behavioral effects during sound exposure. ( Moreover, no data were provided on how the behavioral analysis was performed.) The pathological analysis ( Marty 2004) showed no difference between sound- exposed and control animals.
III. Biology of Fishes
A. Fishes of the Pacific Coast and River Systems
The fishes of the Pacific Coast region that are potentially impacted by pile driving in estuaries, bays, lakes, streams and rivers are listed in Table 2 ( page 17). There is a wide diversity of species that include both cartilaginous fishes ( sharks and rays – class Chondrichthyes), and bony fishes ( class Osteichthyes). The bony fishes include the more advanced teleosts ( ray- finned fishes such as salmon, tuna, perch, and most commercially important species), as well as representatives of more primitive chondrostean fishes, including sturgeons. The vast majority of fish species on the Pacific Coast ( as throughout the world’s oceans and fresh water systems) are teleosts. ( Indeed, teleost fishes make up approximately 23,000 of about 27,000 extant fish species [ Helfman et al. 1997]. It is worth noting that the number of living species of fish far exceeds the number of living species of all other vertebrate groups combined.)
Among these fishes, several are listed as threatened or endangered under the federal Endangered Species Act. These include several species of the genus Oncorhynchus ( chinook salmon, coho salmon, chum salmon, and steelhead), bull trout ( Salvelinus confluentus) delta smelt ( Hypomesus transpacificus), and the tidewater goby ( Eucyclogobius newberry). The salmonids and the smelt are all in the taxonomic order Salmoniformes, while the goby is unrelated to salmonids.
16
B. Fish Hearing and Its Importance
There is a long historic record of human awareness that fishes produce and use sounds in a wide variety of behaviors ( reviewed in Moulton 1963). Studies of fish hearing and sound production ( bioacoustics), and the importance of sounds to the lives of fishes, were not initiated, however, until the early part of the 20th century ( see Moulton 1963 and Tavolga 1971 for historic reviews). The level of investigation of fish hearing and sound production ( e. g., bioacoustics) increased considerably in the second half of the 20th century ( see Popper and Fay 1999; Zelick et al. 1999; Popper et al. 2003; Ladich and Popper 2004).
Table 2: Target Fish Species for Acoustic Exposure Criteria in Pacific
Estuaries, Bays, and Rivers ( Data provided by Warren Shaul of Jones and Stokes Associates)
Estuarine Life Stages
Riverine/ freshwater Life Stages
Species
( A- adult, E- egg, L- larvae, J- Juvenile)
Priority 1: ESA Listed Species
Chinook Salmon Oncorhynchus tshawytscha ( Salmonidae)*
A, J
A, E, L, J
Coho Salmon Oncorhynchus kisutch ( Salmonidae)*
A, J
A, E, L, J
Chum Salmon Oncorhynchus keta ( Salmonidae)*
A, J
A, E, L, J
Steelhead Oncorhynchus mykiss ( Salmonidae)*
A, J
A, E, L, J
Cutthroat Trout Oncorhynchus clarkii ( Salmonidae)*
A, J
A, E, L, J
Bull Trout Salvelinus confluentus ( Salmonidae)*
A, J
A, E, L, J
Delta Smelt Hypomesus transpacificus ( Osmeridae)*
A, J
A, E, L, J
Tidewater Goby Eucyclogobius newberry ( Gobiidae)**
A, E, L, J
A
Priority 2: EFH Species ( species listed under ESA are not listed again in EFH)
Puget Sound Pink Salmon Oncorhynchus gorbuscha ( Salmonidae)*
A, J
A, E, L, J
Leopard Shark Triakis semifasciata ( Triakidae)
A, J
Soupfin Shark ( Tope) Galeorhinus galeus ( Triakidae)
A, J
Spiny Dogfish Squalus acanthias ( Squalidae)
A, J
California Skate Raja inornata ( Rajidae)
A, J, E
Big Skate Raja binoculata ( Rajidae)
A, J, E
Longnose Skate Raja rhina ( Rajidae)
A
Ratfish Hydroglagus coliei ( Chimaeridae)
A, J, E
Lingcod Ophiodon elongates ( Hexigrammidae)**
A, J, E, L
Cabezon Scorpaenichthys marmoratus ( Cottidae)**
A, J, E, L
Kelp Greenling Hexagrammos decagrammus ( Hexigrammidae)**
A, J, E, L
Pacific Cod Gadus macrocephalus ( Gadidae)**
A, J, E, L
Pacific Whiting ( Hake) Merluccius productus ( Merlucciidae)**
A, J, E, L
Sablefish Anoplopoma fimbria ( Anoplopomatidae)**
J
Black Rockfish Sebastes melonops ( Scorpaenidae)**
A, J
Bocaccio Sebastes paucispinus ( Scorpaenidae)**
J, L
Brown Rockfish Sebastes auriculatus ( Scorpaenidae)**
A, J, E, L
Calico Rockfish Sebastes dalli ( Scorpaenidae)**
A, J
California Rockfish Scorpaena guttata ( Scorpaenidae)**
J, L
Canary Rockfish Sebastes pinniger ( Scorpaenidae)**
A?, J
China Rockfish Sebastes nebulosus ( Scorpaenidae)**
A, J
Copper Rockfish Sebastes caurinus ( Scorpaenidae)**
A, J, E, L
Darkblotched Rockfish Sebastes crameri ( Scorpaenidae)**
A, J
Greenstripe Rockfish Sebastes elongatus ( Scorpaenidae)**
Kelp Rockfish Sebastes atrovirens ( Scorpaenidae)**
J
17
Species EstuSatrai gnees L ife Rievre rLinifee/ fSrteasghews at
( A- adult, E- egg, L- larvae, J- Juvenile)
Quillback Rockfish Sebastes maliger ( Scorpaenidae)**
A, J, E, L
Redbanded Rockfish Sebastes babcocki ( Scorpaenidae)**
A
Redstripe Rockfish Sebastes proriger ( Scorpaenidae)**
Rosethorn Rockfish Sebastes helvomaculatus ( Scorpaenidae)**
A, J
Rosy Rockfish Sebastes rosaceus ( Scorpaenidae)**
Rougheye Rockfish Sebastes aleutianus ( Scorpaenidae)**
A, J?
Sharpchin Rockfish Sebastes zacentrus ( Scorpaenidae)**
A, J?
Shortspine Thornyhead Sebastolobus alascanus
A, J
Splitnose Rockfish Sebastes diploproa ( Scorpaenidae)**
A, J
Stripetail Rockfish Sebastes saxicola ( Scorpaenidae)**
A
Tiger Rockfish Sebastes nigrocinctus ( Scorpaenidae)**
A, J
Vermillion Rockfish Sebastes miniatus ( Scorpaenidae)**
A, J
Yelloweye Rockfish Sebastes ruberrimus ( Scorpaenidae)**
A
Yellowtail Rockfish Sebastes flavidus ( Scorpaenidae)**
A, J
Arrowtooth Flounder Atheresthes stomias ( Pleuronectidae)**
A, J
English Sole Parophrys vetulus ( Pleuronectidae)**
A, J, E, L
Pacific Sanddab Citharichthys sordidus ( Bothidae)**
J, E, L
Butter Sole Pleuronectes isolepis ( Pleuronectidae)**
A, J
Curlfin Sole Pleuronichthys coenosus ( Pleuronectidae)**
A
Dover Sole Microstomus pacificus ( Pleuronectidae)**
A, J
Flathead Sole Hippoglossoides elassodon ( Pleuronectidae)**
A, J
Petrale Sole Eopsetta jordani ( Pleuronectidae)**
A, J
Rex Sole Glyptocephalus zachirus ( Pleuronectidae)**
A
Rock Sole Pleuronectes bilineatus ( Pleuronectidae)**
A, J
Sand Sole Psettichthys melanostictus ( Pleuronectidae)**
A, J
Starry Flounder Platichthys stellatus ( Pleuronectidae)**
A, J, E, L
Northern Anchovy Engraulis mordax ( Engraulidae)**
A, J
Pacific Chub Mackerel Scomber japonicus ( Scombridae)**
A, J
Jack Mackerel Trachurus symmetricus ( Carangidae)**
A, J
Pacific Sardine Sardinops sagax ( Clupeidae)*
A, J
Market Squid Loligo opalescens
A, J
Priority 2: Other Commercial Species
Pacific Herring Clupea pallasii ( Clupeidae)*
A, J, E, L
Priority 3: Sensitive Native Species
White Sturgeon Acipenser transmontanus ( Ascipenseridae)*
A, J
A, J, E, L
Golden Trout Oncorhynchus aguabonita ( Salmonidae)*
A, J, E, L
Green Sturgeon Acipenser medirostris ( Ascipenseridae)*
A, J
A, J, E, L
Longfin Smelt Spirinchus thaleichthys ( Osmeridae)*
A, J
A, E, L
Tule Perch Hysterocarpus traskii ( Embiotocidae)**
A, J
Shiner Perch Cymatogaster aggregate ( Embiotocidae)**
A, J
Pile Perch Rhacochilus vacca ( Embiotocidae)**
A, J
Barred Surfperch Amphistichus argenteus ( Embiotocidae)**
A, J
Splittail Pogonichthys macrolepidotus ( Cyprinidae)*
A, J
A, E, L, J
Sacramento Blackfish Orthodon microlepidotus ( Cyprinidae)*
A, E, L, J
Priority 4: Nonnative Sport- Fishery Species
American Shad Alosa sapidissima ( Clupeidae)*
A, J
A, J, E, L
18
Species EstuSatrai gnees L ife Rievre rLinifee/ fSrteasghews at
( A- adult, E- egg, L- larvae, J- Juvenile)
Channel Catfish Ictalurus punctatus ( Ictaluridae)*
A, J, E, L
Striped Bass Morone saxatilis ( Percichthyidae)**
A, J
A, J, E, L
Bluegill Lepomis macrochirus ( Centrarchidae)**
A, J, E, L
Redear Sunfish Lepomis microlophus ( Centrarchidae)**
A, J, E, L
White Crappie Pomoxis annularis ( Centrarchidae)**
A, J, E, L
Black Crappie Pomoxis nigromaculatus ( Centrarchidae)**
A, J, E, L
Largemouth Bass Micropterus salmoides ( Centrarchidae)**
A, J, E, L
Small mouth Bass Micropterus dolomieui ( Centrarchidae)**
A, J, E, L
* Physostomous ( see Glossary, page 61)
** Physoclistous ( see Glossary)
It was also in the latter part of the 20th century that investigators became more acutely aware of the possibility that human- generated sounds may have an effect on the lives of aquatic organisms ( see reviews in NRC 1994, 2000, 2003; Richardson et al. 1995), and that the species affected not only include marine mammals ( the subjects of greatest interest) but also fishes and other aquatic vertebrates ( e. g., marine turtles, aquatic and diving birds) and possibly invertebrates ( e. g., crabs, lobsters). The concerns about potential effects of exposure to human- generated sounds include impacts on communication with conspecifics ( members of the same species), effects on stress levels and the immune system, temporary or permanent loss of hearing, damage to body tissues, mortality, and mortality or damage to eggs and larvae. Moreover, concerns not only include immediate effects, but also potential long- term effects that might now show up for hours, days, or even weeks after exposure to sounds.
1. Sound Production and Communication
Teleost fishes produce sound in several ways, none of which involves a larynx or syrinx- like structure as used by terrestrial vertebrates. Instead, fishes use a variety of different methods to produce sounds that range from moving two bones together to more complex mechanisms involving exceptionally fast muscles connected to the swim bladder. In this latter instance, the muscles contract at frequencies high enough to produce sound ( see Zelick et al. 1999). The gas- filled swim bladder ( or gas bladder) in the abdominal cavity may serve as a sound amplifier ( although it has other very critical functions as well -- see Steen 1970). Sounds produced in this way usually have most of their energy below 1,000 Hz.
Fish use sounds in a wide variety of behaviors including aggression, protection of territory, defense, and reproduction ( reviewed in Tavolga 1971; Demski et al. 1973; Zelick et al. 1999). There is also evidence that at least one species of marine catfish ( Arius felis) uses a form of " echolocation" to identify objects in its environment by producing low frequency sounds and listening to their reflections from objects ( Tavolga 1976). Data in the literature suggest that it is the temporal pattern of fish sounds, rather than their frequency spectrum, that is most important for acoustic communication by fishes ( Winn 1964; Spanier 1979).
19
2. Hearing Capabilities of Fishes
Fishes are able to detect and respond to a wide range of sounds. The techniques for determining hearing capabilities of fishes are similar to those used in studies of other animals, or even humans. One set of measures involves “ asking” a fish what it hears and then measuring some kind of behavioral response from the animal whenever a sound is detected. Such responses may be conditioned ( trained, such as hitting a paddle when a sound is detected) or unconditioned ( untrained, such as change in heart rate). Alternatively, the response of the fish can be determined by measuring electric potentials in the brain that are generated when the ear detects a sound ( i. e., the Auditory Brainstem Response or ABR), much as is done to measure whether new born human babies are able to detect sounds.
In either case, the first goal of measuring hearing is to determine the range of frequencies ( or bandwidth) that a fish can detect, and then determine the lowest levels of the sound detected at each frequency ( the “ threshold,” or lowest signal that an animal will detect in some statistically determined per cent of signal presentations – most often 50%). The graphic representation of the threshold as a function of frequency is called an “ audiogram.” Figure 3 ( page 68) shows audiograms for fishes similar to those found in the Pacific Coast region, or that have ears with similar structures to a number of those species.
Several aspects of fish hearing are apparent from Figure 3.5 The figure clearly shows that species differ in the range of frequencies, or bandwidth, that they are able to detect, and in the lowest sound pressure level ( threshold) that they are able to detect. The fish with the ability to detect the widest bandwidth is the scaled sardine, a species that is probably representative of the sardines and anchovies on the Pacific Coast. Greatest sensitivity ( lowest threshold of detection) is found in the Atlantic cod, a relative of the Pacific cod.
While not as extensively studied, a variety of behavioral and physiological investigations of fish hearing show that a number of species ( and perhaps all) have the same basic acoustic capabilities as other vertebrates, including mammals ( see Popper et al. 2003 and Ladich and Popper 2004 for review of fish hearing capabilities). For example, fishes are able to discriminate between sounds of different magnitudes or frequencies, detect a sound in the presence of other signals, and determine the direction of a sound source ( sound source localization). Indeed, these higher level capabilities are far more important to a fish than just detection of sound ( as illustrated by the threshold measures) because fishes must discriminate between sounds of predator vs. those of prey, determine the direction of a sound emitted by potential predators or prey, and determine the nature of one sound source in the presence of others. Most importantly, fishes must detect the presence of a signal that is important to them even when there are
5 A critical caveat to the threshold data presented in Figure 3 ( page 68) arises from the acoustics of the experimental tanks in which data were obtained, how the test signals were calibrated, and the hearing mechanisms of the species studied. As discussed in detail in the text, hearing “ generalists” appear to primarily detect the particle motion component of the sound field. Producing a controllable particle motion signal in a small tank ( as used in most, if not all, of the earlier studies on fish hearing) is very difficult, and such signals were rarely calibrated. Thus, while the data in Figure 3 and in earlier works are often presented in terms of sound pressure, the more important question for fishes without hearing specializations is their response to particle motion, and these data are not available. While the general frequency hearing range for the species presented is valid, and the sensitivity a reasonable general indicator of hearing capabilities of generalists, future studies need to involve much better control over the acoustic signal used in determining thresholds and better calibration of signals.
20
extraneous background noises. Clearly, addition of human- generated sounds to the background noise can make the environment so loud that fish are not able to detect important signals ( e. g., that of a predator) because of the strong anthropogenic masking sound.
It has generally been agreed that fish can be divided into two groups – hearing generalists ( or “ non- specialists”) and hearing specialists ( see Popper et al. 2003 and Ladich and Popper 2004 for detailed discussion). 6 These groups are not related to the taxonomic relationship between fishes. Instead, both hearing specialists and generalists are found distributed through many fish taxonomic groups.
Hearing specialists have special adaptations that enhances their hearing bandwidth and sensitivity ( i. e., lower their hearing threshold) ( see page 24 for a discussion of hearing mechanisms). Examples of specialists include goldfish, catfish, some squirrelfish, herrings and relatives, and many other taxonomically diverse species. Quite often, hearing specialists will detect signals up to 3,000 – 4,000 Hz, with thresholds that are 20 dB or more lower than the generalists. The Pacific Coast fishes that are known hearing specialists are the sardine and related species ( taxonomic order Clupeiformes - e. g., herrings, shads, menhaden, anchovies). While it is likely that there may be other hearing specialists among Pacific Coast fishes, this cannot be determined without additional experimental studies on hearing capabilies of these species.
The majority of fishes do not have specializations to enhance hearing and are therefore called hearing generalists ( e. g., Popper et al. 2003; Ladich and Popper 2004). Based on what we know about the ears and auditory systems of species related to those on the Pacific Coast, we can tentatively conclude that the majority of the native fishes on the Pacific Coast are hearing generalists. Moreover, the auditory anatomy of the ears of some of these species indicates that they are most likely generalists.
At the same time, it must be pointed out that data on hearing capabilities exist for perhaps only 100 of the 27,000 or more extant species of fish ( see Popper et al. 2003) and so any extrapolation of hearing capabilities between different species, and especially those that are taxonomically distant, must be done with the greatest caution. Thus, studies of hearing capabilities of at least a number of the species on the Pacific Coast may be of considerable value in trying to understand whether or not the sounds generated by pile driving are within the hearing range of the species in question, and whether there are other hearing specialists in the region.
The species for which hearing capability data should be obtained must not only provide insight into their own hearing capabilities but also provide insight into related species. Thus, the species studies should include those that vary in auditory system structure ( e. g., ear) and in other ways, such as whether the fish has a swim bladder or not, and if the swim bladder is or is not connected to the gut for gas exchange ( physostomus vs. physoclistus – see Glossary, page 61). To facilitate data acquisition, the best approach is to use physiological recordings from the brain ( the aforementioned ABR) as opposed to the far more time- consuming behavioral studies done in the past that involved training animals to perform behavioral tasks when they heard sounds.
6 Despite this division into “ specialists” and “ non- specialists,” it is becoming more apparent as we gain more knowledge of fish hearing that there may be a “ continuum” of hearing capabilities in fish with “ specialist” and “ non- specialist” being the opposite ends of the continuum.
21
As indicated above, there are no data on hearing capabilities of any of the fish species in Pacific Coast estuaries and bays that are potentially of concern with regard to human- generated sound ( Table 2, page 17), although there are data for one or two species of lesser concern such as the American shad ( Alosa sapidissima) ( reviewed in Ladich and Popper 2004). It is likely that the hearing generalists among this group of fishes detect sounds only to 1,000 – 1,500 Hz. Studies of hearing capabilities ( albeit very limited and very much in need of replication) suggest that the sharks and rays probably do not detect sounds at frequencies above 800 to 1000 Hz ( e. g., Banner 1967; Nelson 1967; Myrberg 2001; Casper et al. 2003). Moreover, sturgeon is an unknown with regard to hearing capabilities.
There are no reliable hearing data on mackerels or rockfish, and it is not possible to predict their hearing capabilities based on morphology because there are no appropriate morphological data in the literature. The only even remotely related data available are for the Japanese horse mackerel ( Trachurus japonicus), which is reported to be able to detect sounds from 70 to 3,000 Hz ( Chung et al. 1995). 7
The very limited data in the literature on plaice and other related species of flatfish suggest that the Pacific Coast species are likely to have poor hearing sensitivity ( high thresholds) and a relatively narrow bandwidth. For example, Chapman and Sand ( 1974) reported that the plaice, Pleuronectes platessa, is able to detect sounds at frequencies up to 200 Hz. In contrast, Zang et al. ( 1998) suggest that the marbled sole ( Pleuronectes yokohamae) can detect sounds to 1,000 Hz with lowest thresholds at around 300 Hz. This relatively poor hearing capability is likely related to these fishes not having a swim bladder, a structure that appears to widen the bandwidth and lower auditory thresholds in many species.
Salmonids are one of the most important groups of fish because of their ESA listed status and commercial value, and yet the extent of data on their hearing is limited to the Atlantic salmon ( Salmo). Hawkins and Johnstone ( 1978) showed that this species can detect sounds to frequencies somewhat above 600 Hz ( unpublished data on the rainbow trout [ the land- locked version of steelhead, Onchorhyncus mykiss] shows hearing capabilities that are similar to those for Salmo [ Popper et al. in prep. a]), while more recent data show that Salmo are also able to detect sounds to well below 20 Hz ( Knudsen et al. 1992, 1994). It has been suggested that their response to sound at frequencies generally below about 35 Hz could be useful as a way of keeping fish from entering small areas such as irrigation ditches ( Knudsen et al. 1994). It appears, however, that these fish only respond when they are very close to the infrasound source, most likely because very low- frequency sound will not propagate in shallow water ( Rogers and Cox 1988).
One must be careful about extrapolating from Atlantic salmon to Pacific Coast salmonids. Data on the anatomy of the ear of several salmonid species ( Popper 1976, 1977)
7 This work, and that of Zhang et al. on flatfish were only seen in abstract form and it was therefore not possible to determine the methods used in the study, which was written in Japanese. The hearing bandwidth of the mackerel in the Chung study is substantially wider than for any other non- specialist fish. Moreover, the bandwidth for the flounder reported by Zhang et al. ( 1998) is far wider than that reported for another species of the same genus by Chapman and Sand ( 1974). Therefore, without a careful analysis of the methods and results these data must be viewed with considerable caution.
22
suggest that the auditory system is similar in all salmonids, and this is supported by recent work on the rainbow trout ( Oncorhynchus mykiss) ( Popper et al. in prep. a) a member of the same genus as many ESA listed Pacific Coast salmonids. Still, without at least some additional hearing data this extrapolation must be done with considerable caution. Thus, it would be of great value to have hearing data on at least a few of the species in Pacific Coast aquatic habitats. Moreover, such data would be of particular value if it were for animals of different life stages and sizes since it is possible that hearing capabilities change with age and/ or size of the fish, as has been suggested for the hake, a relative of the cod ( Lombarte and Popper 1994, 2004). Thus, a strategy to select fish of different size, life history, and morphology for future studies, should be considered.
At the same time, the specific species to study would depend on their regulatory status and factors that could be decided later. The critical point to make is that we have too few data on hearing in any of the groups of fishes found in the Pacific Coast region, and there are not even data in the literature on related species from which it would be possible to cautiously extrapolate to most of the Pacific species.
3. Auditory Structures
The basic mechanism for transduction of sound into electrical signals compatible with the nervous system is the sensory hair cell ( Fig. 4, page 69). This cell is ubiquitous in the ears of all vertebrates. The same cell is also found in the lateral line system, a series of receptors along the body of the fish that detects water motion relative to the fish that arise from sources within a few body lengths of the animal ( e. g., Coombs and Montgomery 1999; Popper et al. 2003) ( see page 40 for a discussion of the lateral line).
The body of the sensory hair cell is typical of most other cells; however, the hair cell also has an apical group of projections called the ciliary bundle that extends above the surface of the epithelium in which the cell lies ( the sensory epithelium, often referred to as the “ macula”). Bending of the ciliary bundle causes the opening of very tiny channels in the cilia and the entry of ions from the surrounding fluid into the cell ( e. g., Hudspeth and Corey 1977). This produces a series of very rapid chemical events that culminate in the release of chemicals called neurotransmitters from the cell body. These neurotransmitters then stimulate the neurons that contact ( innervate) the sensory cells. The neurons, in turn, send electrical signals to the brain that provide information about the sound.
Fishes, like other vertebrates, have two inner ears that lie within the cranial ( brain) cavity just lateral to the brain as shown in Figure 5 ( page 70). Unlike terrestrial vertebrates, fishes have no middle or external ear. The structure of the fish inner ear is similar to that found in all other vertebrates ( Ladich and Popper 2004). The basic mechanisms of stimulation of the hair cells in the inner ear, and the conversion of acoustic energy to electrical signals compatible with the nervous system, are the same in all vertebrates.
The inner ear ( Fig. 6, page 70) has three semicircular canal ducts, which are small tubes that lie in nearly orthogonal planes to one another. These canals serve to detect angular acceleration ( e. g., rotational acceleration of the head) in fishes just as they do in terrestrial
23
vertebrates. In addition, fishes have three fluid- filled otolith organs ( utricule, saccule, and lagena), each of which contains a dense calcified otolith that overlies a sensory epithelium that contains numerous sensory hair cells. These otolith organs subsume two roles for fish. First, they serve as vestibular organs and measure the position of the head in the vertical direction relative to gravity and in other directions relative to the acceleration of the body, just as they do for terrestrial vertebrates. Second, they are involved in sound detection. The earliest work suggested that the primary auditory end organs in fishes were the saccule and lagena, but there is a growing body of evidence that now suggests that all three of the otolithic end organs have roles in hearing ( reviewed in Popper et al. 2003; Ladich and Popper 2004).
Each otolithic end organ may have many thousands of sensory hair cells. Fishes, unlike most tetrapods other than amphibians, continue to produce sensory hair cells throughout much of their lives ( Lombarte and Popper 1994, 2004; Higgs et al. 2003). In addition, there is evidence that fishes, unlike mammals, can replace sensory cells that have been damaged as a result of exposure to certain drugs ( Lombarte et al. 1993), although there have been no studies to determine if fishes can replace sensory cells that have been killed as a result of exposure to sound.
4. Hearing Mechanisms
Hearing is based on the detection of the mechanical motions in the medium imparted by sound. In fishes, the otolith organs are stimulated directly by the acoustic particle velocity associated with underwater sound fields. In addition, the organs can be stimulated indirectly by particle motions created when sound pressure fluctuations are transformed into motion by a gas- filled accessory organ such as the swim bladder ( see below).
In effect, hearing is based on relative motion between the fish’s body and the otolith. As indicated earlier, the sensory hair cells have an apically located tuft of “ cilia” ( Fig. 4, page 69). Because the body of fish is primarily composed of water, it will move at approximately same amplitude and phase as the impinging sound. The otoliths, however, are stiffer and about three times denser than the rest of the body ( including the sensory epithelium). Thus, the otoliths will move at different amplitude and phase from that of the epithelium, and this differential motion causes the intervening ciliary bundles on the sensory hair cells embedded in the epithelium to move, resulting in the detection of sound.
Similarly, the air- filled swim bladder ( or other gas bubble in the body) will be stimulated by the pressure component of the sound field. The swim bladder then serves as a small transducer that re- radiates energy in the form of particle motion, which is again detectable by the inner ear.
In hearing generalists, the primary acoustic energy is provided by the direct stimulation of the ear, though it is possible that some additional energy is re- radiated from the swim bladder and that this could lower hearing thresholds and/ or increase the bandwidth of detection. As a result, hearing generalists are likely to primarily be sensitive to the particle motion component of the sound field. While earlier studies ( e. g., Fig. 3, page 68) tend to report hearing data in terms of
24
pressure, it will be important that future studies examine the response of fish to particle motion in order to get a fuller, and more accurate, assessment of hearing capabilities of these fishes.
In contrast to hearing generalists, hearing specialists have evolved a number of different mechanisms to acoustically couple the swim bladder ( or other gas- filled structure) to the ear, thereby allowing the auditory systems to detect the pressure component of the sound field. These mechanisms directly transmit motion of the swim bladder or other gas- filled structure, which is induced by sound pressure, to the inner ear, thereby providing a substantial pressure input to supplement the direct detection of particle displacement. This coupling increases hearing sensitivity ( i. e., lowers thresholds) and bandwidth of detection as compared to generalists ( see Popper et al. 2003 for review).
Specializations that enhance hearing vary among different species. They may range from having an extension on the swim bladder that has its rostral termination very close to the ear, as in some croakers and drums ( family Sciaenidae) ( Ramcharitar et al. 2001) to a direct mechanical connection between the swim bladder and ear as found in the otophysan fishes ( catfish, goldfish, and relatives). Finally, there are some species that have an extension of the swim bladder, or a separate bubble of gas, that is tightly associated with the ear, or which lies near the ear ( e. g., all herrings and shads and relatives, mormyrids).
IV. Effects on Fish from Exposure to Sound
Interest in the effects of human- generated sound on aquatic organisms has grown considerably in the past decade ( e. g., NRC 1994, 2000, 2003; Richardson et al. 1995; NRDC 1997). While these reports, and a handful of research studies, have primarily focused on marine mammals, several have raised the issue that sounds that potentially affect marine mammals may also affect other aquatic organisms, including fishes and invertebrates ( e. g., NRDC 1994, 2000; Popper 2003; Popper et al. 2004). The basis for concern about the potential effects of sound on fishes is the well- documented effects of exposures to sound on behavior, hearing, and overall physiology in terrestrial animals and humans ( Lenhardt 1986; NIH 1990). While this extrapolation is certainly valid, there is also a more direct, albeit limited, body of literature that demonstrates the effects of exposure to sound on fishes, including the aforementioned observations that sound from pile driving can injure and even kill fish close to the source. These more specific studies are the basis of the following discussion.
Results of the few peer- reviewed studies on the effects of sound on fishes are discussed in this section. The specific studies are outlined, by type, in Table 3 ( page 26) to give an overview of the investigations and to show gaps in the literature that must be filled if we are to understand overall effects of sound on fishes, and the specific effects of exposure to sounds produced by pile driving. The information in this table should be used with that found in Table 4 ( page 46) to understand specific needs with regard to identifying potential effects of pile driving on fish.
25
A. Behavioral Responses and Masking of Biologically Relevant Sounds
There have been very few studies of the effects of anthropogenic sounds on the behavior of fishes. Data are lacking not only on the immediate behavioral effects on fishes close to a source, but also effects on fishes further from the source. Moreover, nothing at all is known about the long- term effects of pile driving on fish behavior or the effects of cumulative exposure to loud sounds.
Several studies have demonstrated that human- generated sounds may affect the behavior of at least a few species of fish. For example field studies by Engås et al. ( 1996) and Engås and Løkkeborg ( 2002), while not actually observing the behavior of fish per se, showed that there was a significant decline in catch rate of haddock and cod that lasted for several days after termination of air gun use, after which time the catch rate returned to normal. The authors concluded that the catch decline resulted from the sound of the air guns, and that the sound probably caused the fish to leave the area of ensonification, although there was no direct data to support this conclusion. More recent work from the same group ( Slotte et al. 2004) showed parallel results for several additional pelagic species including blue whiting and Norwegian spring spawning herring. Slotte et al. found that fishes in the area of the air guns appeared to go to greater depths after insonification8 compared to their vertical position prior to the air gun usage. Moreover, the abundance of animals 30- 50 km away from the ensonification increased, suggesting that migrating fish would not enter the zone of seismic activity. A non- peer reviewed report by Gausland ( 2003), however, suggests that the declines in catch rate observed in these studies may have resulted from other factors and are not statistically different than the normal variation in catch rates over several seasons.
Similarly Skalski et al. ( 1992) showed a 52% decrease in rockfish catch when the area of catch was exposed to a single air gun emission at 186- 191 dB re 1 μPa ( mean peak level) ( see also Parsons 1987; Pearson et al. 1992) They also demonstrated that fishes would show a startle response to sounds as low as 160 dB, but this level sound did not appear to elicit decline in catch.
Table 3: Citations of selected studies examining the effects of exposure to sound on fishes that have most relevance to pile driving. Note that there are no data for non- teleosts ( e. g., sturgeon, elasmobranchs).
Issue
Hearing Generalists
Hearing Specialists
Mortality
Yelverton et al. 1975 ( guppy, bluegill, trout, bass, carp; explosive blasts)
Yelverton et al. 1975 ( goldfish, catfish, minnow; explosive blasts)
Hastings 1995 ( goldfish and gouramis; pure tones)
Physical Injury
Yelverton et al. 1975 ( guppy, bluegill, trout, bass, carp; explosive blasts)
Govoni et al. ( 2003) ( larval fish; explosive blasts, no pathology seen)
Yelverton et al. 1975 ( goldfish, catfish, minnow; explosive blasts)
Hastings 1995 ( goldfish and gouramis; pure tones)
Auditory Tissue Damage
Enger 1981 ( cod; pure tones, 1 – 5 hr)
Hastings et al. 1996 ( oscar; pure tones, 1 hr)
McCauley et al. 2003 ( pink snapper, air gun)
Hastings 1995 ( goldfish; pure tones, 2 hr)
8 See glossary under “ ensonification” for differentiation between ensonification and insonification.
26
Issue Hearing Generalists Hearing Specialists
Permanent Threshold Shift ( PTS)
No data available
No data available
Temporary Threshold Shift ( TTS)
No relevant data available
Smith et al. 2004a, b ( goldfish; band- limited noise)
Scholik and Yan 2001 ( fathead minnow; band- limited white noise)
Popper and Clarke 1976 ( goldfish; pure tones)
Behavioral Changes
Wardle et al. 2001 ( Exposed fish and invertebrates on reef to continuous air gun with no significant behavioral changes)
No data available
Eggs and Larvae
Banner and Hyatt 1973 ( Cyprinidon and Fundulus showed somewhat decreased egg viability and larval growth in tanks with increased noise)
Kostyuchenko 1973 ( Increased egg mortality up to 20 m from seismic source)
Booman et al. 1996 ( Variable results with some stages showing decreased growth in a few species when exposed to air guns)
No data available
Miscellaneous
Skalski et al. 1992 ( Sebastes catch decreased after one air gun blast)
Engås et al. 1996 ( Haddock and cod catch reduction after seismic survey blasts)
Engås and Løkkeborg 2002 ( Haddock and cod catch reduction area after seismic survey blast
Slotte et al. 2004 ( herring & blue whiting do not enter the area of air gun during use)
Smith et al. 2004a ( no change in corticosteroid levels after continuous exposure to band limited noise)
Wardle et al. ( 2001) used a video system to examine the behaviors of fish and invertebrates on a coral reef in response to emissions from seismic air guns that were carefully calibrated and measured to have a peak level of 210 dB re 1 μPa at 16 m from the source and 195 dB re 1 μPa at 109 m from the source.. They found no permanent changes in the behavior of the fish or invertebrates on the reef throughout the course of the study, and no animals appeared to leave the reef. There was no indication of any observed damage to the animals..
Finally, Turnpenny et al. ( 1994) examined the behavior of three species of fish in a pool in response to different sounds, but results are not useable because of lack of calibration of the sound field at different frequencies and depths, and many other problems with experimental design ( Ellison 1996). In enclosed chambers that have an interface with air, such as tanks and pools used by Turnpenny et al., the sound field is known to be very complex and will change significantly with frequency and depth.
27
While not totally germane to fishes, there is some evidence that an increased background noise ( for up to three months) may affect at least some invertebrate species. Legardère ( 1982) demonstrated that sand shrimp ( Crangon crangon) exposed in a sound proof room to noise that was about 30 dB above ambient for three months demonstrated decreases in both growth rate and reproductive rate. In addition, Legardère and Régnault ( 1980) showed changes in the physiology of the same species with increased noise, and that these changes continued for up to a month following the termination of the signal.
Indeed, we are now aware that fishes, as mammals and probably all other vertebrates, glean a great deal of information about their environment from the general sound field. In other words, whereas visual signals are very important and useful for things near the animal and in the line of sight, substantial information about the unseen part of an animal’s world comes from acoustic signals. 9
One may therefore think of fishes as using two “ classes” of sound. The first is the well- known group of communication signals used to keep in touch with other members of a species and detect the presence of predator or nearby prey. The second are the sounds of the environment that, for a fish, might include the sounds produced by water moving over a coral head, waves breaking on shore, rain, and many more physical and biological sources. Bregman ( 1991) coined the term “ Auditory Scene” to describe the acoustic environment. The acoustic environment has become of increasing importance in the overall understanding of hearing for all animals during the past 15 years. Moreover, it is becoming increasingly clear that one of the major roles of the auditory system is to discriminate between, and determine the position, of sounds in the auditory scene, using a mechanism called “ stream segregation” ( Bregman 1991; Fay and Popper 2000; Popper et al. 2003) whereby an organism is able to distinguish between two sounds (“ streams”) that differ in some way such as direction of the source, frequency spectrum, etc.
B. Stress – Physiological Responses
The impact of stress is much more difficult to define because it is hard to quantify this measure in fish since it has not been extensively studied; however, increased background noise is known to increase stress in humans ( e. g., NIH/ CDC 1990; von Gierke and Eldred 1993; Pearsons et al. 1995). There is evidence that the effects on non- auditory aspects of an animal's physiology, such as an increase in stress levels, can come from increased background noise or exposure to a sudden increase in sound pressure ( e. g., Hattingh and Petty 1992). In turn, increased stress does impact overall human health and well- being, and it is reasonable to suggest that sound is also likely to cause stress in fishes. Thus, a considerable concern with regard to aquatic organisms ( as to humans and other terrestrial organisms) not only related to the impact of exposure to sound on the function of the auditory receptor, but also to the impact of any sounds that are above ambient levels on overall health and well- being.
9 Whereas terrestrial animals can also gain a good deal of information from chemical signals, such signals are less useful in water where there is a very high and rapid dilution factor. As a consequence, chemical signals are not as useful in water as in air, unless animals are very close to one another ( e. g., within visual range) ( Tavolga 1971).
28
An early study by Gilham and Baker ( 1984) used vibration of the aquarium walls to measure stress responses in rainbow trout. Although the stressors were not quantifiable, this study demonstrated that a general stress response occurred in fish between 1 and 5 days after signal onset, as demonstrated by significant increases in serum cortisol levels. Other studies have demonstrated that exposure to non- traumatic stressors ( i. e., crowding, spawning, rapid environmental changes, suboptimal water quality or physical environment, altered conductivity, and pollution) can predispose fish to opportunistic infections ( e. g., Walters and Plumb 1980; Noga et al. 1998; Wedemeyer 1999; Pickering 1981). Mesa ( 1994) found that salmonid smolts going over a hydroelectric dam had significant stress effects, and this made them more susceptible to predation than fish that had not gone over the dam. While the stimulus in this study was vastly different than just high- level sound, the results do suggest that if exposure to sound results in highly stressed fish, even if fish do not die as a direct result of stimulation, they may be more susceptible to predation or other environmental effects than non- stressed fish.
Smith et al. ( 2004a) demonstrated that corticosteroid levels do not change at a statistically significant level in response to exposure to high sound levels in goldfish ( Carassius auratus). Corticosteroid level is a measure of stress, and suggests that stress levels in these animals were not influenced by continuous exposure to band limited noise in the 0.1 – 10 kHz frequency band with an overall rms pressure level of 170 dB re 1 μPa. While these results are of considerable interest, they provide only the first data in what must be a larger- scale examination of effects of sounds on stress levels in fishes. Indeed, Smith et al. ( 2004a) recommended additional studies on goldfish, and similar studies must be done on other species as well. One must be cautious, however, in extrapolating between species and between different experimental paradigms in trying to understand the effects of potential stressors on physiology.
C. Temporary and Permanent Hearing Loss
There are two classes of effects of exposure to sound on the ear. Exposure to low levels of sound for a relatively long period of time or exposure to higher levels of sound for shorter periods of time may result in temporary hearing loss, referred to as temporary threshold shift or TTS ( e. g., Lonsbury- Martin et al. 1987). The level and duration of exposure that causes TTS varies widely and can be affected by factors such as repetition rate of the sound, pressure level, frequency, duration, health of the organisms, and many other factors. By definition, hearing recovers after TTS. The extent ( how many dB of hearing loss) and duration of the TTS may continue from minutes to days after the end of exposure, and the extent of TTS depends on many variables.
The second possible effect is referred to in the literature as permanent threshold shift or PTS. PTS is a permanent loss of hearing and is generally accompanied by death of the sensory hair cells of the ear ( e. g., Saunders et al. 1991).
Laboratory studies have been used to determine whether there may be temporary or permanent changes in hearing ability in animals exposed for short or long periods of time to different types of sound ( e. g., pure tones or white noise). TTS has been found using behavioral or physiological tests for several fish species, including goldfish ( Carassius auratus) and fathead
29
minnows ( Pimephales promelas) ( e. g., Popper and Clarke 1976; Scholik and Yan 2001, 2002; Smith et al. 2004a, b). 10
In a recent set of studies, Smith et al. ( 2004a, b) tested hearing in goldfish and tilapia ( Oreochromis niloticus) to determine more detailed parameters of hearing loss, including the effects of different exposure durations and recovery times. They demonstrated that goldfish had a 5- dB temporary threshold shift after only 10 minutes of exposure to band- limited noise ( 0.1 to 10 kHz, approximately 170 dB re 1 μPa overall spectral sound pressure level), and that goldfish with a three- week exposure to the same stimulus experienced a 28- dB threshold shift and took over two weeks to return to normal hearing. Similarly, Scholik and Yan ( 2001) demonstrated that fathead minnows did not recover to control thresholds even as long as 14 days after termination of a 24- hour exposure to white noise from 0.3 to 2.0 kHz with an overall spectral sound pressure level of 142 dB re 1 μPa. 11
Both Smith et al. ( 2004a) and Scholik and Yan ( 2002) reported no TTS in two species of hearing generalists ( tilapia and bluegill sunfish, Lepomis macrochirus).. In their experiments, however, the sound stimulus consisted of only pressure, whereas primarily the acoustic particle velocity in a sound field stimulates hearing generalists. Moreover, these researchers controlled only the sound pressure level during ABR measurements to determine the audiograms before and after sound. Thus these results are not relevant because hearing sensitivity and temporary threshold shifts in hearing generalists are meaningful only in terms of acoustic particle velocity or acoustic intensity. It is unlikely that TTS would be found in a hearing generalists exposed only to sound pressure.
Two recent investigations used exposure to high- level anthropogenic sources to test hearing loss and effects on other aspects of fish biology. In one study, Popper et al. ( in prep. a) used a U. S. Navy SURTASS low frequency active ( LFA) sonar transducer and tested hearing in the rainbow trout ( Oncorhynchus mykiss), a salmonid of the same genus as many endangered Pacific Coast salmon, and channel catfish ( Ictalurus punctatus) ( to represent hearing specialists). They found that exposure to three emissions of the sonar, each for 108 seconds ( with nine minutes between signals) with a received sound level at the fish of 193 dB re 1 μPa ( rms) ( frequency bandwidth about 160- 325 Hz), produced up to 10 dB of temporary threshold shift. Preliminary evidence suggests that the fish recovered from this TTS within 24- 48 hours. Moreover, there was no mortality to fish exposed to the SURTASS LFA source, even up to four days ( 96- hours) post- exposure. Preliminary results from necropsy and histopathology on experimental and control animals showed no damage to other organ systems, including the ear.
10 It is important to note that the sound levels expressed in these TTS studies were done based on sound pressure level, but should more correctly be determined in terms of cumulative energy exposure. Future experiments need to be done in such context to allow comparison between studies, animal groups, and, most importantly, different signal parameters ( e. g., bandwidth, duration, temporal variation). The importance of the studies cited here lie with the observations that TTS does take place in fish, and that the effects of TTS may last for a considerable time after the termination of the sound source.
11 It must be noted that it is unclear in this study if the units on the sound pressure level were reported correctly. For band limited white noise, the correct units for spectral density level are μPa2/ Hz and one has to multiple by the total bandwidth to get the overall sound pressure level. A spectral density level of 142 dB over a bandwidth of 1700 Hz would have an overall sound pressure level of 174 dB ( re 1 μPa), which is on par with the level found to cause TTS in goldfish, another hearing specialist, by Smith et al. ( 2004a, b). This study needs to be duplicated to confirm that TTS occurs in fathead minnows when exposed to white noise at such a low sound pressure level.
30
In the second study, Popper et al. ( in prep. b) tested three species, including another salmonid ( broad whitefish, Coregonus nasus), after stimulation with five blasts of a seismic air gun with a received mean peak sound level of about 205 dB re 1 μPa ( a received mean SEL of about 175 dB re 1 μPa2- s). The broad whitefish showed no TTS to this signal, whereas northern pike ( Esox lucius) ( a hearing generalist) and lake chub ( Couesius plumbeus) ( a hearing specialist) showed 10- 15 dB of hearing loss and complete recovery within 24 hours after exposure.
While extrapolation between these results and exposure to sound produced by pile driving, or between these species and most Pacific Coast species, must be done with considerable caution, they do suggest that limited exposure to high- level sounds of some types does not result in death, and that any hearing loss is likely temporary.
In humans and other terrestrial vertebrates, exposure to extremely high sound pressure levels, for even a short period of time, may result in permanent hearing loss. This occurs because the sound mechanically destroys the sensory hair cells of the inner ear and/ or fractures or dislocates the ossicular chain of the middle ear ( Roberto et al. 1989; Patterson and Hamernik 1997). It is significant that exposure to sound at lower levels, but for longer periods, as in a noisy work environment, can also lead to permanent hearing loss through death of sensory cells ( Kryter 1985; Hamernik et al. 1994). At the same time, the data on the effects of exposure to these types of sounds on fishes are very limited as compared to data for terrestrial vertebrates; but, there is a small body of peer- reviewed literature showing that such exposures to sound can destroy the sensory cells in fish ears ( Enger 1981; Hastings et al. 1996; McCauley et al. 2003). No study, however, has yet examined the relationship between destruction of hair cells and permanent hearing loss in fishes.
While looking for evidence of frequency discrimination in the ears of Atlantic codfish ( Gadus morhua), Enger ( 1981) found that some sensory cells lost their ciliary bundles ( and were potentially destroyed) after one- to five- hours exposure to pure tones at frequencies from 50 to 400 Hz with a sound pressure level of 180 dB re 1 μPa ( rms) This is 100- 110 dB above threshold in the most sensitive hearing frequency range for codfish. Enger used a waveguide instrumented with a sound projector at each end to produce an exposure that had negligible particle velocity.
In a similar study, Hastings ( 1995) reported damage to auditory hair cells in goldfish ( Carassius auratus) exposed to continuous tones having levels of 189, 192, and 204 dB re 1 μPa ( peak) at 250 Hz and 197 dB re 1 μPa ( peak) at 500 Hz for approximately two hours. These sound pressure levels are at least 120- 140 dB above auditory thresholds for goldfish at these frequencies. This study also used a waveguide that allowed exposure to sound with negligible particle velocity. Four fish were exposed at each set of conditions, and damage was found to correlate with sound pressure level at a 95% confidence level. This study also included several controls ( fish placed in the waveguide and held for 2 hours, but not exposed to sound). In addition, goldfish exposed to 182 dB re 1 μPa ( peak) at 500 Hz had no hair cell damage. This is approximately 120 dB above auditory threshold at a frequency in the most sensitive range.
Hastings et al. ( 1996) demonstrated similar effects on the ears of the oscar ( Astronotus ocellatus). Sensory cells in the ears of four out of five animals were damaged after one hour of continuous exposure to a 300- Hz pure tone at 180 dB re 1 μPa ( peak), while no damage occurred
31
after one- hour exposure to the same tone applied for 20% of each minute. It is important to note that damage was not evident in animals after one day, but was evident in the animals that were kept alive for four days following exposure. These results suggest that damage from exposure to sound takes some time to become visually apparent. The particle velocity in the waveguide used by Hastings et al. ( 1996) was about one- fifth that which would be associated with the same acoustic pressure in open water; however, because the stimulus was a plane wave, the average acoustic intensity for these exposures, 1.67 J/ m2- s, can be calculated. This gives an estimated total average acoustic energy exposure of 6000 J/ m2 ( after one hour) for hair cell damage, and 1200 J/ m2 resulting in no hair cell damage.
McCauley et al. ( 2003) investigated the effects of exposure to the sounds of a seismic air gun on the Australian fish, the pink snapper ( Pagrus auratus). Fish were in a cage and exposed to several air gun emissions at different distances. The animals were kept alive for different time intervals after exposure. After the animals were sacrificed, their ears were examined for signs of damage using electron microscopic techniques very similar to those used by Hastings et al. ( 1996). The results clearly showed extensive damage to the sensory hair cells of the ear in several of the animals. The extent of damage increased with the post- exposure period up to at least 58 days ( the maximum survival interval described).
While the McCauley et al. ( 2003) study further substantiated the potential for destruction of hair cells in fish when exposed to high level sounds, both the McCauley et al. ( 2003) and Hastings et al. ( 1996) studies were careful to provide a number of important caveats to their work. These included ( a) use of only a few species, which may not be representative of other species; ( b) the inability of the caged fish to escape from the sound field; and ( c) the relatively long duration of exposure as compared to exposures to what would be expected from other types of human- generated sounds of that sound pressure level.
One difference between these studies that needs to be controlled in future investigations is the relationship between acoustic pressure and particle velocity in the sound stimulus. While it was possible for Hastings et al. ( 1996) to calibrate both pressure and particle velocity in their test chamber, this was not done by McCauley et al. ( 2003). The importance of having full characterization of the stimulus in these and future studies is to enable correlation of results with the specific component( s) of the sound stimulus and thus comparison of results between studies.
It again needs to be pointed out that hair cell damage observed in these four species ( codfish, goldfish, oscar, pink snapper) was only a visual manifestation of what may have been a much greater effect, and that observable physical evidence took days to show up. It may be more important to evaluate the more immediate effects of the sounds on hearing capabilities of the fish, as was done by Popper et al. ( in prep. a, b). Even if there is only TTS as a result of a loud sound, temporary deafness could result in a fish being unable to respond to other environmental sounds that indicate the presence of predators and facilitate the location of prey and mates. Effects, however, depend on the use of sound by that species in those situations.
While it is clear from the data discussed above that sounds of some types over time can affect the ear and hearing, it is important to note that at this stage of our knowledge, and the very limited data, that one must be extremely cautious in extrapolating results between species or received signals. Thus, results for one species may not be indicative of the results one would
32
obtain for another species using the same type of signal, and the results from one type of signal ( e. g., air gun) may not be germane to another signal ( e. g., pile driving).
The reasons for not being able to extrapolate results are many and include: ( a) differences in the hearing systems of different fish species and too little knowledge about the effects of exposure to sound on such different auditory systems; ( b) limited data on the precise nature of a stimulus ( e. g., pressure and/ or particle velocity) which might affect the hearing apparatus; and ( c) the time course ( e. g., continuous vs. transient) and frequency components of different signals. To be able to reasonably extrapolate between species and signals, much more will need to be known about the effects of sounds on different auditory systems.
Furthermore, it will be imperative to have common ways of expressing exposure to sound so that it is possible to compare stimulus parameters between signals of different types, especially for impact and impulse sounds. Indeed, several studies on guinea pigs and chinchillas have shown that the Leq and SEL metrics do not correlate with TTS and PTS when the stimulus includes impulsive or transient sounds. Lataye and Campo ( 1996) found that Leq was not an accurate predictor of noise- induced hearing loss ( NIHL) in guinea pigs when the temporal characteristics of the sound were varied. Their results indicate the existence of a “ critical intensity” as a demarcation between metabolic and mechanical damage mechanisms, which is not accounted for in the time- averaged Leq. Hamernik et al. ( 1974) showed that when continuous and impulse noises were combined, chinchillas experienced higher levels of TTS and hair cell damage than the additive effects of either component. In more recent studies, Hamernik and Qiu ( 2001) also found that NIHL did not correlate with Leq in chinchillas when the sound exposure included impact signals mixed with band- limited noise even though all exposures had the same total energy.
Both Hamernik and Qui ( 2001) and Hamernik et al. ( 2003) reported that the kurtosis metric ( a statistical measure used to estimate the deviation of the signal amplitude distribution from a normal distribution) correlated with TTS, PTS and outer hair cell loss for exposures that contained high- level transients ( impacts or noise bursts). They found no correlation with energy metrics. Given the results of these studies, it is premature to provide any guidance on exposure levels that could cause TTS or PTS in any fish species based on research reporting TTS when exposed only to pure tones or white noise signals.
Finally, it should be noted that the same concerns regarding stimulus parameters and extrapolation between species applies to all other aspects of the effects of exposure to sound on fishes ( or any animal, for that matter). Some of these other effects are discussed below.
D. Structural and Cellular Damage of Auditory and Non- Auditory Tissues
Compared to data for the effects of exposure to sound on fish hearing capabilities and the ear, there are even fewer peer- reviewed data regarding the effects on other aspects of fish biology. Little work has been done to determine the non- auditory effects of sound on fish. It is widely known that exposure to sounds at high levels can alter the physiology and structure of terrestrial vertebrates ( e. g., Fletcher and Busnel 1978; Saunders et al. 1991). Indeed, there are strong standards set by the Occupational Safety and Health Administration ( OSHA) recognizing
33
that high levels of background sound have an impact on human well- being ( e. g., NIH 1990; von Gierke and Eldred 1993; Pearsons et al. 1995). These changes may include cellular changes, organ system changes, or stress level effects caused by exposure to sound. Exposure to sound at ultrasonic frequencies (~ 750 kHz and higher) have even induced cardiac arrhythmias in humans and premature ventricular contractions in frogs ( Dalecki et al. 1991); however, these effects have not been observed at lower frequencies that characterize the sound produced by pile driving.
While there are far fewer data on the impact of exposure to sound on the health and well- being of laboratory animals, and far less known about the impact of exposure on wild animals ( including aquatic animals), it is reasonable to suggest that long- term exposure to relatively high levels of sound could impact all vertebrates ( e. g., Richardson et al. 1995) ( just as does shorter term exposure to higher level signals). For example, one of the organ systems of most concern with marine mammals is the lungs, and the resultant damage that may occur in this organ due to the presence of air. Most fishes do have at least one large air chamber, the swim bladder, which provides the same discontinuity between water and air as does the lung in marine mammals.
Studies on terrestrial mammals have indicated that gas- filled structures ( i. e., lung) or gas pockets ( such as could occur in the gastrointestinal tract) within a body make it susceptible to damage by sound ( Richmond et al. 1973; Fletcher et al. 1976; Yang et al. 1996; Bauman et al. 1997; Dodd et al. 1997; Elsayed 1997). Tissue damage can occur when sound passes through the interface from a fluid tissue structure ( e. g., adipose tissue and muscle) to a gas void because the gas is more compressible then the fluid, resulting in a relatively large increase in the motion of the connective tissue between the two. In addition, sound will cause gas organs such as the swim bladder and lung to oscillate and push on the surrounding tissues. The amplitude of these oscillations can be quite large at high sound pressure levels or even at lower sound pressure levels if the gas organ is excited at its resonance frequency.
In fishes, gas oscillations induced by high sound pressure levels can even cause the swim bladder to tear or rupture, as has been indicated in response to explosive stimuli in several gray literature reports ( e. g., Alpin 1947; Coker and Hollis 1950; Gaspin 1975; Yelverton et al. 1975). While similar results have been observed in fish exposed to pile driving sound ( Caltrans 2001, 2004; Abbott and Bing- Sawyer 2002), there was no swim bladder damage in response to exposure to SURTASS LFA sonar ( received level of 193 dB rms re 1 μPa) or seismic air guns in two recent studies ( Popper et al. in prep. a, b). Both of these sources produce transient sounds having approximately the same frequency bandwidth as pile- driving signals.
It has been suggested that the large negative overpressure characteristic of pile driving sounds may be more damaging to the swim bladder than the initial positive overpressure ( Trasky 1976; Caltrans 2004) because of its relatively large expansion during the negative phase. Bailey et al. ( 1996), however, found that a sound pulse having a large positive peak overpressure was at least as damaging as one having a large negative peak overpressure of approximately the same level and time duration, to the lungs of mice submerged in water. Damage increased with magnitude of pressure incident at the lung, but histology showed no qualitative or quantitative differences between the effects of positive and negative pressures. Mouse lungs had increasing hemorrhage with increasing exposure levels regardless of the polarity of the peak overpressure. These findings indicate that injury would correlate with the work done on the lung tissue, which would be equivalent to the total energy in the sound wave.
34
Although it is possible to electronically invert pressure waveforms when they are recorded because of the characteristics of the instrumentation and potentially misinterpret the results of such tests, Bailey et al. ( 1996) created a sound pulse having a large negative peak overpressure by reflecting the pulse with positive peak overpressure from the interface between water and air at the top of their test tank. The sound pressure reflection coefficient at this interface is – 1. Therefore, because they used a signal and its reflection from the water surface and found no difference in damage based on polarity of peak overpressure, they obtained legitimate results even if their recording instrumentation happened to invert the pressure waveforms.
@@
Govani et al. ( 2003) also concluded that the total energy in the sound wave, regardless of pressure polarity, was responsible for observed effects of submarine detonations on juvenile pinfish ( Leiostomus xanthurus). Moreover, Stuhmiller et al. ( 1996) concluded that incidence of blast injury to the lung and lethality correlates with total energy in the wave normalized by lung volume in terrestrial animals. Their results were so strong that they recommended that potential for blast injury to the lung could be determined from this simple model without additional animal testing.
Other structures within the body can be affected by exposure to sound because of their small size or dynamic characteristics. There is some evidence to suggest that sound at sufficiently high- pressure levels can generate bubbles from micronuclei in the blood and other tissues such as fat ( ter Haar et al. 1982). In fish, blood vessels are particularly small in diameter so bubble growth by rectified diffusion ( Crum and Mao 1996) at low frequencies could create an embolism and burst small capillaries to cause superficial bleeding. This type of bubble growth may also occur in the eyes of fish where the tissue might have high levels of gas saturation ( see non peer- reviewed reports by Turnpenny et al. 1994; Gisiner 1998).
Another type of tissue damage caused by exposure to high- level, transient sound is traumatic brain injury ( TBI) or neurotrauma. In humans, TBI can occur with no marks of external injury, but manifests itself with instantaneous loss of consciousness or sustained feelings of anxiety and confusion, or amnesia, and may result in death ( Elsayed 1997; Knudsen and Oen 2003). The underlying physical mechanisms for these manifestations are cerebral edema, contusions and lacerations, as well as hemorrhages in the meninges ( protective tissues around the brain), brain substance, nerve roots, and ventricles ( fluid- filled spaces within the brain and spinal cord) that may result from extreme relative motion between the skull and brain during exposure to high overpressures. Hastings ( 1990, 1995) reported “ acoustic stunning” in four blue gouramis ( Trichogaster trichopterus) exposed for approximately eight minutes to a 150- Hz pure tone with a peak pressure of 198 dB re 1 μPa. Three out of four of these fish recovered. The loss of consciousness exhibited by these fish could have been caused by neurotrauma, especially since this species has a bubble of air in the mouth cavity located near the brain that enhances hearing capability of this species ( Yan 1998; Ladich and Popper 2004). Thus fish with swim bladder projections or other air bubbles near the ear ( e. g., butterfly fish, squirrel fish, and many other species) could be susceptible to neurotrauma when exposed to high sound pressure levels.
Elsayed ( 1997) conducted a series of investigations using terrestrial animal models to examine biochemical responses in tissues to blasts. He and his colleagues have found two
35
responses that correlate with blast overpressure: ( 1) depletion of antioxidants and ( 2) lipid pre- oxidation. Cernak et al. ( 1996) also examined biochemistry related to neurotrauma in blast injury. They also found lipid pre- oxidation products as well as increased levels of lactate and calcium ions and decreased levels of glucose and magnesium and zinc ions. Changes in lactate and glucose levels indicate changes in metabolism and energy in the damaged tissue, while changes in ion concentrations indicate cellular disruption and damage. Cernak et al. ( 1996) postulate that afferent neural signals from injured organs ( such as lungs) could impair central nervous system ( CNS) function and contribute to further damage over time. The biochemical mechanisms of acoustic traumas and barotraumas, as well as their acoustic thresholds, however, remain undefined.
Studies reported by several investigators ( e. g., Hastings 1990, 1995; Turnpenny et al. 1994; Caltrans 2001, 2004; Abbott and Bing- Sawyer 2002) also describe non- auditory damage to fish caused by sound including evidence of capillary rupture in the skin, neurotrauma, eye hemorrhage, swim bladder rupture, and death. Hastings showed that pond- size goldfish could not survive 2- hour continuous wave exposures at 250 Hz and a sound pressure level of 204 dB re 1 μPa ( peak), and blue gouramis could not survive 0.5- hour continuous wave exposures at 150 Hz and 198 dB re 1 μPa ( peak).
The work by Turnpenny et al. ( 1994) resulted in the investigators suggesting significant damage to caged fishes resulting from exposure to sound. They reported mortalities 24 hours post- exposure in brown trout and whiting after exposure for only five minutes to various tones at levels as low as 170 dB ( re 1 μPa) ( assumed to be rms, but not reported as such) at frequencies from 95 to 410 Hz. However, this non- peer- reviewed study has significant problems with the experimental design, acoustic environment, data analysis, and poor ( or non- existent) controls. The acoustic design of the test chamber ( a 30 cm x 30 cm x 30 cm mesh cube ensonified by four different sound projectors and located near the water surface) was such that the sound pressure level could not be controlled as indicated in the report ( see Ellison 1996). Indeed, no other studies have reported any physical damage or mortality in fishes after exposure to this low of a sound pressure level for only five minutes. Sound pressure is zero at a water surface, so it is likely that there was significant mechanical energy not taken into account in the test chamber because of severe pressure gradients that created oscillatory fluid motion.
At the same time there are very few studies known to have involved a professional fish pathologist to do full necropsy and histopathology after noise exposure ( Popper et al. in prep. a; Marty 2004). Popper et al. ( in prep. a) showed that there were virtually no effects on any body system in rainbow trout and channel catfish as a result of exposure to SURTASS LFA sonar at received levels of 193 dB re 1μPa ( rms). Marty ( 2004) examined fish following their exposure to four minutes of driving of concrete piles at the Port of Oakland. While Marty found some pathology, it did not differ between sound- exposed and control groups, suggesting that any pathology may have been from prior effects or disease. Indeed, these results highlight the necessity of involving a professional fish pathologist in studies of this type because it is exceedingly easy to impose artifacts in the tissue as a result of mishandling or poor tissue preparation.
Significantly, the studies by Hastings ( 1990, 1995), Turnpenny et al. ( 1994), Caltrans ( 2001), Abbott and Bing- Sawyer ( 2002) and Caltrans ( 2004) do not appear to have involved a
36
professional fish pathologist. In addition, Abbott and Bing- Sawyer ( 2002) sacrificed fish by putting them in plastic bags and placing them on ice, whereas Caltrans ( 2004) sacrificed fishes in other ways but then froze the animals. While differences were reported between exposed and control animals in pile driving studies ( though one study ( Caltrans 2001) showed damage in 70% of controls that were attributed to handling), it is possible that handling post stress from exposure to sound resulted in the reported pathology, rather than being caused directly by exposure to pile driving signal per se. Moreover, it is widely known that freezing damages tissues and cells and such damage could make recognizing actual effects of exposure to sound exposure difficult because it would “ mask” effects of sound exposure in both experimental and control animals. 12
1. Juvenile and Adult Fish
Key variables that appear to control the physical interaction of sound with fishes include the size of the fish relative to the wavelength of sound, mass of the fish, anatomical variation, and location of the fish in the water column relative to the sound source. Yelverton et al. ( 1975), in an important and well- conducted study reported in the gray literature, provides the most definitive study of the gross effects of sound generated by underwater blasts on fishes.
Explosive blast pressure waves consist of an extremely high peak pressure ( called overpressure) with very rapid rise times (< 1 ms). Yelverton et al. ( 1975) exposed eight different species of fish, five with ducted swim bladders ( physostomes) and three with non- ducted swim bladders ( physoclists – see glossary, page 61) to blasts. The physostomes were top minnow ( Gambusia affinis), goldfish ( Carassius auratus), carp ( Cyprinus carpio), rainbow trout ( Salmo gairdneri [ now Onchorhyncus mykiss]), and channel catfish ( Ictalurus punctatus), and the physoclists were guppy ( Lebistes reticulates), bluegill ( Lepomis macrochirus), and large mouth bass ( Micropterus salmoides). The test specimens ranged from 0.02 g ( guppy fry) to 744 g body mass ( large carp) and included small and large animals from each species. The fish were exposed to blasts having extremely high peak overpressures with varying impulse lengths. Yelverton et al. ( 1975) found a direct correlation between body mass and the magnitude of the “ impulse,” characterized by the product of peak overpressure and the time it took the overpressure to rise and fall back to zero ( units in psi- ms), which caused 50% mortality as shown in Figure 7 ( page 67). Trasky ( 1976) also reported significant differences between adult fishes, and salmon and herring fry in the lethal blast overpressure from buried seismic charges
Similar to the findings of Stuhmiller et al. ( 1996) for blast injury to lung and lethality in terrestrial animals, the results of Yelverton et al. ( 1975) indicate that a metric related to the amount of sound energy received, such as the sound exposure level or integral over time of the acoustic intensity, rather than just peak pressure correlates with swim bladder and other tissue damage as well as mortality in fish. In fact Yelverton et al. ( 1975) concluded that peak pressure alone did not correlate with damage because peak pressure was kept constant and the impulse duration was varied or vice versa in their study. The injuries they observed included swim bladder rupture, kidney damage, and liver damage. While these data are important, there is concern regarding the experimental design used by Yelverton et al. ( 1975). The primary concern
12 A much more effective, and accepted, method of preservation for necropsy is to place the tissue into buffered formalin.
37
is a lack of controls to evaluate fish handling procedures. Without these controls there is an issue in quantifying damage to fish strictly attributable to exposure to the explosions compared to damage attributable to handling.
Additional studies, primarily in the gray literature, using explosives suggest that there is far more damage to fishes with swim bladders than to species, such as flatfish, that do not have such air chambers ( e. g., Coker and Hollis 1952; Gaspin 1975; Baxter et al. 1982; Goertner 1994). It has also been shown that the effects on fish decline rapidly with distance from the explosion ( e. g., Houghton and Mundy 1987; Goertner 1994) as the peak overpressure decreases and the impulse duration increases. Similarly, a study by Kearns and Boyd ( 1965) suggested that the extent of fish kill decreases with increasing distance of the fish from an air gun source, and another unpublished study indicated no mortality from seismic air gun shots at considerable distance ( 4000 m) from the source ( Thomsen 2002). There is evidence that the effects of explosions vary by species, even when all test fish have a swim bladder ( Govoni et al. 2003). Based on these and other studies ( e. g., Yelverton et al. 1975), it is clear that there is considerable variability in the effects of explosive blasts on fishes, and that the variables include received sound energy, presence or absence of gas bubbles ( e. g., swim bladder), mass of fish and perhaps body shape ( e. g., Teleki and Chamberlin 1978), and biomechanical properties of the swim bladder wall.
2. Eggs and Larvae
In considering fishes, it is important to not only think in terms of adults, but also in terms of fish eggs and larvae. Whereas it is possible that some ( though not all) species of fish would swim away from a sound source, thereby decreasing exposure to sound, larvae and eggs are often at the mercy of currents and move very slowly, if at all. Eggs are often stationary and thus could be exposed to extensive human- generated sound if it is presented in the area, including sound transmitted through water ( i. e., eggs within the water column) or substrate ( e. g., eggs deposited within substrate, such as salmonid redds).
Data on effects of sound on developing eggs and larvae are very