University of León researchers develop AI to predict crown fires

Researchers at the University of León have developed an artificial intelligence tool named CROWNFIRE-AI designed to predict the risk of crown fires up to 16 days in advance. Crown fires, which occur when flames reach the upper canopy of trees, are considered among the most dangerous types of wildfires due to their high intensity and rapid spread, often moving two to four times faster than surface fires.

The platform integrates satellite imagery from the European Space Agency’s Sentinel-1 and Sentinel-2, NASA’s MODIS data, meteorological forecasts, and topographical information to generate regional risk maps. This technology aims to assist forest managers and emergency services in identifying high-risk areas to better allocate prevention resources.

The development, led by researcher José Manuel Fernández Guisuraga in collaboration with the University of Valladolid and the company Vexiza SL, was trained using a database of 18,874 reference points from 33 major fires in Spain recorded between 2017 and 2024. The study, published in the journal Ecological Informatics, notes that such extreme fire events are increasing globally due to climate change, prolonged droughts, and heatwaves.