A Comprehensive Analysis of Contaminant Transport in the Vadose Zone Beneath Tank SX-109
A. L. Ward, G. W. Gee, M. D. White
Pacific Northwest National Laboratory, Richland, Washington
PNNL-11463 (UC-702), Pacific Northwest National Laboratory, Richland, Wa., 1997.
Summary
The Vadose Zone Characterization Project is currently investigating the subsurface distribution of gamma-emitting radionuclides in S and SX Waste Management Area (WMA-S-SX) located in the 200 West Area of the U.S. Department of Energy's Hanford Site in southeastern Washington State. Spectral-gamma logging of boreholes has detected elevated 137Cs concentrations as deep as 38 m, a depth considered excessive based on the assumed geochemistry of 137Cs in Hanford sediments. Routine groundwater sampling under the Resource Conservation and Recovery Act (RCRA) have also detected elevated levels of site specific contaminants downgradient of WMA-S-SX. The objective of this report is to explore the processes controlling the migration of 137Cs, 99Tc, and NO3- through the vadose zone of WMA-S-SX, particularly beneath tank SX-109.
A series of hypothetical cases were simulated using STOMP, a multi-dimensional flow and transport simulator. The simulations represent the migration of contaminants from a region surrounding tank SX-109. Effects of varying recharge, specific gravity of the leaked fluid, distribution coefficients (Kds), and varying complexity of sediment layering were simulated under known or assumed tank farm conditions from 1964 through the year 2015. Also considered were cases with and without preferential flow, and with and without a water line leak of 946 m3 (250,000 gal) over a 75-day period. Fluid densities of 1.0, 1.4, 1.65 Mg m-3 were simulated to evaluate the effect of wetting front instability. Leak volumes ranging from a 15 m3 (4,000 gal) to 946 m3 (250,000 gal) were simulated. Most of the simulations assumed a leak of 500 m3 (132,000 gal). Imposed on the leaks was a recharge of meteoric water (rain and snowmelt) at steady rates of 100, 10, and .5 mm yr-1. In addition, a variable recharge rate, dependent on annual precipitation, was simulated. A range of Kds was assumed for 137Cs. Values ranged from 0 mL g-1, chosen to reflect the effect of high salt concentration, pH, and high temperature on 137Cs mobility, to 37 mL g-1. A time-dependent Kd was selected to mimic the inverse dependence of sorption on Na+ concentration. The Kd of 99Tc and NO3- was fixed at 0 mL g-1. All but two of the simulations were run for 51 years (1964 through 2015). Two of the simulations for 99Tc and salt (NaNO3) were run for several hundred years to look at peak concentrations of mobile contaminants at the water table.
The simulations illustrate the effect of the key parameters on leak migration in vadose zone of WMA-S-SX tanks. Water flux past the base of the tanks ranged from three to five times the imposed flux because of tank shedding. Gravity fingering, resulting from fluid density contrasts, was demonstrated for the higher density fluids. Failure to include small-scale textural variations resulted in predominantly vertical transport and reduced travel times to the water table. When a sloped layer of coarse sand was incorporated in the model at a depth of 23 m, the simulated 137Cs plumes showed peaks slightly above that depth, which was similar to most of the profiles observed in boreholes that had elevated 137Cs. A Kd of 0 mL g-1 moved the 137Cs unrealistically deep and into the water table, while Kds of .5,3 and 37 mL g-1 and various combinations kept 137Cs from arriving at the water table in 50 yr, except under the most extreme conditions (high density, high recharge, preferential flow). A time-dependent sorption appeared to simulate more realistically the movement of 137Cs than did assuming a constant Kd. An initially low Kd followed by gradually increasing values contained 137Cs within a 40-m depth, with the center of mass at 21.3 m and a variance of 2 m2 about the center of mass.
Recharge rates also had a profound effect on the contaminant migration. Under high recharge rates, 99Tc and NO3- moved rapidly to the water table, arriving within 30 years of the initial leak. Reducing the recharge rate to .5 mm yr-1 prevented 99Tc from arriving at the water table. Given the unavailability of site-specific data and the uncertainty in the contaminant release history, the simulations successfully reproduced the major features of observed contaminant migration in WMA-S-SX, including elevated 99Tc concentrations in the groundwater as well as 137Cs peaks and distributions in the vadose zone below the tanks. Simulations coupled with site data can be used to assess impacts from tank leaks and to evaluate mitigation strategies to prevent leaks. However, model calibration for the immobile contaminants will require more information than can be obtained from spectral-gamma logging of the vertical boreholes.
Error processing SSI file