U.S. Geological Survey Water-Supply Paper 2375
National Water Summary 1988-89--Floods and Droughts:

Floods and Droughts

     Located in the Great Lakes region, Michigan is composed of two large peninsulas. The Lower Peninsula is bounded by Lakes Huron and Erie on the east and by Lake Michigan on the west. The Upper Peninsula is bounded by Lakes Michigan and Huron on the south and by Lake Superior on the north. The climate of Michigan is affected by the Great Lakes. The large surface area and volume of the Great Lakes have a stabilizing effect on temperature. Prevailing westerly winds cause milder winters and cooler summers than at similar latitudes farther west.

     Surface-water supplies are constantly replenished by precipitation, which is evenly distributed throughout the year. About 55-60 percent of the annual precipitation of 31 inches is received during the growing season. Summer precipitation generally results from convective storms, whereas winter precipitation generally results from frontal systems. Annual snowfall in the Upper Peninsula is among the largest east of the Rocky Mountains.

     Major flooding is not frequent in Michigan. When flooding does occur, it is most likely in late winter or early spring when sudden warming and rainfall combine with snowmelt, saturated or frozen soils, and decreased evapotranspiration to produce large quantities of runoff. Flooding can occur during any season. About 6 percent of the State is considered to be susceptible to flooding, and damage by flooding is $60-100 million annually. A major flood in the central Lower Peninsula in March 1904 was one of the first floods recorded in the State. The cause of this flood was rain, snowmelt, and frozen soil. Other memorable floods were recorded in April 1947 April-May 1960, and April 1975. The most extensive and damaging flood of record was in the central Lower Peninsula in September 1986. This flood caused damage of about $500 million.

     Mild droughts are common in Michigan, but severe droughts are infrequent and generally of short duration. The even distribution of precipitation and moderate humidity are helpful in meeting the large demand for moisture by crops. Although rainfall is abundant during the summer, runoff is decreased because of increased evapotranspiration and absorptive capacity of soils. The most severe drought of record was during the 1930's, when only a few streamflow-gaging stations were in operation. The recurrence interval of the 1930-37 drought was 50-70 years, depending upon locality. The second most severe drought, that of the 1960's, was well documented. The recurrence interval for the 1960-67 drought ranged from 40 to 65 years. The effects of both droughts were statewide.

     A flood-hazards program, which is both service oriented and regulatory, is operated by the State of Michigan. Coordination among Federal, State, and local agencies is an important component of the program. About 650 communities or local units of government that have flood-prone areas participate in the National Flood Insurance Program.

     The National Weather Service (NWS) provides flood forecasts for 74 locations. Most of the data necessary for these forecasts are provided by a network of volunteer observers and numerous automated telemetering devices installed at streamflow-gaging stations operated by the U.S. Geological Survey. No fully automated flood-warning systems are operated within the State.

Figure 1. Principal sources and patterns of delivery of moisture into Michigan. Size of arrow implies relative contribution of moisture from source shown. (Source: Data from Douglas R. dark and Andrea Lage, Wisconsin Geological and Natural History Survey.)


     The climate of Michigan is affected by several types of airmasses. Tropical maritime airmasses, which originate in the Gulf of Mexico, are the principal source of moisture (fig. 1). About 75 percent of Michigan's annual precipitation is associated with these airmasses. Polar maritime air-masses, which originate in the north Pacific Ocean and, at times, in the Atlantic Ocean, generally lose much of their moisture before reaching the Great Lakes. Arctic airmasses from the Arctic Ocean and polar continental airmasses from northern Canada deliver little moisture.

     In addition to the oceans, important moisture sources include local and upwind land surfaces, as well as lakes and reservoirs, from which moisture evaporates into the atmosphere. Typically, as a moisture-laden ocean airmass moves inland, it is modified to include some water that has been recycled one or more times through the land-vegetation-air interface.

     Although latitude, which determines the quantity of solar radiation, is the major climatic control, the Great Lakes and differences in land-surface altitude also are important. The combination of the three climatic controls gives most of Michigan a semimarine type of climate despite its midcontinent location. During summer, winds are predominantly from the southwest because of a semipermanent Bermuda high-pressure system centered over the Southeastern United States. During winter, winds are predominantly from the west or northwest, but they change frequently as low- and high-pressure systems move through the area. The eastern Upper Peninsula is an exception because easterly winds prevail during the late fall and early winter. This exception is the result of early winter high-pressure systems that move eastward across Canada and of major storm tracks that push southward (Nurnberger, 1985).

     The Great Lakes are a secondary or regional source of moisture. Lake-effect precipitation is most prevalent in near-shore areas but also affects areas farther inland. The slow response of the Great Lakes to temperature changes and the dominating westerly winds retard the arrival of both summer and winter. In the spring, the cooler temperatures within a few miles of the shoreline slow the development of vegetation. In the fall, tempering of the cold air by warmer lake water results in additional time required for crops to mature or to reach a stage less vulnerable to frost damage.

     Air-temperature data from weather stations at similar latitudes in Michigan and Wisconsin illustrate the lake effect on temperature. On the western side of Lake Michigan, the mean temperatures for January at Madison and Milwaukee, Wise., are 15.6 and 18.7 °F (degrees Fahrenheit), respectively (U.S. Weather Bureau, 1951-69; National Oceanic and Atmospheric Administration, 1970-80). On the eastern side of Lake Michigan, the mean temperatures for January at Muskegon and Lansing, Mich., are 23.1 and 21.6 °F, respectively, illustrating the warming effect of the lake. The lake effect on temperatures during summer is reversed, and temperatures are slightly cooler closer to the lake. However, the lake effect during summer is less pronounced than during winter.

     Average annual precipitation in Michigan is about 31 inches, 55-60 percent of which is recorded during the growing season. Summer precipitation is primarily in the form of showers or thunderstorms, whereas steadier, less intense precipitation dominates the winter. The number of thunderstorms observed annually ranges from about 25 in the Upper Peninsula to about 40 in the Lower Peninsula. The Upper Peninsula of Michigan receives among the largest annual snowfall totals east of the Rocky Mountains, except for some isolated areas in the northern New England States. Annual snowfall ranges from about 30 inches in the extreme southeast to about 160 inches along the northwestern edge of the Upper Peninsula. This gradation is not uniform, however, because areas adjacent to the eastern shores of the Great Lakes receive more precipitation than areas just a few miles inland.

     Surface-water supplies are replenished by precipitation, which is fairly evenly distributed throughout the year even though periods of no precipitation can last as long as 1 month. Most of the State receives 1.5-2.0 inches of precipitation per month from December through March; 2.5-3.0 inches per month during April, October, and November; 3.0-3.5 inches per month during May, July, August, and September; and 3.5-4.0 inches in June. Because of moderate humidity, evaporation is slow.


     Most major floods and droughts described herein have large areal extent and substantial recurrence intervals-greater than 25 years for floods and greater than 10 years for droughts. Numerous other floods and droughts have occurred in Michigan that were of lesser magnitude and that generally were less widespread than those described but, nonetheless, had a significant impact. Major floods and droughts, and those of a more local or less severe nature, are listed chronologically in table 1; rivers and cities are shown in figure 2.

     A record of stream response to precipitation extremes in a watershed is invaluable for water-resources planning. History indicates that streamflow maximums and minimums are continually surpassed; thus long-term, continuous streamflow monitoring is of great value. Streamflow data before 1931 are scarce, especially for unregulated streams. Before that time, most of the State's gaging stations were operated on regulated streams in conjunction with hydropower operations. The most useful streamflow data for this study began in 1931 when gaging stations on unregulated streams became more numerous. Data from 95 gaging stations were used to determine the areal extent and severity of historical floods in Michigan, and data from 40 stations were used for the drought analysis.

     To depict floods (fig. 3) and droughts(fig. 4) graphically in Michigan, six streamflow-gaging stations were selected from the statewide gaging-station network. The six gaging stations have long periods of record, are located on unregulated streams, are representative of hydrologic conditions in major areas of the State, and were operational during water year 1988. Streamflow data are collected, stored, and reported by water year (a water year is the 12-month period from October 1 through September 30 and is identified by the calendar year in which it ends).

     Because of the State's peninsular configuration, rivers flow relatively short distances from their source areas to the Great Lakes (fig. 2). Most of the basins (93 percent) are entirely within State boundaries (Miller and Twenter, 1986, p. 277). The Great Lakes drain into the St. Lawrence River and ultimately into the Atlantic Ocean. In this report, the upper Grand, Maple, lower Grand, and Thornapple River basins are collectively referred to as the Grand River basin. The Pine and Tittabawassee River basins are denoted as the Tittabawassee River basin.

     A discussion of floods and droughts in Michigan would not be complete without mention of water levels in the Great Lakes. The large storage capacity of the Great Lakes generally accommodates most of the variations in water supply. However, water levels are subject to seasonal and annual fluctuations. In the early 1950's and the early 1970's, the average annual levels were record highs following record-low levels in the mid-1930's and the mid-1960's. Record-high water levels occurred again in the mid-1980's as a result of more than a decade of greater than normal precipitation and less than normal air temperature, which translate into less evaporation and transpiration. The greater than normal streamflow that contributed to the rise of the lakes is graphically shown by positive departures from normal in figure 4. Great Lakes diversions and damage caused by high water levels are described by Hitt and Miller (1986).


     Documentation of major floods in Michigan before 1904 is limited. Earlier floods that have been referenced include 1843, 1852, 1861, and 1875 in the Grand River basin; 1873 and 1876 in the Saginaw River basin; 1854, 1858, 1868, 1869, and 1887 in the Kalamazoo River basin; 1902 in the Clinton River basin and Detroit area; and 1863 and 1902 in the Ontonagon River basin.

     The areal extent and severity of five major Michigan floods are shown in figure 3. Annual-peak-discharge data for the six representative gaging stations and the magnitude of discharges having 10-year and 100-year recurrence intervals at each station also are shown. Most floods have caused personal hardship and property damage; many have caused deaths. The five major floods discussed in this section were among the most severe in Michigan in terms of magnitude, areal extent, loss of life,and property damage.

     Late winter and spring floods are, by far, the most common in Michigan. More than 90 percent of the annual peak discharges of the Red Cedar River at East Lansing (fig. 3, site 3), the Muskegon River at Evart, and the Sturgeon River near Sidnaw (fig. 3, site 1) have occurred from December 1 through June 1. Typically, frontal systems produce a light to moderate, but steady and widespread, rainfall on a saturated snowpack. The upper soil layer typically is frozen and impervious to moisture infiltration, Runoff is increased by the melting snowpack and the frozen soils. Flood stages also are commonly increased by backwater from ice jams, as river ice accumulates where it is unable to flow around bends or past obstacles.

     Summer and fall floods that are caused by intense, localized thunderstorms can be equally or more devastating than those caused by widespread rainfall on snowpack and frozen soils. Two examples of late summer floods are the September storms in 1985 and 1986, which produced substantial runoff and damage.

Figure 2. Selected geographic features, Michigan.

Table 1.   Chronology of major and other memorable floods and droughts in Michigan, 1904-89
[Recurrence interval: The average interval of time within which streamflow will be greater than a particular value for floods or less than a particular value for droughts. Symbol: >, greater than. Sources: Recurrence intervals calculated from U.S. Geological Survey data: other information from U.S. Geological Survey, State and local reports, and newspapers]
Flood or
Date Area affected
(fig. 2)
Flood Mar. 24-27,1904 Grand, Saginaw, Kalamazoo, River Raisin, St. Joseph, and Huron River basins.
25 to >100
Rain on snowpack and frozen soils. In Grand Rapids, 2,500 homes surrounded by floodwater; damage, $2 million. In Lansing, 1 death.
Flood Mar. 1912 Lower Peninsula
10 to 25
Runoff from snowmelt. Tittabawassee River at Midland as much as 5.5 feet above flood stage for 11 days; considerable damage at Saginaw.
Flood Mar. 1916 Lower Peninsula
10 to 25
Thousands of acres of farmland inundated by 2-8 feet of water in Saginaw River basin. Damage in Saginaw, $200,000.
Drought 1930-37 Statewide
50 to 70
Most severe in State's history. In 1930, precipitation 9 inches less than normal.
Drought 1939-42 Lower Peninsula
15 to >50
Alternating periods of normal and less than normal precipitation. Crop damage in 1941.
Flood Aug. 22-24,1942 Ontonagon River
Rainfall intensity of 2-3 inches per hour for 5 hours. Lives lost, 3; damage to bridges, culverts, and roadfill, $100,000.
Flood Apr. 4-11, 1947 Kalamazoo, Grand, Saginaw, Clinton, Detroit, and St. Clair River basins; River Rouge basin.
25 to 100
Most damaging since 1904. Result of thunderstorms and snowmelt. Damage in Flint, $4 million.
Drought 1947-50 Upper Peninsula and northern Lower Peninsula.
5 to 45
Greater than normal temperatures. Forest fires destroyed thousands of acres of timber.
Drought 1952-56 Southern Lower Peninsula
5 to 25
Temperatures greater than normal for 4 years. In 1953, precipitation 9 inches less than normal.
Drought 1955-59 Statewide, except southeastern Lower Peninsula.
15 to 45
Most severe in Upper Peninsula.
Flood May 19-20,1959 Au Gres and Rifle Rivers
As much as 4.5 inches of rainfall in small area. Unit runoff near storm center exceeded 1,500 cubic feet per second per square mile. Damage to bridges, culverts, and roadfill, $108,000.
Flood Apr.-May 1960 Upper Peninsula
25 to >100
Record peak discharges at 23 gaging stations. Damage to homes, businesses, and transportation systems, $575,000.
Drought 1960-67 Statewide
40 to 65
Second longest of record. Precipitation least since 1931. Conditions eased in 1965 in northern Lower Peninsula and Upper Peninsula.
Flood June 25-27,1968 Clinton, Detroit, Huron, and River Raisin basins.
10 to 100
Worst since 1947. Several dams breached. Overland flooding, sewer backup, and basement flooding to about 4,000 structures. Lives lost, 4; damage, $11.5 million.
Flood Apr. 18-24, 1975 Kalamazoo, Grand, Flint, and Shiawassee River basins.
10 to 100
Rain on snow. Most severe in Lansing. Stage highest since 1904; about equal to flood of 1947. About 4,700 homes damaged. Damage, $50 million.
Drought 1976-80 Statewide
10 to 20
Particularly severe in Upper Peninsula. Eased in 1978 in Upper Peninsula and in 1979 in northern Lower Peninsula.
Flood Mar. 14-24,1982 River Raisin and St. Joseph River basins.
10 to >100
Severe in multistate area. In St. Joseph basin, discharges largest since 1950. Two counties declared disaster areas. Lives lost, 1.
Flood Apr. 20-24, 1985 Escanaba, Michigamme, Chocolay, and Manistique River basins.
25 to >100
Exceeded that of 1960 in some areas. Roads closed, dams damaged, and 1,900 homes affected. Damage, $3.5 million.
Flood Sept. 6-9, 1985 Flint River
25 to >100
Frontal system stalled near Flint; successive storms tracked along same path. Basement flooding of 2,500 homes. Six counties declared disaster areas. Damage, $63 million.
Flood Sept. 10-15,1986 Central Lower Peninsula
25 to >100
Fourteen record discharges at gaging stations. Fourteen dams failed, 30,000 homes damaged, crops severely damaged. Lives lost, 6; damage, $500 million in 30 counties.
Drought 1986-89 Statewide
New streamflow minimums at many sites. Continuing in Upper Peninsula and northern Lower Peninsula in 1989.

Figure 3. Areal extent of major floods with a recurrence interval of 25 years or more in Michigan, and annual peak discharge for selected sites, water years 1903-88. (Source: Data from U.S. Geological Survey files.)

     Flooding is frequent in the southern two-thirds of the Lower Peninsula. Much of this area consists of population centers built on glacial lakebeds along Saginaw Bay (Lake Huron), Lake St. Clair, and Lake Erie, where land-surface relief is minimal, and soils are relatively impermeable. During wet periods, floods are common; during dry periods, some small streams have no flow. Much of the State's flood-prone lands are within this area. Flood damage in Michigan is estimated to range from $60 to $100 million annually (Great Lakes and Water Resources Planning Commission, 1987, p. 61).

     One of the most disastrous and extensive floods in the southern Lower Peninsula was in March 1904. Runoff resulting from rainfall during March 24-27 was compounded by snowpack and frozen soils. The rain was caused by a frontal system that moved landward from Lake Michigan. Much of the snowfall during the winter had compacted and formed an ice layer at the ground surface. Near Williamston, more than 100 inches of snow fell between November 1903 and March 1904. Ground frost prevented infiltration of snowmelt.

     Flooding in March 1904 was most prevalent in the Grand River, Saginaw River, Kalamazoo River, and River Raisin basins. Flooding in the St. Joseph and Huron River basins was less severe. Few gaging stations were in operation in 1904 to document the magnitude of the flood; however, on the basis of available data, peak discharges in the Grand and Saginaw River basins were greater than discharges expected to recur once in 100 years. Recurrence intervals in the St. Joseph and Huron River basins ranged from 25 to 50 years. Overall, in the southern Lower Peninsula, the flood peaks resulting from this flood are the highest associated with spring flooding since recordkeeping began.

     As a result of the 1904 flood in Grand Rapids, about 14,000 people were temporarily homeless, 2,500 homes were surrounded by floodwater, 30 factories were closed, and about 10,000 people became unemployed. The estimated damage was $2 million (U.S. Weather Bureau, 1904). In Lansing, the flood of 1904 was the most extensive in 135 years of local history. One fatality was reported, and damage was $200,000 (U.S. Weather Bureau, 1904). At Bay City, the flood was described as the most severe since 1887. Numerous dams were washed away or badly undetermined. Highway and railroad bridges sustained considerable damage; railroad traffic was stopped entirely because bridges and sections of track were washed out. In Kalamazoo, the flood inundated about 2 mi2 (square miles) and caused damage of $50,000 (U.S. Weather Bureau, 1904). Temporary closings of numerous factories idled about 1,300 people. Transportation services were hindered, but no lives were lost.

     The flood of April 4-11, 1947, was the most damaging at many locations since the flood of 1904. The meteorological conditions that led to flooding began with a snowfall in March 1947. On April 1, an eastward-moving frontal system caused thunderstorms in the extreme southern Lower Peninsula. On April 2, rainfall was increased by the slow movement of the frontal system and by an abundance of warm, moist air from the Gulf of Mexico. A second frontal system that had originated in the Southwestern United States reached Michigan on April 4. Thunderstorms were moderate to intense during April 4-6. Jackson received almost 5 inches of rain, and a wide area between Benton Harbor and Detroit received more than 3 inches. In the Flint area, average precipitation was 2.3 inches. As with the flood of 1904, melting snow in some areas combined with rainfall runoff to increase streamflow. Frozen soil may have limited moisture infiltration in some areas.

     The areas affected by the April 1947 flood included the Kalamazoo River, Grand River, Saginaw River, St. Clair River, Clinton River, and River Rouge. Many streams within an area bounded by Kalamazoo, Flint, Mt. Clemens, and Detroit had peak discharges with recurrence intervals of greater than 25 years. In the Kalamazoo River basin, Battle Creek at Battle Creek (fig. 3, site 2) had a peak discharge of 3,640 ft3/s (cubic feet per second), which corresponded to a recurrence interval of about 50 years. In the Flint River basin, the recurrence interval of peak discharge for Farmers Creek near Lapeer (fig. 3, site 4) was about 50 years. Streams in several smaller areas had discharges with recurence intervals equal to or greater than 100 years.

     In Flint, many industries, including automotive industries located near the river, were affected by the April 1947 flood. Damage in this area totaled about $4 million (Wiitala and others, 1963). The peak discharge of the Flint River recorded at Flint had a recurrence interval of about 100 years. At Northville, flooding on the Middle Branch River Rouge was the most damaging on record (U.S. Army Corps of Engineers, 1971, p. 24). The floodwaters filled basements and inundated the first floors of some residences. In the Clinton River basin, the peak discharge associated with the April 1947 (fig. 3, site 6) flood was the largest in 53 years of record; however, a flood in 1902 in southeastern Michigan before streamflow records began may have exceeded the 1947 flood in magnitude.

     Record floods were widespread in the Upper Peninsula on April 24-26 and May 7-12, 1960. The April flood affected primarily the western Upper Peninsula; rainfall in the central and eastern Upper Peninsula was moderate, but flooding was minimal. Although most snowpack in open areas had melted, melting of snowpack in timbered areas contributed to the runoff. The May flood affected the central and eastern Upper Peninsula-an area that still had substantial antecedent moisture from the April storm. Both floods resulted from frontal systems that formed in the Western United States. The frontal systems collided with warm, moist air from the Gulf of Mexico and caused intense rainfall. Rainfall was 3-5 inches during April 24-26 and 4-6 inches during May 6-12. The unusually long duration of these storms was caused by stagnation of the low-pressure system centered over Lake Michigan.

     The two 1960 floods had large areas where recurrence intervals of peak discharge ranged from 25 to 50 years; each flood had small areas where recurrence intervals were greater than 50 years (fig. 3). The April flood in the Montreal, Black, and Presque Isle River basins in the extreme western end of the Upper Peninsula had a recurrence interval of 100 years. The May flood in the Manistique River basin in the central Upper Peninsula had a recurrence interval greater than 100 years. Of the 34 gaging stations in operation in the flood-affected area during 1960, record peak discharges were recorded at 23. Except for parts of the Ontonagon River basin that were inundated in 1942, many of these peak discharges remain as the maximum for the period of record. Because much of the area was neither densely populated nor industrialized, losses from flood damage were relatively small. Damage was estimated to be $575,000 and was limited mainly to flooding of residences and businesses and to washouts of roadways and bridges.

     During April 18-24, 1975, a major flood affected the southern Lower Peninsula. Rainfall during April 18-19, 1975 ,was intense; rainfall totals ranged from 3 to 5 inches. Near Williamston and East Lansing, 4-5 inches of rain fell in 7 hours on April 18. Precipitation of that intensity has a recurrence interval of about 100 years. Antecedent moisture was increased by a snowfall of as much as 13 inches over most of the area 2 weeks before the rainstorm. Soils had become saturated, and temperatures had increased sufficiently to cause streams to have relatively large discharges before the flood-producing rain fell.

     Flood peaks occurred between April 19 and 22, 1975, primarily in the Kalamazoo, Grand, Flint, and Shiawassee River basins and several small basins in the Port Huron and Mt. Clemens area. The magnitude of the flooding differed among localities. Near Williamston, the flood magnitude was slightly less than that of a flood having a 100-year recurrence interval. At East Lansing, the flood of April 1975 had a recurrence interval of about 40 years. On the basis of streamflow records for the Red Cedar River at East Lansing (fig. 3, site 3), the April 1975 flood level was the highest since the flood of 1904 and was approximately equal to that of April 1947. The river reached a stage of 12 feet, which was 5 feet above the flood stage (7 feet) established by the NWS. Two gaging stations in the upper Shiawassee River basin recorded discharges having recurrence intervals greater than 50 years. Flooding having a recurrence interval greater than 25 years affected the Lower Peninsula in a band from near Kalamazoo to near Port Huron (fig. 3).

September 1986 flood damage by the Muskegon River at Big Rapids, Mich. (Photograph courtesy of Scott Harmsen, Bay City Times)

     Flooding in 1975 was most severe in the Lansing metropolitan area and, to a lesser extent, the Flint area. Damage to private and public property in all areas affected by the flood was $50 million (David Charne, Michigan State Police, oral commun., 1989). In Lansing, about 175 homes sustained damage totaling at least one-half their value, 4,500 homes received lesser damage, and additional losses were incurred by schools, utilities, hospitals, and transportation systems (Miller and Swallow, 1975).

     The September 10-15,1986, flood was caused by rainfall from a low-pressure system that developed over the central Great Plains. Northeastward movement of the system produced a warm front that extended across the central part of the Lower Peninsula. The precipitation was caused by warm, moist air south of the front that collided with cold air from the north. The absence of upper atmospheric winds caused the storm to remain relatively stationary over the State for several days. In the areas of greatest rainfall, quantities ranged from about 8 to 13 inches. More than 10 inches of rain fell in 2 days within a 3,500 mi2 area.

     New period-of-record maximums were recorded at 14 gaging stations. The Pere Marquette River at Scottville attained a new maximum discharge (6,440 ft3/s), more than twice the previous maximum discharge (2,970 ft3/s) recorded in 1969. In the Tittabawassee River basin, the Chippewa River near Mt. Pleasant (fig. 3, site 5) had a peak discharge of 6,660 ft3/s. This peak discharge is the largest since the 1904 peak discharge of 7,110 ft3/s. The Tittabawassee River at Midland peaked at 38,700 ft3/s, which exceeded the previous maximum of 34,800 ft3/s on March 28,1916. In 1986, the river crested more than 4 feet above the 1916 peak. The discharge of the Saginaw River at Saginaw was less than the discharge of the 1904 flood. Many of the peak discharges, including those on the Chippewa and Pere Marquette Rivers, had recurrence intervals greater than 100 years.

     The flood of September 10-15, 1986, resulted in unprecedented damage. The flooding caused 6 deaths, injured 89, contributed to the failure of 14 dams, threatened 19 additional dams, and caused basement flooding or structural damage to about 30,000 homes (Miller and Blumer, 1988). Four primary road bridges and hundreds of secondary road bridges and culverts failed, making 3,600 miles of roadway impassable. Total damage to homes, businesses, public structures, and harvest-ready agricultural crops was $500 million (David Chame, Michigan State Police, oral commun., 1989). A 30-county area of the State was declared a Federal disaster area. Crop damage was severe, especially in the Saginaw River basin, where dikes were breached and thousands of acres of sugar beets, beans, potatoes, corn, and other vegetables were ruined. Of Michigan's 12 million acres of cultivated land, about 1.5 million acres were affected. In addition to the extensive crop losses, more than 1,200 farm-related structures were flooded.


     Mild droughts are common in Michigan, but severe droughts are infrequent and generally of short duration (Numberger, 1980). The normally even distribution of precipitation and moderate humidity are helpful in meeting the large demand for moisture by crops. Dry weather can last as long as several weeks. Rain-free periods generally do not destroy an entire crop but can result in slowed growth or decreased yields.

     A drought that is only temporarily eased and then resumes may not seem to be severe meteorologically. From a hydrologic view, however, drought-easing precipitation may not be sufficient to replenish soil moisture, percolate to the water table, and eventually return streamflow to normal. Thus, if a drought is considered to continue until streamflow returns to normal, a hydrologic drought may include more than one meteorological drought.

     The maps in figure 4 show the severity of the State's five historically most extreme droughts and the areas that were affected. The hydrographs show annual departures from long-term-average streamflow for six of the gaging stations used in the drought analysis. Drought recurrence intervals are calculated on the basis of the magnitude of cumulative streamflow deficiencies. Droughts are easily recognized in the 1930's, 1940's, 1950's, 1960's, 1970's, and, most recently, from 1986 to present (1989), but there is no evidence that droughts have a cyclic pattern in Michigan (Fred V. Nurnberger, Michigan Department of Agriculture, oral commun., 1988).

     A combination of several meteorological droughts starting in the 1930's led to the most severe hydrologic drought, both in magnitude and duration, in Michigan's history. The drought had a recurrence interval that was greater than 25 years for the entire State but was most severe in the Lower Peninsula. The maximum recurrence interval at any locality was about 70 years.

     During the summer and fall of 1930, precipitation at many locations was less than 30 percent of normal. Statewide, the total precipitation for the year was about 9 inches less than normal, and the lack of precipitation caused streamflow to decrease rapidly. Soil was reported to be unusually dry and hard. Winter precipitation only temporarily relieved the drought, and subsoil moisture remained abnormally dry. In the summer of 1931, many crops were stunted, and many wells were dry. In the summer of 1932, crops again were affected by the dry conditions, but not to the same extent as in 1931. Precipitation was normal during the winter of 1932 but returned to less than normal in 1933. During the growing seasons in 1930, 1934, and 1936, precipitation was about 5 inches less than normal. Because of the severity of the 1930-37 drought, 41 counties were recognized by the Federal Drought Relief Administration as needing assistance. Numerous deaths were attributed to extreme heat in July 1936. The 1930-32 and 1933-37 dry periods cannot easily be distinguished on the basis of streamflow. For this reason, the drought is considered herein to span the interval 1930-37. As indicated by the annual-departure graphs for gaging stations having record during the drought (fig. 4, sites 3-5), streamflow was substantially less than normal in most of those years.

     During 1947-50, a drought developed in the Upper Peninsula and the northern one-half of the Lower Peninsula. In the springs of 1947 and 1948, much of the southern Lower Peninsula experienced wetter than normal conditions, which helped to avert drought in that area. Deficient streamflow in the drought-stricken part of the State is readily apparent in the annual-departure graphs for the Middle Branch Ontonagon River near Paulding, the Manistique River near Manistique, and the Muskegon River at Evart (fig. 4, sites 1,2, and 4). The calculated recurrence intervals of streamflow deficiencies measured during the drought ranged from 5 to 20 years in the Lower Peninsula to as much as 45 years in the western Upper Peninsula.

     The 1947-50 drought was characterized by greater than normal temperatures, particularly during the summer of 1947. The drought was moderate over much of the State, and precipitation was severely deficient only in the western Upper Peninsula. Crops in general were not damaged, but as a result of the dry conditions in October 1947, numerous forest fires destroyed thousands of acres of timber in northern Michigan.

     During 1952-56, the southern one-half of the Lower Peninsula experienced a drought at the same time streams in the northern Lower Peninsula and the Upper Peninsula (fig. 4, sites 1, 2, and 4) had greater than normal flow. The areas most severely affected by drought were the Clinton River, River Rouge (Detroit basin), and Kalamazoo River basins. In south-central Michigan, precipitation was about 9 inches less than normal during the summer of 1953, and 1955 marked the end of 4 consecutive years of greater than normal temperatures. In drought-stricken areas, recurrence intervals for this drought ranged from about 5 to 25 years.

     The drought of 1955-59 affected primarily the area that had not been affected by the 1952-56 drought. However, two areas in southern Michigan had dry conditions in both 1952-56 and 1955-59. In parts of the Clinton (fig. 4, site 5) and Flint River basins, streamflow was less than normal beginning in 1957. In the St. Joseph, Kalamazoo, and Grand River (fig. 4, site 3) basins, streamflow was less than normal beginning in middle to late 1956. Recurrence intervals for this drought ranged from 15 to 35 years in the Lower Peninsula and from 15 to 45 years in the Upper Peninsula.

     The longest drought since the 1930's occurred during 1960-67 in the southern Lower Peninsula and 1960-65 in the northern Lower Peninsula and the Upper Peninsula. Many stream, lake, and ground-water levels were at or near record lows during the drought. Precipitation during 1962-63 was the least since 1931. During the summer of 1965, the lack of rainfall would have been more pronounced except for the abnormally cool temperatures. Statewide, the precipitation deficiency was not as severe as during 1936. Deficient streamflow is evident for all sites in figure 4. Recurrence intervals ranged from 40 to 65 years. Crops in the central part of the Lower Peninsula were severely damaged during 1965. Several counties were designated drought-disaster areas.

     A multistate drought that began in late 1986 (water year 1987) has received substantial attention. During 1987 and 1988, greater than normal temperatures and uneven moisture distribution were the causes of new minimum streamflows at many sites. In 1988, annual streamflow was less than normal at gaging stations statewide (fig. 4, sites 1-6). In 1989, streamflow returned to normal in many parts of the southern and central Lower Peninsula but remained less than normal in parts of the Upper Peninsula and the northern Lower Peninsula. The drought affected water use throughout the State.

Figure 4. Areal extent of major droughts with a recurrence interncal of 10 years or more in Michigan, and annual departure from average stream discharge for selected sites, water yuears 1930-88. (Source: Data from U.S. Geological Survey files.)


     A comprehensive water planning process has been initiated by the Great Lakes and Water Resources Planning Commission. The goal of this process is to streamline and coordinate the management of all water- and land-related resources.

     Flood-Plain Management.--The State of Michigan operates a flood-hazards program that is both service oriented and regulatory. Statewide services are provided by using hydrologic-engineering, water-resources, and community-planning expertise. Regulatory functions include a permit process and hydrologic-engineering review, inspection, and coordination activities. The flood-hazards program is administered by the Department of Natural Resources, Land and Water Management Division. The goal of this program is to minimize personal injury, loss of life, and property damage from flooding. The activities listed below are directly related to Federal, other State, and local agency programs. About 6 percent of the land area in the State is susceptible to flooding. For this reason, 650 communities or local units of government that have flood-prone areas have participated in the National Flood Insurance Program administered by the Federal Emergency Management Agency. Individual elements of the program include the following:

  1. Regulation by flood-plain authority of the placement of encroachments such as bridges and culverts in riverine flood plains. The objective is to preserve the capacity to carry floodwater and to prevent obstruction to flows that increase flood hazards. Of the approximately 700 projects that are reviewed annually, about 60 percent are approved with conditions or modifications.
  2. Determination of flood risk by the subdivision review authority, using hydrologic analysis, for proposed housing subdivision developments.
  3. Provision of information by the flood-hazard information service that can be used at the design stage of a project to avoid flood-plain problems later. Information is provided for about 400 real-estate requests per year.
  4. Provision by the National Flood Insurance Community Assistance Program of technical assistance to communities in developing and administering flood-hazard regulations that are consistent with the requirements of the National Flood Insurance Program. About 650 flood-prone communities or local units of government participate in the program statewide.
  5. Flood-hazard mitigation of existing structures resulting from a program that recommends methods, materials, and techniques to decrease future flood damages. This service is available during disaster and nondisaster periods.

     Flood-Warning Systems.--The NWS currently (1989) provides flood-forecast information at 74 locations in Michigan. This information is made available to radio and television stations, emergency service offices, and State police posts. A network of volunteer observers and 33 automated telemetering devices installed at strategic U.S. Geological Survey gaging stations provides river stage and rainfall data to aid in this effort. In addition, the Michigan Department of Natural Resources independently provides flood forecasts for a large part of the Grand River basin as part of a cooperative agreement with the NWS River Forecast Center in Minneapolis, Minn. As a result of major floods in 1985 and 1986 in southern Michigan, several communities have considered acquiring automated flood-warning systems, although none have been installed to date.

     Water-Use Management During Droughts.--No State or regional water-conservation policies or drought contingency plans have been established in Michigan. Although water resources are generally abundant, Michigan occasionally is affected by droughts. The potential effect of consumptive water use on streamflow during a drought has been investigated in the River Raisin basin in southeastern Michigan (Fulcher and others, 1986). Inventories of all major water users were completed for the basin, and the natural low-flow characteristics of the river were estimated from streamflow records. By combining the inventories and low-flow characteristics, the effects of consumptive water use were calculated throughout the basin. The results indicate that consumptive water use substantially decreases the base flow throughout the river basin and, in fact, can dewater the river completely in some stream reaches. On the basis of the significant consumptive water losses in the River Raisin, an evaluation of other basins in the State to determine the effects of consumptive water use on natural streamflow would be beneficial.


  • Fulcher.G.W., Miller, S.A., and Van Til, R.L., 1986, Effects of consumptive water uses on drought flows in the River Raisin: Lansing, Michigan Department of Natural Resources, Land and Water Management Division, 32 p.
  • Glatfelter, D.R., Butch, O.K., and Stewart, J.A., 1984, Floods of March 1982, Indiana, Michigan, Ohio: U.S. Geological Survey Water-Resources Investigations Report 83-201, 40 p.
  • Great Lakes and Water Resources Planning Commission, 1987, Water resources for the future-Michigan's action plan: Lansing, Mich., 151 p.
  • Hitt, K.J., arid Miller, J.B., 1986, Great Lakes set record high water levels, in U.S. Geological Survey, National water summary 1985-Hydrologic events and surface-water resources: U.S. Geological Survey Water-Supply Paper 2300, p. 35-40.
  • Knutilla, R.L., and Swallow, L.A., 1975a. Flood of April 1975 at East Lansing, Michigan: U.S. Geological Survey Open-File Report 75-299.
  • _____1975b. Flood of April 1975 at Meridian Township, Michigan: U.S. Geological Survey Open-File Report 75-301.
  • _____1975c. Flood of April 1975 at Williamston, Michigan: U.S. Geological Survey Open-File Report 75-289.
  • Miller, J.B., and Blumer, S.P., 1988, Flood of September l0 to 15, 1986, across the central Lower Peninsula of Michigan, in U.S. Geological Survey, National water summary 1986-Hydrologic events and ground-water quality: U.S. Geological Survey Water-Supply Paper 2325, p. 64-65.
  • Miller, J.B., and Swallow, L.A., 1975, Flood of April 1975 at Lansing, Michigan: U.S. Geological Survey Open-File Report 75-300.
  • Miller, J.B., and Twenter, F.R., 1986, Michigan-Surface water resources, in U.S. Geological Survey, National water summary 1985-Hydrologic events and surface-water resources: U.S. Geological Survey Water-Supply Paper 2300, p. 277-284.
  • National Oceanic and Atmospheric Administration, 1970-80, Climatological data annual summary-Michigan: Asheville, N.C., National Climatic Data Center, no. 13 [published annually].
  • Noecker, Max, Wiitala, S.W., and Knutilla, R.L., 1969, Flood of August 1942 in the Ontonagon River basin, Michigan: U.S. Geological Survey open-file report, 37 p.
  • Nurnberger, F.V., 1980, Meteorological drought in Michigan-A review of the past 50 years, in Michigan State University Institute of Water Research, Conference on Supplemental Irrigation-Advantages and disadvantages: Lansing, March 18, 1980, 21 p.
  • ______1985, Climate of Michigan: Lansing, Michigan Department of Agriculture, 4 p.
  • Stoimenoff, L.E., 1960, Floods of May 1959 in the Au Gres and Rifle River basins, Michigan: U.S. Geological Survey Open-File Report 60-135, 14 p.
  • U.S. Army Corps of Engineers, 1971, Flood plain information, Middle River Rouge, Northville, Michigan: Detroit, Mich., Detroit District, 44 p.
  • U.S. Geological Survey, 1986, National water summary 1985-Hydrologic events and surface-water resources: U.S. Geological Survey Water-Supply Paper 2300, 506 p.
  • ______1990, National water summary 1987-Hydrologic events and water supply and use: U.S. Geological Survey Water-Supply Paper 2350, 553 p.
  • U.S. Weather Bureau, 1904, Climatological data, Michigan section: Washington, D.C., Department of Agriculture, Climate and Crop Service Report, March 1904.
  • ______1951-69, Climatological data, annual summaries, Michigan: Department of Commerce, several volumes.
  • Wiitala, S.W., Vanlier, K.E., and Krieger, R.A., 1963, Water resources of the Flint area, Michigan: U.S. Geological Survey Water-Supply Paper 1499-E, 86 p.

Prepared by Stephen P. Blumer, U.S. Geological Survey; "General Climatology" section assistance by Fred V. Nurnberger, State Climatologist; "Water Management" section by David A. Hamilton and Richard C. Sorrell, Michigan Department of Natural Resources, Land and Water Management Division

FOR ADDITIONAL INFORMATION: District Chief, U.S. Geological Survey, 6520 Mercantile Way, Suite 5, Lansing, MI 48911

U.S. Geological Survey Water-Supply Paper 2375, p. 335-344