Soil water content effect in the USLE/RUSLE model to estimate erosion in Uruguay

Abstract

USLE/RUSLE has been tested and used in Uruguay since 1980, but erosion is overestimated when compared with experimental long-term data. It is hypothesized that this is due to not considering the effect of soil water content. The objectives of this work were to prove the hypothesis and to develop a procedure to include the effect of SWC into the erosion estimation with the model. Calculating the RUSLE soil water content subfactor, with experimental data, and including it in the model, the linear regression coefficient between estimated and measured data changed from 1.21 to 0.99 and R² from 0.93 to 0.97. The best soil water balance procedure to estimate SWC was the one of Témez (1977) calibrated in Uruguay by Genta et al. (2005). Using this balance estimates to calculate the CA subfactor and including it in erosion estimations with USLE/RUSLE, soil erosion overestimation changed from 25 % to 4 %, and R² , Nash-Sutcliffe index and PBIAS changed from 0.94 to 0.97, 0.73 to 0.97 and -56 to 4, respectively.