SPATIAL UNCERTAINTY IN PREDICTION OF THE TOPOGRAPHICAL FACTOR FOR THE REVISED UNIVERSAL SOIL LOSS EQUATION (RUSLE)

Abstract

The Revised Universal Soil Loss Equation (RUSLE) is a model widely used to predict soil loss. An importantcomponent of RUSLE is the combined topographical factor (LS), which is the product of the slope length factor (L) and theslope steepness factor (S). It is important to identify the main sources of uncertainty in the LS factor and reduce the uncertaintywhere practical. Moreover, the uncertainty of the LS factor may vary across space, and this spatial uncertainty may requireerror management. For this reason, the spatial effects of slope steepness and slope length should be quantified, and theiruncertainty propagation should be modeled. This article presents a general methodology for spatial uncertainty assessmentof the RUSLE and its application results to the uncertainty analysis of LS as an example. A sequential indicator simulationwas used to develop spatial prediction maps of slope steepness and slope length based on collected field data. The uncertaintydue to slope steepness, slope length, and model parameters were propagated through topographical factor LS using theFourier Amplitude Sensitivity Test (FAST). Spatial variance partitioning was performed to generate error budgets, anduncertainty sources were identified. Slope steepness contributed the largest variance in predicting topographical factor LS,followed by slope length. The variance contributions from the model parameters and measurement errors were relativelysmall. The results provide modelers and decisionmakers with spatial uncertainty information for the purpose of errormanagement.