Factors influencing site occupancy of bats on managed-pine landscapes in the Coastal Plain of the southeastern United States

Abstract

Bats are ecologically important inhabitants of forest landscapes in the southeastern United States. These forests account for the majority of regional land cover, much of which is actively managed for forest products. Although many owners of managed-pine forests are committed to preserving biodiversity, responsible management is hampered by a paucity of data concerning bat ecology in these landscapes. To address the lack of data and improve management, my objectives were to evaluate an acoustic identification method and elucidate factors that influence foraging bats in Coastal Plain managed-pine forests. I used cross-validation to assess performance of K-nearest neighbor analysis as a method to classify unknown bat calls. Overall accuracy of K-nearest neighbor was 83% and species-specific accuracy ranged from 51% to 96%. Results suggest that K-nearest neighbor should be considered a viable call identification method. I used Anabat II detectors to survey and determine presence of multiple bat species across managed-pine forest landscapes. I used program PRESENCE to evaluate plausibility of occupancy models based on Akaike’s Information Criterion and estimate influence of factors on detection and occupancy. Plausibility of detection and occupancy factors was supported over null models. Detection probabilities of resident bats decreased with increasing amounts of vegetation at sample points. Results suggest that investigations concerning the influence of factors on bat occupancy should account for variable detection. Increasing insect abundances increased the probability of occupancy by bats, but the most plausible measure of insect abundance was species-specific. Small-scale vegetation abundance was negatively related to occupancy. Management activities in managed-pine forests that reduce small-scale vegetation clutter and increase insect abundance are likely to benefit resident bat species. Plausibility of landscape-scale factors influencing probability of occupancy was supported, but plausible factors were species-specific. Landscape-scale factors with evidence supporting their potential influence included un-thinned stands of intermediate age, distance to water sources, road density, patch richness, amount of edge, and percentage of stands older than 30 years. Bats respond to landscape conditions, but managing a single landscape attribute likely will not benefit all members of the bat community. However, maintaining a diverse landscape apparently provides habitat for a diverse bat community.