Modelling wildfire occurrence at regional scale from land use/cover and climate change scenarios

ABSTRACT: Wildfire occurrence is expected to increase in future climate and Land Use Land Cover (LULC) change scenarios, especially in vulnerable areas as the European Mediterranean Basin. In this study future probability of wildfire occurrence was estimated for a 20-year time period (2041?2060, centered on 2050) by applying a statistically-based regression model using LULC-derived contact areas with the forest cover (interfaces) as proxy for the human-related factor and a combination of Live Fuel Moisture Content and seasonal climate-related variables as predictors. Future wildfire occurrence was mapped under RCP 8.5 high emissions scenario in four Spanish regions with heterogeneous socioeconomic, LULC and natural fire-related characteristics at 1 km2 target spatial resolution. Results showed increased wildfire probability in ?19?73% of 1 km2 cells, observing regional differences in the variable effects. This approach could be applied to other spatial scales offering tools for planning and management actions and to obtain different possible future scenarios.This paper was funded by the LUC4FIRE project (CSO2015-73407-JIN), supported by the Spanish Ministry of Economy (MINECO) and the Environmental Remote Sensing and Spectroscopy Laboratory (Speclab) at the Spanish National Research Council (CSIC). We want to thank three anonymous rewievers for useful comments to improve the manuscript. We also acknowledge for the provision of fire data the General Directorate of Environment, Castile and Leon (Spain), the General Directorate of Citizen Security (Fire-fighters service) in Madrid, and the Fire Prevention Service of the Generalitat in Valencia. And ESA Climate Change Initiative for the provision of CCI-Land Cover product (https://www.esa-landcover-cci)