Remote sensing for monitoring the effects of forest fires

Forest fires are a phenomenon which occurs regularly in the Mediterranean basin, and particularly in the Italian Peninsula. They generally depend on the climatic features of the area, on the vegetation type and on the dryness of existing fuel. The present work illustrates the capacity of remote sensing and GIS methodologies to monitor the effects of fires on forest vegetation. Specifically, both the delineation of areas affected by fires and the monitoring of vegetation re-growth are considered. Two main methodologies are identified, the first based on the analysis of multitemporal scenes (pre and post event), the second based on the use of an unique scene (post event). In the former case change detection techniques are applied, while image segmentation analysis are utilized in the latter case. The two approaches may be applied both starting from the original spectral bands of a scene and starting from specific transforms aimed at ighlighting some vegetation features. A classical example of the second method is provided by the combination of red and near infrared data to compute vegetation indices. The paper concludes with an overview of the methodology developed by LaMMA within the INTERREG-MEDIFIRE project, aimed at delineating forest areas affected by fires.