A new expression of the slope length factor to apply USLE-MM at Sparacia experimental area (Southern Italy)

Predicting soil loss due to water erosion by empirical models is useful to assess the severity of the phenomenon in an area of interest and to predict the effect of alternative soil erosion control practices. The USLE scheme cannot be used at the Sparacia experimental area (Sicily, South Italy) to predict event soil loss per unit plot area, A(e), because experimental data suggest that, generally, A(e) does not increase with plot length, lambda. The USLE-MM scheme uses the runoff coefficient, Q(R), as an additional independent variable in order to develop an empirical model allowing prediction of storm soil loss values that do not necessarily increase with lambda. According to this model, A(e) is expected to increase with lambda when the rest of factors do not vary between plots of different length, but an inverse relationship between these two variables can be predicted when Q(R) decreases as X increases, which is a common occurrence at the study area. Plot length was found to be enough to predict the mean Q(R) for a set of erosive events. Testing the model with an independent data set supported its applicability for predictive purposes. Collecting additional plot data is necessary to develop simple plot Q(R) predictive procedures at the event temporal scale. (C) 2011 Elsevier B.V. All rights reserved.