Robust Prediction of Kd from Soil Properties for Environmental Assessment

Soil solid/liquid partition coefficients (Kd) are commonly used in quantitative environmental assessments as a means to predict retention of contaminant metals and radionuclides in soils. There are Kd data for most elements in the literature, and they can be measured for specific sites. However, there is also a need for robust, defensible relationships to predict the variation in Kd as a function of soil characteristics. This article reports relationships for As, Cd, Ce, Cl, Co, Cr, Cs, Cu, Fe, Ho, I, La, Mn, Mo, Nb, Nd, Ni, Np, Pa, Pb, Pu, Ra, Sb, Se, Sm, Sn, Sr, Tc, Th, Tl, Tm, U, W, and Yb, most based on a wide range of soil properties. Soil pH, clay content, and organic carbon content were the independent variables considered in all cases, and methodology variables were invoked for Cs, Mn, and Ni. The underlying Kd data were considered representative of steady state conditions: most were measured by desorption of the indigenous element at realistic field moisture contents. The equations, developed by forward or backward stepwise regression, were all statistically highly significant. Estimates from the equations were compared to data for seven site-specific soils and generally were within 95th percentile bounds.