This report gives the results of an investigation to determine the chemical and physical characteristics of ground and surface water in Van Buren County and to relate these characteristics to the agricultural use of land. Chemical inputs to the hydrologic system, including those from precipitation, animal wastes, septic tanks, and fertilizers, are assessed. Land-use, geologic setting, and hydrologic conditions are given as a necessary framework for interpretations.

The land surface in Van Buren county in southwestern Michigan is flat to rolling, and ranges in altitude from 580 to 1,050 feet. About 30,000 acres or 7.4 percent of land is irrigated. Annual precipitation ranges from 34 to 36 inches. Two rivers--the Paw Paw and South Branch Black--drain most of the county. During this study, the maximum discharge of the Paw Paw River was 2,500 cubic feet per second; the minimum discharge was 202 cubic feet per second. The average discharge of the South Branch Black River during a 17- year period of record has been 106 cubic feet per second.

Glacial deposits are the principal source of ground-water supplies. These deposits range in thickness from 100 to 600 feet and consist of till, outwash, and materials of lacustrine and eolian origin. In places the deposits fill buried valleys that are as much as 400 feet deep. The Coldwater Shale of Mississippian age, which underlies the glacial deposits, is mostly shale and usually yields only small amounts of salty water.

Of the glacial deposits, outwash is the most productive aquifer. Most domestic wells obtain water from outwash at depths ranging from 15 to 160 feet. Irrigation wells capable of yielding 1,000 gallons per minute generally are about 200 feet deep. In places in the western part of the county, glacial deposits, which are several hundred feet thick, are mostly clay and yield little or no water.

Areal variations in the chemical and physical characteristics of ground and surface water are related to land use and chemical inputs to the hydrologic system. Information on fertilizer application, septic-tank discharges, animal wastes, and precipitation and dry fallout show that 72.7 percent of nitrogen input is from fertilizer, 21.3 percent from precipitation and dry fallout, 4.5 percent from animal wastes, and 1.5 percent from septic tanks.

Streams and lakes generally contain a calcium bicarbonate type water. The dissolved-solids concentration of streams ranged from 56 to 749 milligrams per liter, and that of lakes, from 28 to 310 milligrams per liter. Water of streams is hard but at most locations is suitable for most uses. Total nitrogen concentrations as high as 15 milligrams per liter were found at two sites. Pesticides (Simazine and Atrazine) were detected at some sites. Relationships between land use and water quality of streams indicate that the nitrate yield of land increases as the percentage of crop land, pasture, and feeding operation increases in a drainage area. Water of lakes also is suitable for most uses, although Alachlor, Atrazine, Silvex, Simazine, Treflan, or 2,4-D were detected in 26 samples.

Ground water is of a calcium bicarbonate type, although sodium, sulfate, and chloride are the predominant ions at some locations. Dissolved-solids concentrations ranged from 112 to 878 milligrams per liter; concentrations of trace metals exceeding those common in water were detected at some locations. Nitrate concentrations in the southern eight townships were generally higher than in the northern ten townships; water from 22 percent of the wells in the southern townships exceeded the 10 milligrams per liter drinking-water standard of the U.S. Environmental Protection Agency. High nitrate concentrations probably are related to fertilizer applications, but equally important seems to be the quantity of irrigation water applied. Oil-field activity in the northern part of the county seems to have increased the chloride concentrations of ground water in some places.

Model simulations of the ground-water flow system matched measured conditions if hydraulic conductivities of 10 to 35 feet per day and a recharge rate of 11.8 inches per year were used. Simulations of 500 gallons per minute pumping from an unconfined aquifer indicated only 2 to 3 feet of drawdown in the vicinity of the pumping well. Similar pumping from a confined aquifer, however, produced about 10 feet of drawdown. Simulations of increased pumping for the irrigated area in the southern part of the county indicated that significant drawdown might be expected in some parts.