War On Ice

No one could doubt Robert Fry’s aviation experience.

He had already made national headlines in 1928 as a Marine pilot in China, when he wrestled his plane through a violent storm and made an almost impossible forced landing in the middle of a shocked force of hostile Chinese troops. There was therefore no question that he knew how to handle the TWA Fokker trimotor aircraft that departed from Kansas City, only a little delayed on the stormy evening of March 31, 1931.

Besides Fry’s copilot, Jess Mathias, the Fokker carried six male passengers, most of them businessmen of some importance. But one was considerably more famous than the others, and in a much different area. Knute Rockne, the legendary Notre Dame football coach, was on his way to Los Angeles to make an instructional football film and perhaps also to schmooze Hollywood producers, with an eye toward getting other film work.

Not long after takeoff from Kansas City, Fry found himself in rapidly worsening conditions. Ice began to collect on the wings and struts, affecting the plane’s handling and stressing the wooden structures. Before Fry could even think about getting out of the storm, one wing ripped away and the aircraft plummeted into a wheat field near Bazaar, Kansas. The occupants died instantly, their bodies and the wreckage burning almost beyond identification. Rockne’s remains were found with a rosary wrapped in his fingers. A nation grieved for the gridiron hero.

The cause of the accident—aircraft icing—is one of the most intransigent and treacherous enemies of all air travelers, “a hazard for aircraft ever since Wilbur and his brother first flew,” says Gene Hill of the Federal Aviation Administration (FAA). Ice rarely brings an airplane down from its weight but acts more insidiously by reducing a plane’s ability to fly. When a foreign material such as ice forms on a wing, it changes its shape, disrupting the flow of air and thus the lift capacity that makes flight possible. Ice also increases drag, the resistance to the smooth flow of air—effects made even worse because ice on an airplane can be hard to detect and adequately quantify. It is conceivable that Fry did not even know that ice had accumulated on his wings.

Paradoxically, a flight bearing revolutionary deicing technology that, if used, might have prevented the accident, lifted off the day before the Fokker left on its fateful flight. A Lockheed Vega test aircraft, Miss Silvertown , which took off from Akron, Ohio, bore a “boot,” a thin, oil-coated rubber sheet affixed to the leading edges of its wings, the brainchild of William Geer, a retired former research chemist with the B. F. Goodrich Company. The oil coat helped to retard ice formation, while the boot could be inflated temporarily to break off the ice that did form. Geer’s boot worked flawlessly. Miss Silvertown flew safely through the worst winter ice conditions that Akron offered.

The New York Times declared that the pneumatic boot system had brought victory over icing, and airlines enthusiastically embraced the innovation. For years their schedules had been completely dictated by the vagaries of the atmosphere; they flew only in perfect conditions. But a plane on the ground was a plane not earning money. With Geer’s deicing system, airlines would no longer have to divert or delay flights just for a little wintry weather. By 1936 the president of TWA proclaimed that the danger of ice had been “virtually eliminated.” He would soon regret that statement.

In March 1937 one of TWA’s own DC-2s, its ailerons and wingtips coated with ice, went down near Pittsburgh, killing 16 people—a brutal reminder that deicing boots were not a silver bullet. To this day, definitive solutions to this devilishly tricky menace have proven maddeningly elusive. Although the basic weapons against icing were developed early in the struggle, the growth of the aviation industry into a year-round, all-weather enterprise has forced a continual, often painful process of refinement and reevaluation. More than once, the combatants have declared victory, only to have their confidence shattered by another tragedy. This saga has proceeded in fits and starts over the entire century of heavier-than-air flight, involving scientists, engineers, regulators, and aircraft manufacturers.

Ice takes three basic forms: clear or “glaze,” which, as the name implies, is transparent; the opaque and milky “rime”; and “mixed,” a combination of the two. The type that forms depends on air temperature and a number of factors. It can develop year-round: as long as air temperature is at or below freezing and some liquid water is present, ice can form from ground level up to about 20,000 feet (in the cloud-forming regions of the atmosphere).

Contrary to popular belief, aircraft are not endangered by ice crystals or snow in the air, which are swept around and past a plane by the airflow. The culprit is liquid water in various forms, either freezing or “supercooled”—actually below the freezing point but still liquid. When an airplane flies through a cloud containing these forms of water, the droplets condense instantly upon contact with the plane’s surfaces, and ice accumulates.

Although ice can present a hazard to any type of airplane, a particular aircraft’s response to ice is determined by myriad variables, including its design, the power of its engines, its anti-icing or deicing equipment, specific weather conditions, and even the skill and experience of its pilot. What works for one aircraft may not work for another, or even for the same plane in different conditions.

Nor does it take much ice to cause serious trouble. In 1939 Jerome Lederer, the first safety director of the Civil Aeronautics Board, noted that “strange as it may seem, a very light coating of snow or ice, light enough to be hardly visible, will have a tremendous effect on reducing the performance of a modern airplane,” a warning that still rings true today. As the National Transportation Safety Board pointed out in a recent alert to pilots: “Research results have shown that fine particles of frost or ice, the size of a grain of table salt and distributed as sparsely as one per square centimeter over an airplane’s upper wing surface, can destroy enough lift to prevent that airplane from taking off.”

Defense against icing comes down to two basic strategies: either preventing ice from forming on an aircraft in the first place (anti-icing), or removing it after it has formed (deicing). Anti-icing usually entails using heat, either electrically generated or bled from the plane’s engines; deicing involves melting or breaking up already formed ice by a number of mechanical or chemical techniques.

The first phase of the fight against ice began shortly after World War I, as maturing civil aviation required pilots to fly in less than perfect conditions. Airmail flyers were among the first casualties, their small aircraft overwhelmed by wintry conditions that brought them swiftly to the ground. Charles Lindbergh reported that ice was the greatest danger he faced on his historic 1927 crossing of the Atlantic—enough at one point to make him seriously consider turning back. The U.S. Air Mail Service and Army did some basic research in the mid-1920s, but their resources could not compare to those of the National Advisory Committee for Aeronautics (NACA), which turned its attention to the issue in 1928.

NACA built a small refrigerated wind tunnel at its Langley Memorial Aeronautical Laboratory in Virginia to study how ice formed on an airfoil. Researchers began to experiment with ways of preventing or interrupting the process. Initially they tried coating the test wings with various substances that might retard ice formation: glycerin, grease, different oils, goose fat, soap, molasses, honey, wax, and even corn syrup, which, oddly enough, was the only material that offered some protection, though not quite enough for NACA to recommend its use. NACA scientists also conducted flight tests on a Fairchild F-17 monoplane modified so that its engine exhaust heat was diverted to an airfoil mounted under the wing. Although both the heat and chemical ideas showed promise, some daunting technical difficulties remained: the weight of most engine exhaust heating schemes proved too much for the still mostly wooden airplanes of the era, and effective chemical solutions, let alone a way to apply them efficiently, had not been worked out. NACA had reached something of a research impasse, but it continued to work on the problem at a low level. For the time being, as two of its research pilots, Thomas Carroll and William McAvoy, reported, “safety . . . lies in avoidance.”

Lewis Rodert, a brash, somewhat abrasive NACA engineer, and his colleagues at Langley kept pressing for answers. A strong advocate of the thermal deicing approach, Rodert was also a firm believer in real-world flight tests over tunnel research. In several aircraft, first at Langley and then later at the new Ames Aeronautical Laboratory in California, he set out to perfect a workable thermal deicing system, a goal made even more pressing by the advent of World War II. Unlike civilian flyers, bomber crews and fighter pilots could not limit their flying to acceptable weather.

NACA’s test pilots set out to fly into the worst conditions possible. On one of their early flights in a specially modified Lockheed 12A, they encountered “violent turbulence, snow-and-rain static which stopped radio communication, and occasional dangerous electrical discharges,” as Rodert put it in his official report. On another flight, the plane was struck by lightning that partially melted a propeller and elements of the airframe. Sometimes ice coated and broke off radio antennas, interfering with communications, and it often made windshields nearly opaque. Yet Rodert and his pilots remained steadfast in their determination to get the work done. There was a war on, and lives to be saved.

Rodert developed a system that ducted heat from the Lockheed’s exhaust into the leading edges of the wings. This technology was applied to the B-24 and B-17 bombers, while efforts continued throughout the war to improve and modify it for other military and civilian aircraft. Rodert’s team also worked on keeping ice from forming on propellers and windshields, not to mention in carburetors, where it could stop engines by choking off airflow. In 1947 President Harry Truman awarded Rodert aviation’s most prestigious award, the Collier Trophy. Truman had a special interest in Rodert’s achievements because his presidential aircraft, the Douglas DC-6 Independence , happened to be one of the first production aircraft equipped with a thermal deicing system. The trophy committee declared that Rodert’s work meant that “ice has been virtually eliminated as a major menace in air transportation”—all too eerie an echo of the words of TWA’s president a decade before. These words would prove to be just as premature.

Not all the deicing action took place in the stormy skies of Northern California during the 1940s. When in 1941 NACA opened another new research facility, the Lewis Engine Research Laboratory near Cleveland, it set out to include a dedicated refrigerated wind tunnel, known as the Icing Research Tunnel (IRT), the biggest and most advanced of its kind. NACA engineers tapped the air-conditioning genius Willis Carrier to devise a means of chilling 10 million cubic feet of air inside the tunnel to as low as -40 degrees. At these temperatures, small droplets of water sprayed into the airstream formed ice on any model or surface placed in its path. When the IRT opened in 1944, engineers and scientists began studying propeller icing and devised a new air scoop for C-46 transport planes, whose engines had been choked with ice while flying over the Himalayas between India and China.

By the dawning of the jet age in the late 1940s, airlines, the military, and aircraft manufacturers relied on three basic weapons against ice. In addition to heat and boots, British engineers during World War II had developed a system known as “weeping” wings, in which anti-icing chemicals were exuded from tiny holes on the wing. On the ground, planes were sprayed with heated solutions of glycol and water under high pressure to remove ice and snow and delay their buildup.

New aircraft and engine designs meant that these techniques needed reevaluation. In certain ways a jet aircraft’s greater power made them less susceptible to icing than propeller aircraft; they also spent more time at higher altitudes where ice was rare, and they generated lots of handy excess heat that could be used for thermal deicing. But different airfoils and plane configurations accreted ice in new ways.

At the IRT, researchers started to codify icing “envelopes”: the specific temperatures, droplet sizes, cloud types, and other parameters that combined to create icing conditions. These data enabled the Civil Aeronautics Administration (the precursor to the FAA) to formulate a set of regulations in 1953 distinguishing the types of aircraft that would—and would not—be allowed to fly into icy weather.

By the middle of the 1950s, cautious yet confident talk again appeared about the end of the aircraft icing problem. Deicing boots and thermal systems were sophisticated and effective. As historian William Leary writes, “Ironically, the very success of the NACA’s icing research led administrators to question the need to continue the program.” The first era of icing work had ended.

Throughout the 1960s and 1970s, researchers conducted little work on icing. In the meantime, icing still continued to claim victims. On January 18, 1960, engine icing brought down a Capitol Airlines plane in Virginia, killing 50. Wing icing doomed 24 on a DC-3 flight in Chile in April 1961, as well as 12 on an Icelandair Vickers Viscount at Oslo, Norway, in April 1963. Aviation experts considered these to be statistical aberrations, mostly attributable to the bad luck or poor judgment of pilots flying into difficult weather.

As the aviation industry continued to grow, it began to dawn on pilots, engineers, and designers that the final word on the subject of icing had yet to be written. A number of new commuter and small air businesses were setting up shop, most of them not operating huge jets but smaller turboprop planes, which flew at lower altitudes where ice was more prevalent. “There was a whole host of new problems to deal with in terms of different kinds of aircraft, such as rotary wing and commuter aircraft,” writes Leary. A conference held at Lewis Research Center in 1978 led to the revival of NASA’s icing research program and the modern era of the icing wars.

The National Transportation Safety Board (NTSB) issued a major report in 1981 that linked 178 accidents between 1976 and 1979 to icing problems, more than half of them involving fatalities. The report noted that icing incidents might be underreported because some crash investigations lacked the evidence to be conclusive, a fact particularly true for the decades between World War II and the creation of the NTSB in 1966. While allowing that earlier research results and the resulting FAA regulations were “still basically valid,” the report recommended that the FAA review its icing regulations and requirements.

On January 13, 1982, an Air Florida 737 attempted to take off from National Airport in Washington, D.C., in the middle of a major snowstorm. It had undergone routine ground deicing, but repeated delays and the temporary closing of the airport had kept it on the ground—snow steadily falling—for almost an hour after the last procedure. When finally cleared for takeoff, visibility was only a quarter of a mile. The airplane lifted off briefly, stalled, and went down, skidding over the 14th Street Bridge and crushing four cars and five people before plunging into the Potomac River. Of the 78 people aboard, only five survived, pulled from the icy river by helicopter in a daring televised rescue. The NTSB held ice to be the prime cause of the crash.

Engineers at NASA’s Lewis Research Center returned to the issue with renewed vigor, intending not to invent wholly new ice protection systems but to better understand the process of icing in general. By working to understand the physics of ice formation, the specific meteorological conditions that caused it, and how ice affected aerodynamic performance, they believed they might better predict icing and avoid it.

The Icing Research Tunnel was therefore back in business. Models of airfoils and aircraft sections were subjected to winds of up to 400 miles per hour and supercooled clouds of water droplets at temperatures as cold as 40 below zero inside a test section measuring six feet high by nine feet wide by 20 feet long. A tiltable turntable enabled models to be mounted in a great range of positions to achieve various angles of attack.

The second major thrust of the program involved aggressive flight research: inserting scientists into situations in which they could study icing in its natural state, as well as giving test pilots the chance to experience flying in icy conditions. A series of flights collected new data for comparison with those from the 1940s and 1950s, helping to validate and refine the FAA’s certification requirements. The flight testing confirmed that ice tunnels accurately simulated natural conditions.

A de Havilland Twin Otter became NASA’s icing workhorse research aircraft, making its maiden flight in 1981. A twin-engine turboprop, loaded with various deicing and anti-icing equipment along with cameras and meteorological and aerodynamic instruments, it was “anything but glamorous . . . kind of like taking a ’57 Dodge station wagon to the prom,” noted pilot Bill Rieke. A typical mission carried one to three researchers in addition to the pilot and copilot.

“Every now and then, you would get yourself into something that either hadn’t been forecast or nobody knew about, and it would get pretty bad pretty quickly,” recalls test pilot Richard Ranaudo. On one flight, ice built up so rapidly that “the drag of the airplane had gone up something like one and a half times,” and Ranaudo had to abort that particular test in a hurry, even before making the planned measurements. “Had we lost an engine under those circumstances, there’s a very good chance that we would have been swimming back from Lake Erie that day,” he reflects.

The NASA team also investigated Electro-Impulse Deicing (EIDI), a system developed in the 1970s but originally conceived in 1937 by Rudolf Goldschmidt, a London-based German expatriate, whose novel concept involved generating a pulse of electromagnetic energy to knock ice off an airplane. Although Goldschmidt patented the idea, he had never built a prototype and tested it. Soviet scientists revived the idea in the 1970s. Films of EIDI tests show ice simply vanishing from wings. Some again believed that the magic bullet of ice protection had finally been discovered.

“Actually, it worked very well,” says Ranaudo, who piloted many test flights with the system. “My experience was that it was far superior to pneumatic boots.” But EIDI had some serious problems, particularly because it needed to be built in, at considerable expense, as part of a wing’s structure. According to Bill Rieke, it also “makes a hell of a racket. It’s analogous to taking a hammer and beating the inside of the leading edge of the wing.” Mostly, it may have been an instance of an idea being too far ahead of its time.

Icing continued to throw curve balls at scientists. “With new testing techniques,” observed the FAA’s Gene Hill, “we [were] seeing phenomena now that we were not aware of”—a bitter truth that became abundantly clear on October 31, 1994, when American Eagle Flight 4184 crashed in Roselawn, Indiana. The NTSB investigation, along with tests in the IRT, flights with the Twin Otter, and other analyses, confirmed that the aircraft had succumbed to a poorly understood and rare phenomenon dubbed “supercooled large droplets” (SLD). Nonscientists simply call it freezing rain.

Most ice on aircraft derives from water droplets with an average diameter of about 50 microns or less (about that of a human hair). SLD starts at that diameter and can attain diameters of 1,000 microns, or one millimeter. “These large drops can form ice behind where the protection exists on the wing,” explains Tom Ratvasky, an icing research engineer at NASA’s Glenn Research Center in Cleveland. Thus, even on an airplane equipped with the best deicing or anti-icing systems available, ice can still form on the wings and flight surfaces farther back from the leading edge, beyond the protected areas. Because the American Eagle ATR 72 flew through large droplets long enough for sufficient ice to accumulate toward the rear of the wings, ridges of ice formed in front of the ailerons, out of range of the deicing boots, destroying roll control.

The SLD problem wasn’t the only variation of the icing threat to materialize in the 1990s. Another major concern was tailplane icing, in which ice accretes to the horizontal stabilizer and affects pitch control, causing an airplane to nose down suddenly. Not completely unknown before, it began occurring more frequently as the numbers of turboprop commuter aircraft grew throughout the 1990s.

Ironically, the advancing sophistication of modern aircraft can actually increase their vulnerability to SLD and tailplane icing. Not only do they spend more time flying at the more dangerous lower altitudes, but their thinner and more advanced wings tend to collect ice more readily than older designs. “Some aircraft designed with these modern materials are designed for a very high degree of aerodynamic efficiency, so keeping them ultraclean is very important,” observes Ranaudo. Wing size is another factor. “Small aircraft accrete ice at a proportionately higher rate than large aircraft,” explains Mike Bragg, director of the University of Illinois Icing Research Group. “So they’re going to have more problems in terms of larger effects on aerodynamics.”

More flight and tunnel research at Glenn Research Center helped engineers to understand both SLD and tailplane icing and to develop ways for pilots to respond. Aside from the technological fixes, pilot awareness is an important aspect of the icing battle. “We’ve done a lot of work on developing education and training materials for different types of pilots,” says Ratvasky. Among videos and other materials, NASA has made interactive training courses on icing available to all pilots online.

A major recent breakthrough has been the development of sophisticated software, such as NASA’s LEWICE code, that can model the aerodynamics and physics of ice formation: “Probably one of the more significant investments we’ve made in the last 15 years,” says Thomas Bond, formerly director of Glenn’s icing branch and now chief scientist for environmental icing at the FAA. “We’ve spent a fair amount of effort in research staff time in developing computer models that represent how ice forms on the leading edge of a wing.” LEWICE comes in two- and three-dimensional versions, both widely used by aircraft and equipment manufacturers in their design and certification work.

Computers also offer the key to better simulation techniques for pilots. Now at the University of Tennessee Space Institute, Richard Ranaudo has worked with NASA and Birhle Applied Research of Hampton, Virginia, to develop a highly sophisticated icing flight simulator known as the Ice Contamination Effects Flight Training Device (ICEFTD), which uses a database compiled from all those Twin Otter flights to give pilots a feel for how icing cripples a plane—and to practice ways of getting through it. Ranaudo and his partners are also developing software and systems that can not only detect the aerodynamic changes created by ice and alert a pilot to them in advance, but also recommend ways to continue flying the plane safely if ice has formed.

Better data on how ice forms and under what conditions have also facilitated the development of more sophisticated computer-based weather forecasting techniques. Vague forecasts of icing conditions over an area four or five states wide are not too useful to a pilot. “It’s really impractical to use as an avoidance technique,” Ratvasky says. “Pilots will look at it and say, ‘okay, there may be icing.’ But if we have a means for getting more refined information to a pilot, they can circumnavigate it.” Drawing on NASA research, the National Center for Atmospheric Research has devised new computer models that are making it much easier for pilots to locate specific icing conditions and avoid them. This is particularly vital for general aviation pilots flying smaller planes more susceptible to hazardous icing. Since the Roselawn crash in 1994, no large airliners have been brought down by ice—but there have been more than 100 fatal icing incidents involving lighter, smaller commuter and private aircraft. In the end, particularly for these more susceptible aircraft, the only foolproof defense against ice is to avoid it, just as NACA pilots Carroll and McAvoy wrote 80 years ago.

The IRT continues its frosty work today, as do other refrigerated tunnels around the world. NASA budget cuts have grounded the Twin Otter, at least for icing research, but that venerable aircraft will soon be replaced by a jet-powered Lockheed S-3 Viking, NASA having managed to pick up the former carrier-based submarine hunter for free when it was retired from Navy service. The S-3’s greater speed, power, and altitude capabilities will also help researchers investigate another major icing problem only recently identified: ice that forms inside an engine, bringing critical power loss and even failure.

For the foreseeable future, the traditional responses of boots, heated air, and ground chemical deicing will continue to be used, perhaps supplemented by newer strategies such as the application of “ice-phobic” materials. Meanwhile, the EIDI idea hasn’t gone away; in 2001 the Electro-Mechanical Expulsion Deicing System (EMEDS), which combines a form of EIDI with electrothermal deicing, became the first new deicing system to be certified by the FAA in 40 years.

“We’re still a significant way away from understanding the overall effects of icing in a very rigorous way so that we can predict the impact on an airplane,” says Bond. “We’re starting to figure out how to model those kind of events, but we still have a long way to go.”

Long way to go or not, no one argues that the icing research effort, and the deicing and anti-icing systems that it has produced, have saved countless lives. “It’s something that the public doesn’t know about where their tax dollars have done a lot of good,” Rieke points out. The battle against aircraft icing has always been a long, twilight struggle, and the irony is that, although we know all too well the names of its victims, we’ll never know just how many more lives have been saved. But even if final victory may never be achieved, the work of the small but dedicated international icing research community continues to bring solutions closer to the light.