Confined aquifers overlain by till confining units (herein termed till aquitards) provide drinking water to thousands of Minnesota residents. Quantification of the recharge (leakage) rate through till, specifically to buried-valley aquifers of glacial origin, is essential to assess their long-term sustainability for drinking-water supply. The U.S. Geological Survey in cooperation with Iowa State University and other Minnesota agencies conducted a two-year study starting in 2015 to characterized hydraulic and geochemical properties of till aquitards and buried-valley aquifers at two sites in central and northeastern Minnesota within the Des Moines and Superior lobes, respectively. The objectives of the project were to estimate recharge, assess the sustainability, and suggest the contamination potential of underlying buried-valley aquifers. My hypothesis is that vertical recharge (leakage) through the overlying till aquitard is small to the buried-valley aquifers, such that the groundwater withdrawal exceeds the recharge (leakage) rate and unsustainable groundwater mining is occurring. Nineteen piezometers were installed in four nests to depths of 340 ft (104 m) in till and sand and gravel at study sites in Litchfield (Des Moines lobe, New Ulm Formation, Villard Member) and Cromwell (Superior lobe, Cromwell Formation, Automba Member and Aitkin Formation). Hydraulic heads were measured and hydraulic conductivities (K) were estimated with slug tests in order to calculate vertical recharge flux to underlying aquifers and estimate groundwater age. Groundwater and pore water were analyzed for major ions, enriched tritium (3H), and stable isotopes (δ18O, δ2H) to corroborate the age of groundwater in the system.

The results from the two study sites suggest three different types of groundwater flow in the till aquitards and their underlying buried-valley aquifers. At the Litchfield study sites, downward-directed, vertical hydraulic gradients of 0.25 and 0.37 in the confining unit are paired with geometric mean K = 8 x 10-7 ft/s (2 x 10-7 m/s) (LFO1) , and geometric mean K = 2 x 10-9 ft/s (6 x 10-10 m/s) (LFO2) suggesting vertical recharge to the aquifer of 78 and 0.34 in/yr (198 and 0.86 cm/yr) and groundwater ages of 2 and 1,026 years. Vertical recharge to the aquifer at LFO1 was limited to 8 in/yr (20 cm/yr) when used to calculate total recharge to the aquifer. The groundwater ages predicted by these calculations are corroborated by the enriched 3H data, which shows pre-bomb groundwater in the aquitard at LFO2 and a possible 3H bomb peak in LFO1, thus suggesting much younger water in the latter. Cl concentrations ranged from 12 to 294 mg/L at LFO1 and LFO2 also suggesting inputs from anthropogenic sources. Values of δ18O and δ2H at either nest do not indicate glacial-age groundwater. Recharge to the aquifer below the aquitard, based on an 8 mi2 (20 km2) area, is between 1,112 and 47 million gallons per year (MGY), which would be sufficient to sustain pumping of 340 MGY in the buried-valley aquifer. In contrast, the hydraulic gradient of 0.07 is upward-directed at CWO1/2 at the Cromwell site and precludes downward flow through the confining unit (geometric mean K = 7 x 10-7 ft/s / 2 x 10-7 m/s). Groundwater age in the till aquitard is pre-bomb, but not of glacial age, and discharge may occur upward into the Automba Member sand and gravel. The lack of a significant difference in hydraulic head between the underlying slate aquifer and the buried-valley aquifer, and the presence of 3H in the latter only, suggest that recharge to the latter is likely occurring somewhere up gradient in the sand and gravel aquifer and not upward from the slate aquifer. Denitrification occurs at depth at both the Litchfield and Cromwell study sites, suggesting that NO3-N contamination of groundwater will not be an issue at either one.

Overall, the results of this study suggest that sites in the same or different glacial lobe, Formation, or Member may differ in their hydraulic properties and their groundwater flow system. Vertical hydraulic gradients may not necessarily be downward. Although my hypothesis that glacial-age groundwater occurs at depth was not confirmed, it is clear that estimating recharge to buried-valley aquifers from field data alone is difficult and that simulation of groundwater flow of the aquitard-aquifer system in three dimensions will be needed to improve prediction of aquifer sustainability of similar buried-valley aquifer systems in Minnesota.