Lake and Groundwater Paleohydrology: Use of Groundwater Flow Theory to Explain Past Lake Levels in West-Central Minnesota

Abstract

Investigation of a simple analytic model of an interfluvial water table demonstrates that a shift in groundwater recharge N changes the water table elevation the most near the middle of the interfluve. Consequently, lakes lying farthest from rivers are most vulnerable to lake-level change. The partial derivative of groundwater head with respect to N, the "positional sensitivity," is quantified for the simple model as a function of position across the interfluve. Despite its simplicity, the positional sensitivity of the model has some predictive value for water-table and lake-level changes in a sandplain in west-central Minnesota.

Lake levels are also a function of surficial hydrology. "Lake pumping" is symbolized by y and defined as the net removal of water from a lake by hydrologic processes acting at the lake surface, namely evaporation minus direct precipitation and minus any input from overland runoff that reaches the lake. Investigation of a simple analytic groundwater model of a circular lake next to an infinitely long river shows that the sensitivity of the lake level to a change in y is proportional to the radius of the lake and its distance from the river. The analysis also indicates that lakes lying in highly permeable substrates are not very sensitive to changes in y. The response of a lake level to a shift in climate depends on characteristics of surficial and groundwater hydrology that are unique to that lake. Determination of the past levels of several lakes, rather than just one, should help provide a more nearly unique reconstruction of past hydrology and climate. Analysis of the sediments of several closed-basin lakes lying in the Parkers Prairie sandplain in west-central Minnesota indicates that lake levels were lowest about 8.5 to 8 ka. I manipulate the N and y of a steady-state analytic-element groundwater model such that the modeled water table coincides with the paleo-lake levels for a given past time. Model results indicate that lake levels at 8.5 to 8 ka can be explained primarily by reducing N to 4O% of the modern value, coupled with a y of about 20 to 30 cm yr-1. By 6 ka N had increased to 50 to 80% of the modern value allowing most lakes to rise in level, but y may also have increased forcing at least one lake to remain nearly dry.