Long-term monitoring of soundscapes and deciphering a usable index: Examples of fish choruses from Australia

Abstract

Similar to geophysical and anthropogenic noise, biological contributions to soundscapes vary considerably in frequency, time, and intensity. Fish choruses are a perfect example, contributing significantly to marine biological noise and are used here as an analogue for variations in soundscapes. Their species-characteristic signals vary thus, so do their choruses, which can raise ambient noise levels by up to tens of decibels, for prolonged periods. Multi-species choruses can occur, with varying degrees of temporal and frequency partitioning, or none at all. Australian datasets of underwater noise have been acquired for nearly two decades and multiple fish calling patterns have been detected. Detecting, delineating, and understanding these patterns is non-trivial and a metric relating their contribution to the soundscape with biodiversity or habitat would be an invaluable tool. In recent years, several acoustic indices have been derived, proving useful in the terrestrial domain. Investigation of their application in marine environmental studies has also begun. However, such a plethora of widely varying sources and changing patterns can affect acoustic indices. This paper describes a simple and automatic suite of tools to help identify signals of a wide range of time patterns which are potentially underrepresented or missed by acoustic complexity or biodiversity indices.