Evaluation of a generic multisurface sorption model for inorganic soil contaminants.

The performance of a multisurface sorption model approach, composed of well-accepted surface complexation models in combination with published "generic" parameter sets, is evaluated for its possible use in risk assessment. For that purpose, the leaching of a broad range of potential soil contaminants (Ni, Cu, Zn, Cd, Pb, Ba, Cr, Co, Mo, V, Sn, Sb, S, As, Se) from eight diffusely and industrially contaminated soils is predicted simultaneously without any parameter optimization. The model approach includes aqueous speciation in combination with sorption to organic matter (NICA-Donnan model), Fe/Al(hydr)oxides (Generalized Two-Layer Model), and clay (Donnan model). Independent data generated by pH-static leaching experiments, performed with individual subsamples over a wide pH range (pH 0.4-12), provide a sensitive evaluation of the model performance. Root-mean-squared error values between predicted and measured log concentrations over the entire pH range, RMSE(log), are < 0.5 for Cu, Ni, Cd, Co, S, and Se, and RMSE(log) < 1 for Zn, Ba, Cr, Pb, Sn, Mo, Sn and As. The approach without parameter optimization has led to recommendations for further research with particular emphasis on identification of leaching mechanisms for Pb, Mo, Sb, and V and further expansion of the data sets to reduce the uncertainty of the available generic sorption parameters for Sn, Sb, Ba, Cr, and V.