HR: 1340h
AN: H23A-1014 [Abstracts]
TI: Direct estimation of hydraulic conductivity using integrated inversion of resistivity data during a
tracer test
AU: * Fowler, D E
EM: dylanf@clemson.edu
AF: Environmental Engineering & Earth Sciences, Clemson University, 340 Brackett Hall,
Clemson, SC 29634-0919, United States
AU: Moysey, S M
EM: smoysey@clemson.edu
AF: Environmental Engineering & Earth Sciences, Clemson University, 340 Brackett Hall,
Clemson, SC 29634-0919, United States
AB:
We present an innovative integrated inversion technique that directly uses surface resistivity measurements to
estimate hydrologic parameters during a tracer test. Determination of hydraulic conductivity (K) and dispersivity
using only measured surface voltage potentials is a noninvasive and cost effective method for subsurface
characterization. Unlike typical hydrogeophysical estimation problems, the need to invert the voltage data to
produce an intermediate resistivity image is eliminated in our approach, thus our problem to be better
constrained than typical hydrogeophysical estimation problems since we use many resistivity data points to
estimate only a few key parameters (compared to the hundreds or thousands of parameters contained in a
geophysical image). In addition, by optimizing for a minimal set of parameters we can eliminate dependence on
prior knowledge of the subsurface that is required to regularize typical inverse problems. Another significant
advantage of our approach is that very few current and potential electrodes are required, allowing for dynamic
characterization of the subsurface during transient processes.
Here we present a synthetic study of a tracer test performed in a homogeneous aquifer under a natural gradient.
During the test, only one current electrode pair and a limited number of potential electrodes placed at the surface
are used to monitor the migration of a saline plume. A coupled forward model for flow, solute transport, and
resistivity developed in COMSOL Multiphysics is used to calculate the concentration and potential distribution for a
given value of aquifer hydraulic conductivity. In practice, the value of K is unknown and our goal is therefore to
estimate it from observed electrical potentials. For this purpose, the coupled flow, transport, and resistivity
models are used to determine the K value that minimizes the difference between the observed and simulated
electric potentials. Using the COMSOL Optimization Lab, we have been able to estimate K to within 0.5% using
only the surface resistivity measurements with an initial tracer concentration of 0.5M. Subsequent sensitivity
analyses indicate that injection concentrations above four times that of the background ion concentration in the
aquifer, set here to 0.0025M, can yield an estimate within half an order of magnitude from the true hydraulic
conductivity thus improving accuracy with increases in concentration. Further sensitivity analyses are being
performed to investigate the signal strength in regards to background noise and tracer depth below the surface.
DE: 0545 Modeling (4255)
DE: 1832 Groundwater transport
DE: 1835 Hydrogeophysics
DE: 5109 Magnetic and electrical properties (0925)
SC: Hydrology [H]
MN: 2007 Fall Meeting