Water for irrigation is needed to improve crop yields from sandy soils in the Brooten-Belgrade area. Ground-water supplies of sufficient quantity and suitable quality for irrigation are available in much of the area.

Quaternary glacial drift, as much as 300 feet thick, is underlain by Precambrian crystalline rocks and possibly by Cretaceous sedimentary rocks. Sand and gravel aquifers are buried at various depths in the drift and can be located by test drilling. One buried aquifer, possibly capable of high yields, is within 250 feet of the land surface in the vicinity of Belgrade.

Glacial outwash comprises the upper part of the drift in most of the project area and is locally more than 100 feet thick. The outwash is made up of crossbedded sand and gravel that is interbedded in places with silt and clay deposits and has a saturated thickness of as much as 65 feet. Locally, the transmissivity of the surficial aquifer is as much as 60,000 gallons per day per foot, but elsewhere is generally less than 30,000 gallons per day per foot. The aquifer should yield more than 300 gallons per minute and locally more than 1,000 gallons per minute to individual wells in much of the northern and southwestern parts of the area.

Recharge to the surficial aquifer is almost entirely from precipitation. Significant ground-water losses occur as base flow and underflow, and through evaporation and transpiration.

Water in the buried and surficial aquifers is of the calcium magnesium bicarbonate type and is of suitable quality for irrigation.

An analog model, simulating yearly 30-day pumping periods and hypothetical volumes and distributions of withdrawals, showed the effects on the surficial aquifer of withdrawals of about 20,000 acre-feet per pumping season for 20 years. Predicted water-level declines caused by withdrawals of 20,000 acre-feet per pumping season were generally less than 5 feet in the surficial aquifer and years Predicted water-level declines caused by withdrawals of 20,000 acre-feet pumping season caused predicted water-table declines of more than 10 feet in large parts of the area and caused lake-level declines of as much as 8 feet. The model indicated that water removed from aquifer and lake storage accounted for less than 50 percent of all withdrawals; the remainder was accounted for by water recovered from stream base flow and by water diverted from evaporation and transpiration.