On the potential of multi-temporal analysis of MODIS and AVHRR data for near real time mapping and monitoring of flooded areas

Floods are devastating natural disasters which may cause very high costs in lives and damages. Among the other potential contributes, satellite remote sensing may help for mapping and monitoring flooded areas. In order to support flood risk management, information coming from satellite, especially during crisis phases, need to be timely provided. Besides, suitable techniques, able to reliably detect flooded areas with a low rate of false alarms, are requested. In this context, a new technique using visible AVHRR data has been recently proposed. In this work, in order to further confirm the reliability and the sensitivity of the proposed approach we analyze the August 2002 flood which hit wide territories of Germany. Afterwards, we exported this approach on MODIS data, in order to exploit the higher spatial resolution in the visible and near-infrared channels offered by such a sensor, for increasing the accuracy in both near real time flooded area mapping and monitoring