Mobile-Immobile Model Simulation of Non-equilibrium Tracer Transport through Undisturbed Soil Columns under Transient Flow Conditions

Physical non-equilibrium transport processes in soil may lead to preferential leaching of contaminants into the ground water system. The application of dual- and multi-permeability models to analyze preferential transport still remains difficult due to the large number of model parameters required that can only be determined with a high level of uncertainty. Mobile-Immobile Models (MIM) require fewer parameters and have been frequently used for simulating non-equilibrium transport under steady-state flow conditions. In this study, we utilize a single region transport model (SRM) and a modified MIM to simulate several tracer transport experiments under transient flow conditions in undisturbed soil columns. Analyses of the transient experiments were performed using inverse parameter estimation algorithm of modified HYDRUS (Simunek et al. 2002). Bulk soil water contents, pressure heads at discrete depths, cumulative outflow and effluent solute concentrations were included into the objective function. The results revealed that for mild non-equilibrium transport in columns of Ap and Bv-horizons, the MIM performed only slightly better than the SRM. For transport under highly non-equilibrium preferential flow conditions (Ah columns), the MIM generally performed better than the SRM. An important observation in our study is that the MIM could not predict solute transport under transient flow conditions when using parameters obtained from the steady-state experiments. Moreover, simultaneous fitting of water and solute parameters proved to be somewhat better than sequential fitting. Similarly, using a water transfer term based on water content gave better results as compared to using a water transfer term based on pressure head. Overall, the MIM could reproduce physical non-equilibrium transport processes under transient flow conditions satisfactory, except for a tendency to over-estimate the observed cumulative tracer loss.