Date of Award


Degree Type


Degree Name

Doctor of Philosophy


Ecology and Evolutionary Biology

Major Professor

Michael R. Pelton

Committee Members

Dr. Frank van Manen, Dr. Mark Knot, Dr. Mike Huston, Dr. Lou Gross, Dr. Arthur Echternacht


The American black bear (Ursus americanus) historically occurred throughout the forested regions of the continental United States of America (USA), and was absent only in central Nevada and portions of southern California and Arizona. Today, black bears only occur in a fraction of their former range, particularly in the Southeast where bear populations are highly fragmented and isolated.

Past research efforts have investigated the habitat preferences of black bears in several regions of the continental USA, but these studies were either limited to relatively small geographic areas or focused on selection preferences of individual bears. As a result, there is a general lack of knowledge about the composition of environmental conditions that demarcate suitable bear range or how bears respond to environmental variability at the macro-scale. To that end, I investigated the relationship between spatial environmental variability and black bear occurrence in the continental USA. The objectives of this study were to: (1) determine if spatial environmental variability influences the distribution of black bears, (2) identify spatial environmental variables correlated with black bear distribution, and (3) develop a model to predict the spatial distribution of suitable bear habitat and potential relocation areas.

I used logistic regression to assess the correlation between spatial environmental variability and bear occurrence, and to develop a model predicting the spatial occurrence of bears. I divided the continental USA into two areas; bottomland hardwood areas of the Southeastern coastal plain, termed lowland environments, and the remainder of the continental USA, termed upland environments. I found that bear presence was correlated with 10 landscape-scale variables in upland and lowland environments of the continental USA. In upland environments, bear presence was positively correlated with all macrohabitat types (except grassland-shrubland mosaic), lands actively managed as wild lands, snowfall > 122 cm, and increasing levels of spring-summer normalized difference vegetation index (NDVI). In contrast, bear presence was negatively correlated with forest fragmentation, road density index, human densities > 10 persons/km2, greater distances from streams, and increasing levels of wetness.

In lowland environments, bear presence was positively correlated with deciduous and evergreen forests, grassland-shrub land mosaic, perforated through edge levels of forest fragmentation, lands actively managed as wild lands, human density <43 persons/km2, increasing soil nitrogen levels, and increasing levels of spring NDVI. In contrast, bear presence was negatively correlated with herbaceous-woodland wetland, sparsely vegetated areas, increasing road density index, and increasing levels of wetness.

The relative probability of bear occurrence models indicated that there is about 2.8 million square kilometers of suitable bear habitat in the continental USA, distributed in about 1,400 distinct patches. However, only 306 of the suitable habitat patches are ≥ km2, corresponding to 2,721,803 km2of suitable habitat. The models identified 981,061 km2 of vacant suitable bear habitat in the continental USA. These habitats are distributed in 394 patches, of which 155 (743,558 km2) are adjacent to occupied bear range and 239 (237,503 km2) are isolated from occupied range. The probability of occurrence models identified 34 habitat patches as priority (≥5,000 km2) reintroduction areas.

This study describes bear-habitat use patterns across a broad spatial extent and identified landscape variables that may influence bear occurrence. In so doing, the study provides a new interactive model that can be adapted and used to predict the relative probability of bear occurrence due to changing spatial conditions and identify potential reintroduction areas. In addition, the results of this study may contribute to future research efforts on corridor analyses or metapopulation dynamics.

Files over 3MB may be slow to open. For best results, right-click and select "save as..."