From hillslopes to watersheds: Variability in model outcomes with the USLE

The Universal Soil Loss Equation (USLE) has been the de-facto standard for soil erosion management studies since its seminal publication in the 1970's. Widespread use of the model is due, in part, to its simple empirical modelling structure and parsimonious parameterization; however, these benefits have also led to criticism of its relevance as a tool for estimating soil erosion. While the USLE has a strong empirical basis, it is regularly used beyond its intended design space, i.e., predicting soil loss from planar hillslopes, to predict distributed soil erosion rates at large spatial extents, which introduces uncertainty in model outcomes. In this paper, we use a case study for up-scaling the USLE to a large spatial extent to assess the variability in model outcomes from different model user's design choices. Our analysis demonstrates that a standardized and accredited methodology for up-scaling the USLE is needed to reduce uncertainty in model outcomes.