EO4Biodiversity: Applying GeoAI to improve habitat mapping for biodiversity net gain

Biodiversity is essential to ecosystem health, supporting clean water, food security, medicine, and resilience to climate change. Reducing biodiversity loss and enhancing habitats through mechanisms such as biodiversity net gain (BNG) is becoming increasingly important. Accurate habitat mapping is central to these efforts, yet existing landcover and habitat datasets often lack the classification detail and spatial accuracy required to support robust BNG assessments.
EO4Biodiversity combines Earth Observation (EO) data with existing datasets and GeoAI techniques to refine habitat classification for Standardised Biodiversity Unit (SBU) calculations required for environmental planning in the UK. The project uses satellite imagery, elevation models, soil data, vector data and ground data alongside machine learning to improve the identification of habitat types.

As the project is in its early stages, we will outline the approach taken to integrate and evaluate these data sources, share initial insights into classification performance, and discuss the challenges and opportunities of applying EO and GeoAI to support biodiversity-sensitive development planning.