FLOODMOIST - Flood mapping and soil moisture retrieval for improved water management

Context and objectives

Flood prediction systems are of key importance for properly managing the event and organizing rescue operations. Unfortunately, the models which are used make errors with respect to timing, flood extent or stage height. In this spin-off project we will investigate how radar remote sensing observations of soil moisture and flood extent can be jointly assimilated into flood prediction systems. The overall goals of this spin-off project are 1) to explore new strategies to extract hydrology-related information from microwave remote sensing (i.e. soil moisture and flood extent) and 2) to demonstrate the merit of jointly assimilating soil moisture and flood extent information into coupled hydrologic-hydraulic models.

Project outcome

  • Development of a robust roughness model that can be used in variety of environmental conditions over the world;
  • The model can be used in a soil moisture retrieval algorithm to obtain qualitative estimates of soil moisture using passive microwave observations;
  • Development of wo methods for characterizing the uncertainty in SAR-based flood mapping. A first method focused on speckle-related uncertainty, whereas the second method aimed at estimating the total uncertainty. It was concluded that both methods yielded comparable results. The potential of producing probability maps from these uncertainty images was also shown;
  • Soil moisture and flood extent observation derived from satellite observations were used to jointly assimilate a coupled hydrological-hydraulic model. This improved the forecasting of the flood for a single event;
  • Development of an alternative method to calibrate inundation models with SAR data.