Directness of resource use metrics affects predictions of bear body fat gain

Many North American ursids rely on an annual hyperphagic period to obtain fat reserves necessary for winter survival and reproduction. Identifying causes of variation in body fat gain may improve understanding of how bear resource use affects body condition. We used data from southcentral Alaska to model changes in percentage body fat of adult female American black bears (Ursus americanus) in 1998 and 2000 and brown bears (Ursus arctos) in 2000. We used year, proportion of radio locations in different habitats, distance to streams containing salmon (Onchorynchus spp.), and degree of radio location clustering as predictors for black bears and elevation, distance to streams containing salmon, and degree of radio location clustering as predictors for brown bears. Degree of location clustering was the only predictor variable supported by parameter coefficients in black bear models, supporting our hypothesis that metrics of energetics perform better as predictors of body condition than habitat use. With every unit increase in location clustering black bear body fat increased 2 %. No predictor variables influenced variation in brown bear change in body fat. Some variables previously found useful for predicting bear presence (e.g., habitat) were not useful in predicting changes in body fat, an important biological outcome for these species. Rather than assuming fitness benefits of habitat-level selection, we recommend including metrics of energetics that might more directly influence biological outcomes.