| Although most bats in the southeastern United States depend on forests for roosting and foraging, we know little about the ecological requirements of bats that live in this region. The objective of this study was to use radio telemetry, acoustic sampling, Akaike's information theoretic procedures, occupancy modeling, and discriminant function analyses to: (1) examine multi-scale roost-site selection for three forest bat species [eastern pipistrelles (Perimyotis subflavus), eastern red bats (Lasiurus borealis), and northern long-eared bats (Myotis septentrionalis)], (2) test the effects of timber harvest on bat foraging ecology in riparian areas, and (3) compare and relate methods of assessing vegetative clutter to the probability of detecting bats. We conducted our study from 2004--2007 in a dense deciduous forest undergoing low-intensity timber management in the southern Appalachian Mountains of western North Carolina, USA. We radiotracked eight red bats to 19 roosts, seven pipistrelles to 15 roosts, and 16 male and 18 female northern long-eared bats to 50 and 52 roosts, respectively. We recorded 48,456 bat passes in riparian areas during 8,309 hours on 832 detector-nights and assessed bat detection probabilities and vegetative clutter at 71 points. Macrohabitat factors were important to male red bats and pipistrelles whereas female northern long-eared bats displayed mainly microhabitat roost-site preferences. Our results indicated that maintaining a diversity of age classes should provide roosting habitat for pipistrelles, red bats, and northern long-eared bats. Leaving large diameter trees and snags of preferred genera (Quercus, Robinia, Carya) during harvests should ensure a continuous supply of suitable roost structures for reproductive female northern long-eared bats. Pipistrelles and female northern long-eared bats may also benefit from retention of mature stands near streams. Riparian areas near small streams in our study area served as foraging habitat for .4 bat species and forested buffers affect the foraging activity of bats in riparian areas following timber harvest in adjacent forests. Quantitative measurements of individual variables (specifically midstory live stem count and canopy crown volume) were the most effective measures of clutter relative the other methods we tested because they were good predictors of bat detection and were most effective in discriminating among survey points of different ages and forest types. In future studies of bat foraging habitat, quantitative measures should be used to assess clutter to facilitate comparisons among habitats or studies. |
