“It was a night of absolute terror.…” That's how Mayor Don Barnett recalled the devastating hours during the evening of June 9 and the early morning of June 10, 1972, when Rapid City, South Dakota, and surrounding communities in the picturesque Black Hills suffered the second-deadliest flash flood in United States history (after the Johnstown Flood of 1889). In a span of six hours, 15 inches of rain fell upstream of Rapid City—averaging about four times the six-hour amounts that are to be expected once every 100 years in the area—causing a wall of water to come rushing through the canyons surrounding the city and into the city itself.
The day began warm and humid. The Daily Weather Map that morning showed both a temperature dew point of 64°F at 7:00 a.m. EST (6:00 a.m. local time). With the air at the surface completely saturated, fog had formed, reducing visibility to one-quarter mile. A stationary front draped across the state, allowed strong, low-level easterly winds to funnel warm, moist air against the Black Hills, destabilizing the atmosphere and setting the stage for disaster.
Caption: NOAA/NWS Daily Weather Map for 7:00 a.m. EST (6:00 a.m. MDT), June 9, 1972. A stationary front draped across South Dakota allows for the influx of moisture-laden air at the low levels of the atmosphere. This, combined with weak upper-level winds, helped set the stage for one of the worst flash floods in United States history.
The National Severe Local Storm Forecast Center (the forerunner of the Storm Prediction Center) did not initially include the Rapid City area in its convective outlook that was issued just after 5:00 a.m. In an updated outlook issued at 9:00 a.m., however, the center expanded its area of expected thunderstorm development to include the Rapid City area.
At noon, the local National Weather Service Office in Rapid City issued the following forecast for that afternoon:
A few isolated thunderstorms may approach severe limits late in the afternoon and evening over Eastern Montana, Northern Wyoming, and Western South Dakota.
As towering cumulus clouds began developing over the hills west of Rapid City, the atmosphere continued to destabilize. By 3:00 p.m., a line of thunderstorms developed to the southeast of Rapid City, moving west northwest. About 40 minutes later, precipitation echoes became visible on the FPS-77 radar located at Ellsworth Air Force Base, just outside of Rapid City.
Shortly after 5:00 p.m., as rain began to fall at Pactola Reservoir, about 20 miles west of Rapid City, journalist Robb DeWall from KOTA-TV took a break from his duties at the station to get dinner. He later recalled, “It was very humid outside and clouds were building to the west.” He also noted that by 5:45 p.m., “There were some gusts of wind, and skies had become heavily overcast, and there were ‘rain squalls.‘” While enjoying a couple of highball drinks in the restaurant lounge of a local motel, DeWall noticed that a TV set was on, tuned to his station. “I recall … our weather announcer was saying there was a chance of scattered showers over the hills,” he said. “A comment was made that someone ought to call the weatherman and tell him it was raining right now in Rapid City. People laughed. I ordered supper.”
The fury that was about to be unleashed on local residents—and DeWall's response to it, however—were no laughing matters. By the time he finished dinner and returned to the studio in downtown Rapid City, having passed on an after-dinner drink, DeWall noted, “The weather looked quite foreboding.” This was around 6:20 p.m. Just minutes later, staff members from the Institute of Atmospheric Sciences at the South Dakota School of Mines and Technology, who had been monitoring a radar used as part of cloudseeding experiments, alerted the local NWS office of the presence of strong thunderstorms visible on the radar near the town of Hermosa, about 20 miles south of Rapid City. Dr. Arnett Dennis from the IAS later recalled that the clouds west of Rapid City “had merged into an almost stationary mesoscale convective system, centered near Rapid Creek below Pactola Dam, and extending some miles both north and south from there.”
By this time, the local police department had alerted Mayor Barnett to the dangers posed by the heavy rainfall in the northern Black Hills, which had already reached four inches in just two hours in some locations. As Barnett later recalled:
I called my best advisor, Mr. Leonard Swanson, the City Public Works Director, and we met at City Hall. Heavy rains were falling, and Mr. Swanson and I drove to Canyon Lake Park where a city worker and his family lived in the park caretaker's home, immediately below the dam. Swanny ordered the caretaker, a Parks Department employee, to immediately take his family, leave their evening meal on the table, and get out of the park. The entire family survived the flood. Not a trace of the home (it was city property) or the contents was ever found.
Similar stories of survival and heroism came out of this event. One 71-year-old woman kept her invalid granddaughter alive by holding her on a mattress as it floated on the water in her flooded house.
The local NWS office issued its first flood warning—for the northern Black Hills—at 7:15 p.m. Forty-five minutes later, after consulting with the NWS River District Office in Sioux City, Iowa, the office expanded the warning to include Rapid City. By 9:00 p.m., telephone service throughout the region began to fail. Northwestern Bell Telephone Company would later report that communications had been disrupted to nearly a dozen exchanges, as well as five Minuteman Missile sites. The company set up two manual telephone circuits “to take priority inward emergency calls from the world into Rapid City, via Sioux Falls.” Operators handled nearly 21,000 calls on June 10—more than five times the normal volume.
At 10:00 p.m., DeWall began broadcasting updates to viewers on KOTA-TV. For the next 13 hours, he would serve “as the voice for the area's Emergency Broadcast System,” The Rapid City Journal recalled in an article following his death in 2009. “Because of that work, he became part of the flood story.” DeWall would later share his feelings about that night in a letter to Barnett:
On that grim “morning after” on June 10th, broadcasting, to a strong degree, helped avert possible panic in the community … and through which, the public was being given the uncolored truth about just what had happened, what was happening at the moment, and what had to be done to help the situation as much as possible.
By 10:15 p.m., Rapid Creek, which cuts right through the city that shares its name, began to overflow its banks. Half an hour later, Canyon Lake Dam, just southwest of the city, failed after the water level rose nearly 12 feet. For the next several hours, flood waters rushed through both canyons and populated areas, destroying much of what was in its path. The flood destroyed every single building in the town of Keystone, just outside of Mount Rushmore National Monument.
Just after midnight on June 10, the flood waters crested in downtown Rapid City, with an estimated flow of 50,000 cubic feet per second. By 5:00 a.m., the flood waters had receded along Rapid Creek, but not before claiming 238 lives and injuring more than 3,000 people. Total damages were estimated at the time at $160 million. The flood destroyed 1,335 homes and some 5,000 automobiles. With a population of around 50,000, nearly everyone in Rapid City either knew someone who had perished, or whose relatives had.
While an NWS survey team, which was charged with assessing the agency's performance during the event, found that flash flood warnings issued by the local NWS office “were timely and useful,” and that prompt action taken by local officials and the media “undoubtedly saved a great number of lives,” it also outlined a number of recommendations to improve the forecast and warning process, as well as to improve the availability of weather radars in the area for use by the local NWS office, which at the time did not have its own radar.
While a flood of similar magnitude could happen again in the Black Hills, the United States Geological Survey estimates the recurrence interval of such a flood in the area at 500 years, meaning that in any given year there is a 0.2 percent chance of a similar flood occurring in the area. While nature may turn a blind eye to history, improvements in science, technology, and partnerships in the decades since the Rapid City Flood of 1972 allow not only for better understanding and detection of flash floods today, but for earlier warnings with improved dissemination.
Contributing Editor SEAN POTTER is a New York-based Certified Consulting Meteorologist (CCM), Certified Broadcast Meteorologist (CBM), and science writer with an interest in weather history.