Greater Confinement Disposal Boreholes

As featured in the inaugural GoldSim newsletter.

Background

GoldSim modelers at Neptune and Company, Inc. have devised a variety of models to aid in environmental decision making. One such model is a contaminant transport and regulatory compliance model of a radioactive waste disposal site at the Nevada Test Site (NTS) near Mercury, northwest of Las Vegas, Nevada. In addition to the classical waterborne transport of radionuclides, Neptune's model includes transport of subsurface materials to the ground surface by burrowing animals (of widely divergent taxa), and the transport of contamination to the surface by plants. Since the desert at the NTS is so dry, biotic transport mechanisms are likely to play a critical role in the movement of contaminants from waste forms at depth to the accessible environment. The greater confinement disposal (GCD) wastes are deeply buried (21 to 36 m below the ground surface), and contaminants are hypothesized to move upward from that depth to near the ground surface in water, advecting with and diffusing within a small amount of interstitial water. The upward flux is inferred from the observed strong gradient of interstitial water potentials from depth to the ground surface. Once the contaminants are within reach of plant roots and burrowing animals, they can be brought to the surface and pose a potential threat to human health through exposure to surface soils. 

The GCD boreholes were designed to isolate wastes to a greater degree than the standard excavated pit or trench used for low-level radioactive waste. The design is simple: using a large augering bit (10 or 12 ft in diameter) on a drilling rig, create a hole in the deep alluvial fan deposits of the Area 5 Radioactive Waste Management Site (RWMS) at the NTS. The total depth of the borehole is about 36 m, and waste packages are stacked in the hole up to a depth of about 21 m. Screened alluvium is used to backfill the spaces around the packages and the top 21 m of the borehole. The great depth to the top of the wastes essentially obviates intrusion by plants and animals, as well as any homesteader who may choose to dig a basement. The only intrusion scenario not protected against (it is essentially impossible to do so) is the inadvertent driller, as is the case with all subsurface disposal technologies to date. 

Conceptual Model of Contaminant Transport

The conceptual model of a GCD borehole is closely patterned after the real thing: the waste is assumed to fill a cylindrical space at the bottom of the hole. This space is surrounded on the bottom and the sides by undisturbed alluvial deposits and on top by the backfill. The backfill is essentially the same material as the original alluvium, except that the larger cobbles have been screened out, the natural imbrication and minor cementation have been lost, and the backfill was apparently not compacted. For the purposes of modeling, however, the backfill is assumed to be equivalent in material properties to the natural alluvium. The material is essentially homogeneous and isotropic, and that's not just a modeling convenience—it really is. 

It can be argued that there is no significant movement of water in the alluvium below a meter or two from the surface until the saturated zone is reached some 250 m below the surface. If this is truly the case, then there is no credible pathway for contaminants to get out of this disposed configuration, shy of direct intrusion. That would be a very short performance assessment! 

To keep things interesting, however, the model assumes the validity of the counterargument: that there could be a very small, very slow upward flux of interstitial water from below the waste horizon to the surface, driven by the existing gradient in matric pressures. That is, it seems that the water content actually decreases from below the wastes to within a meter or two of the surface. If there were enough water to provide a continuous phase of interstitial water, so that advection and diffusion were possible, then perhaps this small upward flux could exist. We assume that it does. 

Water is assumed to move by this slow advective process and by diffusion as well. Since the vertical advection-dispersion has a lateral component, and since diffusion has no preferred direction, a lateral ring of alluvial material is included in the model. The principal reason for this feature is that since the wastes contain contaminants at concentrations above the saturation in water (such as in metallic form) more contamination can find its way out of the container through the sides than just through the top. To ignore this would be to risk underestimating the amount of contamination leaving the waste space. 

If water is capable of transporting contaminants to within a couple of meters of the surface, then they are within reach of the hardy plants and animals that inhabit Frenchman Flat. Plants are limited to grasses and shrubs, and significant animals include rodents, ants, and termites. Plants can bring contaminants to the surface by uptake in the root systems, fixing them in the aboveground parts of the plant, and shedding those plant parts as litter on the ground surface. Animals can transport bulk materials through the excavation of burrows, where materials are assumed to be brought directly to the surface from various subsurface layers, and the burrows are also subject to collapse so that the mass balance of material is preserved. 

GoldSim Model

The modeled subsurface consists of a vertically oriented cylinder of waste and alluvial overburden nested inside another cylinder of surrounding alluvium. Both the interior cylinder and the annular ring are subdivided into several cells, in order to model contaminant transport by advection of water, diffusion in the water phase, advection of alluvium by burrowing animals, and plant-induced transport. 

Contamination brought to the surface by plants is modeled not by moving any contaminated medium, but rather by the simple addition to surface cells (and complementary subtraction from subsurface cells) of predetermined amounts of each contaminant. This transfer of contamination is implemented using a GoldSim Consequence element, once the appropriate amounts of contamination have been calculated based on plant uptake factors and productivity rates. (We have since developed more computationally efficient ways of implementing plant-induced contaminant transport and look forward to the opportunity to build those into this model.)

Transport by animals is done simply by creating an advective connection between upper subsurface cells and the topmost surface soil cell in both the inner cylinder and the surrounding lateral ring. These advective connections move bulk alluvium (part dirt and part water). Complementary connections are made to account for burrow collapse, with materials cascading from the topmost cell to the one below, to the one below that, and so on. 

As much as possible, the parameters defining this model are expressed as probability distributions. This includes, for example, the initial radionuclide inventory, the maximum rooting depths for the various plant types, the shape of burrow densities as a function of depth, plant uptake factors, soil/water partition coefficients, burrow excavation rates, water advection rates, diffusion coefficients, and such. 

Future Work

Neptune and Company, Inc. anticipates adding to this model various dose assessment methodologies, corresponding to different regulatory drivers, based on surface soil and water contaminant concentrations. The model is a management system tool that will support effective long-term management of the NTS low-level radioactive waste disposal facilities. 

Go to Neptune's GoldSim page.

Send comments regarding these pages to John Tauxe.
Last modified: 12 March 2002