Integration of optical and radar remotely sensed data for mapping forest fires in Mediterranean regions

In Southern Europe wildfires can be a key factor for the ecosystem dynamics and they can seriously thread human lives and infrastructures1. Monitoring can support post-fire management (e.g. to identify the location and assess the extent of the burned surfaces, to evaluate the damage to the forest stands, to follow vegetation recover and to plan restoration interventions) and remotely sensed data are a key source of information for supporting these tasks2. In this framework, mapping the areas affected by fires (burned areas) is the first key step that can be accomplished at the local, regional and global scale with remotely sensed data. We developed a burned area mapping fuzzy algorithm tuned over Mediterranean regions and Landsat Thematic Mapper (TM) data; the approach relies on the interpretation and the integration of spectral indices with fuzzy sets theory. The exclusive use of optical remotely sensed images poses issues due to, for example, the presence of clouds, the limited number of images, the errors produced by spectral overlap between burned areas and sparsely vegetated areas. Hence, we tested the feasibility of integrating into the algorithm information extracted from radar images. Over a test site in Portugal, we acquired a Landsat TM image and a multi-temporal dataset of SAR images to be processed and implemented into the burned area fuzzy algorithm. This contribution presents the preliminary results