A comparison of ultrasonic detectors and radiotelemetry for studying bat–habitat relationships

Abstract

AbstractUltrasonic bat detectors and radiotelemetry are 2 methods used to examine habitat selection for bats, but no one has empirically examined if conclusions drawn from these 2 methods are comparable and, although previous work has presented concerns regarding the basic assumptions of inferring habitat selection with acoustic surveys, these assumptions have not been tested. Therefore, during summer 2008, we examined use of bat detectors and radiotelemetry to infer habitat selection among 4 forest habitat types in southwestern Georgia, USA by simultaneously radiotracking evening bats (Nycticeius humeralis; n = 14) and using stratified sampling to acoustically monitor bat activity (n = 95 sites). We used compositional analysis for radiotelemetry data and generalized linear models for acoustic data to assess habitat selection at Johnson's (1980) 2nd and 3rd order spatial scales. At the 3rd order scale, both acoustic sampling and radiotelemetry indicated preference of mature pine and open stands over hardwood stands and pine plantations. Radiotelemetry revealed hardwood stands were preferred at the 2nd order scale, which was not reflected in acoustic data. Qualitatively, conclusions about habitat selection are not consistent between these 2 techniques. We conclude that bat detector surveys may be appropriate for examining habitat selection at the stand‐level, although more work is needed to confirm this, but have limited or no inferential ability at larger scales. Our results support the general inability of acoustic surveys to infer habitat selection due to lack of adherence to underlying assumptions. Therefore, we recommend using radiotelemetry for habitat selection studies and that researchers limit habitat selection inferences derived from acoustic surveys. © 2011 The Wildlife Society.