HR: 12:10h
AN: H32A-08    [Abstracts]
TI: Flow dimensions corresponding to stochastic models of heterogeneous aquifers
AU: * Walker, D D
EM: ddwalker@uiuc.edu
AF: Illinois State Water Survey, 2204 Griffith Drive, Champaign, IL 61820 United States
AU: Cello, P A
EM: cello@uiuc.edu
AF: University of Illinois, Urbana-Champaign, Hydrosystems Laboratory 205 N. Mathews, Urbana, IL 61801 United States
AU: Valocchi, A J
EM: valocchi@uiuc.edu
AF: University of Illinois, Urbana-Champaign, Hydrosystems Laboratory 205 N. Mathews, Urbana, IL 61801 United States
AU: Loftis, B
EM: bruce@ncsa.uiuc.edu
AF: Rosen Center for Advanced Computing, Purdue University, 302 W. Wood Street, W. Lafayette, IN 47907 United States
AB: Traditional approaches to characterization and modeling of highly heterogeneous aquifers faces many technical challenges. One alternative strategy begins with the Generalized Radial Flow interpretation of hydraulic tests, which infers an additional parameter, the flow dimension, to describe the complex geometry of groundwater flow. The flow dimension can be inferred from standard hydraulic tests by examining log-log diagnostic plots of drawdown versus time. This study examines the behavior and variability of the flow dimension, n, for several stochastic models of heterogeneous hydraulic conductivity, K(x). This is accomplished through Monte Carlo analysis of numerical models simulating aquifer tests in two-dimensional systems. For ln K(x) distributed as a multivariate Gaussian variable up to a variance of 1.0, the apparent flow dimension of an aquifer test converges to n = 2 if the scale of the test is large relative to the scale of correlation. The variability of the apparent flow dimension depends on the variance and integral scale of hydraulic conductivity, suggesting that the variance and integral scale of an aquifer might be identifiable from a set of aquifer tests. For ln K(x) with variances greater than 1.0 the results suggest that the mean of the apparent flow dimension may be less than two initially, then converges to n = 2. For ln K(x) distributed as fractional Brownian motion (fBm), the apparent flow dimension averages to n = 2 and its variability increases with time. A percolation network model showed high variability of flow dimension among individual realizations, but the mean apparent flow dimension is between 1.4 and 1.6, with an increasing trend. The trend and the stabilized value apparently are functions of the contrast between the percolating and nonpercolating conductivities. These results suggest that the flow dimension may be a useful diagnostic for selecting models of heterogeneity.
DE: 1828 Groundwater hydraulics
DE: 1869 Stochastic hydrology
DE: 4440 Fractals and multifractals
DE: 5114 Permeability and porosity
SC: Hydrology [H]
MN: Fall Meeting 2005