Automated detection of low-frequency rumbles of forest elephants: A critical tool for their conservation.

Abstract

African forest elephants (Loxodonta cyclotis) occupy large ranges in dense tropical forests and often use far-reaching vocal signals to coordinate social behavior. Elephant populations in Central Africa are in crisis, having declined by more than 60% in the last decade. Methods currently used to monitor these populations are expensive and time-intensive, though acoustic monitoring technology may offer an effective alternative if signals of interest can be efficiently extracted from the sound stream. This paper proposes an automated elephant call detection algorithm that was tested on nearly 4000 h of field recordings collected from five forest clearings in Central Africa, including sites both inside protected areas and in logging concessions. Recordings were obtained in different seasons, years, and under diverse weather conditions. The detector achieved an 83.2% true positive rate when the false positive rate is 5.5% (approximately 20 false positives per hour). These results suggest that this algorithm can enable analysis of long-term recording datasets or facilitate near-real-time monitoring of elephants in a wide range of settings and conditions.