Using occupancy models to determine mammalian responses to landscape changes.

Determining impacts of anthropogenic landscape changes on wildlife populations is difficult. Besides the challenges of designing field studies to document conditions before and after landscape changes occur, assessment of population responses (e.g. changes in population density) often provide poor inference because of sampling limitations. Estimation of occupancy, however, only requires data on detection or non-detection of a species and might provide better inference. To demonstrate the utility of occupancy models, we used data from an American black bear (Ursus americanus Pallas) population in North Carolina, USA to test our research hypothesis that documented declines in site occupancy of black bears would be greater near a new four-lane highway. We used multi-season occupancy models to estimate site occupancy based on bear visitation to survey sites before and after completion of the new highway and as a function of distance to the highway. Site occupancy declined from 0.81 to 0.35 between the two study phases, but was not a function of distance to the highway. Therefore, the impact of the new highway on occupancy extended to the entire study area. Our case study demonstrates that occupancy models can provide powerful inference regarding the potential impacts of landscape changes on species occupancy. As urban areas and transportation infrastructure are rapidly expanding in developing regions of the world, the need to determine how these changes affect mammal populations and how they might be mitigated increases accordingly. Because field sampling for occupancy models only requires detection data, surveys can be conducted for extensive geographic areas, thus making these surveys particularly applicable to studies of large mammals.