Modelling of the Transport of a Reactive Contaminant in Spatial Variable Soil Systems

Abstract

The mathematical description of the transport of dissolved contaminants (solutes) in porous media is, in chemical engineering, soil science etc., usually based on solving the transport equation, with the appropriate initial and boundary conditions. For one-dimensional transport in direction z (depth) and neglecting production terms and concentration induced fluid density gradients this equation is:

$$ \frac{{\partial F}}{{\partial t}} + \theta \frac{{\partial c}}{{\partial t}} = \theta D\frac{{{\partial ^2}c}}{{\partial {z^2}}} - v\theta \frac{{\partial c}}{{\partial z}} $$

(1)

be bz (symbols: see appendix). The parameter assessment is generally done by transport experiments in packed soil columns and comparison of breakthrough with results obtained by solving equation (1) (cf Van der Zee et al, this conference). Extrapolation of results found for homogeneous columns to natural systems is often difficult, due to the observed, scale-dependent heterogeneity of naturally occuring porous media (De Haan and Van Riemsdijk, this conference). Depending on the correlation distance of the porous medium at the scale (travel distance) of interest, generally the heterogeneity is accounted for by identifying several sub-domains (mobile/immobile water) or by increasing the value of D in equation (1). However, considering vertical transport in a field, the field-averaged dispersion process for a conservative solute is controlled by the spatial variability of the hydraulic characteristics, and this leads to a non-sigmoid solute front, instead of the sigmoid front usually found if equation (1) is solved (a review is given by Bresler et al., 1982, further reference given there).