Utilisation des donnees RADARSAT pour I'observation de I'orientation des labours comme parametre du ruissellement

摘要

In agricultural regions, the assessment of surface conditions using remote sensing data is very useful for the management and prevention of risks related to excessive runoff. To this day, imaging radar is the only operational remote sensing instrument sensitive to an important runoff indicator: surface roughness. On agricultural fields, roughness is described by a random component that is related to the cropping practice, and a linear component due to the direction of the mechanical tillage. The aim of this study is to evaluate the effect of tillage orientation on the backscattering coefficient, using RADARSAT imagery. Several SAR images and the corresponding ground measurements were acquired on the Pays de Caux and over the Solnan watershed (France), during winter of 1998 and 1999. The images were calibrated and geocoded. The collected field data were soil moisture and surface roughness (rms height and row orientation). Statistical analysis were made in order to establish relations between the surface parameters and the satellite data. It is demonstrated that, on very rough fields, there is little correlation between radar backscatter and rms height. However, there is a significant relationship between tillage direction and backscatter. An empirical model that combines rms height and orientation was tested in order to gain a better understanding of the effects of both these roughness parameters on radar backscatter. Results show that it is possible, using RADARSAT imagery to map surface roughness and orientation of tillage. This information can be very useful for the modeling of runoff in agricultural areas. This work was pan of the FLOODGEN project supported in part by the RADARSAT Users Development Program and the CEO program of the European Commission.