When a Team of Meteorologists and Combat Pilots Set Out to Understand Thunderstorms, They Made Flying Safer for Everyone

The sky was a very dangerous place in the early days of commercial aviation. By flying into storms to learn how they worked, these experts made air travel and weather forecasting much more predictable

Summer 2026

OPENER - Lightning strikes in Peckham, Oklahoma. — Lightning strikes in Peckham, Oklahoma. Mitch Dobrowner

The 22 passengers who boarded Pennsylvania Central Airlines Flight 19 out of Washington, D.C. on August 31, 1940, would have had little reason to worry about arriving safely at their final destination. It had been 17 months since the last fatal commercial airline accident in the United States, a record at the time. The flight’s captain, Lowell V. Scroggins, had 11,000 hours of flying experience, more than seven times what the Federal Aviation Administration now requires for airline pilot certification. And the aircraft, a rugged DC-3, the era’s quintessential airliner, had flown in that morning from Detroit without incident; the day before, it had undergone a routine inspection of its propellers, wings, radios and all other critical equipment. Even its interior had received a good scrub and polish. Just before 2:30 p.m., Flight 19 fired up its twin 1,100-horsepower engines and began its ascent.

About 55 miles northwest, in Lovettsville, Virginia, Dorothy Everhart was tinkering in her home when suddenly, according to an incident report later compiled by the newly established Civil Aeronautics Board, she heard what she thought was lightning strike her house. Everhart shut off the electricity and stepped onto the back porch. To the west, she could see black clouds assembling over a clump of mountains. A lone airplane was headed straight into the storm. Everhart lived along a well-trafficked flight route, and she was accustomed to airplanes overhead, but this plane was flying “lower than most of them go,” she recalled. Then a brilliant flash bleached the sky, momentarily blinding her, but she couldn’t miss an “awful roaring.”

Another Lovettsville resident, Lydia Jacobs, heard the strike, too, followed by a sound resembling a “scream” or a “siren.” When Jacobs peeked out her window, she saw a streak of fire shooting across the clouds. Flight 19 appeared to Jacobs as “a burnt-up building floating through the air.” Within moments, Flight 19 plunged to the ground and slammed into a nearby alfalfa field, where it was demolished on impact, killing everyone on board.

At the time, Flight 19 was the deadliest commercial airline accident in American history. But it was not unique. Soon after Flight 19, a string of fatal accidents followed. On November 4, 1940, a DC-3 flying from Oakland, California, to Salt Lake City was caught in a snowstorm and struck a mountain, killing all ten people on board. A month later, a DC-3 stalled out while trying to land in overcast and icy conditions in Chicago, killing all three crew members and seven of its 13 passengers. On January 23, 1941, a DC-3 crashed after an aborted landing during bad weather in St. Louis, where one passenger and one crew member died. The next month, five passengers and three crew members aboard a DC-3 en route to Georgia were killed while landing in rain and fog.

The wreckage of Pennsylvania Central Airlines Flight 19 after a lightning strike. NOAA

Mechanical errors, delays, detours and cancel-lations were common in those days, but no threat was greater to commercial air travel than the weather. Scientists knew that clouds could be turbulent environments, but there was little guidance for pilots about how to navigate through them, and meteorological forecasting was still a primitive science. Pilots would take off expecting smooth rides only to stumble into dark clouds: At best, the consequences might be a misadventure through stormy skies that left passengers nauseated (in the 1930s, airliners were routinely hosed down between flights); at worst, the results were catastrophic.

Flight 19 happened to be carrying Senator Ernest Lundeen of Minnesota, and its crash and the many accidents that followed spurred American government officials into action. In March 1941, Oklahoma Representative Jack Nichols assumed the chair of the House Select Committee to Investigate Air Accidents. The committee investigated for the next two years, compiling a series of reports that together amounted to 15 cubic feet of paperwork. Throughout, the evidence highlighted a pervasive and exceedingly familiar natural force: thunderstorms. In fact, thunderstorms were later found to be responsible for 56 airplane accidents between 1938 and 1945, a staggering number considering that commercial air travel was radically curtailed in 1942 to concentrate resources on the war effort.

Thunderstorms were as enigmatic as they were destructive, seeming to manifest suddenly, flooding towns and spawning tornadoes that ripped through America’s heartland. They churned violently at the Earth’s surface and whipped rain and hail. Even the brightest meteorological minds barely understood them, and to begin to decipher their inner workings would require abundant and diverse data that simply eluded the capabilities of ground-bound meteorologists: precise readings of temperature, humidity, wind speed, precipitation, degrees of turbulence, and the rate at which these factors varied across space and time. And the best way to collect that data was to seek out the worst storms and, in the name of science, fly airplanes straight into them.

The herculean government initiative that followed would prove to be one of the most influential scientific efforts of the 20th century. Given an appropriately swashbuckling name, the Thunderstorm Project drew on the expertise of America’s top weather scientists and the know-how of the U.S. Army Air Forces, the U.S. Navy, the U.S. Weather Bureau (which became the National Weather Service) and the National Advisory Committee for Aeronautics (the precursor to NASA), and it helped transform the rudimentary practice of airborne weather forecasting into a precise science of the sky. Commercial air travel was made demonstrably safer, and an untold number of lives were saved. The global economy as we know it grew from the reliability bequeathed to us. Mike Kurz, a forecaster for the National Weather Service, has described its findings as “the cornerstone of today’s understanding of thunderstorms and related weather phenomena.” Remarkably, the project has also largely been lost to history. “It’s not taught in college, when you’re going through meteor-ology,” Kurz told me recently. “I’ve gone back and looked in some textbooks, and it’s literally, in some, a footnote—-maybe a brief paragraph.”

What does remain in a historical record cobbled together from incident reports, oral histories, congressional hearings, scientific publications and interviews with contemporary storm chasers illustrates a dynamic portrait of American ingenuity. It’s a story about the vast potential of government action and cooperation across civilian and military agencies. And it’s a story about people for whom ignorance and hope and superstition were not an option when it came to tackling a problem that North Carolina Representative Alfred Lee Bulwinkle, among the initiative’s strongest supporters on Capitol Hill, described as a “three-dimensional one of baffling complexity.”

When, in 1945, Francis Reichelderfer, chief of the U.S. Weather Bureau (and, as a former competitive hot air balloonist, no stranger to storms), was tasked with finding the right leader for the Thunderstorm Project, he knew just the man for the job.

Thunderstorms first captivated Horace Byers in late summer 1929, while he was riding a train from Berkeley, California, where he had just graduated college, to Cambridge, Massachusetts, where he would begin graduate studies at MIT. A magnificent storm roared continuously over the Great Plains, and Byers watched through black-framed glasses, mystified by the unending crackle of lightning.

At MIT, Byers studied under the Swedish meteorologist Carl-Gustaf Arvid Rossby, a disciple of the “Bergen School,” named for the Norwegian city where its founders developed their theories. Bergen School scientists had discovered that storm systems were composed of “fronts,” masses of cold and warm air with predictable life cycles. Under Rossby, Byers became one of the nation’s top forecasters, and by 1940 he was established at the University of Chicago. During World War II, Chicago became a hub for training military weather officers, and Byers impressed Reichelderfer by helping develop weather training and research initiatives, including forecasting for wartime planning.

Reichelderfer named Byers the Thunderstorm Project’s director by August 2, 1945, before the war was even formally over. Byers’ first task was to recruit the best minds to answer a seemingly straightforward question: How do thunderstorms behave? At the time, scientists estimated a thunderstorm’s size in part by observing its basic visual dimensions—that is, by its height, width and depth. But these measurements did little to indicate how bumpy a flight through a given storm would be. Byers set out to discern the storm’s fourth dimension: intensity.

To understand that, Byers and his researchers needed to decipher a storm’s internal structure, so they began recruiting former World War II pilots to seek out the most dangerous parts of the sky. They would do so in P-61 “Black Widow” airplanes, which had been designed to fly harrowing night missions hunting down enemy bombers. Now the pilots would hunt storms, penetrate them and—if they survived—take careful notes of what they experienced. “No storm was to be avoided because it appeared too large or too violent,” Roscoe Braham, one of the project’s leads, later recalled. All told, the Thunderstorm Project recruited 11 pilots and 7 navigators for this uniquely precarious mission.

Key takeaways: Modern air travel safety was forged by flying directly into danger

Between 1938 and 1945, thunderstorms were linked to at least 56 airplane accidents in the United States. In 1946 and 1947, the U.S. government’s Thunderstorm Project recruited combat pilots to fly hundreds of perilous missions into storms in planes armed with scientific instruments.

Scientists found that thunderstorms follow a predictable life cycle and contain especially turbulent rain-heavy cores. They also discovered navigable, relatively stable corridors within them. Crucially, delaying takeoff by even five minutes could avoid peak danger. The project proved the critical worth of radar in turning forecasting into a practical tool that could make routine flight safe and reliable.

A squadron of Black Widows prepares for takeoff. NOAA

Louis DePas checks wind vane readings at the University of Chicago while Horace Byers, on the ground, corroborates the measurements on an automatic recorder. University of Chicago Photographic Archive, Hanna Holborn Gray Special Collections Research Center, University of Chicago Library

Byers and his colleagues narrowed down what they believed were the main characteristics “reflecting the character and intensity of the thunderstorm process.” Storms, they wrote, contained updrafts and downdrafts (“the lifeblood of the thunderstorm, although little was known of the downdraft”); horizontal winds and temperature gradients; electric fields and variations in rainfall and intensity; changes in air pressure and fluctuations in wind speed at ground level; and lastly, overall levels of turbulence and gustiness.

To measure these phenomena, the Black Widows were affixed with a suite of instruments, from photographic recorders and sensitive temperature gauges to “airborne electric fields meters.” But by far the most important instrument was a newish technology called radio detection and ranging, better known as radar. Radar emits radio waves that bounce off objects and reverberate back to the source, allowing scientists to calculate the distance and relative position of those objects using a returning signal’s travel time and other measurable properties. The technology had been developed in secret before World War II and had been critical in wartime planning. The British relied on radar during the Battle of Britain to detect incoming German bombers, and in 1941 they discovered that they could use radar to track rain, snow and water vapor. The next year, scientists at MIT wrote to Reichelderfer that they had used radar to identify thunderstorms from as far as 160 miles away. By early 1944, many Army Air Forces fighters and bombers were equipped with nascent radar systems.

Byers was already well aware of radar’s forecasting potential. He had helped direct a training program for wartime weather radar officers and had even mentioned the promising technology in his forthcoming meteorology textbook (before he was ordered to remove it because of national security concerns). For the Thunderstorm Project, he believed radar would prove immediately valuable, not only for detecting storms but as a means of triangulating the vast web of data the pilots and scientists would collect. “The problem of tying all of these data together to obtain a unified picture of each thunderstorm required exact radar tracking or radio-direction finding in connection with planes and instruments,” Byers wrote later. For that reason, in addition to ground-based radar networks dispersed throughout the experiment sites, radar beacons were also housed in the noses of the Black Widows, so that the planes could be “located in time and space with respect to the thunderstorm.”

Radar officers monitor a hydrogen balloon’s real-time position to track wind speed and direction. NOAA

A “V-Beam” radar antenna is installed in Wilmington, Ohio. NOAA

Finally, to measure the lower levels of storm systems, where the planes couldn’t fly, Byers’ scientists would use a range of unusual instruments deployed from the ground. One was a mechanical hygrometer (early versions resembled a violin), which measured the air’s relative humidity using a strand of human hair—specifically blond hair, which was especially reactive to humidity, expanding when it encountered moisture and contracting when the air was dry. Another was the “kytoon,” a balloon affixed with small, kite-like panels, allowing it to stay aloft in gusty skies while measuring conditions.

By early summer 1946, Byers and his team had selected two sites for experiments. The Black Widows would first take flight later that summer over Central Florida, a region known for its unusually frequent thunderstorms. The following summer, operations would move to the southwestern Ohio town of Wilmington, which also reported frequent thunderstorms and was close to two military airfields (in 1948 they merged to become Wright-Patterson Air Force Base).

The only thing left to do was await the forthcoming storm season and then fly into the belly of the beast.

During the next two summers, Black Widow crews—an Army Air Forces pilot and a navigator plus, by 1947, a third crew member for helping with observations—made hundreds of rocky passes through storms, bouncing and shuddering through turbulence that routinely sent them rocketing or plummeting more than 500 feet in an instant. The pilots and navigators strained not only to keep the planes level and their instruments functioning but also to frantically note their observations, scribbling them in notebooks or speaking into tape recorders.

Because Byers and his scientists required data from multiple points within a storm, each mission involved five Black Widows flying in a vertical stack formation at 5,000-foot intervals. The flights could be harrowing. One August afternoon in ’47, 20,000 feet over southern Ohio, First Lieutenant Tom Mahon felt his eyeballs jitter in his skull as his retrofitted fighter plane was tossed around by an especially wily storm. He could just make out the time, 2:37 p.m., when a gust blasted his airplane with an uppercut of such violent precision that it flipped the plane nose over tail. Standing vertically in the air, Mahon’s Black Widow stalled out, and he nearly tumbled back to earth.

First Lieutenant Joe T. Hargett, flying 5,000 feet below him, smashed into the same turbulent wave. But rather than causing Hargett’s plane to stall out, the gust transformed it into a rocket ship, exerting bloodcurdling G-forces on the crew as they shot 5,000 feet upward into the storm in less than a minute. First Lieutenant Robert L. Smith, flying a third plane at 10,000 feet, collided with that same force and was bounced upward 2,000 feet. Miraculously, after a second engine-stall, Mahon wrestled control of his airplane and guided his crew to a safe landing. The other four P-61s gritted their way through the mission, managing to continue recording observations through every phase of the storm. (Incredibly, no one was killed during the project’s two-summer span.)

Black Widows before separating to capture meteorological readings at varied altitudes. NOAA

A crew member points out damage to a P-61 “Black Widow” from a storm flight in June 1947. — A crew member points out damage to a P-61 Black Widow from a storm flight in June 1947. U.S. Air Force

With each successive flight, a clearer picture of thunderstorms emerged. The storms were not chaotic or random; they followed a particular, albeit violent, life cycle. Roscoe Braham, a cloud physicist studying under Byers in Chicago and a former World War II bomber pilot himself, was critical in deciphering this cycle. With the aid of radar, which allowed him to coordinate data collected by the planes and the ground operators, Braham identified three distinct stages of a thunderstorm’s life: the cumulus stage, when its winds began to swell upward; the mature stage, when the storm was its most fearsome; and a final stage, after about 30 minutes, when the storm weakened and eventually died out. Learn to identify a storm’s stage, the scientists realized, and they could predict how it would behave. This led to a signature finding: The ground beneath a thunderstorm experienced the most violent gales just after heavy rain began to fall, which caused winds to bounce off the Earth’s surface and then spin and accelerate; the closer one was to the storm’s core, the more devastating the winds. Because takeoffs and landings were (and still are) the most dangerous parts of flying, if airlines could be alerted that a thunderstorm was brewing near an airfield, they could delay their flights accordingly. Byers and Braham concluded that delaying a takeoff by even five minutes could reduce the likelihood of encountering strong turbulence, which would in turn dramatically reduce the chances of an accident.

The project’s analysts also discovered that pilots could adjust the sensors on their radars to pinpoint the heaviest rainfall in a storm, where turbulence was most likely. Because radar beams bounce off water droplets, pilots and researchers could locate not only the presence of oncoming storms but also specifically where within the storm rain, hail and ice were most prevalent. They observed that radar would send back an “echo,” indicating the most highly concentrated areas of water within the storm. The pilots then corroborated these radar findings with their own observations of turbulence and concluded that turbulence was often worst where rainfall was heaviest. (The turbulence in a storm is caused not by rain but by what causes rain—strong updrafts of wind. This moisture-laden air then cools to form rain, which is forced back down to the ground in powerful downdrafts.)

Much to their surprise, the scientists also learned that a thunderstorm, which could cover anywhere from 20 to 200 square miles, was not always one storm. It could comprise multiple cells, each of which might brew its own mini-storm system. If a 200-square-mile storm sounded daunting to air travelers, what the scientists discovered next inspired hope: Although the cores of thunderstorms could be exceptionally rough, the areas between these cells were relatively calm. This finding would allow pilots and ground operators to identify turbulent hot spots inside a storm and navigate around them with minimal disruption. What’s more, it would transform night flying, as radar could read storms in complete darkness.

Byers, Braham and the rest of the team spent two years piecing together their findings. Probably their most significant finding pertained to radar. “Radar is, without question, one of the best instrument aids to thunderstorm flying currently available,” they wrote in their book-length report, released in June 1949. With sophisticated radar and trained radar operators, they wrote, the sky could be traversed not only safely, but with a reasonable degree of predictability.

The final report did not point to major revelations from the ground-released instruments: Thunderstorms proved too powerful even for kytoons, as Byers noted that overwhelming winds at lower elevations rendered any measurements “difficult and impossible to evaluate.” Still, one can appreciate the ebullient creative effort and experimental resourcefulness of Byers’ team in using everything at their disposal to try to wrangle the storm.

Despite the project’s accomplishments, its recommendations were not immediately implemented by a commercial airline industry that was rapidly expanding after the lifting of wartime travel restrictions. It would take a number of near misses, and one tragedy, to mark a turning point. One such near miss was a September 1954 air traffic jam so terrible that it was dubbed “Black Wednesday”: 300 airliners were stuck circling in bad weather above New York, delaying some 45,000 passengers. Then, in June 1956, two airliners crashed into each other 21,000 feet over the Grand Canyon, killing 128 people. Investigators determined that seemingly innocuous fluffy clouds likely played a role in the pilots’ inability to see each other. The incident prompted the creation of the Federal Aviation Agency (now the Federal Aviation Administration), and long-range radar instrumentation and surveillance equipment to monitor air traffic were soon standardized in airports across the nation.

After the Thunderstorm Project, Byers returned to the University of Chicago, where he remained until 1965. The meteorologist R.H. Simpson (after whom the category-based Saffir-Simpson Hurricane Wind Scale is partially named), described Byers as “the balance wheel in the administration of one of the greatest meteorology programs the world has ever known.” Among the scientists Byers mentored was Tetsuya “Ted” Fujita, whom Byers recruited from Japan and who came to be known as “Mr. Tornado” for his groundbreaking research on tornadoes (the EF scale for measuring tornado strength is named after him: Enhanced Fujita).

Byars — Byers was "the balance wheel in the adminstration of one of the greatest meteorology programs the world has ever known." University of Chicago Photographic Archive, Hanna Holborn Gray Special Collections Research Center, University of Chicago Library

Byers, who finished his career overseeing the College of Geosciences at Texas A&M, retired in 1974. But his final meteorological contribution came three years later, when he co-authored a paper with Fujita analyzing an airplane crash at New York’s Kennedy airport in June 1975. The plane experienced “unusually sharp wind changes under a thundershower,” they wrote—the result of a “downburst,” a phenomenon that Byers and his scientists had never observed during the Thunderstorm Project. In the wrong conditions, Fujita and Byers found, the ultraviolent downward gust could cause an airplane to undergo an “unexpected sinking,” an especially dangerous problem when already flying slow and low to the ground. Such downbursts, they discovered, could be detected—and thus avoided—on radar, in the form of rare “spearhead echoes.”

The finding was its own kind of echo to Byers’ initial work on the Thunderstorm Project, when he and Braham improved decision-making for takeoffs and landings. Because the method sometimes calls for delays—even short ones—when a storm is brewing, today we may take such precautions for granted, or worse. Stuck on a tarmac, awaiting the all-clear, perhaps pulling out our phones to track a storm threat in real-time on our pocket radar devices, we rarely appreciate why such safeguards were imposed to begin with. Yet on the whole, the sky can now be traversed with an exceptional level of safety. Hulking airliners ferry millions of Americans every day, crossing continents and oceans with such routine consistency that commercial air travel’s harrowing infancy has largely been replaced by its mundanity—the monotony of its food, the inadequate in-flight entertainment, the endless waiting. Today, to be aloft in the sky is not merely expected. It’s boring.

Few people have much awareness about whom they have to thank for that. But you can find small testaments to Horace Byers and his team if you know where to look. One is a historical marker in St. Cloud, Florida, the site of the project’s first phase of test flights. Another is in rural southwestern Ohio, the site of the second. Last spring, I drove out to Wilmington, Ohio, where I found a marker tucked away on a back road across the street from a local air park, which serves as a home base for cargo carriers like UPS and Amazon Air.

Coincidentally, a severe thunderstorm was forecast for that afternoon. I could see the clouds twisting and the sky beginning to darken. No sooner had I made it back to my hotel than I heard a loud rumble and crackle. There was a pause, as if the sky were taking in a deep breath. Then it exhaled. Winds began whipping, and rain started falling in sheets.

I stepped outside. For a moment, I closed my eyes and listened: whooshing wind and pattering rain and arrhythmic, percussive blasts of thunder and lightning. But something was missing. All day, big, heavy cargo planes had flown overhead, emitting a steady drum of white noise. I’d gotten used to the whine of their engines.

Now a forecast had grounded all air traffic, and there wasn’t a plane in the sky.

Cover image of the Smithsonian Magazine Summer 2026 issue

Subscribe to Smithsonian magazine now for just $19.99

This article is a selection from the Summer 2026 issue of Smithsonian magazine

When a Team of Meteorologists and Combat Pilots Set Out to Understand Thunderstorms, They Made Flying Safer for Everyone

Subscribe to Smithsonian magazine now for just $19.99

You Might Also Like

Get the latest Science stories in your inbox.