Regional Variability and Driving Forces behind Forest Fires in Sweden

Extreme forest fires have been a historic concern in the forests of Canada, Russian Federation, and USA, are now an increasing threat boreal Europe, where recent fire events 2014 2018 drew attention to Sweden. Our study objective was understand vulnerability Swedish by spatially analyzing historical burned areas, link with weather, landscape, fire-related socioeconomic factors. We developed extensive database 1 × km2 homogenous grids, monthly areas were derived from MODIS FireCCI51 dataset. The consists various socio-economic, topographic-, forest-, weather-related remote sensing products. To include new factors IIASA’s FLAM model, we random model assess spatial probabilities areas. Due Sweden’s geographical diversity, dynamics vary between six biogeographical zones. Therefore, applied each zone separately. As outcome, obtained across Sweden observed well captured model. result accuracy differs respect zone; area under curve (AUC) 0.875 0.94 for zones few fires, but above 0.95 higher number events. Feature importance analysis their variability provide valuable information reasons behind fires. Fine Fuel Moisture Code, population road densities, slope aspect, stand volume found be among key modeling approach can extended hotspot mapping other regions thus is highly policy-relevant. Visualization our results available Google Earth Engine Application.