Prediction of soil erosion in a Lake Victoria basin catchment using a GIS-based Universal Soil Loss model

Soil erosion patterns in watersheds are patchy, heterogeneous and therefore difficult to assess. The problem can be overcome by using predictive models. However, wide spread soil erosion model-factor parameterization and quantification is difficult due to the costs, labor and time involved. The objective of this study was to evaluate different methods of USLE input parameter derivation and to predict soil loss within a microcatchment of the Lake Victoria basin (LVB). The highest soil loss was predicted for annual cropland use (93 t ha−1year−1), followed by rangeland (52 t ha−1year−1), banana–coffee (47 t ha−1year−1), banana (32 t ha−1year−1) and forest and papyrus swamp. In the terrain units, soil loss was highest within the back slopes (48 t ha−1year−1) followed by the summits (42 t ha−1year−1) and valleys (0 t ha−1year−1). For the soil units, soil loss was highest in the Chromic Luvisols (52 t ha−1year−1) followed by Petroferric Luvisols (37 t ha−1year−1), Mollic Gleysols (5 t ha−1year−1) Dystric Planasols (0 t ha−1year−1) in large part because soil classification often correspond to various slope position.