Modeling of Non-Aqueous Phase Liquid Migration and Recovery in Heterogeneous Aquifers

Abstract

In recent years, researchers have attempted to incorporate soil variability into models that predict the fate of environmental pollutants in the subsurface, and the emphasis in modeling has shifted from a deterministic to a stochastic approach. Numerous authors (for example, Gelhar 1986; jury et al. 1987; Dagan 1990; Russo 1991) have developed and reviewed the theory of subsurface stochastic flow and transport. Stochastic theory is based on the assumption that a heterogeneous soil property can be treated as a single sample, or realization, taken from a random field. Quantification of spatial variability is generally limited to estimating the mean and variance of the field variable, which is said to be second order stationary if its mean and variance are constant throughout the domain. In general, only one measurement of the field variable is available at a given point in space, but estimates of the mean and variance may still be obtained under the ergodic hypothesis (Lumley and Panofsky 1964; Sposito et al. 1986).