Approaches to Assess Cumulative Impacts
The previous chapters of this report have reviewed a variety of “approaches to assess cumulative effects of multiple stressors on marine mammal populations that, in turn, have direct and indirect effects on vital rates and population health” as stipulated in the statement of task (see Chapter 1). There are very few situations where one can link exposure to stressors directly to effects on marine mammal populations. Several approaches are discussed, beginning with those of limited use for marine mammals and then moving on to those with greater utility for this task.
APPROACHES WITH LIMITED APPLICATION FOR EVALUATING CUMULATIVE EFFECTS IN MARINE MAMMALS
The primary experimental method used to evaluate cumulative effects of stressors involves factorial experiments that manipulate two or more stressors in animals that can be held in controlled settings. As discussed in Chapter 4, many stressors are likely to interact, and their effects should only be assumed to be additive if there are sound biological (as opposed to purely statistical) reasons for this assumption. The committee’s review of meta-analyses of these experiments concluded that there are no obvious generalities that could help us to predict the effects of interactions between stressors on marine mammals in the wild. There are so many stressors affecting marine mammals and the ecosystems upon which they depend that the traditional approach of starting with impacts of individual stressors and then studying interactions when small sets of stressors are added together is not practical. Halpern et al. (2007) found that all of the marine ecosystems they surveyed were threatened by at least nine stressors, leading to hundreds of potential interactions that would need to be studied. This is not practical for marine mammals.
Alternative Model Species
The difficulties of studying cumulative effects in protected, large, long-lived animals such as marine mammals has led some to argue for consideration of other easier-to-study taxa as surrogate model species (Caro and O’Doherty, 1999). However, as Chapter 3 discusses, terrestrial mammals may differ enough in responses to stressors that they may not be good model systems for marine mammals. For example, investigations in pinnipeds have shown that increased oxidative stress during fasting and diving is ameliorated by oxidant-induced hermetic responses that increase antioxidant capacity more than would be predicted using studies from terrestrial mammals (reviewed by Vázquez-Medina et al., 2012). There also are serious questions about extrapolating information about interactions between marine stressors from nonmammalian marine model species to apply to marine mammals. As homeotherms, the response of marine mammals to temperature is very different from that of animals whose temperature matches the ambient. As animals that breathe air, marine mammals are much less sensitive to water-borne compounds than animals that extract oxygen from water. In this report the committee urges caution when extrapolating from non–marine mammal species in assessing cumulative effects of stressors on marine mammals.
There are significant logistical and ethical problems with experiments that intentionally expose marine mammals in the
laboratory to stressors such as pathogens. However, studies have been conducted on stressors such as sound, toxins, and temperature. Chapter 2 reviews studies on effects of sound on marine mammals. De Swart et al. (1996) and Ross et al. (1996b) fed harbor seals with herring from either relatively uncontaminated areas of the Atlantic Ocean or from the contaminated Baltic Sea. Baltic herring was immunotoxic to the seals, potentially reducing their resistance and increasing risk from infectious diseases. Yeates and Houser (2008) determined how low the temperature of air or water had to go before the metabolic rate of their bottlenose dolphin subjects became elevated. Water temperature had a stronger effect than air temperature, and little synergy was observed between the two. These studies of physiological responses to stressors illustrate that laboratory studies can demonstrate causal relationships between stressors and effects.
There may be further scope for laboratory research on effects of stressors on marine mammals, but there is a major advantage for research on wild animals. Marine mammals are exposed to such broad and poorly quantified arrays of stressors that it would be difficult to attempt to reproduce these combinations of stressors in the laboratory. By contrast, if one wants to study the effect of adding one stressor, such as sound, to a population influenced by many stressors, then one can select subjects from the wild population that are exposed to the current combination of stressors. Exposure to intrinsic stressors will vary with life history, and exposure to extrinsic stressors will vary in time and space. If the goal is to study animals whose allostatic load is high, this suggests selecting times when both intrinsic and extrinsic stressors lead to the energy demand exceeding supply (McEwan and Wingfield, 2003). This goal suggests an alternative to fully sampling the range of exposures in the wild. However, studies that involve adding one stressor to a wide sample of subjects in the wild actually do evaluate the cumulative effects of all the stressors to which the subjects are exposed. One cannot count on the same being true for studies of animals that are maintained in laboratory environments where animals are well fed and free from predation and many other stressors. These considerations suggest that wild marine mammals may be more appropriate subjects for studies of cumulative effects than captive animals.
SAMPLING STRATEGIES THAT DEPEND ON RANGING PATTERNS
The opportunities and obstacles for making critical measurements depend on the ranging patterns of the species under study. There are four main patterns for marine mammals that are relevant for sampling strategies for assessing cumulative effects of stressors in marine mammals.
Accessible Resident Populations
Species with home ranges that are small and near shore can be studied in a cost-effective manner by biologists using small vessels to sight individuals that can be identified by markings. These kinds of studies have proven valuable for tracking birth, growth, and death of nearly every individual in a population (e.g., Brault and Caswell, 1993). The overall exposure of the population can be measured on a seasonal or annual basis for a range of stressors based on environmental sampling. Comprehensive health assessments also are able to measure the dosage of individuals for some stressors, along with data on responses to stressors. These studies have been conducted with several populations of bottlenose dolphins that live in coastal waters of the southeastern United States, providing demographic data that can be compared across sites. Comprehensive health assessments involving suites of biomedical sampling (Wells et al., 2004) have also taken place at several of these sites, providing critical data for evaluating the dosage and effects of stressors that impact only one or a few of the sites. For example, Schwacke et al. (2014b) compared results from dolphins oiled after the Deepwater Horizon event to those from a population in Sarasota Bay, Florida, far from the oiling, and Venn-Watson et al. (2015) compared oiled dolphins to those that had stranded in other areas. For populations with limited home ranges, these concurrent studies in several populations provide a powerful tool for studying effects of stressors whose exposure varies across the locations.
Some species associated with deep oceanic areas have small enough home ranges for observational methods to provide important longitudinal data in areas where deep water is close to shore. For example, some beaked whale species are thought to have limited home ranges near seamounts or undersea canyons. Claridge (2013) was able to obtain important life-history data from populations of Blainville’s beaked whale (Mesoplodon densirostris) in Bahamian waters. Similar data have been obtained for pilot whales in the Strait of Gibraltar where a small population of pilot whales resides (Verborgh et al., 2009). These situations may give a biased view, however. For example, pilot whales in most other study sites range so widely that there are relatively low rates of resighting individuals in one location.
Species with Predictable Locations for Birth on Land
Pinnipeds that come ashore in between foraging trips at sea and that give birth on land offer special opportunities for study. Long-term studies of identified individuals in this case can more easily involve sampling, weighing, and tagging than studies for species where animals do not come ashore. The foraging trips may take days to months—durations that are well within the scope of established tag attachments. Some of these species are suitable for the analysis of body condition through measuring buoyancy during drift dives. New et al.
(2014) showed how data on weight and survival of mothers and pups could be coupled with tag data measuring how foraging affects body condition. These data can be incorporated into the kind of model developed in Chapter 5 to relate how variation in stressors leads to variation in reproduction and calf survival. The main obstacles to studying interactions between stressors in these species involve development of more studies of identified individuals, and development of ways to measure exposure and response to stressors. These species are among the most promising for development of studies using the model from Chapter 5.
Species That Are Accessible at Some Points Within Large Home Ranges or During Annual Migrations
Some migratory species of cetacean congregate near shore for enough of their annual cycle to be studied by shore-based researchers. When accessible, these populations can be studied by observing individual animals that have distinctive marks. For species with several such sites, comparing sightings can allow movements to be tracked, but this is biased by the observation sites and is likely to lead to an incomplete view of the population range. For example, the population of right whales in the Northwest Atlantic is well studied from sightings during the summer foraging season, enough to estimate risk of extinction (Caswell et al., 1999). A subset of the population migrates to coastal waters off the southeastern United States, but little is known about where the other segment winters. Similarly, long-term observations of a small population of killer whales that are routinely sighted in Puget Sound, Washington, has provided solid evidence of a decline, enough to list the population as endangered (Ford, 2013). However, this population ranges as far as California during the winter, and little is known about their exposure or response to stressors during this part of the year. In these cases, focused tagging efforts may be needed to supplement local field studies. Obtaining measurements and attaching tags to these animals will be more challenging than working with animals that haul out on land. In addition many of these migrations occur on an annual basis, requiring longer tag attachment times than for most species that give birth on land, to cover the time at sea away from the nearshore site. Many species that have large home ranges or migrate annually have been tagged with satellite tags, but this is expensive, so the sample size is low. Few tags are available with longevity sufficient to cover an entire migration period, but the success rate and length of attachment duration are increasing as the technology evolves (Mate et al., 2007).
Open Ocean Species
Species that are widely distributed in the open ocean are the most challenging for studies of cumulative effects. It is difficult to develop longitudinal studies that involve resighting individuals over such large areas, and it is more difficult to sample or tag animals on the high seas than on land or in shallow coastal waters. Some solutions have been developed for these problems. Remote tagging and biopsy methods have been developed, but these are more limited than those available onshore or where one can handle the animals. Further development of sampling and tag attachments will be required to apply the approaches recommended in this report for open ocean species. Researchers studying the stress to pelagic dolphins of encirclement in tuna nets used the encirclement itself to enable handling, sampling, and tagging dolphins in a floating restraint system (Scott and Chivers, 2009), but this is unlikely to be possible for larger whales. Smith et al. (1999) report on a systematic and standardized effort to photo-identify and biopsy sample humpback whales throughout the North Atlantic. Similar scales of effort would likely be required for sampling exposure and response to stressors for populations of marine mammals that span ocean basin scales. The methods recommended in this report for studying cumulative effects will need considerable development to be applicable for these species.
Combining the difficulty of studying these four groups of marine mammals with the vulnerability of their populations suggests a broad set of priorities. The marine mammal species most at risk of extinction over the past few decades have not been the migratory large whale species, but rather populations of river dolphins, such as the baiji or Chinese river dolphin (Lipotes vexillifer) (Turvey et al., 2007). A range of anthropogenic stressors have been implicated in the decline and extinction of the baiji, with physical injury as a result of interactions with fishing gear being the most important. The limited home ranges of the resident species make them more vulnerable to localized concentrations of stressors. By contrast, the harder-to-study migratory and open ocean large whale species may be less vulnerable. Even though most of these species were exploited during the era of commercial whaling, some populations are large and/or recovering (Whitehead, 2002; Thomas et al., 2016), and the scale of their distribution and movements may render them less vulnerable to local exposure to stressors. This combination of difficulty of study and lower vulnerability may lower the priority for this group for studies of cumulative effects. However, some migratory baleen whale populations, such as the right whales of the western North Atlantic, are exposed to many stressors and have a small and declining population (Kraus and Rolland, 2007). Their coastal distribution puts them at higher risk and makes them easier to study, promoting their priority.
APPROACHES TO ASSESS COMPONENTS OF THE PCOMS FRAMEWORK
Chapter 5 presented a framework for analyzing cumulative effects of stressors on marine mammals. Here we describe approaches to assess cumulative effects organized by the different components of this framework. This sec-
tion focuses on methods to estimate critical parameters in the context of studying relationships between exposure to stressors and (1) behavioral or physiological responses, (2) health, or (3) vital rates.
Measuring Exposure to Stressors
Lioy and Rappaport (2011) identified two different ways by which biomedical researchers could estimate exposure to chemical stressors that influence human health: a geographical approach and a subject-oriented approach. The geographical approach focuses on different external sources of exposure to a contaminant, which must be summed up to estimate aggregate exposure. Identifying external sources can help prioritize ways to reduce exposure. However, it can involve massive effort and can miss internal sources of chemical stressors, which may be very important for health (Rappaport, 2011). A subject-oriented approach samples directly from the subjects to measure contaminants or their biomarkers. This subject-oriented approach suggests the utility of sampling blood or other tissues in order to estimate the dosage of stressors at the animal to evaluate their impact on health and vital rates (Rappaport, 2011). Placing the sampler on the subject frees the study from needing to track the changing location of the subject, and to associate exposure with time spent in each location. The pros and cons of geographical and subject-oriented approaches to measuring stressors in marine mammals are similar to those identified by Rappaport (2011) for humans.
Spatial and Temporal Distribution of Stressors in the Environment
The geographical approach to identify potential risks from the complex combination of stressors in the world’s oceans requires mapping the distribution of the species of concern along with mapping stressors in space and time. An assumption of this geographical approach is that stressors must overlap with the species to exert a cumulative effect. For example, risk of physical injury from fishing or shipping can be estimated by the flux of categories of ships or the density of fishing gear that pose different threats of injury (e.g., fast versus slow ships, gillnets versus other nets). Similarly if predators, competitors, or anthropogenic sources need to be relatively nearby to be perceived as a threat, then data on the distribution of these stressors may provide a useful estimate of exposure. However, mapping noise from acoustic stressors cannot always be derived from information about the location of intense sources alone. Underwater sound can propagate so well that the same sound produced in the Indian Ocean can be detected off California and off Bermuda but at different levels (Munk et al., 1994). The best way to estimate exposure to one or several intense acoustic stressors is to combine acoustic propagation modeling with measurements of levels of sound produced at known ranges and of the transmission loss in the environment. Acoustic propagation models can use source and transmission loss data to predict the sound field around these sources and to guide selection of recording sites to best ground-truth predictions. In cases where sources cannot be so readily identified or measured, ambient noise can be monitored directly. Increasing numbers of acoustic observing systems are coming online globally (Miksis-Olds and Nichols, 2016), providing useful data on integrated exposure to noise from all acoustic stressors.
Similarly, the risks from biological or nonbiological toxins cannot always be derived simply from mapping occurrence of sources of toxins or concentrations in the environment. The processes by which toxins are released, transported, and distributed from sources through environmental media and potentially through the food web to marine mammals are complex and will depend on a number of variables related to the toxin, the habitat, and the species of marine mammal. In some cases, it is possible to examine environmental samples from water, sediment, or prey to predict exposure for marine mammals, but, for toxins that can be detected directly in marine mammal tissues or fluids, direct collection and measurement in marine mammal samples is a preferred approach for characterizing dosage. As discussed in Chapter 3, persistent organic pollutants (POPs), many inorganic contaminants, and harmful algal bloom toxins have been routinely measured from a variety of remotely collected tissue samples. Metabolomic analyses of respiratory samples and proteomic and transcriptomic analysis of tissue samples hold promise for the development of biomarkers that indicate cumulative dosages of many toxins. Respiratory samples also hold promise for detection of markers indicative of pathogenic infections. Similar to toxins, exposure to pathogens can often be better characterized by direct sampling of the animal as the presence of a pathogen in the environment does not necessarily translate to an exposure risk. The actual exposure the animal experiences will depend on a variety of factors, including the presence of transmission vectors, or social structure and aggregation (e.g., colonial breeding) that affect contact rates with infected conspecifics. However, while direct measurement from actual tissues from marine mammals is a preferred approach to measure dosage for toxins, this approach requires extensive sampling effort and analyses that are often very costly. In this regard, it would be beneficial for researchers from multiple disciplines and agencies to collaborate and leverage efforts across projects to collect and analyze samples, building a baseline of data that allows examination of geographic trends for multiple stressors.
Prey limitation is a key factor influencing body condition and, as Chapter 6 emphasizes, is a critical part of the interaction web for marine mammals. Marine mammals are well adapted to use sensory cues from echolocation, vibrissae, and more standard mammalian senses to detect, select, and capture prey. Human methods using ship-based echosounders and nets to map prey are crude by comparison and cannot
yield a complete view of availability of preferred prey for marine mammals. However, Friedlaender et al. (2016) have shown that inclusion of prey density and distribution can explain variation in dive behavior of foraging blue whales in a way that greatly increases the power to detect responses to other stressors, such as anthropogenic sound. Further development of methods to measure prey fields may improve these estimates. However, there are considerable obstacles to measuring prey fields in a way that accurately estimates prey limitation for marine mammals. Well-funded long-term censuses of commercially important fish have not solved the challenge of mapping their distribution, even for informing the management of those commercial stocks. There are very few stock assessments of species that are important prey for marine mammals but not important for human fisheries. In addition, measuring prey fields may not provide a complete estimate for the stressor of prey limitation. For example, if prey change their behavior or localized distribution so they are less accessible, then a foraging marine mammal may experience prey limitation even when the prey are present in the area. Here also, the specifics of how, when, and where marine mammals forage may be needed to assess the level of stress from prey limitation. Exposure to prey limitation as a stressor may be estimated by such measures of prey availabilty, although such data are often limited and difficult to interpret for generalist predators. All of these considerations emphasize the importance of developing measures of foraging success of individual marine mammals over time.
Predation pressure is a stressor that can be an important driver, but measurement of predation risk is difficult for marine mammals. Two important predators of marine mammals are sharks, such as great white sharks (Carcharodon carcharias) and the killer whale (Orcinus orca) (Jefferson et al., 1991). When killer whales are hunting small marine mammals in coastal waters, kills can often be observed visually for an estimation of predation pressure (Baird and Dill, 1995). Baird and Dill (1996) were able to follow killer whales and observe predation events to estimate rates of predation from the predator’s perspective. However, these observations are not the same as estimating the risk of predation from the point of view of marine mammals targeted by the predator. Springer et al. (2008) discussed reasons why killer whale predation on large whales may be underestimated by visual observation. Some preliminary work has demonstrated the ability of tags to detect predation events on tagged pinnipeds. Horning and Mellish (2014) analyzed data from 36 Steller sea lions tagged with life-history tags (Horning and Hill, 2005) and were able to conclude that 15 of these sea lions had been killed by a predator. This tagging work identified a new unsuspected shark predator of these sea lions, but this approach is not appropriate for all species, and its cost limits the sample size, making it unlikely to provide robust estimates of predation risk even for species where it can be used. When predation events cannot be studied directly, another method for estimating the risk of predation is to measure when predators interact with prey. Some investigators use scars from shark or killer whale attacks as indicators of predation pressure (Heithaus, 2001), but this is problematic as the scarred individuals are the ones that got away. Accurate estimation of predation pressure for marine mammals remains a significant challenge.
Animal-Oriented Approaches to Measuring Extrinsic and Intrinsic Stressors
Mapping of stressors allows one to estimate exposure at specific locations. However, many marine mammals range over wide areas. If their path is not known, stressor maps may not suffice to estimate exposure. And, as discussed above, broad geographical overlap is not enough to predict exposure for stressors that concentrate in a narrow part of the geographical area, in particular substrates such as sediment, or in prey that must be ingested. As Chapter 3 notes, in these circumstances, the preferred approach is often to sample tissue from a marine mammal to characterize its dosage of chemical stressors. Tissues can currently be sampled from animals that are held for health assessment, but capabilities for sampling critical tissues such as blood are limited for many marine mammal species. New methods will need to be developed for this subject-oriented approach to reach its full potential for marine mammals.
Passive and active personal dosimeters have become established as useful methods for measuring the dosage of stressors. Here the stressor is either absorbed into a passive matrix (O’Connell et al., 2014) or measured by an active device on the animal or human (Boziari et al., 2010). Dosimeter tags have been developed to measure the dosage of some stressors on marine mammals. Acoustic sensors have been placed on marine mammal tags to quantify the dosage of sound at the animal (Johnson and Tyack, 2003). Optical sensors have also been deployed on tags on marine mammals, both to form images of prey (Hooker et al., 2002) and to measure bioluminescence from potential prey (Vacquié-Garcia et al., 2012). A variety of sensors have been used to detect attempts to capture prey (Plötz et al., 2001; Miller et al., 2004a) or the ingestion of prey (Austin et al., 2006), which may provide direct measures of foraging rates.
Managing Information on Stressors and Ecological Drivers
The obstacles described above for measuring prey limitation and predation pressure highlight the difficulties of assessing single components of interaction webs. The movement toward Integrated Ecosystem Assessments may support broader studies of interaction webs that focus on all human and natural nodes (Samhouri et al., 2014) and that prioritize focal ecosystem components (Levin et al., 2014). However, it will require substantial investments from funders in order to
improve the estimates and accuracy of the various exposures to drivers and their effects.
As discussed in Chapter 7, long-term monitoring across broad spatial and temporal scales (including both passive and active surveillance) could help improve understanding of the geographic and temporal patterns of stressors as well as associated adverse effects, and also could help in detecting emerging health issues in marine mammals that are potentially indicative of a population at risk. In addition, understanding patterns of dosage and exposure for multiple stressors could help to inform future study designs to elucidate potential cumulative effects. This information will be most powerful if it is made widely available to scientists and managers through a centralized data management system that can interface with other databases that allows integration of marine mammal health data with ecosystem and oceanographic data.
Such a data management system, the Marine Mammal Health Monitoring and Analysis Platform (MM Health MAP), has been proposed and is in the early developmental stages (Simeone et al., 2015), being led by the U.S. National Marine Fisheries Service’s (NMFS’s) Marine Mammal Health and Stranding Response Program (MMHSRP) and the U.S. Marine Mammal Commission. The goal of the MM Health MAP is to support mandates under Title IV of the U.S. Marine Mammal Protection Act (MMPA) to gather data on marine mammal health trends and correlate these with biological, physical, and chemical variables.1 However, the successful development and implementation of the MM Health MAP will depend on support not only from the NMFS but also from other federal managers, as well as cooperation and collaboration across the marine mammal research community. These efforts require willingness of, and financial support for, independent research groups to make data available. Other management and funding agencies should also encourage data management policies that lead to broader analyses and synthesis of information, including incorporation of data and model products into such databases. Similar levels of cooperation between the research community and public-sector agencies involved in tracking emerging diseases and specifically zoonotics have been observed (IOM and NRC, 2009). One such example is the PREDICT program within the U.S. Agency for International Development’s Emerging Pandemic Threats Program. The PREDICT program is one of the world’s most comprehensive zoonotic disease surveillance and capacity development programs; they have developed training for staff and low-cost detection tools for new viruses from targeted virus families in 32 laboratories in 20 developing nations. Such efforts, supported by modern data management practices and information sharing, have helped characterize human and ecological drivers of disease spillover from animals to people, and strengthened models for predicting disease emergence in wildlife (Jonna Mazet, personal communication).
To ensure comparability of the marine mammal health and stressor exposure data across studies and over space and time, such a system would require standardized information and proper quality assurance plans for the various analytical results. One of the components of the MMHSRP, which was established under the 1992 amendments to the MMPA, has been to coordinate analytical quality assurance of data from chemical analyses of marine mammal tissues. The quality assurance program for analysis of POPs, fatty acids, and trace elements in marine mammal tissues has been implemented through the National Institute of Standards and Technology and includes interlaboratory comparison exercises, as well as the development of control materials and standard reference materials for marine mammal tissues. Similar quality assurance measures would need to be identified and, if not in existence, would need to be established for other types of health data (e.g., stress hormones) in order to ensure accuracy and interpretability of results across laboratories. Such efforts would broaden understanding of stressor exposure across regions, provide necessary information to managers to assist in evaluating potential stressor mitigation strategies, and inform researchers interested in hypothesis generation for future analytical studies.
Finding 8.1: Improving the estimates of the exposure to and dosage of stressors, and their effects, will require better data availability, standardization, and management. The merger of both stressor and ecological driver-related data through a centralized database would facilitate integration and analyses.
Measuring Change in Behavior and Physiology
Most studies on the effects of sound on marine mammals focus on end points related to disturbance, such as behavioral changes. Where concern has focused on acute effects, such as strandings of beaked whales in response to sonar, it can be very useful to document levels of sound below which no short-term response occurs that poses a risk of stranding. Fernández et al. (2005, 2012) argue that exposure to sonar may also pose a risk of decompression sickness (DCS). Analyses of dive profiles using physiological models of gas dynamics during diving have been used to estimate the risk of physiological changes that could lead to DCS (Kvadsheim et al., 2012). Diving responses of beaked whales to actual sonar exercises have not been quantified, but they have been measured for experiments that used controlled exposures of sonar to tagged beaked and other whales. The behavioral responses to sonar observed in these experiments led to modeled end-dive N2 tensions thought not to pose a significant risk of DCS. However, sonar exercises involve more intense and prolonged exposure than occurred during these experiments, which were designed to minimize risk of
injury. Therefore, while the exposure levels linked to these experiments do not pose a significant risk of DCS, the study cannot rule out that behavioral and physiological responses to actual sonar exercises could cause DCS. Testing for DCS in animals that strand coincident with sonar exercises may benefit from careful measurement of the distribution, volume, and gas composition of bubbles, as this may help discriminate between decompression and decomposition in stranded marine mammals (Bernaldo de Quiros et al., 2012).
For many other responses, there is a critical need to develop methods to evaluate the effects of chronic exposure. Analysis of health in terms of energy stores is a promising way to do this, as it can integrate with energetic models of survival and reproduction (e.g., New et al., 2013b). Further development of methods to estimate the energetic consequences of changes in foraging behavior and the physiology of metabolism will strengthen the promising approaches of Biuw et al. (2003) and New et al. (2014). For example, Wilson et al. (2006, 2008) advocate use of accelerometry to estimate metabolic rates of tagged subjects, and Fahlman et al. (2016) and Roos et al. (2016) describe improvements in methods that use respiration to estimate the metabolic rate of cetaceans.
Another important approach for measuring physiological changes resulting from exposure to stressors involves measuring glucocorticoid stress hormones. A few studies have measured changes in stress hormone levels of marine mammals exposed to sound (Romano et al., 2004; Rolland et al., 2012). Methods are being developed to sample stress hormones from a variety of tissues, such as blubber biopsy, feces, and blows. These methods are critical for practical sampling of animals in the wild, and data from these tissues need to be calibrated against data from blood, which is the standard.
The Functions Relating Exposure to Stressors to Behavioral or Physiological Responses
Short-term tags are well suited to experiments studying responses to acute exposure to intense sounds, and these experiments can produce probabilistic dose–response functions (e.g., Figure 1a in Box 2.2). Once these responses are characterized, monitoring programs can be developed to evaluate responses to longer-term and larger-scale exposures (e.g., Moretti et al., 2014). However, few of these studies have estimated exposure to other stressors that might influence cumulative effects. To evaluate cumulative effects of other stressors in addition to noise, these studies would need to include measurements of exposure to other stressors and responses to them.
The levels of exposure for an individual marine mammal to stressors such as noise, prey limitation, perceived threats, and disease may vary considerably as the animal moves over time periods of minutes to days. The biological responses to a sound stimulus are likely to vary as a function of behavioral states, such as traveling or foraging, and of physiological states, such as oxygen reserves or acute disease infection, that may vary on scales of seconds to days or more. These time scales require behavioral and physiological measurements along with estimates of stressor exposure that are local to the animal. These kinds of data on behavioral and physiological states have been used in experiments to evaluate the effect of behavioral context and the responses of marine mammals to acoustic stimuli (e.g., Goldbogen et al., 2013); this approach may offer some promise for studying cumulative effects involving other stressors.
There is also a data gap for studying effects of chronic exposure to sound. Short-term experiments can expose the same subjects several times to the same or different acoustic stimuli (Antunes et al., 2014; Miller et al., 2014). These experiments enable testing whether responses differ for the first exposure versus later ones, which is a first step in studying responses to repeated sounds. Some studies have taken advantage of unplanned events to study the impact of reductions in chronic noise on marine mammals. For example, Rolland et al. (2012) happened to be studying stress hormones in right whales before and after the terrorist attacks on the World Trade Center and Pentagon on September 11, 2001. Noise levels and the occurrence of ships passing near the whales were greatly reduced due to a pause in commercial shipping after these events; during this period of low noise and ship activity, the levels of stress hormones were lower than those recorded before September 11, 2001, or for the same period in other years. However, this opportunistic study lacks the controls required for a standard experimental design. New designs for experiments and opportunistic studies will be required to document the effects of planned changes in chronic noise and disturbance associated with ship passage induced by changes in shipping lanes or in shipping technology.
Use of Health Indices to Detect and Manage Species at Risk
Chapter 5 developed the Population Consequences of Multiple Stressors (PCoMS) framework that uses health parameters to help integrate effects of multiple stressors over longer time periods than those captured by individual physiological or behavioral responses to acute stressor exposures. Measuring these health parameters can improve the ability to model the linkages between stressor dosage or exposure and long-term effects on populations. Changes in health integrate short-term changes in exposure to multiple stressors, providing a longer-term measure that can more readily be linked to changes in vital rates. Because changes in health can be measured more rapidly than changes in vital rates, health may help provide an early warning indicator for individual animals. If enough individuals in a population are sampled for health, as Chapter 7 discusses, this information
on population health may provide an early warning indicator for populations at risk.
Comprehensive Health Evaluation
Comprehensive health assessments are of particular value because they provide information on multiple aspects of an animal’s condition and are therefore more likely to detect a compromised health state. In addition, health assessments that utilize an array of indicators can help to identify specific causal factors for compromised health and can inform management decisions about which steps to take to reduce risks. Comprehensive health assessments have been developed for pinnipeds and some cetacean species, such as bottlenose dolphins (Tursiops truncatus). In pinnipeds, contaminant burdens measured in tissues, and pathogen exposures sampled from nasal and rectal swabs, can be included in physiology workups for tag deployments and recoveries that also include body condition, stress hormones, and immune markers (e.g., Goldstein et al., 2013; Peterson et al., 2015; Peck et al., 2016). For example, recent work using nasal swabs showed that tagged elephant seals were exposed to the H1N1 virus between instrument deployments and recoveries in 2010 (Goldstein et al., 2013). Comprehensive health assessments have also been conducted for coastal populations of bottlenose dolphins in several sites in the southeastern United States (Wells et al., 2004; Fair et al., 2006; Schwacke et al., 2010). In some cases, these studies have identified adverse health effects in association with stressor exposure. For example, a high prevalence of anemia, low thyroid hormone levels, and immune suppression were associated with polychlorinated biphenyl exposure in bottlenose dolphins inhabiting an estuary near a hazardous waste site in Brunswick, Georgia (Schwacke et al., 2012). Most of these studies rely on sampling of blood but may also include sampling of other tissues or body fluids, and ultrasound examination of organs. Baseline data from these kinds of assessments are critical for studying stressor dosage and responses to stressors.
Understanding the health status of a population aids in the identification of threats that can be effectively mitigated to support recovery, whether or not they have been major contributing factors for the population’s decline. For example, health studies of highly endangered Hawaiian monk seals found that the species was immunologically naïve to morbillivirus, which posed a significant epidemic threat, and furthermore that the lack of genetic diversity could potentially limit the ability of the species to respond to other newly introduced diseases such as toxoplasmosis, West Nile virus, and influenza (NMFS, 2016b). In response, NMFS identified an action to “Detect and prevent catastrophic disease outbreak and disease-related mortality” as a priority in the 5-year action plan for recovery of this species that was on the brink of extinction. A disease outbreak preparedness plan, including the development of a morbillivirus vaccination program, has now been implemented as part of ongoing health research activities.
Assessing Health in Populations That Cannot Be Handled
Current methods and technologies limit comprehensive health assessments to a few species that can be temporarily captured, restrained, and evaluated. This limitation has led to the development of less comprehensive health assessments for other species, often including two types of readily accessible indicators of health: body condition and stress hormones. As these measures can be obtained using visually observed indicators for body condition, or remote sampling for stress hormones, they can be collected for many marine mammal species.
As discussed in Chapter 5, body condition is an indicator of health and allostatic or homeostatic load that can be measured directly for species that can be handled. Methods are more limited for species that cannot be handled. These include visual observations of condition and use of tags to estimate changes in buoyancy of wild marine mammals. Pettis et al. (2004) estimated body condition by scoring the concavity of an area just behind the blowhole that accumulates fat and that is visible in some photographs taken to identify individual whales. C.A. Miller et al. (2012) used aerial photographs taken directly over a right whale to more precisely measure the body shape and quantify the condition of right whales. Unmanned aerial or underwater vehicles may offer more cost-effective ways to obtain such images optimized for measuring features of interest. The tagging method for estimating body condition involves measuring the vertical acceleration of diving animals during drifting periods of the dive. Drift dives, however, do not occur in all species. More detailed research on the forces acting on swimming marine mammals may allow estimation of the static buoyancy force and percentage of lipid in animals that are not passively drifting, but are gliding during ascent and descent phases of normal dives (Miller et al., 2004b; Watanabe et al., 2006; Aoki et al., 2011). This may broaden the number of species that can be studied using this method.
As discussed in Chapter 4, chronic activation of the hypothalamic-pituitary-adrenal axis may be an important mechanism by which cumulative effects of different stressors exert effects on health and vital rates. Glucocorticoid (GC) stress hormones have usually been measured from blood samples, but an array of other matrices for stress hormones, including blubber, feces and exhaled blow, and baleen and earplugs in baleen whales are also being studied for analysis of stress. These other matrices provide longer-term
measures of GC levels than blood and may be more useful for investigating long-term stress dosage and effects. Feces and exhaled blow can be collected noninvasively for some species, and blubber can be sampled by biopsy darting in almost all marine mammal species. The promise of these new matrices cannot be fulfilled without cross-sectional and/or longitudinal studies that help to establish distributions for expected values across different species, age classes, sexes, and reproductive states. Pregnancy changes corticosterone levels in blubber, so such samples also need to measure progesterone to control for this effect.
Remote Assessment of Health
Pettis et al. (2004) conducted an early effort to develop a scale for assessing the health of individual right whales in the western North Atlantic. They took advantage of an extensive photo-identification catalog to score body condition, skin condition, presence of “rake marks,” and cyamids near the blowhole. This assessment scheme was limited to features that were visible from photographs used to identify individual whales. The development of indices that include information from biopsies, blow, and feces will enrich the power of health assessments that are limited to remote sampling.
Health studies that include assessment of body condition as well as collection of contaminant and health biomarkers have been identified as a priority action for the recovery of highly endangered Southern Resident killer whales (NMFS, 2016c). The goal of these health studies is to compare the health of Southern Residents with other killer whale populations to identify potential sources of decreased survival and/or reproduction. High concentrations of emerging contaminants, and specifically flame-retardant chemicals, have been reported in these apex predators (Rayne et al., 2004). Therefore, the health studies are particularly focused on identifying sources for the emerging contaminants and understanding potential associated health effects in order to guide water quality recommendations and reduce contaminant inputs into Southern Resident killer whale habitat.
Finding 8.2: Assessment of health is central to the PCoMS model proposed in this report. Comprehensive health assessments of a cross section of a marine mammal population can also help managers decide when the population is at risk and help them decide which management actions can most effectively support recovery.
Stressor Exposure: Health Response Function
The PCoMS model presented in Chapter 5 has the capability to analyze the short-term links between a health effect and the combination of stressors to which an animal has been exposed. As a sample of wild animals moves through their habitat and/or experiences seasonal changes, they are likely to be exposed to a wide distribution of the stressors that are present in their environment at that time. If the dosage or exposure to the stressors and the effects of each combination of stressors can be measured, then, as Chapter 6 notes, this approach offers the potential for a much larger sample of dose–response measurements than can be tested in experiments, perhaps improving the ability to identify which combinations of stressors have an observable effect on health.
The desired characteristics of the health variables introduced in Chapter 5 are that they can be measured in wild marine mammals, they integrate effects of repeated exposures to multiple stressors, they change over shorter time scales than vital rates, and yet they can influence the vital rates of each individual. The committee has argued that free-ranging marine mammals are influenced by so many stressors, each of whose effects may vary depending on life-history stage of the animal, and that the number of combinations of stressors is too large for experimental studies of how all combinations interact. The committee’s proposed PCoMS framework uses a small number of health variables to integrate the effects from multiple stressors and to improve current understanding of the mechanisms by which combinations of stressors affect vital rates.
Exposure to many of the stressors discussed here varies on an hourly to weekly basis, and even exposure to toxic compounds that have stable concentrations in one area will vary as marine mammals move from area to area. Marine mammals are long lived and give birth at most once per year. This means that studies linking exposure to stressors with reproductive success cannot sample effects more frequently than yearly. By contrast, some of the health variables proposed here have much finer time resolution—more appropriate for linking to stressor exposures. For example, Biuw et al. (2003) state that for estimating body condition from buoyancy in drift dives “biologically realistic changes in drift rate (are) expected to be detectable over a period of 5-6 days.”
If changes in health and exposure to stressors can be sampled over shorter time periods than vital rates, then longitudinal studies may be able to repeatedly measure stressor–health combinations many times within a breeding cycle. Longitudinal studies are particularly well suited for situations where tags can be attached for significant parts of the annual cycle and can sample the health variables of interest. Tags can currently sample body condition in the few species with drift dives but are not able directly to sample the other health variables discussed here. Development of long-term tags that can sample such variables could support this approach for studying cumulative effects. Initial scoping for development would be useful, but breakthroughs are not expected in the next 5-10 years. For these other variables and for species where it is not possible to use tags to measure body condition, it may be more productive to conduct cross-sectional studies where exposure to stressors and the health variables are measured in a large number of individuals within a population. Rather than measuring changes in health
as the pattern of exposure to stressors changes, this approach would sample each individual at a single time point, linking the stressor and health values observed at that time. This approach assumes that the values of stressors observed are close to those that led to the health value measured at the same time. The cross-sectional approach may be less able to detect adverse outcome pathways that involve sequential exposures to stressors over longer time periods.
These kinds of longitudinal and cross-sectional studies are relatively well established for coastal populations of marine mammals in which individuals are small enough to be handled and where relatively comprehensive health assessments have been established. Remote biopsy methods have been developed, but the data obtained by this method are more limited than those available from onshore populations or when one can handle the animals. However, there are precedents for large-scale efforts to sample large, highly mobile whale species. For example, Smith et al. (1999) report on a systematic and standardized effort to photo-identify and biopsy sample humpback whales throughout the North Atlantic. They report that “during 666 days at sea aboard 28 vessels, 4,207 tail fluke photographs and 2,326 skin biopsies were collected.” Their assessment was that “an oceanwide approach to population assessment of baleen whales is practicable.”
One of the goals of the statement of task for this committee is to identify how exposure to nonacoustic stressors may affect a marine mammal’s response to an acoustic stressor. In this context, evaluation of the health status of potential subjects for response studies may help to identify those individuals that may be particularly sensitive or vulnerable to an acoustic stressor. A basic element of the allostasis model is that animals already carrying a large allostatic load may be driven into allostatic overload by a relatively small additional exposure to a stressor. This would suggest that subjects already in adverse health status may be the most vulnerable to even small doses of another stressor. Note, however, that this does not mean that the subject will be the most sensitive in the sense of most likely to show a behavioral response at low exposure levels (Gill et al., 2001). For example, Beale and Monaghan (2004) have shown that birds under nutritional stress may be less likely to stop feeding and move away from a threat than birds of better body condition that may more easily be able to afford the lost foraging opportunities. This emphasizes the importance of measuring the response to stressor in terms of changes in health as well as observing behavioral reactions.
Health Response: Vital Rates Function
The functional relationship between health and vital rates is an important link in the PCoMS model. Parameterizing this relationship will require measuring health and vital rates in the same individuals and populations. Several different methods are used or have been proposed for studying vital rates.
As Chapter 7 notes, vital rates have been estimated for wild marine mammal populations where the same individuals can reliably be resighted. Many demographic parameters can be estimated from focused mark–recapture surveys of animals that can reliably be sighted nearly every year and for which it can be determined whether adult females have given birth. Birth rates and survival of the young are highlighted in Chapter 7 as early demographic indicators of problems; these are most easily studied in species that give birth on land where it can be observed or where young animals are easily distinguished. Several new methods may be appropriate for species where this is not possible, and these will be discussed next.
Matrices That Store Information on Age-Specific Reproduction and Age at Death
One common method for determining the age of mammals involves counting growth layers in tissues such as teeth, baleen, or wax laid down in the ear canal of baleen whales (called the ear plug). Growth layers in teeth have been used to determine the age of dolphins (Hohn et al., 1989), polar bears (Calvert and Ramsey, 1998), and pinnipeds (Scheffer, 1950). Not only can these tissues be used to age marine mammals, but recent work has shown that ear plugs and baleen can provide time records of reproductive and stress hormones as well as contaminants over the lifespan in the case of the ear plug (Trumble et al., 2013) and over several years in the case of baleen (Hunt et al., 2014). Baleen and earplugs are laid down in layers that differ during different parts of the annual cycle, such as feeding, migration, and breeding, making it possible to track each year of life of the animal. Both of these tissues are relevant only for baleen whales—more work on tissues such as teeth that lay down layers throughout the lifespan would help broaden this approach to other marine mammals. In many organisms that lay down these kinds of layers, characteristics of the layer may also indicate the nutritional state of the organism at the time of deposition (Fritts, 2012), potentially providing information on changes in condition.
Life History Tags
Problems with estimating age-specific mortality, and especially causes of mortality in open ocean species, led Horning and Hill (2005) to develop an electronic tag that is implanted internally, recording life-history data through the life of a marine mammal, and that releases and transmits data upon expulsion from the dead animal. Insertion of a tag into the peritoneal cavity requires surgery, but Horning et al. (2008) report that 4 California sea lions (Zalophus
californianus) and 15 juvenile Steller sea lions (Eumetopias jubatus) recovered well under veterinary care after the tag insertion. The sea lions were then released into the wild and tracked with satellite tags. The behavior of sea lions with implanted tags was monitored for up to half a year and was similar to that of sea lions tagged only with satellite tags. Distinct signatures of temperature and light identify when an animal has been killed by a predator (Horning and Mellish, 2014). Temperature data from 15 of the 36 sea lions tagged by Horning and Mellish (2014) indicated that they had been killed by predators. These sea lions were followed for a total of 111 years, so 15 deaths indicate a relatively high predation rate.
The costs and risks of surgical insertion of the life-history tag limit the sample sizes achievable for this kind of tagging, and it may not be appropriate for many marine mammal species. Surgical implantation raises ethical and animal welfare concerns that would require evidence of a clear benefit to these populations that would be sufficient to outweigh the welfare cost. However, this research showed that tags can be developed to record data from within an animal until it dies. This mode of tagging suggests a new approach for active personal dosimeters. The dosimeters described above are designed to measure the dosages of stressors to which an animal is exposed. The potential of a tag that can sample the internal milieu of a marine mammal throughout the lifespan would be greatly expanded if, as with earplugs, it could also sample life-history events, stressor dosage, and response to a variety of stressors. Passive personal dosimeters have been designed with materials optimized for absorbing and storing chemical compounds of interest (Paulik et al., 2016). Tags placed inside the body are best located to measure physiological parameters, such as hormones, and dosages of stressors, such as contaminant loads. For species that do not have tissues from which age-specific samples can be recovered, such as the earplug, there may be benefit in designing passive samplers that can sample compounds of interest at known times throughout the lifespan. Some compounds and other stressors, such as sound, can be detected actively by sensors on an electronic tag, but development of active sensing in lifetime tags will face considerable obstacles in terms of power requirements and space limitations.
Stressor Exposure: Vital Rates Function
Modeling each component of the PCoMS model is very challenging, but it is necessary in most cases, because a direct link cannot be made between stressor exposure and vital rates. However, in cases where a direct link can be made, it may be possible to bypass all the intermediate modeling stages. Such studies have been attempted for several seabird species whose demography and movements have been well documented. Some studies have used the approach taken by Forcada et al. (2006) to compare annual variation in demographic parameters to natural variation in more than one stressor on a year-by-year basis. For example, Rolland et al. (2009) used 26 years of demographic data from a study of black-browed albatross on Kerguelen Island to study the impact of fishing bycatch under various climate conditions. Levels of ocean warming expected for the next century were predicted to enhance the growth of this population, potentially compensating for controlled increases in fishing effort. This analysis was useful to inform management of fisheries in the presence of climate change. However, the authors did not explicitly model potential interactions between stressors.
Few studies on marine mammal populations have used methods similar to those just discussed for seabirds and summarized by Barbraud et al. (2012). However, the demographic parameters for populations of pinnipeds that breed on land could be studied using similar methods. Similar analyses should be possible for species such as resident coastal cetacean populations with long-term studies of identified individuals whose tissue can be sampled and whose vital rates are estimated (Bowen et al., 2010). Exposure to environmental stressors such as ocean temperature and interactions with fisheries can be characterized for marine mammals using spatiotemporal sampling of parameters such as effort statistics similar to those used in the seabird studies. For example, Caillat and Smout (2015) studied the potential effects of prey availability, grey seal numbers, and exposure to biotoxins on the fecundity and pup survival of harbor seals off the east coast of the United Kingdom. They found that a single (but different) dominant stressor explained the observed variations in each demographic rate. It may be possible to identify interactions between these stressors in other populations that have undergone more dramatic changes in abundance.
The potential for tissues such as baleen whale earplugs or manmade sampling devices to provide a lifetime record of age-specific fertility, age at death, and exposure to some stressors suggests the potential for a new approach to studying the relationship between exposure to stressors and vital rates in marine mammal populations. Given the low probability that long-term studies of vital rates and spatiotemporal mapping of exposure to stressors will provide sufficient data over long enough time intervals for marine mammal populations, we recommend research on natural matrices that may provide a lifetime record of stressors and effects. The development of tags to accomplish the same goal for species without such natural matrices faces significant obstacles but is worth scoping as a potential opportunity for the long term.
Finding 8.3: Natural and artificial matrices have potential as tools for documenting dosage of chemical stressors and changes in hormone levels over long enough time periods to test the relationship between stressor dosage and response in terms of health or vital rates. Natural matrices that are laid down in semiannual layers from birth to death are particularly promising.
Measuring the Lifetime Exposure of an Organism to Stressors
Wild (2005) argued for the importance of tracking exposure of stressors throughout the lifespan. He developed the concept of an “exposome”—defined as the lifetime exposure of an organism to stressors from the prenatal period to death. It is clearly a great challenge to measure the exposome, but a series of papers have emphasized the importance of gathering exposure data on stressors, in both the internal and the external environments, throughout the lifetime (e.g., Lioy and Rappaport, 2011). Rappaport (2011) suggests an approach to measuring the exposome by repeated sampling of blood at critical times of life, with each sample analyzed for “important classes of toxic chemicals, notably, reactive electrophiles, metals, metabolic products, hormone-like substances, and persistent organic compounds.” He argues that as the extent of this sampling increases, economies of scale should create positive feedback for growth of exposome sampling. A similar sampling scheme for accessible marine mammal populations using cross-sectional studies supplemented by individuals sampled throughout their lifespan could help to define combinations of stressors that cause adverse cumulative effects. Longitudinal, spatially comprehensive collection of data on exposure to and effects of multiple stressors could be excessively costly. However, ongoing research studies being funded and/or conducted by multiple federal agencies (e.g., National Oceanic and Atmospheric Administration, U.S. Navy, Bureau of Ocean Energy Management, and the U.S. Geological Survey) and independent researchers could be leveraged and expanded to simultaneously collect samples and conduct analysis to assess exposure to and effects of multiple stressors. The value of a centralized database would be increased with additional information from active surveillance (see Chapter 7).
Health: Vital Rates Function
Most of the health indices discussed in this report can be measured directly for species that can be handled for sampling. The committee has suggested several other approaches for tagging or sampling other matrices in the wild that can be used to assess health. Vital rates can also be estimated directly for species where individuals can regularly be resighted and where birth of the young can be detected reliably. For other species, the committee suggests some new approaches that also include tagging animals with artificial matrices or sampling natural matrices that lay down tissue in layers that can be used for aging and that can store hormones.
The best example of estimating the function relating health to vital rates comes from New et al. (2014), who took advantage of studies of elephant seals on beaches where lipid and lean mass could be measured from pregnant females as they left and returned from foraging trips. Their pups were weighed soon after birth and after weaning. These measurements allowed New et al. (2014) to estimate the energy transferred from mother to pup, and to relate pup natal mass to survival. The relationships between the health variable of body condition, expressed as maternal lipid mass, to the pup’s weaning mass, and between the pup’s weaning mass and the vital rate of pup survival enabled the evaluation of the relationship between health and vital rates for this species.
The committee found no examples of similar studies relating health to vital rates in other marine mammals but does suggest some new approaches that may enable such studies. A major problem for these studies is the long time period required to measure vital rates. The discovery that baleen whale earplugs provide a lifetime record of reproductive hormones for each year of life may enable studies of the vital rate of reproduction, and the age at death can be measured from the earplug, providing age-specific mortality. The earplug has been shown to store the health variables of contaminants and stress hormones, and some tissues that are laid down in layers also provide indications of body condition. If large enough samples of earplugs can be recovered and analyzed for health and vital rates, this could enable a new way to evaluate the relationship between these critical parameters. This is the only shortcut found by the committee for retrospective studies of health and vital rates where one can use tissue from dead animals to study these relationships from birth to death. This possibility is promising enough to justify exploration of other matrices, such as teeth and baleen, that may provide similar timelines of health and vital rates.
Recommendation 8.1: Future research initiatives should support evaluation of the range of emerging technologies for sampling and assessing individual health in marine mammals, and identification of a suite of health indices that can be measured for diverse taxa and that best serves to predict future changes in vital rates. Potentially relevant measures include hormones, immune function, body condition, oxidative damage, and indicators of organ status, as well as contaminant burden and parasite load. New technology for remotely obtaining respiratory, blood, and other tissue samples and for remote assessment (e.g., visual assessment of body condition) should also be pursued.
Establishing baseline values of these parameters and their associations in species will provide critical information for assessing individual and population health. Assessment of health is not only central to the PCoMS model proposed in this report, but comprehensive assessments of stressor exposure and health of a population of marine mammals can also help managers decide when the population is at risk, and help them decide which management actions can most effectively support recovery. Long-term studies of known individuals will be important in this regard. Cross-sectional
sampling and repeated sampling from the same individuals of blood or other tissues during critical life-history phases can help to document dosages and health effects of stressors.
Recommendation 8.2: Agencies charged with monitoring and managing the effects of human activities on marine mammals should identify baselines and document exposures to stressors for high-priority populations. High-priority populations should be selected to include those likely to experience extremes (both high and low) of stressor exposure in order to increase the probability of detecting relationships. This will require stable, long-term funding to maintain a record of exposures and responses that could inform future management decisions. Information on baselines and contextual variables is critically important to interpreting responses.
Recommendation 8.3: Standards for measurement of stressors should be developed along with national or international databases on exposure of marine mammals to high-priority stressors and associated health measures that are accessible to the research community.
Recommendation 8.4: Techniques should be developed that will allow historical trajectories of stress responses to be constructed based on the chemical composition of the large number of baleen whale earplugs and baleen samples in museums or similar matrices in other species. Artificial matrices should be studied for their potential to absorb materials (hormones or chemical stressors) and thereby provide a record of exposures and responses to stressors.
There are opportunities to explore the potential for natural or artificial matrices (that store chemical stressors and hormones over long enough time periods) to test the relationship between exposure to the stressors and response in terms of health or vital rates.
Such techniques with museum samples could provide critical information on the relationships between contaminants, stress, and reproductive intervals. Natural matrices that are laid down in semiannual layers from birth to death are particularly promising.
This page intentionally left blank.