U.S. Geological Survey Water-Supply Paper 2375
National Water Summary 1988-89--Floods and Droughts:
Vermont's climate is primarily continental and is sometimes modified by the Atlantic Ocean, about 100 miles to the southeast. Systems of low-pressure centers, or cyclones, move across the State and depart through the northeast. These systems cause frequent weather changes and fairly dependable precipitation.
Airmasses are large pools of air having similar characteristics. Three types of airmasses dominate the State: (1) cold, dry air originating in the Canadian and arctic areas (polar continental); (2) warm, moist air from the Gulf of Mexico and adjacent subtropical Waters (tropical maritime); and (3) cool, damp air of the northern Atlantic (polar maritime). Tropical continental air has little effect.
Tropical maritime air brings the greatest moisture. Most precipitation is associated with frontal systems, where the moist air is pushed over a cold air wedge (warm front) to cause precipitation, or an advancing wedge of cold air (cold front) lifts the warm air above condensation levels. Convective showers, often thunderstorms, contribute considerable summertime precipitation. During some years, one or more tropical cyclones, which include tropical storms or hurricanes, bring excessive rains.
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.
Average annual precipitation in Vermont ranges from 32-36 inches in the upper Connecticut River valley and Lake Champlain valley to 36-44 inches in much of the remainder of the State. In the higher altitudes, average annual precipitation ranges from 50 inches to the 75 inches recorded on Mt. Mansfield. The 104-year record at Burlington indicates extremes of 22.6 inches during 1914, which is 33 percent less than normal (1951-80), and 50.2 inches during 1973, which is 49 percent greater than normal. Seasonal totals are similar; there are no distinct wet or dry seasons.
Monthly precipitation during the winter generally is less than during the rest of the year. Much winter precipitation is in the form of snow; average seasonal totals range from 60 to 70 inches along the Connecticut River valley and near the western border to about 100 inches locally at higher altitudes. At Burlington, the seasonal snowfall has ranged from 31.8 inches during 1912-13 to 145.4 inches during 1970-71.
Widespread, steady rainfall from frontal systems, tropical cyclones, or "northeasters" can result in flooding of large areas. Extensive and disastrous floods are rare but can result from intense spring rains combined with warm, humid winds that rapidly release water from the snowpack. Such was true for the devastating flood of March 11-12, 1936. During this flood, total rainfall and snowmelt ranged from 10 to 16 inches over the southeastern one-half of the State. Rainfall alone can cause disastrous flooding similar to that in November 1927. During that flood, rainfall totals of 5-9 inches were common, and much more occurred at higher altitudes. Intense rainfall caused extensive flooding on September 21, 1938, when the "great hurricane" reached landfall in the southern area of the State. Severe thundershowers more commonly cause localized street and cellar flooding.
Droughts originate from unusually persistent anticyclonic circulation in which dry continental air prevails, in combination with a decrease in coastal and tropical cyclone activity. Summer dry spells warrant crop irrigation, but prolonged droughts are rare. The most prolonged statewide drought of the century began during 1960, persisted into 1969, and prompted concern for agriculture and the water supply. Weather stations, however, reported rainfall quantities to be 63 percent of the 1921-50 average in the driest of those years, and most stations recorded 70 percent or more.
MAJOR FLOODS AND DROUGHTS
Major floods and droughts discussed are those having recurrence intervals greater than 25 years for floods and greater than 10 years for droughts and those that were areally extensive. These major events and those of a more local nature are listed chronologically in table 1; rivers and cities are shown in figure 2. Floods (fig. 3) and droughts (fig. 4) are depicted at six streamflow-gaging stations selected from the statewide gaging network. These gaging stations have long periods of record, are currently in operation, and are representative of hydrologic conditions in the principal physiographic areas of the State. The gaging stations also are located on unregulated streams, so the data reflect natural runoff conditions. 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).
Floods and droughts, as well as causing deaths, property damage, and disruption of services, adversely affect surface-water quality. Floods degrade water quality because large quantities of pollutants are washed into streams. The pollutants generally are flushed rapidly downstream by high flows. Droughts during the summer often lead to fishkills as a result of high water temperatures and inadequately diluted effluent from point and nonpoint sources. Potential pollution hazards increase during droughts because ground-water levels and streamflows decrease; as a result, less flow is available for dilution of pollutant loads.
Table 1. Chronology of major and other memorable floods and droughts in Vermont, 1770-1988.
The areal extent and severity of major floods (based on data from 16 gaging stations) are shown in figure 3. Also shown are the annual-peak-discharge data for six representative gaging stations, the theoretical 10-year and 100-year recurrence intervals, and the dates of memorable floods.
Four major floods have occurred in the State during this century (fig. 3). Other significant floods are known to have happened before 1900, but availability and accuracy of records differ widely. Documented floods have been evaluated in detail. As a result of the 1927 flood, a more extensive cooperative State and Federal program was established to collect streamflow data. The 1927 flood predated this gaging-station network, but reliable records exist in engineering reports, books, newspapers, and an early film.
Flooding may occur during any season of the year. Various conditions provide rainfall that can cause flooding. Moisture originating in the Gulf of Mexico and moving northeast creates long periods of continuous rainfall. Coastal storms along the Atlantic seaboard periodically inundate New England with intense precipitation. In late summer and fall, tropical storms and occasionally hurricanes bring intense rainfall and the threat of flooding. Local, slow-moving thunderstorms can cause flash floods on small streams.
Most floods in the State were preceded by conditions that increased the likelihood of severe flooding. Prolonged, moderate rainfall saturates the soil and causes rapid runoff during intense rainfall. Similarly, frozen ground results in increased runoff because it prevents water infiltration into the soil. Snowmelt due to abovefreezing temperatures results in runoff and, during rainstorms, substantially adds to runoff volume. Flooding due to ice jams can result from midwinter thaws or spring breakup.
The severity of the flood of November 3, 1927 (water year 1928), was affected by antecedent conditions. Rainfall was 50 percent greater than normal in October, soils were saturated, and streamflows were above normal. Remnants of a tropical storm reached the State on November 3 and deposited record rainfall quantities, which resulted in the most severe flooding in the State's history (Kinnison, 1929, p. 61).
The 1927 flood was the most destructive hydrologic event in the State's history. Eighty-four deaths were recorded, and damage estimates were $35 million (Wemecke and Mueller, 1972, p. vii). About 3 percent of the population was left homeless when 264 houses were destroyed and 1,400 were damaged. Transportation was disrupted by extensive damage to highways and the destruction of 1,200 bridges (Kinnison, 1929, p. 91).
The maximum recorded rainfall on November 2-3, 1927, was 9.6 inches at Somerset, and estimates were even larger for the higher altitudes of the Green Mountains. An area of 500 mi2 (square miles) received 9 inches of rainfall, and 1,600 mi2 received more than 8 inches (Kinnison, 1929, p. 57). Damage was greatest in the valleys of the Winooski, Lamoille, Missisquoi, Passumpsic, and White Rivers. Flood peak discharges having recurrence intervals greater than 100 years occurred on many streams. A peak discharge of 120,000 cubic feet per second was recorded on the White River at West Hartford (drainage area 690 mi2) (fig. 3, site 6). During the flood, the Winooski River valley alone sustained 55 deaths and $13.5 million in damage. Residents in Montpelier were forced to the upper floors of downtown buildings to escape the rapidly rising waters. The Winooski River crested 12.1 feet above the current (1988) NWS flood stage.
Events leading to the March 11-21, 1936, floods were typical for early spring. A large snowpack covered much of the State. Water content in the snowpack ranged from 2 inches in the Lake Champlain valley to 10 inches in the higher altitudes of the Green Mountains. Streams were ice covered, and the ground was frozen. On March 11, as much as 4 inches of rain fell on southern areas of the State. The precipitation and moderate temperatures, which melted the snowpack, contributed considerably to runoff. The flood resulting from this storm was compounded by the breakup of thick ice on streams.
A second storm entered the State on March 16, 1936. March 16-19 rainfall quantities were similar to those of the earlier storm; southern Vermont received 4 inches, and northern areas 1 inch. Runoff from this second storm flowed into river systems that were still burdened with waters from the March 11 flood (Grover, 1937, p. 14). The intense rainfall and snowmelt combined to produce widespread flooding. The Connecticut River below Fifteen Mile Falls had peak flows of record; however, at White River Junction, the discharge of the White River into the Connecticut River was much less than during the 1927 flood. This smaller contribution of the White River to the flow of the Connecticut River at White River Junction was the reason the 1927 floodmark was not exceeded during 1936 at this location. Discharge of other streams was generally of a lesser magnitude than during the 1927 flood. However, many streams in northern and southern areas had flood recurrence intervals of greater than 50 years. In the central area of the State, peak discharges had recurrence intervals of 25-50 years.
The March 11-21, 1936, floods affected the entire northeastern region of the United States. Extraordinary floods occurred from the Middle Atlantic States to Maine. No deaths were recorded, but damage estimates were $100 million in New England and $1 million in Vermont (Grover, 1937, p. 14).
Small to large quantities of rain fell intermittently during September 12-20, 1938. On September 21, a hurricane moved northward along the eastern seaboard. The hurricane turned northeastward and tracked along the Connecticut River valley, entered Vermont near Brattleboro, then turned northwestward and moved diagonally across the State to near Burlington. Rainfall directly attributed to the hurricane ranged from 2.8 inches in Burlington to 5.5 inches in Brattleboro. Total rainfall amounts, including those from the frequent rains before the hurricane, were 8.2 inches in Brattleboro but only 1.9 inches in Newport.
During the September 21, 1938, flood, peak flows on the Black, Williams, Saxtons, and West Rivers were greater than those for the 1927 and 1936 floods. The peak discharge on the Batten Kill at Arlington (fig. 3, site 1 ) had a recurrence interval greater than 50 years. In southern Vermont, the towns of Wilmington, Londonderry, and Jamaica sustained damage greater than during the 1927 flood. In the central area of the State, flood peaks during September 1938 did not exceed those of the 1927 flood. The Dog River at Northfield Falls (fig. 3, site 2) and the White River at West Hartford (fig. 3, site 6) recorded discharges having recurrence intervals of about 50 years. Damage by winds accompanying the rain was severe in the central area. High winds toppled trees, damaged property, and disrupted services. Vermont lost about one-half of its sugar maple trees as a result of wind damage. In the north, lesser quantities of rainfall resulted in no serious flooding. Recurrence intervals for flood peak discharges were less than 10 years for most northern streams.
The flood (hurricane) of 1938 was, until that time, the worst natural disaster in U.S. history. The storm caused about 700 deaths and $400 million in damage across New England. Only one death was reported in Vermont, but when wind and flood damages are combined, the 1938 hurricane was the most destructive storm in Vermont's history (Ludlum, 1985, p. 137).
On June 28-30, 1973, weather in the State was affected by two frontal systems. A north-trending frontal system moved in from the west and joined a moist, southeasterly flow from the Atlantic Ocean. As a result, many parts of the State recorded the largest rainfall totals since the 1927 flood. The town of Jamaica, in southern Vermont, reported 7.5 inches. The eastern slopes of the Green Mountains received as much as 5 inches of rainfall, and the Lake Champlain valley area recorded about 2 inches. Runoff from this storm was rapid because the soil was saturated by the preceding 3 months of greater than normal precipitation.
Peak discharges during 1973, in the central and northeastern areas, were the largest since the 1927 flood. The peak discharge of the Passumpsic River at Passumpsic (fig. 3, site 5) had a recurrence interval of greater than l00 years. The Dog River at Northfield Falls (fig. 3, site 2) and upper Lamoille River at Johnson (fig. 3, site 3) had peak discharges with recurrence intervals greater than 50 years. The Winooski River at Montpelier crested 2.6 feet above flood stage. Highway damage was extensive in the south-central, southeastern, and northeastern areas of the State. The town of Ludlow on the Black River was seriously damaged. Three persons were killed in the 1973 flood, and damage was estimated at $64 million. Sizable crop loss was reported, and damage to State highways was estimated to be $10 million. The entire State was declared a disaster area (Ludlum, 1985, p. 249).
Localized floods cause extensive damage on a small scale. Frequent, local flooding has been noteworthy throughout the State's history. On April 18, 1982, the Missisquoi River near East Berkshire (fig. 3, site 4) had a peak discharge that was second only to that during the 1927 flood. The recurrence interval was greater than 100 years. This discharge was largely unnoticed because the Missisquoi River flows through rural areas near the Canadian border. Streams south of this area had peak discharges with recurrence intervals of 10 years or less.
On June 7, 1984, a band of severe thunderstorms caused extensive flooding in central areas of the State. Peak discharge on the Lamoille River at Johnson (fig. 3, site 3) equaled that of the 1936 flood: the recurrence interval was about 35 years. The storm caused $16.5 million in damage, and the area was declared a disaster area.
Flash flooding caused derailment of a train on July 6, 1984, in Williston when a small stream, swollen by 6 inches of rainfall, washed out a culvert in a railroad embankment. Five deaths resulted, and 137 persons were injured.
At least $60 million has been spent on flood-control projects in Vermont since 1927 (Wemeeke and Mueller, 1972). Despite these efforts, damage from a major flood, such as in November 1927, can be sizable. Population growth and encroachment onto the flood plain counteract some of the protection provided by flood-control projects. Flood-control projects planned after the 1927 flood were only partly completed because of excessive costs and the reluctance to provide large tracts of land for storage of water during floods (Wemeeke and Mueller, 1972).
Three flood-control dams were built in the Winooski River basin. The effectiveness of these structures in decreasing flood crests was documented during several floods. The 1936 flood occurred soon after completion of two detention reservoirs upstream from Montpelier. During the flood of March 1936, the full capacity of these reservoirs was not utilized (Grover, 1937). The flood crest was decreased, thereby minimizing flood damage to the Montpelier area. A local newspaper account quoted residents' descriptions of the flood as having a smaller peak but a longer duration than previous floods (Montpelier Evening Argus, March 23, 1936). Other flood-control projects were completed on tributaries of the Connecticut River in the 1950's and 1960's. These structures combined to decrease flood crests on the Connecticut River.
The areal extent and severity of major droughts (based on data from 16 gaging stations) are shown in figure 4. Drought duration and severity were determined from cumulative departure from average stream discharge for selected sites. The graphs in figure 4 show annual departure from average stream discharge at six gaging stations. A negative departure indicates less than average flow for a given year.
Precipitation in Vermont is commonly abundant throughout the year. In late summer, a potentially dry period, the State frequently benefits from intense rainfall attributed to tropical airmasses. Summer thunderstorms add to the water budget, maintaining adequate ground-water levels and streamflow (Ludlum, 1985, p. 133). When precipitation decreases or is nonexistent for a month or more, however, agriculture, industry, and public or private water supplies are adversely affected.
A severe drought during 1930-36 affected the entire State (fig. 4). In the northern part of the State, the drought was moderate and had recurrence intervals that ranged from 10 to 25 years among gaging stations. Drought conditions in the rest of the State had recurrence intervals greater than 25 years. This drought coincided with severe drought conditions present in large parts of the central and eastern United States.
Drought conditions in Vermont during 1939-43 were moderate. Only in the extreme southwestern area of the State did recurrence intervals exceed 25 years.
During the 1947-51 drought, the northern area experienced the most severe conditions in the State. Drought recurrence intervals were greater than 25 years. Conditions were moderate in central areas, and drought recurrence intervals were less than 10 years in the south.
The drought of 1960-69 affected the entire State and was the most severe of this century in Vermont. The recurrence interval of the drought was greater than 50 years (fig. 4). This drought was regional in scope, encompassing most of the northeastern United States. Precipitation in the State was less than normal every year during 1960-68, which was the longest continuous spell of deficient precipitation since 1895. Streamflow deficiency was greatest during 1965 (fig. 4, sites 1-6). In 1969, the drought ended abruptly. The winters of 1969-72 produced record snowfalls, and greater than normal precipitation was recorded in 8 of the 11 years during 1969-79.
A drought affected Vermont during 1979-80; drought conditions were moderate throughout much of the State during the summer of 1980. In the northwestern area, however, the situation was sufficiently severe that State and local officials offered drought assistance to dairy farmers. Water was trucked in to provide relief to drought-stricken dairy herds.
Flood-Plain Management.--Regulation of flood plains is a function of local governments. No State statutes authorize direct regulation of flood-plain areas, nor does the State require that local governments adopt and administer such regulations.
Local flood-plain-management regulations are necessary for participation in the National Flood Insurance Program. About 76 percent, or 158 of the 208 communities identified as flood-hazard areas in Vermont, participate in this program.
The Department of Environmental Conservation of the Vermont Agency of Natural Resources is the primary source for flood information. It serves principally as technical advisor and provides engineering data and flood-planning information to local governments and other State agencies. The Department prepares model flood-plain-management ordinances, assists local governments and State agencies in reviewing proposed construction in flood plains, distributes flood maps and flood-level data, and coordinates the National Flood Insurance Program in the State. The Vermont Emergency Management Office, in cooperation with the Federal Emergency Management Agency, provides flood-preparedness and flood-recovery assistance (Roy Gaffney, Department of Environmental Conservation, oral commun., 1988).
In 1953, the U.S. Congress approved the Connecticut River Flood Control Compact. The Compact is an agreement between States within the Connecticut River basin to promote interstate cooperation in flood control. Under the Compact, five reservoirs were constructed in Vermont. These reservoirs are under the authority of the U.S. Army Corps of Engineers, Waltham, Mass.
Flood-Warning Systems.--Information on local flood-warning programs is limited. The National Weather Service River Forecasting Center at Hartford, Conn., and the National Weather Service at Albany, N.Y., and Burlington, Vt., broadcast flood-stage and weather information pertinent to floods by radio and television. Current flood stage and forecasts of time and altitude of peak stages are broadcast for some major rivers. Small-stream flood watches, based on current weather information, are broadcast for local areas. At this time (1988), no automated flood-warning programs are available in the State. Flood-warning systems are commonly believed to be the responsibility of local police and fire departments.
Water-Use Management During Droughts.--Water-supply needs during droughts are primarily addressed at the local level. The Vermont Emergency Management Office has contingency plans to assist communities during drought emergencies.
Prepared by Jon C. Denner, U.S. Geological Survey; "General Climatology" section by Robert Lautzenheiser, New England Climatic Service
FOR ADDITIONAL INFORMATION: Chief, New Hampshire-Vermont Office, U.S. Geological Survey, 525 Clinton Street, RFD 12, Bow, NH 03301