Theoretical Analysis of Errors in Rill Erodibility Parameter Estimation Used in WEPP Model

Abstract

Rill erosion is of great significance in hill-slope erosion. Soil erodibility is a key parameter for soil erosion prediction using WEPP model. The estimation of soil erodibility needed by WEPP was based on the potential (maximum) detachment rate of rill erosion, measured from sediment yield in long rill channels resulted from runoff plots. The relationship between sediment yield and rill length was analyzed in this paper, along with the source of errors in estimation of erodibility. A theoretical approach was described for determining the error of erodibility parameter estimation used in WEPP. Analysis indicated that for a well-defined rill channel, the error of erodibility estimation is mainly from the error of potential detachment rate estimation. Instead of being the average value over the entire rill length, the ideal estimated value of the potential detachment rate is the derivative of sediment load as a function of rill length at the initial position when the length variable equals to zero. Computations revealed that the error in erodibility estimation using existing method in WEPP model could be as high as 90% with the lowest error being grater than 50%. Therefore, it is suggested that both the method and the erodibility data for the WEPP model need to be reexamined and reevaluated. Possible ways for minimizing parameter estimation error was recommended.