Cambridge AI and satellite data enable rapid forest species mapping

A team at the University of Cambridge has demonstrated a straightforward method for mapping forest composition by combining artificial intelligence with satellite imagery. The study, published in *Science of Remote Sensing*, shows how readily available data can be leveraged to identify tree species across large landscapes.

Using embeddings generated by Tessera, a geospatial foundation model, the researchers processed high‑resolution satellite images of the Trentino region in the Italian Alps. Tessera’s embeddings translate raw imagery into a form that the model can interpret, enabling accurate classification of individual tree species across the study area. The approach proved both scalable and efficient, requiring only standard satellite datasets and open‑source AI tools.

The findings suggest that similar techniques could be applied worldwide to monitor forest biodiversity, assess ecosystem health, and support conservation planning. By simplifying the mapping process, the Cambridge team has provided a practical tool that could be adopted by environmental agencies and researchers with limited resources.