GSA 2020 Connects Online

Paper No. 253-7
Presentation Time: 11:50 AM

HYDROGEOLOGIC INVESTIGATION OF FLOW INTO AND THROUGH UNNAMED LAKE, BEMIDJI, MN


DEVLIN, J.F.1, JONES, Matthew M.1, OSORNO, Trevor C.1, HEYER, Bryan R.2, TROST, Jared J.3 and BEKINS, Barbara A.4, (1)Geology Department, University of Kansas, Lindley Hall Room 215, 1475 Jayhawk BLVD, Lawrence, KS 66045, (2)Geology Department, University of Kansas, 1414 Naismith Dr., Slawson Hall Room 270, Lawrence, KS 66045, (3)U.S. Geological Survey, Minnesota Water Science Center, 2280 Woodale Drive, Mounds View, MN 55112, (4)Water Mission Area HQ, U.S. Geological Survey, Menlo Park, CA 94025

Over the 35 years of research at the oil spill site near Bemidji, Minnesota, only a handful of studies have examined Unnamed Lake, despite proximity to the contamination source. Previous studies suggest that discharging groundwater is introducing oil metabolites to the lake, but hydrogeological details are lacking. In this study, conventional hydrogeological methods were combined with Streambed Point Velocity Probe (SBPVP) methods (mini tracer test) to better understand groundwater flow into and through the lake.

Darcy-based methods indicated that the seepage velocity in the surrounding aquifer was on the order of 0.05 to 0.1 m/d, and values in the lake bed, discharging to the lake, were in the range of about 0.5 to 5 m/d, depending on the assumed hydraulic conductivity (K). Based on hydraulic gradients, upward flow was suggested on the western side of the lake (inflow) and downward flow on the eastern (outflow) side, as expected in a flow-through lake. However, the SBPVP measurements near the lake-sediment interface indicated downward flow near the west side of the lake and upward flow near the east side, contradicting the hydraulic gradient data. Expected flow patterns were only found along the lake edges, where the muck thinned out or was absent.

A reconciliation of the conventional and SBPVP data sets is possible with consideration of a thick organic muck layer in the lake. Work completed in 2019 and 2020 indicated that water beneath the muck layer was under pressure, suggesting the layer served as an aquitard. The SBPVP data collected in 2020 suggests that the interface between the muck and the underlying mineral sediment may act as a high permeability conduit for flow under the lake. Preliminary modeling suggests that downward flow through this conduit could occur even while hydraulic gradients between the lake water and the underlying aquifer indicate upward flow.

This work has led to an unusual conceptual model for a flow through lake. A low density aquitard comprising organic muck constrains much of the groundwater discharging into the lake to a narrow zone near the edges of the lake, and possibly into a permeable conduit at the muck-lakebed interface, where seepage velocities are relatively high. This pathway is important to recognize if contaminant impacts on the lake are to be understood in the future.