Identifying Great Gray Owls Using Territorial Calls
Our research team has been deploying autonomous recording units (ARUs) since 2016 to detect Great Gray Owls across breeding territories in the Jackson Hole Valley. Recently we began looking at the potential to use territorial calls of Great Gray Owls to identify unique individuals. We used data from 26 ARUs deployed from 2019-2022 across 14 territories representing 14 unique individuals. We compared two methods for identifying individuals using territorial call characteristics; one method involved hand measuring call characteristics from spectrograms of the call, and the second was an automated method using mel-frequency cepstral coefficients (MFCCs).
Mel-frequency cepstral coefficients measure the amplitude of a call across multiple points in time and were first introduced in their use for human voice recognition. We found that by using MFCC features we could correctly identify Great Gray Owl individuals using territorial calls 97% of the time as compared to 78% with hand measuring call characteristics from spectrograms. We plan to continue to incorporate territorial calls from additional territories and years to identify more unique individuals using this method. Ultimately these results can be used to gain a better understanding of the movements of Great Gray Owls between territories within and across years, and provide information on population dynamics.