Operational performance of current synthetic aperture radar sensors in mapping soil surface characteristics in agricultural environments: application to hydrological and erosion modelling
Résumé
SAR (Synthetic Aperture Radar) sensors are often used to characterise the surface of bare soils in agricultural environments. They enable the soil moisture and roughness to be estimated with constraints linked to the configuration of the sensors (polarization, incidence angle and radar wavelength). These key soil characteristics are necessary for different applications, such as hydrology and risk prediction. This article reviews the potential of currently operational SAR sensors, and those planned for the near future, to characterise soil surface as a function of users' needs. It details what it is possible to achieve in terms of mapping soil moisture and roughness by specifying optimal radar configurations and the precision associated with the estimation of soil surface characteristics. The summary carried out for the present article shows that mapping soil moisture is optimal with SAR sensors at low incidence angles (<35°). This configuration, which enables an estimated moisture accuracy greater than 6% is possible several times a month taking into account all the current and future sensors. Concerning soil roughness, it is best mapped using three classes (smooth, moderately rough, and rough). Such mapping requires high-incidence data, which is possible with certain current sensors (RADARSAT-1 and ASAR both in band C). When L-band sensors (ALOS) become available, this mapping accuracy should improve because the sensitivity of the radar signal to Soil Surface Characteristics (SSC) increases with wavelength. Finally, the polarimetric mode of certain imminent sensors (ALOS, RADARSAT-2, TerraSAR-X, etc.), and the possibility of acquiring data at very high spatial resolution (metre scale), offer great potential in terms of improving the quality of SSC mapping.