How high is high? Using percentile thresholds to identify peak bat activity

Abstract

Passive acoustic monitoring is a common tool used for monitoring bat activity levels. Identifying periods and locations of peak levels provides insight into bat ecology and has important management implications. One limitation of passive acoustic monitoring is the relative nature of the data, often relying on subjective interpretation of descriptive terminology such as “higher” or “lower”. We propose the use of percentile thresholds (PTs) for objectively identifying peak activity. By compiling a reference data set, it is possible to define percentiles of the observed activity levels and these percentiles can provide objective thresholds for comparing activity levels. We used acoustic recordings from sites in Canada and calculated PTs based on the distribution of the number of calls per hour across all nights and sites for three species of bat. Given species ecologies (e.g., hibernating, migrating), we were able to use PTs to objectively identify peak activity levels on a species-specific basis. Percentile thresholds are also a replicable method of describing within-night activity by evaluating species-specific activity patterns and important times of night. Our analyses and examples represent a proof of concept. The next step is to move towards a standardized distribution to generate PTs. Creating a public repository of acoustic data sets to evaluate activity of a species in the context of its entire range would allow us to standardize terms such as “high” activity in an objective manner.