Population estimation, habitat associations and range expansion of black bears in the Upper Midwest

North American Black Bear (Ursus americanus) populations have grown and expanded throughout the Upper Midwest in the last 20 years. In this dissertation I present data on the abundance of black bears in Wisconsin, explore the statistical properties of population indexing methods, and analyze black bear habitat associations in Michigan and Minnesota and Wisconsin.;I estimated the abundance of Wisconsin's 2006 black bear population using a tetracycline-biomarker capture-recapture technique. Volunteers deployed 2,235 tetracycline laced baits of which 971 were consumed by bears. I collected 2,587 ribs in 2006, and 2,378 in 2007. All ribs were analyzed for tetracycline, 69 were positive in 2006 and 44 were positive in 2007. I estimate the mean state wide population to be 33,657 individuals. I discuss the biases associated with this technique compared with other approaches and present alternative analysis to account for bias from unequal probability of capture.;Population indices based on visits to detection stations commonly are used to monitor carnivore populations. The functional relationship between index responses and population density is poorly understood. I created a flexible simulation environment to study the response of detection station population indices under a wide variety of species life history and study design characteristics. Proportional indices exhibited non-linear responses to changes in population density while cumulative indices responded linearly. Responses are functions of home-range size, individual detection probabilities and spatial arrangement of animals and sampling stations. Cumulative indices have significant statistical advantages over proportional indices including higher power, linearity, consistent index response across a wide range of densities, and ability to aggregate data to meet minimum sample size requirements.;I used data from capture-recapture programs in Minnesota, Wisconsin and Michigan to quantify habitat associations of black bears across the three states. My model is based on 4,733 bait locations with 1,913 successful bear detections. I modeled landscape-scale habitat association with black bear occupancy using Public Land Survey townships. I estimated occupancy = 0.744 (SE = 0.017) and detection probability = 0.560 (SE = 0.011). Detection probability associated negatively with Julian date of baiting and occupancy was most closely associated with forest and wetland habitat types.