Automated detection of thermal features of active volcanoes by means of infrared AVHRR records

An innovative, Advanced Very High Resolution Radiometer (AVHRR)-based technique for improved automatic detection of volcanic hotspots and thermal anomalies is proposed in this paper. It is mainly based on a multitemporal analysis of historical, long-term satellite records. Such a technique basically rests on the Robust AVHRR Techniques (RAT) approach, which has been already successfully applied to several natural and environmental emergencies (e.g., fires, floods, earthquakes). In this work, the proposed technique has been tested on an extended set of eruptive events of Mt. Etna and Stromboli volcanoes. Results achieved, in terms of reliability (low false alarm rate) as well as of effectiveness (detection sensitivity), are described in detail. Moreover, the potential in low-level thermal anomaly detection, as possible pre-eruptive thermal signs, is also addressed and preliminary results obtained for a couple of events, discussed. The study cases here presented show the benefits of such a technique especially when different observational conditions (time/season of pass, atmospheric moisture content, solar illumination, satellite angles of view, etc.) are considered, making such a method globally applicable. ture prospects, also in terms of possible operational scenarios, coming from the implementation of such an approach on the new generation of satellite sensors (such as SEVIRI aboard Meteosat Second Generation) are also discussed.