NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

National Research Council (US) Committee on Potential Impacts of Ambient Noise in the Ocean on Marine Mammals. Ocean Noise and Marine Mammals. Washington (DC): National Academies Press (US); 2003.

Cover of Ocean Noise and Marine Mammals

Ocean Noise and Marine Mammals.

Show details


The environment, whether in sea or on land, is filled with natural sounds, although increasingly many locales have sound contributed by anthropogenic sources as well. The extent to which sound in the sea impacts and affects marine life is a topic of considerable current interest both to the scientific community and to the general public. Scientific interest arises from a desire to understand more about the role of sound production and reception in the behavior, physiology, and ecology of marine organisms. Anthropogenic sound, including sound necessary to study the marine environment, can interfere with the natural use of sound by marine organisms. Public interest arises primarily from the potential effects of anthropogenic sound on marine mammals, given the broad recognition of the importance of sound in the lives of marine mammals.

For acoustical oceanographers, marine seismologists, and minerals explorers, sound is the most powerful remote-sensing tool available to determine the geological structure of the seabed and to discover oil and gas reserves deep below the seafloor. Society as a whole has reaped substantial intellectual and practical benefits from these activities, including bottom-mapping sonars and technology leading to the discovery of substantial offshore oil reserves.

Scientists and the public are also acutely aware that sound is a primary means by which many marine organisms learn about their environment and that sound is also the primary means of communicating, navigating, and foraging for many species of marine mammals and fish. Indeed, the study of sounds of marine organisms provides insight into important aspects of their biology.

The public's interest in the impact of human-generated ocean noise on marine animals has greatly increased. Concerns include whether human-generated sounds may interfere with the normal use of sound by the marine animals or whether the human-generated sounds may cause the animals physical harm. At issue is whether the human-generated sounds affect the ability of marine animals to pursue their normal activities and the long-term ability of these animals to survive, reproduce, and maintain healthy populations.

It is also critical to note that sound is an essential tool for ensuring national security. The development of underwater sound as a method for detecting submarines began during World War I and accelerated rapidly during World War II. During the Cold War, acoustic antisubmarine warfare became the principal deterrent against missile-carrying submarines roaming the high seas. Since the end of the Cold War ocean acoustics has continued to retain its military significance, but now militaries seek to expose submarine and submerged mine threats in shallow-water areas.

It is in this context of parallel developments and applications in ocean acoustics, marine seismology, oil exploration, and animal bioacoustics that concerns about the effects of sound on marine life have emerged. While researchers had been aware for quite some time of the sounds produced by marine life, it was not until the Acoustic Thermometry of Ocean Climate (ATOC) project (Baggeroer and Munk, 1992), in which high-intensity, low-frequency (defined for this report as sounds below 1,000 Hz) sounds were transmitted over long distances, that the public's attention focused on the possible impacts of human-generated noise on marine mammals, although noise with potential impacts had been regulated since the passage of the Marine Mammal Protection Act in 1972. Suddenly, it seemed, nearly all sources of anthropogenic sound came under intense scrutiny as potential threats to the existence and well-being of undersea life. These have included not only the aforementioned oceanographic, naval, and seismic surveying tools but also additional sources of unintentionally generated noise, such as commercial shipping, offshore construction, and recreational boating. As a result, research support for marine mammal bioacoustics, principally from the Office of Naval Research (ONR; Gisiner, 1998), grew substantially, and the permitting process necessary for conducting ocean acoustics experiments that allow incidental takes, administered by the National Oceanic and Atmospheric Administration (NOAA) and the U.S. Fish and Wildlife Service, received increased scrutiny. Two National Research Council (NRC) panels (NRC, 1994, 2000) were convened especially to address those issues associated with low-frequency sound, with particular attention paid to the ATOC project (NRC, 2000). The current NRC committee, which is responsible for generating this report, was convened at the request of the interagency National Ocean Partnership Program, with support from ONR, the National Science Foundation, NOAA, and the U.S. Geological Survey. It was requested in the context of growing concern over noise in the ocean [Natural Resources Defense Council (NRDC), 1999] and with the recognition that there was a need to focus on a broader range of issues than those associated with the ATOC project.

Although the thrust of this study and those that have preceded it (NRC, 1994, 2000) is the impact of anthropogenic sounds, it must be realized that sound in the sea is produced by a large and extraordinarily diverse number of naturally occurring nonbiological and biological sources. Natural nonbiological sounds are as diverse as the wind and waves, rockslides, geologic events, thunderstorms, and water moving over a coral reef. Many of these sources of sound have existed since the formation of the earth and oceans, and it is highly likely that these sounds have had some impact on the evolution of the auditory system, animal communication, and ecology (Fay and Popper, 2000). Biologic sounds are equally diverse and are emitted intentionally or unintentionally by numerous organisms. Unintentional sounds include, for example, those produced by schools of fish swimming through the ocean or release of air by large groups of fish as they adjust their buoyancy (Moulton, 1960, 1963). Intentional sounds, including whale songs, dolphin clicks, and fish vocalizations, are believed to be produced in various species for communication, echolocation, and perhaps even acoustic “imaging” of the environment to assess the physical characteristics of their habitat.

Sound detection by vertebrates clearly arose in the aquatic environment (Fay and Popper, 2000). The earliest known vertebrate fossils had ears (Jarvick, 1980), although there is no way of knowing if these ears functioned for sound detection or only served for detection of head motion and balance. Ears and functioning auditory systems are found in all aquatic vertebrates.1 Auditory capabilities of bony fish are reasonably sophisticated, and a number of species not only detect sounds but can also determine sound source direction, detect signals in the presence of noise sources (maskers), and discriminate between sounds (e.g., Popper and Fay, 1999; Fay and Popper, 2000). Moreover, there is considerable similarity in the structure of the ear in aquatic and terrestrial vertebrates, and it is clear that the basic structure of the ear, including the sensory hair cell that converts sound to signals in the nervous system in all vertebrates, evolved very early in vertebrate history (see Popper and Fay, 1997; Fay and Popper, 2000).

The questions then to ask are why hearing evolved and why one would expect hearing to be particularly sophisticated in marine animals.2 The aquatic environment has limited or no light, and even in areas where there is considerable light, the range of visibility is rather limited as a result of the attenuation characteristics of light in water. As a consequence, if early aquatic animals had only visual systems, the range of information about the environment around them would have been constrained by their field of vision. With the evolution of the auditory system, the sensory world of the organism expands to greater distances and the animal develops an acoustic image of the world around it, just as humans sense the world around them using sound, even when vision is not available. The evolution of an auditory system that can discriminate among sounds, determine the direction of a sound source, and detect sounds even when the environment is reasonably noisy greatly increased the survival potential of aquatic animals. It has been argued that humans and animals glean a great deal about their environment from the “acoustic scene” and that this scene provides an immense amount of subtle information (see Bregman, 1990; Fay and Popper, 2000). Indeed, Bregman's ideas can be extended to argue that the most important aspect of hearing is not communication per se but learning about the acoustic scene in order to detect objects and organisms in the environment and the ability to discriminate between sounds and the location of different sounds, a process called “stream segregation” (Bregman, 1990; Fay and Popper, 2000).

In essence, sound and sound detection would seem to be critical parts of the lives of marine mammals and fish.3 Many of these animals use sound for communication between members of their species. But equally important is the idea that probably all of these species use sound to learn about their environment and to survive. Therefore, there should be concern not only about the impact of anthropogenic sounds on communication but also about the impact on general determination of information in the environment.

A fundamental question is whether the impact of anthropogenic sounds on marine mammals and the marine ecosystem is sufficiently great to warrant concern by both the scientific community and the public. As discussed in detail in this report, the data currently available suggest that such interest is indeed justified. However, as will also be shown, the data are still quite limited, and it will be important to develop a research program that will provide substantially more data on this topic. Only when these data are available will it be possible to draw concrete conclusions regarding this question. The statement of task and the committee's response provide the framework for obtaining these data.


This study will evaluate the human and natural contributions to marine ambient noise and describe the long-term trends in ambient noise levels, especially from human activities. The report will outline the research needed to evaluate the impacts of ambient noise from various sources (natural, commercial, naval, and acoustic-based ocean research) on marine mammal species, especially in biologically sensitive areas. The study will review and identify gaps in existing marine noise databases and recommend research needed to develop a model of ocean noise that incorporates temporal, spatial, and frequency-dependent variables.

In its interpretation of the statement of task, the committee felt that there were several key guidelines that should be followed and several key questions that must be addressed. First, to researchers the term “ambient noise” typically refers to the overall background noise caused by all sources such that the contribution from a specific source is not identifiable. For example, considering only shipping noise in this context, Cato (2001) states that “traffic [shipping] noise is the low-frequency general background noise resulting from contributions from many ships over an ocean basin, but in which the contribution of no individual ship is distinguishable.” However, the committee felt that this conventional definition was too restrictive and that sound caused by identifiable, often transient, typically nearby sources should be included in its considerations as well. The term “ocean noise” was therefore defined by the committee as encompassing not only the usual background ambient noise but also the noise from distinguishable sources (Box 1-1). Second, the committee agreed that, although its work would concentrate primarily on the effects of noise on marine mammals, it should consider other species as well (e.g., fish) that are part of the ecosystem and food web on which marine mammals depend. Third, the frequency band to be studied was determined to range from 1 to 200,000 Hz (200 kHz), since this is the entire bandwidth that various marine organisms are capable of detecting.

Box Icon

Box 1-1

Sources of man-made noise in the ocean. TRANSPORTATION Aircraft (fixed-wing and helicopters)

Five key questions were considered to be essential to achieving the goals described in the statement of task:

1. What is the noise budget in the ocean?

It is well known that noise in the ocean arises from a variety of sources, including ships, breaking waves, and living organisms. Far less is known about the relative contributions of each of these sources (referred to, in this report, as the noise budget) to the total noise field in various parts of the world's oceans, including seasonal differences, or about the more detailed spatial and temporal variability of the noise field. Furthermore, within a particular source category (e.g., ships, seismic surveys) the contribution from subsets should be understood. For example, within the major category of ships the contribution from different types of vessels has not been quantified.

2. What are the long-term trends in noise levels?

It is clear that prior to the Industrial Revolution (ca. 1850), the contribution of anthropogenic activity to the noise budget was negligible and that ocean noise levels were determined by naturally occurring sources (e.g., wind, waves, earthquakes, organisms). Little is known about the changes of these levels with time as a result of the increased maritime activity associated with the onset of industrialization. To what extent has this trend been influenced by factors such as the number of ships, their size, and propulsion? In more recent years, changes in the noise budget would also have to take into consideration other sources of anthropogenic sounds discussed in this report.

In order to understand long-term changes in the noise budget caused by human activity, a baseline can be obtained from noise measurements in areas with few human-generated contributions, for example, several places in the southern hemisphere far removed from shipping lanes and where low-frequency sound from long range is blocked by bathymetry.

3. Are existing models of ocean noise still valid?

Probably the most widely used models of the ambient component of ocean noise continue to be the curves developed by Wenz (1962; see also Richardson et al., 1995). These provide a summary of average ambient noise spectra from various sources, as shown in Plate 1. But according to Ross (1993), “they are not particularly useful in predicting or explaining ambient noise measured in a particular location at a particular time.” Furthermore, considerable additional noise data have been acquired and theoretical developments have occurred during the past 40 years (Gisiner, 1998), so that updated and improved versions of the Wenz curves could be developed. What are the effects of specific properties of noise sources, including rise times, tonal content, bandwidth, and power levels?

4. What are the effects of transient and long-term noise exposure on marine mammals and the ecosystems on which they depend?

Specific conclusions on the effects of noise-induced hearing loss on terrestrial mammals have been drawn. Recent experiments have shown that (1) the noise need not be painful to cause permanent loss; (2) the damage is approximately proportional to noise energy integrated over time; (3) high-frequency noise is more dangerous than low-frequency noise; (4) narrowband noise is more dangerous than broadband noise; and (5) there is large intersubject variability in the resistance to noise, even among genetically identical animals (Liberman, 2001). Comparable data are not available for marine mammals, although it is clear that such data are needed in order to understand the impact of anthropogenic sound on these organisms.

Despite the lack of data for marine mammals, some general comments can be made about the impact of noise on aquatic organisms by introducing the concept of zone of influence (Richardson et al., 1995; Gisiner, 1998; NRDC, 1999). Essentially, the effect of noise on the animal depends to a large degree on the proximity of the animal to the noise source and the received level of the signal by the animal. At very short ranges (that have yet to be determined), a sufficiently loud source may cause severe physiological damage and perhaps death. At greater ranges, geometrical spreading and absorption reduce the signal level substantially and the same source may cause hearing loss and short-term behavioral changes, which can contribute to death under particular circumstances (Evans and England, 2001). A quantitative evaluation of the radii of these zones for different species as well as an understanding of effects analogous to those described for terrestrial mammals have yet to be determined.

It should also be noted that marine mammals are part of a larger ecosystem upon which they depend. Included in this ecosystem are other organisms, particularly fish and possibly marine reptiles and invertebrates, which use sound in their normal behavior and that may also be impacted by anthropogenic sounds. Thus, in addition to understanding the direct impact of such sounds on marine mammals, it is important to understand the impact of these sounds on fish and other organisms.

5. What are recommendations for future research?

None of the four preceding questions currently has a concrete and final answer. It is therefore crucial that specific areas for future research, leading to more conclusive answers, be identified. Research recommendations from previous NRC studies (1994, 2001) are included in Appendix D and should be reviewed and considered with those presented here. Progress has been made in many of the areas described in the previous reports, but much more must be accomplished to improve our ability to predict and assess the impact of ocean noise on marine mammals.


The following section presents the sonar equation and discusses its application to biological receivers. This section is not intended to be a thorough review of this topic but, rather, to introduce many of the terms and ideas that will be addressed throughout the remainder of this report. Additional terms along with measures of the properties of acoustic sources and acoustic fields are discussed in the Glossary. For additional study of the fundamentals of ocean acoustics and biosonar, interested readers can refer to one of several textbooks on these topics (e.g., Busnel, 1963; Urick, 1975; Tolstoy and Clay, 1987; Brekhovskikh and Lysanov, 1991; Burdic, 1991; Au, 1993; Frisk, 1994; Jensen et al., 1994; Richardson et al., 1995; Medwin and Clay, 1998).

The quantitative description of the acoustic pressure wave to which an animal is exposed is obtained through the use of the sonar equation (Urick, 1975; Jensen et al., 1994). Specifically, the received acoustic level (RL) from a source with source level (SL) is given by

RL = SL TL + AG, (1-1)

where TL is the transmission loss from source to receiver and AG is the processing gain associated with the animal's reception system. All of the components of the sonar equation are expressed in decibels (dB), which are proportional to the logarithms of the corresponding linear values. The decibel is used largely for convenience, since the individual components of the equation may span a broad dynamic range, and furthermore, the logarithmic operation expresses multiplicative processes in terms of seemingly simpler additive operations. In addition, a logarithmic scale is typically used for sound levels because human perception of loudness increases logarithmically. Specifically, the decibel is inherently a relative quantity, that is

Image p20007774g20001.jpg


where the reference pressure level used in underwater acoustics is 1 µPa (see Glossary for further explanation). The SL is defined as the pressure at a unit distance, typically 1 m, from the source, while the TL describes all of the geometrical spreading and attenuating effects of the medium associated with propagation, scattering, and absorption as the signal travels from a position 1 m from the source to the location of the animal. The AG represents the enhancement of the received signal that can occur through the application of signal-processing techniques and perhaps multiple sensors in the receiving system. Combining all of these terms, the ability of the animal to detect the signal can be interpreted in terms of the animal's hearing sensitivity, that is, the minimum detectable value of RL, which expresses its minimum threshold4 hearing level as a function of frequency (Figure 1-1) (Wartzok and Ketten, 1999; Finneran et al., 2002). In the ocean environment, an additional term must be introduced into the sonar equation, namely an ocean noise term (NL), which is defined with respect to the same reference pressure and frequency bandwidth as SL and RL. The actual excess signal level (SE) available to allow detection and interpretation of the signal is given by

FIGURE 1-1. Audiograms for individual land mammals, cetaceans, and odontocetes.


Audiograms for individual land mammals, cetaceans, and odontocetes. Underwater audiograms for (A) odontocetes and (B) pinnipeds. More than one curve is shown for some species because data reported in different studies were not consistent. Note that for (more...)

SE = RL NL = SL TL + AG NL. (1-3)

The animal will be able to hear and respond to a signal of a particular frequency only if SE is greater than zero. An interesting observation is that the superposition of odontocete hearing sensitivity (the audiogram) on the Wenz curves (Plate 2) indicates that the hearing thresholds of these animals correspond to the quiet ocean ambient noise spectral levels over the animals' frequency bands of hearing sensitivity. In other words, in the absence of human noise, the ocean is very quiet for them; they seem to have adapted to the natural noise that surrounds them.

Transmission loss in Equations 1-1 and 1-3 is a complicated function of the source and receiver geometry, frequency, and environmental parameters of the water column and the seabed (Brekhovskikh and Lysanov, 1991; Frisk, 1994; Jensen et al., 1994). In general, transmission loss with increasing source-receiver range is dominated by two important effects. First, the sound speed in the sea is not constant but varies with both depth and range, immediately altering the simple spherical spreading loss associated with a point source in free space. Sound waves interact with both the moving sea surface and the seabed, which is a complicated multilayered structure that supports acoustic waves. All of these factors combine to create a channel, or waveguide, for the sound waves that are trapped between the surface and the bottom in shallow water or focused by the sound speed structure in deep water as they propagate outward from source to receiver. This channeling effect causes the envelope of the signal to spread cylindrically, rather than spherically, outward at ranges much greater than the waveguide thickness, D (which equals the water depth in shallow water environments). Second, the intrinsic absorption properties of seawater cause the sound wave to be further attenuated by heat, viscous, and molecular relaxation losses (Medwin and Clay, 1998). As a result, the transmission loss can be expressed generally as:

TL (dB re 1 m) = 20log10 r + αr, when r < D (1-4)

TL (dB re 1 m) = 10log10 r + 10log10 D + αr − 3, when r > D, (1-5)

where r is the horizontal range between source and receiver (in m), and the absorption coefficient α (in dB/m) is approximately proportional to the square of the frequency (Figure 1-2; Frisk, 1994) with the impact of absorption shown for an idealized case. Equations 1-4 and 1-5 are valid only for omnidirectional, single-point sources; the geometrical spreading for other types of sources (e.g., line sources such as vertical source arrays) may be significantly different.

FIGURE 1-2. Ideal transmission loss.


Ideal transmission loss. Transmission loss in an ideal 5,000-m-deep ocean with perfectly reflecting surface and bottom. This chapter details the calculation for transmission loss. The differences between the curves for 100 Hz and 1,000 Hz are due to frequency-dependent (more...)

Waveguide effects are important in determining the distance traveled and the character of acoustic energy as it propagates through the ocean. The key factor that influences the character of the propagation in deep water is the variation with depth z of the sound velocity profile c(z). Amazingly, the small relative variations in sound speed, which are typically less than 4 percent, have a profound influence on the structure of the sound field. Ducting by the sound speed structure dominates over any interactions with the boundaries in sound propagating from a deep source (about 1,000 m) in the classical SOFAR (sound fixing and ranging) channel found, for example, in the North Atlantic Ocean. The complexities of sound propagation in the sea must be carefully and accurately taken into account when evaluating the contribution of a particular sound source to the overall ocean noise field and are presented in more detail in Chapter 4.

In coastal regions and coral reefs where water depth is very shallow compared to that of the deep ocean, propagation of sound is more complex (Frisk, 1994). In these areas sound propagates over distances greater than a few water depths only by repeatedly interacting with the surface and bottom. At both the surface and bottom, a sound wave reflects back onto itself, and these reflections interfere with the original wave to produce an interference pattern in the water column. A sound source transmitting at a single frequency will produce a discrete number of vertical interference patterns, each with a different number of maximum and minimum pressures from top to bottom (Ferris, 1972). Each vertical interference pattern, or standing wave in the vertical direction, propagates in the horizontal direction at its own speed. However, if the frequency of a standing wave is too low, it will not propagate. This lower frequency limit is called the cutoff frequency, and standing waves with frequencies below the cutoff cannot propagate in the horizontal direction. Therefore, at a given water depth, an absolute cutoff frequency exists that is equal to the cutoff frequency for the vertical interference pattern having the fewest number of maximum and minimum pressures in the vertical (Rogers and Cox, 1988). A simple mathematical model of the shallow water environment can be devised by assuming it consists of a homogeneous ocean overlying a fluid-like, homogeneous bottom. For this model the absolute cutoff frequency (in Hz) below which no sound can propagate in shallow water, is given by

Image p20007774g24001.jpg


where cw is the speed of sound in water, cs is the speed of sound in the bottom sediment, and h is the water depth in meters. Real ocean bottoms are much more complicated than the simple homogeneous model described, and the bottom can become part of the medium in which the sound propagates (Figure 1-3). The propagation efficiency of the seabed, however, is far less than that of the water column because the intrinsic absorption of the bottom is typically about 1,000 times that in seawater. Because of variations in water depth and in ocean bottom properties (as well as variations in the sources of noise themselves), ocean noise in shallow water can be highly variable from one location to another (Urick, 1984; Zakarauskas, 1986).

FIGURE 1-3. Cutoff frequencies estimated for propagation of sound in shallow water environments composed of a homogeneous ocean overlying a fluid-like, homogenous bottom.


Cutoff frequencies estimated for propagation of sound in shallow water environments composed of a homogeneous ocean overlying a fluid-like, homogenous bottom. Sound at frequencies below the cutoff frequency (indicated by the shaded regions) will not propagate (more...)

In many cases of waveguide propagation in the ocean, the upper boundary of the waveguide is formed by reflection from the underside of the ocean surface. Therefore, the sea surface plays a fundamental role in acoustic propagation. Interaction of sound with the ocean surface also is important from a biological perspective, since marine mammals must come to the surface to breathe. The sea surface under calm conditions is a nearly perfect reflector of ocean-borne sound at all incident angles over a wide band of frequencies.5 Because the overlying mass of air provides very little resistance to particle motion (its acoustic “impedance” is small compared to that of seawater), the sea surface yields completely to the incoming underwater sound field. At this interface the ocean acoustic particle motion in the vertical direction is maximum and the acoustic pressure becomes zero, known as pressure release. Actual open-ocean surface conditions are complicated by factors such as the presence of near-surface bubbles and moving, wind-generated roughness. Animals that sense acoustic pressure can reduce their received sound levels by going to the ocean surface. As a result, comparisons of the density of marine mammals near sound sources and in other locations where the underwater sound levels are high may be biased by the animals moving close to the surface in the presence of the sound in order to reduce the received sound pressure level.


This report describes sound origins, trends, effects on marine mammals, and current modeling efforts (Figure 1-4). Chapter 2 provides descriptions of the natural and human sources of ambient noise in the ocean and the possible reasons and evidence for long-term trends in ocean noise. Chapter 3 describes what is known about the impacts of marine noise on mammals, including masking, sensitization, and habituation. Chapter 4 summarizes existing modeling efforts and ocean noise databases, particularly those that integrate the known information about noise with behavioral databases on marine life. Chapter 5 synthesizes findings and recommendations of the committee for future research.

FIGURE 1-4. Components necessary to understand the effects of ocean noise on marine mammal behavior.


Components necessary to understand the effects of ocean noise on marine mammal behavior.



The only exception may be the jawless fish, lampreys and hagfish, where there is a functioning ear but no evidence to indicate whether they can or cannot detect sound. In these species, the ear may strictly serve as an organ of balance.


How the ear evolved is another issue of considerable interest, but one that will not be considered here. Readers are referred to van Bergeijk (1967), Baird (1974), Ridgway et al. (1974), and Fay and Popper (2000) for useful discussions of this issue.


It should be noted that there have been very few studies on sound detection by marine invertebrates and so we do not yet know if any of these species detect sound.


It should be noted that the concept of the threshold is a statistical one and represents the minimal detectable level for an organism in some percent of trials—often 50 or 75 percent of trials. The threshold for an individual animal may change by a few decibels, even within the course of a testing session, and the threshold at any given moment may depend on motivational level and distractions in the environment (Holt et al., 2002).


Although underwater sound incident on the underside of a flat ocean surface is perfectly reflected for all intents and purposes, airborne sound that is nearly vertically incident on the sea surface can couple into ocean-borne sound, as discussed in Chapter 2.

Image p20007774g195001
Image p20007774g196001
Copyright 2003 by the National Academy of Sciences. All rights reserved.
Bookshelf ID: NBK221259


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...