Assessment of a Robust Satellite Technique for forest fire detection and monitoring by using a Total Validation Approach (TVA).

A multi-temporal satellite data analysis, named RST (Robust Satellite Technique), already successfully used to study different natural and environmental hazards, was recently applied and tested in a pre-operational way, for forest fires monitoring. In this paper, the results of RST applied to NOAA-AVHRR data are evaluated in the case of a real time monitoring of winter fires in Northern Italy. The collaboration with Local Civil Protection offices, which made dedicated light aircraft surveys and voluntary teams available for direct ground observations, permitted us to exploit (for the first time) a Total Validation Approach (TVA) in order to correctly evaluate apparent false alarms and to fully appreciate the sensitivity of the methodology also in the case of very small active fires. TVA, based on a systematic study of the origin of each hot spot identified by RST, allowed us to recognize most of them as actual thermal anomalies (associated to small fires, to variations of thermal emission in industrial plants, etc.) and not as false alarms simply because not associated to officially documented forest fires. On this basis, a suitable scheme for forest fire detection able to offer, at the same time, high sensitivity (to very small active fires) together with the highest reliability (false alarm rate close to 0%), was drawn and successfully tested. Its extension to sensors aboard other polar (EOS/MODIS) or geostationary (MSG/SEVIRI) satellites will be also discussed.