Modeling areas burned in forest fires over 2000-2015 in Catalonia (Northeast Spain)

Forest fires have become one of the significant global challenges in the 21 st century, posing a grave threat to European Mediterranean landscapes and ecosystems. In this study, we applied the wildFire cLimate impacts and Adaptation Model (FLAM) to simulate historical burned areas in the Catalonia region of Spain. FLAM is a process-based model that operates with daily time steps at different spatial scales. It employs mechanistic fire modeling algorithms to parameterize the impacts of climate, fuel availability, and human activities on the likelihood, occurrence, and extent of wildfires. Our study focused on optimizing, calibrating, and downscaling FLAM probability algorithms to suit the regional conditions. Our analysis of historical wildfire data shows that the Catalonia region experienced approximately 288 fires, destroying over 63,000 hectares (ha) of forestland between 2000 and 2015, with frequent fire episodes occurring during the fire season (June-August). According to our modeling results, the FLAM model was able to reproduce the total monthly burned area (ha) in each pixel across the calibration period (2004-2009) and capture the general representation of the seasonal pattern of wildfire activity with recurring large fires every extreme fire season (July) throughout the entire study period (2000-2015). We evaluated the model performance using Pearson's correlation coefficient (r) and the Mean Absolute Error (MAE) for monthly data within each year. The model demonstrated varying Pearson's r and MAE values across different years. However, we found strong correlations and low MAE values in thousands of hectares in specific years, such as 2005 (r = 0.95, MAE = 2.72), 2007 (r = 0.87, MAE = 2.73), 2010 (r = 0.79, MAE = 2.02), 2011 (r = 0.75, MAE = 1.75), and 2015 (r = 0.94, MAE = 1.15). While the model effectively reconstructed seasonal burned area patterns in Catalonia, further research is needed to accurately model the inter-annual variations and spatial patterns of burned areas.