Using Soundscapes to Assess Changes in Coral Reef Social-Ecological Systems

Abstract

Coral reefs are among the world’s most diversified marine ecosystems. While their rich biodiversity is essential to many social and economic activities, they face unprecedented cumulative impacts from climate change and anthropogenic development. Aligning governance structures with ecological processes has become key to sustaining ecosystem services. Nevertheless, a major obstacle to informing decision-making is the lack of empirical data on interacting ecological and social processes. Therefore, novel methods for facilitating the assessment of coral reef social-ecological systems are necessary.

Unlocking the soundscape of coral reefs with artificial intelligence: pretrained networks and unsupervised learning win out

“…Statistical comparisons of individual acoustic indices represent the predominant method for automated analysis of reef soundscape data indices (17,29,30). To assess the performance of individual indices on our data, we deliberately selected the individual acoustic indices that reported the most significant differences between each habitat class.…”

“…Machine learning (ML) presents a powerful alternative for automating the analysis of bioacoustic data across multiple ecosystems and taxa, so far primarily focused on generating species detectors (18)(19)(20). However, applications of ML for soundscape ecology, rather than individual bioacoustic signals, are much less developed for most ecosystems and are almost entirely untested on coral reefs (21)(22)(23). The optimum ML approaches to employ and insights these approaches can generate remain unknown.…”

“…1a). Finally, we benchmarked our ML approaches against the predominant method for automated reef soundscape analysis of recent years: individual acoustic indices (17,29,30). To ensure that our results are representative of reef soundscapes broadly, we used recordings from three biogeographically distinct locations: Indonesia, Australia and French Polynesia (Fig.…”

Worldwide soundscape ecology patterns across realms

“…PAM can help to address fundamental questions in ecology (Ross et al 2023). We could investigate the general relationship between anthropophony and biophony (Lin et al 2023), but also specifically monitor impactful human activities (e.g., illegal hunting, bottom trawling, habitat restoration) (Sethi et al 2020) to model human impact on biodiversity and ecosystem services, and to support law enforcement and conservation planning. One aspect of this relationship is pervasive noise pollution effects on wildlife (Slabbekoorn et al 2018), that have been extensively studied in the marine and terrestrial realms (Boyd et al 2015;Buxton et al 2017;Cox et al 2018;Francis, Ortega, and Cruz 2009;Jerem and Mathews 2021;Solé et al 2023) and increasingly so for freshwater (Mickle and Higgs 2018).…”