HYDRASENS - Integrating radar remote sensing, hydrologic and hydraulic modeling for surface water management

Context and objectives

The rationale of this project is the increasing need for improved tools supporting water management at the catchment level as required by different national and international legislations. In 2000, the Water Framework Directive has been introduced in Europe which set the political frame for implementing Integrated Water Management (IWM) at the catchment level. The European Commission is currently also preparing a directive on the assessment and management of floods. This new directive will aim at the reduction and the management of the risks that floods pose on human health, the environment, and the infrastructure. As it is foreseen that in the coming decades the risk for flooding will be higher, and the economic damage will increase appropriate measures are needed that can reduce its likelihood, or limit its impacts. These measures should be based on a thorough understanding of the impact of management measures on the hydrologic and hydraulic functioning of catchments.
The overall goals of this project are 1) to explore new strategies to integrate radar remote sensing, hydrologic, and hydraulic modelling for water management purposes through data assimilation, and 2) to demonstrate the applicability of advanced data assimilation schemes for a set of water management problems. The major outcome will be fundamental knowledge about radar data assimilation in hydrology. The major deliverables will be a set of data and operational procedures supporting water management.

Project outcome

  • The potential of remote sensing for retrieving hydrological relevant information for updating hydrologic and hydraulic models was investigated with a focus on soil moisture and flood extent.
  • Two alternative ways were explored that circumvent roughness measurements which are an important issue in moisture retrieval from SAR data: a first investigated the impact of roughness uncertainty on retrieved soil moisture, based on possibility theory, a second introduced the concept of effective roughness parameters which can be related to the backscatter coefficient and seems promising;.
  • An improved methodology for determining soil moisture at the microscale, based on the GPR-technique was developed and it was possible to capture the spatial soil moisture patterns. However, it was more complex to derive a scaling function that allowed to relate soil moisture;
  • Improvements to the SAR based flood mapping extent mapping have been introduced through incorporating high accuracy DEM information;
  • A loosely-coupled hydrologic/hydraulic model was developed. and data assimilation schemes, both the EnKF and the PF, were implemented. First, different calibration algorithms were compared from which was concluded that PSO slightly performed better. In a second phase, soil hydraulic parameters of a physically-based hydrologic model were calibrated based on SAR-derived soil moisture values. It was found that a high spatial variability was observed for the different parameters.
  • Then soil moisture information was assimilated into hydrologic models. Two different methodologies were tested or proposed. First, different alternative approaches of the particle filter were tested. Second, a methodology for assimilating coarse scale soil moisture observations in a high resolution hydrologic model was developed. With respect to the hydraulic modelling, first a simple hydrologic model, i.e. a cascade model was coupled to a hydraulic model, i.e. HEC-RAS. Then remote sensing data, used to monitor water storage were used in the calibration of the model. In a second phase, flood maps were assimilated through the PF in order to improve flood inundation modelling.
  • Concerning data assimilation, it was shown that, for a single event, the joint assimilation of soil moisture and flood extent improves the forecasting of the flood wave;
  • It was also shown that the use of remote sensing information during the calibration phase of hydrologic models.