Geostatistical estimation of parameters for hydrologic transport modeling

Abstract

A significant part of the work of evaluating a geologic formation as a potential repository for hazardous wastes is the modeling of contaminant transport in the surrounding media in the event the repository is breached. The transport equations that are commonly used are deterministic functions. However, because the data can vary within the area being considered, there is a degree of uncertainty associated with the results obtained from the contaminant transport models. There are several ways to incorporate uncertainties into the transport equations, but they assume that distributions and parameters such as variances and covariances are known. This paper discusses the application of geostatistical spatial estimation techniques to estimate quantities used in transport modeling. The techniques are illustrated on data from an electric analog simulation of a two-dimensional ground water system. Geostatistical methods were used to estimate potential and hydraulic conductivity surfaces from data generated from the simulation of the ground water system. Although the two surfaces were highly dependent through Darcy's Law, they were estimated independently. Independent verification of the two surfaces showed that they approximately satisfied the required conservation of mass condition that: ∇ ⋅ v = 0.