STRATIGRAPHIC ANALYSIS OF GLACIAL SEDIMENTS USING OUTCROP, WELL BORE, AND NATURAL GAMMA RAY LOGGING

The postglacial topography of southern Michigan overlies a complex sedimentary sequence. Pleistocene glacial activity created heterogeneous aquifer systems characterized by variable depositional energy and sediment load, anastamosing outwash flows, and superposed channels. In Ann Arbor, Michigan, USA, an extensive subsurface data set has been developed to facilitate an extensive groundwater remediation project. Borehole logs, natural gamma logs, static water levels, and concentration measurements from more than 120 monitoring wells distributed over 10 km² are available to constrain models of the hydrostratigraphic architecture. In addition, glacial stratigraphy is exposed within a local sand and gravel pit located on the flank of the Fort Wayne Moraine. The purpose of this study is to characterize the sedimentology and stratigraphy of the gravel pit exposures, and to correlate those exposures to the available subsurface data.

Ten stratigraphic sections were measured along the east and south face of the quarry across a composite thickness of approximately 30 meters. Stratigraphic cross sections were constructed using an exposed till contact as datum. The base of the till is sharp, subhorizontal, and laterally continuous across 260 m of exposure. This surface delineates a macroscopic change from the fine-textured till above to coarse, clastic aquifer below and may be correlatable into the subsurface. Below the contact, stratigraphic units are defined by erosive surfaces, abrupt changes in particle size, and 1-6 m thick fluvial fining-upward sequences. Most units contain coarse grained particles and are poorly sorted, ranging from fine sand to boulder size. Occasional lenses of moderately- and well sorted medium grained sand may provide preferential pathways for transmission of groundwater.

Long range objectives of this study are to refine the understanding of dominant large-scale subsurface transport vectors within glacial sediments, and to produce a realistic conditioning set for finer scale stochastic modeling. The macroscopic allostratigraphic correlations developed here will be extended throughout the network of monitoring wells and finer scale observations will be used to help condition geostatistical models of aquifer parameters within the larger hydrostratigraphic framework.