Review of Artificial Production of Anadromous and Resident Fish in the Columbia River Basin

read full document >

Part I: A Scientific Basis for Columbia River Production Programs

Note: This web page contains only the title page, transmittal letter, introduction and conclusions. Click here for the full report.

Scientific Review Team
Independent Scientific Advisory Board

Ernest L. Brannon James A. Lichatowich
Kenneth P. Currens Brian E. Riddell
Daniel Goodman Richard N. Williams

Willis E. McConnaha, Chair

Program Evaluation and Analysis Section
Northwest Power Planning Council
851 SW 6th Avenue, Suite 1100
Portland, Oregon 97204

Cite as:
Brannon, Ernest L., et. al., (Currens, Kenneth P.; Goodman, Daniel; Lichatowich, James A.; McConnaha, Willis E.; Riddell, Brian E.; Williams, Richard N.). 1999. Review of Artificial Production of Anadromous and Resident Fish in the Columbia River Basin, Part I: A Scientific Basis for Columbia River Production Program, Northwest Power Planning Council, 139 pp.

Scientific Review Team
Independent Scientific Advisory Board

April 22, 1999

Mr. Todd Maddock, Chair
Northwest Power Planning Council
851 S.W. Sixth Avenue, Suite 1100
Portland, OR 97204

Dear Mr. Maddock:

I am pleased to transmit to you the first phase of the Scientific Review Team's (SRT) report on hatchery programs in the Columbia River Basin. This version updates our December version and adds considerations of resident fish artificial production.

The report includes an historical overview of artificial production within the basin, a review of the state of the science with respect to effectiveness of artificial production as a means to augment harvest, a review of the state of the science with respect to ecological and genetic effects of artificial production on wild spawning populations, and a set of recommendations in the form of guidelines that might guide policy on artificial production so as to be consistent with a sound scientific foundation.

The proposed guidelines at this point are based on our review of the published science, which completes this phase of our task in the Review of Artificial Production. In the next phase we will determine whether analysis of existing monitoring data resolves any of these hypotheses more conclusively, in the specific context of the Columbia Basin.

The Artificial Production Review is consistent with other recent scientific reviews that considered this topic (the Northwest Power Planning Council's Independent Scientific Group (ISG), 1996; the National Research Council (NRC), 1996; and the National Fish Hatchery Review Panel (NFHRP), 1994). There is considerable consensus among the three previous scientific reviews and our own, regarding the status of the science associated with artificial production.

We understand very well that the priorities of fisheries management have changed significantly in recent years, so the needs that hatcheries should serve are also changing. The ecosystem based management approach places artificial production in the basin in a very different role than employed in the recent past. In particular, the reality of increasing numbers of Endangered Species Act listings of anadromous and resident fish puts a much higher emphasis on wild stocks and naturally spawning stocks. This increases the concern over the potential for artificial production to cause genetic and ecological harm to such stocks. But it also raises the possibility that hatcheries may serve some more positive role in this era of new priorities.

The historical track record does not reassure us that hatchery programs can respond to changing realities and changing scientific knowledge. We hope that ways will be found to ensure that in the future this important component of the fisheries management system will evolve more rapidly and constructively. Even by the standards of the old objective of simply augmenting harvest to mitigate for lost and impaired habitat, artificial production of anadromous salmonids in the Columbia basin would need careful re-evaluation because of the perceived less-than-effective return of adults. However, under the new ecosystem management paradigm, performance assessment of hatchery programs in light of the potential impacts on native fish becomes most critical.

Other major factors besides hatcheries, including some factors that are beyond human control, surely are playing some role in the salmon decline. And some hatcheries have performed considerably better than average, contributing substantially to harvest, and maintaining runs of the hatchery stock. One hope in undertaking the SRT evaluation is to identify causes for the differences in performance between hatcheries, in order to develop recommendations for improving average hatchery performance, as well as to add to our understanding of how hatcheries should fit into the new ecosystem management concept.

There has not been adequate monitoring and evaluation of Columbia Basin salmon stocks. Monitoring and analysis to date has been sufficient to document rather disappointing numbers. But the monitoring and analysis has not been sufficient to pin down the causes of the poor performance, nor has it been sufficient to determine why some hatcheries perform better and some perform worse. The challenge in the concept of ecosystem management is how to reconcile local objectives of providing harvest opportunities, sometimes with exotic species, with larger ecosystem objectives.

Ecological and genetic science does provide plausible hypotheses about what some of the avoidable causes of poor performance might be among the anadromous salmonid hatchery fish. These include: inbreeding, domestication, inappropriate life history timing, inappropriate physiological and behavioral conditioning, over-reliance on too small a number of life history types, and exceeding carrying capacity in some key portions of the natural environment. Our report reviewing the science explains how these plausible ecological and genetic mechanisms could be operating in hatchery stocks.

Given the new management emphasis on wild stocks, it is especially distressing to learn that these mechanisms that might account for poor performance of hatchery fish could also have very negative effects on naturally spawning fish that interact ecologically and genetically with the hatchery product. Our report reviewing the science explains plausible mechanisms of interaction between the hatchery products and wild spawning fish.

Ecological and genetic science suggests that artificial production must be carefully integrated into the functioning of the entire ecosystem. Both the ecosystem and artificial production need to be managed to some normative standard, in order for artificial production to be effective in meeting the basin's goals. Our report reviewing the science offers recommendations on how this integration and management might be accomplished.

Uncertainty places a premium on experimental management to better diagnose the causes of the problems, and to refine the solutions. The effectiveness of experimental management depends critically on a comprehensive program of monitoring and analysis to detect responses to experiments, and to alter management according to what is learned. The programs of monitoring and analysis now in place in the Columbia Basin may not be equal to this task. The region, through the Council, needs to implement a scientifically valid, comprehensive monitoring and evaluation program to assess hatchery procedures, production, impact on natural populations, and achievement of goals.

We note, finally, our analysis under the new concept of ecosystem management in the basin must place hatcheries in the experimental framework, and most certainly in the perspective of ecosystem function we conclude that the outcomes of these experiments are not yet known. Yet artificial production has been implemented on a scale that will continue to commit a large percentage of the region's restoration resources, a large percentage of the available watersheds, and a large percentage of the remaining stocks to a single, unproven technology. There may be merit to reconsidering these priorities.

The SRT will continue its review of hatchery programs, compilation of available data, and analysis of those data. We expect to finalize our review and recommendations by June 30, 1999.

Sincerely yours,

Willis E. McConnaha, Chair
Scientific Review Team

A. Scope of Review
B. Definition of Artificial Production
C. Relationship Between this Review and Development of the Regional Multi-Species Framework
D. Definition of the Columbia Basin Ecosystem

Conclusions and Guidelines
A. Points of General Agreement with Recent Reviews
B. General Considerations of the State of the Science and the Technology
C. Relation to an Ecological Framework
D. Guidelines on Hatchery Practices, Ecological Integration and Genetics


In July 1997, the U.S. Senate directed the Northwest Power Planning Council, with the assistance of the Independent Scientific Advisory Board (ISAB), to "conduct a thorough review of all federally funded hatchery programs operating in the Columbia River Basin?" with the intent to ensure that federal dollars are spent "wisely" and "in a cost-effective manner that maximizes the benefits to the fish resource." The Council is to assess the "operation, goals and principles of state, tribal and federal hatcheries..." with regard to the effectiveness of their role in the broader context of fisheries management. The Council is to recommend to Congress a set of policies that would guide the use of Columbia River hatcheries.

In response to the Congressional directive, the Council consulted with the ISAB and appointed a Scientific Review Team (SRT) to provide an independent assessment of the basin's artificial production program. The SRT includes four members of the ISAB, two additional independent scientists, and a scientist from the Council staff as chair of the team. The SRT will review hatchery programs in the basin, analyze their effectiveness in meeting mitigation responsibilities, assess their success in enhancing salmonid production, and evaluate their role in supplementation of natural salmon and steelhead runs. The SRT analysis will provide the biological basis for the Council's recommendations to Congress.

To provide timely advice for the Council's report to Congress, the SRT is conducting the analysis as three sequential tasks:

    1. A summarization of current scientific knowledge on artificial production and its implications for Columbia River programs.
    2. A compilation of the data relating to the performance of artificial production in the Columbia River Basin.
    3. An analysis of this information in light of current knowledge to assess the performance of artificial production in the Columbia River Basin.

This report addresses the first task, which is to provide background information on the state of the science that relates to artificial production in the basin. The scientific rationale developed in this report takes the form of guidelines that could form the basis of recommendations regarding artificial propagation, in the context of ecosystem management. The second task, assembling the database of all past and current records on artificial production in the basin, has been assigned to a contractor. The third task will be the analysis of hatchery programs and the database, and finalizing the report. Each task will be summarized in separate reports to the Council and then integrated into a final report.

This paper provides the SRT's analysis of the history of artificial production and other hatchery evaluations related to the Columbia basin. Hatcheries initially were used to augment the fishery, later to mitigate for habitat destruction by development activities, and more recently to supplement natural production and conserve salmon using captive brood stock techniques. These roles are defined and discussed in this report, and the state of current knowledge of the genetic and ecological effects of hatcheries is summarized, as well.

The next phase of the SRT report, where applicable, will use guidelines associated with uncertainties as a basis to evaluate hatchery performance. The evaluation should identify production and operational strategies that can assist in development of hatchery policy. Performance evaluation will use production return criteria and/or fulfillment of mitigation objectives as the basis of assessment. The two evaluations, with recommendations emanating from both the scientific analysis and hatchery performance evaluation, will be articulated in a proposed conceptual foundation for the Columbia River Basin's artificial production program. Whether or not this conceptual foundation is adopted as the basis for regional hatchery policy, it is imperative that policy is based on a scientific foundation and that adaptive management is pursued using performance criteria.

A. Scope of the Review

Artificial production has been used in the Columbia River Basin for many purposes during this century. Although basin hatcheries have produced resident species, such as sturgeon and rainbow trout, hatchery production has focused almost exclusively on anadromous salmonids -- primarily coho and chinook salmon and steelhead trout.

These three species also have been the focus of sport and commercial fisheries management in the basin and, ironically, recovery measures, as well. Due to the ecological, economic, recreational and cultural importance of these species, the Council's policy recommendations must address anadromous salmonids primarily, but also apply to a much broader spectrum of species.

In the Columbia River Basin, there are hatcheries for both anadromous and resident fish. Many resident fish hatcheries in the basin, like many anadromous fish hatcheries, are intended for mitigation and are located upstream from Grand Coulee and Hells Canyon dams, which are complete barriers to anadromous fish passage. Netboy (1986) estimated that 40 percent of the original spawning areas for Columbia Basin salmonids had been lost because of blockage due to dams.Return to the River estimates a 55-percent loss (45 percent remaining; Page 353). Of the original salmon and steelhead habitat available in the basin, 55 percent of the watershed area and 31 percent of stream miles are now inaccessible to anadromous salmon, having been blocked by dam construction (NRC 1996, pg. 63). Furthermore, much of this inaccessible habitat was irreplaceable natural spawning habitat, located mainly in headwater regions of the basin. Thus, successful artificial production of resident fish is a necessary and crucial component to fully mitigate anadromous fish losses in these blocked areas. In addition, many native resident fish species are currently federally listed as threatened or endangered under the U.S. Endangered Species Act, based on their imperiled biological status. As with anadromous salmonids, numerous at-risk resident fish populations (e.g. bull trout, white sturgeon, and various resident salmonids) are also the focus of recovery measures.

The scope of our review concentrates on artificial production of anadromous salmonids in the Columbia Basin. However, most, if not all, of the scientific information relating to ecological impacts of anadromous salmonid hatcheries applies to the use of hatcheries that currently produce more than a dozen ecologically and economically valuable species of resident fish. Therefore, resident fish hatchery policy must be consistent with the principles in the conceptual foundation that the SRT will recommend to the Council for anadromous salmonids. In fact, because resident species do not have the distribution range of salmon and steelhead, and thus are not exposed to the same risks facing anadromous salmonids over their migratory corridor, we expect that resident species will be very responsive to the principles guiding policy in anadromous salmonid management. Throughout the review, we make connections between anadromous and resident fish production with regard to principles and technologies.

B. Definition of Artificial Production

Artificial production and hatcheries are generally viewed as synonymous terms in that both refer to the same range of fish culture technologies, encompassing everything from releases of unfed, substrate-incubated fry to captive rearing of migrant juvenile salmonids on formulated diets in concrete raceways. The most common type of fish hatchery is a cluster of buildings and concrete raceways located adjacent to a tributary stream. But a hatchery also can be a gravel-lined incubation box in which artificially spawned eggs are incubated to enhance fish production. Or, a hatchery can be an engineered spawning channel that salmon enter to spawn naturally on graded substrate, where water flow is controlled to enhance egg-to-fry survival. Or, a hatchery even can be an earthen acclimation pond in which fingerlings are fed before dispersing into the natural stream on their own volition for rearing or migration.

In this report, our focus is on the "standard" public hatchery design -- the cluster of buildings beside a tributary, with tray incubators and concrete raceway rearing systems that provide the entire freshwater feed and residence requirements before the fingerlings are released to migrate seaward. Columbia River hatcheries were designed around variations of this "standard" incubation and rearing system. It is a system that has been used for most chinook and coho salmon, and steelhead trout, hatcheries over this century.

This type of hatchery generally controls the entire freshwater juvenile life cycle, except the migratory passage. Adults are intercepted and spawned artificially, based on a breeding plan that varies from simply crossing multiple females with a composite of two or more males, to a breeding matrix that maximizes the available genetic variability. Eggs are usually incubated in trays until hatching or to the point of emergence when yolk stores are nearly exhausted. Some form of substrate is often included in the incubation compartment to reduce alevin activity and prioritize stored energy for growth. At or before the emergence phase, the young fry are placed in troughs or tanks for swim-up and early rearing, and then transferred to raceways for production rearing until they are distributed for release as smolts or presmolts to natural waters. Formulated diets are used throughout rearing, based on nutritional requirements, and fed as mash or graded pellets to accommodate the size of the fish as they grow. The system is well defined in a program to maximize efficiency of operations.

Hatchery performance assessment understandably has been limited within the rather narrow definition of variables in facility design and operations common to such facilities. Because Columbia River hatcheries use the standard technology, performance differences have as much to do with management as with application of the technology itself. Variables such as the source of fish, release strategies, relative size and condition of smolts, water supplies, location of the hatchery and its location on the migratory corridor over the length of the river will affect performance. Therefore, the context of our evaluation is the relative performance of a particular class of hatcheries within the confines of river conditions in the Columbia Basin, under agency management responsibility. Consequently, our assessment will be an assessment of the policy and location as much as the technology involved.

C. Relationship Between this Review and Development of the Regional Multi-Species Framework

As this review is being undertaken, states, tribes and agencies of the federal government in the Pacific Northwest are collaborating on a multi-species planning process for fish and wildlife in the Columbia River Basin. The multi-species planning process is guided by a framework that links Columbia Basin fish and wildlife policy to a vision that balances the many values provided by the natural resources of the Columbia River and its tributaries. The multi-species framework will be based on an ecological, conceptual foundation that recognizes that the river and its species are interrelated parts of a whole.

The multi-species framework will include principles, goals and objectives that reconcile seemingly inconsistent and uncoordinated approaches to fish and wildlife policy in the region. These principles, goals and objectives will be expressed in a set of scientifically supportable alternatives for the future of the Columbia River -- especially as it relates to management of fish and wildlife resources. As they are developed, these alternatives will be analyzed for their ecological impacts, based on an explicit conceptual foundation. The conceptual foundation includes a set of scientific principles that define the scientific context for the analysis.

Once it is developed, the multi-species framework will provide systemwide direction and specific strategies for fish and wildlife programs, as well as objectives against which results can be evaluated. The conceptual foundation for artificial production developed in this review should be consistent with the set of scientific principles guiding development of the multi-species framework. In this sense, a conceptual foundation for artificial production is a refinement of the more general conceptual foundation for the multi-species framework, and serves to focus scientific principles on decisions about how to use artificial production. We believe that a scientifically supportable conceptual foundation, such as that guiding development of the multi-species framework and potentially refined by our assessment, should be the basis for developing future hatchery policies.

D. Definition of the Columbia Basin Ecosystem

Natural and cultural attributes define an ecosystem (ISG, 1999). The modern Columbia River ecosystem is far different than the ecosystem that existed before the encroachment of modern civilization -- as that ecosystem was different from the one that existed before Native Americans began exploiting the Columbia River fishery. Man's actions irrevocably altered the Columbia River ecosystem, and those impacts define the parameters for ecosystem management today.

As major hydroelectric facilities multiplied in the Columbia River Basin, the free-flowing river became a series of linked reservoirs. This new environment favored species previously limited by higher velocities and cooler water temperatures. Predator and competitor species assumed new levels of abundance in the river system previously dominated by salmonids. For example, the northern pikeminnow (previously named the northern squawfish), a native salmon predator, increased in number -- and impact on salmonids -- as a result of the increased reservoir habitat. (Zimmer 1953, USFWS 1957, Thompsom 1959, Beamesderfer and Rieman 1991, Poe et al. 1991, Rieman et al. 1991.

Perhaps even a more serious impact in the evolutionary sense, however, were the many exotic fish brought into the basin by private, state and federal entities, such as American shad, channel catfish, largemouth and smallmouth bass, blue gill, yellow perch, brown trout, brook trout and lake trout (Simpson and Wallace 1982). While many of these fish were introduced for sportfishing diversity before the ecological impacts were fully appreciated, and have now become an important part of the species selection offered the sportfishing public, they nevertheless have permanently changed the Columbia River ecosystem. Although most of the exotic species were introduced half a century ago, interactions among the various non-native and native fish species are likely to continue to evolve toward a new equilibrium (as yet unknown).

Substantive and even drastic changes in species composition and habitat utilization have occurred over the last several decades. Preliminary surveys in the lower reaches of the Yakima River over the last half of 1996, for example, revealed that about two-thirds of the species encountered were exotics, and smallmouth bass represented over 60 percent of all fish intercepted (Monk 1997). Sampling gear tended to exclude fish larger than 10 centimeters in length, but as an index of general abundance, the survey demonstrated how dominant these species have become in some areas of the Basin. The impact of these newcomers, through competition or predation on endemic species, is unknown, but the success of exotic species has come at a cost to native fish.

Major changes in the operation or configuration of the hydrosystem also will affect interactions among fish species. Changes that increase normative conditions, such as natural river drawdown in the lower Snake River or a major drawdown of John Day Dam (system configuration alternatives currently being studied by the National Marine Fisheries Service or the U.S. Army Corps of Engineers) would promote an equilibrium that favored coldwater native fish species over warmwater native and non-native fish species.

This dynamic mix of native and non-native species defines the modern Columbia Basin ecosystem. Where anadromous species have been eliminated by barrier dams, mitigation has been in the form of replacement resident fisheries, and sometimes those fisheries include exotic species. Moreover, resident fish hatchery programs often will not be complementary with the ecosystem management perspective adopted for anadromous hatchery production. Nor can mitigation in these cases necessarily imply that hatchery production will be temporary until natural production can sustain the population. In some cases resident fish populations have been established where none existed before and will be entirely dependent on artificial production. Thus, in some cases the concept of using supplementation hatcheries to rebuild naturally reproducing populations does not necessarily apply to resident fish .

Another difference between anadromous and resident fish hatcheries is performance measurement. Traditionally for anadromous programs, we have measured hatchery production success in terms of harvest return. This is often an inappropriate statistic in resident fish hatchery performance. More realistic performance goals for resident fish might include catch and release only, or simply having fish in the system available for viewing. Thus, for both anadromous and resident production, performance criteria should be matched with the fisheries management objective of the specific program or facility and in recognition of the community of fishes that are now in the basin. What constitutes the Columbia Basin ecosystem, therefore, is basic in how hatchery production is viewed, and why "normative" is such a key concept that accommodates the biological realities with the cultural and economic changes that define the present ecosystem.

Conclusions and Guidelines

To briefly review, hatchery production of Columbia River Basin salmon started before the turn of the century for the purpose of augmenting harvest of chinook salmon for the commercial fishery. Science initially had only a small role in the process -- primarily the development of fish husbandry. Over time, the role of science evolved to include formal attention to nutrition, genetics and pathology. That attention, however, centered primarily on the technology of fish husbandry, with little attention to concerns about hatchery fish interaction with wild fish, or with the natural (post-release) environment.

With the new paradigm of ecosystem function, science articulated a refreshed interest in community balance, food chain dynamics, population structure, and integration of hatchery fish as a functional component of the ecosystem. Standard hatchery procedures no longer were accepted as a means of addressing augmentation or mitigation, and much greater emphasis is placed on developing a new conceptual foundation under which artificial propagation should proceed. The architects of this new conceptual foundation cannot be oblivious to the fact that the Columbia and Snake rivers are systems substantially altered from the historical conditions in which anadromous salmonids evolved.

This report is not a commentary establishing the role of artificial production in future Columbia River fisheries management, or recommending the degree to which hatchery production should contribute in the basin. That is the responsibility of the state and tribal fisheries managers. This report concerns the state of the science that relates to artificial production, and in that regard presents guidelines that we believe should be the foundation of recommendations on the appropriate use of artificial production in the future. Following the points of general agreement with the three recent scientific reviews that broadly addressed hatchery operations (ISG, NRC, NFHRP), the guidelines are presented in two parts. First are guidelines based on our scientific assessment of artificial production that includes hatchery practices, ecological, and genetic considerations. Second are guidelines that address what we consider the necessary research to resolve problems and questions about the technology and management of hatchery programs.

To provide further background for our guidelines, it is appropriate to discuss generally what is known and not known about hatchery effectiveness and hatchery effects. The level of knowledge varies considerably for different aspects of the effects of hatchery management and policy. The three major divisions of this material are knowledge about:

  1. Effects of hatchery practices on the egg to smolt phase of the life cycle of hatchery fish;
  2. Effects of hatchery practices on the post-release phase of the life cycle of the hatchery product, and,
  3. Effects of the hatchery product on the wild stocks with which they interact ecologically and genetically.

The amount and certainty of the knowledge in each of these three divisions is so different that we cannot reasonably attain the same level of conclusiveness about them from our review of the science.

In order to attune our conclusions with the differing levels of certainty, we offer our advice in the form of recommendations, guidelines and hypotheses. And we urge our audience to be sensitive to the distinction and the implications of these guidelines, recommendations and hypotheses. Basically, the current state of the science can support firm "recommendations" about practices to enhance the performance of the hatchery product in the egg-to-smolt phase of the life cycle. The science is less comprehensive and conclusive about the post-release performance of the hatchery product. For this aspect we can offer tentative "guidelines" for practices that we are reasonably sure will generally improve post-release performance somewhat, but we can't be sure how much, and we can't offer assurances that these guidelines in themselves will be sufficient for meeting objectives of the program. The science is even less conclusive about the effects of the hatchery fish on wild stocks. The science can identify specific genetic and ecological mechanisms that must be operating in the interaction between hatchery and hatchery fish, but the degree of quantification and empirical verification in this important aspect of our knowledge is so low, that for the most part we can only state important, plausible "hypotheses." The monitoring, analysis, and experimentation necessary to arrive at conclusions about these hypotheses should be elevated in priority for the future. In the meanwhile, we advise managers and policy makers to adopt a precautionary approach in the decisions where these hypotheses have a bearing.

The picture is further complicated by the very real possibility of working at cross-purposes by simultaneously attempting to manage for improved egg-to-smolt performance within the hatchery, improved smolt-to-adult returns of hatchery fish, enhancement of wild stocks through supplementation, preservation of wild stocks with captive breeding, and minimization of potential negative effects of augmentation hatchery operations on wild stocks. Practices that are good for one objective could be bad for another. A successful overall hatchery policy will have to be cognizant of these possible effects "between compartments." We will discuss these trade-offs in the third, synthesis phase, of our review.

A. Points of General Agreement with Recent Reviews

The three recent independent reviews of fish and wildlife recovery efforts in the Columbia River Basin addressed hatcheries among other issues -- one report addressed hatcheries specifically. These reviews collectively represent a concerted effort to assess hatchery production from the scientific perspective. There was consensus among the three panels, which underscores the importance of their contributions in revising the scientific foundation for hatchery policy. The ten general conclusions made by the three panels are listed below.

  1. Hatcheries generally have failed to meet their objectives.
  2. Hatcheries have imparted adverse effects on natural populations.
  3. Managers have failed to evaluate hatchery programs.
  4. Rationale justifying hatchery production was based on untested assumptions.
  5. Supplementation should be linked with habitat improvements.
  6. Genetic considerations have to be included in hatchery programs.
  7. More research and experimental approaches are required.
  8. Stock transfers and introductions of non-native species should be discontinued.
  9. Artificial production should have a new role in fisheries management.
  10. Hatcheries should be used as temporary refuges, rather than for long-term production.

Given the present degree of uncertainty about hatchery success, the SRT agrees that unified hatchery management policies should include plausible hypotheses that test some of the uncertainties inherent in these conclusions.

In particular, with respect to the hypothesis that the future role of hatcheries in fisheries management will evolve considerably, we note that the priorities of fisheries management have changed significantly in recent years, so the needs that hatcheries should serve are also changing. The ongoing reality of increasing numbers of Endangered Species Act listings of anadromous and resident fish puts a much higher emphasis on wild stocks and naturally spawning stocks. This increases the concern over the potential for artificial production to cause genetic and ecological harm to such stocks. But it also raises the possibility that hatcheries may serve some positive role in this era of new priorities.

B. General Considerations of the State of the Science and the Technology

The goals sought by hatchery programs have changed over the years. However, like earlier hatchery programs, the most recent efforts of augmentation, supplementation and captive brood stock production may have succeeded in their numerical production objectives with regard to juvenile releases. The issue is that the effect of that production on increased return has generally not been demonstrated, and that effects on naturally spawning stocks have not been adequately investigated. Agencies have evaluated some hatchery procedures, such as the effect of size and time of release on return success, but there has been a general lack of effort at the programmatic level. Only recently has natural production in the Columbia basin been given priority.  Previously, the approach of concentrating artificial production of Pacific salmon downstream from lower Columbia dams was considered a viable mitigation option for providing the necessary production from the system, based on general trends in hatchery production returns. However, if evaluations demonstrating the consistent production benefits of hatcheries have been undertaken, they have not been published in the peer-reviewed literature. Such publications are required to provide fair analysis of these hatchery programs. Issues of genetics, stock transfers and limiting effort to avoid overfishing wild stocks mixed with hatchery fish are symptomatic of the previous philosophy downplaying the role of natural production and the human alteration of the natural Columbia River Basin ecosystem. Given the present emphasis on the ecosystem approach, these issues are now important and should be given priority in the development of the new conceptual foundation for artificial production.

In the past, weak native runs have been replaced with other fish in the development of hatchery programs, such as the original plan regarding Sooes River fall chinook salmon. Such action is inconsistent with present values. Diversity is now believed to be one of the keys to the long-term success of salmonid populations, and adaptive traits should never be willfully abandoned. In situations where a stock has been extirpated, managers need to have the option of introducing non-native fish to establish the nucleus on which restoration can take place. Even in this situation, however, the donor stock chosen should not be simply based on egg availability. Careful analysis is required to assure environmental relationships between donor and target streams are as compatible as possible for the stock selected. Frequently, appropriate donor stocks will come from ecologically similar and geographically adjacent streams or watersheds.

Stock transfers and introductions can place serious risk on native fish stocks and should be discontinued from hatchery programs except when the purpose is to restore an extirpated run or population. Introductions also might be justified when genetic diversity is so low as to threaten the persistence of a population.

The primary role of hatcheries in the basin is mitigation for the loss of harvest as a result of reduction of habitat from economic development of the Columbia and Snake rivers. Given the present encroachment of habitat modification and degradation into the riparian and adjacent lands of these river systems, it is unlikely that natural production in a recovered ecosystem would satisfy commercial, tribal, and sports harvest interests. The options, therefore, are (1) to be content with lower production from managed natural populations and use hatcheries in a more temporary role for rehabilitation, or (2) to manage for greater harvest potential from a combination of natural production and hatcheries mitigating for habitat no longer accessible. Mitigation hatcheries are a long-term commitment involving significant cost. Although Columbia Basin hatcheries have not satisfied their objective of sustaining production thus far, nonetheless they now account for the majority of production in the basin.

Changing the manner in which hatcheries address their role is the hope that hatcheries can succeed. Based on past hatchery performance in the basin, such expectation is bereft of proof. But abrogation of the concept based only on the past is also imprudent when hatchery management has made such serious mistakes and the fish still persist. As Reisenbichler (1998) reasoned after observing fish in the hatchery environment, ".. substantial adaptation to hatchery conditions [occurs]... and holds promise that modifying hatchery conditions can reduce deleterious genetic differences between hatchery and wild fish." The hope is that with care given to appropriate changes in the hatchery environment, the response of hatchery fish can be compatible and complementary to the natural population structure of the native species. The normative ecosystem is an equitable mix of natural and cultural features with environmental equity to sustain all life stages of a diverse mixture of healthy wild anadromous salmonids, concurrent with cultural and economic development of water resources. Hatcheries can have a mitigation role in the normative ecosystem. These may become rehabilitation programs that secure the endurance of native runs. They may also become perpetual programs to supply commercial or angling opportunities. The challenge is to redevelop the concept of a hatchery to assure enhanced production to meet both ecological and economic objectives.

C. Relation to an Ecological Framework

It is imperative that priority be given to the development of a set of scientific guidelines that serve as a conceptual foundation for the Columbia basin hatchery program. These also must be consistent with the eight elements of the basin-wide ecological framework (NPPC Document 98-6) that is to guide management of the Columbia River as an ecological system. The eight ecologically based elements are listed below.

  • The abundance and productivity of fish and wildlife reflect the conditions they experience in their ecosystem over the course of their life cycle.
  • Natural ecosystems are dynamic, evolutionary, and resilient.
  • Ecosystems are structured hierarchically.
  • Ecosystems are defined relative to specific communities of plant and animal species.
  • Biological diversity accommodates environmental variation.
  • Ecosystem conditions develop primarily through natural processes.
  • Ecological management is adaptive and experimental.
  • Human actions can be key factors structuring ecosystems.

The set of scientific principles that relate to artificial production, and emphasized by the latter two elements listed, are meant to minimize unintentional human influences on ecosystem structure. These principles can be divided along technological and managerial lines, differentiating between how hatchery fish are produced and how hatchery fish are used.

D. Guidelines on Hatchery Practices, Ecological Integration and Genetics.

Management of all hatcheries should be consistent with the life history of the cultured stock and the environmental conditions of the watershed, especially the annual temperature regime of the relevant section of native habitat represented in the stock of fish propagated. Life history strategies demonstrate the optimum course of action in the complexity of selective pressures exerted on them (Brannon, in press). Proper management, therefore, must include only measures that are consistent with those life histories, or severe impacts on the native populations should be expected. Management policy on such conventions as stock introductions (listed above), size and time of release, magnitude of release, genetic agenda, and recovery strategies are of major importance to the success of hatchery programs. Details on these issues are in the following guidelines, but it needs to be understood that in many cases where scientific principles are advocated, applied evidence is not available to demonstrate the precept. In these cases, it may be more appropriate to view the guidelines as hypotheses that need to address problems they exemplify -- as safeguards against unforeseen events that could destroy the viability of the runs managers are attempting to conserve. Some theories are troublesome to practitioners because their experiences do not support the axiom. Concerns about inbreeding are an example. Many populations of salmonids are small and inbred by the nature of the environment describing their habitat. In fact, where certain traits are critical to their survival, such as an innate complex orientation pattern to reach a destination, specificity rather than diversity defines fitness. This appears contrary to the theory, but in the broader range of the species, diversity is still the key to species stability. Measures taken to maintain the diversity present, or to prevent potentially negative effects of induced inbreeding, even within naturally inbred lines, are precautions that safeguard against artificially imposing a deleterious artifact of hatchery production on a population.

Present technology is bringing into application measures that improve the quality of fry at the time of emergence and at readiness of juveniles to enter the migratory phase. Providing required nutritional needs in a form available in artificial diets were some of the first advancements in hatchery technology (Hublou, 1963), and nutritional develops have continued (Forster and Hardy, 1995).

Substrate and darkness during incubation to maximize energy efficiency for growth are now employed routinely. These conditions were found to more accurately simulate natural incubation environments and produce larger fry at emergence than open tray or basket incubators (Brannon, 1965). Other technologies are also being employed, and their appearance in the list only reaffirms the importance placed on them.

Guideline 1. Technology should be developed and used to more closely resemble natural incubation and rearing conditions in salmonid hatchery propagation.

In developing hatchery technology, hatchery programs should work toward the goal of providing environments that resemble natural conditions during artificial propagation. These may include:

  • Incubation in substrate and darkness;
  • Incubation at lower densities;
  • Rearing at lower densities;
  • Rearing with shade cover available;
  • Exposure to in-pond, natural-like habitat;
  • Rearing in variable, higher velocity habitat;
  • Non-demand food distribution during rearing;
  • Exposure to predator training;
  • Minimize fish-human interaction;
  • Acclimation ponds at release sites;
  • Volitional emigration from release sites.

Rationale: Lower rearing densities, minimum exposure to humans, and shade cover over raceways enhances fish quality and maintains a behavior more similar to that of wild fish. Also, volitional migration when the fish are ready to begin their journey to sea is a technology practiced at some hatcheries, promoting natural transit behavior and less impact on the carrying capacity of the receiving stream or other water body. These are positive advancements in hatchery production operations that are encouraged to continue. Other practices need research on potential indirect effects. For example, although accelerated rearing can easily overcome any size deficiency of the fry experienced at the time of emergence, what isn't known are the potential impacts accelerated rearing will have on the normal biological development from embryo to fingerling, or the impact that large hatchery fish have on their wild counterparts.

Guideline 2. Hatchery facilities need to be designed and engineered to represent natural incubation and rearing habitat, simulating incubation and rearing experiences complementary with expectations of wild fish in natural habitat.

Rationale: Hatchery technology in the Columbia basin has relied primarily on standard tray incubation and concrete raceway technology based on engineering designs that emphasize efficiency and convenience for fish culture operations. Qualities associated with natural habitat have not been incorporated in such designs, and fish reared in standard concrete raceways learn behavior conducive to those situations, and out of harmony with what they will experience when released into natural conditions. Comparatively poor survival success of hatchery fish is attributed in part to such experiences atypical of natural conditions. Technology needs to design facilities that utilize engineered earthen stream channels that represent natural habitat with cover, glides and pools, woody debris and flow patterns mimicking natural habitat. Incubation and rearing could take place in the same channel facility, at densities appropriate to encourage natural feed (supplemented with formulated diets) and provide learning opportunities under simulated natural conditions. Training would include exposure to size variability among other species that share the habitat, and limited exposure to predation.

Guideline 3. New hatchery technology for improving fish quality and performance needs to have a plan for implementation and review at all hatchery sites, where appropriate, to assure its application.

Rationale: Assuring that technological advances in hatchery propagation are part of hatchery operational plans is critical to the implementation of changes meant to improve the quality and performance of hatchery fish in the natural environment. Often such implementation occurs only among those hatcheries where a willingness to make changes exists, given that information on new technology is even transmitted. It is important that technological advancements are first verified and the mechanism through which such technology enhances quality or performance is well understood. Then there needs to be a process for implementing the technology, with accountability for its installation and review to make it as routine as feed delivery, assuring its application and evaluation.

Guideline 4. To mimic natural populations, anadromous hatchery production strategy should target natural population parameters in size and timing among emigrating anadromous juveniles to synchronize with environmental selective forces shaping natural population structure.

Rationale: Hatchery programs have tended to concentrate on large-size fish at the time of release, as well as varying the timing of release, to facilitate higher return success. Although such rationale is understandable from the standpoint of improving hatchery fish survival, such practices introduce atypical migrants that create an alteration in the natural continuity of events around which population strategies have evolved. With the exception of fall chinook that normally show variation in migratory distribution patterns, such practices with other anadromous salmonids are believed to have negative effects on fitness of wild fish, and may perturb population structure to the disadvantage of natural populations. Based on interpretations of population structure and life history patterns (Brannon, in press), avoiding atypical size and time at migration among hatchery fish is desirable, even with the immediate disadvantage it may have on hatchery return success. The point is that hatcheries should focus on mimicking the natural environmental selective forces within the target watershed so hatchery-produced emigrating juveniles exhibit the same size distributions as juveniles from the natural population.

Guideline 5. To mimic natural populations, resident hatchery production strategy should target population parameters in size and release timing of hatchery-produced resident juveniles to correspond with adequate food availability and favorable prey to maximize their post-stocking growth and survival.

Rationale: Post-stocking mortality of a wide array of resident fish species could be reduced by implementing release strategies that match released fry or fingerlings with periods of adequate production and availability of planktonic and invertebrate food items. Attention to vulnerability of stocked resident fish fry or fingerlings as prey, and abundance and behaviors of potential predators in receiving waters can also significantly improve initial post-stocking survival.

Guideline 6. Supplementation hatchery policy should utilize ambient natal stream habitat temperatures to reinforce genetic compatibility with local environments and provide the linkage between stock and habitat that is responsible for population structure of stocks from which hatchery fish are generated.

Rationale: Temperature is a crucial factor affecting adult salmonid return timing and spawning (Brannon, 1987), and is an important factor affecting the length of time juveniles spend in stream residence before migrating to sea. This fundamental influence has formed the framework around the evolution of salmonid population structure. Temperature demonstrates its pivotal effect on the evolution of life history forms through temporal influences on egg incubation and juvenile growth as the basis for differentiation of adult timing and juvenile residence behavior, respectively. It is argued, therefore, that temperature is one of the most critical environmental factors affecting life history forms peculiar to their respective stream system. Temperature is the environmental parameter motivating the evolution of stock predispositions selectively reinforced over time to represent genetically distinct units. Temperature regimes during early life history are typically altered from the natural pattern by hatchery use of ground water for incubation. Hatchery management policy should adhere to using the ambient temperature regime of their natal environments to maintain the compatibility of hatchery fish with the natural system and the effectiveness of hatchery contribution to the natural spawning population. In some cases, wild fish spawn on spring-fed reaches of streams, and the appropriate incubation temperatures in those situations would be incubation substrate temperatures. However, when it comes to the rearing phase where the growth rate is determined by temperature (Brett et al, 1969), it is the daily ambient mean temperature that is important to follow.

Guideline 7. Salmonid hatchery incubation and rearing experiences should use the natal stream water source whenever possible to enhance homestream recognition.

Rationale: Another factor associated with the natal habitat and homing accuracy is the homestream odor profile that provides the fingerprint ultimately identified with the homestream spawning and incubation site. Hatchery programs not only use ground water for incubation, but hatcheries are usually away from the natal environment to which local stocks have adapted. The assumption is that by planting the fish in the proper location, hatchery fish will home to that stream on return. While this is true, imprinting is sequential (Brannon and Quinn 1990; Quinn et al. 1990), and the incubation environment is the first odor cue on which alevins imprint and the ultimate identity sought by returning fish (Brannon 1982). Strays are common in some hatchery populations and lack of having imprinted during the incubation phase is suggested as being responsible for higher stray rates. To assure the continuity between hatchery fish genetics and local stream habitat, the water sources closely linked with the natal environment are most desirable. This guideline is most difficult to incorporate with present hatcheries because the capital structure and water system have been established without those priorities. New facilities, however, should be located on sites with access to appropriate water sources.

Guideline 8. Hatchery release strategies need to follow standards that accommodate reasonable numerical limits determined by the carrying capacity of the receiving stream to accommodate residence needs of non-migrating members of the release population.

Rationale: Standards should include impact considerations on the wild fish residing in the system, and should be based on life history requirements of the cultured stock. Hatchery releases of cultured fish into receiving streams occur under the assumption that the river is used primarily as a migratory conduit to the estuary. This is true for only those fish (smolts) at emigration readiness. Fish not ready to migrate will take up transitional residence in the stream, causing the potential negative interactions with wild fish present. Care should be taken to limit release numbers consistent with the estimated rearing capacity of the system to minimize impacts on wild fish. Moreover, the practice of releasing fish to make space for other broods should be discontinued. Release of hatchery fish must fit a schedule consistent with life history requirements of the natural population from which the brood lot was derived.

Guideline 9. Hatchery programs should dedicate significant effort in developing small facilitates designed for specific stream sites where supplementation and enhancement objectives are sought, using local stocks and ambient water in the facilities designed around engineered habitat to simulate the natural stream, whenever possible.

Rationale: Hatcheries are most often developed around the concept of a central facility from which fish are outplanted to many other streams or acclimation ponds, not always using native stocks in each instance. The rationale is usually related to the major capital expenditures for hatcheries under the old hatchery concept. It is much more desirable to locate smaller, stream-specific operations to maintain stock identity with the particular stream targeted. Nothing larger than a station capacity of 100,000 eggs or 25,000 fingerlings would be required on smaller tributary systems. This would require no more than a rearing channel to accommodate such small inventories, but small numbers in natural-like habitat is the ideal for supplementation of native salmonids. Even fry releases can be a feasible option to consider under these circumstances associated with the natural habitat, when conditions for supplementation can call for such limited, and perhaps temporary, artificial application. Again, this hypothesis is impossible with present facilities located where they are and with capital commitments in water and concrete. However, with new artificial production facilities, part-time stations of this nature would address both the biological and ecological requirements that future operations must satisfy.

Guideline 10. Genetic and breeding protocols consistent with local stock structure need to be developed and faithfully adhered to as a mechanism to minimize potential negative hatchery effects on wild populations and to maximize the positive benefits that hatcheries can contribute to the recovery and maintenance of salmonids in the Columbia ecosystem.

Rationale: As an integral component in a complex ecological system, salmonid stocks have evolved with their environments. Spawning time, emergence timing, juvenile distribution, marine orientation and distribution are not random, but rather occur in specific patterns of time and space for each population (Brannon 1984), and include behavior that evolved under historical abundance constraints in natural populations. The appropriate seed stock is key to producing viable, healthy fish for the respective system. Given the ecosystem concept for management protocol in the Columbia Basin, population genetics and the natural environment salmonid stocks have evolved under have to become blueprints in hatchery programming. Differences between the genetics of wild stocks and hatchery fish (Ryman and Sthl, 1980; Allendorf and Utter, 1979) are considered by the SRT as a major source of poor hatchery fish performance in the wild. Development and adherence to strict genetic guidelines and breeding protocols consistent with local population structure is essential for effective hatchery contribution to wild production and maintenance of local genetic diversity.

Guideline 11. Hatchery propagation should use large breeding populations to minimize inbreeding effects and maintain what genetic diversity is present within the population.

Rationale: One of the potential negative effects of artificial production is that relatively small breeding populations are involved in hatchery programs. Even when 100,000 fingerlings are scheduled for supplementation, that number represents a little over 25 females for brood stock, and a relatively limited representation of the gene pool. In the Idaho captive rearing project where juveniles are intercepted and reared to maturity as a means to avoid demographic risks of cohort extinction, only enough parr are captured to provide 20 spawners for each population, which is even a smaller representation of the gene pool. The risks in using small breeding populations are loss of diversity and magnifying the effect of deleterious genes. Hatchery survival can increase the contribution of the artificially propagated fish out of proportion with number, with the result that over time the hatchery population will become increasingly more represented among the natural spawners. The issue is not just inbreeding, because many healthy natural populations are very site- specific in unique environments and represent inbred lines. The risk is that hatchery production can accelerate the potential harmful effects of inbreeding by involving only a small portion of the returning adults in the artificial breeding population. To avoid these negative effects of hatchery production, a large number of spawners should be included in the breeding protocol. When the run is relatively small, this may require live spawning, and removing only a portion of the eggs from each female and subsequently releasing the fish to continue spawning naturally.

Guideline 12. Hatchery supplementation programs should avoid using strays in breeding operations with returning fish.

Rationale: In situations where strays constitute a substantial proportion of hatchery return populations, care should be taken to avoid inter-stock hybridization because of the loss of adaptive traits in the resulting progeny. Reisenbichler (1998) demonstrated examples of reduced fitness from hybridization. Stock hybridization breaks down genetic homeostasis and disrupts co-adaptive gene complexes, which lowers the fitness of the local stock. A policy needs to be developed to minimize the contribution of strays to the local hatchery stock. In the situation where a hatchery is supplementing a native population, inter-stock hybridization should be avoided to prevent loss of adaptive fitness.

Guideline 13. Restoration of extirpated populations should follow genetic guidelines to maximize the potential for re-establishing self-sustaining populations. Once initiated, subsequent effort must concentrate on allowing selection to work by discontinuing introductions.

Rationale: When undertaking restoration projects where populations have been extirpated, restoration strategies need to be given careful consideration and reference to genetic guidelines. Where neighboring populations represent appropriate characteristics, stock transfer may be the best strategy. When suitable stocks are not available, or when information is insufficient with which to match a donor stock, then inter-stock hybridization may be an alternative. Inter-stock hybridization breaks down co-adapted gene complexes and releases genetic variability on which selection can work. Restoration can use different genetic-based approaches, depending on the situation, but the characteristics of the donor stock(s) are critical. The key is to follow through with the strategy selected and allow sufficient time for the founders to be selectively established by avoiding continued introductions in the target stream.

Guideline 14. Germ plasm repositories should be developed to preserve genetic diversity for application in future recovery and restoration projects in the basin, and to maintain a gene bank to reinforce diversity among small inbred natural populations.

Rationale: One of the most important considerations in the Columbia Basin fisheries management plan is to preserve the existing genetic diversity. Diversity is inherent to the stability of the species. The various systems, with their component population networks, are the sanctuaries of variability. Recovery and enhancement of natural production in the basin will not be a rapid process, and in the meantime further loss of diversity may occur, with some populations becoming extinct. It is critical, therefore, to launch an immediate program to preserve germ plasm by collecting and cryopreserving milt from all naturally spawning populations that can be reached. The technology is available and currently is being employed with some ESA-listed salmonid stocks. This effort needs to be expanded and given greater priority. Germ plasm should be collected from each population on more than one broodyear to develop as complete a repository as possible. The availability of germ plasm for future use in maintenance of diversity or restoration of extirpated runs will be invaluable in the long-term ecological framework of the managed river.

Guideline 15. The physical and genetic status of all natural populations of anadromous and resident fishes need to be understood and routinely reviewed as the basis of management planning for artificial production.

Rationale: Knowing the status of the endemic stock where hatchery fish are involved is imperative under the ecological framework of fisheries management. Information should include life history, population structure and the habitat utilized. This knowledge must include, in addition to the traditional numerical status of the run, details on its population structure, distribution patterns, size and timing of migration, and the level of genetic specificity and diversity within the population. The habitat status associated with the population must also be known, including the area available, the condition of the habitat, new areas that can be developed, and the carrying capacity. This information is essential to the management of all native anadromous and resident species in the Basin, which will require ecological expertise at the programmatic and hatchery levels.

E. Guidelines on Research and Monitoring

Good management is the key to successful integration of hatcheries into a functioning and dynamic ecosystem. Research to improve artificial production, the extent of its application, and its limitations is basic to the effective management of hatcheries in the basin. In this regard, monitoring is also a critical element in the management process. Knowing what is successful and what must change is impossible without appropriate monitoring programs.

Guideline 16. An in-hatchery fish monitoring program needs to be developed on performance of juveniles under culture, including genetic assessment to ascertain if breeding protocol is maintaining wild stock genotypic characteristics.

Rationale: The NPPC needs to design a scientifically valid monitoring program for the basin hatcheries. Special attention should be paid to the collection of valid data that applies to routine assessment of juvenile performance in the hatchery incubation and rearing phase, up to the point of release. Genetic monitoring of the stock inventory would include descriptive evaluation at first feeding and at release time to assess if hatchery propagation is altering genotypes from that of the wild population.

Guideline 17. A hatchery fish monitoring program needs to be developed on performance from release to return, including information on survival success, interception distribution, behavior, and genotypic changes experienced from selection between release and return.

Rationale: The NPPC needs to design a scientifically valid monitoring program for hatchery fish performance after release from the culture facilities. In addition to return success, attention should be paid to relative interception distribution (tag analysis) of hatchery fish to compare performance parameters with native fish. Special attention should also be given to descriptive genetic assessment at time of return to determine if genotypes surviving are representative of genotypes released, and compatible with the native stock. With the advent of the PIT tag system, opportunities to gather more specific information exists. Significant insights can be gained on straying, migratory route and timing that are key to honing hatchery programs.

Guideline 18. A study is required to determine cost of monitoring hatchery performance and sources of funding.

Rationale: A study should be undertaken to consider how much monitoring programs will cost and what reallocation of effort in the production programs would be required to fund adequate monitoring efforts where additional funds cannot be secured.

Guideline 19. Regular performance audits of artificial production objectives should be undertaken, and where they are not successful, research should be initiated to resolve the problem.

Rationale: Routine audits of hatchery production objectives should be established (for example, every five years) to determine if they are achieving their objectives. In those cases where programs or hatcheries are not showing any production benefit, they should be re-prioritized to research-only until the problems can be resolved. In some cases, research may disclose that the objectives are not attainable. In those situations, emphasis can then be redirected, programs changed, or discontinued.

Guideline 20. The NPPC should appoint an independent peer review panel to develop a basinwide artificial production program plan to meet the ecological framework goals for hatchery management of anadromous and resident species.

Rationale: With the development of the broad ecological framework in the basin placing emphasis on hatchery management in the arena of conservation fisheries and ecosystem function, it will be necessary for practitioners and fisheries scientists to work together in developing the appropriate hatchery program plans to achieve the ecosystem goal. Problems that have prevented hatcheries from achieving their goals, or insights on what may be impossible to achieve in the ecosystem approach at the hatchery level, cannot be ascertained without major contribution from hatchery managers experienced in the system. Also, the inherent conflict between the concept of ecosystem management and the concept of management for harvest mitigation has to be resolved within the ecosystem framework. Those resolutions, and the development of the hatchery program plan addressing specific actions needed to achieve the goal, are essential elements early in the planning process. The responsibility will require appointment of an independent peer review panel that can give careful and appropriate consideration, through solicitation of agency, tribal and public interests, to past management experiences.

Given the new management emphasis on wild stocks, special consideration must be given to the possibility that some of the maladaptive traits developed by hatchery fish in hatcheries could be expressed even more deleteriously when those fish attempt to spawn naturally (in a supplementation program) or when they interact genetically (as strays) with natural spawning populations, or as they interact with natural stocks ecologically throughout the post-release portion of the life cycle. While these possible risks are in some sense the most alarming, they are also the most poorly documented, and the quantitative strength of the underlying forces are not well understood. Therefore, a large research and monitoring effort needs to be directed at these questions of genetic and ecological effects of hatchery fish on naturally spawning stocks. The results of these studies are needed to lay to rest some of the fears about worst-case scenarios, and they are also needed to teach us how to modify hatchery management to achieve the most positive kinds of interactions with wild stocks.