Transient transport of reactive and non-reactive solutes in groundwater*1

Abstract

A numerical model capable of predicting the transient changes in concentration levels of a solute along a homogeneous aquifer system is presented. The advection-dispersion equation (ADE) is utilised in predicting the concentration levels for cases of continuous and instantaneous release modes. The Crank–Nicholson equation is employed in the presented finite difference model. The numerical calculations are carried out using the implicit Gauss–Seidel method with over- and under-relaxation coefficients depending on the state of convergence. The correction terms resulting from the removal of zero- and first-order truncation errors in the ADE with a reaction term have significantly improved the performance of the numerical scheme. Comparisons between the numerically predicted concentrations with analytical and measured values were carried out for cases of non-reactive (tracer) and reactive (organic) solutes with continuous injection in homogeneous isotropic soils. The overshooting problems experienced in the numerical calculations are minimised by refining the finite grid size. The analysis of results has shown that the model can produce reliable simulations for cases of non-reactive solutes. While for the case of solutes undergoing adsorption, accurate concentrations can be predicted by adjusting the influent pore water velocity through the use of a retardation factor, which is suitable for aquifers with low organic carbon content and undergoing hydrophobic partitioning.