Advective Transport of Interacting Solutes: The Chromatographic Model

Abstract

In the past decade affordable computers have become capable of calculating transport of several chemically and biologically interacting components in porous media. Examples of such components are natural ions, ligands, nutrients and contaminants. Many transport codes have been written, as is evident from reviews (e.g. Kirkner and Reeves 1988; Reeves and Kirkner 1988; Kinzelbach et al. 1989). Numerical difficulties and long computation times have been addressed in various ways, depending on the choice of the geochemical problem. Subroutines, e.g. PHREEQE, MINEQL or other members of the MINEQL family, such as MINTEQ, GEOCHEM or HYDROQL, are called at each time step to establish chemical equilibrium in all cells of the spatial discretisation (nodes; Walsh et al. 1984; Cederberg et al. 1985; Novak et al. 1988; Berninger et al. 1991). When applied to the nodes sequentially, these subroutines need more than 90% of the computation time. Soon excessive time is spent in these subroutines when the number of nodes is increased, unless the computer code has been vectorized, thus being able to establish chemical equilibrium in all nodes at once (Vogt 1990). Lichtner (1992) and Ortoleva and co-workers (1987) went in another direction, avoiding super- or mini-supercomputers by introducing very potent approximations in the transport equations for the case of mineral precipitation and dissolution.