Predicting water erosion in arid lands using the GIS-based RUSLE model: A case study of Bedour catchment, central Tunisia

Abstract

This study was conducted with a view to quantifying soil erosion in arid lands of Tunisia. To do this, we have opted to use the RUSLE model based on geographic information systems. By collecting data on rainfall, soils, vegetation, slopes and conservation practices separately as a layer and determining the pixel values for each of these factors, a quantified assessment of erosion in the basin is obtained. The data superposition and computing, following the model equations and protocol, allowed us to know the spatialized water erosion values at the pixel level. For the whole catchment, the study showed values oscillating between 0 and 163 Mg·ha -1 ·year -1 with an average annual rate of 3 Mg·ha -1 ·year -1 . With such a low R (rainfall erosivity) factor (between 21.43 and 21.88 MJ·mm·ha -1 ·h -1 ·year -1 ) itself related to low monthly and annual rainfall amounts, the region experiences locally very high annual erosion rates. Soil protection through conservation practices has saved the basin from even higher erosion. While plains cultivated and equipped with contour benches often suffer from low rates of erosion (less than 2 Mg·ha -1 ·year -1 ), unused slopes are neglected without protection, resulting in significantly high rates of erosion.