Acoustic Differentiation of Horseshoe Bats in Luzon Island, Philippines: Developing a Call Identification Key and Uncovering Potential Cryptic Species

Acoustic techniques have been frequently applied in field surveys as they offer a rapid and non-invasive approach in studying bats. More importantly, bioacoustics has been invaluable in identifying rare and morphologically cryptic bat species and providing additional data for resolving the taxonomy of species complexes. In the Philippines, however, automated call identification is still hindered by the lack of a robust call library, inter- and intraspecific variability in echolocation calls of many species, and the substantial time needed for manual acoustic data analysis. In this study, we developed a call identification key using over 34,000 reference calls recorded from eight Rhinolophus sonotypes in Luzon. Several call parameters were extracted and utilized to construct classification trees for species differentiation. Rules generated from the trees were used to construct the call identification key. Testing and subsequent refinement of the key resulted in overall 93.0% and 97.5% correct classification rates when tested against an independent set of reference call sequences and set of unknown calls recorded from free-flying bats, respectively. Assessment of morphometrics and acoustic parameters also revealed the presence of distinct small and large size morphs within R. macrotis and R. philippinensis, each echolocating at distinct frequencies. Our results emphasize the reliability of the call identification key for automated field identification of morphologically similar bat species and detection of potential novel cryptic species.