Weather Watch: Icing's learning curve

The first noteworthy icing encounters came in the mid-1920s, when basic flight instruments were developed and the U.S. Air Mail Service pioneered the first mail route between New York and Chicago. This area was—and still is—prime icing country, and of all the hazards faced by Air Mail pilots, icing ranked highest. There were no ice-protection systems yet, so once ice began to cripple an airplane, the only solution was to bail out and parachute to the ground. The lucky ones lived. Charles Lindbergh, a mail pilot on the Chicago to St. Louis route, bailed out twice. Talk about having an icing escape strategy!

By the late 1920s, the U.S. Army Air Corps was growing, and the government decided to fund icing research. The National Advisory Committee for Aeronautics (NACA, NASA’s predecessor) built the first icing wind tunnel. Test results came pouring in. They seem obvious now, but were breakthroughs then. Heated pitot tubes prevented ice accumulations. The temperatures and humidities conducive to icing were discovered, as were the shapes that icing formations could assume, along with their effects on lift and drag. Anti-icing agents—corn syrup, oil, paraffin, honey, glucose, Vaseline, and goose grease—were tested. They didn’t work (but the corn syrup showed promise). One experiment was prescient: Scientists channeled exhaust heat through a pipe mounted behind a biplane’s wing leading edges. Another test discovered that fuel tank vents could ice over, causing fuel tanks to collapse. That’s how we got “NACA vents,” with their ports submerged in the wing skin, where ice could not accumulate. Carburetor icing was also studied.

B.F. Goodrich invented the first inflatable leading-edge deice boots, and by the 1930s they were installed on the growing fleets of fledgling airlines. The results were mixed. In the aviation classic Fate Is The Hunter, author Ernest K. Gann is a rookie co-pilot flying a Douglas DC–2 over Knoxville, under the tutelage of a Captain Hughen. The DC–2 is at 7,000 feet, and icing up. “The leading edge of the wing is now one long, unbroken bar of ice,” Gann writes. The ship has propeller and windshield alcohol deice protection—another invention of the day, much better than corn syrup—but that’s not the problem. It’s the boots. “Yes, the boots are working. But they are expanding and contracting beneath the sheath of ice and consequently useless.” It’s the first noteworthy mention of a phenomenon called ice bridging, but certainly not the last. Gann and Hughen survive, but only after losing 3,000 feet and 50 knots, and then restoring full power by intentionally backfiring the engines and climbing to on-top conditions. Haven’t read the book? Get a copy. You won’t regret it.

Icing intensity definitions came about in the 1940s, when the then-U.S. Weather Bureau mounted three-inch-diameter test cylinders on the Mount Washington, New Hampshire, weather observatory. By documenting the time it took for various ice accretions to form, we got the now-familiar trace-light-moderate-severe spectrum. The trouble was, the tests were abstractions, assuming a “typical” airplane flying at 200 mph/174 knots, a single wing leading edge cross-section, and accumulations defined by grams per cubic meter of supercooled water per hour. More baselines for icing definitions came out from the U.S. Air Force in 1956; yet another came out of a joint civil-military publication in 1964; still another in 1968; and still others in 1981 and 1998. Some included operational considerations in their definitions. For FAR Part 135 operations, FAR Part 135.227 even allowed for operations in known or forecast “light” or “moderate” icing conditions if equipped with certain ice-protection equipment. As for flights under FAR Part 91, operations in icing conditions are not addressed. But common sense should rule. We should all know by now that flying in icing conditions is never a good idea—no matter what kind of rules you fly under, your aircraft type, or your certificate or rating levels.

Each aircraft responds to icing differently, based on wing chord and profile as well as many other aerodynamic properties. What may be trace icing on a Piper Aztec’s big, fat, Clark Y airfoil may rate as moderate on a slim Mooney wing. And if you linger long enough in trace or light icing, it may well become moderate or severe, depending on the nature of the ice buildup—rime, clear, mixed, or large-droplet.

Yes, large-droplet. The October 1994 crash of an American Eagle ATR–72 turboprop in Roselawn, Indiana, led researchers to discover a new type of especially hazardous form of clear icing. Large-droplet icing conditions can cause droplets to splatter and run far back on the wing chord, forming ridges of ice aft of areas protected by deice boots or leading-edge panels heated by engine bleed air. Even worse, the ridges created low pressure areas ahead of the ATR’s ailerons, causing them to deflect and create asymmetric wing stall. It didn’t help that the crew used a notch of flaps to help lower the ATR’s unusually nose-high attitude, either. Thanks to the Roselawn crash, we now see predicted areas of supercooled large-droplet (SLD) icing published on many aviation weather websites—a practice begun by the Aviation Weather Center (www.aviationweather.gov).

There have been other discoveries. The January 1982 takeoff crash of Air Florida Flight 90 out of Washington, D.C.’s (then) National Airport revealed several causes that impact the way we fly today. The airplane—a Boeing 737—was improperly deiced prior to taking the runway for takeoff. It also had snow and ice accumulations on its wings and other surfaces. The pilots failed to turn on the heat to the engines’ pressure-ratio-sensing probes. As a result, the crew saw false high-power readings, believed them to be correct, and thus accelerated slowly during takeoff, and stalled during the initial climb. Now we have more detailed rules concerning “holdover” times between ground deicing intervals when it’s snowing, or when supercooled droplets are falling.

NASA’s Glenn Research Center in Cleveland has also been yielding especially productive safety information. Its March 2000 report on its studies of tailplane icing proved that ice accumulations on horizontal stabilizers can cause tail stalls, which in turn cause an airplane’s nose to pitch suddenly downward. The danger is especially great when an airplane is at increasing power settings and with flaps extended. The recovery actions are counterintuitive: aft control column, power to idle, and flaps up. Information about the phenomenon was widely distributed to the pilot community.

And what about ice bridging? It doesn’t exist—at least not now. The proceedings of a joint NASA/FAA De-Icing Boot Ice Bridging Workshop, published in November 1997 with test information from longtime ice-protection manufacturers B.F. Goodrich and France’s Aerazur, concluded that there was “no evidence” of ice bridging. Gann’s DC–2 had boots that operated at lower pressures, and could take up to 30 seconds to fully inflate. That’s long enough for ice to accumulate on inflated boots, but today’s boots inflate (and deflate) faster, and work best when inflated at the first sign of ice. Automatically cycling boot systems usually shed ice successfully on the first inflation cycle. Any ice that doesn’t shed on the first cycle continues to increase in thickness and sheds subsequently.

True, many pilot operating handbooks and airplane flight manuals continue to advise letting various levels of ice accrete before inflating deice boots. Usually, they say to wait until ice builds to the range of a quarter to a half inch before activating the boots. But this appears to be as much a carryover from the past as it is a nod to FAA recommendations that manufacturers provide at least some guidance on the issue. But icing wind tunnel and inflight testing is conclusive: Ice bridging is no more. The experts say that those who insist they see it are really seeing residual ice that hasn’t yet shed.

Icing continues to give up its mysteries. These days, the NASA Glenn Center is examining what’s thought to have caused several cases of double-engine flameout in jets flying near thunderstorm tops. The suspicion is that tiny ice crystals emanating from the cloud tops can enter a fanjet’s engine intakes, melt, and then refreeze as ice buildups on stators and other internal components, disrupting the airflow to the point that it halts combustion. The scary thing is that this sort of icing happens at the periphery of storm cells, at high altitudes—precisely where you’d be if you were circumnavigating thunderstorms. And yet, we know very little about this new form of icing.

Or has it been around all along, and we’re only now aware because of some fairly recent inflight coincidences? It makes you wonder. Icing has been around as long as we’ve been flying, and yet we still seem to know so little about some of its darker corners.

Email [email protected]