PNNL

Abstract

Laboratory-scale experiments suggest that the properties of hypersaline fluids may influence transport behavior through an interaction between fluid and hydraulic properties. Yet, the importance of these mechanisms during field-scale transport is largely unknown, thereby limiting accurate prediction of future contaminant migration. To assess the importance of these interactions on field-scale solute transport, tank leaks were simulated by performing a series of injections using solute-free and hypersaline waters, in two consecutive years in the Hanford site in Richland, WA. Starting in May 2000, five 4000-L injections of Columbia River water were made with no-flow periods occurring every 3-5 days. The third injection contained 1000 ppm of Br- and a suite of isotopic tracers. In May 2001, the experiment was repeated with five 4000-L injections of saturated sodium thiosulfate containing 2500 ppm of Cl-. Water content distributions were measured in 32 wells (18 m deep) arranged in a concentric pattern over a 16-m diameter. Water extracts from soil cores were analyzed for anions including F-, Cl-, Br-, NO3-, P2O5, SO-2, and S2O3-. Differences between wetting front location and solute front location were apparent and were dependent on fluid constitution. Resident concentration profiles were generally asymmetric with a large solute mass occurring at 5 - 7 m, and a smaller mass at 9 - 10 m. Multiple peaks were also observed and were coincident with finer-textured lenses embedded in the sand matrix, emphasizing the need to consider local-scale textural discontinuities in the conceptual models of the field-scale transport at the Hanford Site. Results show no evidence of fingering due to fluid properties.