After a review of the current state of the art in LAI retrieval with optical and radar remote sensing data, this study investigates the capabilities of satellite remote sensing imagery in operational crop growth monitoring. This study demonstrated that the availability of an extensive crop field delineation database (like existing for the entire Belgian country) is of crucial in interest in order to retrieve crop specific information. LAI remote sensing retrieval was achieved during the year 2003 on a large Belgian agricultural area (4500 km<sup>2</sup>) for Sugar beet, Winter wheat and Maize crops. In order to increase the monitoring temporal frequency, an integration of SPOT-HRV, ENVISAT-MERIS and ERS2-SAR sensors was carried out, with a good level of accordance. The retrieval results were compatible with the concurrent field measurements as well as with the outputs given by the WOFOST crop growth model.
Previous experiments demonstrated the relationships between the radar backscattering coefficient, σ<sub>o</sub> and crop parameters such as fresh biomass, plant height and Leaf Area Index (LAI). Topsoil water content also influences the backscattered signal and is as such a required input parameter in the physical and semi-empirical models that extract vegetation parameters from σ<sub>o</sub>. In an operational environment, it is not possible to measure soil moisture over an entire agricultural region. As the vegetation cover hampers the radar remote sensing of soil moisture, near surface soil moisture can be simulated using a hydrological model. In this paper, it is investigated whether soil moisture values obtained through the hydrological model TOPLATS can be used in a crop parameter retrieval algorithm. The data set used for this investigation was collected from March to September 2003 in the Loamy Region, Belgium. During this period, 18 agricultural fields were sampled for vegetation parameters and soil moisture. In addition, 11 ERS-2 images of that period were acquired of which 6 coincided with the field measurement dates. Because the necessary catchment data were not available, TOPLATS was calibrated on a point scale for every field with in situ soil moisture. The calibrated TOPLATS model was applied to simulate soil moisture values at the ERS-2 acquisition dates for which no soil moisture field measurements were available. In parallel, the Water Cloud model was calibrated using the biophysical parameters measured on the field in order to retrieve LAI estimates from ERS SAR time series. In a second step, the simulated soil moisture values corresponding to the SAR acquisition dates were used as input in the Cloud model as substitutes of field measurements, and the propagation of the soil moisture estimate error in the LAI retrieval algorithm was studied. Finally the experimental results were discussed in the perspective of a regional crop monitoring system and the operational feasibility is assessed.