Sound recorders can successfully capture the remarkable diversity of vocalizing birds, frogs and mammals – much more effectively than other sampling techniques or classical observational methods. Jörg Müller and his team implemented sound recorders in 43 of our plots, then asked skilled experts to identify hundreds of species from selected snapshots and also employed a artificial intelligence method to quantify the diversity from these recordings. Both methods – expert-based and AI – performed very well in describing the recovery of species composition along the chronosequence. Such techniques, including a metabarcode analysis of light trap captures included in the study, represent very effective and powerful rapid assessments of biodiversity changes and thus help tracking the targets of ecosystem restoration. This pioneering work was now published in Nature Communications:
MEDIA COVERAGE:
Nature: How AI can help to save endangered species
The Economist: AI can catalogue a forest’s inhabitants simply by listening
MONGABAY: Sound recordings and AI tell us if forests are recovering, new study from Ecuador shows
Bloomberg: To Track a Forest’s Recovery, Artificial Intelligence Just Listens
Jocotoco: Artificial intelligence models capture the biodiversity of the Chocó
Nature Podcast: Sounds of recovery: AI helps monitor wildlife during forest restoration
Detschlandfunk: Tropische Artenvielfalt mit Hilfe von KI bestimmen: Interview Prof. Jörg Müller
Science Daily: AI models identify biodiversity from animal sounds in tropical rainforests
Uni Würzburg: AI Models Identify Biodiversity in Tropical Rainforests
Part of the team on a long trail to the plots. The species composition – and thus sounds by vocalizing animals – in secondary forests in the foreground are predictably different from the old-growth forest in the background. But species composition grows back with forest age.