Date of Award

January 2019

Degree Type

Open Access Thesis

Degree Name

Master of Science (MS)

Department

Biological Sciences

First Advisor

Luke E. Dodd

Second Advisor

Stephen C. Richter

Third Advisor

Kelly Watson

Abstract

Bat populations of eastern North America continue to decline due to the cumulative effects of White-nose Syndrome, habitat loss, and anthropogenic disturbances across the landscape. Unique stressors exist on military installations, such as noise created during training activities. Given the scarcity of data that exists for these widespread ownerships, I created predictive models for nightly bat activity related to local habitat, landscape, and military use parameters. Bat activity was assessed during the summers of 2016 and 2017 using full-spectrum acoustic detectors across the Wendell H. Ford Regional Training Center (WHFRTC), a ca. 4200 ha military landscape in the Interior Rivers and Valleys ecoregion of Muhlenberg County in western Kentucky. Local habitat data was collected on site, and landscape-level data was extracted using geographic information systems. Ongoing military activities as reported by the Kentucky Army National Guard were summarized on the temporal and spatial bases that I hypothesized might impact bat activity. I then used regression techniques in combination with Akaike’s Information Criterion (AIC) to identify the most parsimonious model for predicting bat activity across the landscape. Distinct habitat models were developed to explain bat activity based on: landscape-level habitat characteristics, site-level habitat characteristics, and both temporal and spatial disturbance models resulting from military training. The principal response variable considered was the total number of bat passes observed per detector-night. This variable was further divided into calls identified by Kaleidoscope Pro as Indiana bats (Myotis sodalis) and northern long-eared bats (Myotis septentrionalis). In total 13 unique bat species were identified across the WHFRTC property, including two focal threatened and endangered species, northern long-eared bats and Indiana bats. Predictive models from each model suite identified parameters that influence bat activity within the three focal categories. The minimum distance to a firing range, firing activity in relation to sunset time, tree community composition, and distances to various natural and modified landscape types were identified as important predictors for bat activity. When planning future training activities on the landscape, precautions can be taken to minimize the potential detrimental impacts on foraging and commuting bats. Additionally, habitat management, such as planting native vegetation and removing excess snag trees can promote the foraging potential of the area. Overall, bat species conservation goals should be clearly defined by the managing agency in order to effectively and efficiently protect species of concern and the preferred habitat.

Beyond my assessment of bat activity across the military landscape of WHFRTC, I conducted an additional study to identify if bats would alter within- or across-night behaviors in response to an auditory predation cue in the form of broadcast owl calls. Bat activity was assessed using passive acoustic monitoring at WHFRTC from June to August of 2017. Bat detectors were paired with waterproof speakers that broadcast one of three possible treatments every ten minutes throughout nights. Treatments included predation cues (owl calls), noise (nocturnal frogs and insects), and a control (silence). On a given sampling night, six detectors were deployed with a randomly selected auditory treatment so that all treatments were represented. Approximately 9,000 bat passes were recorded during 990 detector-hours of sampling. Total bat activity was not altered by predation cues or noise (p > 0.05). For the most commonly recorded species, big brown bat (Eptesicus fuscus), eastern red bat (Lasiurus borealis), and tri-colored bat (Perimyotis subflavus), activity did not change in response to auditory treatment between nights or within nights (p > 0.05). However, subtle differences in hourly accumulation trends were realized across species. While big brown bat and tri-colored bat activity patterns suggest unimodal trends, eastern red bat activity was more consistent throughout the night. The results of this study suggest that bats do not respond to the auditory predation cues of a nocturnal avian predator; the robust sampling framework and effort presented here provides a benchmark for future auditory investigations of predator avoidance by bats.

Share

COinS