Red River Rising CoverRed River Rising: The Anatomy of a Flood and the Survival of an American City

By Ashley Shelby

Chapter 1: The Way the Winter Ends


On either side of any highway, the land of the Red River valley unfolds across the earth in what seems, in places, like an endless repetition of the same acre. In the dark, cold mornings of a northern winter, the grain elevators slowly materialize from the shadows as hulking, sluggish monsters. The sky is big, and sunrise first appears there as slender ribbons of pale purple cirrus clouds, seemingly as taut and crisp as taffeta. The land on both sides of the Red River of the North is naked, the topographic equivalent of a confession.
The people of North Dakota take the land on its own terms. The grain elevators stand full or they stand empty-depending on the year’s wheat crop; farmers raise barley and ranch Hereford cattle and don’t complain about a bad year and don’t celebrate a good one; the American Crystal Sugar plant in Grand Forks turns tons of sugar beets into sugar cubes and hardly anyone who buys them at the giant supermarkets in Fargo knows or even cares where they come from. It’s not that there is nothing to complain of in North Dakota; it’s just that this is life, not a tourism campaign. North Dakota is a state that has little to brag about; and, in general, very little bragging goes on here. There is one thing, however, that its people will boast of: the weather. Plains weather summons merciless droughts that so chap the land that the soil splits and swallows crop seed, and any hope of harvest. Plains weather reacts so angrily to the collision of a frosty Alberta clipper and a warm Gulf jet stream that it spits out twenty or thirty inches of snow on fifty-mile-per-hour winds. Plains weather may be mild and salubrious in the morning, then arctic and saturnine in the afternoon. People in North Dakota are on intimate terms with weather, dependent upon it, sensitive to its subtle shiftings, and proud of its random ferocity.

The only thing more impressive than the weather is the good fight the people of North Dakota put up against it each year. In Grand Forks, when the Red River swells during spring thaw, people worry little and sandbag a lot. This is the way winter ends. Nature is not romantic here-it is stark and present. Although North Dakota raises churches and monasteries in much the same way it raises Scotch Fife and Velvet Chaff wheat, even the monks know better than to ascribe the whims of nature to God. Nature is an independent force. And Nature hit North Dakota hard as 1996 ended and through that winter and spring. The Red River of the North at Grand Forks, high up on the North Dakota-Minnesota border, did more damage that year than it had anywhere else along its course, than it ever had.

The author of the 1997 Grand Forks flood has been in dispute for years. Some in Grand Forks believed it was simply winter, a familiar guest who always overstayed his welcome. Winter has always been tough on North Dakota. Back when men used horse and sleigh to get around town in the winter, snow banks could grow so high and so dense that a man could ride his horse and cart over the drifts and never break the crust. When Grand Forks was a frontier town, farmers strung wire or rope between their barns and their farmhouses in late fall so they could feel their way between them during winter’s blinding blizzards.

Despite an illustrious history of demon winters and successful adaptation to a frozen land, there are many people in Grand Forks who believe that if Nature gets the credit for the 1997 flood, then the National Weather Service then should be considered its ghostwriter.

Hydrologists identify at least five factors that can predetermine spring flood conditions: a wet fall, an unusually cold winter, heavy snow accumulation, a cool spring followed quickly by a warming trend, and heavy rainfall during the thaw. The Red River valley had already been dealt a very rainy fall; by November 1996, the autumn moisture level in Grand Forks was more than twice the average amount. Then came an early cold snap and three blizzards. By December the people of Grand Forks had already dug out of enough blizzards to feel as fatigued as they normally did in late April. The head of the University of North Dakota’s Regional Weather Information Center, professor Leon Osborne, Jr., told the local newspaper in mid-December, “We’ve already had a month of January weather, and January isn’t even here yet.”

January weather consists first of unobstructed prairie winds blasting across the plains in startling gusts and often bearing fifty-below-zero wind chills. It then delivers blizzards that stretch their frozen arms across the coteaus and prairies and drop inch after inch of snow, paralyzing whole cities in their embrace.

The threshold between accident and disaster is slender, yet the landscape on either side is manifestly different. To this day in Grand Forks, there is a difference of opinion regarding the fundamental character of the 1997 flood. Some consider it simply a natural disaster, dealt to Grand Forks by the same hand that delivered the blizzards of the previous winter. To consider it simply another act of Nature’s hand is comforting-it implies history, regular promises to be kept. Others, however, consider what happened in the spring of 1997 a man-made calamity.

The fall had been cold and wet, and the winter composed of ice-rinsed winds and blizzards. Grand Forks had been buried under eight blizzards that season, each one so ruthless that it became a distinct personality with its own character traits. The Grand Forks Herald named them like hurricanes: Andy, Betty, Christopher, Doris, Elmo, Franzi, Gust, and Hannah.

In its new Grand Forks office-just a year old-the National Weather Service was performing well. Its meteorological forecasts were nearly flawless during that fall and winter. It issued blizzard warnings promptly, providing users ample time to prepare, and although its employees were themselves snowed in during several blizzards and worked with little or no staff relief for consecutive shifts, operations continued without interruption. Moreover, the agency’s precipitation forecasts were uncannily precise. It was as if the weather events of that winter were not composed by Nature, but were instead dictated by the National Weather Service; if an NWS meteorologist had pointed to the sky and said, “Snow ten inches in Grand Forks,” you half-believed it would.

“You say it’s going to snow on that square piece of land, by God it’s going to happen,” Gregory Gust, a meteorologist at the Grand Forks National Weather Service office said. “We try to do it that way.”

The agency’s predictive skills with blizzards seemed to extend to river forecasting. For the Upper Midwest, those predictions are made at the North Central River Forecast Center (NCRFC), an NWS regional office in Chanhassen, Minnesota, near Minneapolis. Here, scientists formulate river forecasts and relay them to local NWS ffices throughout the region; those offices interpret the results to the public. The coverage area includes the Red River of the North, among 850 other basins. In 1996 the NCRFC had predicted a crest of 44.5 feet for the Red at Grand Forks. The river ultimately crested at 45.8 feet-very accurate, as river forecasts go. However, in the early 1990s, the NCRFC had displayed a tendency to overshoot the mark. In 1994 it forecasted a crest of 42 feet. When the river finally crested, it was ten feet lower than the widely publicized forecast.

It might seem of little consequence if the National Weather Service overestimates a flood stage; how much harm could there be, some might ask, in giving a city some room to breathe? A lot. Overestimating a flood crest can be as grave a mistake in terms of financial impact as underestimating can be in terms of human impact. Small cities all along the Red’s course rely heavily on the NWS, designing costly flood protection plans around its forecast crest number. Reinforcing a dike with clay and sandbags-adding even only a couple of feet of protection-can cost a town millions of dollars. Small towns on the Red often exhaust their meager financial resources paying for flood protection supplies and machinery.

The National Weather Service, then, is under immense pressure to produce accurate flood crest numbers that will neither underestimate nor overestimate the actual flood crest level. Either way, a city on the Red could lose millions of dollars and blame the loss on the National Weather Service. In Chanhassen, the NCRFC hydrologists in charge of forecasting the Red River regularly field phone calls during flood season from frantic local emergency managers who tell them that if the crest prediction rises one more foot, their cities will go bankrupt. Those hydrologists often hear the emergency managers’ questions echoing in their heads as they close their eyes to sleep: Are you sure? Since the 1994 ten-foot overshoot, NCRFC hydrologists were intent on avoiding another costly overestimate.

On January 23, 1997, the St. Paul district of the U.S. Army Corps of Engineers hosted an interagency “Winter Planning Meeting” to discuss flood potential in the Upper Mississippi River, Red River, and Great Lakes Basins. Representatives from the Corps, the National Weather Service, the United States Geological Survey, and the Federal Emergency Management Agency attended, along with local officials. Each representative presented data thought to be relevant to the task at hand-preventing a catastrophic flood. The buzz about this year’s flood potential had grown louder at the agencies with each blizzard that descended upon the Red River valley. Few in the Corps’ or National Weather Service’s delegations had seen so much snow on the ground this early in the season.

Another interagency meeting was held a few days later, this time in Fargo. Both meetings conveyed one overriding message: this was going to be a rough flood season. Fargo weather office warning coordination officer Lou Bennett and service hydrologist Wendy Pearson began making the rounds to small communities along the Red, trying to relay the message the National Weather Service had adopted as its temporary mantra: expect to see more water this spring than you’ve ever seen in your life. At area shopping malls, Bennett manned booths under the slogan “Flooding: When, Where and How High.” He participated in television and radio interviews about the flood potential, and attended, with Pearson, town meetings in small North Dakota towns like Harwood and Kindred.

“We’ve never seen this much water in the snow before,” Pearson told them. “We’re expecting a serious situation.”

People in Grand Forks noticed an unsettling influx of strangers during the first days of February. Federal disaster officials had arrived to begin reconnaissance work and to initiate flood mitigation activities. At the same time, the Federal Emergency Management Agency (FEMA) and National Flood Insurance Program (NFIP) began issuing press releases, television public service announcements, and newspaper appeals urging Grand Forks residents to purchase low-cost federal flood insurance. On February 5, FEMA announced that it was providing an emergency enrollment plan for North Dakota towns on the Red River, cutting the required sign-up time for local and county governments by more than half. The public service spots ranged from outright pleas to slickly produced scare tactics filled with images of past catastrophes-people clinging to treetops while water raged around them. But by the middle of February virtually no one in Grand Forks had purchased flood insurance.

On February 13, the National Weather Service issued a flood outlook for the Red River valley, nearly a month earlier than usual: “Above normal soil moisture and high snowpack water equivalent represent a high spring snowmelt flood potential for the tributaries to the Red River, and a severe spring snowmelt flood potential for the Red River from Wahpeton, North Dakota to the Canadian border.”

Later, some people in Grand Forks would criticize this first narrative outlook as wordy and imprecise, but employees of the National Weather Service would later wonder, “What part of ‘record flooding’ did they not understand?” In fact, the agency had attached a kind of glossary, titled “Terminology” to the outlook. There, “Severe Snowmelt Flood Potential” was defined as “crests from near record levels to exceeding record levels . . . assuming normal future precipitation for the remainder of the snow season.” Buried in the sixth page of the seven-page outlook was this caveat: “Additional above normal precipitation and a rapid melt, combined with spring rains would increase the flood threat.” For anyone without a degree in hydrology, however, this was eye-glazing verbiage. It could have meant anything.

On February 26, the NCRFC hosted an interagency flood coordination meeting at its offices in Chanhassen. It was the third held in just over a month, and the flood season hadn’t even officially begun. The hydrologists assigned to the Red River announced to representatives from the U.S. Army Corps of Engineers and the United States Geological Survey (USGS), among others, that they would be issuing their first numerical flood outlook in two days. This was a big event in this circle of scientists, forecasters, and engineers. But would it catch the attention of the public?

The next day, the Red River hydrologists traveled from their Chanhassen offices to Fargo for yet another interagency flood coordination meeting. In attendance were mayors, city engineers, and emergency managers from towns up and down the Red River valley, FEMA representatives, Corps engineers and reconnaissance teams, local National Weather Service office employees, Small Business Administration and National Flood Insurance Plan representatives, North Dakota’s congressional delegation (senators Kent Conrad and Byron Dorgan, and representative Earl Pomeroy), and dozens of others.

“You’d look over the people out there and it was a Who’s Who of the flood fight in the Red River valley,” NCRFC hydrologist Mike Anderson said. Anderson led the agency’s team of four hydrologists and hydrometeorologists assigned to the Red River. For Anderson and his NCRFC colleague Steve Buan, the Red River valley was home; they’d both grown up there and still had family scattered from Wahpeton to Grand Forks. They knew the Red River. They even loved it.

“I can tell you right now,” Anderson told the assembled group in Fargo, “that forecast tomorrow is going to say flood of record or greater.” He made sure to punch the last two words, and then paused. “Does everyone understand that? We’re going to hit the flood of record at every location in the valley”-he paused again-”or greater.” To the city of Fargo, that meant 38 feet or more. To the city of Wahpeton, it meant 17.9 feet. To Grand Forks, it meant 49 feet.

On the 1997 Spring Flood Potential Map, which the National Weather Service issued the next day, the Red River valley was represented as a misshapen splotch of crimson. On this map, red indicated areas where flooding would be “severe.” While other parts of the Upper Midwest showed moderately high flood potential, the Red River valley was the only area highlighted in red.

That same day, the number forty-nine was heard in Grand Forks for the first time. In this, its first numerical flood outlook for the season-its first solid number-the National Weather Service had calculated a crest of 49 feet for the Red River at Grand Forks. This was two-tenths of a foot higher than the city’s official highest flood crest, 48.8 feet in 1979. The river, though, had been higher than that; 1897’s unofficial crest of 50.2 feet was considered Grand Forks’ highest. But the United States Geological Survey (USGS), the federal agency charged with constructing historical hydrographs for America’s rivers, chose not to use hundred-year-old data on a river as dynamic and as ever-changing as the Red. So the 1979 crest of 48.88 feet was considered, for all intents and purposes, Grand Forks’ flood of record.

The Red River’s depth at Grand Forks varies, though not by much, during the year. Its normal depth is anywhere between eight and twelve feet. In Grand Forks, the river’s flood stage is 28 feet. By all accounts, 49 feet would be a massive, dangerous flood, one that would test the limits of Grand Forks’ levees and dikes, which could be fortified and topped to 52 feet.

The National Weather Service crest outlook was considered a “potential” crest, calculated with present snow cover plus anticipated normal precipitation for that area, for that time of year. Mike Anderson and the NCRFC hydrologists had chosen to issue this forecast two weeks earlier than they ever had before. They knew it was going to be bad and wanted people to know.

A week after the National Weather Service released its 49-foot outlook, the agency’s director, Elbert “Joe” Friday, appeared on CNN to warn that spring flooding in the Red River valley would be “higher than ever before” and that people in the valley should expect “record flooding” in their towns. These kinds of statements constitute some of the strongest language in the agency’s terse vernacular, but Friday went one step further to make sure the people in the Red River valley knew what he meant: “They’re going to see more water than they’ve ever seen before in their lives.” In Minnesota, NWS weather forecasters (responsible for precipitation forecasts, not river forecasts) were staging a major media campaign to emphasize the severity of the upcoming flood season. On March 1 Craig Edwards, an NWS forecaster for twenty-five years, told the Minneapolis Star Tribune, “This situation is the most serious I’ve seen since I’ve been here.” Another article quoted hydrologist Gary McDevitt: “If we should get into a rain situation, it could be disastrous.”

This kind of language kicked other federal agencies into higher gear. “We hope that residents will seriously consider taking preparedness steps now so that they are protected in the coming months should these forecasts hold true,” said Roger Free, a FEMA coordinating officer.

Still, in Grand Forks, the National Flood Insurance Program enrollment logs were almost blank. The flood seemed so far away, and a crest of 49 feet seemed manageable. Many people living down near the dikes, in neighborhoods like Lincoln Drive, Riverside Park, and Central Park, kept reminding themselves that they had never even had seepage in their basements before, despite repeated flooding-the dikes had done their job. Even some insurance agents were dissuading people from buying flood policies. They considered it a waste of money.

On the far end of the University of North Dakota campus on the west edge of town, meteorology professor Leon Osborne was working in what could be described as a weather lab. The Regional Weather Information Center (RWIC) houses a prototype weather information management center, outfitted with Doppler radar, access to a network of automated observing stations similar to those used by the National Weather Service, a soundproof radio room, a television studio, an editing bay, and a graphics and animation station. The 2,000-foot laboratory is a weather analysis and forecasting crucible; it is, one could argue, a kind of condensed version of the North Central River Forecast Center in Chanhassen-but with academic wiggle room, constructed in the hope that the scientist-philosophers who work there will wed science with function.

Leon Osborne grew up in a north Texas farming community and came to UND after receiving a degree in physics from Utah State University and a masters in meteorology from the University of Oklahoma. Osborne, who looks like an ex-football player who might have let too many years pass between visits to the weight room, is the equivalent of the dean’s list mainstay. He is a member of Sigma Pi Sigma physics honor society, Chi Epsilon Pi national meteorology honor society, a professional member of the American Meteorological Society, and a Sigma Xi member. He was, at UND, the big man on campus.

The RWIC would be the greatest accomplishment of Osborne’s career, and the winter and spring of 1996-1997 would be its first test. Osborne and his staff were under enormous pressure. That fall the center had gone live. It had finally received state and federal funding to implement a highway weather forecasting system. The staff had divided nearly every road and interstate in North and South Dakota into sixty-mile segments. Now, every three hours, a forecast was generated for each stretch. Any driver in the Dakotas could now dial a cell phone and hear a precise, very local forecast. Osborne and his staff of eight employees-four on each shift-were prepared to produce fifteen thousand forecasts a day, twenty-four hours a day. Both states had planted hundreds of blue signs on the shoulders of major highways promoting the service, which Osborne had christened Advanced Transportation Information Systems.

However, Osborne had been told in no uncertain terms that if his system went offline for any reason-be it blizzard or flood-the funding would disappear, because he needed to prove the system’s “survivability” was as close to one hundred percent as possible. Newly appointed federal secretary of transportation Rodney E. Slater had visited the RWIC that winter, arriving just ahead of a blizzard.

“Slater looks over his shoulder,” Osborne remembered, “and says ‘Leon, tell me: How bad is this blizzard going to be?’” Osborne stepped over to a computer and began to formulate a forecast. Slater stopped him and pointed to an unsuspecting young staffer, an underling, sitting alone at another computer.

“No,” Slater said, “I want him to tell me.”

“If the system went down, there went our funding,” Osborne said later. “There went our program.” The RWIC was the most important project Osborne had ever produced in his life, and he was determined to make it a success. But the very conditions he was forecasting could also be his undoing. To ascertain the survivability of his weather system, he needed to know how high the Red would crest. How would the Center itself be affected by a flood? Although the RWIC offices were on the far west end of Grand Forks, a relatively safe distance from the river, its hub for telephone communications-its lifeline, really-was in the U.S. West telephone building a block from the Red River. Only the slightest amount of flooding, just one levee breach or crevasse, could put Osborne’s program offline and bury it for good.

“I’d seen enough snow that you had to be blind not to see that there was going to be a major flood,” he said. More specifically, “Given ideal snowmelt conditions, and based upon the amount of snow that we had and we were expecting to see-based upon the global weather patterns-we were expecting to see a flood that would be a record flood. If there were a record flood, and the city was not able to prepare for it, we would lose our program. The conclusion from our analysis was that we were shit out of luck.” That was not precise enough; Osborne needed a reliable number, and he was unconvinced by the National Weather Service’s 49-foot outlook. With sodden soil, ten feet of snow on the ground and more on its way, that outlook seemed absurd to him. Osborne set out to produce his own flood crest forecast.

On March 5, after Blizzard Gust had blown snow dust through the town on violent winds, a nominal spring arrived, mild and gray. The air grew warm, and the snowdrifts in town, still so high they formed nine-foot walls on both sides of city streets, started melting. The soil was saturated-an archive of the fall’s rains and the winter’s snows-and the snowmelt simply collected in giant puddles atop the earth. Interstate 29, the main corridor connecting Grand Forks to the rest of the world, had been closed for a total of eighteen days because of the blizzards. The town’s civil resources-plows, National Guardsmen, the power company-had been overworked from battling blizzard after blizzard and the townspeople, too, were exhausted. The fight had been a good one, and they had triumphed, but they were tired. Everyone was looking forward to a warm spring.

After the ruinous Mississippi valley flood of 1965, the Corps of Engineers and the National Weather Service had made a deal: you show me your data, and I’ll show you mine. That flood was considered, at the time, the Mississippi’s worst ever for the seven-hundred-mile stretch between Royalton, Minnesota (about one hundred miles north of Minneapolis) and Hannibal, Missouri. Much of the anatomy of that flood was eerily similar to what had been happening in the Red River valley, and what would happen in the next weeks: an exceptionally cold and wet autumn that saturated the soil, a midwinter thaw, an above-average March snowfall, and an unexpected spring freeze that halted snowpack runoff. The 1965 Mississippi flood had caused $225 million in damage.

Until that point, there had been little communication between the two federal agencies. The Corps studied the mechanics of the river. The National Weather Service-then called the U.S. Weather Bureau-predicted its behavior. It often takes a disaster to force the federal government to reexamine the way it conducts business, and the 1965 Mississippi flood, simply because of its severity, brought to light the fact that greater understanding of American river systems might be gained if the two agencies shared information.

So it was in that spirit that two Corps hydraulic engineers from the St. Paul district office, Richard Pomerleau and Terry Zien, began their reconnaissance work for the National Weather Service. Pomerleau, in particular, seemed to be everywhere and able to do anything. A graduate of the University of Minnesota’s Institute of Technology and a senior hydraulic engineer for the Corps, Pomerleau is also a nationally known expert in ice engineering. He has developed a highly respected emergency management spill model for the Upper Mississippi River. He is a skilled reconnaissance pilot, adept at gathering data under dire conditions.

“If we needed something,” NCRFC hydrologist Steve Buan said, “he’d be able to figure out a way to figure out how to measure something. If he didn’t have the equipment to measure it, he’d go back to physics principles and come up with a number.” Richard Pomerleau is, in Buan’s words, “MacGyver.” Like MacGyver, Pomerleau and other Corps engineers have sophisticated tools in their toolkits. Likewise at the National Weather Service.

“In years past, you went out there with a coffee can, stuck it in the snow, melted it down and made your best guess, by hook or crook, how much water you could expect,” said Paul Jacobs, northwest Minnesota’s emergency coordinator for the Department of Public Safety. But now the coffee can had now been replaced by remote sensing flyovers, satellites, automatic river gauges, and three-dimensional cameras attached to the bellies of survey planes.

The National Operational Hydrologic Remote Sensing Center, operating under the NWS’s umbrella, is located next door to the NCRFC in Chanhassen. Its small staff is concerned with only one thing: snow. Its pilots collect snow data and analyze it. The mere amount of snow that has fallen in a certain location is empty data to a flood forecaster; he or she must know how much water that snow contains, and that will vary depending on the snow’s character. Light, fluffy snow contains far less moisture than heavy, wet snow; it is entirely possible that a ten-inch snowfall in Grand Forks could yield twice as much moisture as a ten-inch snowfall in Bismarck. The moisture content-called snow melt equivalent-is an essential piece of data needed by flood forecasters because it determines the amount of runoff that will flow into a river during the spring thaw. Without this data, an accurate flood crest forecast is impossible.

The NCRFC, and numerous other National Weather Service river forecasting centers, utilize the Remote Sensing Center’s small staff during flood season, requesting the services of two twin-engine planes, which make regular data-collecting flights. These planes, flying at five hundred feet, measure the gamma radiation emanating from the ground. The lower the radiation, so the theory goes, the more moisture on the ground, be it in the form of standing water or snow.

On other survey flights, three-dimensional cameras capture images of the floodplain, measuring elevation virtually house by house. Automatic river gauges-Grand Forks’ was housed in a shed just off the Sorlie Bridge downtown-kept track of river depth, and transmitted the numbers to hydrologists in a continual stream of data. In the short history of advanced hydrologic data collection, this information had been routed directly only to hydrologists and others who were responsible for prediction. In 1997, though, a few Grand Forks city officials were wearing pagers clipped to their belts that displayed an hourly river level reading. The computer system at the city’s emergency operations center could dial the modem on the Grand Forks river gauge and retrieve real-time river conditions. It was proving to be a year in which to show off how exact flood prediction could be. Any erratic river behavior would be telegraphed, literally and figuratively. Nature, it seemed, wouldn’t be delivering any blows this year that the National Weather Service couldn’t anticipate.

Anticipation was one thing; communication was another. The burden of communicating the severe flood threat to the city of Grand Forks rested on the shoulders of the National Weather Service and its hydrologists. If the citizens didn’t know what was coming, none of this data meant anything.

“Technology has become a major player in getting us the information,” Paul Jacobs said, “but we still depend on the people who use the technology and who are trained to interpret it.” However, the further the coffee-can technique receded into the past, the less the men and women at the National Weather Service trusted human judgment-even their own.

The reconnaissance missions that Corps engineers Richard Pomerleau and Terry Zien were running consisted of regular low-altitude flights over various waterways in the Red River basin. Their main purpose was to gather hydraulic and hydrologic data in areas that contain Corps reservoirs. As a general rule, the Corps does not gather hydrologic data; what Pomerleau and Zien were doing for the National Weather Service was really a kind of favor. The Corps is concerned with hydraulics-the study of the mechanics of fluids, a somewhat fixed science. Hydrology, on the other hand, is concerned with the mutable behavior of rivers, the properties of flow, and the effect of water on the earth. Yet certain hydrologic “laws” could be considered fairly firm, and therefore provide the foundation for a predictive model-laws such as: in flood, the river will fall in the same way it rose. This particular “law” was so reliable, it might as well have been a divine commandment.

When the Red River entered flood stage in the beginning of April, the Corps’ recon team began to gather data specifically for the National Weather Service. Pomerleau and Zien faxed their handwritten daily reports on the basin to NCRFC hydrologist Mike Anderson-meticulous reports whose details ranged from snow appearance (”The snow cover in the Wahpeton-Breckenridge area has taken on a grayish cast as the snow pack solidifies and compacts. The discoloration is probably due to increased water content on the surface”) to the memories of local residents (”Farm lady said that the 1969 peak was higher but of short duration.”) Anecdotal details like these were useful, even to the scientists at the National Weather Service. Despite the agency’s high-functioning models and its contracted gamma flights, it still utilizes over eleven thousand “cooperative observers”-many of them retired farmers or elderly farm widows. In many ways, these observers still provide the “coffee can” measurements that Paul Jacobs said had long since been replaced by high-tech tools.

The National Weather Service provides observers equipment-like rain gauges and yardsticks-and training materials; these may include warnings such as “Freezing rain (glaze ice) should never be reported as snowfall,” or Byzantine instructions for extracting crucial data like snow moisture content. Although National Weather Service observers were “trained,” many of them were elderly and living alone in remote locations, and, despite having lived through the worst weather North Dakota could offer, were perhaps more vulnerable during severe blizzards.

During Blizzard Elmo in January, two meteorologists from the Bismarck office received a phone call from a co-op observer named Florence Newsom. Her home in Hurdsfield, a tiny town halfway between Bismarck and Devil’s Lake, was surrounded by ten-foot drifts, and she was unable to get to the Fischer & Porter rain gauge to change the monthly tape. The county road leading to her home had been closed for two weeks. The two NWS meteorologists-Vern Roller and Dan Markee-decided to visit Florence, who was eighty-two years old and living alone, and change the tape for her. They arrived at the blocked county road two miles from her house, minutes ahead of the Wells County snowplow. The plow became stuck when it tried to move one of the snowdrifts. After giving the plow driver a ride into town, Roller and Markee decided to walk the two miles to Florence’s farmhouse. As they climbed over snowdrifts hardened by the frigid winds and nearly sixteen feet high, a “hungry-looking” coyote followed them. When the meteorologists finally reached the farmhouse, they found it was, indeed, surrounded by a levee of snow.

“After changing the Fischer & Porter tape,” the meteorologists’ report reads, “a cup of hot chocolate and a cookie were greatly appreciated.”

People in the Red River valley liked the regional NWS employees; they were neighbors. Most of them had grown up near the river. They raised their families in the valley. The NWS was the farmer’s friend, the city emergency manager’s ally, a trusted comrade.

But after the flood of 1997, many people in Grand Forks found it hard to look any NWS employee in the face without feeling anger rising from their chest, into their throat, the words catching there. The National Weather Service, in the minds of many citizens, had ruined their lives.

On March 14, 1997, the National Weather Service issued a revised flood outlook for towns along the Red River of the North, raising the crest numbers for nearly every forecast point along its course-from its headwaters at Wahpeton down through Fargo, Halstad, Drayton, and Oslo. The only city on the Red that did not see an increase in its flood crest outlook number was Grand Forks. It remained at 49 feet. Ten days later, the Corps of Engineers told Grand Forks’ city engineers to build the dikes to 52 feet, giving the city three feet of “freeboard”-breathing room.

At the RWIC, Leon Osborne and his staff had come up with a flood crest number. It was different from the number pegged by the National Weather Service. Far different. Osborne’s analysis of the flooding situation in Grand Forks was not, in the purest sense of the word, hydrologic. Osborne had performed a kind of climatological and meteorological medley: he produced an aggregate number based on all the radar the RWIC had generated in the last few months; he created a high-resolution grid of precipitation that had occurred over the course of the winter; he sent staff out to perform snow moisture spot checks; he kept tabs on the distribution of snowfall; he utilized many of the same co-op reports the NWS was utilizing, then adjusted his own data to reflect what the observers were reporting. With this information, Osborne then compared the collected data with previous snowmelts, specifically the 1996 thaw.

“On the left-hand side of the equation, I have 1995-1996 snowfall and the 1996-1997 snowfall; on the right-hand side of the equation, I have crest level of 1995-1996, divided by the crest level of 1996-1997. It’s a simple ratio.” By applying the dynamics of the 1996 snowmelt to the nascent 1997 snowmelt-the snow moisture numbers, the precipitation numbers, and so on-Osborne could extrapolate, could project how the coming thaw would unfold. His analysis suggested one thing: “an incredible amount of water.”

His forecast: 52 feet. Although only three feet higher than the NWS outlook, it might as well have been thirty. Even if the river stayed at exactly 52 feet, perfectly level with the tops of the dikes, it would take only the slightest change-an escaped canoe, a block of ice fallen from the undercarriage of the Sorlie Bridge, an unexpected rainfall-for the river to start pouring over the dikes.

But Osborne did not come up with his number by using a hydrologic model, something widely considered absolutely necessary to accurately predict a flooding river’s crest number. He didn’t take an accurate and thorough measurement of the water content of the snow on the ground.”I don’t have to,” Osborne said. “I have a pretty good idea at what temperature the snow was formed. There are standard relationships; given the nucleation temperature of the snow and the environment in which it resides after it falls, I can tell what the sublimation rate is going to be.”

These are the words of a scientist confident in his own expertise, one who is even a little cocky. Osborne liked to use his tools to enhance his performance, not to dictate it, and for that reason felt comfortable filling in gaps in the data with his own educated guesses-something that was simply not done at the National Weather Service.

“Yeah, there was some subjectivity to it,” Osborne said. “Human judgment came into it when we aggregated the precipitation because there were data voids. The radar wasn’t hitting everywhere.” Osborne could do this-could rely on his own judgment when certain data was absent-because the numbers he was generating were being used only to evaluate the flood risk to the Regional Weather Information Center itself, not to the city of Grand Forks. In fact, he made it a point not to step on any toes. “We were trying to be very respectful of the National Weather Service; I mean, that’s their job, not ours. The last thing I ever wanted to do was to get crosswise with the National Weather Service.”
Osborne would not be able to avoid doing just that.

On March 18, the hall at the National Press Club in Washington, D.C., was filled with news cameras, microphones, and network and newspaper reporters. Two National Weather Service officials stood at the podium: director Dr. Elbert “Joe” Friday, Jr. and Hydrologic Information Center chief Frank Richards. They were preparing to announce the potential for one of the worst flood years the United States had ever seen. The agency had invited FEMA director James Lee Witt and American Red Cross director Elizabeth Dole to join them at the news conference, hopeful that the presence of the U.S. government’s disaster relief agency and one of the largest private relief organizations in the world would convey the urgency of its message. Both had declined to appear, and the National Weather Service “elected to go it alone.” Friday and Richards very much hoped that the wire reporters would write stories that emphasized the certainty of severe flooding, and stressed the importance of proactive flood protection strategy, especially in the Red River valley.

“These could be the highest floods in those areas in the one hundred and fifty years we have been keeping records,” Dr. Friday said when he stepped to the microphone at the Press Club. “You’re going to see hundreds of square miles underwater. We want to make sure this doesn’t come as a surprise to anyone.”

Upper Midwestern states “have been depositing snow all winter like a savings bank,” Richards added. “We’ll be drawing it out of that account this spring.”

The headlines and lead paragraphs that appeared in more than two hundred newspapers across the country the next day likely made Friday and Richards happy. USA Today: “The highest floods in 150 years in the northern Midwest could occur this spring as a result of a deep blanket of snow.” Associated Press: “Friday predicted record-breaking floods on the Red River of the North in North Dakota and northwestern Minnesota.” Reuters: “Spring flooding could be more widespread this year than in any year in the past decade, with the upper Midwest’s snowbound winter leaving that region especially vulnerable.” And in the small papers across North Dakota, the warning was duly noted. Bismarck Tribune: “Flood alert puts focus on N.D.” Wahpeton Daily News: “NWS says major flood is probable.” Ten days later, on March 28, the National Weather Service issued another revised flood outlook for the Red River valley. But the crest number for Grand Forks again remained the same. 49 feet.

The National Guard began “dusting” the Red River on March 31. A river that flows north tends to dam itself with chunks of ice during the spring thaw; a dammed river will flood. Dusting a river with sand softens the ice-the sand attracts the sun’s thermal energy-and speeds up thawing, reducing ice jams. The 1997 dusting operation was the most extensive in state history. Helicopters made flyovers, and National Guardsmen hung out of the chopper doors and dumped dark sand onto the frozen river. On March 30, the National Weather Service issued an official flood warning-akin to a thunderstorm warning-for the Red River of the North, which meant the flood threat was immediate.

In Grand Forks the air was growing warmer and signs of spring emerged. Grass began to appear from beneath the melting snow, and the snowdrifts that had lined the streets all winter were shrinking. The balmy weather gave many people hope that the flood wouldn’t be so bad. National Weather Service hydrologist Wendy Pearson remembered being challenged by a reporter who questioned whether the 49-foot crest number might be a bit of an exaggeration.

“He said, ‘Out on the street people are saying we’re not going to have such a bad flood,’” Pearson said. “They changed their minds and felt things wouldn’t be so bad.”

But in Utah a little blizzard was building, gathering strength as it pushed its way across the plains. It moved toward Grand Forks like a frozen tumbleweed, growing larger as it approached.


All rights reserved by the Minnesota Historical Society Press