Acoustic Surveys for Bats are Improved by Taking Habitat Type into Account

Abstract

Passive monitoring of bat species via acoustics is a growing field and as a result there are various software programs available that allow for species identification. However, accuracy of these programs is variable and creating a local call library is essential when trying to identify acoustically similar species. In Dinosaur Provincial Park, Alberta, Canada, 3 Myotis species are difficult to distinguish acoustically. We created an echolocation call library from known Myotis evotis, M. lucifugus, and M. ciliolabrum flying in open spaces and near clutter to test whether or not recording in the bats' local habitat type improved call library quality and identification accuracy. Bat calls recorded within open spaces (coulees) differed from those recorded near cluttered spaces (tree edges) for M. ciliolabrum and M. lucifugus. Accuracy of species identification also increased when we used models based on bat calls recorded in cluttered habitats. Using a simple model with recordings from different habitat types within a study site, we were able to improve identification accuracy and model performance. © 2019 The Wildlife Society.

Effects of Environmental Clutter on Synthesized Chiropteran Echolocation Signals in an Anechoic Chamber

“…Although our results would suggest that clutter effects are not as apparent as receiver angle and proximity to the source signal, research does indicate that forest structural characteristics do effect echolocation. Structural habitat metrics of vegetation have been linked to bat detection, altering call parameters, and correct identification [6,8,11,20,39]. Increased canopy cover and mature forest habitat as compared to more open areas have shown a negative correlation with bat call identification [20,39].…”

“…Ultrasonic acoustic lures also are an emerging technology that shows promise to attract bats during mist-net surveys in part to overcome decreases in catch-per-unit effort from WNS [2][3][4][5]. The degree of microhabitat complexity, or "clutter", which typically consists of the boles of trees and attached foliage in the recording or broadcasting environment, presents a challenge for both technologies and represents the most important ecological constraint for foraging bats [6][7][8]. Clutter can absorb or reflect ultrasonic signals, alter recorded parameters, create echoes, or block all or a portion of the signal [9][10][11][12][13].…”

“…They found that there was a high degree of interspecific and intraspecific variation in call parameters between the three clutter conditions, and that using a clutter-derived call library to train the neural network improved species classification accuracy when provided calls also recorded in cluttered conditions. Similarly, Findlay and Barclay [8] found that using calls recorded in more cluttered conditions improved identification accuracy for Myotis species as many are adapted to foraging in high-clutter forest habitats [20]. Avoiding forest habitats may significantly bias acoustic surveys by underrepresenting clutter-adapted bat species, such as the federally threatened northern long-eared bat and the federally endangered Indiana bat (Myotis sodalis), thereby misrepresenting habitat associations [21,22].…”

The Weekend Effect on Urban Bat Activity Suggests Fine Scale Human-Induced Bat Movements

“…Since the independent variable was the day of the week, we first excluded weeks of incomplete data (bat detectors malfunctioned, no weather data, or weeks overlaid with two seasons). It is important to note that the recorded bat activity naturally varied among sites because of differences in obstacles within the recording acoustic space such as canopy cover or sound transmission interference [ 55 , 56 ]. Therefore, we treated each site separately in the statistical analyses.…”

Testing Accuracy and Agreement among Multiple Versions of Automated Bat Call Classification Software

“…Acoustic monitoring is less invasive than capturing and handling bats and reduces the risk of disease transmission, but leaves greater uncertainty in species identification compared to having bats in the hand. Limitations on the accuracy of bat echolocation call identification include call variation within species, overlapping call characteristics among species, recording quality, environmental conditions, and habitat type, among other factors (Barclay 1999, Frick 2013, Goerlitz 2018, Russo et al 2018, Findlay and Barclay 2020). Inaccurate call identification could lead to inappropriate management decisions, for example if a federally listed species is thought to be absent but is present, or vice versa.…”