Capability of convection–dispersion transport models to predict transient water and solute movement in undisturbed soil columns

The capability of the convection–dispersion model coupled with the Richards equation to predict transient transport of water and solutes in porous media is studied in a laboratory soil column. The solutes used are 3H2O as a conservative tracer and 14C-labeled dibutylphthalate (DBP) as a sorbing solute. To account for multiple nonideal transport processes, a model was developed that may consider hysteresis of the water retention curve, sorption nonequilibrium (sorption kinetics and sorption hysteresis) and mobile and immobile water domains during solute transport. The model parameters were determined independently from the transient experiments, but on the same soil column to avoid uncertainty in transferring results from one sample to another. The hydraulic properties of the soil column were estimated by a multistep outflow experiment using inverse modeling. Transport parameters were estimated form breakthrough curves of 3H2O measured under steady-state flow conditions. The sorption parameters of DBP were determined using batch techniques. Matrix potentials in the soil could not be predicted using a simple hysteresis model. Whereas the prediction of the tracer transport was satisfactory, it was not successful for DBP, inspite taking all nonequilibrium processes into account. DBP showed an early breakthrough and fluctuations in the outflow concentrations, which were induced by changes in the hydraulic boundary conditions. These fluctuations are caused by exchange processes during redistribution. The results suggest that models based on equilibrium assumptions may not always be suitable to predict the observed transport patterns. A calibration of the multinonequilibrium convective–dispersive model with the measured data, however, was possible.