Predicting the Extent of Wildfires Using Remotely Sensed Soil Moisture and Temperature Trends

Abstract

Recent climate trends evidence a rise of temperatures and an increase in the duration and intensity of droughts which is in turn leading to the occurrence of larger wildfires, which threaten the environment as well as human lives and beings. In this context, improved wildfires prediction tools are urgently needed. In this paper, the use of remotely sensed soil moisture data as a key variable in the climate-wildfires relationship is explored. The study is centered in the fires registered in the Iberian Peninsula during the period 2010-2014. Their prior-to-occurrence surface moisture-temperature conditions were analyzed using SMOS-derived soil moisture data and ERA-Interim land surface temperature reanalysis. Results showed that moisture and temperature conditions limited the extent of wildfires, and a potential maximum burned area per moisture-temperature paired values was obtained (R ² = 0.43). The model relating fire extent with moisture-temperature preconditions was improved by including information on land cover, regions, and the month of the fire outbreak (R ² = 0.68). Model predictions had an accuracy of 83.3% with a maximum error of 40.5 ha. Results were majorly coherent with wildfires behavior in the Iberian Peninsula and reflected the duality between Euro-Siberian and Mediterranean regions in terms of expected burned area. The proposed model has a promising potential for the enhancement of fire prevention services.