Nocturnal habitat selection of bats using occupancy models

Abstract

ABSTRACTThreats to North American bat species necessitate the quantification of bat habitat. To characterize bat habitat, we recorded bat echolocation calls from state forests in Indiana, USA. We identified calls using automated classification software and constructed detection histories for 3 species of conservation concern: Indiana bat (Myotis sodalis), little brown bat (Myotis lucifugus), and northern long‐eared bat (Myotis septentrionalis). We analyzed these data using single‐ and multi‐species occupancy approaches to estimate occurrence and detection probabilities at the genus level (Myotis) and for each species separately. We constructed models using a suite of habitat covariates (e.g., forest cover, edge, streams, hibernacula, public roads) and determined the effect of these features on occupancy for each species (or group). We validated all models using training and independent data. Occupancy of all species was affected by the proportion of forest in surrounding areas. For most models, bat occupancy was greatest in areas with a lower proportion of forest within 1 km. Occupancy was also often greatest in recently harvested or intact, mature second‐growth stands as opposed to stands 10–30 years following harvest. The spatial configuration of water and roads also affected species occupancy. Our methods allowed us to assess the factors that affect occupancy for the genus and individual species, thus allowing for the identification of features useful in classifying important habitat at both taxonomic levels. The results of this study can aid land managers in identifying habitat that may be used by foraging bats and in planning future timber harvests in managed forests. © 2017 The Wildlife Society.