Click for next page ( 12


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



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

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

OCR for page 11
Marine Ecosystems . A Conceptual Framework Before attempting to define the Bering Sea ecosystem and its analogs, it is useful to examine the nature of oceanic ecosystems from a conceptual point of view. The approach taken here is not the standard "ecosystem ecology," but rather a practical hybrid combining dynamics (from population biology), a multispecies perspective (from community ecology), and an acknowledged role for the effects of physical environmental fluctuations (an ecosystem perspective). What we attempt to do is transmit to the reader a sense of an "ecosystem perspective" for management a context that recognizes the interactions among biotic and abiotic elements of ecosystems and their effects on human institutions. This chapter provides a conceptual framework for oceanic ecosystems with a particular focus on the Bering Sea and introduces three conceptual questions that are important to later . c .lscusslons. Are there fundamental differences between oceanic and terrestrial ecosystems and if so, what are they? How does one deterrence the relationship between natural and anthropogenic forcing in an ecosystem context? How does one assess the "condition" of an ecosystem? The chapter then discusses ecosystem management and explores whether it can be seen as a solution to perceived problems in the Bering Sea. Finally, it describes the Bering Sea ecosystem and its analogs. CONCEPTUAL FRAMEWORK Ecosystems and the Concept of Order Understanding an ecosystem first requires some justifiable delineation of what constitutes the ecosystem. The complete description of biotic and abiotic variables that interact to define a functional ecological system represents a tremendous challenge. Ecologists recognize that 11

OCR for page 11
12 The Bering Sea Ecosystem global interdependencies operate among atmospheric, hydro spheric, and lithospheric processes, such that no absolute independence exists for any given ecosystem defined at a subgIobal scale. Yet ecosystems are operationally defined and meanin~fuliv .sill~lierl on more Mann amble scales smaller than the entire planet (Steele 19741. ~, _ _ ~_ _ ^.,~.~e,_~._ , , , Such delineations of ecosystems are made in practice by setting explicit geographic bounds around some subset of nature. For example, Sherman and Alexander (1986) identified several separate large marine ecosystems (LMEs) as appropriate units for analysis and management in the marine environment. These delineations of ecosystems are based on distinct hydrographic boundaries, bottom topography, and trophic dependencies of interacting populations. Such abstractions of nature not only are necessary to allow feasible study of community and ecosystem processes but also acknowledge that the earth's biota is spatially partitioned into meaningful groups of interacting species. Nevertheless, despite success at setting spatial bounds to ecosystems that may correspond well to biogeographic, habitat, and hydrographic distinctions, it is still unportant to recognize and analyze those processes that cross the boundaries. The necessary practice of setting bounds to an ecosystem before it can be meaningfully studied and analyzed is actually just the first step in the process of abstraction, averaging, and pooling that must be done to simplify the ecosystem for effective study of organizing processes (Levin, ~ 992~. The complexity of any natural ecosystem is just too great to allow even its complete description, let alone meaningful analysis of its individually recognizable components. Decisions about which variables to pool or average and over what scales in time and space are critical to the success of efforts to understand the dynamics of ecosystems and the causes of underlying patterns. These decisions are informed and guided by the results of conceptual and mathematical models, which should be constructed and iteratively modified in collaboration with empirical observations and experiments (Kareiva and Anderson, 19881. Unlike many other aspects of ecology, modeling is a necessary part of ecosystem study in that it identifies those areas of most sensitivity and critical Importance for unproved measurement and the time and space scales needed for those critical measurements. In addition, the models sharpen our understanding of nature by making processes explicit and by allowing some degree of prediction, which cannot be achieved through description of patterns alone. Interestingly, the supplest models without elaborate reductionist detail are often the most reliable predictors of ecosystem pattern (Ludwig and Walters, 19851. Mathematical modeling of the effects of interactions between species on the dynamics of species populations has a long history in ecological research. Some of the earliest work was done by Lotka (1925) and Volterra (1931) on predator-prey interactions and by Gause (1934) on effects of competitors. These approaches were later expanded by May (1973), Caswell (1978), Yodzis (1988), and others to encompass more complete communities of interacting species. Perhaps the major cross-cutting challenge to scientists interested in evaluating ecological patterns is the integration of the multiple spatial and temporal scales involved (Levin, 1992). There is no single "correct" scale at which to view ecosystems. Processes that occur at one scale have impacts observable on other scales. Patchiness on virtually all scales is a

OCR for page 11
Marine Ecosystems: A Conceptual Framework . . . .. ~ 13 characteristic of ecosystems that complicates their complete description and poses an unplicit challenge to ecologists to explain these patterns (Denman and Powell, 19841. Our snapshot views of temporally dynamic ecological systems represent one of the most serious limitations to proper understanding of causes of ecosystem dynamics. The evolutionary history and the more immediate past in ecological time are important contributors to the patterns observed in nature. Yet the historical memories present in any ecosystem are difficult or impossible to extract in the short time horizons of both management institutions and scientific study plans. Because of the significance of processes on multiple spatial and temporal scales to the generation of ecological pattern, and especially the unknown contributions of historical events, scientists studying ecosystems are always faced with incomplete and imperfect information. As a result, ecosystem processes reside in what Magnuson (1990) called "the invisible present, " hidden from view or understanding because they occur slowly or because effects lag years behind causes. Pomeroy et al. (1988) elaborated on this theme and emphasized the concept of scale as being critical to the understanding of ecosystems. One feature of the biosphere that clearly emerges from ecological studies is the range of scales of space and time over which events occur. Each kind of event may have a characteristic time or half- time, ranging from nanoseconds for intracellular processes...to millions of years for the evolution of continents and ocean basins and their contained ecosystems. We are really dealing with a continuum, of functional response over a range of time scales, and it is often necessary to think about a substantial part of that range in order to properly describe and understand ecological processes. Central to conventional ecosystem theory is the concept that structure (the way process distributes itself in space) results from differences in process rates. Thus, according to this theory, ecosystems are structured according to the relationships between processes occurring at different time and space scales. Although organisms as taxonomic entities are not the essential "stuff" of ecosystems, their populations, as variable conduits, clearly define the fundamental nature of ecosystem dynamics. It is the populations of species that participate in creating or modifying energy fluxes that, as a result, occur at different rates in space and time. What we are talking about here is ecology in four dimensions. An ecosystem cannot be defined or conceived of without clear reference to tune. In the words of Allen and Hoekstra (1992),"... complex behavior will arise when very different rates are pressed together in the formulation of a population equation. . Complexity arises from the interaction of differently scaled processes. . . " The ecosystem is thus a hierarchical system that responds differently over a range of time and space scales. Lower-level components (those operating at short time scales) respond significantly to disturbances at higher levels. However, higher levels "see" only the averaged or integrated responses of lower-level components. This does not mean that everything in an ecosystem occurs in a top-down fashion. O'Neill et al. (1986) pointed out that in hierarchical organizations, lower-level behaviors are essential to the functioning and persistence of higher

OCR for page 11
14 The Bering Sea Ecosystem level structure, although, because of their slower rates, they are unable to affect the behavior of the higher level. Apollonio (1994) brought hierarchy theory into the realm of marine fisheries. He distinguished between long- and short-lived species in terms of their ecosystem-organizing capacities as follows: "~Long-lived] species, because of their biological attributes, have a role in structuring or shaping their systems...That is to say, their presence induces predictability into the system. And, by inference, their reduction or removal reduces internal control mechanisms and increases variability and unpredictability." The point that Apollonio (1994) was making here is that, according to hierarchy theory, ecosystems are structured in part by the natural frequencies of their component parts. "Structuring" tends to occur as a function of these cycle times or natural frequencies. The implication is that ecosystems that are dominated at the top by long-lived species tend to be more "structured" and predictable than those that are not. It is interesting to chronicle the history of fishing in the Bering Sea in the second half of the twentieth century with respect to these concepts. Whales and seals were harvested first; then rockfish such as Pacific Ocean perch; then herring, crab, and yellowfin sole; and then pollock. The interesting thing is Me apparent succession in the targets of harvest from long-lived to short- lived species. Because the exploitation history affects the organization of a marine ecosystem, it should be taken into account in attempting to develop ecosystem management policy. Cascading Trophic Interactions in Ecosystems Development of the science necessary to support ecosystem management probably will require a new synthesis of previously separate approaches to the study of ecology a melding of relevant aspects of population ecology, community ecology, and traditional ecosystem ecology. The goal is to be able to predict, understand, and interpret changes in populations of component species of an ecosystem, especially those that have value to human society. Such understanding must be based in large part on knowledge of the interaction of those important populations with other species in the ecosystem. Such questions are the essence of community ecology, and the experimental work done on trophic cascades by Paine (1980) in marine intertidal systems, the research done to evaluate the indirect effects of removal of sea otters from subtidal communities (Estes and Palmisano' 1974; Sunenstad et al.7 1978)' and the experimental manipulations by Carpenter and colleagues (e.g., Carpenter et al., 1985; Carpenter and Kitchell, 1988) in lakes represent the best examples of how community ecology must be involved to develop appreciation of processes controlling dynamics of individual species. Parsons (1992) presented examples from the Atlantic, Pacific, and Antarctic oceans where the removals of marine predators by fisheries have had various cascading short- and long-term effects on ecosystems (e.g., the effects of the harvest of antarctic blue whales on changes in abundance of other krill-eating and krill-dependent species and the inverse relationship between the abundance of planktivorous right whales and the abundance of planktivorous sand lance in the Gulf of Maine). These reports tend to substantiate the existence and importance of top-down effects in structuring marine ecosystems. In addition' they demonstrate that these effects can occur at a number of different hme scales.

OCR for page 11
Marine Ecosystems: A Conceptual Framework The Effects of CInnate on Structuring Marine Ecosystems 15 Several marine ecosystem scientists have written that the nonlinear and seemingly unpredictable nature of marine ecosystems is a manifestation of the capacity for self-organization (Allen, 1985; Mann, 1988). The classical example of this kind of behavior is the so-called Russell Cycle of the western English Channel (Figure 2.1) (Mann, 1988; Mann and Lazier, 1991). In essence, the Russell Cycle involved a major flip in the 1930s in the western English Channel from one fairly stable configuration of organisms to another quite different mix of species; then the system reversed itself in the late 1960s. These flips seem to have been induced by changes in climate associated with the North Atlantic Oscillation. The essence of this theory is contained in Allen's (1985) interpretation of the work of Ilya Prigogine (see, for example, Nicolis and Prigogine, 1977; Prigogine, 1980) on chemical systems. Nicolis and Prigogine consider ecosystems to be thermodynamically open dissipative systems. Natural systems are open, usually have strong inputs and outputs, and have strong internal coupling between elements. Although they may appear to be stable (dynamic stability), they are in fact accompanied by incessant fluxes of energy and materials and are maintained far from equilibrium. Prigogine coined the term dissipative for structures that are formed in response to the fluxes themselves. They are formed within the system and obtain the energy for their formation from the dissipation of energy flowing through the system. In fact, the flow of substances through the system can cause tension in the organization of the system, similar to the tension that occurs between tectonic plates prior to an earthquake. In this way, systems can flip from one state to another, causing quite drastic alterations in system structure, and yet the alterations in system structure may be triggered by changes in system fluxes that are themselves relatively small. In the words of Margalef (1986), the impetus for these changes in system fluxes can come from climate-induced energy "kicks" which tend to disrupt or decouple a number of ecological relationships within the ecosystem. Summary Marine ecosystems are complex structures in which significant (re-)ordering occurs in both top-down and bottom-up directions. The key aspects are the roles of rates (e.g., natural frequencies) and time (e.g., cascading trophic interactions and infrequent but often rapid changes in physical forcing due to climate change). In addition, most of the marine fisheries ecosystems with which we are presently concerned have very substantial pieces missing; i.e., major components with which the current fish community evolved and interacted are no longer there in any significant numbers (Apollonio, 1993). Some populations, such as large whales, have been harvested to the point of extinction. Others, such as marine fish, have both been reduced in absolute abundance and had their population structure altered by the significant truncation of their age structure. Because of its reactive nature, fisheries science has seldom if ever considered a marine community before exploitation and frequently does so only after heavy exploitation. This combination of factors has significant effects on the Bering Sea ecosystem.

OCR for page 11
16 2500 2000 1 500 o. ~ 1 At 500 The Bering Sea Ecosystem ~ 3171 - '~ \^ Total fish larvae ~,\/~ Callionymus All/ ~ flatfish - ~:Gadoids _ ~ ~ O 1 ~. ~ 1 930 Cal/ionym~j | flaffish ~ ~1 1 1960 1970 (iadoids l 1940 1 950 Figure 2.1 The Russell Cycle of the western English Channel (Mann and Lazier, 19911. OCEANIC AND TERRESTRIAL ECOSYSTEMS . - . . . . . . In thinking about the Bering Sea ecosystem, it is important to recognize that extrapolation from terrestrial to marine, and especially to oceanic, ecosystems can be misleading. In particular, marine and terrestrial ecosystems differ markedly in their functional responses tO environmental change (Steele, 1985, 19911. Important differences include the following: _ _ _~ _ _____ ~--r ~ ~- - i--- Random environmental forcing of terrestrial ecosystems shows constant variance per unit frequency out to about 50 years, while that of oceanic ecosystems shows increasing variance with decreasing frequency from hours to millennia. Thus, terrestrial ecosystems for periods up to 50 vears are inherentIv more predictable than oceanic ecosystems. . J - ~ ~ ~ On time scales of years to decades, oceanic systems are more closely coupled to variations in their physical environment than are terrestrial systems. Unlike many terrestrial ecosystems, natural fluctuations in oceanic ecosystems between states of high and low abundance are usually related to long-term climate trends, rather than to internal dynamics. Since the amplitude of environmental variability is much smaller in the ocean than on land, marine organisms have less robust internal processes to handle variability at the short periods commensurate with their life spans. Marine populations have different ways of dealing with short-term variance and also respond differently at longer time scales. Except at the benthic or coastal boundaries, components of oceanic ecosystems

OCR for page 11
Marine Ecosystems: A Conceptual Framework 1 7 are subject to circulation and mixing, which affect their dispersal, migrations, and aggregations. In particular, there is a close interaction between reproductive processes and particular patterns of ocean currents and mixing. Because most organisms in oceanic ecosystems tend to move or to be moved extensively in three dimensions that include the water column, and they move between trophic levels as they develop, oceanic ecosystems are much less well defined than terrestrial ecosystems and at a given location tend to be more changeable than those on land. As a consequence of these differences, Steele (1991) concluded that it is inappropriate to apply terrestrial perspectives to marine communities, especially in the context of management or conservation. In particular, he noted: Large switches in open sea marine communities can last several decades. While these changes can have major economic consequences, they cannot be considered as ecological disasters or even as being deleterious in any way within the marine systems. Such switches have occurred without human involvement, for example before heavy fishing was a factor, but they may be increased in frequency or amplitude by human actions. However, this is not to suggest that fishing cannot significantly affect the composition of marine fish communities. Events on Georges Bank off New England and Nova Scotia provide a rather convincing example of the effects of intensive exploitation on the composition of fish assemblages. Most of the important groundfish species on Georges Bank have been seriously overfished, and the high fishing effort continued well into the 1990s, despite declines in the catch per unit effort (NOAA [National Oceanic and Atmospheric Administration], 1993). In addition, dogfish and skates increased in numbers and biomass while groundfish and flounders declined; in 1963, dogfish and skates were only about 25 percent by weight of the fish surveyed, while in recent years they have increased to 75 percent (NOAA, 1993). ENVIRONMENTAL AND ANTHROPOGENIC EFFECTS Oceanic ecosystems are affected by environmental change as well as by human activities such as fishing. The oceanic environment changes over a broad range of time scales. In considering large systems, of particular interest are periods of years to many decades; changes over such long periods are usually referred to as climate changes. The surface layers of the ocean (to 100 m or so) are the principal habitat for organisms in most oceanic ecosystems. Changes in atmospheric circulation and the exchange of energy between the atmosphere and the sea surface cause horizontal flow, vertical mixing, and variation in temperature and salinity of the surface layer. Temperature also serves as an index of more fundamental change in ocean conditions, apart from its direct effects. The direct effects of climate changes in near-surface temperature (and to a lesser extent, salinity) are principally on distributions of organisms. Thus, prolonged warming of the surface

OCR for page 11
18 The Bering Sea Ecosystem layer can be associated with invasions of warmth-preferring organisms and reductions in abundance or relocation of the system's usual residents (see, for example, Wooster and Fluharty, 1985~. In some cases, regime shifts can occur. For example, after the collapse of the Peruvian anchovy population in 1973, the continuing warm conditions allowed a sardine-mackerel community to establish itself for a number of years (Pauly et al., 19891. Distributions are also affected by changes in horizontal flow, as are locations of frontal and more diffuse transitional zones between ecological provinces. For example, the Russell Cycle could be interpreted as a decadal-scale relocation of the transition zone between temperate and boreal provinces. Locations and intensities of zones of convergence ant} divergence are also affected, the latter being related to the supply of nutrients to the euphoric zone. Thus, environmental change can lead to changes in phytoplankton species composition and the primary productivity underlying the ecosystem. Although the mechanisms have not been well clarified, the effects of environmental change on recruitment have been established (see, for example, Hollowed et al., 19871. A study group of the International Council for the Exploration of the Sea (ICES) (anonymous, 1992) concluded that most fluctuations in fish populations seem to be driven by recruitment rather than by exploitation, with the major fluctuations in recruitment (and subsequently in biomass) being more successfully linked to physical oceanography than to the biological consequences of fish catches. Nonetheless many marine and anadromous fish copulations have been suhstantiallv , . . . ^. . . , ._` _ , . .. ~ ~ ~ . . _ reduced by tensing te.g., And [National Research councils, lYY4a, lY96; NMFS, 1992~. In fact, as discussed in more detail below, it is plausible that anthropogenic removals from the northeast Pacific, Bering Sea, and Aleutians in the 19SOs and 1960s could have caused a rapid increase in pollock populations, which, in turn, could have caused the current declines in sea lion and bird populations of the region. Particular sets of environmental conditions appear to favor the production of strong year classes, and different species have different environmental preferences. Thus, over periods of several years, the mix of species in an ecosystem can be altered by environmental change (Hollowed and Wooster, 19941. This effect is, of course, in addition to that caused by removal of selected species by fisheries. It is often alleged that overfishing is a major cause of ecosystem damage. Certainly large-scale fishing operations reduce populations of target and incidentally caught species and thus affect predators, prey, and competitors of these species. These effects are commonly seen as evidence of overfishing. For example, Goni et al. (1993) stated, "After several decades of intense fishing in the Eastern Bering Sea, the Aleutian Islands, and the Gulf of Alaska, a number of fish stocks have been overfished, with signs of depletion and ecosystem disturbance apparent in associated marine populations." Although the meaning of the term is ambiguous, overfishing in the context of fishery management is generally defined as "a [fishing] rate in excess of sustainability 'I (Rosenberg et al., 19931. In 1992, 95 stocks were managed under federal fishery management plans and had approved definitions of overfishing under the Magnuson Fishery Conservation Act Management Act. Of these, 68 stocks (72 percent) were defined as overfished on the basis of fishing rate without reference to stock abundance. A related term, overutilization, is applied when more fishing effort than is necessary is used to achieve long-term potential yield (NMFS, 1992~.

OCR for page 11
Marine Ecosystems: A Conceptual Framework 19 While from an economic point of view such overutilization is certainly inefficient, it would not necessarily appear to threaten the continued viability of overutilized stocks. From a biological point of view, species overfishing can be defined as a level of exploitation that jeopardizes the ability of a stock to replace itself. Ultimately, this concept of overfishing implies the threat of species extinction, or at least reduction below a minimum viable population. Although this is an unlikely outcome with respect to oceanic finfish, it could apply in the case of marine mammals, as it did, for example, to the Steller sea cow, which became extinct through exploitation by humans. It is often difficult to tell whether declines in commercial fish populations are due to fishing, environmental changes, or both. Species with relatively short life spans, such as sardines, anchovies, and herrings, are well known to experience large fluctuations in abundance. Populations of such species can decline rapidly even in the absence of fisheries and, under favorable environmental conditions, can quickly recover to high abundance despite very low stock levels caused by either heavy fishing or natural variation (Baumgartner et al., 1992). The ICES study group (anonymous, 1992) concluded that "the collapse of the North Sea herring in 1976-1978 has been linked to overharvesting. However, the recruitment failure which followed can be explained by anomalies in North Sea circulation affecting recruitment processes, as well as by depressed spawning biomass. Whatever the cause, the stock rebounded strongly less than half a decade later.' It can be argued that species overfishing has also occurred when the rate of recovery from low levels has been significantly slowed. This is a problem particularly for the case of stocks with relatively long replacement times. A classic example is Pacific Ocean perch, which was reduced to very low abundance in the northeastern Pacific Ocean by heavy fishing and is only slowly recovering. In this instance, the reproductive potential survived the heavy harvest, as evidenced by the recent increase in population, but it seems clear that the fishery could have been managed to produce more fish at a steadier rate than actually happened. For a definition of species overfishing to be based on rebuilding, it is necessary to know how a rate compares with rebuilding in the absence of fishing. Applying the definition would be difficult because of interspecies differences in replacement times, frequency of strong year classes, and response of recruitment to biomass reduction and to environmental change. In most cases where overfishing and mismanagement have been blamed for declines of nontarget species, as for the pollock fishery and Steller sea lions in the eastern North Pacific, the effects (alleged or actual) have been indirect, and ecosystem (or ecological) overfishing might be the appropriate term for such effects. Reduction of one component of an ecosystem by fishing can have consequences for other components, especially for predators, competitors, and prey of the target species. Predictability of these consequences is poor for several reasons, including the nonlinear and poorly understood relationships among the components and the accompanying effects of natural variability forced by environmental change. This discussion suggests that many cases of alleged overfishing, although serious from an economic or social point of view, do not necessarily present any threat to the continued existence or replacement of a target species (with the exception of marine mammals). Ecosystem overfishing, on the other hand, can also have a variety of serious effects, but it is difficult to demonstrate because the effects of significant removals need not be direct and are not always consistent with intuition.

OCR for page 11
20 The Bering Sea Ecosystem THE CONDITION OF AN ECOSYSTEM It has been charged that the Bering Sea ecosystem has been damaged by human activity. This charge is based on (~) the greatly increased level of human activity, especially commercial fishing, since the 1960s, and (2) unexpected and unexplained changes in important components of the ecosystem (e.g., some pinnipeds and seabirds). Indeed, human removals from certain populations have been large (e.g., more than a million tons [t] of pollock per year), with likely effects on other species (predators, competitors, prey) and probable changes in the proportions of various species components of the ecosystem. Regarding unexplained populations changes, some declines appear to have been unprecedented, but some other species have increased, and the relative importance of alternative causes of these changes (e.g., fishing versus environmental change) has not been determined. Doubtless the Bering Sea ecosystem has changed and has been changing throughout its existence. Before the eighteenth century, the changes must have resulted mainly from environmental forcing and internal restructuring. Human activity since the late eighteenth century eliminated one species (Steller sea cows) and greatly reduced other populations (e.g., some whales and fur seals). Commercial fisheries for finfish and crustaceans increased enormously from the middle of the twentieth century. The charge of ecosystem damage has been most often applied during the last few decades, to the large pollock fishery, and to the declines of Steller sea lions, harbor seals, and certain seabirds. What criteria are appropriate to determine the condition of an oceanic ecosystem? Change is certain, and the desirability of a specific change is a value judgment reflecting human, not ecosystem, preference. Reversibility is said to be a desirable attribute, but irreversible changes can occur in ecosystems free of human influence (e.g., extinctions). The rapidity of change is not necessarily diagnostic (crashes can occur in populations in the absence of human activity), nor is surprise, which can reflect misunderstanding of ecosystem behavior as much as the nature of the ecosystem itself. Similarly, the rate of response to external forcing may be a function of species composition a whale-dominated ecosystem probably would have response times different from those of one dominated by pollock. The most useful diagnostic may be comparison of the observed rates of change of selected ecosystem components with those expected to occur. Expectation may be based on historical records from that system or changes observed in analogous systems. Apart from this, there appear to be few, if any, objective criteria that can be used to determine if an oceanic ecosystem has been adversely affected by humans. ECOSYSTEM MANAGEMENT Ecosystem management in the marine environment lags substantially behind the record of terrestrial progress both in conceptual development and in application. Management in the oceans is still typified by a focus on maximizing yields or economic profits from individual resources without an understanding of the ecosystem processes required to sustain those resources. This substantial disparity between the goals of management in marine and terrestrial environments may be due to our lack of perception of some of the fundamental differences

OCR for page 11
Marine Ecosystems: A Conceptual Framework 21 between marine and terrestrial ecosystems mentioned earlier or due to other conceptual problems such as (1) gross undersampling of the marine environment and consequent poverty of information on its biological diversity; (2) widespread ignorance of the functional dynamics of marine (especially pelagic) ecosystems; (3) the openness and interconnectedness of marine ecosystems on a scale that exceeds greatly the bounds of any management authority; and (4) a perception that intense exploitation of marine resources has sufficient economic and social value to outweigh any risk of damage to future ecosystem services or any alternative management goal of ecosystem integrity. Only in shallow coastal habitats dominated by sessile organisms emergent from the substrate (coral reefs, seagrass beds, marshes) is ecosystem management practiced to a degree that approaches management of terrestrial lands. For example, the Great Barrier Reef Marine Park Authority in Australia created an effective zoning and management scheme to protect and manage the ecosystem services of the Great Barrier Reef (Kelleher and Kenchington, 1992). The management of natural resources, including especially timber resources and marine fish stocks, has been undergoing a major change during the past decade (Sherman, 1991). The traditional management of natural resources has been based on formulations of the dynamics of the individual species in question. In fisheries, single-species models, especially those of Ricker (1958) and Beverton and Holt (1957), have been instrumental in determining management decisions with an intent of achieving optimal sustainable yields of the fish stocks in question. This approach has been recently modified to reflect the realization that species do not exist in isolation, independent of other species in the ecosystem. The National Marine Fisheries Service has developed and employed multi species models and has expressed commitment to the principles of ecosystem management. Although ecosystem management is widely perceived as the goal of modern fishery management, it is not clearly defined in an operational sense. Ecosystem management usually refers to either attempts to manage human activities to achieve specific ecosystem characteristics or products, or attempts to keep the whole ecosystem in mind when managing (or harvesting) particular components. The latter is usually easier to put into practice than the former (NRC, 1995). A committee of the Ecological Society of America recently described ecosystem management (Ecological Society of America, in press). The unanimous judgment of that committee was that ecosystem management does not imply any specific management goal, such as preserving biodiversity or promoting one value basis for a resource over any other, but rather that ecosystem management simply reflects a formal acknowledgment that species populations are nested within an ecosystem such that interdependencies among ecosystem components must be identified and included to achieve any specified management goal. However ecosystem management is defined, establishing preference among incompatible management outcomes requires negotiation among stakeholders, who may have conflicting values and priorities. The management goal of sustaining ecosystem composition, structure, and functioning implies that human actions can be mandated to produce desired consequences. But practical implementation of ecosystem management as defined above is difficult and may be impossible in marine ecosystems, for several reasons: . Control over human activities is inadequate.

OCR for page 11
22 The Bering Sea Ecosystem Many key ecosystem elements are not monitored or harvested commercially, and the cost of monitoring them is often considered unacceptable. Within any oceanic ecosystem, the interactions among organisms and between them and the abiotic environment are poorly understood and difficult to quantify and are likely to remain that way for the foreseeable future. Because the interactions are nonlinear and the systems are inherently dynamic and complex, their predictability is inherently limited. inasmuch as needed understanding of cause and effect is generally unavailable, protection of the ecosystem requires management to be risk-averse. Conservatism is also necessary because the desired control of human activities often cannot be fully achieved. The conservative approach can be relaxed if management is experimental or adaptive. For these approaches to be successful, tight control, comprehensive monitoring, and quick response by management to changes in forcing both human or environmental and ecosystem reaction are essential. Another key element to the concept of ecosystem management is a point made by Norton (19921: human cultures and institutions change much faster and, generally, at a smaller spatial scale than the ecological systems they are related to. As Norton pointed out, resource management generally operates on an annual cycle, whereas environmental management, governed by constraints necessary to protect the self-organ~zing and self-regulatory resource system, must operate on much longer time scales. Marine ecosystems are influenced by actions taken by a wide spectrum of management agencies. Coastal estuaries, bays, and oceans to a distance of three miles offshore in the United States fall within the purview of individual states. States take responsibility for management of fisheries resources within this coastal fringe, although regional management councils play an increasing role in integrating state programs on migratory stocks shared among several states. The National Marine Fisheries Service of the National Oceanic and Atmospheric Administration has responsibility for management of fishery resources from 3 to 200 miles offshore. Similarly, states can manage the exploitation of minerals, including of} and gas, within 3 miles of the coast, whereas the federal (Department of the Interior) Minerals Management Service makes management decisions for the 3 to 200 mile zone. Various federal laws mandate federal agency review of state land use management to protect certain marine habitats, notably marshes and submerged aquatic vegetation (seagrasses, etc.) and to protect endangered and threatened species. State environxnental management authorities influence the integrity of coastal marine ecosystems by permitting discharges of nutrients, oxygen-consuming materials, and other chemicals into rivers that flow to the seacoast. These same authorities regulate emissions of greenhouse gases, which threaten the ecological integrity of all marine ecosystems via changes in global climate. States also have responsibility for regulation of land development so as to prevent erosion of soils and enhancement of turbidity of coastal marine waters and sediment deposition on the seafloor. Marine ecosystems include some of the best and some of the least understood of all the worId's ecosystems. The processes that organize ecosystems on intertidal and shallow subtidal rocky shores are particularly well understood. The intertidal zone has been subjected to intense observation and a wealth of imaginative experimental manipulations, revealing much about the functional basis of these ecosystems (e.g., Connell, 1972; Paine, 19801. On the other hand,

OCR for page 11
Marine Ecosystems: A Conceptual Framework 23 marine pelagic and deeper sea-bottom communities and ecosystems are rarely studied as functional ecosystems. Experimentation is difficult and sometimes impossible in these environments. Only a small minority of the ecological players are described and ecosystem dynamics are typically unstudied. At the federal level, the National Marine Fisheries Service has made a commitment to development of ecosystem-management by choosing practices that preserve the health and sustainability of marine ecosystems (NMFS, 1991). However, management of marine fisheries in the context of the broader ecosystem is rare. A widespread international movement exists to urge management of marine resources as ecosystem units (LMEs) (Sherman, 19931. This approach represents ecosystem management of marine resources, but it is difficult to implement without better understanding of the processes that control function and structure of these ecosystems. One of the best ways to improve our understanding of process in these systems is to use management as a perturbation experimentiand thereby practice adaptive management of these systems. This approach holds great promise for advancing both our process understanding and our management. The EME approach recognizes the wide spatial scales over which most marine ecosystems extend. Nevertheless, these scales differ for different species and the management of even a single EME challenges existing management structures. Many natural marine ecosystems extend into international waters, where complex treaty agreements may be required for their effective management as ecosystems. Many LMEs have important component species with life stages that depend on coastal habitats and thus respond to management decisions made at the level of the state (within the United States). Where pollution from terrestrial runoff is degrading an EME, the range of management authorities that require coordination is even greater. Ecosystem management as applied to pelagic marine ecosystems demands a fundamental change in our management institutions and integration among them. Transition from commodity to ecosystem management in the sea also will require public involvement to be successful. There is now abundant evidence that humans have changed and degraded many marine ecosystems by many means (Norse, 1993). Overexploitation has caused extinctions and dramatic reductions in populations of many of the larger marine animals, especially whales and other marine mammals, sea turtles, and seabirds. Intentional and accidental introductions of exotic species through importation for aquaculture, release of ballast waters, and other mechanisms have changed many marine ecosystems (Carlton and Getter, 19931. Eutrophication of coastal estuaries and bays has affected ecosystem structure and functioning. Habitat alteration through fishing practices such as trawling, dredging, and dynamiting is widespread in coastal habitats (e.g., Riemann and Hoffmann 1991~. The influence of fishing both on targeted species and on species incidentally caught has altered the composition and functioning of many marine ecosystems (Norse, 1993; NRC' 1994a). Despite the substantial and compelling list of human alterations of natural marine ecosystems' the economic and social importance of fisheries resources may continue to drive marine resource management in its traditional commodity-based mode. Social and economic pressures contribute strongly to marine fisheries management. Skepticism over the argument that present exploitation schemes threaten long-term sustainability provides a selective force for setting short-term goals of maximizing yields. Several fisheries scientists argue that these social and economic pressures for short-term gain have resulted already in gross overexploitation of many of the world's fisheries resources (Ludwig et al., 19931. This view is not shared by

OCR for page 11
24 The Bering Sea Ecosystem everyone responsible for managing the fisheries, and there may be resistance to changing the current practice of marine fisheries management to one based on maintaining ecosystem structure and functioning. Fundamental conflicts between the usual goals of maximizing yields in marine fisheries management and the goals of the Endangered Species Act, Marine Mammal Protection Act, and National Environmental Policy Act remain unresolved even at the federal level in the National Marine Fisheries Service, as illustrated by the present groundfish management by the North Pacific Fishery Management Council, which does not explicitly orient management to provide additional food resources for the threatened Steller sea lion or for other declining marine mammals and seabirds in the Bering Sea and Gulf of Alaska ecosystems (NPFMC, 19931. At the state level, ecosystem management is even less contemplated and ahnost never practiced. Public opinion must play a vital role in resolving these conflicts and in setting limits to the application of ecosystem management in the seas. THE BERING SEA ECOSYSTEM AND ITS ANALOGS The Bering Sea ecosystem is not unique among oceanic ecosystems in its response to environmental and anthropogenic forcing. Most such systems in temperate and higher latitudes have experienced similar decadal-scale environmental fluctuations, along with continued, usually increasing, fishing pressure, accompanied by significant changes in the abundance of commercially important and other fish populations. In most cases, the systems include marine mammals as top predators. Indeed, the distribution of pinnipeds overlaps that of most Commercially important fish stocks. This is perhaps not surprising as pinnipeds often feed directly on these stocks or share prey consumed by these fish stocks. Oceanic ecosystems with substantial fisheries and pinniped populations are of particular interest. The most similar system is that of the eastern Gulf of Alaska, where Steller sea lion populations are stable or increasing off both southeast Alaska and British Columbia. Otariids (sea lions and fur seals) also occur in eastern boundary current systems (off California, Peru- Chile, and southwestern Africa), as well as off the Kuri} islands and Japan. In these and other regions of high oceanic productivity, other seal species (phocids) are found. Two seasonally ice- covered seas are similar to the Bering Sea in many ways. The Okhotsk Sea has both Steller sea lions and northern fur seals and the same phocid species as the Bering Sea. The Barents Sea, although having no otariids, has most of the same phocid species. The effects on other species in the ecosystem of reductions in key species abundances by fishing have rarely been established. In particular, the indirect effects of fishing on pinniped and seabird populations have been poorly documented. Some information on the coexistence of pinnipeds and fisheries is available from Peru and South Africa. Off Peru, the fur seal population was driven to low abundance by commercial exploitation and despite restrictions some killing of this species and the local sea lion continues. Anchovy has been an important component of the diets of both species. Studies on the fur seal suggest that the diet reflects the availability and abundance of prey, so that when anchovy is scarce, a wider range of prey is taken, as occurred after the collapse of the anchovy population in the early 1970s. According to Maj~uf and Reyes (1989), "Recovery of these marine mammal populations to their former levels is...unTikely, since the biomass of anchoveta necessary to

OCR for page 11
Marine Ecosystems: A Conceptual Framework 25 maintain large numbers of sea mammals is no longer available." Nonetheless, even with continuing exploitation mortality, the biomass of pinnipeds in the Peru Current region increased by a factor of three from 1955-65 to 1975-85 (Muck, 19891. The South African fur seal is the only species of seal breeding on the coasts of South Africa and Namibia, mostly on the west coast, where the fishing grounds are particularly rich (David, 1989). Since the beginning of the century there has been heavy exploitation, which reached its zenith in 1983 (when the 10-year average harvest was 75,000 pups), after which the industry collapsed, largely for political reasons. The species feeds principally on fishes (74 percent) and squids (17 percent). The quantities consumed are comparable with those caught by humans and by predatory pelagic fish; seals probably account for 5 to 10 percent of the total biomass of vertebrates and cephalopods removed by humans and other predators (Crawford et al., 1992). Yet in the presence of heavy fishing and continuing exploitation mortality, the fur seal population has increased throughout the century, by 1987 exceeding 1 million animals (David' 1989). Another relevant analog may be the North Sea, where heavy fishing for gadoids and other species has occurred throughout most of this century. While there are fewer species of pinnipeds (in this case, phocids), and their populations are smaller in the North Sea than in the Bering Sea, pinnipeds are important predators of gadoids in both regions (Perez and Loughlin, 1986). While Bering Sea fur seals and sea lions have declined during the last several decades, in the North Sea, grey seal populations have been stable or increased. The same is true of common seals up to and following the sharp decline caused by the phocine distemper virus in 1988 (anonymous 1992). These observations, incomplete as they are, suggest that comparative studies of analogous oceanic ecosystems, in part because of their differences in biologies, fisheries, and environmental conditions, could add to the understanding of events in the Bering Sea ecosystem. DELINEATION OF THE BERING SEA ECOSYSTEM Any definition of the principal ecosystem of concern is likely to be equivocal. The offshore boundaries of the system(s) are nebulous, may differ for different trophic levels, and may shift with time as environmental changes occur. Sherman and Alexander (1986) proposed that logical units for research and management related to marine living resources are LMEs, which they defined as extensive areas of ocean space of at least 200,000 km2, characterized by distinct hydrographic regimes, submarine topography, productivity, and trophically dependent populations. It is relatively easy to think in these terms when the LME is clearly bounded (e.g., in an enclosed sea) or has characteristic boundaries and features that are relatively repeatable (e.g., eastern boundary coastal upwelling ecosystems). The situation is less clear in the Bering Sea, for which three relevant LMEs have been identified (Sherman, 1993): the eastern Bering Sea, the western Bering Sea, and the adjacent Gulf of Alaska. In this scheme, the Bering Sea is divided by the shelf break and, in the north, by the international boundary; the "Gulf of Alaska" extends offshore for 200 miles and south to the Straits of Juan de Fuca. These domains appear to reflect political more than ecological realities. A more natural division of the Bering Sea is that of Hela and Laevastu

OCR for page 11
26 The Bering Sea Ecosystem (1962), who identified Kamchatka-Kurile Water, West Bering Gyral, Alaska Coastal Water, Alaska Gyral, and Northwest American Coastal Water. For convenience, this report considers four domains of the Bering Sea: (~) the continental shelf and slope, east and west, (2) the Aleutian Basin, (3) the Aleutian Islands, and (4) the adjacent Gulf of Alaska (Figure 2.2). The boundaries of these domains are basically physiographic and represent generalized ecological boundaries. The principal exchanges of water take place through Bering Strait into the Chukchi Sea, and in the western Bering Sea where the Kamchatka Current flows south into the Oyashio mixing region; this water is supplied by diffuse flow through the Aleutians. Atmospheric systems that control surface ocean conditions in the region, such as the Aleutian Low, extend well beyond these domains and are influenced by events at much greater distances, such as E1 Nino Southern Oscillation events in the equatorial Pacific. The more mobile organisms at higher trophic levels can move among these domains. For example, in the eastern Bering Sea, fish such as salmon and halibut, birds, and mammals move back and forth into the northern Gulf of Alaska. The co-location of a population of animals with an ecological domain varies with time, with life history stages and their migrations, and with environmental conditions. Furthermore, the distributions of populations of interest are rarely identical with those of their prey, competitors, and predators. Both walleye pollock and Steller sea lions are important components of the ecosystems of concern, but the ecosystems they inhabit are not identical, although they overlap substantially, with the extent of overlap varying over time. Thus, the boundaries of the several ecological domains of interest to this study tend to be ill defined and changeable. This limitation must be kept in mind as we attempt to trace the ecosystem consequences of changes in one or another species component, whatever the causes of these changes.

OCR for page 11
Marine Ecosystems: A Conceptual Framework 27 IY.-E 160 ~164-E ~-E I72-t 176-E i~. '?c.W 172-W ~ ~164-W "-w '~w . 152' W 64 62 ~ _ Or j4 SO. _ __ /~ CAPE Us,, ~NAVARIN - ST. LAWRENCE ISLAND A/ Flit/ Is/ ST MATTHEW ISLAND ISLAND PRIBILOF ISLANDS ~5~0t Bar ~ 4,/ ~ f / 9' /.~ . . 3 If I- c L ~ 4 , ~1 - D I ~! 1 1 ~1 16~E 172-E 176-E ~I76'W 1 S ~ 5 ! I I ~l I ~, ~, ~ 172-W 168'W t64-W - Figure 2.2 Four domains of ~e Bering Sea: (~) cont~nental shelf and sIope, (2) Aleutian Basin, (3) Aleutian Islands, and (4) the northern Gulf of Alaska. 62 58 56' sir 50.