Comparison of Four Dispersion Coefficient Models in Mathematical Analysis of Contaminant Transport in Heterogeneous Soils

Media found in nature are heterogeneous and contaminant dispersion coefficient in such media may have spatial variations. In this study، dispersion coefficient is presented in constant and spatially-varied forms. The spatially-varied dispersion coefficient has been defined by means of three types of linear، power and exponential functions and each of them has been entered separately in the transport equation and the obtained models are solved using the initial condition and Dirichlet boundary condition. For comparison of the models results، data of the continuous injection of sodium chloride ions in a 12. 5m column containing different soil textures have been used. Concentration-location curves at different times showed that at early times after the injection dispersion coefficient of the linear model was more compatible with the experimental data and that of the power model was less compatible with them. With elapsing time، results of the power model were better than those of the two other models. In general، with increasing time the similarity between the results of models and experiments was reduced. By fitting dispersion coefficient values obtained from the experiments at different points of the column length، the coefficients of determination for the linear، power and exponential models were estimated as 0. 501، 0. 856 and 0. 714، respectively. This showed that power and exponential models had better estimation than the linear model. At all times، results of the constant dispersion coefficient model had less similarity with the experimental results than the other models.