Among the most extensive environments on Earth are tropical forests. They have a remarkably diverse range of species and are important to both the global carbon cycle and the climate. However, overexploitation and deforestation are prevalent issues in many tropical forest regions.
Therefore, the importance of reforested regions in the tropics for biodiversity and the climate is growing. It is possible to keep a close eye on the degree of biodiversity development in these places by using automated animal sound analysis. Researchers published their findings in the journal Nature Communications.
Recordings on Former Cocoa Plantations and Pastures
The crew, which was a part of the DFG research group Reassembly, worked on defunct cocoa plantations and pastures in northern Ecuador, where the forest is slowly returning. There, they looked at the possibility of using artificial intelligence (AI) and autonomous sound recorders to autonomously identify the species groups of mammals, birds, and amphibians.
The research results show that the sound data reflect excellently the return of biodiversity in abandoned agricultural areas.
Jörg Müller, Head, Ecological Station Fabrikschleichach, Julius-Maximilians-Universität Würzburg
The study was conducted under the direction of Müller and his colleague Oliver Mitesser.
Overall, communities of vocalizing species mimic recolonization quite effectively–because the communities precisely follow the recovery gradients. A preliminary collection of 70 artificial intelligence bird models was able to characterize the whole species communities of birds, amphibians, and several calling mammals. Even nocturnal insect modifications might be meaningfully connected with them.
AI Models are Being Further Refined
The team is currently focusing on refining and increasing the variety of AI models used. The goal is to be able to automatically record even more species. The concepts will also be used in other protected places in Ecuador, including the Sailershausen JMU Forest and the Bavarian Forest, Germany’s oldest national park.
“Our AI models can be the basis for a very universal tool for monitoring biodiversity in reforested areas,” Müller noted.
The Würzburg professor envisions potential uses, such as certifications or biodiversity credits. Biodiversity credits work in the same way as carbon emissions trading works. They are provided by programs that aim to preserve or promote biodiversity. They are acquired by organizations or companies seeking to offset the harmful effects of their operations.
Sponsors and Participants
The study was carried out within the framework of the German Research Foundation (DFG)-funded research group Reassembly.
In addition to the JMU researchers, ornithologist Dr Martin Schaefer, managing director of the nature conservation foundation Jocotoco, and Professor Nico Blüthgen from the Technical University of Darmstadt - the DFG research group's spokesperson and a JMU alumnus–were also involved.
Professor Zuzana Burivalova of the University of Madison (USA) and Rainforest Connection, a startup that specializes in AI models for tropical bird detection, also contributed.
Müller, J., et al. (2023) Soundscapes and deep learning enable tracking biodiversity recovery in tropical forests. Nature Communications. doi:10.1038/s41467-023-41693-w