by Lee Sandlin
Fujita’s research projects were part of the vanguard of this new thinking. His first major project was an exhaustive analysis of a tornado that had passed through Fargo, North Dakota, in June 1957. He wanted to do what nobody had ever done before—document the full life cycle of a single tornado. He thought of a novel way of attempting it. It happened that this tornado had been unusually slow moving; it hadn’t raced through Fargo but had almost dawdled, at an average speed of barely twenty miles an hour. This meant that almost everyone in town had stopped to watch it, and many of those people had time to bring out their cameras. With the help of a local television station, Fujita was able to gather more than 150 amateur photographs of the tornado. He used them to create a kind of filmstrip that showed the tornado at one-minute intervals over almost the entire duration of its existence. He then spent the next two years watching the filmstrip. At the end of that time, he had evolved the first model for a tornado’s growth, maturity, and decay. Many of the most basic elements of a tornado’s structure as it is understood today—such as its initial descent from a “wall cloud” within the supercell—were identified for the first time by Fujita in his filmstrip.
Fujita settled into being an American. He brought his wife and children over from Japan and bought a house in Chicago near the university. He became an American citizen. He grew to be a celebrity within the meteorological community, and he published often; his papers on tornadoes, hurricanes, and typhoons were regarded as consistently original and brilliant. He decided to publish under the name T. Theodore Fujita and was pleased when people called him Ted.
In the 1960s, Fujita made exhaustive studies of tornado damage tracks. Initially, he walked most of the tracks on foot, but he ultimately came to prefer being flown the length of the track in a high-wing Cessna. He was particularly insistent on that model because its unusual wing position didn’t block his panoramic view of the damage. By his own count, he conducted detailed examinations of more than three hundred tracks. His published reports set an extraordinarily high standard. One man who accompanied Fujita on some of these flights, a professor of civil engineering named James McDonald, later wrote that Fujita “had an uncanny ability to sort out the damage indicators on the ground and come up with plausible flow patterns.”
Through his exhaustive studies, and through his curious attentiveness to the exact nuances of destruction, he came up with an idea for a classification system for tornadoes based on how much damage they caused. He devised a scale of wind speed with twelve steps, like a musical scale, from calm winds to winds at the speed of sound, and he began assigning the tornadoes he studied to points along the scale. The steps began at category 0, with winds under a hundred miles an hour, causing damage that he classified as light. Category 1 was moderate damage, followed by considerable, severe, devastating, and incredible. Incredible damage was category 5, with estimated wind speeds topping three hundred miles an hour. That was where the scale stopped. Fujita believed that this was the upper limit of what was physically possible: damage that could be considered category 6 he called inconceivable. No tornado on record, not the Tri-state Tornado nor the strange moving mountain that destroyed Irving, Kansas, in 1879, could be rated a 6 or higher.
Fujita introduced his scale in 1971. It was immediately and universally accepted. It quickly became routine to speak of a given tornado as, say, a Fujita scale 2, or just an F2. But nobody imagined how soon the scale would turn out to be essential.
Around one o’clock on the afternoon of April 3, 1974, a tornado touched down deep within the Illinois farm country. It was an unseasonably hot and humid day; the forecast from the National Weather Service was for violent thunderstorms with possible tornadoes all through the heartland. But this tornado hardly counted as violent. It was a faint white sprite of the sultry air; it danced for a few hundred yards across the green depths of a gigantic cornfield and then dissolved against the backdrop of a purple thundercloud. It probably did no more damage than scatter a hayrick or knock over a road sign. For anybody who might have glimpsed it in the distance from an interstate highway, it was doubtless a memorable sight, but on the Fujita scale it rated on the lowest rung, F0. Among professional meteorologists it wouldn’t have been remembered at all, if everything about that day hadn’t turned out to be unforgettable.
The weather data for April 3 showed that there was a very large and powerful area of low pressure moving east across the country from the Rocky Mountains. At the surface, this system was drawing in a flow of warm humid air from the Gulf of Mexico. In the upper air it was drawing in hot dry air from the southwestern deserts. This was an unusual pattern. More typically what happens is that the upper-air winds come in from the north and are very cold. This causes a large falloff in temperature between the surface and the higher altitudes; convection columns form and bleed off their moisture in the form of clouds and rain—just as James Espy had envisaged more than a century before. But on April 3, both the surface air and the upper-level winds were warm, so something else happened. The convection columns didn’t form; instead, the hot air remained near the surface. The atmosphere, as meteorologists have come to say, was “capped.”
The cap remained in place as the day wore on. The cap was so strong, in fact, that it prevented the ordinary formation of clouds. This meant that the undiluted heat of the sun caused the trapped air at the surface to warm up even more. By early afternoon the potential energy within the system was spiking, and the atmosphere was becoming catastrophically unstable. Extremely hot and humid convection columns at last began punching up randomly through the cap, and cumulonimbus clouds bloomed in the skies all over the eastern Mississippi valley and the Deep South.
As the atmospheric cap disintegrated all across the eastern half of the country, exceptionally violent supercells swelled up, died off, and regenerated in the inexhaustible convective flow of released potential energy. By late afternoon tornadoes were spinning out of the clouds and touching down everywhere. There were dozens of them: long writhing snakes, gigantic black blasting chimneys, shapeless gray whirls of debris, vast churning cliff-face wall clouds, graceful tapering cones glowing with late-afternoon gold. By nightfall, tornadoes had touched down in thirteen states. At one point during the evening, so many tornadoes were simultaneously on the ground in Indiana (sixteen, by one count) that meteorologists in a move of desperation issued a single tornado warning for the entire state.
There had been outbreaks of tornadoes before. There were records of them as far back as the Enigma Outbreak of 1884. There had been two major outbreaks in 1953 alone—thirty-three tornadoes in the plains over three days in early May, forty-eight across the Great Lakes and New England a month later. But there had never been a case where so many violent tornadoes had touched down all at once. In Ohio, two tornadoes touched down simultaneously and swirled together into one large F5 funnel cloud about a half mile wide, which shortly afterward destroyed the town of Xenia. It picked up a school bus and dropped it through the roof of the school, where it fell directly onto an auditorium stage; the children had been rehearsing a play there a few minutes before. Roughly two hours later, the town of Tanner, Alabama, was destroyed by an F5 tornado, and as the survivors and rescuers searched frantically through the rubble, a second F5 tornado following the same path blasted through the town and leveled whatever was still standing. Another F5 tornado in Alabama was on the ground for almost two hours and traveled more than a hundred miles; it was one of the most powerful tornadoes recorded since the Tri-state Tornado of 1925.
By the end of the night, more than three hundred people were dead, thousands were injured, and there were tracks of damage and devastation from Illinois to the East Coast. Tornadoes were still touching down after midnight as the storm system was blown toward the Atlantic. One of the last trailing tornadoes touched down in Kingsport, Tennessee, around three in the morning. It was a mile wide, a vast bulging apparition of terror rumbling through the darkness, but it was little more than a hollow bag of gas. Its winds w
ere well under a hundred miles an hour, and on the Fujita scale it rated an F0.
The Super Outbreak, as it quickly became known, was the worst single tornado event in American history. Almost 150 tornadoes touched down out of a single weather system. Fujita had thought that there might turn out to be an average in America of four or five major tornadoes annually—that is, tornadoes rating higher than an F3. In that one night, there had been twenty-four F4s and six F5s.
Fujita made the study of the Super Outbreak a major project. He assembled microscopically detailed surveys of the damage tracks of every tornado in the outbreak, and he mapped them all. The result was an extraordinarily rich image of the ever-shifting and fluid progress of a major storm system. The map also demonstrated that Fujita’s own tornado scale had to be used judiciously. There was no steady and predictable path of destruction even with the strongest tornadoes. A tornado classed as an F5, for instance, usually had clusters and pockets of incredible calamity interspersed with large zones of much milder damage. The damage varied from mile to mile and sometimes from square foot to square foot: most of the houses on a block might be blown into splinters, leaving one in the middle completely untouched. Evidently, even the most violent tornadoes had weird areas of calm within them, while the weakest tornadoes might have pocket domains of uncontrolled fury.
Most meteorologists would have been inclined to write these variations off as random flukes, unpredictable and inexplicable. But Fujita had been struck by a novel—even hair-raisingly original—thought. Tornadoes were not solid, unitary objects. There wasn’t necessarily one central vortex of low pressure at the heart of the funnel. A tornado might be made up of a dance of multiple vortices, spinning and weaving around a central core. The meandering motions of these vortices, as they approached and retreated from the ground back to the upper altitudes, were what created the intricately variable tangles of wreckage that the most powerful tornadoes left behind.
It was ideas like this that led people to start calling Fujita “Mr. Tornado.”
Fujita studied wind damage patterns at hurricanes; he mastered satellite imagery analysis to describe cloud and debris evidence at major tornadoes. He was taken in a high-altitude research jet to circle thunderstorm tops as they rose and collapsed (he believed there was a correlation between a sinking thunderhead and tornado formation). But his most extended research project in his later years was the search for an explanation for a plane crash. On June 24, 1975, an Eastern Air Lines 727 was on final approach to Kennedy International Airport during a violent thunderstorm; without warning the plane swerved out of its flight path, struck the approach lights, burst into flames, and broke up on the ground. Of the 124 people on board, only 11 survived. Eastern Air Lines asked Fujita to find out what had gone wrong. Fujita’s analysis of the data revealed a mystery: the plane had been buffeted by an extraordinarily violent gust of wind that hadn’t been recorded anywhere else on the airfield. So where had the wind come from? He was struck by a radical new idea. Suppose an extremely concentrated downdraft had formed in the storm cell directly overhead. It might have been so localized that it wasn’t even detectable a few hundred yards away, but the wind shear of its gust front, for the few seconds it lasted, would be strong enough to knock a commercial jetliner out of the sky. Fujita called it a microburst.
Fujita made use of a new piece of technology to test his theory—the Doppler radar. This was an advanced radar design that did more than passively reflect back the simple presence of objects; it discerned the way these objects were moving relative to the radar signal. When directed at storms, it allowed for the first time a fine-grain analysis of the complex movement of rain and winds within a thunderhead. Fujita spent the next several years using Doppler radar units to search for microbursts. Most of his colleagues were skeptical—if not openly contemptuous. It seemed absurd to them to suggest that there could be such a strong, complex, and violent phenomenon taking place within a thunderstorm that had somehow gone totally unnoticed throughout all of history. They were openly astonished when Fujita emerged from his multiyear study with Doppler images of microbursts erupting from storm cells, falling with extreme speed, and rapidly spreading and dissipating across the ground directly below, like milk poured out from a carton onto a tabletop. Fujita, with the backing of the entire meteorological community, used his evidence to lobby the aviation industry and the government for Doppler radar to be made standard equipment at airports.
Fujita always had his opponents. Few seem to have disliked him personally, but there were always complaints about how he got his results. He was not one for precision. He would guess. This made the exact calibrations of damage in the Fujita scale something of a sham, since his criteria for assessing the tornado were actually impressionistic, even poetic—what was, after all, the objective distinction between “devastating” and “incredible” damage? (After his death, the scale was revised, and it is now customary to refer to a tornado not as an F2 but as an EF-2, for “Enhanced Fujita.”) Fujita was also notorious in his later years for refusing to use computers in order to model tornadoes. Computers, he liked to say, simply didn’t understand these things.
At bottom he was another intuitionist, like Robert Miller, like John Finley, like James Espy. He was relying not on a ground of theoretical or mathematical knowledge but on a kind of imaginative sympathy for the storm itself. He made no secret of his fascination with the unquantifiable. Only the wildest and most free ranging of intellects could have conceived the dynamics of a thunderstorm in such a way as to suggest the existence of the multiple-vortex tornado and the microburst, but brilliantly illuminating as these ideas were, Fujita’s models for them were hard for other scientists to use. They weren’t so much convincing as stunning.
Fujita didn’t entirely care. He said he would be more than satisfied if only half his ideas were found valuable. He called his memoirs An Attempt to Unlock the Mystery of Severe Storms, and he never claimed that his attempt had been successful. There were many aspects of tornadoes that he never fully engaged with. There was in particular the one issue that everybody else still believed was the sole and essential question: the exact mechanism of tornadogenesis. Fujita offered several models for it, but none of them were met with any lasting enthusiasm. He could also be reluctant to shed his own ideas when they were proven wrong. He had, for instance, begun with the general assumption shared by most meteorologists: that tornado damage was the result of a near vacuum within the funnel that caused buildings to explode. But there was never much empirical evidence to back up this theory; Fujita’s own maps of damage tracks showed no sign of any exploded buildings. It gradually became obvious to the younger generation of meteorologists that Fujita and his colleagues had been mistaken, but Fujita himself, well into his sixties and seventies, remained convinced of the existence of a deadly vacuum at the tornado’s heart.
But this was Fujita’s great legacy: freeing up the minds of meteorologists to reconceive the nature of violent storms from scratch. It’s understood now that tornadoes are so destructive not because of some catastrophic differential in air pressure but from the simple effects of wind. While there is an area of extremely low pressure within the tornado, this is incidental. There is in fact only one irreducible element in any tornado: the vortex that draws in winds from the surrounding atmosphere. It’s the astonishing strength of these winds—technically known as surface inflow jets—together with the barrages of flying debris they carry with them, that causes all of the tornado’s damage. Everything else we think of as characteristic of tornadoes—in particular the funnel cloud itself—is a secondary effect that is easily occluded or may not appear at all. The tornado isn’t a peculiar kind of cloud but a certain configuration of the air, a moving vortex within the storm, which in its purest form is invisible.
In fact, as Fujita’s followers and successors began to think of it, there was something fundamentally misleading about conceiving of tornadoes as distinct phenomena. They are only an aspect of the fantasticall
y complex and violently evolving dynamics of a supercell thunderstorm. They are rarely singular. They form in clusters and waves; they breed and die off within the larger movement of a storm front like bubbles in the froth. The number of tornadoes that form or half form, that blur and merge and separate within any given storm cell, defies any exact count. The sheer chaos of a severe storm renders precision impossible.
This way of thinking about tornadoes has suggested a dire possibility to contemporary meteorologists: that true prediction of tornadoes might never be a practical reality. The problem of tornadogenesis will doubtless be solved, but the solution might not help. It could turn out that large, rapidly unfolding storm systems simply can’t be modeled in sufficient detail to predict which cells will or won’t spawn tornadoes and in what number. The ultimate goal of tornado forecasting, then, could turn out to be a will-o’-the-wisp.
But the study of tornadoes sooner or later breaks down any and all certainties. Even the most obvious truths about them eventually turn out to be backward. For decades after 1974, meteorologists took it as a self-evident fact that the Super Outbreak was a freak outlier, a “five-hundred-year event” (meaning something that might possibly happen once or twice a millennium). This certainty endured despite a growing mountain of evidence that major storm systems can routinely generate dozens of tornadoes in very brief periods of time. As a single storm system passed across Oklahoma and Kansas on May 3, 1999, for instance, there were 66 tornado touchdowns. Then, in April 2011, over the course of three days, a gigantic storm system sweeping across the eastern United States and Canada spawned more than 350 tornadoes. It became known as the Super Outbreak of 2011, and nobody talked any longer about five-hundred-year events. Instead, the suspicion grew that Super Outbreaks might not be rare at all but something closer to the norm.