See more from this Session: Symposium--Reactive Transport Modeling In Soils: II
Wednesday, November 3, 2010: 1:25 PM
Long Beach Convention Center, Room 202B, Second Floor
Reactive transport simulations provide a systematic framework for integrating hydrologic and biogeochemical conceptual process models into a quantitative description of subsurface behaviors. However, subsurface environments are open and complex and subject to multiple interpretations and conceptualizations. The approach of this research is to postulate multiple plausible conceptual models, calibrate each model to observations and then make predictions with the calibrated model ensemble. Parametric and conceptual model uncertainty of the alternative models is jointly evaluated using a maximum likelihood formulation of Bayesian Model Averaging (BMA). The BMA method is being applied to small-scale tracer test results to test the hypotheses that conceptual model uncertainty dominates parametric uncertainty and that BMA improves predictive performance. The tracer tests were conducted in a shallow alluvial aquifer at the Naturita UMTRA site that is contaminated with hexavalent uranium (U(VI)). Two tracer tests were conducted by extracting uranium contaminated groundwater and then either increasing or decreasing the alkalinity of the pumped groundwater and finally injecting the solution into the aquifer. Increasing the alkalinity from 8 to 23 meq/L initially caused the U(VI) to increase from 4 μM to 11 μM which was followed by a decrease in U(VI) to as low as 2 μM and then gradually relaxing to the initial concentration. The experiment with the decreased alkalinity followed the opposite trend with U(VI) concentrations initially decreasing followed by an increase above ambient conditions. The BMA approach is being applied to alternative representations of the U(VI) adsorption reactions, the desorption rate limitations and the ion exchange reactions which affect key major ions such as calcium. These models are being calibrated to selected breakthrough curves for the increased alkalinity experiment and the predictive performance will be evaluated for both other wells sampled during the increased alkalinity experiment and for the experiment with decreased alkalinity.