Prepared in cooperation with the
Ohio Department of Natural Resources
Record peak stage and peak streamflow were recorded at two USGS streamflow-gaging stations, one on Ohio Brush Creek and the other on Shade River. The peak streamflow at each gaging station exceeded estimates of the 100-year-recurrence-interval peak streamflow. The storms produced heavy rainfall in a band along the Ohio River stretching from western Kentucky to West Virginia, causing floods on many tributaries to the Ohio River. As these and other tributaries drained into the Ohio River, communities along the Ohio River experienced some of the most severe flooding in more than 30 years.
Some of the flooding resulted in tragic consequences in Ohio. The Ohio State Highway Patrol attributed five deaths to the flooding: two in Adams County, and one each in Brown, Pike, and Gallia Counties (fig. 1). These individuals were sweptinto floodwaters while attempting to drive through flooded roadways (Baird and Dempsey, 1997). Areas of southern Adams and Brown Counties received 10 --12 inches of rain, which were the largest recorded accumulations during the March 1--2 storms in Ohio. Areas of Athens and Vinton Counties received about 6 inches of rainfall during the storm. Four inches or more of rain fell over most of the counties located along or near the southern border of Ohio. Eighteen counties were declared Federal and State disaster areas, which qualified them for Federal and State disaster assistance: Adams, Athens, Brown, Clermont, Gallia, Hamilton, Highland, Hocking, Jackson, Lawrence, Meigs, Monroe, Morgan, Pike, Ross, Scioto, Vinton, and Washington (fig. 1).
Although the March 1997 flood and the accompanying damages were covered by the news media and by bulletins and real-time data provided by the U.S. Geological Survey (USGS) Internet sites, there was a need to document the southern Ohio flood and subsequent hydrologic analyses in one publication. Hence, the USGS, in close collaboration with the Ohio Department of Natural Resources (ODNR), began to work in March 1997 to synthesize information on the flood into a single publication.
Top of Document
February 1997. Precipitation in southern Ohio during February was below normal, with rain being the dominant form. The NWS Southwest, South Central, and Southeast Divisions received 78, 67, and 67 percent of normal precipitation, respectively (fig. 2), (Cashell, 1997b). The greatest accumulations of precipitation in southern Ohio during the month was on February 4--5; typical total accumulations were 0.5 to 1 inch in most areas and some locations received nearly 1.5 inches. The following 3 weeks of February were generally dry, although light precipitation fell on many days. On February 27, areas of southern Ohio received rainfall totaling about 0.25 inch. Light snow cover of 1--2 inches was reported in some areas of southern Ohio during the middle of the month. However, daily high temperatures for the latter half of February in these areas generally were well above freezing (40° to 60°Fahrenheit) and most snow cover had melted.
During the evening of February 28, 1997, a low-pressure center was established over the Upper Midwest region of the United States, with a stationary front extending southeast from the low to the Tennessee-Kentucky border. A cold front extended south from the low into Texas. The stationary front remained near this position until just after midnight on March 1, when it began to move north towards the Ohio River.
The former stationary front moved north as a warm front and slowed (nearly stalling) over central Ohio by the afternoon of March 1. Although the warm front moved out of the southern Ohio/northern Kentucky area, several boundaries of unstable air remained over northern Kentucky. As typically happens with a surface low-pressure and frontal system with these characteristics, warm, moisture-laden southerly air was pumped into the "warm sector" (the region ahead of the cold front and south of the warm front) from the Gulf of Mexico. This warm and moist air interacted with the boundaries of unstable air and triggered thunderstorms.
What caused these storms to produce such heavy rains was that the frontal system slowed dramatically and strong, persistent southerly winds continued to pump warm air into northern Kentucky and southern Ohio, sustaining the thunderstorms. The extremely slow movement of the cold front kept the thunderstorms over the Ohio Valley for nearly 36 hours. If the cold front had moved faster, it would have pushed the thunderstorms more quickly to the east, and rainfall amounts would have been much less.
What resulted on March 1--2, 1997, was a broad area of thunderstorms that produced as much as nearly 12 inches of rain in some areas of southern Ohio. The 48-hour total rainfall for this storm is shown in figure 3. The storm produced rainfall over the southern half of the State, the heaviest rain being concentrated to the south and east of a line generally extending from Cincinnati to Bellaire, Ohio.Top of Document
A factor contributing to the rapid rise of the Ohio River is that the heaviest rainfall during the March 1--2, 1997, storms was concentrated in a relatively narrow band along the Ohio River. Thus, a large amount the runoff produced by the storms was downstream from most of the U.S. Army Corps of Engineers flood-control reservoirs, on major tributaries to the Ohio River.
Communities along the Ohio River in Kentucky, Indiana, and Illinois also experienced flooding as a result of the March 1997 flood. In Kentucky, the worst flooding was generally located within a 40-to 50 mile-wide band running approximately parallel to and south of the Ohio River (Kevin J. Ruhl, U.S. Geological Survey, written commun., April, 1997). The city of Falmouth, Ky., at the confluence of the South Fork and main stem of the Licking River, was devastated by floodwaters; hundreds were left homeless and four deaths were attributed to the flooding (Kevin J. Ruhl, U.S. Geological Survey, written commun., April, 1997).
Ranges of recurrence intervals that bracket the estimated recurrence interval associated with the peak streamflow are included in table 1page 1 page 2 (gif format) to provide an indication of the relative magnitude of the March 1997 flooding at each of the gaging stations. The location and corresponding recurrence-interval range for the streamflow-gaging stations listed in table 1 are shown on figure 1.
Peak streamflows for the March 1997 flood at the streamflow-gaging stations were determined by use of standard USGS techniques (Rantz and others, 1982). USGS personnel were able to obtain direct measurements of the streamflow at most of the gaging stations in southern Ohio during the March 1997 flood. At one gaging station, an indirect determination of the peak streamflow was made by use of the slope-area method (Benson and Dalrymple, 1967; Dalrymple and Benson, 1967).
Record peak stage and peak streamflow occurred during the March 1997 floods at two streamflow-gaging stations:
Ohio Brush Creek near West Union (03237500, Adams County) --- The peak streamflow and corresponding stage at the USGS gaging station on Ohio Brush Creek near West Union, Ohio, were the highest recorded since the gaging station was established (in water year 1927). The March 1997 peak streamflow of 77,700 ft 3 /s was greater than the estimated 100-year peak-streamflow value and, therefore, is considered to have been greater than a 100-year flood.
On Sunday morning March 2, 1997, the water level was high enough that highway officials closed the State Route 125 bridge over Ohio Brush Creek (about 6 miles east of West Union) because of concerns about the structural integrity of the bridge (fig. 6 )(gif format). By the afternoon, the water level had dropped, and officials reopened the highway. USGS field personnel were able to obtain a direct measurement of the streamflow from the State Route 125 bridge (fig. 7)(gif format).
Shade River near Chester (03159540, Meigs County) --- The peak stage and peak streamflow for the March 1997 flood were the highest recorded since the gaging station was established (in water year 1966). The March 1997 peak streamflow of 15,600 ft 3 /s exceeded the estimated 100-year peak-streamflow value and, therefore, is considered to have been greater than the 100-year flood.
During the March 1997 flood on the Shade River, all roads leading to the gaging station were inundated, and no direct measurement of streamflow could be made. Therefore, the slope-area indirect method of peak-streamflow determination (Benson and Dalrymple, 1967; Dalrymple and Benson, 1967) was used to determine the peak streamflow at this gaging station. The indirect measurement site was approximately 4,100 feet upstream from the gaging station.
Major flooding occurred at other streams in southern Ohio. Selected locations are discussed in the following paragraphs:
Raccoon Creek at Adamsville (03202000, Gallia County) --- The March 1997 peak streamflow of 16,500 ft 3 /s on Raccoon Creek at the USGS streamflow-gaging station was the largest streamflow since May 1968. The recurrence interval corresponding to the March 1997 peak streamflow is estimated to be greater than 25 years but less than 50 years. Historic log cabins at the recreated 1800's village along the bank of Raccoon Creek at The Bob Evans Farms property in Adamsville were damaged by the floodwaters (figs. 8 and 9) (gif format).
White Oak Creek near Georgetown (03238500, Brown County) --- The March 1997 streamflow of 17,900 ft 3 /s at the USGS streamflow-gaging station White Oak Creek near Georgetown (fig. 10)(gif format) was the largest streamflow since July 1980 (19,500ft 3 /s). The March 1997 flood was the second largest streamflow recorded at this location since the peak streamflow of 22,400 ft 3 /s in March 1964. The recurrence interval for the March 1997 peak streamflow is estimated to be greater than 10 years but less than 25 years.
Ohio River at Cincinnati (03255000, Hamilton County) --- The March 1997 peak streamflow of 625,000 ft 3 /s was the largest streamflow at the Ohio River at Cincinnati USGS streamflow-gaging station since March 1964. The recurrence interval corresponding to the March 1997 peak streamflow is estimated to be greater than 10 years but less than 25 years. Many low-lying areas along the Ohio River in the Cincinnati area -- the most densely populated part of southern Ohio -- were inundated during the March 1997 flood (fig. 11)(gif format).
Shortly after the floodwaters receded in southern Ohio, personnel from the Ohio Department of Natural Resources, Division of Water obtained high-water marks for selected streams. The high-water marks were initially identified and flagged in the field, and third-order-accuracy surveys were made at a later date to determine elevations of the high-water marks. These data (table 2 page 1 page 2) (gif format) were collected to document water-surface elevations along various streams.Top of Document
The human impact of the March 1997 flooding in southern Ohio
was appreciable. Nearly 20,000
people were evacuated during the flooding, and 5
people lost their lives. The fatalities were in Adams, Brown,
Pike, and Gallia Counties. The Ohio Emergency Management
Agency (OEMA) reported that 93 roads were closed as a result of
the flooding on March 2, 1997. OEMA estimated that on March 5,
1997, 1,200 residents of southern Ohio were without natural gas,
2,032 were without electricity, and 1,785 were without telephone
service. OEMA also reported that 37 boil-water advisories were
in effect for various southern Ohio communities, owing to
contamination of water supplies by floodwaters on March 6, 1997.
Preliminary estimates prepared by the Ohio Governor's Office of Budget and Management indicate that the economic impact of the March 1997 flood on Ohio communities may approach $180 million. The following breakdown of the estimates (based upon information from the Ohio Emergency Management Agency, May 1997) supports this determination: Individual Assistance, $12 million, Public Assistance, $45 million; Small Business Administration Loans (Businesses, Homes, Churches), $98 million; Agriculture, $8 million; Mitigation, $12 million; and State-agency expenditures, $5 million.
Ohio Brush Creek near West Union -- The March 1997 peak stage of 31.15 feet at the gaging station on Ohio Brush Creek near West Union, in Adams County, surpassed the previous peak stage (27.91 feet) that occurred on March 10, 1964, by more than 3 feet. A major flood also occurred on the Ohio River in 1964, and Adams County residents along the Ohio River were forced to evacuate. According to the March 12, 1964, issue of "The People's Defender," a West Union newspaper, more than 175 families were affected by the Ohio Brush Creek flooding and more than 300 individuals sought emergency housing in local school buildings. State Route 52, in southern Adams County along the Ohio River, was closed because of the flooding of the Ohio River in 1964 and 1997.
Backwater caused by the Ohio River has contributed to flooding on Ohio Brush Creek. During the 1937 Ohio River flood (the peak of record at many Ohio River gaging stations), the gaging station on Ohio Brush Creek near West Union had been temporarily discontinued because of funding restrictions. Connell Moore, Jr., a lifelong Adams County resident who lived adjacent to Ohio Brush Creek in 1937, stated that the flooding in Ohio Brush Creek was due to backwater caused by the Ohio River. He recalls that his father, Connell Moore, Sr., who was employed by the USGS as a gaging station observer on Ohio Brush Creek from 1940 to 1970, was able to row a boat and touch the bottom of the old State Route 348 bridge that crosses Ohio Brush Creek at the station. The deck of the old State Route 348 bridge, which is no longer in service, was submerged by the flooded Ohio Brush Creek in March 1997.
Raccoon Creek at Adamsville -- The village of Adamsville was originally established adjacent to Raccoon Creek in 1800, although residents began to relocate to the vicinity of Rio Grande in the 1870's as a result of recurring flooding. This left only a few dwellings in the community of Adamsville. Beginning in 1971, The Bob Evans Farm moved log cabins from various locations in southeastern Ohio to Adamsville in order to recreate an 1800's village. Two of these cabins (the Wickline and the Post Office) were swept away by the 1997 flood (fig. 9).
The largest flood of record at the USGS streamflow-gaging station Raccoon Creek at Adamsville occurred in May 1968. Late spring rains produced flooding that forced the evacuation of homes and businesses in the town of Vinton, about 12 miles upstream from Adamsville. The peak stage of the May 1968 flood at the gaging station was 28.69 feet and, according to an issue of the "Gallipolis Daily Tribune" dated May 27, 1968, the State Route 35 bridge (currently State Route 588) over Raccoon Creek at Adamsville was under water for the first time since construction in 1952.
In 1937, Raccoon Creek peaked at 25.20 feet, almost 4 feet less than the March 1997 flood. The January 26, 1937, issue of the "Gallipolis Daily Tribune" reported the Raccoon Creek flood as "smashing all records for height and devastation. " The 1937 flood on Raccoon Creek was 4.1 feet higher than the previous peak stage (21.10 feet), which had been recorded at the gaging station in April 1920.
Shade River near Chester -- The gaging station on the Shade River near Chester is situated in a remote area of eastern Meigs County (fig. 1). The March 1997 peak stage of 31.38 feet exceeded the previous record of 27.39 feet, which was established May 25, 1968. The flood of 1968 affected much of southern Ohio, and it was reported in the May 24, 1968, edition of the "Daily Sentinel" (a Pomeroy newspaper) that thousands of southern Ohio residents were forced to evacuate their homes. It also was reported that the flood caused the closure of all major highways in Meigs County. Another flood that occurred in February 14, 1966, on the Shade River resulted in the closure of major highways in Meigs County. The peak stage of the February 1966 flood was 24.67 feet, 6.71 feet less than in 1997.
Ohio River at Cincinnati -- Floods on the Ohio River have been numerous and have been documented since 1773. The January 1937 flood resulted in a record peak stage of 80.0 feet at the Ohio River at Cincinnati gaging station surpassing the March 1997 peak stage by more than 15 feet. The 1937 flood caused an estimated $32 million in damage (1937 dollars) and forced approximately 50,000 people from their homes (McNutt, 1995). Floodwaters toppled nine fuel tanks in the Mill Creek valley spilling about 135,000 gallons of gasoline. The gasoline ignited, resulting in a fire that covered an area 1 mile long by 3.5 miles wide.
Legislation had been enacted in 1936 authorizing the U.S. Army Corps of Engineers (COE) to construct flood-control projects throughout the Ohio River Basin, and the 1937 flood spurred an acceleration of the effort. The U.S. Army COE currently manages 78 flood-control projects in the Ohio River Basin, 55 of which are located in the stream system above Cincinnati. In addition, the construction of the Mill Creek barrier dam, which was completed in 1948, provides additional flood protection for Cincinnati from Ohio River backwater.
The gage on the Ohio River at Cincinnati has recorded stages in excess of 60 feet eight times since 1937, but this level was last exceeded in 1964 when the river reached a stage of 66.20 feet, almost 2 feet higher than the peak stage of 64.48 feet in 1997. In 1964, there were 34 flood-control reservoirs operating upstream from Cincinnati. The U.S. Army COE estimates that these flood-control projects helped to reduce the March 1997 peak stage at Cincinnati by about 2.5 feet (Steven Holmstrum, U.S. Army Corps of Engineers, written commun., 1997).Top of Document
Record peak stage and peak streamflow were recorded at two USGS streamflow-gaging stations, Ohio Brush Creek near West Union (Adams County) and Shade River near Chester (Meigs County). The peak streamflows at both locations exceeded estimates of 100-year-recurrence-interval peak streamflow. The peak streamflow and the corresponding water level at the USGS gaging station on Ohio Brush Creek near West Union, Ohio, were the highest recorded since the gaging station began operation (in water year 1927). The peak stage and peak streamflow for the Shade River near Chester gaging station also were the highest since the station began operation (in water year 1966). The peak streamflow at the Shade River near Chester was determined by use of an indirect method (slope-area computation) because all roads leading to the gaging station were inundated during the flood.
The largest accumulations of rainfall in Ohio for the March 1--2, 1997, storms were recorded in southern Adams and Brown Counties and ranged from 10 to 12 inches. Parts of Athens and Vinton Counties received about 6 inches of rainfall during the storm. Generally, rainfall amounts of 4 or more inches fell on most of the counties along or near the southern border of Ohio.
The most severe flooding within the state of Ohio was observed in streams located within 50 to 70 miles north of the Ohio River. Recurrence intervals at selected USGS streamflow-gaging stations throughout southern Ohio ranged from less than 2 years to greater than 100 years. The streamflow-gaging stations where floods having higher recurrence-interval floods occurred were on Ohio River tributaries in counties that border the Ohio River. These tributaries drain into the Ohio River, resulting in the flooding of many communities along the Ohio River and some of the worst flooding in over 30 years.