Modeling Grizzly Bear Habitat Suitability in Idaho

Many of the issues confronting wildlife managers and scientists are challenging the conventional spatial boundaries defined by administrative units. This holds especially true in the management of large carnivores such as wolverines, wolves, mountain lions, and grizzly bears. Individual grizzly bears range over 400 to 1000 square kilometers in a lifetime, while viable bear populations may require 10 to 30 times as much space. Such scales require a very broad view of habitat conditions. Not insignificantly, understanding these bears requires regional GIS databases that transcend state and even national boundaries.

Idaho is currently grappling with a number of issues related to grizzly bear management, including the potential reintroduction of a population into its central mountain wilderness areas and the management of humans in areas currently occupied by grizzly bears in the Panhandle and in the Yellowstone ecosystem. There has been reoccurring debate over the extent and location of "suitable" habitat. In addition, there are concerns about fragmentation and insufficient overlap between physically productive habitat and wilderness areas secure from substantial human intrusion. Scientists from the University of Idaho's College of Forestry, Wildlife, and Range Sciences GIS Lab and from the National Biological Service's Cooperative Park Studies Unit are trying to answer to these questions and develop a prototype for looking at the suitability of habitat for large carnivores elsewhere.

This research has drawn upon regional GIS databases, including GAP data for the state of Idaho, to model grizzly bear habitat suitability. These data were rasterized and combined in ARC/INFO grid format. Since grizzlies, like most other large carnivores, die primarily because humans kill them, a large part of this model deals with human-related features such as townsites, roads, and trails. This information is integrated into a measure of potential human activity for each map pixel and treated as an analogue of grizzly bear death rate. Information on vegetation, topography, and ungulate populations is integrated into seasonal measures of potential habitat productivity and treated as an analogue of birth rate. These two metrics are then combined in a way that culminates the analogy—by subtracting the standardized index of human activity from the standardized index of habitat productivity, the resulting measure is a direct analogue to population dynamics.

This model has already produced information of value to management deliberations. Maps have been produced that show seasonal habitat productivity for the entire state, as well as the location of "suitable" habitat defined by increasingly restrictive criteria. These maps show that, by most standards, there is abundant well-protected grizzly bear habitat in central Idaho that could potentially support a reintroduced bear population. They have also highlighted the potentially precarious status of existing grizzly bear populations, especially in the Panhandle. These results, as well as a description of the method, are parts of a manuscript that is currently being reviewed prior to submission to a journal for publication.

Even though significant progress has been made with this project, some major work remains ahead. In particular, we are prioritizing efforts to relate model outputs to parameters more directly relevant to management considerations, including actual grizzly bear birth and death rates. To date, we have partially confirmed the model by comparing outputs with delineations of currently occupied habitat and by assessing statistical relationships with bear sightings. We anticipate substantial future progress by extending the method to well-studied bear populations in areas such as the Yellowstone ecosystem and the northern Rocky Mountains of Montana.

David J. Mattson
National Biological Service
University of Idaho, Moscow