Rudi Hoeben

Assimilation of active microwave observation data for soil moisture profile estimation.

This paper discusses the potential of retrieving information about the soil moisture profile from measurements of the surface soil moisture content through active microwave observations of the Earth. Recently, Mancini et al. [1999] have shown through laboratory experiments that the volumetric moisture content of the first few centimeters of a bare soil can be determined within 5 ol accuracy by means of C and L band active microwave observations and inverse modeling. Here we use active microwave observations of the surface soil moisture content in a data assimilation framework to show that this allows the retrieval of the root zone soil moisture profile. The data assimilation procedure developed is based on the Kalman filter technique. Kalman filtering allows reconstruction of the state vector of a system when this system is represented by a dynamic model and when at least part of the state variables are observed regularly. The dynamic model of the system used here is based on the one-dimensional Richards equation. The observation equation is based on the Integral Equation Model [Fung et al., 1992; Fung, 1994] and is used to link the radar observations to surface soil moisture content. It is shown that even in the presence of model and observation noise and infrequent observations, accurate retrieval of the entire moisture profile is possible for a bare soil. ? 2000 American Geophysical Union

The importance of the spatial patterns of remotely sensed soil moisture in the improvement of discharge predictions for small-scale basins through data assimilation

Abstract In this paper, we investigate to which degree information concerning the spatial patterns of remotely sensed soil moisture data are needed in order to improve discharge predictions from hydrological models. For this purpose, we use the TOPMODEL-based Land–Atmosphere Transfer Scheme (TOPLATS). The remotely sensed soil moisture values are determined using C-band backscatter data from the European Space Agency (ESA) European Remote Sensing (ERS) Satellites. A baseline run, without soil moisture assimilation, is established for both the distributed and lumped versions of the land–atmosphere scheme. The modeled discharge matches the observations slightly better for the distributed model than for the lumped model. The remotely sensed soil moisture data are assimilated into the distributed version of the model through the ‘nudging to individual observations’ method, and the ‘statistical correction assimilation’ method. The remotely sensed soil moisture data are also assimilated into the lumped version of the model through the ‘statistical correction assimilation’ method. The statistical correction assimilation method leads to similar, and improved, discharge predictions for both the distributed and lumped models. The nudging to individual observations method leads, for the distributed model, to only slightly better results than the statistical correction assimilation method. As a consequence, it is suggested that it is sufficient to assimilate the statistics (spatial mean and variance) of remotely sensed soil moisture data into lumped hydrological models when one wants to improve hydrological model-based discharge predictions.

Multifrequency radar observations of bare surface soil moisture content: A laboratory experiment

This paper reports on a laboratory experiment that investigates the use of active microwave observations to estimate volumetric soil moisture content. The experiment, held in 1995, was set up at the European Microwave Signature Laboratory, Joint Research Centre of the European Communities, Ispra (Italy). Full polarimetric radar observations of a 2 m diameter cylindric container filled with a sandy loam soil were performed. During successive wetting and drying cycles, different soil moisture profiles were generated and observed in situ by means of time domain reflectometry probes. The radar data are analyzed based on the Integral Equation Model that simulates radar backscattering given known surface characteristics, such as moisture content and roughness. This allows the evaluation of the predictive power of the simulation model. We find general good agreement between measurements and simulations, but problems occur at high incidence angles. The model is then used to invert soil moisture information from radar measurements. It is shown that, in spite of the complexity of the model involved, it is possible to retrieve under certain circumstances reliable soil moisture estimates with similar accuracies as the in situ measurements.

Soil moisture inversion from ERS and SIR-C imagery at the Zwalm catchment, Belgium

Several models have been presented in the previously which should allow soil moisture inversion from bare soil radar backscattering. Three widely used models (i.e., the IEM (Fung et al. 1992), the models of Dubois et al. (1995) and of Oh et al. (1992)) have been applied to ERS and SIR-C data obtained over the Zwalm catchment in Belgium. Results show an inability of the models to estimate surface soil moisture accurately through direct inversion. However, applying the method of effective roughness parameters (Su et al. 1997) on the multi-temporal ERS tandem mission data resulted in a significant improvement of the soil moisture retrieval. The same method applied to multi-frequency SIR-C data was not able to improve the estimations. It is believed that the inversions suffer from the high soil roughness sensitivity of the backscattering at higher incidence angles.

Sensitivity of radar backscattering to soil surface parameters: a comparison between theoretical analysis and experimental evidence

The understanding of the sensitivity of radar backscattering to surface parameters is essential in applying microwave remote sensing to the retrieval of geo- and bio-physical parameters. A theoretical model, the integral equation model, is used to investigate the sensitivity of radar backscattering to soil surface parameters. This model is first tested against a dataset retrieved under well controlled conditions at the European Microwave Signature Laboratory, Joint Research Centre, Ispra, Italy. Then the surface roughness parameters are normalised with respect to wavelength and incidence angle to improve insight into the sensitivity of radar observations to surface roughness. In this way the roughness space can be subdivided in two regions where surface slope controls the dependency of the signal to roughness.

Assimilation of active microwave measurements for soil moisture profile retrieval under laboratory conditions

The authors discuss the potential of retrieving information on the soil moisture profile from measurements of the surface soil moisture content through active microwave observations. They use active microwave observations of the surface soil moisture content in a data assimilation framework to show that this allows the retrieval of the entire soil moisture profile. The data assimilation procedure demonstrated is based on the Kalman filter technique. Kalman filtering allows reconstruction of the state vector when at least part of the state variables are observed regularly. The dynamic model of the system used is based on the 1D Richards equation. The observation equation is based on the integral equation model of A. K. Fung et al. (1992) and is used to link the radar observations to surface soil moisture content. Recently, M. Mancini et al. (1999) reported about laboratory experiments investigating the use of active microwave observations to estimate surface soil moisture content. The present authors apply the data assimilation scheme to the radar measurements of these experiments to retrieve the entire soil moisture profile in the soil sample used, and compare these results with the soil moisture profile measurements (using TDR). It is shown that with a limited number of radar measurements accurate retrieval of the entire soil moisture profile is possible.

Assessing the Applicability of Hydrologic Information from Radar Imagery

During the last two decades, the potential of radar remote sensing in the retrieval of the water content of the near-surface unsaturated soil zone has been explored. This water content is usually referred to as soil moisture. The inversion of radar observations into soil moisture values has been hampered by an insufficient characterization of the soil roughness. However, some studies have focused on relating temporal changes of the radar signal to hydrologic relevant information. One result is presented here, where we show that variable source areas, which are mainly responsible for runoff in a catchment, can be visualized through a principal component analysis. A second part of this paper shows how soil moisture information obtained from radar imagery can be incorporated into hydrologic models. A first example uses an extended Kalman filtering technique, which adjusts the state variables from a hydrologic model. Through this technique, one-dimensional soil moisture profiles are retrieved with high accuracy. In a second example, we show a data-assimilation method which uses both the statistics and the spatial distribution of radar-retrieved soil moisture values, to adjust the modeled soil moisture profile. This methodology enables a better modeling of the rainfall-runoff behavior of the catchment.