Application of flood monitoring from satellite for insurances

Flood monitoring from satellite data provides the opportunity to quickly and precisely overview flooded areas. The extent of the flooding and affected areas can be delivered to water authorities, civil protection agencies or insurances. Evaluations include information to facilitate damage assessment, to better estimate risk in future, and to prepare protection measures. For demonstrating the potential of satellite based evaluations, results for the Elbe flood that occurred in central Europe in August 2002 are presented. This flood had a geographical and economic large impact. In order to identify the affected areas, a strip of 450 km along the river Elbe was investigated. Satellite data of different sensor systems (Landsat, SPOT, ERS) had to be analyzed to derive a complete flood extent data set. For different river sections satellite data of different dates were used to perform the best fit to the flood peak along the river. Damage assessment can be supported using the flood extent overlaid with a city map in high resolution. Maps of complete Dresden have been requested by a German insurance company to support their damage assessment. The evaluation of flooded areas further allows improvements for planning as e.g. the declaration of risk areas. Observed flood extents can serve as validation source for simulations of inundations using hydrodynamic models. The combination of additional information to the flood extent layer opens up various applications. Landuse classified in industrial areas, settlements, agriculture, and forest, which were derived from up-to-date satellite images, are intersected with the inundated areas. This intersection results into the detection of affected areas (and economic values), which have to be protected against flood events in future. Multitemporal evaluations using a set of images of different days in succession document the spatial and temporal dynamics of a flood event. They allow to better rate damages as well as consequential costs.