Mechanistically-Based Field-Scale Models of Uranium Biogeochemistry from Upscaling Pore-Scale Experiments and Models

Effective environmental management of DOE sites requires reliable prediction of reactive transport phenomena. A central issue in prediction of subsurface reactive transport is the impact of multiscale physical, chemical, and biological heterogeneity. Heterogeneity manifests itself through incomplete mixing of reactants at scales below those at which concentrations are explicitly defined (i.e., the numerical grid scale). This results in a mismatch between simulated reaction processes (formulated in terms of average concentrations) and actual processes (controlled by local concentrations). At the field scale, this results in apparent scale-dependence of model parameters and inability to utilize laboratory parameters in field models. Accordingly, most field modeling efforts are restricted to empirical estimation of model parameters by fitting to field observations, which renders extrapolation of model predictions beyond fitted conditions unreliable. The objective of this project is to develop a theoretical and computational framework for (1) connecting models of coupled reactive transport from pore-scale processes to field-scale bioremediation through a hierarchy of models that maintain crucial information from the smaller scales at the larger scales; and (2) quantifying the uncertainty that is introduced by both the upscaling process and uncertainty in physical parameters. One of the challenges of addressing scale-dependent effects of coupled processes in heterogeneous porous media is the problem-specificity of solutions. Much effort has been aimed at developing generalized scaling laws or theories, but these require restrictive assumptions that render them ineffective in many real problems. We propose instead an approach that applies physical and numerical experiments at small scales (specifically the pore scale) to a selected model system in order to identify the scaling approach appropriate to that type of problem. Although the results of such studies will generally not be applicable to other broad classes of problems, we believe that this approach (if applied over time to many types of problems) offers greater potential for long-term progress than attempts to discover a universal solution or theory. We are developing and testing this approach using porous media and model reaction systems that can be both experimentally measured and quantitatively simulated at the pore scale, specifically biofilm development and metal reduction in granular porous media. The general approach we are using in this research follows the following steps: (1) Perform pore-scale characterization of pore geometry and biofilm development in selected porous media systems. (2) Simulate selected reactive transport processes at the pore scale in experimentally measured pore geometries. (3) Validate pore-scale models against laboratory-scale experiments. (4) Perform upscaling to derive continuum-scale (local darcy scale) process descriptions and effective parameters. (5) Use upscaled models and parameters to simulate reactive transport at the continuum scale in a macroscopically heterogeneous medium