DRAWING THE DOG: DETERMINISTIC HYDROSTRATIGRAPHIC MODELING OF A COMPLEX GLACIAL AQUIFER SYSTEM USING AN ALLOSTRATIGRAPHIC APPROACH

Complexity and glacial sediments go hand in hand. Consequently, modeling three-dimensional variability in glacial aquifer systems can require considerable experience, perseverance, and imagination, even when abundant subsurface data are readily available. In practice, subsurface information is usually sparse, and evaluating model uncertainty constitutes a significant challenge. This study employs hybrid models incorporating stochastic variability within a deterministic hydrostratigraphic framework to model spatial variability of physical hydrogeologic properties and assess contaminant transport prediction uncertainty in a complex glacial aquifer system.

The approach is illustrated with a case study in Ann Arbor, Michigan, USA, where plumes of groundwater containing 1,4-dioxane have migrated several kilometers in different directions through 80m of underlying glacial drift. The deepest known plume appears to be advancing toward the Huron River beneath a groundwater Prohibition Zone established in 2005. 1,4-Dioxane is readily soluble in water but resistant to microbial degradation and adsorption to soil particles. Thus, it provides a tracer-like record of solute transport. More than 130 monitoring wells and 20 extraction wells have been drilled to detect, trace, and remediate 1,4-dioxane in the area. These wells form the basis of an allostratigraphic interpretation of the three-dimensional distribution of aquifer and aquitard units, constrained by available hydraulic head and contaminant concentration data, that is described in this presentation. Stochastic modeling of aquifer and aquitard properties within that deterministic hydrogeologic framework is described in a companion presentation.