Estimating WEPP Cropland Soil Erodibility from Soil Properties

Abstract

In the late 1980s, the USDA Agricultural Research Service along with other federal agencies and multiple universities collaborated to develop a new physically based soil erosion model, the Water Erosion Prediction Project (WEPP) Model. The WEPP model was intended to replace the Universal Soil Loss Equation and was to include estimates for upland runoff and erosion, sediment delivery to first order channels, and runoff and sediment routing through a downstream channel network. The WEPP technology estimated erosion from raindrop splash and sheet flow (interrill erosion) and concentrated channel flow (rill erosion). To make these erosion estimates, WEPP required new soil erodibility values for interrill erodibility (K_i), rill erodibility (K_r), and critical shear (𝜏_c) for concentrated flow erosion. The WEPP Core team determined that they needed to estimate these three erodibility values from measurable soil properties for a wide range of soil conditions. To develop relationships between WEPP soil erodibility variables and other soil properties, a field study was carried out using rainfall and runoff simulation to measure the three erodibility values for 36 soils. Sites were identified on USA croplands from Washington to Georgia and Maine to California for erodibility measurements. Concurrently, the USDA Soil Conservation Service (SCS) carried out detailed soil surveys and laboratory analyses for all sites to provide a large database of soil physical, chemical, and engineering properties. Correlation and regression analyses were carried out to develop relationships between SCS-measurable soil properties and WEPP soil erodibility values. This presentation provides a summary of those analyses, and subsequent predictive equations that were developed. The predictive equations that were finalized in the WEPP User Summary used sand, very fine sand, clay, and organic carbon contents to predict cropland soil erodibility, but the Coefficient of Determination (r²) values were 0.55 or less. More complex predictive equations were developed with soil physical, chemical, mineralogical and geomorphic properties with r² values up to 0.81. Most of the better predictive equations included terms for soil texture, and clay mineralogy, often with additional chemical properties. A set of simplified equations are proposed using only the readily available properties of soil texture, organic carbon, cation exchange capacity, slope steepness, and taxonomic order for use within the WEPP Model. The simplified equations are: <fig><graphic xlink:href=23010v_files/23010v-00.jpg id=ID0EIPAC></graphic></fig> where K_i is interrill erodibility (kg s m^-4), K_r is rill erodibility (s m^-1), τ_c is critical shear (Pa), M is Wischmeier and Smith‘s (1978) soil textural property (Percent²), OC is the soil organic carbon (Percent), VfSa is the soil very fine sand content (Percent), Silt is the soil silt content (Percent), Slope is the location slope steepness (Percent), Clay is the soil clay content (Percent), and CEC is the soil cation exchange capacity (MEQ/100g).