National Academies Press: OpenBook

The Bering Sea Ecosystem (1996)

Chapter: 6 Causes and Effects in the Bering Sea Ecosystem

« Previous: 5 Human Use - Fisheries
Suggested Citation:"6 Causes and Effects in the Bering Sea Ecosystem ." National Research Council. 1996. The Bering Sea Ecosystem. Washington, DC: The National Academies Press. doi: 10.17226/5039.
×
Page 196
Suggested Citation:"6 Causes and Effects in the Bering Sea Ecosystem ." National Research Council. 1996. The Bering Sea Ecosystem. Washington, DC: The National Academies Press. doi: 10.17226/5039.
×
Page 197
Suggested Citation:"6 Causes and Effects in the Bering Sea Ecosystem ." National Research Council. 1996. The Bering Sea Ecosystem. Washington, DC: The National Academies Press. doi: 10.17226/5039.
×
Page 198
Suggested Citation:"6 Causes and Effects in the Bering Sea Ecosystem ." National Research Council. 1996. The Bering Sea Ecosystem. Washington, DC: The National Academies Press. doi: 10.17226/5039.
×
Page 199
Suggested Citation:"6 Causes and Effects in the Bering Sea Ecosystem ." National Research Council. 1996. The Bering Sea Ecosystem. Washington, DC: The National Academies Press. doi: 10.17226/5039.
×
Page 200
Suggested Citation:"6 Causes and Effects in the Bering Sea Ecosystem ." National Research Council. 1996. The Bering Sea Ecosystem. Washington, DC: The National Academies Press. doi: 10.17226/5039.
×
Page 201
Suggested Citation:"6 Causes and Effects in the Bering Sea Ecosystem ." National Research Council. 1996. The Bering Sea Ecosystem. Washington, DC: The National Academies Press. doi: 10.17226/5039.
×
Page 202
Suggested Citation:"6 Causes and Effects in the Bering Sea Ecosystem ." National Research Council. 1996. The Bering Sea Ecosystem. Washington, DC: The National Academies Press. doi: 10.17226/5039.
×
Page 203
Suggested Citation:"6 Causes and Effects in the Bering Sea Ecosystem ." National Research Council. 1996. The Bering Sea Ecosystem. Washington, DC: The National Academies Press. doi: 10.17226/5039.
×
Page 204
Suggested Citation:"6 Causes and Effects in the Bering Sea Ecosystem ." National Research Council. 1996. The Bering Sea Ecosystem. Washington, DC: The National Academies Press. doi: 10.17226/5039.
×
Page 205
Suggested Citation:"6 Causes and Effects in the Bering Sea Ecosystem ." National Research Council. 1996. The Bering Sea Ecosystem. Washington, DC: The National Academies Press. doi: 10.17226/5039.
×
Page 206
Suggested Citation:"6 Causes and Effects in the Bering Sea Ecosystem ." National Research Council. 1996. The Bering Sea Ecosystem. Washington, DC: The National Academies Press. doi: 10.17226/5039.
×
Page 207
Suggested Citation:"6 Causes and Effects in the Bering Sea Ecosystem ." National Research Council. 1996. The Bering Sea Ecosystem. Washington, DC: The National Academies Press. doi: 10.17226/5039.
×
Page 208
Suggested Citation:"6 Causes and Effects in the Bering Sea Ecosystem ." National Research Council. 1996. The Bering Sea Ecosystem. Washington, DC: The National Academies Press. doi: 10.17226/5039.
×
Page 209
Suggested Citation:"6 Causes and Effects in the Bering Sea Ecosystem ." National Research Council. 1996. The Bering Sea Ecosystem. Washington, DC: The National Academies Press. doi: 10.17226/5039.
×
Page 210
Suggested Citation:"6 Causes and Effects in the Bering Sea Ecosystem ." National Research Council. 1996. The Bering Sea Ecosystem. Washington, DC: The National Academies Press. doi: 10.17226/5039.
×
Page 211
Suggested Citation:"6 Causes and Effects in the Bering Sea Ecosystem ." National Research Council. 1996. The Bering Sea Ecosystem. Washington, DC: The National Academies Press. doi: 10.17226/5039.
×
Page 212
Suggested Citation:"6 Causes and Effects in the Bering Sea Ecosystem ." National Research Council. 1996. The Bering Sea Ecosystem. Washington, DC: The National Academies Press. doi: 10.17226/5039.
×
Page 213
Suggested Citation:"6 Causes and Effects in the Bering Sea Ecosystem ." National Research Council. 1996. The Bering Sea Ecosystem. Washington, DC: The National Academies Press. doi: 10.17226/5039.
×
Page 214
Suggested Citation:"6 Causes and Effects in the Bering Sea Ecosystem ." National Research Council. 1996. The Bering Sea Ecosystem. Washington, DC: The National Academies Press. doi: 10.17226/5039.
×
Page 215
Suggested Citation:"6 Causes and Effects in the Bering Sea Ecosystem ." National Research Council. 1996. The Bering Sea Ecosystem. Washington, DC: The National Academies Press. doi: 10.17226/5039.
×
Page 216
Suggested Citation:"6 Causes and Effects in the Bering Sea Ecosystem ." National Research Council. 1996. The Bering Sea Ecosystem. Washington, DC: The National Academies Press. doi: 10.17226/5039.
×
Page 217
Suggested Citation:"6 Causes and Effects in the Bering Sea Ecosystem ." National Research Council. 1996. The Bering Sea Ecosystem. Washington, DC: The National Academies Press. doi: 10.17226/5039.
×
Page 218
Suggested Citation:"6 Causes and Effects in the Bering Sea Ecosystem ." National Research Council. 1996. The Bering Sea Ecosystem. Washington, DC: The National Academies Press. doi: 10.17226/5039.
×
Page 219
Suggested Citation:"6 Causes and Effects in the Bering Sea Ecosystem ." National Research Council. 1996. The Bering Sea Ecosystem. Washington, DC: The National Academies Press. doi: 10.17226/5039.
×
Page 220
Suggested Citation:"6 Causes and Effects in the Bering Sea Ecosystem ." National Research Council. 1996. The Bering Sea Ecosystem. Washington, DC: The National Academies Press. doi: 10.17226/5039.
×
Page 221
Suggested Citation:"6 Causes and Effects in the Bering Sea Ecosystem ." National Research Council. 1996. The Bering Sea Ecosystem. Washington, DC: The National Academies Press. doi: 10.17226/5039.
×
Page 222
Suggested Citation:"6 Causes and Effects in the Bering Sea Ecosystem ." National Research Council. 1996. The Bering Sea Ecosystem. Washington, DC: The National Academies Press. doi: 10.17226/5039.
×
Page 223
Suggested Citation:"6 Causes and Effects in the Bering Sea Ecosystem ." National Research Council. 1996. The Bering Sea Ecosystem. Washington, DC: The National Academies Press. doi: 10.17226/5039.
×
Page 224
Suggested Citation:"6 Causes and Effects in the Bering Sea Ecosystem ." National Research Council. 1996. The Bering Sea Ecosystem. Washington, DC: The National Academies Press. doi: 10.17226/5039.
×
Page 225
Suggested Citation:"6 Causes and Effects in the Bering Sea Ecosystem ." National Research Council. 1996. The Bering Sea Ecosystem. Washington, DC: The National Academies Press. doi: 10.17226/5039.
×
Page 226
Suggested Citation:"6 Causes and Effects in the Bering Sea Ecosystem ." National Research Council. 1996. The Bering Sea Ecosystem. Washington, DC: The National Academies Press. doi: 10.17226/5039.
×
Page 227
Suggested Citation:"6 Causes and Effects in the Bering Sea Ecosystem ." National Research Council. 1996. The Bering Sea Ecosystem. Washington, DC: The National Academies Press. doi: 10.17226/5039.
×
Page 228
Suggested Citation:"6 Causes and Effects in the Bering Sea Ecosystem ." National Research Council. 1996. The Bering Sea Ecosystem. Washington, DC: The National Academies Press. doi: 10.17226/5039.
×
Page 229
Suggested Citation:"6 Causes and Effects in the Bering Sea Ecosystem ." National Research Council. 1996. The Bering Sea Ecosystem. Washington, DC: The National Academies Press. doi: 10.17226/5039.
×
Page 230
Suggested Citation:"6 Causes and Effects in the Bering Sea Ecosystem ." National Research Council. 1996. The Bering Sea Ecosystem. Washington, DC: The National Academies Press. doi: 10.17226/5039.
×
Page 231
Suggested Citation:"6 Causes and Effects in the Bering Sea Ecosystem ." National Research Council. 1996. The Bering Sea Ecosystem. Washington, DC: The National Academies Press. doi: 10.17226/5039.
×
Page 232
Suggested Citation:"6 Causes and Effects in the Bering Sea Ecosystem ." National Research Council. 1996. The Bering Sea Ecosystem. Washington, DC: The National Academies Press. doi: 10.17226/5039.
×
Page 233
Suggested Citation:"6 Causes and Effects in the Bering Sea Ecosystem ." National Research Council. 1996. The Bering Sea Ecosystem. Washington, DC: The National Academies Press. doi: 10.17226/5039.
×
Page 234
Suggested Citation:"6 Causes and Effects in the Bering Sea Ecosystem ." National Research Council. 1996. The Bering Sea Ecosystem. Washington, DC: The National Academies Press. doi: 10.17226/5039.
×
Page 235
Suggested Citation:"6 Causes and Effects in the Bering Sea Ecosystem ." National Research Council. 1996. The Bering Sea Ecosystem. Washington, DC: The National Academies Press. doi: 10.17226/5039.
×
Page 236
Suggested Citation:"6 Causes and Effects in the Bering Sea Ecosystem ." National Research Council. 1996. The Bering Sea Ecosystem. Washington, DC: The National Academies Press. doi: 10.17226/5039.
×
Page 237

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

CAUSES AND EFFECTS IN THE BERING SEA ECOSYSTEM 196 6 Causes and Effects in the Bering Sea Ecosystem Population changes in some Bering Sea marine mammal, seabird, shellfish, and fish species over the last 30 years may indicate changes that could affect the long-term status of the Bering Sea ecosystem as a natural resource. The causes and significance of these changes are the key issues faced by the committee. The goals of this chapter are to broadly examine the effects of both environmental variation and patterns of human use on the Bering Sea ecosystem, and to develop a scenario for how they may have manifested themselves over the past three decades. The chapter is divided into three sections. The first section attempts to answer the following questions: Is environmental variability a cause of changes in marine mammal, seabird, and fish assemblages in the Bering Sea ecosystem? If so, are the links direct (e.g., a direct population response to physical forcing) or indirect (e.g., bottom-up trophic interactions)? The second section focuses on human effects by addressing the following questions: Is human activity a cause of change in marine mammal, seabird, and fish assemblages in the Bering Sea ecosystem? Has nonextractive resource use, such as waste disposal and pollution, had an effect on the ecosystem? Has extractive resource use—i.e., fishing—affected the Bering Sea ecosystem through removals causing short-term changes in density or distribution, or longer-term effects on population abundance, or through indirect impacts on nontarget species (e.g., bycatch)? The third section presents a conceptual analysis of observed changes in marine mammal, seabird, and fish assemblages in the Bering Sea ecosystem over decadal periods. Some major changes in the Bering Sea ecosystem are known to be the direct result of human extractive use (e.g., the depletion of certain marine mammals and fish stocks of commercial importance), whereas other changes appear to be related to environmental factors. The proposed conceptual model recognizes that both anthropogenic and natural factors have had major influences on the Bering Sea ecosystem—the model suggests that the current Bering Sea ecosystem is a product of the complex interactions between human use and natural fluctuations caused by natural forcing. The analyses in this section build on the framework provided in Chapter 2.

CAUSES AND EFFECTS IN THE BERING SEA ECOSYSTEM 197 ENVIRONMENTAL VARIABILITY Much has been said about the regime shift that occurred in the North Pacific starting in the mid or late 1970s (Chapter 3). A relatively warm period was initiated in the northeastern part of the region, oceanic plankton productivity increased in the Gulf of Alaska, salmon production more than doubled throughout Alaska, and several fish populations experienced unusually strong recruitment and grew dramatically. At the same time, Steller sea lions, harbor seals, and some populations of seabirds in the Alaska region declined. Were such changes unique or had they occurred previously in this region or elsewhere in the world? Appropriate data sets to answer this question are scarce. There are suggestions (see Francis and Hare, 1994; Beamish, 1993) that an earlier period of high productivity occurred in the North Pacific about 50 years ago. The apparent abundance of several stocks of sardines in the Pacific showed similar peaks, in the late 1930s and early 1940s and again after the mid-1970s (Kawasaki, 1991). Steele and Henderson (1984) modeled long-term fluctuations in fish stock and showed that many pelagic fish stocks change rapidly in abundance, with intervening periods of about 50 years. In a later paper, Steele (1991) reviewed changes in several North Atlantic ecosystems, including the North Sea, where there had been dramatic herring declines and demersal fish increases, and concluded that large switches in marine communities can last several decades and that they can occur without human involvement, but can be increased in frequency or amplitude by human actions (see also Chapter 2). While the major cause of such regime shifts appears to be changes in ocean climate of comparable frequency —that is, changes in circulation and mixing, and consequent alteration of surface layer temperature and other conditions—the mechanisms linking environmental and biotic changes are poorly understood. Conventional wisdom is that productivity at lower trophic levels is enhanced (or diminished), for example, by intensification (or diminution) of upwelling, with altered productivity then being felt at higher trophic levels. There may be other significant biotic consequences of climate changes, including: • Changes in the community structure as recruitment of different species is differently affected by ambient conditions. • Changes in distributions and migration patterns, especially of early life history stages, affecting survival of individuals of species and their availability to predators. Many of the changes in the relative abundances of species in both fish and invertebrate communities observed in the 1960s in the Bering Sea appear to have been associated with the development of large commercial fisheries. This is perhaps most clearly illustrated in the overexploitation of groundfish assemblages (e.g., Pacific Ocean perch, yellowfin sole, and herring in the eastern Bering Sea and Gulf of Alaska). But low- frequency changes in ocean and atmospheric climate have also been documented in the Bering Sea and Gulf of Alaska over this

CAUSES AND EFFECTS IN THE BERING SEA ECOSYSTEM 198 period. These environmental changes may have increased the productivity of some species while reducing the productivity of others. Natural Frequencies in the Environment Over relatively short periods, changes in ocean climate will have the most pronounced impacts on the productivity of lower trophic levels. The response of top consumers to such changes may also be evident in the short-term (for example, seabird chick survival), but longer-term effects may also be expected. One important issue in this study is whether the environment is influencing fish population and associated fishery changes. To address this question, it is necessary to first define what is meant by natural fluctuations in the environment and their effects on fish and fisheries in marine ecosystems. Some aspects of Bering Sea variability discussed in Chapter 4 are then revisited in order to see how they may relate to environmental forcing. There are many natural frequencies of variability in the northeast Pacific and Bering Sea. Those most studied by fisheries oceanographers are those that occur at the annual and decadal-scales. The sense of annual scale variability is critical for providing the stock assessments needed for annual fisheries management cycles. However, it is the decadal (and longer)-scale variability that seems to provide more insight into marine ecosystem processes, and such scales are rapidly becoming the focus of fisheries oceanography in both the North Pacific and North Atlantic. A fundamental question being asked is whether climate can cause rather rapid shifts in the organization of marine ecosystems, and if so, on what time and space scales these effects can be measured. Hollowed and Wooster (1992) postulated that there are two mean state of winter atmospheric circulation in the North Pacific (Figure 6.1). These states are identified by the frequency of intense winter ''Aleutian low pressure" events and the resulting ocean temperature and surface wind field responses. In a later paper, Hollowed and Wooster (1994) were able to show a link between years of anomalously low coastal ocean temperatures (caused by intense winter low-pressure events) and simultaneous strong year classes in groundfish species ranging from California to the Bering Sea (Figure 6.2). Fritz et al. (1993) provided evidence that pollock recruitment is positively correlated with temperature (Figure 6.3). Francis and Hare (1994) have used the methods of time series analysis (autoregressive integrated moving average, and intervention models) to analyze or describe the spatial and temporal dimensions of the relationship between salmon production and atmosphere/ocean physics. They found very significant and coherent linkages between relatively sudden interdecadal shifts in the North Pacific atmosphere and ocean physics and marine biological responses as evidenced by indices of Alaska salmon production. Figure 3.10 summarizes some of the results of these analyses. The top panel shows the time series of the North Pacific index (winter atmospheric pressure) during the twentieth century (Trenberth and Hurrell, 1994). The bottom two panels show actual fits of data and intervention analysis models to the index (1900–92) and western Alaska sockeye salmon catch time series (1925–92). Two points stand out from this analysis:

CAUSES AND EFFECTS IN THE BERING SEA ECOSYSTEM 199 Figure 6.1 Two alternating patterns of atmospheric circulation postulated by Hollowed and Wooster (1992). An example of a winter sea level pressure pattern is illustrated for each circulation type (reproduced from Emery and Hamilton, 1985).

CAUSES AND EFFECTS IN THE BERING SEA ECOSYSTEM 200 Figure 6.2 Relationship between sea surface temperature anomalies (upper) and proportion of northeast Pacific groundfish stocks with extreme year classes (lower) (W. Wooster, personal communication).

CAUSES AND EFFECTS IN THE BERING SEA ECOSYSTEM 201 Figure 6.3 Bering Sea pollock spawner-recruit relationship and relative temperature during the first year of life for each year class. W = warm year, C = cold year, and A = average year (from Fritz et al., 1993a). • Coherence is shown between physical and biological variables primarily at the decadal (regime) scale, and not the annual scale. • During the twentieth century, there appear to have been four interdecadal regimes (Figure 3.10) in the North Pacific coupled atmosphere/ocean system: 1900 to 1924, 1925 to 1946, 1947 to 1976, and 1977 to the present. Certainly the well-documented regime shift that occurred during the winter of 1976–77 had a profound impact on many components of North Pacific large marine ecosystems. Venrick and others (1987) show significant shifts in phytoplankton production (integrated chlorophyll-a) just north of Hawaii at about that time (Figure 3.16). Subsequent analysis reveals that the increase may be due to increased production of a deep species of phytoplankton in response to a shift in ocean mixing and a deepening of the mixed layer. Brodeur and Ware (1992) found a significant shift in the zooplankton biomass of the subarctic Pacific (Alaska Gyre), which corresponds to this recent regime shift (Figure 3.17). Later in this section, we point out possible relationships between this climatic regime shift and biological responses in the Bering Sea ecosystem. Baumgartner et al. (1992) looked at natural frequencies in climate-driven fish production of the northeast Pacific over a much longer period. Through the analysis of fish scale deposition rates in anaerobic sediments in the Santa Barbara Basin off southern California, they developed a 1,750-year time series proxy of pelagic fish abundance in the California Current (Figure 6.4,

CAUSES AND EFFECTS IN THE BERING SEA ECOSYSTEM 202 Figure 6.4 Proxy time series of pelagic fish abundance in the California Current (top), power spectra for high- frequency (< 150 year) variability (bottom left) and low-frequency (> 150 year) variability compared with tree ring widths (bottom right) (Baumgartner et al., 1992; t. Baumgartner, personal communication).

CAUSES AND EFFECTS IN THE BERING SEA ECOSYSTEM 203 top panels on Pacific sardine and northern anchovy). They partitioned this variability into low frequency (periods greater than 150 years) and high frequency (periods less than 150 years). The low-frequency variability reflects the centennial-scale climate epochs of the last 200 years (e.g., Medieval Warm Period and Little Ice Age), and relates very closely to bristlecone pine tree ring widths, which reflect prevailing atmospheric temperature regimes (Figure 6.4, bottom right panel). The high-frequency power spectra for both sardine and anchovies show peaks at around 60 and 75 years (Figure 6.4, lower left panel). During the twentieth century, pelagic fish populations of the California Current have tended to vary in response to atmospheric forcing similar to Alaska salmon (see Figure 3.14) (Ware and Thompson, 1991). This being the case, the kind of decadal-scale atmospheric forcing in the North Pacific seen during the twentieth century could have persisted for centuries. As discussed in Chapter 3, Royer (1982) has identified possible relationships between regular variations in the production of a number of commercially important North Pacific marine species and the 18.6-year nodal tidal signal. For example, fluctuations in Bering Sea herring abundance since the early 1960s are noticeably similar to the pattern of sea surface temperature (Figure 6.5). Herring spawn inshore and feed in the coastal zone during the summer. Thus, year-class strength is probably related to the seasonal development of the coastal biological community, which is advanced and more prolific in warm years (Cooney, 1981; Springer et al., 1984, 1987). Likewise, the nodal tide signal has been associated with 60 percent of the variance of year class strength of halibut in the Gulf of Alaska (Parker et al., 1994). Of possible greater importance to the biota than the temperature contribution is the alteration of the relative amplitudes of diurnal and semidiurnal tidal components over this 18.6-year period. Organisms that are dependent on one of these tidal components might suffer under these changing conditions. Nodal tide variability might also affect the tidal mixing in the Bering Sea, as suggested by Loder and Garrett (1978) for the Labrador shelf. The point of this discussion is that climate-driven variability in the large marine ecosystems of the North Pacific is significant, occurs at many different time scales, and affects many components of the ecosystems. It seems clear that climate does cause rather rapid shifts in the organization of these marine ecosystems and that the decadal-scale may be more important than the annual scale in its impact. Environmental Effects on Marine Mammals and Birds Over relatively short periods, changes in productivity will likely have the most pronounced effects at lower trophic levels, but there may also be longer-term effects on top consumers. Lipps and Mitchell (1976) proposed a model linking the evolution of pelagic marine mammals to changes in the trophic structure of the oceans. They suggested that variations in primary productivity caused by upwelling may explain the invasion of the sea by mammals, and subsequent radiations and extinctions. As large-bodied, long-lived, K-selected species, marine mammals would be expected to be relatively insensitive to many types of environmental fluctuations. As discussed in Chapter 4, most species that occur in the Bering Sea have relatively broad distributions in the North Pacific or Arctic Basin, and are therefore adapted to survival under a range of environmental

CAUSES AND EFFECTS IN THE BERING SEA ECOSYSTEM 204 conditions. Conditions that exceed physiological limits and result in the deaths of individuals should occur infrequently, especially in the core of a species' range (e.g., Trites, 1990). However, it is clear that pinnipeds living in more southern regions may be affected by relatively severe climatic events such as El Niño (Trillmich and Ono, 1991). Figure 6.5 Bering Sea herring abundance and sea surface temperature (Wespestad, 1991). Climate-caused physical changes in the environment can have significant effects on marine mammal habitats, and thereby on the abundance and distribution of species. In the Bering Sea, sea ice plays an important role in marine mammal ecology, providing important habitat for some species while excluding others (Fay, 1974). Year-to-year variations in the extent of sea-ice cover change species distributions, which may have a variety of biological effects. An example of ice-cover-dependent interspecific interaction is the relatively high rate of walrus predation on seals that occurred in the Bering Sea in 1979. This was a very light ice year, when the overlap in distributions of these species was greater than usual (Lowry and Fay, 1984). Over a longer time frame, changes in sea-ice cover are related to changes in sea level. The two in combination may have a dramatic effect on marine mammal habitat availability (Davies, 1958). These events must have resulted in major shifts in the distributions of pinnipeds that use either land or ice for hauling out. As discussed previously, the recent significant declines in fur seals, sea lions, and harbor seals probably have been environmentally influenced by climate-induced changes in the abundance and availability of food for juveniles during this critical phase of their life histories. More specifically, the problem probably has something to do with the relative availability of appropriately sized pollock and other forage fish (e.g., capelin, sand lance, and herring), and

CAUSES AND EFFECTS IN THE BERING SEA ECOSYSTEM 205 would be a decadal (rather than annual)-scale effect. Environmental change, however, may not be the only cause of reductions in food for pinniped species. Bering Sea Fish (Invertebrates) and Fisheries A number of the variations in Bering Sea fish populations described in Chapter 4 appear to be related to environmental forcing. In general, there are two ways that this kind of forcing manifests itself. First, it can redistribute marine fish populations and their fisheries in space. The second and perhaps more important way is by affecting the absolute abundance of marine fish populations primarily through mechanisms controlling recruitment of new individuals. A considerable number of commercially important marine fish stocks of the Bering Sea (and Gulf of Alaska as well) and their fisheries are, at any time, supported by a relatively small number of very strong year classes or cohorts. Looking at a number of figures presented in Chapter 4, one can certainly see evidence of this in the eastern Bering Sea and Gulf of Alaska (see Figures 4.1, 4.2, and 4.3 on king and Tanner crab; Figure 4.5, on walleye pollock; Figure 4.13, on Pacific cod; Figure 4.14, on Atka mackerel; Figure 4.18, on flatfish; Figure 4.21, on Pacific Ocean perch; and Figures 4.24 and 4.25, on herring). Because fisheries tend to develop on stocks that are at high abundance (due primarily to strong recruitment), attempts to attribute cause generally focus on the causes of declines in these stocks rather than their increases even though the high stock levels may be anomalous events. The classic example of such a case in the northeast Pacific is that of pollock in Shelikof Strait (Gulf of Alaska). Groundfish. A number of groundfish populations of the eastern Bering Sea, Aleutian Islands, and Gulf of Alaska were significantly affected by the well-documented regime shift (general warming) that occurred in the winter of 1976–77 and subsequently manifested itself in the late 1970s and early 1980s. As pointed out by Hollowed and Wooster (1992), several strong year classes of groundfish occurred in 1977 and 1978 (eastern Bering Sea and Gulf of Alaska pollock in 1978, eastern Bering Sea cod in 1977, and Aleutian Islands Atka mackerel in 1977). In particular, these very strong year classes of pollock and cod supported significant expansions of Bering Sea fisheries. The 1978 year class of pollock essentially supported the expansion of the Bering Sea pollock fishery into the Aleutian Basin (Bogoslof and donut hole). In addition, the pollock fishery of the eastern Bering Sea shelf was redistributed to the southeast in the late 1970s in possible response to changes in ocean climate (see Figures 4.8, 4.9, and 4.10). Although the year classes seem to be more difficult to pinpoint, significant increases in eastern Bering Sea shelf flatfish populations occurred in the late 1970s and early 1980s (yellowfin sole, rock sole, and Alaska plaice; see Figure 4.18). There was also a large increase in arrowtooth flounder in the Gulf of Alaska. At the same time, the one eastern Bering Sea flatfish species that tends to prefer colder water (Greenland turbot) declined significantly. Juvenile Greenland turbot were pushed to the northwest along the outer Bering Sea shelf as the ocean climate warmed (Figures 4.15 and 4.16).

CAUSES AND EFFECTS IN THE BERING SEA ECOSYSTEM 206 Forage Fish. Pacific herring biomass in the eastern Bering Sea exhibits patterns similar to those of a number of other groundfish and invertebrate populations—decadal-scale surges in biomass supported by a small number of very strong year classes (Figures 4.24 and 4.25). As for a number of groundfish populations, the 1977 year class of eastern Bering Sea herring was extremely strong, possibly in response to the concurrent regime shift. It is interesting to note that the three largest year classes of herring (1957, 1958, and 1977) occurred at times of significant warming in the North Pacific (see Figure 6.5), and perhaps also in response to the 18.6-year nodal tidal signal. In the western Gulf of Alaska, a 20-year record of trawl catches of important indicator species revealed that capelin virtually disappeared in the 1980s (Figure 4.27). Data on seabird diets and from fishery surveys in the Bering Sea also indicate that capelin and eulachon, another smelt, essentially disappeared there during the same time (Decker et al., 1995; Fritz et al., 1993; Figure 4.26). Both species were comparatively abundant in the 1970s, when the physical regime was much different than in the 1980s. Deposition rates of fish scales in Skan Bay on the Bering Sea side of Unalaska Island (eastern Aleutians) declined abruptly in the late 1970s after a period of relatively high abundance (R. Francis, personal communication). None of these short histories runs entirely parallel to that of herring, and all are difficult to evaluate in a decadal-scale context. Salmon. The clear link between decadal-scale variations in Alaskan salmon abundance and regime-scale North Pacific climate fluctuations was discussed earlier in this section. Invertebrates. As discussed in Chapter 4, fluctuations in eastern Bering Sea and Gulf of Alaska crab populations (king and Tanner; Figures 4.1, 4.2, and 4.3) and their fisheries occur in response to infrequent strong year classes or cohorts. The most dramatic of these fluctuations occurred in 1981 in Bristol Bay and 1982 in the Kodiak region of the Gulf of Alaska, when red king crab fisheries in both regions crashed. The Bristol Bay resource supported the largest king crab fishery in the world. Beginning in 1966, this fishery increased annually, to an all-time record catch of 58,938 t in 1980 (Figure 5.7). This increasing trend was mirrored in the biomass estimates from stock assessment surveys performed annually by the National Marine Fisheries Service. However, beginning in 1979, the survey began reporting a decline in overall biomass relative to the prior year's estimate. These preseason forecasts were not readily accepted by the industry. Many people believed that the survey "missed the crab." In 1981, however, the catch dropped from almost 59,000 t to 15,876 t. Although such a decline in catch was predicted, the magnitude of the decline was shocking. In 1982, the catch fell to only 1,361 t, and in 1983 the number of reproducing male and female crabs was believed to be so low that no directed fishery was permitted. Similar collapses were observed in all the other major crab fisheries in the Bering Sea (Figure 5.7). The next most valuable king crab was blue king crab, found most abundantly in the Pribilof Islands region. This fishery experienced similar declines in catch, from a high of 4,991 t in 1980, to 1,998 t in 1982, to only 318 t in 1987. Since then there has been no directed fishery for blue king crab in the Pribilof region. The largest of the Tanner crabs,

CAUSES AND EFFECTS IN THE BERING SEA ECOSYSTEM 207 Chionoecetes bairdi, went from an all-time high in 1976–77 to no fishery by 1986. The smaller C. opilio crab became valuable only after the industry scrambled to develop alternative markets. Once again, environmental forcing could have significantly affected all of these dynamics, by creating favorable conditions for those pulses of strong year classes that have supported many of the invertebrate (and finfish) fisheries of the North Pacific and Bering Sea. The declines, or in some cases crashes, were likely the results of disappearances due to mortality (natural and/or fishing) of those strong year classes, and were probably not environmentally driven. HUMAN EFFECTS Pollution Persistent organics, heavy metals, and radioisotopes are major anthropogenic inputs from human industrialization that flow into marine systems. For example, acid rain drainage into the western Bering Sea from coal mining on the Kamchatka Peninsula and into the Gulf of Anadyr in the northern Bering Sea is of concern (National Geographic, 1994). The impact of such anthropogenic contaminants on high-latitude marine ecosystems, such as the Bering Sea, can be especially detrimental, because many biota are long-lived and slow growing, and thus are capable of accumulating pollutants over time. In addition, certain higher trophic animals (marine mammals, seabirds, and fish), which can accumulate these pollutants, are consumed by humans, and thus these contaminants are transferred to humans. Currently, the international Arctic Monitoring Assessment Program (AMAP, 1993), along with U.S. and Russian scientists involved in the program of Long-term Ecological Investigations of the Bering Sea and Other Pacific Ocean Ecosystems (BERPAC) (O'Connor et al., 1992), are focusing on these major contaminants. There is to date no hard evidence that these contaminants have significantly affected the Bering Sea ecosystem. Persistent Organics. Persistent organics are transported primarily from low latitudes to the high latitudes via airborne transport. For example, polychlorinated biphenyls (PCBs), DDT, dioxins, hexachlorocyclohexane (HCH), dibenzofurans, chlordane, and toxaphene have been found in arctic air, surface water, snow, suspended sediments, fish, marine mammals, seabirds, terrestrial animals and humans (Barrie et al., 1992; Lockhart et al., 1992; Muir et al., 1992; O'Connor et al., 1992; Thomas et al., 1992). Long-range transport can occur via rivers and ocean currents as well as the atmosphere (carrying pollutants from industrialized and agricultural areas, including those in Asia, Europe, and North and Central America). Although many of these organic compounds have been banned in many countries, they are still produced and used in many other countries around the world. Persistent organic compounds are preferentially volatized in warm regions of the world and transported to cold regions, such as the Bearing Sea and the Arctic Ocean (Otto, 1981). This process is known as "cold trapping" (Muir et al., 1992). Persistent organic compounds preferentially accumulate in the fatty tissues of organisms. This is especially important in cold northern regions where lipid-rich wildlife is consumed by native populations, thus providing

CAUSES AND EFFECTS IN THE BERING SEA ECOSYSTEM 208 a direct pathway of these pollutants to humans (Kinloch et al., 1992; Muir et al., 1992). Concentrations of HCHs and toxaphene are greater in surface waters of the Arctic than in waters of the temperate or north Atlantic waters (Muir et al., 1992; O'Connor et. al., 1992). Dioxins and dibenzofurans have higher concentrations in ringed seals and polar bears in the Canadian high Arctic than mammals further south (Norstrom et al., 1990). Elevated concentrations of PCBs have been found in both planktonic and benthic fauna and sediments in the Bering and Chukchi seas (Chernyak et al., 1992; Rice et al., 1992), and elevated concentrations of PCBs have been found in the breast milk of native Inuit women (Dewailly et al., 1989). However, no population-level effects have been observed to date. Heavy Metals. Heavy metals are also of concern and have both local and long-range transport mechanisms. The cold Arctic is a sink for heavy metals, and marine mammals are known to contain elevated concentrations of heavy metals, especially mercury and cadmium (AMAP, 1993). Low concentrations of heavy metals have been found in organisms and sediments of the Bering and Chukchi seas (Kolobova et al., 1992; Krynitsky et al., 1992). Other anthropogenic compounds of concern are those influencing acidification and arctic haze in the atmosphere, climate change (e.g., global warming), and depletion of stratospheric ozone. Radionuclide Contamination. Radionuclide releases, both from atmospheric testing fallout and from dumping into the marine environment, are of special concern. The revelations of large-scale disposal of radioactive materials in the North pacific Ocean (e.g., off Kamchatka Peninsula in the western Bering Sea) and the Arctic Ocean by the former Soviet Union (Yablokov et al., 1993) have led to new interest in the study of how these ecosystems respond to environmental stress and global change. Recent information on the extent of Russian radioisotope contamination in the North Pacific and Arctic oceans is of interest to this study. Although current concentrations in the Bering Sea do not indicate a widespread threat, there is a need for national surveys and monitoring networks to establish baselines for specific radioisotopes in the atmosphere) and in aquatic and terrestrial environments (AMAP, 1993). An example of a radioactive isotope used to study long-term transport of released materials is plutonium (with a half-life of 24,000 years). Plutonium isotopes (240Pu and 239Pu) from fuel reprocessing have a specific ratio different from the ratios in bomb fallout. Data presented by T. Beasley (personal communication) at the 1993 Interagency Research Policy Committee Workshop on Arctic Contamination in Anchorage showed that low concentrations of plutonium in surface sediments in the Canada Basin of the Arctic Ocean off Alaska are unequivocally derived from weapons-grade material (240Pu/239Pu = 0.06), most likely discharged from plutonium reprocessing plants on the Ob and Yenesey rivers in Siberia. At the same time, the plutonium isotope composition of shallow sediments in the Bering and Chukchi seas is consistent with bomb fallout sources (240Pu/239Pu = 0.20), even though the still low inventories of plutonium are an order of magnitude higher than in the deep arctic sediments. Cesium-137, a particle-reactive radioisotope, is an anthropogenic product of nuclear testing (half-life of 30.2 years) that was introduced into the biosphere at around 1953, as a result of bomb fallout deposition. The distribution in marine sediments is influenced not only by

CAUSES AND EFFECTS IN THE BERING SEA ECOSYSTEM 209 atmospheric fallout but also by oceanographic currents, bioturbation, and greater cation exchange and solubility in seawater, as well as by additional sources derived from nuclear material disposal, particularly in the former Soviet Union. Recent data indicate only low-level contamination of marine sediments (Figure 6.6; Cooper et al., 1995; Grebmeier et al., 1993). The highest radiocesium concentration (12.9 Bq kg dw-1) occurred in a sheltered embayment (Port Clarence, Alaska) subject to significant freshwater runoff, which would provide a source of terrestrially deposited 137Cs, irreversibly bound to clay minerals. Some radiocesium accumulation has been detected in Bering and Chukchi Sea benthic organisms (J.H. Grebmeier, L. Cooper, and I.L. Larsen, unpublished data, 1994). Low concentrations of dissolved or waterborne 137Cs have also been detected in the water column, and the inventory integrated over depth at various sites in the Bering and Chukchi seas ranged from 8.1 to 12.6 mBq cm-2 in 1988 (Medinets et al., 1992); consequently most of the radiocesium is deposited within sediments. A comparison with marine sediments more heavily by anthropogenic contamination, such as Russian waters of the Black Sea, indicate a combined surface sediment concentration of 17.3 Bq kg dw-1 for 137Cs, with total 137Cs inventories of 536 mBq cm-2 (J. Grebmeier, L. Cooper, and I. L. Larsen, unpublished data, cited in Cooper et al., 1995); therefore typical cesium concentrations and inventories in Alaskan marine sediments of the Bering and Chukchi seas are an order of magnitude less than in Black Sea sediments. Ongoing studies as part of the Office of Naval Research, U.S. Fish and Wildlife Service and Arctic Monitoring Assessment Program will provide additional data to develop long-term monitoring and assessment programs. Offshore Oil and Gas Development Although Alaska is estimated to contain large petroleum resources on its outer continental shelf (OCS) and in state waters (i.e., within three miles of shore) (NRC, 1994b), the only oil produced from Alaska's OCS to date has come from Cook Inlet south of Anchorage. There has been considerable industry interest in the OCS off Alaska's Beaufort and Chukchi seas, north of the Bering Sea, and a great deal of production from state waters and onshore around Prudhoe Bay. Because the oil is piped from the North Slope to Alaska's south-central coast, there has been little or no direct effect of North Slope oil production on the Bering Sea. Industry interest in the Bering Sea has been relatively low and has declined recently (NRC, 1994b), although leases have been sold there and exploration has occurred in the past (Lewbel, 1983; MMS, 1992). For example, a lease in the Navarin Basin (part of the Bering Sea OCS) planned for 1992 was canceled because of lack of industry interest and deferred for further review until 1996 (MMS, 1992). Thus, the likelihood of offshore oil and gas production in the eastern Bering Sea within the next 10 years is close to zero, because it takes at least that long after a lease sale to begin producing oil and gas, and there is little or no exploration taking place in the eastern Bering Sea at present. There is no possibility in the foreseeable future of the kind of extensive oil and gas activities that characterize the OCS of the central Gulf of Mexico. Little or no OCS oil and gas activity is occurring or likely to occur soon in the Russian area of the Bering Sea (West, 1994). It is therefore safe to conclude that OCS oil and gas activities in the Bering Sea have not significantly affected the Bering Sea ecosystem.

CAUSES AND EFFECTS IN THE BERING SEA ECOSYSTEM 210 Figure 6.6 Surface sediment cesium-137 distributions during 1992 and 1993 (Cooper et al., 1995). The black dots indicate radioactive dump sites (Yablokov et al., 1993). The actual effects of offshore oil and gas activities have been described in recent reports by the National Research Council (most recently with respect to Alaska [NRC, 1994b]) and by the Group of Experts on the Scientific Aspects of Marine Pollution (GESAMP, 1992). The 1994 National Research Council report considered potential effects of oil and gas activities in Alaskan waters, especially the Beaufort and Chukchi seas. The most likely adverse effects of these activities—for the Bering Sea as well as for the Beaufort and Chukchi seas—would occur on shorelines (NRC, 1994b), because there is unlikely to be enough development to affect more than small, localized areas in the open sea. Even a major oil spill at sea would probably have the most serious consequences on shorelines reached by the spilled oil. Two major potential sources of adverse effects—in both the western and eastern Bering Sea—would be onshore developments that were related to offshore oil and gas activities, including exploration, development, production, and operation of terminals (possible onshore developments could be related to staging, fabrication, and pipeline terminals, harbors, and housing for workers) and oil spills (from accidents involving development and production wells, and transportation accidents). Particularly sensitive areas onshore include estuaries, lagoons, haulout areas for marine mammals, marine bird colonies, and other places where organisms gather to feed, overwinter, or reproduce. Areas of human habitation and activities would also be particularly sensitive.

CAUSES AND EFFECTS IN THE BERING SEA ECOSYSTEM 211 Direct Effects of Extractive Resource Use Chapters 4 and 5 showed that a number of changes in animal populations observed in the Bering Sea are due simply to direct and intentional removals by humans. It is important to recognize these explainable changes, as distinct from the unexplained population declines that are now of major concern. Marine Mammals. Hunting has clearly had a major impact on populations of Bering Sea marine mammals. Over the past 200 years, nearly all species have been harvested for commercial and/or subsistence purposes. The Steller sea cow was hunted to extinction, and the northern right whale nearly so. Gray whales, bowhead whales, fur seals, walruses, and sea otters were severely reduced, but their populations have recovered to various degrees. Blue whales, fin whales, humpback whales, sperm whales, and sei whales were not heavily hunted until relatively recently, so it is unclear what their long-term recovery will be. Species of relatively low commercial value (e.g., Steller sea lions, ringed seals, bearded seals, spotted seals, and harbor seals) generally have not been severely depleted by hunting, though they are consistently hunted for their subsistence use and because they have important cultural and traditional values for the native populations. However, it is quite possible that hunting was partially responsible for the initial decline of sea lions, harbor seals, and fur seals in some parts of their range. Fish. Fishing has clearly had significant influence on the dynamics of fish populations in the Bering Sea. The effects have been most clearly seen in the K-selected slope rockfish species (in particular, Pacific Ocean perch) harvested in the 1960s from the eastern Bering Sea, Aleutian Islands, and Gulf of Alaska. As was mentioned in Chapter 5, these populations were essentially mined out of the region and, with a few exceptions, have shown little indication of recovery. Most slop rockfish are long-lived (as long as 90 years for the Pacific Ocean perch [Bakkala, 1993]) and have very low rates of production. This natural history suggests that these populations are adapted to capitalize on infrequent environmental conditions that support strong year classes. With populations at such low levels from the heavy fishing of the 1960s, the chance of taking advantage of these rare conditions when they occur could be significantly reduced. The other group of groundfish species that showed similar responses to the heavy and uncontrolled exploitation of the late 1950s and 1960s were the eastern Bering Sea shelf flatfishes, and in particular, yellowfin sole (Figure 4.15). Unlike Pacific Ocean perch, perhaps because of its more r-selected nature, the yellowfin sole population grew in the late 1970s and early 1980s to previous biomass levels. It is clear that fishing has affected the populations of many groundfish, forage fish, and invertebrates of the Bering Sea and Gulf of Alaska. These fisheries grew rapidly in response to large year classes that boosted populations to high levels of abundance, and in later years of lower recruitment, the fisheries accelerated the declines by removing the biomass that had been created by strong recruitment. This certainly seems to be the case for pollock and cod in both

CAUSES AND EFFECTS IN THE BERING SEA ECOSYSTEM 212 the Bering Sea and Gulf of Alaska, Atka mackerel in the Aleutian Islands, herring in the eastern Bering Sea, and king and Tanner crab in the eastern Bering Sea and western Gulf of Alaska. Recalling the discussions in Chapter 2 of ecosystem structure and self-organizing capacities, perhaps the most serious effects of fishing are long-term impacts on ecosystem organization by the rapid removals of large biomasses of long-lived species (Apollonio, 1994). Removals create tension or disequilibrium at the system level, which, because of the low rates of production of the harvested species, will not be relieved by recoveries to previous conditions. As a result, something else fills the energy niche. And so, according to Allen (1985), this tension within an ecosystem can set the stage for rapid reorganization—the greater the tension, the greater the potential for rapid reorganization. Systems under great tension can then be triggered, perhaps by rather subtle environmental shifts, into rapid and significant flips from one state to another, with drastic alterations in system structure. In this way, the history of fishing and hunting can play a vital role in determining the nature of marine ecosystem order. This is clearly true for the Bering Sea ecosystem, as is discussed further in this report. Indirect Effects of Extractive Resource Use Each year commercial fisheries remove large quantities of biomass from the eastern Bering Sea and Gulf of Alaska. (Substantial quantities of biomass are also removed from the western Bering Sea, but—as described in Chapter 5—reliable quantitative information on those removals is not available.) Although this is an obvious feature of fishing, the actual particular impacts of such removals on other species and on the ecosystem as a whole are largely unknown. Management of marine fish harvests is largely based on assumptions of maximum sustainable yield models. It is assumed that fishing will reduce the standing stock of fish such that there is a density-dependent increase in productivity, resulting in a harvestable surplus. Under the sustainable yield model, the annual biomass increment that is removed by fishing will be replaced the following year as a result of the growth of surviving fishes and the recruitment of new individuals to the population. It is also assumed that changes in the abundance of the target fish population as a result of fishing have no impact on the abundance of competitors in the system. That is, it is assumed that the dynamics of the target population are decoupled from the system. This is not likely to be a reasonable assumption in most situations, and is not an ecosystem-based approach as described in Chapter 2. There are several ways in which removals of fish by fisheries may have effects on other ecosystem components. One may be the effect of fishing on competitive interactions, such that the reduction in density of fishes due to fishing results in reduced food availability for fish-eating predators. The amount of fish caught each year depends on a number of factors. In the Bering Sea, annual catch rates for pollock, herring, and yellowfin sole have ranged from less than 10 percent to more than 60 percent of the exploitable biomass removed (Fritz et al., 1993b). Clearly, industrial fisheries, like that for pollock in the Bering Sea, remove massive amounts of fish biomass from the ocean (the Bering Sea total pollock harvest was almost 4 million t in 1988; Figure 4.6). According to fishery managers, pollock stocks in the Bering Sea, Aleutian Islands, and Gulf of Alaska have not been ''overfished" in recent years. Nonetheless, several regions

CAUSES AND EFFECTS IN THE BERING SEA ECOSYSTEM 213 where pollock were once abundant (e.g., Shelikof Strait, the Bogoslof Island area, and the donut hole) have been heavily exploited, and pollock stocks are now very low in those areas. The most dramatic example of possible local overexploitation occurred in Shelikof Strait. A large spawning concentration of pollock was "discovered" in this region in the late 1970s, and an intensive fishery rapidly developed. As a result, pollock catches in the Gulf of Alaska increased from less than 100,000 t to more than 300,000 t and remained high even during the mid-1980s, when it was apparent that there had been poor recruitment and stock abundance was declining (Megrey and Wespestad, 1990). During this period, annual combined removals from fishing and sea lion predation were estimated to be as much as 30 to 50 percent of the total exploitable biomass (Lowry et al., 1989). According to one estimate (Nunnallee and Williamson, 1988), exploitable biomass of pollock in the Gulf of Alaska fell from greater than 3 million t in 1981 to less than t in 1993 (NPFMC, 1993). During this same period of time, sea lion numbers on nearby rookeries showed a dramatic decline, and animals began to show signs of reduced growth rate (Calkins and Goodwin, 1988; Lowry et al., 1989). However, before 1962–71, models suggest that pollock population sizes and year-class strengths were very low, yet there is no evidence that the sea lion population was in decline. It would be of interest to know what the principal fish in their diet were at that time. The effect of fishery removals on competitive interactions will depend not only on the amount of biomass removed, but also, and perhaps more importantly, on the spatial and temporal pattern of fishing. We might expect that fisheries that operate on a broad temporal and spatial scale would have a lesser effect on local density of fishes than those that are compressed in space and time. The development of sophisticated, highly capitalized fishing fleets has in many cases resulted in harvesting that is very intense. For example, during the winter of 1994 the Bering Sea trawl fleet caught approximately 600,000 t of pollock in a six-week period. At its peak, the fleet was harvesting at a rate of 30,000 t per day. The actual effects of fishing on local densities of fish and the effects of changes in local fish density on marine mammal predators are currently unknown. Another way in which fishing may indirectly affect marine ecosystems is through effects on community structure. In the eastern Bering Sea, large harvests of both whales and fishes, coupled with favorable environmental conditions, could have led to the large increase in pollock abundance that is thought to have occurred in the late 1960s and early 1970s (see Chapter 4). This pollock-dominated pelagic system has probably had significant effects on the abundance of other species of fish, which in turn may have affected seabirds and pinnipeds. Similar major shifts in the dominance of the fish community on Georges Bank in the western North Atlantic (from gadoids to elasmobranchs and pelagic species) have recently been attributed to excessive fishing (NOAA, 1993). Numerous detailed studies have shown a direct correlation between die-offs of seabirds related to deprivation of fish; several examples are cited in Croxall (1987) and Nettleship et al. (1984). One example from the Arctic was presented by Lid (1981), who showed that there was a drastic decline in the breeding success of Atlantic puffins, common guillemots, and razorbills in the Norwegian sector of the North Sea as a result of a reduction in local herring through overfishing. Harvests of herring dropped from 10 million t in 1958 to 0.5 million t in 1969. As discussed in Chapter 4, the major die-offs of seabirds in the Bering Sea and Gulf of Alaska, particularly in the vicinity of the Pribilofs, have been associated with increased fishing and reduction in prey for the birds.

CAUSES AND EFFECTS IN THE BERING SEA ECOSYSTEM 214 Probably the most intriguing relationship between trawling and ecosystem patterns is the time-space correlation between intense trawl fishing and the declines in two marine mammal species, the Steller sea lion and the harbor seal. Declines in these two species coincide with the period of tremendous concentration of trawling effort in time and space that accompanied the Americanization of the fishery after establishment of the 200-mile exclusive economic zone (Fritz, 1993a, 1993b). Yet the observed declines in the populations of both Steller sea lions and harbor seals are not uniform across the ranges of the species in the North Pacific. Populations have not declined in the region from southeast Alaska southward. This line coincides with the demarcation between the geographic region where trawl fisheries operate to the north and where trawling is light or absent to the south. The existence of such a correlation is consistent with some effect of trawling on the populations of these two species of marine mammals. On a smaller scale, attempts to correlate the amounts of pollock caught with changes in sea lion numbers have yielded equivocal results. Loughlin and Merrick (1989) compared fish removals from blocks around seven rookeries with the trend in sea lion counts at those rookeries and found significant negative correlations in some cases, but no relationship or positive correlations in other instances. The likely impacts of fishing on the nutritional status of consumers will depend on a number of factors, including the areas and times fished, the size distribution of the catch, the feeding preferences and population sizes of the predators, and the availability of alternate prey. Steller sea lions consume a wide range of sizes of pollock, including those targeted in the commercial catch (Frost and Lowry 1986). Much of this catch has been taken from areas in the Bering Sea, Aleutian Islands, and Gulf of Alaska that are near major Steller sea lion rookeries and haulouts (Fritz et al., 1993a, 1993b). These fisheries are very much pulsed in nature, and they might therefore have a major impact on the local abundance of pollock (and perhaps other nontarget species) in areas that are intensively fished. More information is needed on this matter. The National Marine Fisheries Service has identified the areas within 20 nautical miles of most rookeries and haulouts, and three at-sea zones (eastern Bering Sea near Bogoslof Island, Seguam Pass, and Shelikof Strait) as critical feeding habitat for Steller sea lions. Fritz (1993a) compiled the records of pollock catches within these areas during 1977–92. In the Bering Sea the catch in critical feeding habitat increased from 100,000 to 300,000 t during 1977–86, to 500,000 to 600,000 t during 1987–91. Most of the latter catch was from the Bogoslof fishery, targeting spawning female pollock in the eastern Bering Sea critical feeding habitat. For the Gulf of Alaska, the percentage of the catch from critical feeding habitats increased from less than 5 percent in 1977 to more than 80 percent in 1985, and has fluctuated between 55 percent and 93 percent during 1986–92. Research on the effects of this fishing on sea lions is needed. Relatively small pollock appear to be important foods for several marine mammal species (Frost and Lowry, 1986; Figure 6.7), and they may be particularly significant in the diets of Pribilof fur seals and young Steller sea lions (Merrick and Calkins, in press; Sinclair et al., 1994). Pollock stocks in both the Gulf of Alaska and Bering Sea have shown variable recruitment since the early 1960s, and occasional strong year classes have supported the commercial fishery (Fritz et al., 1993b). If reductions of pollock spawning stocks from fishing affect the abundance of small pollock (either the size of recruitment within a year or the frequency of relatively strong recruitment events), this could have an impact on the growth,

CAUSES AND EFFECTS IN THE BERING SEA ECOSYSTEM 215 Figure 6.7 Size distributions of walleye pollock eaten by five species of marine mammals collected in the Bering Sea, 1975–81 (from Frost and Lowry, 1986).

CAUSES AND EFFECTS IN THE BERING SEA ECOSYSTEM 216 reproductive output, and survival of species dependent on small pollock for their nutrition (Anonymous, 1993). Trawl Effects on Bottom. Wherever trawling is intensive, there is concern about its possible side effects on ecosystem structure and dynamics. Two major issues are often raised: the effects of bottom disturbance, and the effects of bycatch mortality. Bottom disturbance has the potential to alter the benthic invertebrate community, which itself directly or indirectly provides the prey for most demersal fish and crustacean species. Important bottom habitats, such as biogenic or physical structures, may also be affected by trawling. Bottom disturbance can also alter geochemical fluxes of nutrients and materials from the sediments to the water column, thereby influencing water column production. Mortality of individuals caught as bycatch and discarded can be separated into two cases, losses of juveniles of commercially or recreationally important species, and losses of species without direct importance as targets of human fishing. This distinction is meaningful, because for exploited species, bycatch mortality can be estimated and factored into harvest quotas, whereas for other species there is no accounting by managers and there are no models for the ecosystem consequences of the exploitation, except in the special cases where a bycatch species is endangered or threatened, as for sea turtles (NRC, 1990). Bottom trawling is generally confined to regions characterized by soft sediments, because trawl gear snags and tears on rocks. Because soft sediments are inhabited largely by infaunal, buried invertebrates, rather than epifauna, most of the research on direct impacts of trawling on the benthos has focused on the infauna. These studies do reveal substantial impacts of trawls, especially beam trawls, and dredges on benthic infauna. Mortalities vary with species, but range from 15 to 55 percent (ICES, 1992). The subsequent impact of trawling on benthic communities will depend on the rate of the recolonization process of secondary succession on the seafloor. This involves colonization by more opportunistic species, with rapid growth rates, high reproductive rates, and generally short life spans. Despite some level of confidence in predicting the general effect of physical disturbance on infaunal communities, the longer-term implications of such changes induced by trawl disturbance of the seafloor are not at all obvious. For example, three months after disturbance, Peterson and others (1987) could detect no substantial differences in the macroinfauna of estuarine seafloor in areas that had been intensively disturbed by a clam dredging process. For any given ecosystem, such as the Bering Sea, this question would need to be evaluated empirically by designing fishing regulations as an experiment and evaluating the ecosystem consequences subsequently. This has not been done. The effects of disturbance from bottom trawls on habitat are likely to be greater when the habitat is biogenic rather than physically created. There are reports of trawls moving rocks (Caddy, 1973; Fowler, 1987), and so physically created habitat is not immune to the trawl. However, rocky bottoms are generally avoided by trawlers because of the risk of damage to the gear. In addition, biogenic habitats are more fragile and susceptible to major disruption by trawls. For example, reefs of the tube-building polychaete Sabellaria have been destroyed in regions of the German Wadden Sea, where shrimp trawling is intense (Riesen and Reise, 1982). It is reasonable to expect other biogenic structures produced by benthic animals, namely, burrows, mounds, pits, tubes, and so forth, to be similarly disrupted by bottom trawling. This effect is probably transitory, i.e., the animal architects themselves survive to rebuild the

CAUSES AND EFFECTS IN THE BERING SEA ECOSYSTEM 217 features, but relatively long-lasting where the trawl causes widespread mortality of the habitat-providing species, as in the case of Sabellaria. If recolonization and recovery rates of the habitat provider are slow relative to the frequency of trawling, alteration of the biogenic habitat could persist indefinitely until trawling operations cease. In the Bering Sea, there is no information on the importance of biogenic habitat to various demersal species and no basis on which to evaluate the possible ecosystem significance of trawl damage to bottom habitat. Bottom trawling also creates sediment excavation and resuspension to some extent, depending on trawl design, transit, and bottom type. This physical disturbance of the seabed can enhance the turbidity of the overlying water column, facilitate erosion and transport of fine sediments (thereby changing the grain sizes on the seabed), and alter rates of important biogeochemical processes, such as nutrient releases. Churchill (1989) concluded from study of the trawling on the northeast coast of the United States that trawling could contribute substantially to sediment transport where storm-induced bottom stresses were weak, but that no clear short-term impacts of trawling could be detected on erosion rates of outer shelf sediments. The turbidity induced by passage of trawls is obvious near the bottom but may not be of much direct biological significance except in shallow waters where shading may inhibit primary producers in the photic zone. The immediate effect of the passage of a trawl over the seafloor is to create a pulse of released nutrients. This is then followed by reduced rates of nutrient release to the water column, as biochemical gradients within the sediments are reestablished. The net effect of these initial enhancements and subsequent depressions is probably not large (Krost, 1990). Nevertheless, it is possible that primary production could be enhanced by the greater oxygenation of buried sediments and the resulting faster decomposition of organic matter in the seabed. Furthermore, wherever buried pollutants, such as heavy metals or organic contaminants, are concentrated in the upper sediments of the seafloor, trawling could act to remobilize those contaminants and reinject them into food chains. No study of the Bering Sea explicitly addresses the importance of trawling on these sedimentary processes. Because wind-generated waves during storms cause substantial physical disturbance of bottom sediments, to a depth of even 85 m (Cacchione et al., 1987), it seems reasonable to assume that in the stormy Bering Sea the physical impacts of sediment disturbance by trawls are of relatively little added significance in shallow shelf environments. Bycatch, Discards, and Waste. Fisheries that produce substantial bycatch discards or that kill or even excavate and expose large numbers of marine invertebrates by the disturbance of the trawl seem likely to benefit populations of scavengers that can utilize these food resources. The large growth in populations of scavenging seabirds in the North Sea has been attributed to exactly this process (Cramp et al., 1974; Fisher, 1953; Lloyd et al., 1991). The provision of discards to scavenging seabirds can alter seabird communities by favoring not only those species that scavenge but also the most aggressive of those scavengers. Furness et al. (1988) documented the exclusion of smaller birds, such as herring gulls and kittiwakes, by larger species, such as gannets, in feeding around fishing boats in the North Sea. Such enhanced feeding opportunities for the more aggressive scavengers seem likely to explain, at least in part, the observed increases in their populations. However, increased populations of these larger, dominant scavengers may subsequently have negative impacts on other, less aggressive seabirds

CAUSES AND EFFECTS IN THE BERING SEA ECOSYSTEM 218 through interference competition, nest predation, or resource competition. In the North Sea, declines in herring gulls, lesser black-backed gulls, and kittiwakes may reflect such indirect effects of subsidizing scavenger food chains by trawler discards. Enhanced provision of dead, injured, or exposed animals on the seafloor may likewise be expected to benefit populations of certain demersal scavengers and predators, including various crabs, large scavenging amphipods, and some demersal fishes. To our knowledge, no study has attempted to evaluate this influence of bottom trawling for any fishery. The direct effects on those species caught incidentally as bycatch during bottom trawling operations must vary tremendously among fisheries and species of bycatch. As a rule, marine mammals and seabirds are more subject to bycatch in fixed gears such as gill nets than in trawls (ICES, 1992). However, the bycatch mortality of Steller sea lions taken by trawlers in the U.S. exclusive economic zone was estimated as approximately 989 animals per year by the foreign and joint-venture trawl fishery between 1966 and 1988 (Perez and Loughlin, 1991). The kill was especially high in the early 1980s Shelikof Strait join-venture pollock fishery. Estimated takes were 958 to 1,436 in 1982, 216 to 324 in 1983, and 237 to 355 in 1984 (Loughlin and Nelson, 1986). Most of the animals caught were sexually mature females, which would increase the likely effect of this mortality on the dynamics of local sea lion rookeries. Trawl fisheries of the Bering Sea may also have had some significant effect on marine mammals through direct capture in the nets. While the Bering Sea pollock fishery is reported to be among the "cleanest" of all trawl fisheries—with very low bycatch relative to targeted catch (Alverson et al., 1994)—there is bycatch nonetheless. Depending on species, some fraction of discarded bycatch survives, but death is the more common fate. No systematic assessment of the composition and potential ecosystem consequences of bycatch has been conducted for the Bering Sea trawl fisheries. This assessment may be particularly important for forage fishes, such as capelin and sand lance, which are important foods of both fishes and pinnipeds. THE "CASCADE HYPOTHESIS" AND THE BERING SEA ECOSYSTEM No matter how ecosystems are defined, it seems clear that they are hierarchical products of complex and dynamic interactions between organisms and their physical environments, and that these interactions evolve over time and space as described in Chapter 2. To the extent that these characterizations are accurate, then the current state of an ecosystem is also a product of its history. And this history may affect the future evolution of the system's structure and function. We conclude that significant changes in the Bering Sea over the past 50 years can be attributed in part to the history of human exploitation of living resources in this region, and in part to environmental changes. Both human exploitation and changes in the physical environment have occurred at different spatial and temporal scales, confounding our current best attempts to understand the relative importance of each. The upper trophic levels of the eastern Bering Sea and Gulf of Alaska area have undergone significant changes in the past 100 to 150 years, at least partly due to commercial exploitation of mammals, fish, and invertebrates. Figure 6.8 provides a schematic representation of historical trends in the dominant marine mammal, seabird, fish, and crab populations in the Bering Sea. Although initial population sizes are uncertain, there is general agreement that

CAUSES AND EFFECTS IN THE BERING SEA ECOSYSTEM 219 populations of blue, sei, fin, and humpback whales were severely reduced by the end of whaling in about 1975. These species fed primarily on zooplankton and small, pelagic, schooling fishes. At about the same time (i.e., beginning in 1954), large-scale international fisheries for groundfish developed rapidly, and since 1960 groundfish resources have been exploited intensively (Bakkala, 1993) and have dominated the biomass of living resources removed from the eastern Bering and Gulf of Alaska. Large, mobile trawler fleets exerted intensive fishing efforts on localized fishing grounds for a few years, and then moved on to other species when depletions of target populations became unprofitable (Megrey and Wespestad, 1990). Figure 6.8 Schematic temporal change in relative abundance of certain marine mammals, seabirds, fish, and shellfish in the Bering Sea. Lower bar indicative of changes in sea surface temperature (Committee on the Bering Sea Ecosystem). Coincident with the final decline in baleen whales during the late 1950s, 1960s and 1970s, Pacific Ocean perch was also overfished throughout the region and populations declined dramatically (Figure 4.20). Yellowfin sole, pandalid shrimp, herring, and crabs (except C. opilio) declined dramatically during the 1960s and 1970s. Fur seal populations in the Pribilof Islands started declining beginning in the 1950s, stabilized in the mid-1960s, declined again from the late 1960s to the mid-1980s, and appeared to have stabilized since then. Steller sea lions and some harbor seal populations began to decline in the 1970s, but the decline increased during the

CAUSES AND EFFECTS IN THE BERING SEA ECOSYSTEM 220 1980s. This was mirrored to some extent by declines in murres and kittiwakes in the Pribilofs and Gulf of Alaska. On the other hand, pollock abundance increased dramatically in the Bering Sea during the late 1960s and early 1970s, and again during the early 1980s, along with a number of other groundfish species (e.g., Pacific cod and flounders). Although many of the Bering Sea biological resources were directly depleted by human use, the effects of overfishing on the spatial distribution of these species outside the fishing area, yet still within the Bering Sea ecosystem, are unknown. The following conceptual analysis elaborates on this history and discusses changes in abundance and spatial distribution of four major groups of biota in the Bering Sea over the past 30 years, in relation to environmental or human-induced causation. The four groups are marine mammals (great whales, Steller sea lions, fur seals, and harbor seals), fish-eating seabirds, fish and invertebrates (including Pacific Ocean perch, yellowfin sole, pollock, herring, and capelin), and crab (king and Tanner). Marine Mammals The large baleen whales were commercially harvested in the Bering Sea starting in the mid-1800s. Whaling for the balaenopterids (blue, fin, sei, and humpback whales) and sperm whales was intense from the 1950s through the mid-1970s and resulted in major stock depletions. Whale removals occurred over large spatial areas of the Gulf of Alaska, Aleutian Islands, and Bering Sea as far north as Cape Navarin (Figure 6.9). Within the Bering Sea, whale harvesting was intensive in the middle and outer shelf regions, especially along the shelf break in the region of the green belt (Figure 3.15). As noted earlier, over the last 15 to 20 years, sea lion, harbor seal, and fur seal populations have declined or stabilized in several areas of the Gulf of Alaska and Bering Sea; the exception is southeast Alaska and British Columbia, where both sea lion and harbor seal populations have increased (Figure 6.10). Given the detailed analysis of Chapter 4, what can be concluded about the causes of these recent declines? In the case of Pribilof fur seals, it appears that for a period beginning in the late 1960s, intentional harvests of adult females and reduced juvenile survival account for the observed trend in pup production. Concerning the cause of reduced juvenile survival, it seems most likely that entanglement in debris and reduced food availability (due to climate effects) in the North Pacific Ocean were both significant factors. It does not appear that environmental conditions in the eastern Bering Sea have played a large role in the population decline. Furthermore, it appears that the population decline stopped in the 1980s, at least at St. Paul Island. One of the most striking features of the pattern of decline in sea lion numbers is the fact that the most severe declines have occurred in the core of the species' range from the central Gulf of Alaska through the Aleutian Islands. Although sea lions have been hunted, direct human exploitation is not sufficient to explain the decline that has occurred in their numbers in the Gulf of Alaska and Bering Sea since 1960. Commercial harvesting ended in 1972, and although commercial harvest and incidental takes of sea lions would have contributed to the declines in some regions through the 1980s, they cannot account for the magnitude of the observed declines during these years. Subsistence takes have been relatively small and localized. Incidental takes

CAUSES AND EFFECTS IN THE BERING SEA ECOSYSTEM 221 Figure 6.9 Schematic distribution of great whale declines (Committee on the Bering Sea Ecosystem). Figure 6.10 Schematic distribution of pinniped increases and declines (Committee on the Bering Sea Ecosystem).

CAUSES AND EFFECTS IN THE BERING SEA ECOSYSTEM 222 in fisheries have at times been substantial, but they can account for only a relatively small part of the decrease in numbers. Pollution, disease, and predation do not appear to have been important factors. Available data, though limited, do suggest that food availability may have declined at a faster rate than the number of sea lions, and resulted in reduced condition and reduced survival, particularly of juveniles (Anonymous, 1993). Small walleye pollock were important prey items for juvenile Steller sea lions in the 1970s and 1980s, and reduced abundance of juvenile pollock and other forage in the Gulf of Alaska in recent years could have influenced sea lion feeding. Changes in abundance of other small forage fishes, such as capelin, herring, and sand lance, may also have affected sea lion nutrition. Changes in these fish populations may have been brought about through the trophic effects of significantly reorganized ecosystems in response to historical patterns of exploitation and/or climate-induced environmental effects. For the most part, we do not have the data to assess the relative importance of fishery effects (both direct and indirect) and environmental effects on food availability, but both have likely been involved in the decline of sea lion numbers over the several decades. Finally, there is very little understanding of what factors may be causing the number of harbor seals to decline in parts of Alaska. However, there are some similarities with the Steller sea lion situation in the timing of the declines and the areas affected, suggesting that a decrease in food availability may somehow be involved here as well. Seabirds Adult survival of most seabirds is high, reproductive output is low, and sexual maturity is delayed. Thus, seabirds, like marine mammals, have life history attributes that in the long-term tend to make populations resistant to environmental variability and dampen fluctuations in population size (Furness and Monaghan, 1987). In the short-term, however, populations do vary considerably—in Alaska, there are numerous examples of growing and shrinking colonies of a variety of species (Figure 6.11). In fact, most populations for which we have a history of information spanning at least 10 to 20 years have shown remarkable changes in abundance. Change appears to be common rather than exceptional. The question, then, is ''Are the changes natural?" That is, should we be concerned that any of the changes, such as the broadly based declines on the Pribilofs, are related to factors other than "natural" change in the physical and biological environment of the southeastern Bering Sea? The term natural in this sense excludes human-induced changes, such as, through commercial fishing, oil spills, or other involvement. Or are we seeing evidence of the variability in productivity and abundance of birds that is shaped by changes in weather and climate, but because of our short historical time scale, we cannot yet recognize it as such? If so, what is the evidence? Seabirds in Alaska appear to respond to changes in the environment at interannual, decadal, and perhaps even longer time scales. At the interannual scale, there are many cases of nesting failures of seabirds in particular years, for example, that were clearly caused by shortages of prey during periods of anomalous physical oceanographic conditions (e.g., Springer et al., 1984, 1987). At the longer scale, the overall abundance and distribution of red-legged

CAUSES AND EFFECTS IN THE BERING SEA ECOSYSTEM 223 kittiwakes apparently have declined during this century. They are now absent from the western Aleutians (where they were abundant nesters and apparently outnumbered black-legged kittiwakes in the late 1800s), and from Akutan Island in the eastern Aleutians and Sanak Island in the western Gulf of Alaska (where they also once were common) (Byrd, 1978; Turner, 1885). Another example of long-term change is the decline of least auklets on St. George Island, which now are perhaps at only one-tenth of their abundance of 100 years ago. A long, gradual decline has been attributed to the encroachment of soil and vegetation into the limited talus areas of the island, and the resulting loss of important nesting space (Roby and Brink, 1986). Although there are no clear physical environmental correlates to these changes, it seems likely that they might exist. Figure 6.11 Schematic distribution of fish-eating seabird increases and declines (Committee on the Bering Sea Ecosystem). There are now enough years with information on seabird biology and the physical environment in Alaska that we can see what appear to be coupled decadal-scale trends at several locations. Perhaps the most compelling relationship is between kittiwake productivity and water temperature in the eastern Chukchi Sea (Figure 6.12). The coastal food web that supports the birds in summer is very temperature-dependent, and prey populations tend to grow in proportion to the seasonal rise in water temperature (Neimark, 1979; Springer et al., 1984, 1989). In cold years, when warming is delayed, prey production generally is not sufficient to support high productivity of kittiwakes.

CAUSES AND EFFECTS IN THE BERING SEA ECOSYSTEM 224 Figure 6.12 Productivity of black-legged kittiwakes and sea-surface temperature in the eastern Chukchi Sea (Springer, 1991; reproduced with permission of the Minister of Supply and Services, Canada, 1995). Just the opposite relationship exists on the Pribilofs—kittiwake productivity was highest in the mid-1970s, declined through the early 1980s, and recovered by the early 1990s (Figure 6.13). The same pattern is apparent, but less pronounced, for murres. In this case, productivity is inversely correlated with water temperature, and it seems likely that changes in productivity are also mediated by changes in the food web, although the mechanism in that outer continental shelf community is not obvious. In this case, temperature might only be a proxy for other changes in the physical environment that are responsible for fluctuations in prey availability. The abundance of kittiwakes and murres on the Pribilofs is variable, but the trends are downward in all but one case (Figure 6.14). The "official" time series does not begin until 1976, when the first census was taken, but the decline was first noticed in about 1973 (L. Merculieff, personal communication). For kittiwakes on St. George Island (for which the data are most complete), it appears as if kittiwake numbers and ecosystem productivity might be related, with lag times corresponding to life history characteristics of the two species (Figure 6.15). The longer apparent lag between change in productivity and change in numbers of red-legged kittiwakes compared to black-legged kittiwakes is consistent with other life history characteristics of the two species, which indicate that the former are somewhat more K-selected than the latter. In the western Aleutian Islands, numbers of murres and kittiwakes increased between the mid-1970s and mid-1980s at rates that were similar to the rates of decline on the Pribilofs during

CAUSES AND EFFECTS IN THE BERING SEA ECOSYSTEM 225 Figure 6.13 Kittiwake productivity on the Pribilof Islands and sea surface temperature in the southeastern Bering Sea (three-year running mean of kittiwake productivity, and five-year running mean of sea surface temperature) (Anonymous, 1993). Figure 6.14 Trends in abundance of murres and kittiwakes on the Pribilof Islands (Dragoo and Dragoo, 1994; Climo, 1993).

CAUSES AND EFFECTS IN THE BERING SEA ECOSYSTEM 226 Figure 6.15 Trends in productivity and abundance of black-legged (BLKI) and red-legged (RLKI) kittiwakes on St. George Island (Springer, 1993).

CAUSES AND EFFECTS IN THE BERING SEA ECOSYSTEM 227 the same time (Figure 6.16). In both regions, inflection points in the trends apparently occurred in the mid-1970s and the mid-late 1980s. Contrasting with both is the apparent stability in numbers of murres and kittiwakes at Cape Peirce, a coastal colony in Bristol Bay (Figure 6.17). Productivity of kittiwakes at Buldir cannot be readily assessed, because it has been determined only since 1988. At Cape Peirce, information on kittiwake productivity is fragmentary from 1970 to 1984 but continuous since, and it shows features similar to the Pribilofs case—productivity is comparatively high now after recovering from a low period during the early to mid-1980s. Kittiwake productivity at offshore colonies in the Gulf of Alaska declined to very low levels in the early 1980s, as water temperature rose (Figure 6.18). However, it did not subsequently recover (as at the Pribilofs and Cape Peirce), but has continued to be generally poor, perhaps because water temperature in the gulf has remained warm (unlike the Bering Sea, which has been cooling). Murres in the gulf have also suffered poor productivity in recent years, and declines have occurred both within and outside of the area oiled by the Exxon Valdez spill and were apparent before the spill (Piatt, 1994). The seabird decline in the gulf occurred at the same time as a shift in pathways of energy flow through food webs there. Capelin was the dominant prey of several species of seabirds during the 1970s but was absent, or nearly so, in diets in the 1980s (Figure 6.19). The shift corresponded with the other indicators of a widespread decline of capelin in the Gulf of Alaska and Bering Sea. Capelin was replaced in seabird diets primarily by pollock and sand lance. It is not known if the absence of capelin by itself has been responsible for the poor productivity of some seabirds or is simply indicative of generally poor prey availability. The declines in the abundance of an important forage species coincided with changes in the physical environment of the gulf (as measured by water temperature), trends that together provide support for the hypothesis that some recent dynamics of seabirds, and perhaps other species at higher trophic levels, may be related to physical changes in the ecosystem. Fish and Invertebrates The main target of the developing international fishery in the late 1950s was yellowfin sole, a bentho- pelagic continental shelf species (see Chapter 4). During a brief period from 1959 to 1961, catches averaged over 400,000 t annually. After 1962, the yellowfin sole population and catches rapidly declined to low levels as a result of heavy exploitation. With the crash of the yellowfin sole population, trawl fisheries began to harvest eastern Bering Sea Pacific halibut, Pacific Ocean perch, and sablefish—all long-lived continental-slope species. With the rapid overexploitation of these species and herring (eastern Bering Sea herring biomass decreased to one-sixth of its former levels during the 1960s, as shown in Figure 4.24), foreign trawl fleets moved into the Aleutian Islands and Gulf of Alaska to target slope rockfish and, in particular, Pacific Ocean perch. The latter species was subsequently "mined out" of these regions during the 1960s (see Figure 4.20 for stock biomass trajectories). During the 1960s, as stocks of Pacific Ocean perch declined, foreign trawl fleets shifted to walleye pollock (Alton 1981). Reported catches of pollock in the eastern Bering Sea increased rapidly from 175,000 t in 1964

CAUSES AND EFFECTS IN THE BERING SEA ECOSYSTEM 228 Figure 6.16 Trends in abundance of thick-billed (TBMU) and common (COMU) murres and black-legged (BLKI) and red-legged (RLKI) kittiwakes at Buldir Island (Williams and Byrd, 1992; J.C. Williams, unpublished data).

CAUSES AND EFFECTS IN THE BERING SEA ECOSYSTEM 229 Figure 6.17 Trends in productivity and abundance of common (COMU) murres and black-legged (BLKI) kittiwakes at Cape Peirce (Hagbloom, 1994).

CAUSES AND EFFECTS IN THE BERING SEA ECOSYSTEM 230 Figure 6.18 Productivity of black-legged kittiwakes in the Gulf of Alaska (three-year running mean of kittiwake productivity and five-year running mean of sea surface temperature) (Hatch et al., 1993; reproduced with permission of the Minister of Supply and Services, Canada, 1995).

CAUSES AND EFFECTS IN THE BERING SEA ECOSYSTEM 231 Figure 6.19 Diets of seabirds in the Gulf of Alaska, 1975–78 and 1988–91 (Piatt and Anderson, in press).

CAUSES AND EFFECTS IN THE BERING SEA ECOSYSTEM 232 to 1.9 million t in 1972. Much of this shift was due to the more than six-fold increase in eastern Bering Sea pollock biomass between 1964 and 1971 (Bakkala, 1993). Pollock catches in the Gulf of Alaska increased in the early 1980s, to a peak of around 300,000 t in 1984 (due to a more than three-fold increase in biomass between 1974 and 1982), and then fell to near 60,000 t in 1988 as the strong year classes responsible for the surge in the fishery disappeared. Between the late 1970s and mid-1980s, a number of eastern Bering Sea (EBS)/ Aleutian Island (AI) groundfish and pelagic stocks showed rapid increases in biomass (EBS pollock, 2.6 times; EBS cod, 3.1 times; AI Atka mackerel, 4.4 times; EBS herring, 3 times; EBS arrowtooth flounder, 4.7 times; EBS yellowfin sole, 1.6 times; EBS rock sole, 1.9 times; and other EBS flatfishes, 3.8 times) (NPFMC, 1993). The estimated adult biomass of these species combined increased from around 10 million t in the late 1970s to 25 million t in the mid 1980s. Fisheries on these stocks increased accordingly. At the same time, a number of forage fishes, in particular capelin around the Pribilof Islands and Kodiak Island, declined significantly. King crab populations and their associated fisheries, which had declined in the eastern Bering Sea and around Kodiak Island in the early 1960s, grew very rapidly in the eastern Bering Sea in the 1970s and then crashed in both the eastern Bering Sea and Kodiak region in the early 1980s (Figures 4.1, 4.2, and 4.3). Both C. opilio and C. bairdi populations in the eastern Bering Sea declined in the 1970s; C. opilio populations and landings increased significantly in the late 1980s (Figure 4.1). DISCUSSION The Bering Sea ecosystem, like all ecosystems, has undergone natural environmental changes for a very long time. It has also been significantly affected by human activities for hundreds of years, more significantly in the past 200 years (Chapter 5). Therefore, the changes of the past several decades that have been discussed in some detail in this report cannot be interpreted as perturbations to a ''pristine" ecosystem, and they are not unique in many ways. However, until recent decades, good information on many of the major species in the Bering Sea was lacking. Thus the "cascade" changes that began in the 1950s described below probably are only the most recent in a series beginning no later than the nineteenth century. To recapitulate our understanding of the most recent set of changes, from the mid-1950s to the early 1970s, large populations of both fish and mammals (particularly the large whales) were dramatically reduced. This reduction in biomass likely increased the amount of food (zooplankton and small fish) available to other vertebrate predators. Unfortunately, there is not sufficient information available on a time series basis for this period to examine the abundances and distributions of such organisms to verify this. However, it seems reasonable to suggest that the dramatic increase in the abundance of pollock, which apparently occurred during the late 1960s in the eastern Bering Sea, may have been in some way linked to overexploitation and reduction of these other populations. In fact, there is little doubt that the eastern Bering Sea fish assemblage switched to a pollock-dominated system in the late 1960s and early 1970s, and that this domination of pollock biomass has persisted since then. What other changes might have been associated with the concentration of resources into pollock biomass? One effect might have been reduced abundance of forage fish species, such

CAUSES AND EFFECTS IN THE BERING SEA ECOSYSTEM 233 as capelin and sand lance, which compete with young pollock for zooplankton. Unfortunately, it is not possible to evaluate this hypothesis, even in a general way, because we have few data on trends in the abundance of these fishes and other forage species. A major climate regime shift occurred in the North Pacific in the late 1970s. Combined with the historical restructuring of parts of the large oceanic ecosystems of the region, this climate shift could have set the stage for the rapid changes in top-level predator populations that seem to have accelerated in the last 15 to 20 years. There is persuasive evidence of very large increases in piscivorous fish populations resulting from the synchronous strong year classes that occurred in the late 1970s, just after the regime shift. Merrick (1995) proposed a very similar hypothesis to explain the recent declines in the Steller sea lion population of the northeast Pacific and eastern Bering Sea. He suggested that there had been a major shift in the trophic structure of the Bering Sea (and probably the Gulf of Alaska) beginning in the mid-1960s, "a decade prior to the initiation of the current fishery management regime." He argued that if there was an actual cause- effect relationship, it lay in the dietary overlap between whales, fish, and some pinnipeds. The significant increase in pollock biomass of the 1960s (eastern Bering Sea) and early 1970s (Gulf of Alaska) (see Figure 4.4) could then have resulted from the release of euphausiid and calanoid copepod prey (preferred prey for juvenile pollock) due to the fishery removal of whales (fin, sei, and humpback), Pacific Ocean perch, and herring, and the reduction of juvenile pollock predators due to the significant harvest of northern fur seals. Merrick estimated that "the reduction of fin whales, Pacific herring and Pacific Ocean perch in the Bering Sea and Aleutian Islands could have released 1.36 to 2.81 million mt of zooplankton prey a year…which would be sufficient to feed 4.3 to 8.9 billion age-1 pollock for a year." In addition, he estimated that the reduction in fur seals could remove predation on 2.8 billion age-1 pollock. He concluded that the released zooplankton and reduced predation pressure could have supported pollock abundance at the high levels resulting from recent high recruitments. He also reported that around the time of the 1976–77 climate regime shift, preferred prey (forage fish such as capelin and sand lance) began to disappear in a serial fashion as a number of adult groundfish populations, in particular Pacific cod and a number of flatfish species, grew significantly. In summary, Merrick concluded that the combination of favorable ocean conditions, continued zooplankton release, and low predation in the late 1970s and early 1980s resulted in a series of strong pollock year classes. As a result, adult (but not juvenile) biomass again increased and remained high in the eastern Bering Sea (perhaps due to an effective fisheries management regime using light exploitation rates), and predatory Pacific cod and flatfish populations also increased. These fishes reduced the populations of small fish prey such as capelin and sculpins. The reduced biomasses of forage fishes and juvenile pollock abundance were insufficient to meet the dietary needs of Steller sea lions. One piece of this scenario that is not fully understood is why juvenile pollock have been less abundant recently. One possibility is cannibalism, as well as predation by Pacific cod and flatfishes. Another possibility is that although recent environmental conditions have favored pollock recruitment, there have been a few large year classes (especially 1978), rather than a steady supply of juveniles, and the lack of other forage fishes has made the variable supply of juvenile pollock more of a limitation for marine mammals and birds than it normally would have been. Indeed, the year-class strength of pollock has declined recently. What does seem likely

CAUSES AND EFFECTS IN THE BERING SEA ECOSYSTEM 234 is that a combination of cascading trophic interactions and a significant climate shift has caused major—and perhaps irreversible—restructurings of the oceanic ecosystem of the eastern Bering Sea, Aleutian Islands, and Gulf of Alaska. There are several important questions to consider when analyzing the pattern of declining and changing marine mammal, seabird, and fish populations in the Bering Sea: 1. Can we expect species to recover from a perturbation (either human-caused or environmental) in a changing ecosystem? For example, consider the large-scale removal of whales and predatory fishes from the Bering Sea. This removal may have resulted in an explosion of one or more prey species (e.g., walleye pollock), as well as an increase in other predators or increased concentration on pollock by predators that feed on these prey, including humans. The resulting reordering of the competitive food web would likely be characterized by an increase or decrease of many of their other components (e.g., crab, sea lions, and seabirds). As a result of this complex reorganization, whales may encounter new competitors for their prey base (e.g., factory trawlers with sophisticated electronics and huge nets) that limit their recovery notwithstanding direct regulatory protection. 2. Time lags and different time scales can produce nonlinear relationships that make it very difficult to explain the changes that occur at any one time. For example, are we still seeing changes in the Bering Sea ecosystem caused by baleen whale removals in the 1960s and 1970s? Was the large pollock year class of 1978 a result of prey release or competitive interactions at lower trophic levels that enabled pollock populations to explode? Or was it a response to warming seawater temperatures after the 1976 regime shift, allowing an increase in water column production and juvenile success? Were both factors partly responsible? Is the current organization of the ecosystem stable or is it likely to shift again, either to its previous condition or to some other condition? 3. At any one time the system may be controlled (limited) by a few key factors (Berryman, 1993). Because these key factors are generally unknown, it is not often possible to know a priori when humans influence one of them, setting off a series of unexpected, cascading changes throughout the ecosystem. This "cascade hypothesis" (described in Chapter 2) might explain some of the rapid and then persistent changes observed in crab, pandalid shrimps, demersal fish, seabird, and marine mammal populations and assemblages. The key question is how this knowledge can be used to manage human actions in order to reduce our impact on the ecosystem. CONCLUSIONS As a general conclusion, the most likely explanation of events in the Bering Sea ecosystem is that a combination of a decadal or regime shift in the physical environment acted in concert with human exploitation of predators (whales, other fish) to cause pollock to dominate the ecosystem and other predatory fish populations to greatly increase in abundance. As a result, some forage fishes that have higher nutritional value than pollock became less available to some marine mammals and birds, leading to their decline. The increase of adult pollock and other predatory fishes in the past 20 years might also be responsible for keeping the forage fishes

CAUSES AND EFFECTS IN THE BERING SEA ECOSYSTEM 235 relatively scarce. This food shortage might have been exacerbated by pulse fishing of pollock, which might have removed them from some areas for long enough to cause difficulties to marine mammals and birds, especially juveniles. There is no evidence that any human activity other than fishing (including whaling) has had a significant effect on the Bering Sea ecosystem, although better information on various pollutants would be of interest. We emphasize that although this is the most likely explanation or description of events over the past 50 years, it is not one in which there can be a high degree of confidence at present, and it might well be partly but not completely correct. Indeed, until we know more about exploitation rates in the western Bering Sea, we can probably never be very confident of any scenario. Nonetheless, it is worth exploring this scenario in a little more detail. Physical changes have occurred in the Bering Sea ecosystem over the past 50 years. In addition to interannual variations, decadal or regime shifts have occurred; these have included changes in sea surface temperature, the mean extent of ice cover, and atmospheric and oceanic circulation patterns. It is most likely that these changes have altered the distributions and abundances of many members of the ecosystem, especially short- lived species. One of the largest such changes occurred in the winter of 1976–77; since then, the eastern Bering Sea has had warmer surface temperatures and less ice cover on average than in the period before 1976. The brief period of warmer sea temperatures in the 1960s (Figure 6.5) could also have contributed. Human impacts were also being felt by the Bering Sea ecosystem during the past 50 years—indeed, for the past 200 years at least, and probably much longer (Chapter 5). Intensive hunting of whales continued, along with a significant harvest of fur seals. Heavy exploitation of yellowfin sole caused rapid declines in the population in the early 1960s; other long-lived slope species, such as halibut, Pacific Ocean perch, and sablefish, along with herring in the eastern Bering Sea, also suffered population declines as a result of exploitation. It is extremely plausible that the removal of whales and various fish species that were predators on or competitors with pollock combined with a decadal regime shift to cause a large increase in pollock recruitment. As result, pollock populations increased (as did populations of Pacific cod and some predatory flatfishes, especially arrowtooth flounder), and they came to dominate the ecosystem. It is also plausible—but completely untested—that increased competition from young pollock for zooplankton or increased predation by growing populations of piscivorous fishes led to the decline of important forage species such as capelin, sand lance, and squid. There is persuasive although not conclusive evidence indicating that some marine mammals—especially Steller sea lions in the eastern Bering Sea—have and some seabirds—especially murres and kittiwakes, particularly around the Pribilof Islands—have suffered from food shortages, probably affecting juveniles more severely than adults. Indeed, it seems that some populations have decreased as pollock biomass has increased. It thus appears that the declines were not caused by declines in total abundance of pollock (which have not declined), or even in total abundance of all prey species, although declines in total abundances of capelin, sand lance, squid, and perhaps of juvenile pollock might have contributed to their difficulties. But, especially since the late 1970s, fishing for pollock has been concentrated at some times in some places. This would imply that pollock are effectively removed from some areas at some times, and the local populations would probably take at least days or weeks to be rebuilt

CAUSES AND EFFECTS IN THE BERING SEA ECOSYSTEM 236 by in-migration from elsewhere. It is thus possible that food shortages for some mammals and birds—perhaps at crucial times and places for juveniles—have been exacerbated by this intense pulse fishing. Even if a general increase in pollock has caused a decrease in some other, perhaps more valuable, forage species, it is almost certain that removing pollock from an area could not result in a rebound in the populations of those other species in only a few weeks. As long as overall pollock biomass remains high, it is plausible that the biomass of those other species will remain low. In any event, the above scenario—the most plausible that the committee could identify—does not lead to the conclusion that pollock have been subjected to ecosystem overfishing. In other words, it is hard to see how the total rate of exploitation of pollock over the past 25 years is directly (or even indirectly) responsible for the decline of mammals and birds. If this is true, it is then impossible to see how reduction of the total rate of exploitation of pollock would be helpful in the short-term; it is even possible, although highly speculative, that some mammals and birds would be helped by a temporary increase in the exploitation of pollock. It is more likely that marine mammals and birds have been affected by the distribution in space and time of fishing effort on pollock, and thus that they would be helped by a broader distribution of fishing effort in space and time, especially in areas where they are known to feed. We caution that it is by no means certain that this would be effective enough to reverse or even halt their population declines. It is also hard to predict the effects of protecting other marine mammals. Many whale species have been recovering outside the Bering Sea since whaling ceased or was reduced, and some of those are the ones whose declines might have encouraged pollock increases. It is not clear why recoveries of those species in the Bering Sea have not been observed (Chapter 4), but recoveries are possible. Perhaps if they increase and human exploitation continues, the dominance of pollock in the ecosystem will be reduced, particularly if environmental conditions do not favor their recruitment. Perhaps some of the shorter-lived forage species such as capelin and sand lance would increase quickly enough under those circumstances to provide a significant food source for mammals and birds. At least we can say that it seems extremely unlikely that the production of the Bering Sea ecosystem can sustain current rates of human exploitation of the ecosystem and the populations of all marine mammal and bird species that we believe existed before human exploitation—especially modern exploitation— began. Over the long-term, fishing competes to some degree with at least some top-level predators; therefore, if the goal of management is to have as many top-level predators as possible (not a common management goal), then it seems that ultimately fishing will have to be reduced. We note that "fishing" includes commercial, subsistence, and recreational fishing, all of which can compete among each other as well as with marine mammals, birds, and predatory fishes. Even commercial fishing can be divided into various sectors. Allocating effort among these competing groups remains a major challenge to managers. Finally, we emphasize how difficult it will be for human management to cause a large, complex marine ecosystem to achieve and maintain a desirable balance. If the above scenario is even partly correct, it is clear that many significant factors have influenced the Bering Sea ecosystem over the past 200 years, only some of which have been under human influence (let alone control) and only some of which have been adequately documented. Some of the changes that have occurred might be irreversible over normal human time frames (say 100 years or less).

CAUSES AND EFFECTS IN THE BERING SEA ECOSYSTEM 237 This difficulty does emphasize the need for an adaptive approach to management and the need for good, long-term data on physical and biological phenomena.

Next: 7 Implications for Management Policy and Institutional Arrangements »
The Bering Sea Ecosystem Get This Book
×
Buy Paperback | $85.00 Buy Ebook | $69.99
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

The Bering Sea, which lies between the United States and Russia, is one of the most productive ecosystems in the world and has prolific fishing grounds. Yet there have been significant unexplained population fluctuations in marine mammals and birds in the region. The book examines the Bering Sea ecosystem's dynamics and the relationship between man and the ecosystem, in order to identify potential reasons for the population fluctuations as well as identify ways the Sea's living resources can be better managed by government.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!