OBJECT-ORIENTED STOCHASTIC MODELING OF THE PASKAPOO FORMATION BEDROCK AQUIFER SYSTEM

In the prairie region of Alberta, surface waters are fully allocated, and groundwater is a potentially under-utilized resource. To evaluate groundwater resources under various alternative future scenarios (e.g., climate change, increasing water use), fully-coupled process-based hydrologic models are being developed. An archetype for prairie hydrogeology, the West Nose Creek Watershed, was selected for intensive monitoring and modeling. The primary aquifer in the region is a Tertiary sandstone-shale complex (the Paskapoo Formation). Preliminary groundwater modeling indicates that the Paskapoo aquifer system cannot be modeled as homogeneous and isotropic. It is of fluvial origin and was likely deposited by an anastomosing river system. Anastomosing systems are characterized by well-preserved stacked channel sands with high lateral stability and thin laterally extensive sand splays. These sands are encased in overbank flood deposits. The resulting bi-modal lithologic distribution of sandstone and shale is even discernable in many low-quality drilling logs. Assuming that a dominant west-east paleo-flow direction controlled the orientation of channels, the depositional environment explains the apparent anisotropy of the aquifer system. The depositional environment hypothesis will be statistically evaluated (inverse modeling of outcrop data, ~7000 low quality water well driller logs, and ~6 highly-characterized high-tech geophysical boreholes), then realizations of hydraulic parameters for hydrologic modeling will be generated (forward modeling). A state of the art object-oriented modeling program (IRAP RMS) developed by the petroleum industry will be used to generate realizations that will be coupled with a hydrologic model (Hydrosphere). Since the eventual goal is to transfer knowledge from the study basin to the entire Prairie Region of Alberta, all tasks are being completed with attention to streamlining the data-processing and efficient coupling between models.