Monitoring biodiversity has traditionally been complex, costly, and requires specialized expertise. However, with the help of artificial intelligence (AI), changes in the landscape can be detected easily and at an early stage using audio recordings. In the Bio-O-Ton project, researchers are developing a method that allows the condition of habitats to be assessed over large areas with the support of citizen scientists. As an early warning system, this method is intended for future use by state authorities. For its development, the project is seeking citizen scientists to make audio recordings in natural areas.
Detecting changes in nature early on is a key prerequisite for identifying habitats worthy of protection and preserving them in the long term. Researchers led by the Karlsruhe Institute of Technology (KIT) are currently developing a fast and cost-effective method, in collaboration with the Institute for Social-Ecological Research (ISOE), that can help make such changes visible across large areas in Germany.
This initiative is based on data from citizen scientists: Interested members of the public can now use a smartphone to make audio recordings in natural areas and submit them to the “Bio-O-Ton” research project. These field recordings from as many parts of nature in Germany as possible will form a kind of early warning system that researchers hope to use in the future to support authorities in monitoring habitats.
The project is supported by a roundtable of specialized agencies, nature conservation associations, surveyors, and scientific institutions. Together, they first developed a classification system for land assessments, which serves as the basis for training an AI at the project partner ci-tec GmbH. Audio recordings were identified as the ideal measurement method for this purpose.
One-minute sound recordings are intended to capture typical sounds of a natural environment, such as animal noises, wind, or even distant traffic sounds. They provide information that can be scientifically and precisely analyzed, can be used over large areas, and enable an initial assessment of the area without entering it. The researchers combine the audio recordings with satellite data, weather information, and other biodiversity data, creating a comprehensive reference database of previously assessed areas.
Building on this, ci-tec develops AI models capable of assigning acoustic patterns to specific habitat types, while ISOE tests the practical applicability of the method in official biodiversity monitoring through an acceptance study and evaluates potential use cases.
The success of the project is closely linked to the support of citizens. The Bio-O-Ton team is therefore collaborating with “Dawn Chorus,” a citizen science project that invites people to record and share nature sounds using their smartphones. The research team calls on everyone who spends time outdoors in nature to use the app to record nature sounds.
In principle, audio data can be collected anywhere. However, areas whose conservation value has already been mapped by experts are particularly important for training the AI, as changes in biodiversity can be easily tracked there. Since these areas are particularly worthy of protection, audio recordings may only be made from exitsing paths. No prior knowledge is required. Detailed instructions for citizen scientists are available here: https://www.ipf.kit.edu/Citizen_Scientists.php
The research project “Bio-O-Ton – Biodiversity Assessment of Biotope Types via Machine Learning Using Citizen Science Audio Recordings and Satellite Images” is funded by the Federal Ministry of Research, Technology, and Space (BMFTR) as part of the call for proposals “BiodivKI – Artificial Intelligence Methods as Tools for Biodiversity Research” of the Research Initiative for the Conservation of Biodiversity (FEdA).
The project Bio-O-Ton develops a new machine-learning approach to the assessment of biodiversity by combining smartphone sound recordings with satellite imagery.
What are the causes of biodiversity loss? How can we protect biodiversity in the future? How can we create a willingness for change?
Go to Topic Page
