A Real-Time Streaming and Detection System for Bio-Acoustic Ecological Studies After the Fukushima Accident

Acoustic ecologydata have been used for a broad range of soundscapeinvestigations. Counting sounds in a given soundscapeis considered an effective method in ecology studiesthat provides comparative data for evaluating the impact of human communityon the environment. In 2016, Kobayashi and Kudo collected a particularly valuable dataset containing recordings from within the exclusion (i.e., difficult-to-return-to) zone located 10 km from the Fukushima Daiichi Nuclear Power Plant in the Omaru District ( Namie, Fukushima, Japan). These audio samples were continuously transmitted as a live streamof sound data from an unmanned remote sensing station in the area. In 2016, the first portion of their collected audio samples covering the transmitted sound recordings from the station was made available. Such data cover the bioacousticsin the area. This paper describes the methodologies by which we processed these recordings, in extreme conditions, as preliminary eco-acoustic indexes for demonstrating possible correlations between biodiversity variation and preexisting radioecologyobservations. The variations in some of these vocalizations were also studied.