HR: 13:40h
AN: H23I-01 INVITED [Abstracts]
TI: Field-Scale Hydraulic Conductivity (K) and Mass Transfer at the MADE Site in Columbus, Mississippi: A
Review and Continuing Studies
AU: * Molz, F J
EM: fredi@clemson.edu
AF: Environmental Eng. and Geology, Clemson University,
342 Computer Court, Anderson, SC 29625
United States
AU: Guan, J
EM: jguan@clemson.edu
AF: Environmental Eng. and Geology, Clemson University,
342 Computer Court, Anderson, SC 29625
United States
AU: Liu, H
EM: hhliu@lbl.gov
AF: Lawrence Berkeley Natl. Lab, MS 90R1116, 1 Cyclotron Rd., Berkeley, CA 94720
United States
AU: Zheng, C
EM: czheng@wgs.geo.ua.edu
AF: Department of Geology, University of Alabama, Tuscaloosa, AL 35487
United States
AB:
During the late eighties and early nineties, several natural gradient tracer tests were conducted in a shallow unconfined
fluvial aquifer at Columbus Air Force Base in Mississippi. The aquifer matrix was highly heterogeneous (natural log(K)
variance of about 4.5) and consisted of poorly-sorted to well-sorted layered sandy gravel to gravely sand, with variable
silt and clay content (Boggs et al., 1993). Prior to performing the tracer tests, the aquifer was characterized extensively
using a borehole flow-meter. The resulting tracer plumes were highly elongated with dilute leading edges in the
down-gradient direction, and transport appeared to be advection-dominated. Although there is still some controversy,
reasonably successful simulations of the MADE tracer data have settled on an approximate dual porosity conceptualization of
the aquifer matrix. Throughout the aquifer, high K zones (mobile porosity) are visualized as being in contact with low K
zones (immobile porosity), with mass transfer between the zones governed by an effective mass transfer coefficient B. Such a
transfer coefficient is analogous to the matrix diffusion coefficient Dm used to simulate transport in fractured rock with
diffusion into the rock matrix (Foster, 1975). Recently, experiments and geometrically-based reasoning have been presented,
implying that the effective Dm, like dispersivity, increases with travel distance (Liu et al., 2004; Zhou et al., 2005).
Conversely, other studies based on multiple rate mass transfer between mobile and immobile porosities in granular media
(Haggerty et al., 2004) have indicated that B will decrease with travel distance. Thus in geometrically complex granular
media, like those at the MADE site, two opposing effects may be present. To further study this question, new 3-D
simulations of tritium transport are being performed using flow-meter K data and the measured tritium concentrations at
selected times. Results to date indicate that B generally decreases with scale, but changes will depend on the details of
how the flow and mass transfer process at the MADE site is conceptualized. For example, did the tritium tracer injected
initially all enter the mobile porosity, as commonly assumed, or was a significant portion of it forced into the immobile
porosity? Was fluid in the immobile porosity essentially non-moving relative to mobile fluid, or did both fluid classes move
significantly down-gradient. Alternatively, was tracer simply injected into an overall low K region, from which it slowly
leaked out during the course of the 328 day experiment? Simulation results from different scenarios will be presented and
implications discussed concerning the detailed scale-dependence of B at the MADE Site.
DE: 1828 Groundwater hydraulics
DE: 1832 Groundwater transport
DE: 1847 Modeling
SC: Hydrology [H]
MN: Fall Meeting 2005