Evaluation of ArcMUSLE model to prediction of sediment yield in the Upper Rio Negro watershed, southern Brazil

Abstract

Water erosion is one of the main processes which cause soil degradation in tropical environments. This process occurs by the interactions between hydrological, geomorphological and pedological factors that cause the detachment and transport of particles. Through process understanding it is possible to make use of mathematical models in order to estimate the sediment yield in watersheds. The main objective of this paper is to apply the Modified Universal Soil Loss Equations (MUSLE) to simulate a rainfall event and evaluate the fit of this simulation to the measured data of sediment transport.  MUSLE is a deterministic, empiric and originally lumped model that replaces the rainfall erosivity factor by the overland flow factor to simulate sediment yield in individual rainfall events. In addition to overland flow effects, the model considers the soil erodibility factor (K), hillslope length and steepness (LS), crop and management factor (P) and the erosion control practice factor (C) to perform simulations. The ArcMUSLE model was used integrated with a Geographical Information System (GIS) allowing an estimate, entirely for isolated events, of the overland flow and the sediment yield through a spatially distributed form. The model was applied to the Upper Negro River watershed, located between the geographic coordinates 25°55’S and 26°14’S, 48°56’W and 49°23’W, and drains an area of 788km².  For the simulated event, with total rainfall of 60mm, results showed that both, overland flow and sediment transport, were overestimated by the model. Although the overland flow and the sediment yield simulated were, respectively, 8,81% and 30,02% higher than measured data, the results allowed analysis of the spatial variability of sediment yield and to locate sediment source areas in the watershed.