Integrated wildfire susceptibility and hazard mapping: A data-driven approach for Mediterranean regions using the Garrotxa (Catalonia, Spain) as a case study

Wildfires represent a growing threat to Mediterranean landscapes, where socio-ecological factors, land use dynamics and climate variability converge to increase fire risk. This study presents a data-driven approach to wildfire susceptibility and hazard mapping in the Garrotxa region (Catalonia, Spain), identifying the most influential drivers and generating susceptibility maps for present and future conditions. Because georeferenced burned area records exist only for a fraction of the 265 documented ignition events (1900–2023) in the Garrotxa, fire perimeters were reconstructed by simulating spread from each ignition point and its reported size. Validation against fires with real perimeters yielded a mean overlap of 0.71, confirming the reliability of the simulation. These simulated burned areas were integrated with fifteen environmental conditioning factors, whose influence on wildfire occurrence was quantified using the Frequency Ratio method to produce interpretable susceptibility scores. The most influential drivers were wind exposure, land cover, temperature and land surface temperature, while the topographic wetness index and slope exerted weaker effects, consistent with findings from other Mediterranean regions. The resulting map achieved a ROC AUC of 0.73, comparable to a Random Forest benchmark, and 79% of past wildfires fell within the two highest susceptibility classes. Finally, by replacing the climatic conditioning factors with their projected future values, the framework generated wildfire susceptibility maps for the near-, mid- and far-future horizons, showing how fire-prone areas may evolve under climate change. This transferable and interpretable approach reproduces observed fire patterns and supports risk mitigation and climate-adaptive land management in Mediterranean fire-prone areas.