ROLE OF HIGH-RESOLUTION DATA SETS TO IMPROVE SITE CONCEPTUAL MODEL DEVELOPMENT FOR A CONTAMINATED SEDIMENTARY ROCK SITE IN SOUTHERN, WISCONSIN

There are thousands of sites in the U.S. where dense non-aqueous phase liquids (DNAPLs) have degraded groundwater resources. Comprehensive site conceptual models (SCMs) are foundational to effective long term management and evaluation of remedial options for these sites. Characterization of the flow system (i.e., delineation of hydrogeologic units (HGUs), identification of boundary conditions, quantification of hydraulic gradients) is a critical component of any SCM. This presentation shows a unique approach to flow system characterization at a site in southern Wisconsin where an estimated 72,700 L of mixed organics as DNAPLs were released into the subsurface prior to 1970. The DNAPL penetrated to about 56 m bgs, accumulated in a fractured sandstone and formed a 3 km long dissolved phase plume. The approach involves characterizing the three-dimensional distribution, magnitude and direction of the vertical component of hydraulic gradient using high resolution head profiles (> 3 monitoring zones/10 m) obtained from Westbay multilevel systems (MLSs). The MLSs were specifically designed to minimize cross-connection and were installed at 7 locations along two 4 km long cross-sections to depths between 90 and 146 m bgs. The vertical gradient profiles showed depth discrete changes in the direction of the vertical gradient, thin intervals of large vertical gradient (some >1 m/m) and thicker intervals of no to minimal vertical gradient and 11 laterally extensive contrasts in vertical gradient indicative of contrasts in the vertical component of hydraulic conductivity (K_v). The boundaries of the K_v contrasts did not coincide with lithostratigraphic units, but further research showed that the K_v contrasts were closely associated with sequence stratigraphic units (maximum flooding intervals, unconformities). Integrating the K_v contrasts and sequence stratigraphy allowed for 3-D delineation of HGUs at the plume scale. The robustness of the SCM suggests these high resolution data sets are critical to the construction and calibration of numerical models used to predict contaminant transport. These results also highlight the importance of well construction informed by accurate characterization of HGU boundaries to minimize cross-contamination from shallow sources into deep water supply aquifers.