T41 Innovations and New Frontiers in Hydrologic Modeling II

Sunday, 16 October 2005
1:30 PM - 5:30 PM

TRITIUM TRANSPORT AT THE NEVADA TEST SITE

Paper number: T41-4
Program: Technical Sessions
Time: 2:15 PM – 2:30 PM

Authors

Abstract

Since its advent in 1974, the Radionuclide Migration Project at the Nevada Test Site has spawned several interesting groundwater modeling ventures. Of interest to this research is the Cambric detonation site, where a tracer test was conducted from 1975 to 1991. Burbey and Wheatcraft (1986) built a groundwater/transport model of the Cambric site and at the time of calibration had achieved a good match to the measured data. Since then the predicted concentrations have diverged from the measured concentrations, which appear to exhibit classic heavy-tailed behavior. It was hypothesized that the Fractional Advection Dispersion Equation could accommodate this “heavy tail”; the goal of this research was to test that hypothesis.

In the course of modeling tritium transport and studying the observed tritium concentrations, an apparent mass balance discrepancy was found. At the time when monitoring ceased, 98% of the tritium thought to have existed at the Cambric site had passed through the pumping well. Yet concentrations were still high at that time and actually appear to be leveling off, rather than dropping. A model which succeeds in closely reproducing the entire tritium breakthrough curve requires over 69,000 Curies of tritium, fully 17% more tritium than was though to exist at the site. This suggests that the “heavy tail” is due to either (1) recirculation from the drainage ditch, or (2) underestimation of the source term.

It is concluded that although dispersion does not exhibit Lévy behavior at the Cambric site, fractional-order mobile-immobile transport (MIM) does occur. The Lévy distribution, formulated in the FADE with a fractional-in-space derivative, predicts anomalously early concentrations, which don't occur at the Cambric site. Though the first and peak arrivals can be fit with a traditional retardation factor, a better fit with more realistic parameter values is achieved by incorporating a fractional rate of release from immobile pore spaces—the fractional mobile-immobile model (MIM/FADE).