A new study in Nature Communications shows that soundscape recordings and artificial intelligence can effectively monitor biodiversity recovery in regenerating tropical forests.
Tropical forests play a key role in addressing climate change and biodiversity loss. Large-scale restoration efforts are expanding globally to counter deforestation.
However, assessing the pace and success of biodiversity return in regenerating forests remains challenging and controversial. Quantifying biodiversity gains is vital for sustainable management and funding of restoration projects. However, comprehensive biodiversity surveys are expensive, complex, and require taxonomic expertise.
In this research, an international team tested if bioacoustics and AI could overcome these hurdles and track biodiversity recovery across 43 plots along a forest regeneration gradient in Ecuador. The plots spanned active pastures and cacao farms, abandoned agricultural land regrowing for 1-34 years, and old-growth forests.
The study utilized three approaches: expert identification of vocalizing vertebrates
The team utilized three approaches: expert identification of vocalizing vertebrates from recordings, analysis of recordings with acoustic indices, and an AI model to identify bird species. They also sampled insect diversity through DNA metabarcoding.
The results showed the vertebrate community composition shifted along the gradient, with old-growth forests distinct. Automated measures strongly correlated with the gradient: acoustic indices explained 62% of the variation in the vertebrate community, while the AI-derived bird community explained 69%. Both also significantly correlated with insect composition, reflecting broader biodiversity recovery.
According to the lead author, soundscape analysis is a powerful tool to monitor faunal community recovery across agricultural abandonment to old-growth forests robustly and reproducibly. The study highlights the promise of automated monitoring, even with limited regional AI training data, to cost-effectively assess restoration outcomes at scale.
A co-author added that well-documented soundscape data enables retrospective re-analysis as algorithms improve. Implementing such monitoring can prevent greenwashing and empower sustainable funding for quantified biodiversity gains.
Overall, the research shows bioacoustics and AI provide an effective approach to benchmarking and tracking biodiversity recovery in regenerating tropical forests worldwide. Automated acoustic monitoring could strongly support restoration efforts and outcomes for climate change mitigation and conservation.
Featured Image Credit: Photo by Nandhu Kumar; Pexels; Thank you!