Establishing a Geochemical Heterogeneity Model for a Contaminated Vadose Zone-Aquifer System

The Hanford Integrated Field Research Center (IFRC) is investigating multiscale mass-transfer processes that control seasonally variable concentrations in the 300 Area uranium plume. The plume has displayed remarkable persistence over the past 20 years, and questions remain as to whether causes are hydrologic or geochemical. Key to the understanding and simulation of these questions is information on the spatial distribution of U(VI) contaminant concentrations, and reaction properties that determine solid-liquid distribution. About 750 grab samples were collected during the installation of 35 wells within the 1600 sq m IFRC site. Particle size distribution was measured, and the <2 mm fraction characterized for the following properties: total U, 1000 h bicarbonate extractable U, surface area, and ammonium oxalate- and hydroxyl amine-extractable Fe(III). Adsorption distribution ratios (Kd’s) were measured from synthetic groundwater on bicarbonate extracted sediments that had been washed repeatedly to remove residual bicarbonate. Desorption Kd’s, were measured in eight successive equilibrations with synthetic site groundwater. Correlations between the variables were evaluated, and a geostatistical analysis was performed that included generation of stochastic realizations of the spatial distribution of key properties and variables in the lower vadose zone and upper saturated zone of the IFRC site for reactive transport modeling. We found that high extractable U (> 7.5 µg-U/g of sediment) was localized to middle vadose zone hot spots that did not correlate with grain-size distribution. A secondary maximum of adsorbed U (~5 µg-U/g of sediment) occurred in the lower vadose zone, with concentrations increasing upward to the maximum elevation of the current water table (the “smear zone”). Adsorbed U(VI) was low but detectable in the saturated zone where the plume exists. Monte Carlo analysis was used to estimate the mass of extractable U present in the smear zone and the uncertainty in that estimate. The data and analysis will be used in establishing a spatially variable, kinetic surface-complexation model for the site.