Reactive transport modelling of Ammonium: 1D Conceptual modelling and comparison of reactive transport codes

Contamination of NH4+, as main inorganic contaminant, resulted from disposal of ammoniacal liquor at a former coal carbonisation plant at Mansfield, UK. Previous research for evaluating natural attenuation (Davison, 1998; Davison and Lerner, 2000; Jones, 2001; Jones et al., 1998; Jones and Lerner, 2001) revealed strong retardation of NH4+ in the aquifer due to cation exchange with existing cations on the sediment.
After disposal of ammoniacal liquor, NH4+ entered the aquifer and equilibrium took place between NH4+, the existing cations and the available exchange sites. An existing hydraulic gradient keeps flushing the contaminated aquifer with pristine background water. NH4+ desorbs progressively due to new cation exchange equilibrium when the input of NH4+ decreases or eventually stops.
In order to be able to model the reactive transport of NH4+, a reaction module is set up for future 2D/3D reactive transport modelling of NH4+ using PHT3D. In a first approach a conceptual 1D-model example is considered and two multicomponent reactive transport models, PHT3D (Prommer and Barry, 2001; Prommer et al., 1999) and PHREEQC (Parkhurst and Appelo, 1999), are compared. In the 1D-model example, the flushing of ammonium contaminated groundwater by pristine background water is simulated. The included processes are advection, dispersion and cation exchange between NH4+, Ca2+, Mg2+, K+, Na+ and the sediment. As both models are using the same geochemical module, the comparison focuses on the coupling approach between the transport and geochemical modules.
Results show that large numbers of pore volumes (i.e., a nondimensional time parameter where elapsed time is divided by the hydrodynamic residence time (Brusseau, 1994)) are needed to flush the NH4+ off the aquifer sediment. From the comparison of simulation results compiled with the two codes it is clear that an appropriate choice of the reaction step size for the sequential coupling between the transport and geochemistry modules, is a major point for accuracy of the model predictions.