Utility of acoustic indices for ecological monitoring in complex sonic environments
| Main Authors: | Samuel RP-J Ross, Nicholas R Friedman, Masashi Yoshimura, Takuma Yoshida, Ian Donohue, Evan P Economo |
|---|---|
| Format: | info dataset eJournal |
| Terbitan: |
, 2020
|
| Subjects: | |
| Online Access: |
https://zenodo.org/record/4081631 |
Daftar Isi:
- Abstract With the continued adoption of passive acoustic monitoring as a tool for rapid and high-resolution ecosystem monitoring, ecologists are increasingly making use of a suite of acoustic indices to summarise the sonic environment. Though these indices are often reported to well represent some aspect of the biology of an ecosystem, the degree to which they are confounded by various extraneous sonic conditions is largely unknown. We conducted an aural inventory across 23 field sites in Okinawa to identify the number of unique animal sounds present in recordings. Using these values of ‘measured richness’, we then examined how the performance of 11 commonly-used acoustic indices varied across a range of sonic conditions (including in the presence and absence of insect stridulations, audible wind or rain, and human-related sounds). Our analysis identified both well- and poor-performing acoustic indices, as well as those that were particularly sensitive to sonic conditions. Only two indices reflected measured richness across the full range of sonic conditions examined. A few indices were relatively insensitive to extraneous sonic conditions, but no index correlated with measured richness when masked by sound from broadband stridulating insects. Our results demonstrate considerable sensitivity of most commonly used acoustic indices to confounding sonic conditions, highlighting the challenges of working with large acoustic datasets collected in the field. We make practical recommendations for acoustic index use based on study design, with the aim of identifying the suite of acoustic indices with greatest utility as indicators for rapid biodiversity monitoring and management of the world’s natural soundscapes. Methods The dataset contains the names of audio files collected across 23 field sites between April 2017 and January 2018 as part of the OKEON-Churamori project on the island of Okinawa, Japan. We conducted an aural inventory, manually counting and recording the number of unique biotic sounds (approximately corresponding to species richness) and noting the presence or absence of three potentially confounding sonic conditions: audible geophony (wind, rain etc.), anthropophony (human-related sounds), and broadband sounds produced by stridulating insect (e.g. cicadas, orthopterans). We then calculated 11 commonly used acoustic indices from the literature and compared their performance (correlation with richness) in the presence vs absence of each sonic condition. Our dataset also contains time and date information for each recording, and the mean site-level richness (i.e. across multiple recordings) for each site and for each unique site-by-season combination. See Table A2 and Methods section in the associated manuscript for details on data processing and the calculation of acoustic indices. Usage notes See readme file for descriptions of data table structure.