Hierarchical conservation and management of Indiana bat (Myotis sodalis) habitat may benefit from use of species distribution models. White-nose syndrome has caused additional declines for this endangered bat, requiring use of historical presence locations for habitat-related analy- ses. We created random forest presence/pseudo-absence models to assess the distribution and availability of Indiana bat habitat across the 670,000-ha Monongahela National Forest (MNF), West Virginia, USA. We collated historical roost and capture locations, both individually and in combination, to examine impacts of various biotic and abiotic predictors on roosting and foraging habitat of Indiana bats. Our final concordance map suggests that In- diana bat habitat was abundant (37.2% of the MNF) but localized, with predicted suitable areas often associated with edges of dry-calcareous forests. We observed significant variation between models, with the capture-only model independently identifying the greatest amount of potential habitat (47.8%). However, 21.9% of all potential Indiana bat habitat was identified by complete inter-model agreement. Our SDM outputs may assist land managers in identifying avoidance areas and new survey sites (i.e., capture and acoustic sampling) to support forest management activities.
