Soil Loss Estimation Based on the USLE/GIS Approach Through Small Catchments - A Minor Field Study in Tunisia

suédois. A USLE/GIS approach was applied to estimate soil loss for the two catchments ‘Mrichet’ and ‘Sadine2’ in the semi-arid Tunisian Dorsal. The approach is inspired by the doctoral thesis ‘Water erosion modeling using fractal rainfall disaggregation – A study in semiarid Tunisia’ by Dr. Sihem Jebari. The Universal Soil Loss Equation (USLE) was applied to predict soil loss magnitude and Geographic Information System (GIS) software ArcView and ArcMap was used to simulate the soil loss in spatial distribution. Each one of the USLE-parameters (rainfall erosivity R, soil Erodibility K, topography LS, conservation practice P and land use C) were represented by a thematic raster layer in the GIS. The thematic layers were defined from available data such as satellite images (C and P), pedological maps (K), rainfall intensity records (R) and topographic maps (LS). The model was calibrated primarily considering the grid cell size resolution of the raster layers. The estimated soil loss was, depending on the grid cell dimension used, estimated at approximately 12-16% of the observed soil loss for both catchments. The underestimation of the soil loss is most likely due to underestimated K- and R-factor values. Also the USLE is restricted to rill and inter rill erosion while the comparative observed erosion is likely to include all types of water erosion within the catchment. This implies that the modeled values could be expected lower than the observed. In addition some soil samples were analyzed and K-factor values experimentally determined for comparison with the theoretical. These values were in the same order of magnitude as the theoretically determined.