Increased water demand in and around Windom led the U.S. Geological Survey, in cooperation with the Minnesota Department of Natural Resources, local water suppliers, and Cottonwood County, to study the hydrology of aquifers in the Des Moines River Valley near Windom. The study area is the watershed of a 30-kilometer (19-mile) reach of the Des Moines River upstream from Windom.

Based on stratigraphic analysis, two hydrologically and genetically separate surficial aquifers underlie the study area. The Windom aquifer has a saturated thickness of 34 meters (111 feet), and the Des Moines aquifer has a saturated thickness of 33 meters (108 ft). The surficial aquifers are relatively isolated from deeper aquifers by till, but some leakage probably occurs. Recharge to the aquifers is from areal recharge, from Cottonwood Lake, and from edge recharge. Pumping at the Windom well field induces substantial amounts of Cottonwood Lake water into the aquifer. During this study, the water level in a well located between two Red Rock wells and the river was lower than the river level during two periods. During those periods, water in the Des Moines River had the potential to recharge the aquifer. Discharge from the aquifers is primarily to municipal wells, the Des Moines River, and other surface waters.

Most of the ground-water samples collected in the study area consisted of calcium-magnesium bicarbonate waters. Corn and soybean herbicides and their degradates were detected at low concentrations in 14 of 27 ground-water samples and in all 3 river samples. Metolachlor ethane sulfonic acid was the most commonly detected compound and also was detected at the highest concentrations. Nutrient concentrations in ground-water samples were skewed low with high outliers, and nutrient concentrations in river samples generally were less than analytical reporting limits.

Nearly all recharge to the aquifer in the ground-water simulation was from edge recharge (80 percent). Calibrated net areal recharge ranged from 17 to 30 percent of the average annual precipitation. Isotopic composition of ground water and Cottonwood Lake water indicated about one-half of the water withdrawn from the Windom aquifer is from Cottonwood Lake.

Scenarios tested with the calibrated model involved increased ground-water withdrawals and changes in recharge to simulate drier or wetter weather conditions. Doubling the withdrawals from all wells in the model had a small effect except in the Windom well-field area. Maximum head declines in the Red Rock well field and the Jeffers city well were less than 40 centimeters (15 inches). In the Windom well field, the maximum head decline was 11 meters (36 feet). The Windom well field does not induce recharge from the Des Moines River. The addition of a new well that pumped 2,000 cubic meters per day (0.44 million gallons per day) in the Augusta Lake Valley area caused a 0.83-meter-deep (2.72-foot-deep) cone of depression that extended to the valley walls. The drought scenario and the high-precipitation scenario resulted in head changes in the northern part of the Augusta Lake Valley area, in the southwestern part of the Red Rock area, and near the valley edges.

Long-term withdrawals of water for public supplies may cause a net decrease in ground-water discharge to surface water. Water that does not evaporate, or that is not exported, is discharged to the Des Moines River but with changed water quality. Because ground-water and surface-water qualities in the study area are similar, the ground-water discharge probably has little effect on river water quality.