Ecologically effective ecosystem management will require the development of a robust logic, rationale, and framework for addressing the inherent limitations of scientific understanding. It must incorporate a strategy for avoiding irreversible or largescale environmental mistakes that arise from social and political forces that tend to promote fragmented, uncritical, short-sighted, inflexible, and overly optimistic assessments of resource status, management capabilities, and the consequences of decisions and policies. Aquatic resources are vulnerable to the effects of human activities catchment-wide, and many of the landscape changes humans routinely induce cause irreversible damage (e.g., some species introductions, extinctions of ecotypes and species) or give rise to cumulative, long-term, large-scale biological and cultural consequences (e.g., accelerated erosion and sedimentation, deforestation, toxic contamination of sediments). In aquatic ecosystems, biotic impoverishment and environmental disruption caused by past management and natural events profoundly constrain the ability of future management to maintain biodiversity and restore historical ecosystem functions and values. To provide for rational, adaptive progress in ecosystem management and to reduce the risk of irreversible and unanticipated consequences, managers and scientists must identify catchments and aquatic networks where ecological integrity has been least damaged by prior management, and jointly develop means to ensure their protection as reservoirs of natural biodiversity, keystones for regional restoration, management models, monitoring benchmarks, and resources for ecological research.(KEY TERMS: ecosystem management; ecological integrity; aquatic biodiversity; cumulative effects; conservation reserves; landscape planning; watershed analysis.)

Over 300 native stocks of Pacific salmon, steelhead, and coastal cutthroat trout (Oncorhynchus spp.) are at risk of extinction in the Pacific Northwest. With only limited resources available for conservation and recovery, prioritization of these stocks may become necessary if meaningful measures are to be implemented. We propose criteria by which prioritization may be guided. First, we rank stocks for risk of extinction, either by population viability analysis or by a set of surrogate measures. Then we rank stocks for biological consequences of extinction, using sets of questions designed to establish the genetic and evolutionary consequences and the ecological consequences if a stock were to become extinct. Together, these rankings allow stocks to be prioritized for a range of possible actions, with those stocks at highest risk and bearing the greatest biological consequences of extinction receiving attention first. Application of the prioritization process to 20 Pacific anadromous salmonid stocks worked as intended, although data limitations are considerable. The process is most likely to work successfully when applied to many stocks on which data exist, when several experts carry out the prioritization, and when the results are peer reviewed.

No abstract