Functional evaluation of linearized Langmuir equations to characterize cadmium sorption and transport in selected calcareous soils

Abstract

The objective of this study was to compare the performance of nonlinear (NL) and linearized Langmuir models to fit the experimental data of cadmium (Cd) sorption to soils. Cd-sorption data, collected from 19 soils, were fitted to the NL and four linearized Langmuir models. Functional evaluation of the models was provided to compare Cd transport simulated by HYDRUS-1D. Furthermore, it was determined if sorption parameter values obtained either by NL fitting of the Langmuir model or by linearized alternatives would affect simulation of Cd transport in soil .The NL fit of the Langmuir model performed best. The best linearized model was subjective, either depending on the similarity between obtained sorption parameter values compared with the NL model or the variations in the prediction error in Cd sorbed to the soils. The magnitude of error varied at different concentrations, and the best model in terms of its prediction accuracy was dependent on the concentration (C) at which was needed to estimate the Cd sorption to the soil (S). The simulation of Cd transport in soil, using the values of sorption parameters obtained via either fitting the NL or linearized Langmuir models, showed that at a specific soil extract concentration, 1.24 mg Cd L−1, a linearized model demonstrated simulations most like that of the NL model. The linearized model with the lower prediction error was different from the linearized models that were most similar to the NL model to simulate Cd transport in soil. Furthermore, the linearized model with the most comparable parameter values to the NL model did not demonstrate a lower prediction error. The findings suggest that it is vital to generate sorption data over the concentration range of interest for specific applications.