Icing: the cold, hard (air) facts

On January 2 of last year, a Piper Seneca came to grief over Kentucky, killing four. The accident made national news because the only survivor was a 7-year old girl who walked three quarters of a mile through the woods, barefoot in the dark. I have not seen the final NTSB report, but the available information indicates that engine air induction icing may have been a factor.

On December 8, 2014, an Embraer Phenom 100 crashed into a Maryland home while on approach. The NTSB said “weather data indicate that the accident flight encountered clouds and was exposed to structural icing conditions.”

TBM crash in NJ
Icing brought down a TBM 700 in December 2014.

In December 2011, a Socata TBM 700 crashed in New Jersey while attempting to climb through icing conditions which were prompting a number of urgent pilot reports. Post-crash analysis revealed the outboard section of the right wing and several sections of the empennage had departed the aircraft while in flight due to an overload of the wing structure.

Icing is generally more of a hazard to smaller aircraft. There are technical reasons for this, but no aircraft is immune to the effects of accreted ice. My work in the flight simulation business eventually drove me to gain a better understanding of how in-flight icing occurs and why some types of icing are worse than others.

Icing on the ground is pretty straightforward. Any frozen precipitation on the upper surface of a wing or horizontal tail is a threat to the production of lift, and must be removed before takeoff. Once in flight, things are not so cut and dried. How much ice is too much? In my earlier flying days I tended to think a little was tolerable, and a lot was bad. Problem is, a little can become a lot rather quickly. There is a saying that goes, “In aviation there are three kinds of ice: good, bad, and hazardous. The good ice is found in the galley.” Most of us are not flying airplanes with galleys, but you get the point. So any ice on the outside of an aircraft is either bad or downright hazardous.

It is easy to assume that icing is only a problem in winter, but such is not the case. My most recent encounter was last August when flying over the Cascade Mountains of Washington. The basic ingredients for in-flight icing are supercooled water droplets and an accretion surface which is below freezing temperature. These can exist at any time of year.

“Supercooled” refers to the existence of water in a liquid state when at a temperature below zero degrees Celsius. This seems counterintuitive, but it is part of the physics of clouds. If the droplets are small, rime icing is the result. Rime ice occurs at lower temperatures with droplets which freeze on impact. Clear or glaze ice occurs at warmer temperatures with larger droplets which freeze gradually. Supercooled Large Droplets (SLD) are an extreme form of glaze icing and constitute the most serious hazard. Most real-world icing tends to be a mixed affair with Rime at one end and Clear at the other end of a continuum. This is because droplets are usually of widely varying sizes within any given cloud.

Another factor which determines the presence of icing conditions is airspeed. The faster the aircraft, the colder the air must be for ice to accrete because of the aerodynamic heating which occurs. Jet aircraft are less susceptible to in-flight icing because they fly faster and higher than propeller-driven aircraft.

Altitude matters because the air gets drier as you go higher. Icing is much less likely above a pressure altitude of about 22,000 feet. So jets can blow through an icing layer at greater speed and get above it. They also have better ice protection systems. Of course, those systems must be turned on and functioning to provide any benefit.

In-flight icing effects fall into several categories:

  • Drag increase
  • Thrust decrease
  • Stall speed increase
  • Inoperative sensors
  • Degraded controllability

The first two items on the list conspire together to reduce cruising speed and climb capability. Drag increases because the streamlined shapes of an aircraft are distorted by the accreted ice on the leading edges. The mechanism by which thrust is decreased depends on the type and nature of the propulsion system. Ice on a propeller greatly reduces its efficiency in converting horsepower to thrust. Engine air inlet or induction icing can result in reduced power output up to and including engine failure.

P210 boots with ice
Boots are nice, but you should avoid ice whether you have de-ice equipment or not.

It is commonly thought that a wing produces less lift when it is iced, and that is true when talking about maximum lift. The stalling angle of attack is reduced compared to that of an uncontaminated wing, generally speaking. A reduced angle of attack at stall means a higher stalling speed. However, above that stalling speed the lifting performance is largely unchanged. This is what leads to nasty surprises when slowing an iced-up airplane for approach and landing. Even in cruise flight, an unrecoverable situation can develop if the thrust/drag performance loss becomes great enough such that the airplane slows to what is now a much higher than normal stall speed.

Icing can render a pitot tube useless if it is not heated. An iced-over pitot tube will lock in whatever pressure was registered at the time. Any subsequent altitude changes will cause indicated airspeed changes because the airspeed indicator operates on the difference between pitot pressure and static pressure. In this instance a climb will produce an increase in indicated airspeed. The outside static pressure drops in the climb and the differential increases. Conversely, a descent will produce a decreased airspeed indication. It does not take too much imagination to see where this leads if it actually happens while in the clouds.

We learned about degraded controllability in icing the hard way during the 1980s and 1990s. There were a number of accidents attributed to Ice-Contaminated Tailplane Stall (ICTS) and one very famous accident which brought SLD icing into the spotlight. ICTS occurs when a horizontal tail accretes ice to the point where it can no longer counterbalance the lift generated by the wing during an approach to landing. In this case the tail does not stall in the same manner as a wing stalls. Rather, a tendency develops for the elevator to deflect downward all on its own. This is due to the low-pressure zone that develops on the lower surface of the tail when it is at a large but negative angle of attack. NASA once flight tested a DHC-6 Twin Otter in natural icing and inadvertently got into full-blown ICTS. It took nearly 100 pounds of pull force to counteract the nose-down movement of the elevator.

Recognizing and recovering from ICTS is tricky because of the differences between it and ice-contaminated wing stall. Not all airplanes are susceptible to ICTS. The FAA recently directed a review of FAR Part 91K, 121, and 135 training in this area to make sure that it is given only for those aircraft which are affected. Nearly all aircraft types certificated under FAR Part 25 (heavy aircraft) have been evaluated for ICTS, but only a limited number certificated under FAR Part 23 (light aircraft). T-tail configurations are much less prone to this situation.

An ATR 72 turboprop was lost on October 31, 1994, over Indiana after encountering SLD icing. The subsequent investigation revealed that ice had formed well aft of the deicing boots on the wing leading edges. The right wing must have accreted more ice than the left due to the slipstream effect of the propellers. When the boots were cycled, the ice was removed from the protected area but remained elsewhere. A ridge of ice built up on the right wing, causing a disruption of airflow over the right aileron. When the aircraft was slowed for approach, the angle of attack increased to the point where the right aileron deflected upward and the autopilot disconnected. This was followed by a rapid right roll and eventual loss of control.

SLD forecast map
The red areas indicate the potential for SLD.

SLD icing conditions are rare but can produce a hazardous situation very quickly. This is because the large droplets do not freeze immediately on contact, but instead travel some distance aft of the impact point before freezing. There is therefore a wide range of possibilities for airflow and sensor disruption. No aircraft is certified for flight in SLD icing conditions, because the regulations do not require testing at such large droplet sizes. All forms of icing are to be avoided as much as possible but an SLD encounter definitely requires an immediate exit strategy. The NOAA’s Aviation Weather Center specifically outlines the areas where SLD may occur on the Icing Forecast graphic.

You may have noticed that increased weight was not in the list of in-flight icing effects. Accreted ice obviously adds weight. At 57 pounds per cubic foot, a 3-inch thickness of ice across the entire frontal area of a Boeing 737-800 would weigh 403 pounds. That sounds like a lot until you compare it to a maximum takeoff weight (MTOW) of 155,500 pounds. That 3 inches of ice is only 0.26% of MTOW! Closer to home, a chunk of ice that is 2 by 3-1/2 inches across the wing span of a Cessna T210N would weigh 100 pounds, or 2.5% of a 4000-pound MTOW. So the percentage is higher than for a 737, but still not what you might expect.

I have only had a few icing encounters that were memorable. I intend to keep it that way. One was in a Cessna 414 while penetrating a cold front. The clear ice from the convective activity piled on quickly, and the pneumatic boots cracked it right off again. It was a non-event because all the ice protection was turned on and the conditions were short-lived.

Not so with a 182 that I was flying from Ohio to Oklahoma one November years back. We departed on an instrument flight plan and the headwind grew to 50 knots. It eventually became apparent that a fuel stop would be required. This was a problem in that I had not intended to make a fuel stop. The problem became acute as we began to pick up ice. Airspeed dropped off and a noticeable vibration set in. Realizing that I had to do something, I requested and received clearance to a higher altitude. Fortunately we were able to get on top of the cloud layer and the ice began to sublimate off. However, by then what was going to be an unscheduled fuel stop became a fuel emergency.

The ground speed had dropped even further due to the ice on the airframe and propeller. So declare an emergency I did and we were cleared straight into the Springfield, Missouri, airport with a downwind landing to boot. Of course I was asked to call the tower after I got on the ground. They were understanding about it once an explanation was given. There was a light twin in the area that day which did not fare so well.

Ice on King Air
Just a little bit of rime up there…

Inadvertent flight into icing often (but not always) requires prompt action. If the airplane does not have complete ice protection, the options are to climb, descend, or turn around. Climbing only works if the performance is still there to do it. Descending only works if there is enough room between the freezing level and the ground. It pays to think through the possibilities beforehand.

A flight manual may state that a particular aircraft type is not approved for flight into known icing conditions. OK, so what exactly are known icing conditions? More than two years after AOPA asked the FAA to reconsider its interpretation of “flight into known icing conditions,” in early 2009 the agency released a new letter of interpretation.

While the letter is somewhat lengthy and convoluted, it does among other things say that:

  1. The agency’s goal is to encourage proper flight planning in advance and to avoid unwarranted risk-taking.
  2. Flight which results in the formation of ice on an aircraft is not the sole factor the FAA will use in determining whether enforcement action is warranted in any particular case.
  3. If ice is detected or observed along the route of flight, the pilot should have a viable exit strategy and immediately implement that strategy.

In other words, use your head. If there is a chance of significant icing along the route, either change the route or have an escape plan ready to go at a moment’s notice. Sometimes the only right answer is to cancel the flight. I once agonized over a forecast for widespread icing conditions between me and my desired destination before giving up and getting in the car. That decision was justified when we got to eastern Colorado that evening and saw the trees covered in ice from freezing fog.

In-flight icing is one of the many risks we as pilots must manage in our flying. I consider it to be one of the Big Four weather factors in addition to thunderstorms, fog, and high winds. Having a good understanding of the phenomenon helps to devise effective strategies for dealing with it.


  • Good article, and very important information. Thanks for posting. The absolute worst I’ve seen in my career is around the Great Lakes region extending eastward to the Atlantic Coast.

  • Very good information. Do you happen to know why the T-tail configuration is less susceptible to ICTS? Perhaps the wake off the wing alters the droplet size?

    • The T-tail configuration is less prone to ICTS because it does not see as much downwash coming off the wing. Less downwash means a less negative angle of attack at the tail and therefore less chance of a tail stall.

  • From experience I know that it takes just one harrowing experience with ice to convince a pilot never again to mess with the stuff. Icing forecasts are great for go/no-go decisions, but once aloft be aware that if it’s near freezing or below and you’re about to enter cloud, the possibility of icing exists, and you’re on board weather or a controller’s scope may not pick up the severity of the conditions.

    How much ice can your plane hold and still be flying? Nobody knows. Before penetrating that cloud ask yourself: If ice starts building, do I go up? Down? Turn around? Also be aware that if you’re in densely controlled airspace your options may be severely limited.

    Before my frightening experience, I had done approaches in blinding snowstorms blithely unaware of the potential danger. Now there’s a very careful calculus I perform before entering a cloud in freezing temperatures.

  • Having flown around the Great Lakes, commercially, since 1970 I have been baptized more than a few times by ice. I live under an MOA at KAPN. It is difficult, when flight planing to determine if an MOA will be active when you are in the airspace. The only real way to know is to talk directly to ATC. If on departure I get a 4,000′ or 5,000′ restriction for the next 56nm I have the option of delaying the departure. Finding out the MOA is hot (7 minutes south of the MOA) and forced to descend to 4,000 for the last 56nm is cause for great caution during icing season, especially if your alternates are all in front of the aircraft. Foreflight does a great job of posting TFR’s. Perhaps they could highlight active, or soon to be active, MOA’s the same way? Happy Easter.

Leave a Reply

Your email address will not be published. Required fields are marked *