The transition from wind-driven to mechanized shipping became the first step in what was to be a continued increase in the introduction of sound into the oceans. The oceans are much less transparent to light than to sound; as a result, many marine species use sound rather than light to navigate and communicate. Over the last 40 million years, marine mammals have evolved specializations for using underwater sound. The initial introduction of the propulsion sound of ships was unintentional, but engineers and scientists have also learned, with the development of sonar, how to use sound intentionally for underwater communication, navigation, and research. At some point as humans introduce more sound into the oceans, the conflict with evolutionarily-adapted marine mammal sound-sensing systems seems inevitable. Attention has been drawn to this issue through a series of marine mammal strandings, lawsuits, legislative hearings, and National Research Council (NRC) reports (1993, 2000, and 2003b) and, most recently, the draft report of the US Commission on Ocean Policy (2004).
Two earlier National Research Council reports (1994, 2000), while addressing biological issues of marine mammals and noise, also made recommendations that affected federal legislation and its implementation. The first was issued in 1994 in response to the feasibility test of a proposal to track global warming by monitoring the speed of an acoustic signal across an ocean basin (Munk et al., 1994). The feasibility test was to have set the stage for the full Acoustic Thermometry of the Ocean Climate (ATOC)
experiment, but because of concerns over possible effects on marine mammals only a limited deployment of ATOC was attempted. The 1994 report recommended that there be legislative distinction between different types of “taking” and that the regulatory agencies streamline the permitting process for activities that did not kill or capture marine mammals. Additional streamlining was recommended for nonlethal activities that have negligible effects. The 2000 National Research Council report reviewed the marine mammal research program that was a component of the limited ATOC deployment. In Marine Mammals and Low-frequency Sound: Progress Since 1994, the committee noted that the 1994 amendments to the Marine Mammal Protection Act (MMPA) addressed some of the issues raised in the 1994 NRC report. The 1994 amendments introduced two levels of takes by harassment under the MMPA—level A and level B harassment. Level A harassment was defined in the 1994 amendments as “any act of pursuit, torment, or annoyance which has the potential to injure a marine mammal or marine mammal stock in the wild.” Level B harassment was defined as “any act of pursuit, torment, or annoyance which has the potential to disturb a marine mammal or marine mammal stock in the wild by causing disruption of behavioral patterns including, but not limited to, migration, breathing, nursing, breeding, feeding, or sheltering.” However, the 2000 National Research Council report emphasized the importance of a criterion for significance of disruption of behavior (pg. 68):
It does not make sense to regulate minor changes in behavior having no adverse impact; rather, regulations must focus on significant disruption of behaviors critical to survival and reproduction.
The report (pg. 69) recommended redefining level B harassment as any act that
has the potential to disturb a marine mammal or marine mammal stock in the wild by causing meaningful disruption of biologically significant activities, including but not limited to, migration, breeding, care of young, predator avoidance or defense, and feeding.
Since the report was issued, the term biologically significant has been used in discussions of the 2003-2004 reauthorization of the MMPA (House Report 108-464). The US National Marine Fisheries Service (now National Oceanic and Atmospheric Administration [NOAA] Fisheries) has also used the term in decisions to grant incidental harassment authorizations. Scientific investigation and description of what would constitute “biologically significant” have not been pursued in a comprehensive manner.
The charge to the present National Research Council committee (Box ES-1) was to clarify the term biologically significant. In the broadest sense, it is a straightforward charge. An action or activity becomes biologically significant to an individual animal when it affects the ability of the animal to grow, survive, and reproduce. Those are the effects on individuals that can have population-level consequences and affect the viability of the species. However, those effects are separated in time and usually in space from the precipitating event. What can be observed, with difficulty in the case of marine mammals, are the direct behavioral and in some cases physiological responses of individual animals.
It was recognized that the definition of level B harassment proposed in the 2000 report introduced two kinds of biological significance: one with respect to animal activities, stated directly, and the other implied in the “meaningful disturbance” of those activities. On reflection, it became clear that wild animals rarely engage in activities that are not biologically significant (even play is not frivolous [Bekoff and Byers, 1998]), so the primary
In its 2000 report, Marine Mammals and Low-frequency Sound, the National Research Council recommended that the Marine Mammal Protection Act definition of “Level B harassment should be limited to meaningful disruption of biologically significant activities that could affect demographically important variables such as reproduction and longevity.” Recognizing that the term “biologically significant” is increasingly used in resource management and conservation plans, this study will further describe the scientific basis of the term in the context of marine mammal conservation and management related to ocean noise. Based on input from a scientific workshop, consideration of the relevant literature, and other sources, the committee will produce a brief report that reviews and characterizes the current scientific understanding of when animal behavior modifications induced by transient and non-transient ocean acoustic sources, individually or cumulatively, affect individuals in ways that have negative consequences for populations.
concern should be with determining when human activity elicits behavioral or physiological responses in marine mammals that rise to the level of biological significance.
Changes in behavior that lead to alterations in foraging efficiency, habitat abandonment, declines in reproduction, increases in infant mortality, and so on are difficult to demonstrate in terrestrial animals, including humans, and are much more difficult to demonstrate in animals that may only rarely be observed in their natural environment.
A CONCEPTUAL MODEL TO ADDRESS POPULATION CONSEQUENCES OF ACOUSTIC DISTURBANCE
A conceptual model is proposed that identifies the different stages required to move from marine mammal behavior to a determination of population effects of behavioral change. The model first characterizes an acoustic signal, the resulting behavioral change, and a determination of the “life function” or activity affected. It then describes the resulting change in vital rate, such as life span, and finally suggests population effects—effects on following generations. “Transfer functions” connect the variables. A transfer function is essentially a relationship that allows one to estimate, for example, how a change in migration route leads to a reduction in reproductive success. It was quickly recognized that the high-priority research identified in the earlier National Research Council reports (1994, 2000, 2003b) is essential for building the first stages of the model.
RECOMMENDATION 1: The high-priority research identified by the National Research Council (1994, 2000, 2003b) should be completed. That research is essential for the model proposed in this report.
Through discussions before and during the public workshop held at the National Academies in March 2004, a consensus was reached that the proposed conceptual model includes the components needed to develop a predictive model to determine the biological significance of behavioral change. However, there was also a consensus that we are a decade or more away from having the data and understanding of the transfer functions needed to turn such a conceptual model into a functional, implementable tool.
RECOMMENDATION 2: A conceptual model, such as that described in this report, should be developed more fully to help to assess impacts of acoustic disturbance on marine mammal populations. Development of such a model will allow sensitivity analysis that can be used to focus, stimulate, and direct research on appropriate transfer functions.
To enhance such a model and progress toward determining population effects of acoustic disturbance, all available sources of data on marine mammal behavior and reactions to noise will need to be accessed. In addition to results of normal scientific studies, a veritable wealth of data on marine mammals is collected in compliance with federal regulatory requirements, but those data are not being accessed or used beyond the original intent of their collection (such as for permit issuance). A data-coordination effort could provide substantial benefits and improve our knowledge of marine mammal distribution, critical habitats, behaviors, population estimates, and other items essential for the modeling effort. Although data coordination would be difficult to implement, over the long term the value added by improving the organization and accessibility of data collected for these purposes would provide an efficient means of extracting invaluable information, at relatively small additional cost, for improving understanding and management. Such leveraging of diverse data collection efforts would represent an efficient use of resources and public funding. For example, the UK Joint Natural Conservation Committee has summarized sighting data from commercial seismic surveys, which help in evaluating avoidance responses (Stone, 2001, 2003).
RECOMMENDATION 3: To assist in the development of the conceptual model, a centralized database of marine mammal sightings and their responses to anthropogenic sound in the ocean should be developed and should include
Published peer-reviewed papers in the scientific literature.
Government technical reports.
Data submitted to NOAA Fisheries and the US Fish and Wildlife Service in permit applications.
Data submitted by industry to the Minerals Management Service for regulating off-shore hydrocarbon exploration and production.
All relevant data accumulated by all federal agencies in the course of their research and operational activities, including monitoring.
To facilitate the integration of data from the various sources, federal agencies need to develop standardized data-reporting formats. Survey data should include locations where marine mammals were detected and the track lines when personnel were monitoring for marine mammals, regardless of whether any were sighted. All data entered into such an integrated database must be coded as to quality, and peer-reviewed data and interpretations should be rated highest.
The biological significance of the behavioral response of an animal to an acoustic stimulus is modulated by many seasonal and environmental factors. For example, the lengthening of a foraging trip from a rookery that would be of no particular significance during a normal year could rise to the level of biological significance during an El Niño year. Allostasis, the maintenance of an animal’s physiological stability in spite of change, is a useful way to conceptualize the integration of short-term and cumulative stress and thereby to determine the possible additional effects of anthropogenic noise on marine mammals. Although data for marine mammals are lacking, serum hormone concentrations have been shown to be good measures of stress in terrestrial animals. For animals in which blood sampling is impractical, fecal sampling has been used successfully and is now being applied to some marine mammals. Preliminary studies measuring glucocorticoids in hair samples and enhanced synthesis of RNA coding for stress-induced proteins in skin samples merit further development. Measures of stress may provide critical information on marine mammal physiological status and change in response to disturbance by acoustic and other stimuli.
Correlational observations of behavioral responses to noise and other stressors have indicated general trends in such responses and in some cases have highlighted subjects of concern. To calibrate an animal’s response to a stimulus as required for a predictive model, correlational observations must be replaced with controlled-exposure, dose-response experiments. Such an approach allows researchers to go beyond observational study and determine statistically the likelihood of a particular reaction to a given acoustic stimulus. In marine mammals, such experiments are only beginning to be
conducted. There is a potential for collecting both behavioral and physiological data during controlled-exposure experiments. The resulting data will be essential for integrating behavior and physiology in models of the population consequences of acoustic disturbance.
Additional development of data-logging technology is necessary for support of controlled-exposure experiments. Data-logging packages should be modified to incorporate blood sampling during controlled-exposure experiments. Initial studies on Weddell seals (Leptonychotes weddellii) would be particularly useful in as much as research on their blood chemistry during free dives has already been conducted (Hill, 1986). Eventually the packages would benefit from new less-invasive methods for collecting blood and conducting “on-board” blood-chemistry analysis to record responses of animals in situations less unusual than that of the Weddell seal—a situation in which the animal can be handled before and after tagging.
RECOMMENDATION 4: The use of glucocorticoid and other serum hormone concentrations to assess stress should be developed, validated, and calibrated for various marine mammal species and age-sex classes and conditions. Dose-Response curves for those indicators as a function of sound characteristics need to be established. Development of a sampling package that could take blood samples on a controlled basis and stabilize hormones for later analysis or process samples “on-board” for corticosteroids at various stages of a CEE would be invaluable for determining the stress that the sound is producing. The use of fecal sampling to measure condition or stress needs to be investigated further and developed. Research efforts should seek to determine whether reliable long-term stress indicators exist and, if so, whether they can be used to differentiate between noise-induced stress and other sources of stress in representative marine mammal species (this recommendation was also made in NRC, 2003b).
Although the full predictive model of the path from acoustic stimulus to population effect is unattainable in the near term, various modeling techniques can enhance our understanding of the components of the larger model. One approach involves demographic models in which age- or stage-specific developmental, behavioral, or physiological characteristics of individuals are used to explore changes in population dynamics (Caswell,
2001). Another approach involves individual-based models that can be used to infer population responses by tracing the life history of individuals. For a number of nonmarine mammal species, individual-based models include physiology and behavior; such models have provided insight into how ecological change and human disturbance have altered demographic variables. Although a thorough, detailed model is not now possible for any marine mammal species, this approach can be used to provide preliminary understanding and to identify the most crucial gaps in available data.
Qualitative or categorical modeling that characterizes the strength of links between stimulus and response, response and function, and function and demography on a simple low-medium-high basis can be useful. A focused effort is needed on a modeling exercise that should include quantitative demographic models, individual-based models, and qualitative categorical models. Such an effort should start with, and be calibrated against, expert opinion. The effort should
Probe how successfully current knowledge could be applied.
Identify crucial gaps in our knowledge.
Encourage and provide structure for interdisciplinary synthesis.
Require that all modeling efforts be explicit about uncertainty and the consequences of uncertainty.
Require that all models clearly state their limited purpose and evaluate both their strengths and their shortcomings.
Assess the risk for the species being modeled if the model is to be used for management decisions.
Exploratory models could help to bridge the gap between changes in the physiology and behavior of individuals in response to sound and demographic effects at the population level. Demographic models might be used in an exploratory way to help to bound the problem and establish thresholds for different species. Individual-based models may provide a method for exploring the consequences of changes in individual behavior and social interactions. Those modeling approaches could be used, individually or in combination, to provide greater understanding of the problem, look for important thresholds, speculate on the likely outcome of hypothesized changes, and develop a conceptual framework for formulating management guidelines.
RECOMMENDATION 5: Several marine mammal species for which there are good long-term demographic and behavioral data on individuals should be selected as targets of an intensive exploratory modeling effort that would develop a series of individual-based models and stage- or age-structured demographic models for the species as appropriate. NOAA Fisheries should bring together an independent, interdisciplinary panel of modelers and relevant empirical scientists that would meet periodically to pursue the modeling effort collaboratively in an iterative and adaptive manner with the long-term goal of developing tools to support informed, practical decision-making.
As noted, the full predictive model is at least a decade away from coming to fruition, and the management requirements involved in addressing concerns over ocean-noise effects on marine mammals are extremely pressing. Efforts are under way to address the long-term goal of producing the predictive model outlined here, but an interim plan is needed. One strategy is to implement a management regimen that uses available data, agreed upon management goals, and a conservative approach to the insufficiencies of the available data. The regimen should encourage data acquisition to reduce uncertainty. The NOAA Fisheries Potential Biological Removal (PBR) model is such an example.
RECOMMENDATION 6: A practical process should be developed to help in assessing the likelihood that specific acoustic sources will have adverse effects on a marine mammal population by disrupting normal behavioral patterns. Such a process should have characteristics similar to the Potential Biological Removal model, including
Encouragement of precautionary management—that is more conservative (smaller removal allowed)—when there is greater uncertainty in the potential population effects of induced behavioral changes,
Being readily understandable and defensible to the public, legal staff, and Congress,
An iterative process that will improve risk estimates as data improve,
Ability to evaluate cumulative impacts of multiple low-level effects, and
Construction from a small number of parameters that are easy to estimate.
The PBR model has the potential for being applied more widely than it is now. So far, for most species it has incorporated only direct fishery mortalities and serious injuries in the determination of biological removal. Indirect fishery mortalities, nonfishery mortalities, and mortality equivalents for injury and disruption need to be added to the biological removal in the model to encompass the multitude of effects, including acoustic effects, of human activities on marine mammal populations.
RECOMMENDATION 7: Improvements to PBR are needed to reflect total mortality losses and other cumulative impacts more accurately:
NOAA Fisheries should devise a revised PBR in which all sources of mortality and serious injury can be authorized, monitored, regulated, and reported in much the same manner as is currently done by commercial fisheries under Section 118 of the MMPA.
NOAA Fisheries should expand the PBR model to include injury and behavioral disturbance with appropriate weighting factors for severity of injury or significance of behavioral response (cf. NRC, 1994, pg. 35).
Current knowledge is insufficient to predict which behavioral responses to anthropogenic sounds will result in significant population consequences for marine mammals. The predictive model and even the proposed revisions to PBR will take years to implement. In the interim, those who introduce sound into the marine environment and those who have responsibility for regulating sound sources need a system whereby reasonable criteria can be set to determine whether a particular sound source will have a non-significant effect on marine mammal populations. Collectively, there is sufficient expert knowledge and there are extensive databases to establish such a system and to set the criteria conservatively enough for there to be
broad agreement on the nonsignificant effect criterion. An example of a preliminary application of the approach is the impact-likelihood risk-evaluation matrices developed for typical acoustic equipment used on research vessels in the Antarctic (SCAR, 2004).
RECOMMENDATION 8: An intelligent-decision system should be developed to determine a de minimis standard for allowing proposed sound-related activities. An expert-opinion panel should be constituted to populate the proposed system with as many decision points as current information and expert opinion allow. The system should be systematically reviewed and updated regularly.
The goal of this report is to provide a method for clarification of the concept of biologically significant disturbance. The recommendations made here are intended to provide both a long-term, well-supported, and valid solution and a near-term problem-solving strategy to assist resource managers in coping with this difficult and complex issue.