SIMULATION AND AREAL INTERPOLATION OF REACTIVE SOLUTE TRANSPORT

Reactive solute transport in a spatially variable field is studied. Adsorption is assumed to be described by the nonlinear modified Freundlich isotherm with an adsorption coefficient that depends on pH and organic matter content. Assuming spatial variation of pH and organic carbon content, a chemically heterogeneous field is simulated by the generation of a random field of the Freundlich adsorption coefficient. A three-dimensional reference transport calculation is performed with the complete heterogeneous field as input. The effect of the sequence of calculations and interpolations, and of the size of the data set are examined. A CI (calculations before interpolations) and an IC (interpolation followed by calculations) procedure are defined. Attention is focused on the prediction of average solute breakthrough and on the prediction of concentrations at specific locations. The results show that the optimal procedure and data set size depend on the quantity of interest. For the prediction of average behaviour, a smaller data set is sufficient compared with the prediction of concentrations at specific locations. For the calculation of the average solute breakthrough a more efficient IC procedure can be used in combination with the smaller data set. If the concentrations at specific locations in the field are requested, the size of the data set is more significant. However, for relatively small data sets a CI procedure performs better than an IC procedure.