2014 GSA Annual Meeting in Vancouver, British Columbia (19–22 October 2014)

Paper No. 101-4
Presentation Time: 8:45 AM

CHARACTERIZATION OF GROUNDWATER TRANSPORT PARAMETERS IN REGIONAL FRACTURED-ROCK AQUIFERS USING MULTIPLE ENVIRONMENTAL TRACERS AND A DUAL-DOMAIN MODELING APPROACH


SANFORD, Ward E., U.S. Geological Survey, 431 National Center, 12201 Sunrise Valley Drive, Reston, VA 20192, PLUMMER, L. Niel, U. S. Geological Survey, MS 432, Reston, VA 20192, NELMS, David L., U.S. Geological Survey, Virginia Water Science Center, Richmond, VA 23228, SCHLOSSER, Peter, Lamont-Doherty Earth Observatory, Columbia University, 61 Route 9W, Palisades, NY 10964, CASILE, Gerolamo C., Water Resources Discipline, US Geological Survey, Mail Stop 432, 12201 Sunrise Valley Drive, Reston, VA 20192 and BUSENBERG, Eurybiades, U. S. Geological Survey, National Center, MS432, Reston, 20192

A study has been undertaken to characterize the groundwater flow and transport properties of the upper Potomac River Basin (24,000 km2) in the eastern United States. The basin is composed of Paleozoic sandstones, shales, and carbonates in the Valley and Ridge and Appalachian Plateau Provinces, and metasedimentary and Proterozoic metaigneous rocks in the Blue Ridge Province. A flow and transport model of this area was developed assuming that both hydraulic conductivity (K) and effective porosity (ne) decrease with depth. Water levels in 200 wells in the basin were used to constrain values of K for five general rock types and five constants defining the rate of decrease of K with depth. Optimal values of K decreased by 2-3 and 7 orders of magnitude within 30 m and 100 m of land surface, respectively.

Environmental tracers from over 30 karstic springs were collected from within the Shenandoah Valley section of the model area. CFC-113, SF6, 3H and 3He concentrations were measured in most of these samples. Piston-flow age calculations showed disagreement between the tracers, with CFC-113, SF6, 3H/3He values suggestting mean residence times of 15, 7, and 4 years respectively. A MODPATH analysis was conducted on the contributing areas to the springs and combined with the tracer input history. The results yielded a distribution of travel times from days to decades, with a best-fit ne of 0.03 and a median travel time of about one year, but little improved agreement between the tracers.

Finally, a dual-domain porosity calculation was added to each MODPATH pathline in order to investigate the potential effects of the matrix (and local fracture) porosity (nm), the regional fracture porosity (nf), the domain exchange rate, and the decline rate of nf with depth. Results yielded a dramatic improvement in the agreement between simulated and observed concentrations of the various tracers. The optimal simulation yielded a value of nf of 0.16 at the land surface, a mean exchange rate of four years, and an nf that decreased to 0.01 and 0.001 by 10 m and 30 m depth below land surface, respectively. Results were not very sensitive to the value of nm, which was set to 0.20. The dual-domain pathline model also proved to be much more computationally efficient than an equivalent advection-dispersion model, which would require a much finer vertical discretization.