Managing the Southern Ocean Krill Fishery

Krill (Euphausia superba), a two-inch-long, shrimp-like crustacean, is the center of attention in the waters—often called the Southern Ocean—that surround the Antarctic continent. By virtue of its abundance and its tendency to form high-density aggregations or "swarms," krill plays a key role in the Southern Ocean ecosystem as the primary food source for many species of whales, seals, and marine birds.

Not surprisingly, these attributes also have made krill the object of a significant fishery, an industry that began in the mid 1970s. The krill harvest grew rapidly from 20,000 tons in 1973-74 to a peak of 500,000 tons in 1981-82. Since then the harvest has fluctuated. The 1986-87 catch was approximately 375,000 tons.

The fishery thus far has been dominated by the Soviet Union, which takes 80 percent of the catch. Japan accounts for much of the remainder, while Poland, Chile, Korea, and Spain all harvest small quantities.

Krill is marketed in several forms, primarily as fish meal in the Soviet Union (where fish meal is used as a high-protein additive in animal feed), while most of the Japanese harvest is marketed for human consumption.

The magnitude of the potential krill harvest continues to arouse considerable interest, and with it concern about the possibility of overharvesting. It has been estimated that as many as 150 million tons of krill could be harvested annually in the Southern Ocean—a quantity described as "a shrimp cocktail the size of a city block and piled five miles high." A harvest of this magnitude would exceed the 1986 worldwide harvest of all fish by 70 percent.

Potential for overharvest

Concern about the possibility of over harvesting krill first surfaced during the 1970s as the commercial krill fishery began to flourish. Proponents of moderation warned that the harvest might follow a pattern of rapid uncontrolled growth leading to overharvesting, the collapse of the industry, and the decline of populations of species dependent on krill for food. They pointed out that such a pattern of growth in a fishery supplying the fish meal market had been already demonstrated in another part of the world: between 1958 and 1970 the harvest of Peruvian anchovy grew from less than one million to more than 12 million tons, at which point the overharvested population declined dramatically and catches fell to a small fraction of the peak harvest.

Meanwhile, control of most of the world's fisheries was gradually being transferred from international councils to coastal state jurisdiction as nation after nation established a 200-mile Exclusive Economic Zone (EEZ) in its coastal waters. However, a 200-mile EEZ could not and cannot be established around Antarctica because under the Antarctic Treaty, no country has sovereign rights over the continent.

In 1977, at the urging of biologists studying Antarctica and the Southern Ocean, the fifteen nations party to the Antarctic Treaty recommended that an international convention be established governing the management of fisheries in the Southern Ocean. Among these were nations with active fishing operations in the area, such as the Soviet Union and Japan. Others were concerned primarily about the potential impacts of the krill fishery on species dependent upon krill; these nations included the United Kingdom, Australia, New Zealand, and the United States.

Three years later, the efforts of these nations culminated in the signing of the Convention on the Conservation of Antarctic Marine Living Resources (CCAMLR). The convention encompasses an entire marine ecosystem (the Southern Ocean) and adopts an ecosystem to approach the management of all fisheries, including that for krill.

Traditionally, fishery management has tended to focus only on the particular species being harvested. Single-species models have been used to estimate the maximum harvest of a fish population permissible without depleting the population to the point where harvests in subsequent years will be reduced. Such a harvest level is known as the maximum sustainable yield.

The objectives of CCAMLR—and thus ecosystem management—differ from this maximum sustainable yield objective in several respects. First, the primary objective of CCAMLR is the conservation of marine resources of the Southern Ocean rather than maximization of fishing yield. Second, CCAMLR states that when fishing does take place, the ecological relationships between harvested and dependent populations must be maintained. This objective addresses the potential for fishing to harm other species dependent on the species being harvested. Finally, the convention calls for minimizing the risk of irreversible change to the ecosystem.

CCAMLR applies to all of the species in the Southern Ocean. In fact, most of the current management activities in CCAMLR are directed not at krill but at the restoration of bottom-fish stocks that were severely over harvested before 1980. However, the original impetus for using an ecosystem approach in CCAMLR came primarily from concern about the krill harvest, and the fate of the krill fishery remains the prime policy concern regarding Southern Ocean fisheries.

Problems ahead

The prognosis for CCAMLR success, according to many observers, is not good. Most of its operational procedures are not unlike those of previous conventions—which largely have been unsuccessful. In fact, John Heap, the chief negotiator of CCAMLR for the United Kingdom, has written that "if krill proves to be an economically exploitable resource, the activity will be on a grand scale and past experience gives no grounds to expect that catastrophe will not, despite . . . [CCAMLR], be the eventual outcome—it will just take a little longer."

One of CCAMLR's challenges is that it must contend with differences in interpretation of its objectives among member parties. Australia has led a move to define the objectives more precisely and to change the decision-making process to ensure that they are achieved.

Even if objectives were clearly defined, CCAMLR would still face practical problems in achieving them. Those procedures that do differ from previous regimes, in particular the concept of consensus voting, could thwart efforts to bring about positive change. Under a consensus voting regime, everyone must eventually agree to support a policy decision, despite any reservations. This policy can all too easily depend on the "least common denominator" of the differing views of the parties.

For instance, at present, the de facto management strategy of CCAMLR is to allow krill to be harvested until it can be proven that the level of harvest is harmful or excessive. One proposed method of ensuring that the ecosystem is not damaged by the krill fishery would be to shift the burden of proof so that a stock cannot be fished until it can be shown to be capable of withstanding the fishing pressure. However, for practical reasons—in particular consensus voting—such a change is unlikely.

CCAMLR is further limited by the tremendous scientific uncertainty surrounding the dynamics of the Southern Ocean ecosystem. Because little is known about key features of that ecosystem, such as population sizes and growth rates of important species, it will be some time before efforts to develop mathematical models of the system can provide even qualitative answers to management questions.

To minimize scientific uncertainty, CCAMLR has established a monitoring program to detect and record significant changes in critical components of the ecosystem. Both the populations of species being harvested and the populations of several species dependent on krill, such as fur seals and penguins, will be monitored.

The aim of the ecosystem monitoring program is to help distinguish between changes in the ecosystem due to harvesting and those due to environmental variability. However, with only monitoring data available, it may be extremely difficult to determine whether harvesting is the cause of a change, because it would be necessary to assume that in the absence of the fishery, the system would have remained unchanged. In reality, even in the absence of environmental changes, most biological communities are rarely in equilibrium states. If we add in known environmental changes such as global warming and decreased stratospheric ozone, both of which could affect the Southern Ocean, the problem of determining the cause of ecosystem changes through a monitoring program may become insurmountable.

Alternative strategies

Clearly, if CCAMLR is to achieve its objectives—and, by implication, if krill harvesting is to be controlled—a management approach is needed that recognizes the scientific uncertainty present and adapts accordingly. At this time, scientists can provide assessments of optimal harvest levels only in the form of a probability distribution reflecting the uncertainty of the estimate. Given such a range of potentially correct values, the individuals responsible for setting harvest quotas have often been influenced by short-term economic interest or differing levels of risk aversion to choose the most optimistic level of harvest.

An alternative strategy, relying on the scientists' "best guess" or highest probability assessment, has its own problems. Such an approach is appropriate only if the underlying model used in making the prediction is correct, and in few cases can we be confident of that.

Krill provides a good example of the shortcomings of the "best guess" approach. In the late 1970s, using available data, scientists estimated the lifespan of krill to be two to three years. With such a short life, the annual productivity of krill was thought to be tremendous because a high biomass of krill was present in the Southern Ocean. Given this information, scientists could have provided policymakers with a highest probability estimate of the optimal krill harvest.

Yet a decision based on such an assessment would have been a poor one. Subsequently it was determined that the lifespan of krill may be between six and seven years, double the previous estimate, and the annual productivity figures and the potential yield of krill have been substantially reduced.

If neither giving managers a harvest range from which to choose not relying on the harvest estimate with the highest probability is suitable for effective management under current conditions, what is the solution? It may well lie in testing alternative hypotheses to determine what model or what value of a parameter is correct. This strategy is referred to as "adaptive management" or "experimental management."

A number of scientists are already advocating that experimental management be applied to the krill fishery. If krill in the Southern Ocean were divided by location into five or six populations for management purposes, different management regulations could be applied to each of the populations. Results could then be monitored, and eventually an overall management strategy could be implemented on the basis of the experimental results.

Admittedly, such a system would have its shortcomings. There would probably be a considerable flux of krill drifting with the currents from one management unit to the next. For scientific purposes it would be better if the management units were completely isolated; for conservation purposes it is better that they are not, so that if one unit is overharvested it can be replenished from the adjoining units.

In general, however, the use of an experimental management approach to the krill fishery is ideal for scientific purposes. First, it provides a means of distinguishing between ecosystem changes caused by harvesting and those caused by environmental effects. For example, if changes in krill populations or in the populations of indicator species were of the same magnitude in all management units, despite differences in fishing intensity, it could be surmised that the change was a result of a general environmental effect. Second, data would be available from each area on how species dependent upon krill respond to a range of harvest levels. Therefore scientists could more accurately estimate critical parameters in models.

An experimental approach also may offer strong appeal to policy making groups. For instance, CCAMLR could adopt such an approach without forcing participating countries to modify their fishing activities in the near term.

Currently, krill harvest levels are low and the fishery is concentrated in just one location. Using an experimental management approach, this area could be designated as having no limits on the harvest. Harvest levels for other areas could be set at some proportion of the harvest in the unlimited area. At least one area could be targeted for a low harvest—say, 10 percent of the unlimited area—thus serving as a control area. All areas would be kept open to fishing on at least a minimal level to enhance the amount of information gathered on the krill population size.

The experimental management approach provides a hedge against risk because even if krill were over harvested in some locations, a portion of the system would be preserved. Accurate assessments of maximum harvest levels could be obtained more rapidly than under a modeling approach, and because the assessments would be based on empirical data rather than modeling estimates, the recommended harvest levels would probably be more acceptable to all parties. Though the harvest in some areas would be held to low levels and thus could slow overall growth of the fishery, this loss would be weighed against the assurance of a sustained yield.

Outlook

The rapid growth in the krill fishery that occurred before 1981 will probably not be repeated in the near future. Though there is some growth in the Japanese market for whole peeled krill, this is a low-volume, high-value market and will not result in harvests high enough to be of concern.

Demand for fish meal in the form of krill is expected to increase only moderately. The Food and Agriculture Organization (FAO) predicts that demand for fish meal of all types will increase only slightly over the next 15 years. The importance of krill within this market depends on its price relative to other types of fish meal. The costs of fishing in the Southern Ocean are high; in the case of krill, there are added costs because it must be processed immediately aboard ship to avoid spoilage. Consequently, krill meal is not viewed presently as a highly competitive protein source.

Current no-growth conditions in the krill industry may provide a "window of opportunity" for establishing a management strategy that could cope with possible future rapid growth. Such a strategy could anticipate growth, rather than react to it.

Certainly, the eventual success or failure of CCAMLR will depend on current efforts to implement a management strategy that deals with the extreme uncertainty surrounding the ecosystem. The strategy must reduce uncertainty, provide the basis for current management decisions, and meet the ecosystem objectives of the convention. Bearing these requirements in mind, an experimental, adaptive approach may well provide the best solution.