Flow In Highly Heterogeneous Aquifers: Coping With Uncertainty
Abstract
We discuss a random domain decomposition approach to analyze flow through highly heterogeneous, composite porous media that greatly improves estimates of pressure head statistics. Composite porous media consist of disjoint blocks of geologic ma- terials, each block comprising a single material type. Within a composite medium, hydraulic conductivity can be represented through a pair of random processes: (i) a boundary process that determines block arrangement and extent and (ii) a random process that defines conductivity within a given block. To demonstrate the advantages of this model, we analyze two-dimensional flow in a layered heterogeneous medium composed of two materials whose hydraulic properties and geometries are uncertain. For simplicity, hydraulic conductivities of both materials are treated as statistically ho- mogeneous log-normally distributed random fields. The location of the internal bound- ary between the two layers is treated as a normally distributed random variable. We consider two flow scenarios, (i) parallel and (ii) perpendicular to the layering. In both cases, the hydraulic head and flux statistics obtained from the moment equations for composite media model are virtually indistinguishable from those obtained by Monte Carlo simulations. We conclude by contrasting our model with the existing determinis- tic trend models and with a statistically homogeneous model, which ignores composite nature of the medium.
- Publication:
-
EGS General Assembly Conference Abstracts
- Pub Date:
- 2002
- Bibcode:
- 2002EGSGA..27.1957T