Summary
Adsorption isotherms of Cd or Zn in soils of varying composition are often described using generalized Freundlich equations with scaled model parameters. However, this approach fails to describe the competition of several different metal cations for adsorption sites. To overcome this problem, we developed an empirical model based on cation exchange equations with scaled selectivity coefficients for Cd–Ca and Zn–Ca exchange. A large set of experimental data on the sorption of Cd and Zn to soils suspended in Ca(NO3)2, CaCl2, or Ca(ClO4)2 electrolyte solutions was compiled from the literature. The data covered wide ranges in solution concentrations (Zn: 10−8 to 10−2 m; Cd: 10−9 to 10−3 m), pH (3.9–7.7), and soil composition (organic carbon: 2 to 170 g kg−1; clay: 9 to 732 g kg−1). Selectivity coefficients for Cd–Ca and Zn–Ca exchange were scaled by expressing them as functions of organic carbon content, clay content, and solution activities of H+, Ca2+, and Zn2+ or Cd2+. The exchange selectivity coefficients and corresponding scaling factors were determined by fitting to experimental data. The model describes adsorbed amounts of Cd and Zn within a 95% confidence interval of approximately ± 0.5 logarithms (in mol kg−1). The model adequately reproduced adsorption isotherms of Cd and Zn and also the coupled transport of Cd and Ca through a loamy soil. To test the predictive capabilities of the model, we conducted leaching experiments in columns packed with four contaminated soils, which were not part of the calibration data set. Using the generalized model parameters obtained from fitting the literature data, we calculated the concentrations of Zn, Cd, Mg, K, and Ca in the effluent and compared them with experimental results. The observed concentrations of Cd and Zn in the effluent were predicted rather well; that is, predictions were generally within the expected 95% confidence interval obtained for the model. In the most acidic soil an excellent prediction of the elution patterns of Zn, Cd, Mg, and Ca was obtained. The parameters of the generalized model can easily be fine‐tuned to obtain an accurate description of the sorption and transport of Cd and Zn in a particular soil. However, in many practical situations detailed experimental data are lacking. In such cases, the parameters of the model can be used to predict pore water concentrations and leaching of Cd and Zn in contaminated soils.