Abstract
North American bats are experiencing declines in part due to anthropogenic impacts resulting in habitat loss and disturbance. In eastern deciduous forests, bats rely on forest resources for all or part of the year. Therefore, to promote conservation of bats, it is essential to determine whether current forest management techniques are compatible with habitat use by bats. We evaluated the relative effect of landscape characteristics, including forest management variables, on sex-specific foraging habitat of an insectivorous forest-dwelling bat species, the evening bat (Nycticeius humeralis), and predicted areas of suitable habitat for N. humeralis. A total of 18 variables were assessed using a maximum-entropy (Maxent) machine-learning approach: eight land use–land cover classes, three stand types, two topography measures, normalized difference vegetation index, and four forest management variables. Females showed the highest probability of presence closer to stands treated with prescribed fire, whereas males showed the highest probability of presence closer to reforested stands. In general, males exhibited more flexibility than females in their habitat selection. The Maxent model further indicated that habitat associated with suitability of > 70% was ∼4 times larger for males than females, and predicted an additional area of suitable foraging habitat where no presence locations had been recorded. Our modeling approach may be suitable for other researchers to derive models appropriate for a wide range of bat species.