Comparison of Three Water Erosion Prediction Methods (137Cs, WEPP, USLE) in South-East Brazilian Sugarcane Production

Abstract

Soil erosion is the main degradation process in tropical agroecosystems. Soil erosion rates should be considered in land evaluation and conservation planning assessment. The methods available for erosion prediction are not sufficiently calibrated or validated for tropical soils, climates and crops. Thus, differences in estimated soil erosion values may be expected, even if considering the same input data. Three soil erosion estimation methods (USLE, WEPP and 137^Cs) were applied to the same watershed cultivated with sugarcane in Southeastern Brazil (near Piracicaba S22°38'54" and W47°45'40"). The absolute erosion rate values and the differences in the spatial distribution were evaluated. The overall results suggested that there are important differences in soil loss estimated by the three methods. The differences occurred in both, mean values and geographic locations. The sequence of mean soil loss values was USLE>>137^Cs>WEPP and Standard deviation values USLE>WEPP>,137^Cs, indicating that USLE predicted the highest erosion values and spread out over the widest range. The poor geographical coincidence of the results is evidence that the values resulting from none calibrated soil erosion methods should be considered only as qualitative indications. The method selection should consider overall site variability in relation to known sensitive method factors.