Research project T4/DD/33 (Research action T4)
This proposal corresponds to the objectives of the "Fundamental Research" projects within the program "Satellite Observation of the Earth", subprogram "Global Change", theme "Soil Moisture" (A2).
The current proposal is a continuation of the TELSAT4 project T4/DD/009 "Development of assimilation schemes for soil moisture profiles using remotely sensed data?. During the execution of that project a detailed analysis of multi-frequency polarimetric data, collected during the Non-Vegetated Terrain experiment, held at the European Microwave Signature Laboratory (EMSL), JRC, Italy, was performed in order to assess the accuracy of radar derived soil moisture estimates under a wide range of surface conditions (roughness and moisture content) and radar configurations (frequency, polarisation and incidence angle). The radar derived surface soil moisture content was then used to update a 1-D hydrological model to retrieve information on the moisture content of the entire profile. The updating was performed by means of an assimilation scheme based on the Kalman filter concept, and using the Integral Equation Model as the observation equations. Also during the execution of that project satellite imagery and field data were collected in the experimental catchment of the Zwalmbeek and one of its tributaries (the Wijlegemse Beek). 19 field campaigns were organised in 1997-98, resulting in 13 ERS-2 and 6 RADARSAT SAR images and corresponding in-situ soil moisture data. Also a 350 m long transect, perpendicular to the stream of the Wijlegemse Beek, was installed. Along this transect weekly soil moisture profiles of the root zone (to a depth of 70 cm) were measured using TDR probes. At one location along the transect a permanent soil moisture profile station was installed, yielding 10-min information on the time-variability of moisture fluctuations.
This proposal will extend the findings from T4/DD/009 to the catchment scale. A similar assimilation procedure will be developed. This assimilation procedure will be based on a 3-D hydrological model, capable in simulating the soil moisture dynamics in the root zone by accounting for the external driving forces, such as rainfall and evapotranspiration. Again a Kalman filter will be used to update the predictions of the hydrological model every time surface soil moisture information becomes available from the SAR image analysis. The observation equations are provided through the Integral Equation Model as applied to the entire catchment.
The expected result from this research is a robust algorithm to retrieve root zone soil moisture information at the catchment scale from radar imagery. The procedure is based on 3-D hydrological modelling and an assimilation scheme based on the Kalman filter concept. When successful this algorithm will be the first to allow the estimation of catchment-scale root zone soil moisture content. This will allow improved prediction of crop growth, floods, and catchment-scale evapotranspiration. Possible end users are agricultural services, hydrological services and meteorological institutes responsible for long-term weather forecasting.
