LIVE FUEL MOISTURE CONTENT AND IGNITION PROBABILITY IN THE IBERIAN PENINSULAR TERRITORY OF SPAIN

Abstract

This paper presents an operational algorithm to produce Live Fuel Moisture Content (LFMC) at national scale from MODIS data. The algorithm is based on the inversion of Radiative Transfer Models (RTM) that estimate moisture content based on different simulation scenarios. In addition, logistic regression models were calibrated to convert the derived LFMC values into Ignition Probability (IP) maps. The areas under the curve obtained by the Receiver Operating Characteristic (ROC) plot method provided by the models were close to 0.6. Several statistical analyses were performed in order to ascertain whether the variables proposed to be included in the fire danger model were significantly related to forest fires. A non parametric U-Mann-Withney test confirmed significant differences between fire and non-fire pixels (p<0.001). Fire pixels occurred at significantly lower LFMC values than the non-fire pixels.