Acoustic identification and activity of bats in red-pine stands in the Manistee National Forest

I used ultrasonic detectors to record echolocation calls of eight species of bats and construct a classification model using discriminant function analysis (DFA), based on 10 call parameters, for use in the Manistee National Forest. Cross validation of the model indicated accuracies that ranged from 11% to 100%, depending on species, with an overall accuracy of 74%. In addition to calls collected from bats flying in open habitats, I recorded calls from bats flying in cluttered situations and added them to the DFA. Species identification decreased to 65%. Detectors were also used to determine species presence in red-pine stands and adjacent openings and the effects of thinning red-pine plantations on bat activity in the Manistee National Forest. Bat activity was lower in stands than in openings, and activity was associated with greater insect biomass and decreased structural complexity in the openings. Thinning had no effect on bat activity.