Evaluating Methods for Emergence Counts at Bat Roosts: A Pilot Study Comparing Drone-acquired Thermal Imagery, Acoustic Estimates, and Visual Observations

Karah Jaffe^{1, 2, *}, Richard Carter², and Aaron J. Corcoran³

¹U.S. Fish and Wildlife Service, Kentucky Field Office, Frankfort, KY 40601. ²Department of Biological Sciences, East Tennessee State University, Johnson City, TN 37614. ³Department of Biology, University of Colorado, Colorado Springs, CO 80918.^*Corresponding author.

Journal of North American Bat Research, Special Issue 1 (2024): 30–45

Abstract
Evaluating methods for monitoring wildlife populations is essential for conservation. Here, we conduct an initial assessment of the precision of 3 methods deployed concurrently for surveying bats emerging from roosts: human visual observation, drone-acquired thermal imagery, and soundpressure levels of ultrasonic recordings. We conducted 3 surveys at a roost of Myotis grisescens (Gray Bats) and 10 surveys at a roost of Eptesicus fuscus (Big Brown Bats). Fitting linear and exponential models to the data resulted in strong correlations between visual counts and both drone-acquired (r2 = 0.83–0.91) and acoustic (0.51–0.90) estimates. Nightly counts based on thermal imaging differed from visual counts by a median of 17.5% and 32% at our 2 study sites (range: 1.6–263%), while nightly acoustic estimates differed from visual counts by a median of 4.7% and 15.1% (3.5–144%). We discuss the limitations and advantages of each method and conclude that all can be useful for surveying bats, but each has its own inherent constraints. When possible, we recommend using multiple methods concurrently to increase the reliability of count estimates.

Download Full-text pdf (Open access)