Effective porosity of a carbonate aquifer with bacterial contamination: Walkerton, Ontario, Canada

摘要

Preferential flow through solutionally enlarged fractures can be a significant influence on travel times and source area definition in carbonate aquifers. However, it has proven challenging to step beyond a conceptual model to implementing, parameterizing and testing an appropriate numerical model of preferential flow. Here both porous medium and preferential flow models are developed with respect to a deadly contamination of the municipal groundwater supply at Walkerton, Ontario, Canada. The preferential flow model is based on simple orthogonal fracture aperture and spacing. The models are parameterized from borehole, gamma, flow and video logs resulting in a two order of magnitude lower effective porosity for the preferential flow model. The observed hydraulic conductivity and effective porosity are used to predict groundwater travel times using a porous medium model. These model predictions are compared to a number of independent estimates of effective porosity, including three forced gradient tracer tests. The results show that the effective porosity and hydraulic conductivity values closely match the preferential flow predictions for an equivalent fracture network of ～10. m spacing of 1. mm fractures. Three tracer tests resulted in groundwater velocities of hundreds of meters per day, as predicted when an effective porosity of 0.05% was used in the groundwater model. These velocities are consistent with a compilation of 185 tracer test velocities from regional Paleozoic carbonate aquifers. The implication is that carbonate aquifers in southern Ontario are characterized by relatively low-volume dissolutionally-enlarged fracture networks that dominate flow and transport. The porous matrix has large storage capacity, but contributes little to transport. Numerical models based on much higher porosities risk significantly underestimating capture zones in such aquifers. The hydraulic conductivity - effective porosity prediction framework provides a general analytical framework for a preferential flow carbonate aquifer. Not only is the framework readily parameterized from borehole observations, but also it can be implemented in a conventional porous medium model, and critically tested using simple tracer tests.