Soundscape Characterization Using Autoencoders and Unsupervised Learning.

Passive acoustic monitoring (PAM) through acoustic recorder units (ARUs) shows promise in detecting early landscape changes linked to functional and structural patterns, including species richness, acoustic diversity, community interactions, and human-induced threats. However, current approaches primarily rely on supervised methods, which require prior knowledge of collected datasets. This reliance poses challenges due to the large volumes of ARU data. In this work, we propose a non-supervised framework using autoencoders to extract soundscape features. We applied this framework to a dataset from Colombian landscapes captured by 31 audiomoth recorders. Our method generates clusters based on autoencoder features and represents cluster information with prototype spectrograms using centroid features and the decoder part of the neural network. Our analysis provides valuable insights into the distribution and temporal patterns of various sound compositions within the study area. By utilizing autoencoders, we identify significant soundscape patterns characterized by recurring and intense sound types across multiple frequency ranges. This comprehensive understanding of the study area's soundscape allows us to pinpoint crucial sound sources and gain deeper insights into its acoustic environment. Our results encourage further exploration of unsupervised algorithms in soundscape analysis as a promising alternative path for understanding and monitoring environmental changes.

Getting better temporal and spatial ecology data for threatened species: using lightweight GPS devices for small primate monitoring in the northern Andes of Colombia.

The use of GPS telemetry has been a reliable research tool for the study of primate biology in recent years. Although in the past technological restrictions limited its use mainly to large primates, recent improvements in battery size make it possible to use this technology for small species. We used GPS devices for monitoring two adult white-footed tamarins (Saguinus leucopus) from a free-ranging group, and assessed its applicability for recording spatial and ecological data. GPS devices were operational for 66 and 85 days, recording 221 positions (36.6% acquisition rate; 73% of which were highly accurate) and 3195 activity values for both individuals. Depending on the estimation method, we calculated the home range size for the group to be 19.4 and 22.9 ha, which were within the range for the species in other localities. The animals were active each day for 11 h, with high activity during the early morning. The monkeys showed a constant and alternate use of four sleeping sites with a limited reuse of the same site on consecutive nights. These daily activity and sleeping site use patterns are similar to those reported for other Saguinus species. Based on the kind and quality of the data recorded, we consider GPS telemetry to be an efficient and advantageous method for tracking and obtaining ecological information from S. leucopus. In comparison to other data collection methods, GPS telemetry required fewer personnel and less time commitment to record data without compromising the accuracy of the spatial and activity information we obtained.