6
Steller Sea Lion Decline: Environmental Context and Compendium of Evidence

Evaluation of the main hypotheses proposed for the causes of decline and failure to recover of the western Steller sea lion population depends on understanding how food web linkages affect sea lions. Every species in an ecosystem, including humans, contributes to and is affected by the linkages represented in a food web. Food webs are described by identifying who eats whom, based on direct observation, stomach and scat analyses, or prey item chemical signatures based on stable isotope and fatty acid analyses. Understanding how linkages influence population and ecosystem dynamics is a far greater challenge because the complexity of interactions precludes analysis through static observation. Only through perturbation of one or more populations is it possible to evaluate the dynamic nature of food webs. Though much is known about the descriptive structure of food webs, the dynamic properties are less well understood. Humans are part of the food web; in the current case, they may change food web dynamics through direct takes of sea lions, removal of sea lions’ preferred prey, removal of alternate prey items of sea lion predators, or some combination of the above. For example, humans depleted populations of several whale species, possibly inducing killer whales to increase their predation on sea lions, seals, and otters.

This chapter describes four concepts that provide a context for analyzing the role of food web interactions in Steller sea lion population dynamics and then applies these concepts to evaluate the many hypotheses proposed to explain the Steller sea lion population decline.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 112
6 Steller Sea Lion Decline: Environmental Context and Compendium of Evidence Evaluation of the main hypotheses proposed for the causes of decline and failure to recover of the western Steller sea lion population depends on understanding how food web linkages affect sea lions. Every species in an ecosystem, including humans, contributes to and is affected by the linkages represented in a food web. Food webs are described by identifying who eats whom, based on direct observation, stomach and scat analyses, or prey item chemical signatures based on stable isotope and fatty acid analyses. Understanding how linkages influence population and ecosystem dynamics is a far greater challenge because the complexity of interactions precludes analysis through static observation. Only through perturbation of one or more populations is it possible to evaluate the dynamic nature of food webs. Though much is known about the descriptive structure of food webs, the dynamic properties are less well understood. Humans are part of the food web; in the current case, they may change food web dynamics through direct takes of sea lions, removal of sea lions’ preferred prey, removal of alternate prey items of sea lion predators, or some combination of the above. For example, humans depleted populations of several whale species, possibly inducing killer whales to increase their predation on sea lions, seals, and otters. This chapter describes four concepts that provide a context for analyzing the role of food web interactions in Steller sea lion population dynamics and then applies these concepts to evaluate the many hypotheses proposed to explain the Steller sea lion population decline.

OCR for page 112
FOOD WEB CONCEPTS Bottom-Up and Top-Down Control Food web linkages connect species across different trophic levels (Paine, 1980). The functional significance of any linkage can be viewed from the perspective of either the consumer or the prey. When a population’s size is limited by the availability of prey, it is described as bottom-up control; when the size of a population is determined by predation, it is described as under top-down control. Bottom-up control characterizes populations that decline or fail to expand because there is insufficient food for growth, reproduction, or survival. Top-down control characterizes populations whose size is regulated by the abundance and feeding habitats of the species that prey on them. Direct and Indirect Food Web Linkages Food web linkages can connect two species directly or indirectly if there are one or more intermediate species. Although hypotheses based on direct effects could be sufficient to explain the decline of Steller sea lion populations, indirect effects may influence the pattern of decline and thereby complicate analysis of direct impacts (see Box 6.1). How the number of intermediate species affects the strength of the interaction is generally unknown, but the number of potential indirect linkages is far greater than the number of direct linkages. Therefore, it is insufficient to consider only the availability of specific prey items or the foraging patterns of generalist predators because changes among other members of the food web may indirectly affect sea lion survival. Humans have exploited many large and small predators in marine ecosystems, acting as an agent for top-down control of marine populations. If top-down forcing is important, the depletion of apex predators should have strong effects on food webs through increased abundance of species at lower trophic levels; removal of predators disrupts the trophic cascade (Paine, 1980; Carpenter and Kitchell, 1993). The consequences of these shifts in marine food webs are complex, difficult to predict, and often unrecognized. Scale and Connectivity Most analyses of food web dynamics have focused on the linkages among species in a common ecosystem. There is growing evidence for the importance of linkages across ecosystems. Linkages of this sort occur in a variety of forms and connect otherwise functionally distinct ecosystems

OCR for page 112
BOX 6.1 Some Plausible Examples of Indirect Effects The Bowen et al. (2001) report presented a number of hypotheses concerning the history of Steller sea lion population trends since 1960 to identify the most informative parameters for guiding future research programs. These hypotheses address direct effects that may explain the decline in Steller sea lions. For example, the discussion of trophic consequences of fishing has focused on reduced food availability. A broader classification of “fishing effects” would also include indirect influences of diverse trophic interactions (Wootton, 1994), estimated to be approximately 50% of all ecological interactions (Schoener, 1993; Menge, 1995). One example, captured in the following quotation from Loughlin and York (2000), is the implication that factory trawlers, simply by their presence and activity, can aggregate sea lions and killer whales. “Predation [by killer whales] is often focused in small areas, i.e., where sea lions are localized near large fish processing vessels, resulting in exacerbation of local declines” (p.43). As a second plausible example, bycatch (nontarget animals caught, killed, or injured during fishing operations) could provide a new source of food for bottom-dwelling organisms, including various flat fish. Arrowtooth flounder had several strong recruitment years in the 1980s; biomass peaked in 1995 (five-fold increase over 1980 biomass) and by the 2001 estimate had declined about 20% due to lower recruitment levels in the 1990s (Wilderbuer and Sample, 2001). The years of high recruitment occurred when the pollock fishery used bottom trawling gear that causes incidental mortality of benthic organisms (National Research Council, 2002). The high bycatch of that fishery may have contributed to the resurgence of arrowtooth flounder. These fish may pose a competitive threat to sea lions because they also prey on young pollock, other demersal fish, and invertebrates. In a third indirect effect scenario, the 10-fold increase in jellyfish over the past decade could also deplete the sea lions preferred prey (Brodeur et al., 1999). Plausible interaction pathways could include direct competition with age 0 pollock for zooplankton or predation on the smaller pollock. This massive increase in jellyfish might constrain the recruitment of many commercially and ecologically significant fish stocks. In the Black Sea, studies have documented the top-down control of commercially valuable fish stocks by a comb jelly (Shiganova, 1998). Further sampling and analysis of these food web linkages in the Bering Sea would be needed to fully evaluate the impacts of this increase in jellyfish on recruitment of important sea lion prey species. Indirect effects of fishing activity—especially if concentrated seasonally and spatially—could account for some of the unexplained mortality in the western population of Steller sea lions. In any event, despite the difficult challenge of unraveling their impacts, indirect effects should not be dismissed as either biologically unusual or dynamically trivial.

OCR for page 112
over a wide range of spatial and temporal scales. Many marine species have dispersive life stages that can be carried great distances by ocean currents. Also, large animal movements can link disparate ecosystems in important ways. The altered foraging behavior of killer whales (to include sea otters in their diet) provides a link between the kelp forest ecosystems of the Aleutian archipelago and the food web of the open ocean (Estes et al., 1998). Previous shifts in the abundance of key species continue to affect the dynamics of present-day food webs. For instance, the progressive removal of herbivorous fishes and invertebrates by historical fisheries from Caribbean coral reefs in conjunction with a mass die-off of sea urchins was likely responsible for declines in coral abundance due to overgrowth by algae (Jackson et al., 2001). Thus, it is possible that the Steller sea lion population decline and failure to recover is in part influenced by events distant in either time or space. Alternative Stable States Because food webs have complex multiple linkages, the response of these systems to disturbance is often nonlinear (Ruesink, 1998). For instance, if a major predator is removed, the increased availability of prey resources may allow expansion of other predatory species to a new, relatively stable equilibrium. This alternative stable state, dominated by a different assemblage of species, may inhibit the return of the food web to its previous status (Lewontin, 1969; Holling, 1973; May, 1977; Sheffer et al., 2001). This concept has important implications for management because of the possibility that disturbed ecosystems may not return to their previous state of equilibrium. Hence, even if the causes of the Steller sea lion decline are identified and addressed, the western sea lion population still may fail to reach its former abundance. MULTIPLE WORKING HYPOTHESES At least eight hypotheses have been proposed to explain the rapid decline of the western stock of Steller sea lions. As pointed out in Box 1.1 (Chapter 1), these various hypotheses cannot be accepted or rejected through the method of strong inference. The data necessary to conduct determinative analyses were simply not collected during the years of the rapid decline. However, this is not to say that relevant data are entirely lacking, particularly with regard to current trends in the population. Numerous types of information on Steller sea lions and their environment have been obtained over the years, some fortuitously and some for the specific purpose of trying to better understand sea lion ecology and popu-

OCR for page 112
lation biology. Although none of this information is sufficient to prove or eliminate hypotheses, much of it can be rated according to its consistency with any given hypothesis. When all of the information is assessed in aggregate, a weight of evidence argument emerges that allows ranking of the hypotheses according to conformity with available information. The main hypotheses that have been proposed to explain the Steller sea lion decline are described in Table 6.1. Each hypothesis is presented separately for the sake of clarity, but this should not be taken to imply that the hypotheses necessarily act independently of each other nor does it preclude the possibility that the recent decline results from a combination of the hypothesized causes. FOOD LIMITATION—BOTTOM-UP HYPOTHESES Under the bottom-up scenario, the Steller sea lion decline is attributed to a deficiency in food resources. This deficiency could be manifested as depletion of prey, reduced abundance of preferred prey species, or reduced accessibility to prey due to local depletion or disturbance of fish stocks. Nutritional limitation caused by either a climate regime shift and/or a fisheries effect requires that either the quantity or quality of food is insufficient for the recovery or maintenance of the Steller sea lion population. This could come about from starvation conditions, nutritional impacts on reproductive success, or increasing susceptibility of animals to disease. During the period of rapid decline in the 1980s, the demographics of the western stock gave some indications that Steller sea lions were nutritionally stressed. In 1985, sea lions were on average smaller, were slower to reach reproductive maturity, and had a lower birth rate than in the 1970s (Calkins and Goodwin, 1988; York, 1994). There was also evidence of higher rates of abortion and lower juvenile survival (Pitcher et al., 1998). Nutritional stress may have been a contributing factor in causing the rapid decline of the western population of sea lions, but models indicate that reduced prey availability alone is unlikely to account for the dramatic decline in the size of the population (see Chapter 3). In 1991 the Alaska Sea Grant College Program (1993) sponsored an international conference entitled “Is It Food?” to ascertain what kind of physiological or biochemical changes would be expected in a chronically or acutely food-stressed pinniped. Because of the difficulty in handling Steller sea lions, the first field studies to address these questions were conducted during the summer on newborn pups and adult females on rookeries. They utilized an east versus west comparative approach with the hypothesis that the declining western population would be stressed relative to the stable eastern population. The studies looked at pup growth

OCR for page 112
TABLE 6.1 Eight Major Hypotheses Proposed to Explain the Steller Sea Lion Population Decline. Each hypothesis is characterized by purported demographic mechanism(s) of population change, food web forcing directions, and the acronyms used later in Table 6.2. Although the cause of the sea lion decline likely falls within this breadth of hypotheses, more than one of the listed hypotheses may have contributed to the decline, additively, interactively, or in various degrees of relative importance in different places or at different times. Hypothesis Mechanism of Population Limitation Forcing Direction Acronym 1. Fisheries removal Starvation and/or reproductive failure because of nutritional limitation Bottom-up FR 2. Climate change/regime shift Starvation and/or reproductive failure because of nutritional limitation Bottom-up CE 3. Predation Elevated mortality from attack by predators Top-down PRED 4. Direct take Elevated mortality from shooting or other purposeful killing Top-down DT 5. Subsistence harvest Elevated mortality from shooting for food or other subsistence uses of sea lions Top-down SH 6. Incidental take/entanglement Elevated mortality from entanglement in fishing gear due to injury or drowning Top-down IT/ENT 7. Disease Elevated mortality or reproductive failure caused by parasites, viruses, or bacteria Top-down D 8. Pollution/biotoxins Elevated mortality or reproductive failure from poisonous or toxic substances, either natural or human produced Top-down or Bottom-up PO rates, maternal attendance patterns, blood chemistry profiles, milk quality, at-sea metabolic rate estimates, thermoregulatory measurements, and several other variables that had been outlined in the “Is It Food?” conference. These studies in the mid-1990s found that animals in the western population were at least as healthy as in the southeastern Alaskan populations

OCR for page 112
based on several measurements of body condition such as birth size, pup growth, and adult size. This research was summarized in the “Is It Food? II” conference in 2001 (see DeMaster and Atkinson, 2002). Conclusions based on these results are limited because of sample size (less than 20 adult females and less than 100 pups), seasonality (they were only conducted in the summer on rookeries), and insensitivity to subtle differences between populations. Despite these limitations, the studies suggest that it is unlikely that newborn pup survival has been compromised by acute or chronic malnutrition over the past decade. These studies have now been expanded to juveniles on a year-round basis because new capture methods allow large numbers of juveniles to be handled. All preliminary evidence shows similar results: sea lions in the western population show no indication of being nutritionally stressed relative to sea lions in the eastern population (Richmond and Rea, 2001). The consensus statement drafted from the “Is It Food? II” conference (Alaska Sea Grant College Program, 1993) states that nutritional limitation is probably not a major contributor to the population decline over the past 10 years. Additional studies on animals from a variety of locations would be necessary to establish whether these results apply generally to Steller sea lions throughout the western range. The following sections describe the two mechanisms proposed to cause a decrease in the availability or quality of the food supply for Steller sea lions throughout the history of the decline. These two mechanisms, climate regime shifts and fishery removals, may have had a combined effect that limited the availability of common Steller sea lion prey items during the earlier phases of the decline. Climate Regime Shift The regime shift hypothesis links climate-forced environmental changes to changes in the welfare of Steller sea lions through indirect trophic interactions in the marine food web. Several different mechanisms have been proposed that link climatic regime shifts to declines in Steller sea lions in the 1970s and 1980s. A reduction in the abundance of herring, capelin, and sand lance and a concomitant increase in large piscivorous fish associated with the 1977 climatic regime shift (see Chapter 2) may have adversely affected Steller sea lions by reducing the proportion of high-calorie fish in their diet (see discussion of the junk food hypothesis below). Merrick et al. (1997) showed that declines in sea lion populations correlate with a decrease in sea lion dietary diversity, which may be indicative of a change in the availability of prey species (Anderson and Piatt, 1999).

OCR for page 112
The regime shift hypothesis largely rests on statistical inference, wherein correlation analyses identify statistically significant associations between many 20th-century climate, fishery, and ecosystem survey records across the broad geography of the North Pacific and Bering Sea (e.g., see Anderson and Piatt, 1999; Hare and Mantua, 2000). Potential mechanisms linking climate changes to ecosystem changes in the North Pacific and Bering Sea are reviewed in Chapter 2. Testing the regime shift hypothesis is essentially limited to a “wait and see” approach that cannot distinguish between the impacts of natural environmental changes and other perturbations. Based on a recent shift to cooler upper-ocean temperatures and a weakening of the wintertime Aleutian Low beginning in 1998, the climate may have shifted to a “cool phase” Pacific Decadal Oscillation (PDO) state similar to what existed before the steep decline in Steller sea lion populations (Hare and Mantua, 2000; Schwing and Moore, 2000; Peterson and Mackas, 2001). If PDO regime shifts exert a significant and reversible forcing on sea lion abundance, the western population should begin to recover in response to the 1998-2002 climate trends. Fishery Removals The spatial and temporal scales of commercial fisheries provide important insights for evaluating the possible effects of fishery removals on the nutritional status of Steller sea lions. Fisheries and their potential interactions with Steller sea lions were discussed in detail in Chapter 5. This section first evaluates whether fisheries have depleted prey resources at a regional scale on an interannual basis to the extent that there is insufficient fish biomass to sustain the extant number of sea lions. The second part of this section considers the potential for fisheries to deplete sea lion prey at a local scale. Evidence for Broad-Scale Depletion Fishery and stock assessments indicate that walleye pollock (except the Donut Hole stock), Pacific cod, and Atka mackerel stocks are not overfished (North Pacific Fishery Management Council, 2001a, 2001b). Periodic strong year classes drive much of the change in fish stock abundance. Although parental abundance affects recruitment at low stock levels, some if not most of the recruitment variability in groundfish stocks in the North Pacific Ocean appears to be associated with variability in environmental conditions. Regime shifts toward winters with a deepened Aleutian Low Pressure System tend to be associated with higher frequencies of strong year classes among groundfish stocks (Hollowed and Wooster, 1992, 1995).

OCR for page 112
Are there enough fish to support a healthy population of Steller sea lions? This is a difficult question to answer without making many assumptions, but it addresses only the simplest consequence of the fishery. The more comprehensive question is whether the appropriate species, sizes, and densities of prey are available at spatial and temporal scales necessary for foraging sea lions. As will be shown, this is also a difficult question to answer given the poor state of knowledge about sea lion foraging ecology, fine-scale distributions of fishes, and effects of fishing on fish school dynamics. In the 2000 Biological Opinion (BiOp #3), the National Marine Fisheries Service attempted to determine whether there was sufficient groundfish prey for Steller sea lions by calculating the amount of food consumed by sea lions relative to the biomass of the groundfish in the Gulf of Alaska, Aleutian Islands, and Bering Sea in 1999. This yielded a ratio of biomass consumption to availability of 1:54. Similarly, the agency deduced that a historical high number of 184,000 Steller sea lions would consume about 1.7 million metric tons (mt) annually, for a ratio of consumption to availability of 1:21. This comparison indicated that the current availability of fish biomass is higher for the 1999 population than for the prefishery population of sea lions. Another approach for examining simple food availability is to compare trends in pollock, Pacific cod, and Atka mackerel biomass with Steller sea lion counts. Figure 6.1 shows estimates of exploitable fish biomass. Steller sea lion counts correspond to index sites in the Gulf of Alaska and Aleutian Islands, as there are no index rookeries in the Bering Sea. Also, analyses of sea lion counts were restricted to years in which at least 24 or 35 index rookeries were observed in the Gulf of Alaska and Aleutian Islands, respectively; counts from years with fewer sites tended to be inconsistent with counts from adjacent years. In the Gulf of Alaska, sea lion numbers declined from the 1950s through the 1970s, a period during which pollock abundance was increasing (Figure 6.1a). The most rapid decline of sea lions occurred from 1977 to 1985, when pollock landings peaked in the Gulf of Alaska (Figure 5.2) and there was a large increase in the percentage of groundfish taken in Steller sea lion critical habitats (Figures 5.10 and 5.11). Trawl surveys in 1984-1996 did not show any decrease in high-density pollock abundance. However, many sea lions were taken as bycatch in the fishery. During this same time period in the 1980s, Pacific cod abundance increased (Figure 6.1c). Abundance of all three species declined from the late 1980s to the present. However, regardless of whether one considers pollock biomass alone or combined pollock and cod biomass, there have been more of these fish species available per Steller sea lion since the mid-1980s than prior to 1980 (see Figure 6.2a).

OCR for page 112
In the Bering Sea and Aleutian Islands, Steller sea lion counts were high from the mid-1960s to the late 1970s, at a time when eastern Bering Sea pollock abundance appears to have been low to moderate (Figure 6.1b). Steller sea lions declined sharply in the 1980s at a time when pollock, cod, and Atka mackerel abundances were increasing (Figure 6.1b, d, e). Interestingly, pollock biomass apparently declined during 1989-1992, whereas the Steller sea lion decline abated during 1990-1992 in the Aleutian Islands. As with the Gulf of Alaska, the biomass of pollock and the combined biomass of pollock, cod, and Atka mackerel per sea lion were higher after 1985 than prior to 1980 (Figure 6.2b). As mentioned earlier in Chapter 5, these comparisons are fraught with assumptions and therefore should be interpreted with caution. Several factors tend to lead to overestimation of the fish biomass available to Steller sea lions. For example, in this analysis the committee chose index sites because this subset of rookeries provides the best measure of trends over time. However, by definition, these indices underestimate the total abundance of Steller sea lions because they do not include counted animals on nonindex sites and uncounted animals at sea during the surveys. Additionally, only index sites in the Gulf of Alaska within the range of the western stock of Steller sea lions were considered. Increases in the abundance of Steller sea lions in southeastern Alaska (eastern stock) were not included, although fish biomass estimates are gulf-wide values. Moreover, pollock in the northwestern portion of the eastern Bering Sea may not be available to sea lions, but the values represent total exploitable pollock biomass over the entire continental shelf. Other assumptions of this analysis contribute to underestimation of prey biomass available to Steller sea lions. For instance, biomass estimates of Aleutian Islands pollock were not included because of questions about the discreteness of this stock. In 2000 a bottom-trawl survey estimated 105,500 mt of pollock in the Aleutian Islands (Ianelli et al., 2001). Survey trends indicate that pollock abundance in this area peaked in 1983, declined until 1994, and increased since then. Likewise, pollock biomass in the Aleutian Basin and Bogoslof Island areas were not considered in our analysis. It also should be noted that Steller sea lions target juvenile pollock (Merrick and Calkins, 1996), yet young pollock are not fully sampled by the survey gear and tend to be underestimated in stock assessments. Finally, our estimates do not include other components of the sea lion diet, including fish (e.g., salmon, herring, flatfishes, rockfishes, sand lance, capelin) and invertebrates (e.g., octopus, squid). Given these caveats, this analysis does not provide support for the hypothesis that the recent decline in the Steller sea lion population is due to depletion of sea lion prey by the groundfish fisheries.

OCR for page 112
FIGURE 6.1 Trends of Steller sea lion (SSL) index counts in the Gulf of Alaska (GOA) versus exploitable biomass of (a) GOA pollock and (b) GOA Pacific cod. Trends of SSL index counts in the Aleutian Islands versus exploitable biomass of (c) Bering Sea/Aleutian Islands (BS/AI) pollock, (d) BS/AI Pacific cod, and (e) BS/AI Atka mackerel. Index counts are for those years in which a minimum of 24 rookeries in the Gulf of Alaska or 35 rookeries in the Aleutian Islands were observed. SOURCE: Data from National Marine Fisheries Service, National Marine Mammal Laboratory, Seattle, available at www.afsc.noaa.gov/.

OCR for page 112
discharge firearms within 100 yards of Steller sea lions since 1990, the amount of shooting in recent years is even less well known, but some anecdotal evidence suggests that the practice persists. Hence, the “smoking gun” in the mystery of the Steller sea lion population decline could, at least in part, be the unreported and illegal takes. INFECTIOUS DISEASE Theoretical studies suggest infectious diseases may regulate host abundance by exerting density-dependent effects on host reproduction or survival (Anderson, 1979). However, few empirical studies on free-living animals have determined whether effects of disease on populations are density dependent, as most disease investigations on wildlife have focused on determining the proximate causes of large die-offs. Such die-offs are most often the result of epidemics of disease in host populations that have not previously been exposed to the disease or that have become more susceptible to an introduced agent due to changes in immune status. Although not generally considered regulators of host population density, severe epidemics may reduce host population density to such an extent that stochastic events or previously unimportant ecological factors may further reduce the host population size (Harwood and Hall, 1990). The importance of infectious disease epidemics in causing declines of marine mammal populations is unclear because few die-offs have been investigated sufficiently to determine their cause, and it is often difficult to accurately determine host population numbers. Recent epidemics in marine mammals have caused dramatic mortality, but the effects on host population numbers vary. For example, approximately 17,000 harbor seals (70% of the population) died in the phocine distemper (PDV) epidemic in Europe in 1988, but 10 years later the population had recovered to pre-epidemic numbers (Reijnders et al., 1997). The source of infection for this morbillivirus outbreak in marine mammals is unclear. The 1988 PDV epidemic was believed to have resulted from the introduction of a virus into a naïve population, due to the large numbers of animals without antibodies prior to the outbreak. Based on this assumption, a mathematical model investigating the infection dynamics of this disease in 1992 predicted that reintroduction of the virus resulting in large-scale mortality would not occur for at least 10 years (Grenfell et al., 1992). A new outbreak of PDV is currently occurring in the North Sea (Jensen et al., 2002; cwss.www.de/news/news/Seals/01-seal-news.html). In 1988 it appeared the mortality was lower in seal populations that had experienced previous mortality events (Grenfell et al., 1992). Endemic diseases have more subtle effects and may be more important in regulating marine mammal populations than previously thought.

OCR for page 112
Infectious diseases such as brucellosis are known to cause population declines in terrestrial mammals due to spontaneous abortion and reproductive failure. Although these types of disease organisms have only recently been isolated from marine mammals, they may be prevalent in free-ranging populations (Dunn et al., 2001). Food limitation may increase the impact of macroparasites, resulting in population crashes, as observed in Soay sheep populations in Scotland (Gulland, 1992). Even if only a few individuals show signs of infection, it is possible that parasitic nematodes have a significant influence on the host population size (Hudson and Dobson, 1995). There are a wide variety of parasitic nematodes in marine mammals, but their effects on host population dynamics are unknown (Dailey, 2001). A number of infectious disease agents known to occur in marine mammals of the Arctic and Pacific could cause epidemics, endemic mortality, or decreased reproduction in Steller sea lions. These include morbilliviruses, influenza virus, phocine herpesviruses, caliciviruses, Leptospira spp., Brucella spp., Chlamydia psittaci, Toxoplasma spp., and various species of nematodes (Dierauf and Gulland, 2001). If an epidemic hit a population of Steller sea lions, it would be expected to spread from one area, cause mortality in animals of all ages, and leave survivors with antibodies to the causative agent (Heesterbeek and Roberts, 1995). Although the rapid decline of sea lions is consistent with the first two of these conditions, to date, no antibodies have been detected to morbilliviruses or influenza virus—the two viruses most likely to cause such an epidemic (Burek et al., 2001). Antibodies to these viruses are also absent from sea otters (Enhydra lutris) and northern fur seals (Callorhinus ursinus) that share waters around the Aleutians with Steller sea lions and are susceptible to these diseases (Hanni et al., in press; Terry Spraker, Colorado State Veterinary School, Fort Collins, personal communication, 2000). Thus, it is unlikely, but not impossible, that a viral epidemic caused the rapid decline of Steller sea lions during the 1980s. If morbillivirius did cause an undetected epidemic, however, experience from the European harbor seal epidemics suggests that it would not continue to affect population size in the absence of other factors and that the population should show signs of recovering to pre-epidemic numbers. To definitively eliminate the hypothesis that a viral epidemic caused the rapid decline of Steller sea lions would require more comprehensive tests of banked serum samples and molecular tests for the presence of disease agents in preserved tissue samples. Endemic diseases could inhibit recovery of the Steller sea lion population. Antibodies to phocid herpesvirus, caliciviruses, and leptospires have been detected in Steller sea lions (Barlough et al., 1987; Calkins and Goodwin, 1988; Zarnke et al., 1997; Burek et al., 2001). All of these agents

OCR for page 112
cause mortality or reproductive failure in other marine mammal species, but their effects on Steller sea lions have not been documented because neither sick nor dead animals have been available for examination. In other marine mammals, these endemic diseases cause reproductive failure through abortions or poor pup survival. Higher levels of haptoglobin, a nonspecific indicator of inflammation, have been detected in Steller sea lions in declining populations (Zenteno-Savin et al., 1997). Inflammation is often a consequence of infectious disease; therefore, higher haptoglobin levels could be an indication that declining populations of sea lions have higher rates of infection. The prevalence of these disease organisms in the Steller sea lion population is unknown. Although antibodies to some disease agents have been found in sera, few studies have confirmed the presence of the disease-causing agent in tissue samples. Chlamydia spp. was observed in tissues of an aborted fetus (Spraker and Bradley, 1996), and antibodies to this organism are widespread in adult Steller sea lions, especially females (Burek et al., 2001). Further studies should be directed at determining the prevalence of infection in Steller sea lion populations and the effect of infection on reproduction and survival. Macroparasites such as nematodes, flukes, and tapeworms are common in pinnipeds and may cause mortality in malnourished animals. California sea lions (Zalophus californianus) that suffer food deprivation or feed on unusual prey species during El Niños have heavier parasite burdens than animals in other years and can die from parasitic ulcers (Fletcher et al., 1998). These macroparasites have been found in Steller sea lions, possibly compounding the effects of malnutrition and increasing juvenile mortality. Little is known about the species of parasites or their prevalence and intensity of infection in Steller sea lions. Future studies should identify macroparasites, determine the prevalence and intensity of infection, and determine whether infection intensity correlates with nutritional status. In conclusion, little is known about the prevalence of infectious diseases in Steller sea lions or their morbidity. Both eastern and western populations of Steller sea lions have antibodies to agents that could decrease survival and reproduction. The prevalence and intensity of infections need to be assessed to determine whether they play a role in the decline of Steller sea lion populations. Although a viral disease could have occurred in the 1980s, to date there is no direct evidence of an epidemic. TOXINS Biotoxins produced by harmful algal blooms have caused episodic mortality in a number of marine mammal populations around the United States, from manatees (Trichecus manatus) off Florida to California sea

OCR for page 112
lions off California (Bossart et al., 1998; Scholin et al., 2000). A bloom of saxitoxin-producing algae is believed to have caused a die-off of about 50% of the Mediterranean monk seal population, although controversy still surrounds this event due to the lack of fresh carcasses for examination (Hernández et al., 1998). Harmful algal blooms producing toxins such as domoic acid occur in Steller sea lion foraging habitats, but their effect on these animals is unknown because no carcasses have been found or examined. It is unlikely that a large mortality event occurred, however, because these toxins cause dramatic clinical signs that would have been readily detected. In addition, mortality of other species (including fish and birds) usually occurs, and these were not reported in areas of sea lion declines. Furthermore, to account for the pattern of sea lion decline, the bloom of toxic algae would have to spread from the central Gulf of Alaska to the western Aleutians but not to southeastern Alaska. It is unlikely that such an event would have gone unnoticed. However, retrospective analysis of stored tissues for biotoxins would be necessary to completely rule out the possibility that algal toxins contributed to the rapid decline of Steller sea lions. There is an extensive literature on the effects of toxic contaminants on mammalian reproduction (reviewed by O’Hara and O’Shea, 2001). There are also data on the levels of a number of elements in marine mammal tissues (e.g., cesium, cadmium, mercury, selenium) and persistent organic compounds such as polychlorinated biphenyls (PCBs) and dichloro-diphenyl-trichloro-ethane (DDT), but few data exist on dose-response effects even for well-known contaminants in marine mammals. It is thus not possible to determine whether the levels of contaminants measured in tissue samples affect the survival of Steller sea lions. The levels of some xenobiotic compounds have been determined in a limited number of Steller sea lion tissues. PCB and DDT levels in blubber of sea lions sampled between 1976 and 1981 in the Bering Sea were lower than in sea lions from the Gulf of Alaska (Lee et al., 1996). Levels were lower in females than males, as occurs in most marine mammal species, due to the lactational transfer of lipophilic toxins. Levels of PCBs and DDTs were higher in Steller sea lions than in ringed and harp seals from Arctic waters but were comparable to levels in gray seals from the east coast of Canada and lower than in California sea lions with normal gestation periods. Both gray seals and California sea lion populations are currently increasing. It is thus unlikely that the contaminant levels in Steller sea lions are causing direct mortality in this species, although more subtle effects on physiology could occur. There may be species-specific effects, and combinations of contaminants may have more deleterious effects than single compounds. Thus, it is not possible to eliminate the possibility that contaminants affect

OCR for page 112
the physiology of Steller sea lions by measuring a few compounds in blubber at any one time during development. Positive associations between organochlorine burdens and reduced immune function have been observed in harbor seals, but the overall effect of the health of the population is still unclear (deSwart et al., 1994). If contaminants were causing immunosuppression in Steller sea lions, an increase in prevalence and susceptibility to infectious disease should be observed in declining populations, but these epidemiological observations are lacking. Differences in contaminant burden have been inferred from fecal levels of PCBs in Steller sea lions that could result from regional differences in the prey population (Beckmen et al., 2001). Because there has been considerable military activity in the Aleutians, it is possible that certain sites have localized contamination with unidentified compounds. Estimates of vital rates of Steller sea lions in different locations may uncover differences in local mortality and reproduction that are indicative of toxic contamination. Further epidemiological studies focusing on associations between contaminant levels in tissues of individuals and life history parameters, coupled with determination of the significance of reproductive failure and infectious disease in the dynamics of Steller sea lions are needed to determine whether contaminants could play a role in limiting sea lion recovery. WEIGHT OF EVIDENCE Steller sea lion behavior, physiology, life history, and environment can be analyzed with regard to how they would be expected to change under each of the eight hypotheses described above using a simple positive or negative response variable (Bowen et al., 2001). In the analysis presented below, the disease category is divided into epidemics and endemics because these two processes are expected to produce different responses in sea lions. Biotoxins are included in the epidemic disease category because the response variables should change in the same direction. As discussed in the Bowen et al. review, the directional changes represent the most likely responses based on current information but do not represent predictions. Use of this analysis should be tempered by consideration of the following: For some response variables the direction of change under a given hypothesis is debatable. Change may depend on whether the effect is size selective, is local or regional, or reduces performance rather than increases mortality (e.g., disease, pollution).

OCR for page 112
The magnitude of change may vary with the intensity of the effect. Hence, it is possible to have a false correlation if an effect appears to be consistent but is too minor to affect population size. Data used to determine the direction of each response variable may be biased due to sampling errors such as age or size-specific effects, local effects generalized to the whole population range, or time-frame dependence (what is observed in one year becomes generalized over a decade). In analyzing the weight of evidence for each hypothesis, the committee assumes that each variable changes as the result of direct interaction. The various hypotheses listed in Table 6.1 can be characterized as either bottom-up or top-down forcing mechanisms (e.g., Hunter and Price, 1992). Bottom-up forcing hypotheses are defined by their impact on sea lion food availability and include biomass removed by fisheries, climate change, and regime shifts. Top-down forcing hypotheses are defined by their impact on survival (assuming food is not limiting) and include disease, predation, and human killing of all kinds. Pollution has impacts that could reflect either top-down or bottom-up effects. Grouping hypotheses simplifies the task of pattern assessment because the response variables are predicted to change in a consistent manner depending on whether the direction of forcing is bottom up or top down. For example, foraging time should increase when per capita food availability decreases because prey has been depleted by fishery removals or productivity has changed due to a climate regime shift. Foraging time should decrease as per capita food availability increases if predation or human takes reduce the size of the sea lion population relative to the prey base. The committee extended this analysis by including the available observational evidence for comparison with the expected direction of change (Table 6.2). The directionality of the observed response is determined by comparison of sea lions in the western population with sea lions in the eastern population. For this purpose, characteristics of sea lions in the eastern population are assumed to be representative of the western population prior to the start of the decline. The rationales for the expected response and data sources for the observed response are described in Box 6.4 Table 6.2 lists relevant behavioral, physiological, and life history metrics of Steller sea lions, and features of the associated ecosystem for which data are available. The expected directions of change were derived from observations or ecological models as first described in the “Is It Food?” workshop (Alaska Sea Grant College Program, 1993). Eberhardt (1977) used similar reasoning to provide a general framework for assess-

OCR for page 112
TABLE 6.2. Matrix of Expected/Observed Directional Changes in the Response Variables Under Hypotheses Proposed to Explain the Decline of the Western Steller Sea Lion Stock. The response variable changes are given as either higher (H) or lower (L), for example under a bottom-up forcing, the predicted impact of fishery removals is lower (L) birth mass. However, higher (H) birth masses have been observed. Therefore the entry in the table is L/H. Matches between observed and expected (excluding unknowns) are in bold. Forcing Direction Bottom-up Top-down Uncertain Correlate/Response Variable FR CE PRED DT SH IT/ENT D PO Pups   Birth mass (1) L/H* L/H* H*/H* H*/H* H*/H* H*/H* U/H* U/H* Pup growth rate (1,2a,3a) L/H L/H H*/H H*/H H*/H H*/H U/H U/H Adult female   Body condition (4,5,6) L/H L/H H*/H* H*/H H*/H H*/H L/H L/H Foraging trip duration (7,8) H/L H/L L/L L*/L L*/L L*/L L/L L/L Dive depth (9) H/L H/L L/L L*/L L*/L L*/L L/L L/L Field metabolic rate (9) H/L H/L L*/L L*/L L*/L L*/L L/L L/L General   Foraging range (10) H/L H/L L/L L/L L/L L/L L/L L/L Beach strandings (11) H/L H/L L/L U/L L/L L/L H/L H/L Other piscivores (12) L/NC L/NC NC/NC NC/NC NC/NC NC/NC NC/NC U/NC Food availability (10) L/H L/H H/H H/H H/H H/H H/H H/H FR—fishery removals PO—pollution CE—climate/regime shift PRED—predation DT—direct take (shooting) SH—subsistence harvest IT/ENT—incidental take/entanglement D—disease H—higher H*—higher or no change L—lower L*—lower or no change NC—no change U—uncertain aBased on pup mass at one month. SOURCES: Data came from (1) Brandon and Davis (1999); (2) Merrick et al. (1995); (3) Rea et al. (1998); (4) Davis et al. (1996); (5) Adams (2000); (6) Mike Castellini, University of Alaska, Fairbanks, personal communication, 2002; (7) Brandon (2000); (8) Milette (1999); (9) Andrews et al. (2002); (10) National Marine Fisheries Service (2001); (11) Calkins et al. (1998, p. 241); and (12) Dragoo et al. (2000). Table was modified from Bowen et al. (2001).

OCR for page 112
BOX 6.4 A Brief Rationale for the Directions of Population Change inTable 6.2 Birth mass—Reduced food availability is expected to limit a female’s ability to expend energy on her offspring. Bottom-up forcing mechanisms should thus lead to a reduced birth mass, while top-down forcing mechanisms, if per capita food availability is not limiting, should lead to increased birth mass. The directional effects of disease and environmental pollutants are indeterminate: if these factors increase morbidity, the outcome (fewer sea lions and less competition) would resemble a top-down mechanism, but if they reduce fitness, the outcome (less energy available during gestation), would resemble a bottom-up mechanism. Comparisons of pup birth mass in declining rookeries with birth weights in stable or increasing rookeries in southeastern Alaska during the 1990s have shown that birth masses are similar (Brandon and Davis, 1999). Pup growth rate—An argument similar to that made for expected changes in birth mass. Pup growth rates (Brandon and Davis, 1999) and masses at 1 month (Merrick et al., 1995; Rea et al., 1998) were higher in the western population. Body condition—An argument similar to that made for expected changes in birth mass, except that disease and environmental contaminants should lead to reduced body condition. Adult females in the west have been observed to have greater mass and more body fat (Davis et al, 1996; Adams, 2000; Michael Castellini, University of Alaska, Fairbanks, personal communication, 2002). Foraging trip duration—As per capita food availability declines, the time needed to obtain nutritional resources necessary for maintenance and reproduction should increase. Therefore, bottom-up forcing scenarios suggest increased foraging trip duration, while top-down forcing mechanisms, by increasing per capita food availability, suggest reduced foraging trip duration. Foraging trip length for females with pups was lower at western rookeries in the mid-1990s (Brandon and Davis, 1999; Andrews et al., 2002). Dive depth—Diving is costly in terms of time and energy expenditure and, in the case of deep dives, is the less efficient metabolic pathway associated with anaerobic diving. Thus, sea lions should dive no deeper than necessary to obtain the nutritional resources needed for maintenance and reproduction. Increased dive depths are consistent with bottom-up forcing scenarios, and reduced dive depths are consistent with top-down forcing mechanisms, both through effects on per capita food availability. Foraging range—An argument similar to that made for expected changes in foraging trip duration and dive depth. Measurements of foraging range in 1997 at a rookery in the central Aleutians and one in southeastern Alaska indicate that foraging range is lower in the west (Andrews et al., 2002).

OCR for page 112
Field metabolic rate—As per capita food availability declines, the search and pursuit times required to obtain nutritional resources needed for maintenance and reproduction should increase—hence, an argument similar to that made for expected changes in foraging behavior. Andrews et al. (2002) measured metabolic rate at the same rookeries as described above and found a lower rate in the Aleutians. Beach strandings—Expected changes in the number of stranded or moribund sea lions vary less consistently across categories of forcing mechanisms and in some cases are contingent on a variety of important details. Most forms of food limitation should lead to increased numbers of weakened and emaciated animals onshore. Predation by killer whales (assuming that attacks are rarely unsuccessful and that all or most of the carcasses are consumed) produce no or minimal visual evidence. In contrast, shark predation is associated with diagnostic scars on pinnipeds in areas where shark attacks have been observed. If shooting occurs on- or near shore, elevated numbers of stranded carcasses would be expected, even though sea lions normally sink after being shot and killed. Shooting farther from shore might result in the carcasses sinking or decomposing before they drift ashore. These same arguments apply to subsistence harvest and incidental losses in fishing gear. Eventual mortality from wounds inflicted during a subsistence hunt might lead to elevated numbers of beached carcasses, whereas animals dying from entanglement in fishing gear would likely be too far from shore for this to happen. Other piscivores(surface feeding and diving piscivorous seabirds)—Since piscivorous seabirds feed on earlier life stages of many of the same prey as Steller sea lions, their populations also would be expected to decline under a bottom-up forcing scenario. Most top-down forcing scenarios, in contrast, predict no change in seabird populations. Killer whales or sharks in the northern hemisphere do not consume seabirds, at least so far as is known. A subsistence harvest for pinnipeds should have little effect on seabird populations. The effects of gear entanglement and incidental takes in fisheries is less certain due to the substantial differences in body size between seabirds and marine mammals. Nonetheless, the most likely direct impacts of significant incidental mortality on seabirds in fishing gear are reduced populations. Disease in Steller sea lions would be expected to leave seabird populations unchanged. The expected effects of environmental contaminants are less certain—direct effects on sea lions should leave seabird populations unchanged, whereas effects on the prey base might cause both to decline. Food availability—Prey abundance should decline under the bottom-up forcing scenarios and increase under the top-down scenarios. The effect of environmental contaminants on food availability is unknown.

OCR for page 112
ing marine mammal population status. Box 6.4 provides a brief rationale for the directions of change given in Table 6.2. SYNOPSIS If the rationale for the expected patterns of decline (Box 6.4) is generally correct, the weight of evidence for causality of the Steller sea lion decline since 1990 is most consistent with a top-down forcing scenario. Predation and human-induced mortality provide a good fit to the available data. Currently available assay data for disease and contaminants do not indicate additional mortality from these factors. The virtual absence of beach strandings at first appears inconsistent with unreported shooting. However, it is known that most Steller sea lions sink after being shot in the water. In their study of sea lion stomach contents, Imler and Sarber (1947) shot at least 20 animals in the water and only one (<5%) floated. Likewise, in another feeding study, Fiscus and Baines (1966) killed 34 animals in the water and 23 (68%) sank before they could be recovered. It is possible that the continuing decline in the population is caused by a combination of mortalities from killer whale predation, illegal shooting, incidental takes in fishing gear, and subsistence harvest. Evidence gathered in the 1990s is generally incompatible with a bottom-up forcing scenario. Indicators of sea lion health and foraging behavior suggest that the western population is not food limited when compared to the stable, slowly increasing population in southeastern Alaska. Indirect effects of fishing or other ecosystem changes may influence sea lion population trends (Box 6.1), but support for these potential mechanisms would require a more in-depth understanding of food web interactions in the region. During the rapid decline of the sea lion population in the 1980s, some studies suggested a decrease in sea lion fitness consistent with a bottom-up forcing scenario. Although groundfish fishery yields were high during this period, the biomass of these species was also high. Hence there was no global reduction in the amount of groundfish available to sea lions. The possibility remains that local depletion of some fish stocks, such as Atka mackerel, may have occurred in Steller sea lion habitat, but there is less support for local depletion of pollock stocks. Changes in the abundance of forage fish species related to the regime shift in the late 1970s may have affected sea lion fitness, but these effects do not appear sufficient to account for the large mortality of sea lions in the 1980s. Although a disease epidemic could explain a rapid drop in population, to date there is no indication of such an event based on immunological analysis of serum samples. Similarly, available data do not support widespread mortality from biotoxins or contaminants. Top-down sources of mortality

OCR for page 112
also apply to the earlier phase of the decline, and in the case of human takes, the levels are known to be higher before 1990 than in the most recent decade. Shooting of sea lions was legal to protect fishing gear, and high levels of bycatch mortality were reported in the joint-venture fisheries during the 1980s. Also, it is possible that predation by killer whales was higher than previously estimated. Therefore, even during the rapid decline of the western population, it is likely that a combination of top-down and bottom-up forcing mechanisms were responsible for the high mortality of Steller sea lions.