Biodiversity assessment of biotope types through machine learning based on citizen science sound recordings and satellite images

The extent of biodiversity loss can currently only be detected using complex, mostly manual methods. The Bio-O-Ton project aims to develop a new approach to recording the biodiversity of entire landscape units. Using meadow biotopes in Germany as an example, the project aims to show how the biodiversity of meadows can be classified and changes detected using machine learning (ML) methods with the aid of sound recordings (sound-scape-ecology methodology), satellite data and citizen science data. The results are published openly in interactive thematic maps and can thus be used directly by stakeholders (forestry and agriculture, environmental associations, nature conservation authorities and citizens). As a result, the conventional mapping approaches of habitats worthy of protection can be supplemented and authorities can be supported in a well-founded, cost-efficient monitoring of protected areas.

In the first phase, the project pursues the in-depth development of a concept and the preparation of an application for a second phase, which will focus on the implementation of the aforementioned objectives. In addition, initial scientific work will be carried out in the project to enable an effective and rapid entry into a possible second phase.

Project lead: Dr. rer. nat. Susanne Benz

Karlsruhe Institute of Technology