RST-BASED FLOODED AREA MAPPING AND MONITORING IN NEAR REAL-TIME BY USING MODIS DATA

摘要

Flood forecast and mitigation actions need updated and timely information about precise location, extent and dynamic evolution of the flooding event. Remote sensing technology, based on microwave and optical satellite data, is currently capable of giving reliable contributions towards a rapid detection of affected areas in order to improve flood hazards management and to study remote areas where ground-based observation systems are still lacking. For a near real time monitoring and mapping of flooded areas, fundamental during the crisis and post-crisis phases to support civil protection activities, frequent observations of the Earth's surface can be derived from optical sensors aboard meteorological satellites. Recently, a new Robust Satellite Technique using AVHRR (Advanced very High Resolution Radiometer) observations has been proposed for mapping and monitoring flooded areas, providing good results. Afterwards, the same approach has been exported on MODIS (Moderate Resolution Imaging Spectroradiometer) data, in order to investigate if its higher spatial resolution in visible and near-infrared channels might be exploited to increase the accuracy in both near real time detection and mapping of flooded areas. Preliminary results confirmed the reliability and the sensitivity of the proposed approach but further analyses have to be carried out in order to better assess the actual reliability and efficiency of such a technique. To this aim, in this paper, the extreme flooding event which hit wide territories of Germany and Czech Republic, during the August 2002, has been studied.