127
9
Assessing Land-Use Impacts on Bull
Trout Using Bayesian Belief Networks
Danny C. Lee
Introduction
Management agencies responsible for public lands within the United States increasingly find their activities circumscribed by considerations of threatened,
endangered, or sensitive species (TES species). In the western states, where the
U.S. government owns a majority of the land base, many TES species are critically dependent on habitat conditions within federal lands. Federal regulations
provide specific guidelines when dealing with TES species on these lands. For
example, U.S. Department of Agriculture Forest Service (Forest Service) regulations require the Forest Service to maintain viable populations and promote recovery of TES species throughout their native range. Commonly, the Forest Service
prepares biological assessments for all management activities that may affect TES
species. Such activities include timber sales, road construction, mining leases,
grazing allotments, and recreational enhancements. It is the responsibility of the
forest supervisor and staff to assess the threat posed by each activity; those that
may adversely affect TES species are either modified to mitigate the impact or
abandoned. Species protected under the Endangered Species Act additionally
require that the federal oversight agency (either the U.S. Fish and Wildlife Service
or the National Marine Fisheries Service, depending on the species) concurs with
the impact assessment and proposed mitigation measures.
The contentious nature of species protection combined with the large number of
assessments required by the agencies has prompted a huge demand for information and methods that can provide quantitative estimates of the risks to TES
species posed by land-use activities. Linkages between management activities and
population risk are complex, however, and the level of information available
about them inevitably is incomplete. Frequently, biological assessments are based
largely on the judgment and experience of forest biologists because specific
information on local populations simply is unavailable and unlikely to be obtained
within a reasonable timeframe. Managers need methods that recognize the uncertainty inherent in decisions made in information-poor environments, yet are
efficient, rigorous, and defensible. Such tools must be capable of incorporating
complexity, while simultaneously acknowledging uncertainty.
9
Assessing Land-Use Impacts on Bull
Trout Using Bayesian Belief Networks
Danny C. Lee
Introduction
Management agencies responsible for public lands within the United States increasingly find their activities circumscribed by considerations of threatened,
endangered, or sensitive species (TES species). In the western states, where the
U.S. government owns a majority of the land base, many TES species are critically dependent on habitat conditions within federal lands. Federal regulations
provide specific guidelines when dealing with TES species on these lands. For
example, U.S. Department of Agriculture Forest Service (Forest Service) regulations require the Forest Service to maintain viable populations and promote recovery of TES species throughout their native range. Commonly, the Forest Service
prepares biological assessments for all management activities that may affect TES
species. Such activities include timber sales, road construction, mining leases,
grazing allotments, and recreational enhancements. It is the responsibility of the
forest supervisor and staff to assess the threat posed by each activity; those that
may adversely affect TES species are either modified to mitigate the impact or
abandoned. Species protected under the Endangered Species Act additionally
require that the federal oversight agency (either the U.S. Fish and Wildlife Service
or the National Marine Fisheries Service, depending on the species) concurs with
the impact assessment and proposed mitigation measures.
The contentious nature of species protection combined with the large number of
assessments required by the agencies has prompted a huge demand for information and methods that can provide quantitative estimates of the risks to TES
species posed by land-use activities. Linkages between management activities and
population risk are complex, however, and the level of information available
about them inevitably is incomplete. Frequently, biological assessments are based
largely on the judgment and experience of forest biologists because specific
information on local populations simply is unavailable and unlikely to be obtained
within a reasonable timeframe. Managers need methods that recognize the uncertainty inherent in decisions made in information-poor environments, yet are
efficient, rigorous, and defensible. Such tools must be capable of incorporating
complexity, while simultaneously acknowledging uncertainty.