Study of Backscatter Signature for Seedbed Surface Evolution Under Rainfall - Influence of Radar Precision

We propose a 3D-approach of the soil surface height variations, either for the roughness characterization by the mean of the bidimensional correlation function, or as input of a backscattering model. We consider plots of 50 cm by 50 cm and two states of roughness of seedbed surfaces: an initial state just after tillage and a second state corresponding to the soil roughness evolution under a rainfall event. We show from stereovision data that the studied surfaces can be modelled as isotropic Gaussian processes. We study the change of roughness parameters between the two states. To discuss the relevance of their differences, we find from Monte-Carlo simulations the bias and variance of estimator for each roughness parameters. We study the roughness and moisture combined influences upon the direct backscattering coefficients by means of an exact method based on Maxwell’s equations written in a nonorthogonal coordinate system and by averaging the scattering amplitudes over several realizations. We discuss results taking into account the numerical errors and the precision of radar. We show that the ability of the radar to discriminate the different states of seedbed surfaces is clearly linked to its precision.