SAR imagery to estimate roughness parameters when modelling runoff risk

The influence of the roughness of agricultural soil on runoff and erosion i s a proven fact. Synthetic aperture radar (SAR) sensors should enable discr imination between plots with different cropping patterns. A study of Medite rranean vineyards in southern France was made, with the aim of obtaining a better understanding of the potential for using radar satellite data from E RS-1 when estimating roughness parameters. Roughness measurements enabled m odelling of the backscattering coefficient (sigma degrees) of known surface s, using the electromagnetic Integral Equation Model (IEM). The good correl ation between ERS-1 and IEM data indicated the feasibility of extracting ro ughness parameters by means of remote sensing methods. Seven ERS-1 images were examined, corresponding to different stages in the development of vegetation and roughness. Two images were finally selected a s they offered the possibility of discriminating between two factors: (1) t he orientation of mechanical labour, which can be related to a periodic and stable roughness over time, and (2) cropping practices, corresponding to a random roughness pattern that changes with season. Both roughness parameters derived from SAR satellite data contribute additi onal data to runoff models-a preferred runoff direction as defined by furro w direction, as well as the intensity of this runoff under the influence of random roughness. A rule for the behaviour of sigma degrees in terms of fu rrow orientation is presented.