SOLUTE TRANSPORT AND REACTION IN GROUNDWATER

We have designed laboratory experiments to study two basic problems in hydrogeology: How do chemicals mix so that they react in natural porous media and what model best describes large-scale spreading of solutes in the subsurface? These questions are intimately related because large-scale spreading, pore-scale mixing, and chemical reaction may profoundly affect one another.

I will describe how connected flow paths of relatively high permeability affect large-scale solute transport by considering laboratory visualization experiments, field experiments at the MacroDispersion Experiment (MADE) site, and numerical simulations of transport through different spatial textures of hydraulic conductivity. We have found that rate-limited mass transfer between mobile and immobile regions best describes solute mixing and spreading in aquifers with geologically-realistic flow paths, even though the spatial pattern of permeability in these aquifers may have the same conventional statistics as are widely used in stochastic theories that predict Fickian spreading. Thus, the conventional statistics used to describe spatially heterogeneous permeability fields are not sufficient to predict large-scale flow and transport, and information on the connectedness of geologic media may be necessary not only to choose parameters for flow and transport models, but also to choose the form of the transport model.

We have also developed methods that accurately image changing concentrations during reaction in porous media experiments using colorimetric chemical reactions at the Darcy-scale. These methods enable us to distinguish product from reactants and quantify their concentrations throughout experimental porous media chambers by their different light absorption (e.g. a blue product formed from clear reactants). The results show that the conventional coupling of chemical reaction equations with transport equations developed from observations of conservative solutes may either over predict or under predict the degree of chemical reaction. In some cases, incomplete pore-scale mixing reduces chemical reaction, and in other cases chemical reaction creates density fingering, further mixing the reactants, and greatly enhancing reaction. Finally, I will consider how pore-scale chemical mixing may be affected by Darcy-scale heterogeneity.