Tritium Tracer Experiment and Modeling

A series of modeling investigations were carried out to evaluate results from a tritium tracer experiment in fractured and highly weathered shale saprolite at Oak Ridge National Lab in east Tennessee. The tracer experiment was started by USGS researchers in 1976 when 50 curies of tritium were added to a shallow well in the saprolite and concentrations of tritium were monitored in the injection well and in an array of downgradient monitoring wells (Fig. 1). Based on calculated fracture apertures and flow velocities (determined using the cubic law), researchers expected the tracer to be flushed through the system within in a few weeks or months. Instead, the plume migrated much more slowly (Fig. 2), apparently due to the influence of matrix diffusion, which causes solute to exchange between the fast-flow regime in the fractures and the relatively immobile pore water in the matrix. The plume was initially monitored for 5 years after the injection and then the system was left unattended for 11 more years before it was re-sampled by the UT - Hydrogeology Research Group and ORNL researchers.

A series of numerical modeling studies were then carried out to investigate plume behavior with both an equivalent porous media (EPM) model and a discrete-fracture/dual-permeability model. The EPM was successful in describing the fast moving leading edge of the plume (up to 0.19 m/day), the slower moving center of mass (0.009 m/day), and the very slowly declining concentrations in the "tail" of the plume. Part of the exercise involved fitting the EPM model to data from the first 5 years of the experiment and then "predicting" concentrations measured during the most recent monitoring visit (Fig. 3). The plume also exhibited a high degree of lateral spreading, which could only be fit by using an unusually high value of transverse dispersivity, approximately equal to the longitudinal dispersivity. There are several possible causes of such spreading, including a high degree of anisotropy and seasonal fluctuations in water table shape. One possibility, variations in the fracture network, was also investigated. This study showed that the presence of a few large aperture fractures, which terminate either in the matrix or by intersecting another fracture can cause a high degree of lateral spreading in a steady state flow system (Fig. 4).

Recent publications and theses resulting from this research include:

• Stafford, P.L., L.E. Toran, and L.D. McKay, Influence of fracture truncation on dispersion: A dual permeability model, J. Contaminant Hydrol., 30(1-2), 79-100, 1998.
• McKay, L.D., P.L. Stafford, and L.E. Toran, EPM modeling of a field-scale tritium tracer experiment in fractured, weathered shale, Ground Water, 35(6), 997-1007, 1997.
• Stafford, P.L., Simulation of a field-scale tritium tracer experiment in a fractured, weathered shale using discrete-fracture/matrix-diffusion and equivalent porous medium models, MS theses, Dept. Geol. Sci., Univ. of Tennessee, Knoxville, TN, 1995.
• Webster, D. A., Results of ground-water tracer tests using tritiated water, at Oak Ridge National Laboratory, Tennessee. U.S. Geological Survey Water-Resources Investigations Report 95-4182, 1996.
  
  Figure 1. Photograph of tracer test site near Burial Ground 4 (BG4) at Oak Ridge National Laboratory in eastern Tennessee.
  
  Figure 2. Map of tritium concentration at different times after the pulse injection.
  
  Figure 3. Measured and EPM model predicted concentrations in well 7: a) up to 16 years after injection; b) up to 50 years after injection.
  
  Figure 4. Simulation showing influence of a single truncated fracture on transverse spreading of tritium plume.