Stand- and landscape-level roost selection by forest bats in the Ouachita Mountains

Abstract

Scope and Method of Study: The purpose of this study was to determine roosting habitat preferences by forest bats and to create landscape-level models to determine characteristics of the landscape that were associated with roosting locations. I located 428 roosts from 162 individual bats of 6 species. Using Euclidean distance statistics and the individual bat as the experimental unit, I determined preference and avoidance for 13 habitat types in the Ouachita Mountains of central Arkansas. I also used logistic regression to create scale-dependent models that differentiated habitat classes, topography, and landscape patch metrics of roost locations from random locations.

Findings and Conclusions: In general, most bat species preferred to roost in stands that had been partially harvested and contained few midstory trees. The most preferred habitats were partially harvested mixed pine-hardwood stands and stands > 99 years old. Although highly available, second-growth mixed pine-hardwood stands that were 50-99 years old were only preferred by 1 bat species. Forest habitat classes were included more often that topographic or patch-configuration metrics in landscape models. Fine-scale models were a better fit than broad-scale models for most species. Most models included habitat classes of partially harvested or thinned forest that retained mature overstory trees. For all species combined, bats were more likely to roost in close proximity to water sources and in areas that contained less dense or young forest than random locations.