Estimating the Water-Dispersible Clay Content of Soils

Soil erodibility is influenced by several soil properties including the extent to which the clay fraction will disperse in water. Because early methods for estimating soil erosion were empirical methods and did not utilize water-dispersible clay as a parameter, few data have been collected. The recent development of the Water Erosion Prediction Project (WEPP) model, a process-based model for predicting water erosion that uses water-dispersible clay in the algorithm for computing interrill erodibility, resulted in an increased demand for these data. In order to accommodate this and similar models, a method for estimating the water-dispersible clay content of soils based on existing information is needed. Data collected by the National Soil Survey Laboratory in support of the WEPP were used to identify soil properties that were significantly correlated with water-dispersible clay and to develop equations to estimate the water-dispersible clay content of soils based on those properties. The property most strongly correlated with water-dispersible clay is total clay. Other properties significantly correlated with water-dispersible clay are the water content at 1.5 MPa, dithionite-citrate-extractable Fe and Al, the coefficient of linear extensibility, Wischmeier's M, the very-fine-sand content, the ratio of cation-exchange capacity (CEC) to total clay, Bouyoucos' clay ratio, and the CEC. A simple linear regression of water-dispersible clay vs. total clay revealed that, for the soils included in this study, approximately one-third of the total clay was water dispersible. However, the model only had an R² of 0.604. When the ratio of the CEC corrected for organic carbon (CCEC) to total clay was included in the model, the R² improved to 0.723. However, sorting the data by the ratio of CCEC to total clay instead of including it in the model improved the overall fit of the model and increased the R² to 0.879. Contribution from the USDA-ARS in cooperation with the USDA-SCS and the Agricultural Research Div., Univ. of Nebraska, Lincoln. Published as Journal Series no. 9674.