AI predicts dangerous crown fires up to 16 days ahead

Researchers at the University of León (ULe) have developed CROWNFIRE-AI, an artificial intelligence platform that predicts the risk of crown fires—the most dangerous type of wildfire—up to 16 days in advance. Crown fires spread through tree canopies, advancing two to four times faster than surface fires and are extremely difficult to extinguish Source: infobae. The tool, published in Ecological Informatics, was trained on data from 33 large fires in Spain (2017–2024) and 18,874 reference points, integrating satellite imagery (Sentinel-1/2, MODIS), topography, forest inventories, and weather forecasts Source: infobae. It generates regional risk maps accessible via a free, user-friendly web platform hosted by ULe, requiring no programming skills Source: infobae. The methodology could be adapted to other fire-prone regions after calibration Source: infobae.

Broader context

Experts emphasize that AI is transforming wildfire management, but prevention remains underfunded. José Vicente Oliver, a forestry engineering professor at the Polytechnic University of Valencia, notes that while Spain has cutting-edge extinction technology, "if the majority of resources are destined to extinguishing fires and not to preventing them from occurring, something is being done wrong" Source: abces. AI models can create dynamic risk maps using weather, fuel moisture, topography, and fire history, aiding preventive actions like prescribed burns and grazing Source: abces. However, Oliver cautions that for early detection, "the 112 emergency number remains the best sensor we have" in much of Spain Source: abces.

Key points

• CROWNFIRE-AI predicts crown fire risk with up to 16 days lead time Source: infobae.
• The platform is free and accessible after registration Source: infobae.
• AI supports scenario analysis and decision-making, but prevention needs more investment Source: abces.