Landing an Iced-Up Airframe
If you fly in honest-to-god weather, sooner or later you'll have to do it. Here's a survival guide from a 12,000-hour veteran test pilot.
As the prime winter icing season once again approaches, many of us will be confronted with this sinister hazard. Every year, almost without fail, there are between 30 and 40 accidents involving icing, about half of them fatal. As we've pointed out in previous issues, by heeding the pireps and taking decisive action at the first sign of ice, the icing risk is manageable, especially if you accept the notion that on some winter days, you'll simply have to cancel your flying. The risk of serious icing will be too great.
In this article, we'll examine some of the aerodynamic considerations of flying and landing an iced up airframe. But don't get the impression that I'm suggesting these techniques make it safe to fly in ice. Far from it. I'm offering these observations strictly as a survival guide if you have to put an ice-laden airplane onto a runway some day.
The Great Unknown
Most pilots have heard this caution: When your airplane is carrying ice, you're a test pilot. If you've accumulated a lot of experience in flying iced up airplanes — whether certified for known icing or not — you might not take this warning too seriously. After all, if you've had ice dozens or even hundreds of times and survived it, the warning must surely be an overstatement. Maybe. But I wouldn't count on it.
Permit me a war story. In my flying career, I've been both a giver and a receiver of airframe and engine ice. Back in my Air Force test pilot days, around 1971, I flew the KC-135 water spray tanker, which we used to douse various airplanes to study the effect of airframe icing. The object was to control the amount of ice build-up up on the receiver aircraft to determine its flying characteristics and to see how well it would shed ice. Even though these tests were done under carefully controlled conditions and flown by real test pilots, the results were sometimes unpredictable.
We had been asked to fly the spray tanker over to England, to help the Brits with icing certification of the Concorde. Our flight trials were going well until about the fifth flight, when the British team was trying to determine the maximum amount of ice the engines could handle. We were at 16,000 feet, with the water spray giving them a good load right into the number two engine when suddenly, we heard, "Uh-oh, we have a slight problem here in the Concorde."
The engine had stalled and surged and the crew decided to shut it down. A ground inspection revealed several of the compressor guide vanes had sheared off and gone through the engine. Even 25 years ago, that was a $2 million engine and I doubt if the consortium had budgeted for that. The Brits decided they'd had enough icing tests, thank you. They later certified the airplane using natural icing. The point is, the outcome of that icing test was entirely unexpected, even though it was done under controlled conditions. If you pick up more than a trace of ice, the same may be true for your airplane.
Obviously, the best way to avoid an unpredictable outcome is to stay out of ice in the first place. When the pireps confirm that it's widely present, stay home, drive or go commercial if your only other choice is to fly an unprotected airplane. If you do encounter ice that continues to accumulate, don't hang around waiting for it to stop accreting. Formulate a plan right now. A couple of years ago, when we reviewed 170 icing accidents for an article, we found that many pilots underestimated both the rate of accretion and how it would affect aircraft performance.
In more than a few of these accidents, pilots reported icing to ATC then declined to divert or declare an emergency until it was too late. The accident data strongly suggests that once ice has accumulated to the point that the airplane will no longer maintain altitude, the chances of making it safely to an on-airport landing are poor. Given that the majority of icing accidents seem to involve experienced pilots, it's reasonable to assume that pilots fall into the trap of concluding that one icing event is just like the next. The facts suggest otherwise. Ice — and its effects on airframe and engine — is extremely variable. Just because you've survived 99 icing events, doesn't mean you'll survive the next. Resist the instinct to tell the controller you don't have a problem. If you've got ice, you've got a problem.
Drag and AOA
Even pilots with lots of experience flying in ice don't always understand the aerodynamic penalties of hauling around a load of it. Ice adds both weight and, more significantly, tremendous drag; cleaner airfoils on high performance airplanes may be more efficient collectors of ice and will suffer more from its effects.
Attaching meaningful numbers to the damage ice does to lift and drag is difficult, since it varies with airplane and airfoil. However, icing research done by Dennis Newton and reported in his excellent book Severe Weather Flying, revealed that typically, even a small buildup can reduce the maximum coefficient of lift by 30 percent. This has the effect of decreasing the stall angle of attack, which translates to a higher stall speed. Newton says that even a 1/8-inch buildup raised the clean stall speed of one airplane from 69 knots to 80 knots. A further accretion of 1 1/4 inches increased the stall speed only another 4 knots.
Drag keeps on building up with further accretion, however, and this is what paints many a pilot into a corner from which there is no exit. As drag increases with a continual buildup of ice, the power required to maintain cruise airspeed or even an airspeed above the stall also increases, to the point that even full throttle won't do the job. If this happens, the pilot has no choice but to enter an involuntary descent and hope for the best. Further aggravating the angle-of-attack/drag issue is the fact that in icing conditions, the engine probably won't be capable of delivering rated power. In a carburated engine, carb heat should be on in icing. But the warm air going into the induction system is less dense and reduces the engine output. Fuel injected engines with manual or suck-open alternate air intakes generally draw air from inside the engine compartment and it too is warmer than ambient, resulting in somewhat lower power yield. Ice accumulation on the propeller is no help, either. When these factors are taken together, even a moderate accumulation on an underpowered airplane such as Cessna 172 or a Cherokee can soon result in the inability to hold altitude.
A note on prop de-ice: The theory that a clean prop will haul around a lot of ice, even if the wings aren't protected by boots, is a dangerous misconception. Newton's data revealed that power loss from iced props amounts to about 9 percent, but is rarely more than 20 percent. The additional power required to move an iced wing through the air, with its higher drag, may be as much as 250 percent.
It's long been known that small radius or sharp-edged surfaces are more efficient collectors of ice than are blunt or large radius objects, such as the wing's leading edge. That's why the first place to look for ice is on a sharp-edged projection, such as an OAT probe or a protruding fuel vent. It's also the reason that the tailplane's relatively small leading edge may collect ice twice as fast as the wing. The consequences of this have been understood only recently, following a spate of accidents in which tailplane stalls were suspect.
To understand the tailplane stall, recall your private pilot ground school sessions in which flight mechanics were discussed. The center of lift of the main wing is such that a stable aircraft has a natural nosedown tendency. The horizontal stabilizer's job is to counter this, by exerting downward lift and thus noseup moment. As does the main wing, the tailplane has angle-of-attack limitations, but in reverse to those that apply to the wing. This means that if the tailplane stalls, it essentially stops making the downward lift needed to counter the nosedown moment. Result: The aircraft may pitch down uncontrollably.
Knowing that ice effectively lowers the stall angle of attack on the main wing, it's easy to see how this works on the tail, too. But since the tail is a more efficient collector of ice, it may reach the stall angle-of-attack before the wing catches enough to be a problem to the pilot. In fact, it's conceivable that the tailplane could have enough ice to stall it, when none is present on the wings.
Flaps aggravate the problem. In all airplanes, flap deployment produces downwash that affects the tailplane and elevator. In some airplanes, this downwash is so pronounced that it produces a local airflow that changes the tailplane angle of attack substantially. Ice acts like a stall fence, encouraging the onset of tailplane stall. Tailplane stalls have gotten a lot of press since 1991, when NASA and the FAA conducted a conference on the topic of tailplane icing. Since then, several accidents have been directly linked to tailplane stalls and, retrospectively, other heretofore unexplained accidents may also have been the result of tailplane stalls. Most recently, the NTSB investigated two accidents involving British Aerospace Jetstream turboprops in which tailplane icing was suspect. A third Jetsteam accident in December 1993 took place in icing conditions but, in the final analysis, ice wasn't considered a factor. Other airplanes, including the ATR 42, Saab SF340A and Embraer EMB-100 have been the subject of ADs having to do with tailplane sensitivity in icing conditions.
So much for turboprops. What about the Cherokee and Bonanza crowd? Are these airplanes as susceptible to the hazards of tailplane icing as are the turboprops? Do any light airplanes have a particularly nasty history of tailplane stall incidents and accidents? The short answer is we don't know. Nothing excuses light singles and twins from the same aerodynamics that apply to turboprops but we are unaware of any body of research that shows that any of the 30 or 40 icing accidents each year are due to tailplane stalls.
Our review of four years' worth of icing accidents (1974-75, 1984-85) revealed that about half were fatal crashes. There's nothing typical about fatal icing accidents, other than a pilot encountering more ice than he or the airplane could handle. Most of the fatals do result in uncontrolled flight into terrain but who can say if these were the result of tailplane stalls, wing stalls and spins or some other factor? Unless the pilot survives to tell the tale, accident investigators don't have a lot to go on and they're generally reluctant to speculate.
Not too surprisingly, a twin is more likely to make it to a runway after a severe icing incident than a single is. But once over the threshold, a landing hard enough to damage the airplane is often the result. Many of these pilots reported being unable to flare, even to the extent of pitching to a level attitude. Lack of noseup pitch authority or very heavy elevator forces are consistent with a tailplane stall. But again, given the paucity of data in the accident reports, that's but an informed guess.
Approach and Landing
So how do you translate these aerodynamic basics into a survival strategy for approach and landing? The important thing to remember is this: With ice on an airframe never certified to fly in it, you really are a test pilot.
First, I'll assume you've accumulated ice and, rather than trying to get out of it, you've decided to divert and land ASAP. That's usually a good choice and one that should be made sooner rather than later. The longer you wait, the fewer options you'll have. If you even suspect a diversion may be necessary, start planning for it and have plates at the ready. Know what the weather is along the route and keep up with it by listening to ATIS reports and/or checking with FSS regularly. (While you're at it, pass along any icing — or lack thereof — pireps for your fellow pilots.)
Tell the controller right away what your plans are. If the controller asks "are you declaring an emergency?" think very hard about your answer. By declaring, you'll be given priority and you'll get out of the ice that much sooner. If you decline, the controller will handle you first-come, first-serve, and that may mean a roundabout vector and more ice. As we've said so many times before, it's a myth that declaring an emergency results in a lot of paperwork or regulatory hassles.
If you have a choice, pick an airport with a good approach, preferably an ILS with approach lighting. Having accurate course and vertical guidance will be a real help, especially if the windshield is iced over and difficult to see through. Furthermore, an ILS runway will usually be 5000 feet or longer, enough length to land with minimum or no flaps, which is the recommended procedure.
As far as descending for the approach is concerned, that will take some judgment and what you want to do may not square exactly with how ATC wants to run the program. Depending on how the airspace is set up, the controller's normal procedure might descend you to 2000 AGL on a vector 10 or 12 miles from the airport. But if the ice is still building and the freezing level is right to the surface, maybe you shouldn't give away that altitude so far out.
This is doubly true if you can barely hold altitude going into the approach in the first place. Consider joining the approach higher than normal and descending on or slightly above the glideslope, in a steady, controlled descent with power on. In any case, even if you're visual, don't execute a long, dragged in approach with periodic level-offs. You may find out too late that you don't have the power to arrest the descent.
On the other hand, don't try or accept a slam dunk. A brisk descent through an ice-bearing layer is one thing, but a banshee dive for a lower altitude invites an abrupt pull out and increases the chances of descending right through the MDA. That's apparently what did in the crew of the Express II Jetstream in Hibbing, Minnesota in December, 1993.
Speed and Configuration
How fast to fly the approach? Those with expertise in the field may not agree on the exact numbers but everyone agrees on this: fly it faster than normal. Dennis Newton recommends a 20 percent increase, my own guideline is 10 to 20 knots faster. Even with a tiny coating of ice, you won't know what the stall speed is but if you fly the approach normally, you may learn the hard way that it's much higher than you imagined.
If 20 knots extra is good, won't 40 knots be better? Perhaps not, for two reasons: First, you've got to land out of that approach. Crossing the threshold 40 or 50 knots fast with flaps up will eat up a lot of runway. And if there's ice on the airplane, the runway may be icy, too, or at least wet, further complicating your ability to stop. Second point: The higher the speed, the greater the risk of a tailplane stall. At the higher speed, the main wing angle of attack is lower, meaning the tailplane will have a higher negative angle-of-attack. And like the main wing, if the tailplane has ice on it, it will stall at a higher angle of attack.
As for flaps, again no universal recommendations, other than using full flaps is a bad idea. Full flaps aggravate the tailplane stall situation and they may require a much larger pitch moment to round out and flare for landing. Besides, if you really don't need them, they're just another surface to collect ice. Very few singles are certified for known icing, so chances are the manufacturer won't have flown the aircraft with iced wings in any configuration. You're on your own.
Newton cautions against using any flap setting you haven't first tried with some altitude beneath you. Since it may not be convenient to conduct any testing while in an icy stratus layer, that might mean landing with no flaps. So be it. There aren't many singles that you can't land safely with no flaps and this probably applies to a lot of piston twins, too. Next time you're practicing approaches, add flapless landings to your routine, so you'll know what to expect. Aircraft that are icing certified may or may not have flap recommendations. If yours does, follow the book. If it doesn't, use no flaps or no more than half flaps. Gear can be extended normally, unless you're worried that extending it too soon will produce an unarrestable descent rate. If that's the case, you'll have a difficult choice. Extend it and hope you make the runway before settling short or leave it in the wheels until crossing the threshold or simply plan to land gear up. That may be a better choice than plowing into the approach lights short of the runway.
The Home Stretch
Our review of icing accidents showed that quite a few airplanes, especially twins, make it over the threshold safely, only to suffer damage as the result of a hard landing. As I said, this could be due to tailplane stalls, wing stalls or failure to see well enough through an iced windshield to arrest the approach descent.
To avoid this, keep the power on and the airspeed up right across the threshold to the runway. Fly the airplane onto the runway with only enough roundout to arrest the descent. Don't worry about a two-point touchdown, a greaser or, least of all, a full-stall landing. Flying an airplane onto the runway with no flaps will feel fast but if you maintain directional control while you decelerate, you should have plenty of room to stop.
If you go visual well out from the airport, pick the longest runway, wind permitting. But be very, very careful about circling close in. With iced wings and a higher stall speed, even a moderately banked turn could stall one wing, resulting in loss of control at low altitude. Keep the turns shallow.
As you plan your flights this winter, remember that icing is rarely a bolt-from-the-blue surprise. It's forecast more often than it's present but when it is there, it's usually expected. An inadvertent encounter with ice needn't been the gut-wrenching terror we imagine it to be. Just have a plan of attack and be ready to exercise it and you'll come out just fine.