MODELS OF HETEROGENEITY FOR SHALLOW DOLOMITES IN SOUTHERN LAKE MICHIGAN

The fractured dolomite aquifers surrounding southern Lake Michigan are an important water resource for the Great Lakes region of North America. Characterization of these aquifers is challenging due to high levels of heterogeneity that are poorly suited to traditional approaches to hydraulic testing. An alternative approach to interpreting hydraulic tests is the Generalized Radial Flow approach, which infers a parameter known as the flow dimension to describe the geometry of the flow field during a hydraulic test. Noninteger values of the flow dimension often are inferred for tests in fractured dolomite aquifers, yet subsequent modeling studies typically ignore the flow dimension. Monte Carlo analyses of detailed numerical models of aquifer tests examine the flow dimension for several stochastic models of heterogeneous transmissivity, T(x). These include multivariate lognormal, fractional Brownian motion, a site percolation network, and discrete linear features with lengths distributed as power-law. The behavior of the simulated flow dimensions are compared to the flow dimensions observed for multiple aquifer tests in the southern Lake Michigan region. The combination of multiple hydraulic tests, observed fracture patterns, and the Monte Carlo results are used to screen models of heterogeneity and their parameters for subsequent groundwater flow modeling. The study shows that discrete linear features with lengths distributed as a power-law appear to be the most consistent with observations of the flow dimension in fractured dolomite aquifers.