A MULTIVARIATE MODEL OF FEMALE BLACK BEAR HABITAT USE FOR A GEOGRAPHIC INFORMATION-SYSTEM

Simple univariate statistical techniques may not adequately assess the multidimensional nature of habitats used by wildlife. Thus, we develo ped a multivariate method to model habitat-use potential using a set o f female black bear (Ursus americanus) radio locations and habitat dat a consisting of forest cover type, elevation, slope, aspect, distance to roads, distance to streams, and forest cover type diversity score i n the Ozark Mountains of Arkansas. The model is based on the Mahalanob is distance statistic coupled with Geographic Information System (GIS) technology. That statistic is a measure of dissimilarity and represen ts a standardized squared distance between a set of sample variates an d an ideal based on the mean of variates associated with animal observ ations. Calculations were made with the GIS to produce a map containin g Mahalanobis distance values within each cell on a 60- x 60-m grid. T he model identified areas of high habitat use potential that could not otherwise be identified by independent perusal of any single map laye r. This technique avoids many pitfalls that commonly affect typical mu ltivariate analyses of habitat use and is a useful tool for habitat ma nipulation or mitigation to favor terrestrial vertebrates that use hab itats on a landscape scale.