Location: National Soil Erosion ResearchTitle: Process-based soil erodibility estimation for empirical water erosion models Author
Submitted to: Journal of Hydraulic Research IAHR
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 2/20/2017
Publication Date: 6/21/2017
Citation: Deviren Saygin, S., Huang, C., Flanagan, D.C., Erpul, G. 2017. Process-based soil erodibility estimation for empirical water erosion models. Journal of Hydraulic Research IAHR. doi:10.1080/00221686.2017.1312577. Interpretive Summary: Soil erodibility is a term used to quantify how erodible a soil is. As erosion models evolved from empirically based, such as the Universal Soil Loss Equation (USLE) and the Revised USLE (RUSLE), to process-based, such as the Water Erosion Prediction Project (WEPP) model, the erodibility term also expanded from one, i.e., USLE/RUSLE-K, to three terms, i.e., interrill erodibility, rill erodibility and critical shear stress. Since the WEPP erodibility factors can be derived quickly from lab or field experiments while USLE/RUSLE-K requires many years’ natural runoff data, there is a need to derive USLE/RUSLE-K from WEPP erodibility factors. Using two soils from Washington State, we found that these two soils have distinctly different WEPP erodibility values while the USLE/RUSLE-K are identical, indicating the empirical model is not sensitive enough to reflect how soil properties may have affected the erodibility. We also found that soil moisture condition and management history (by comparing results between those obtained in 1987/88 and from current study) affected process-based erodibility values. In order to accurately assess soil erosion, we need to treat soil erodibility as a dynamic factor to include the soil moisture effects, and be able to account for its change through cropping and management history.
Technical Abstract: A variety of modeling technologies exist for water erosion prediction each with specific parameters. It is of interest to scrutinize parameters of a particular model from the point of their compatibility with dataset of other models. In this research, functional relationships between soil erodibility equations of the USLE/RUSLE and WEPP models were sought for by using new data sets from laboratory experiments to fill a gap stemming from the conceptual differences of the two models in estimating the soil erodibility. In this context, soil erodibility potentials of two different soils collected from WEPP cropland erosion sites in the U.S. Pacific Northwest were quantified under laboratory conditions for three different initial moisture conditions, and the comparisons between “the process-based WEPP-K terms”, namely interrill (Ki) erodibility, rill erodibility (Kr) and critical shear stress (tc), and “the empirically based USLE/RUSLE-K term” were performed. A combination of lab experiments with subsequent WEPP model simulations was carried out to obtain process-based (R)USLE-K values. The results were also compared with RUSLE2-K and the WEPP-K values for interrill erodibility (Ki), rill erodibility (Kr) and critical shear stress (tc) derived from field experiments between 1987 and 1988 in the same soil series. The results indicated that RUSLE2-K was in completely different order of magnitude and 25.78 and 7.10 times higher than the WEPP model K values of the studied soils. When different hydrologic conditions were considered, back calculated process based Dry-K value was found to be more capable to reflect the erodibility potentials of soils in terms of process based approach since it has not shown significant difference from the WEPP field expt-K. Statistically significant differences occurred in terms of the effects of surface hydrologic conditions on soil erodibility, and a back-calculated process based (R)USLE-K value under dry conditions was found comparable with original WEPP-K values in magnitude. Thus, since the effects of surface hydrologic condition on soil erodibility is not considered by the prediction technologies of RUSLE and WEPP models, the approach and the results of this research could be very useful for tapping into a large number of datasets available and building up the next generation process-based erosion models.