Improving The Reliability of Soil Erosion Estimates

Measured soil erosion-resistance parameters exhibit large variability (up to several orders of magnitude) not only between different soil types, but also for same or similar soil types. This variability is not only caused by the inherent, spatial variability in soil properties (e.g., texture, density, moisture, and organic content), but also by the different instrumentation and post-processing techniques employed to quantify soil erosion-resistance. Two widely used measurement methods are the Jet Erosion Test (JET) and Erosion Function Apparatus (EFA). Both JET and EFA measurements have shown that the erosion function (that is, relationship between soil erosion rate and shear stress exerted by flow on the soil surface) is often non-linear, which generally indicates differing erosion mechanics along the erosion function. For example, at higher applied shear stresses, aggregates or chunks of soils are detached (mass erosion regime), which are much larger than the detached particles at small excess shear stresses (particle-by-particle erosion regime), resulting in magnitude differences in the rate of detachment. The rate of soil erosion is most commonly calculated using a linear excess shear stress equation, in which the portion of the hydraulic shear stress exerted on the soil that exceeds a soil critical shear stress, is multiplied by an erosion rate coefficient (also called detachment or erodibility coefficient). Post-processing techniques that fit a linear excess shear stress equation to the measured non-linear erosion function can introduce significant variability in the measured critical shear stress and erodibility coefficient values. In addition to the systemic post-processing discrepancies, the JET and EFA use different hydraulics to erode soil samples and different techniques to determine the applied shear stresses. The inherent differences between devices and methods typically leads to dissimilarities in their respective erosion functions.

We conducted JET and EFA tests on silt and silty sand Unified Soil Classification System soil types obtained from the banks along the Lower American River, CA. Using their standard post-processing techniques, distributions of JET and EFA soil erosion-resistance values for each soil type differed significantly. However, accounting for sources of variability during the post-processing stage resulted in distributions of JET- and EFA-derived soil erosion-resistance values that were similar. Further, we used these soil erosion-resistance distributions to conduct a stochastic analysis with the Bank Stability and Toe Erosion Model (BSTEM) of bank erosion along the Lower American River. Calibration of model erosion-resistance values against observed bank erosion showed the distribution of calibrated values for the silt soil type was similar as that measured. However, for the silty sand soil type the distribution of calibrated erodibility values significantly differed from that measured. We recommend that erosion calculations of fine-grained, cohesive soils should be based on measured data that are carefully analyzed to account for variability introduced by instrumentation and soil heterogeneity and match the expected, study erosion regime. Erosion estimation reliability can further be improved by employing a thorough calibration process.