Airframe Icing: How It Forms and How to Get Out

Airframe icing is what happens when supercooled water freezes onto an aircraft in flight, and it kills lift, adds drag, and raises stall speed faster than most pilots expect. It comes in two unrelated forms that get lumped under one word: structural ice, which builds on the wings, tail, prop, and antennas, and carburetor ice, which forms inside the induction system of a carbureted engine and can appear on a warm day with the airframe completely clean.

Those two hazards share a name and nothing else. Structural ice needs visible moisture and a below-freezing surface. Carb ice needs neither. What follows is how each one forms, why the type of structural ice changes the danger, what ice does to the airplane, what “known icing” legally means, and your options once you are in it. For where icing sits among the other things the sky does to small airplanes, see our overview of the weather hazards that hurt GA pilots.

How structural ice forms

Clouds and precipitation routinely hold liquid water at temperatures below 0°C. Water that stays liquid below freezing is supercooled, and it stays that way until something disturbs it. Your leading edge is that something. When a supercooled droplet strikes the airframe, part of it freezes on contact and the rest spreads and freezes behind it.

Two ingredients are required, and both matter: visible moisture (cloud, rain, drizzle, wet snow) and a surface temperature at or below 0°C. Dry, clear air below freezing gives you nothing to collect. A cloud made entirely of ice crystals gives you very little, because ice crystals mostly bounce off rather than adhere. It is the supercooled liquid that sticks.

Most structural icing lives in a band from 0°C down to about -20°C. Below roughly -20°C, most of a cloud’s liquid water has already frozen to crystals, so accretion tapers off. The most productive slice for a light airplane, 0°C to -15°C, is exactly the altitude a lot of us cruise through in winter.

The worst case is freezing rain or freezing drizzle, which the icing world calls supercooled large droplets. Big drops carry more water and, because they are heavy and slow to freeze, they run back across the wing and freeze behind whatever the leading edge protects. That is how ice ends up on unprotected surfaces of an airplane certified to shed it, and why freezing rain is a get-out-now condition in anything a typical GA pilot flies.

Clear, rime, and mixed, and why the difference matters

The type of ice depends mostly on droplet size and temperature, and the type tells you how it will behave.

Rime ice forms from small droplets that freeze almost instantly on impact, trapping air. It is the milky, rough, opaque ice that builds forward into the airflow. It usually forms in the colder part of the band, roughly -15°C and below, though small droplets can produce it warmer. Rime is lighter and its shape is fairly predictable, but the roughness it adds to the leading edge is what does the aerodynamic damage.

Clear ice forms from larger droplets, warmer temperatures (about 0°C to -10°C), and slower freezing. The water flows aft before it sets, forming a smooth, dense sheet, sometimes with horns that grow into the airstream. Clear ice is heavier, holds on harder, and is easier to miss at night because it is transparent. It is generally the more dangerous of the two.

Mixed ice is the two together, typical of the -10°C to -15°C range, and it builds the opaque, lumpy accretions that combine clear ice’s weight with rime’s roughness.

The label shapes your escape. Clear ice from large drops usually means supercooled large droplets or a warm layer aloft, and the fix is often to get out of the wet layer entirely rather than nudge a few hundred feet. Rime in a stratus deck may clear with a small altitude change into drier air.

Carburetor ice is a different animal

Carburetor ice has nothing to do with the weather freezing onto your wings. It forms inside the carburetor throat, where two effects stack up: fuel vaporizing as it mixes with the incoming air, and the pressure drop through the venturi. Together they can cool the induction air by as much as 70°F. If that cooling takes moist air below freezing, ice forms on the throttle valve and venturi walls, chokes the airflow, and slowly strangles the engine.

The part that catches pilots is the temperature. Because the carburetor makes its own cold, carb ice forms on warm days. The FAA puts serious risk of fuel-vaporization icing roughly between 20°F and 90°F outside air temperature, with the greatest risk around 50°F to 70°F and high humidity. Carb ice can form on a 70°F afternoon with no cloud in sight, and it is most likely at low power, which is why a long descent with the throttle back is a classic setup.

The distinctions worth keeping straight:

	Structural ice	Carburetor ice
Where	Wings, tail, prop, antennas, airframe	Inside the induction system
Needs visible moisture	Yes	No, just humid air
Temperature	Airframe at or below 0°C	Ambient roughly 20°F to 90°F
First sign	Ice you can see, performance loss	Power loss (fixed pitch: dropping RPM), rough running
The fix	Change altitude, exit the conditions	Carburetor heat

Carb heat is the answer, and applying it early, before the RPM sags, is the whole game. Different skill from managing structural icing, same mental file: two ways ice takes an airplane down.

What the ice actually does

You do not need much. FAA and NASA testing found that ice or frost no thicker or rougher than coarse sandpaper on the leading edge and upper wing can cut lift by around 30 percent and increase drag by up to 40 percent. That is a barely-visible layer, not a coating you would photograph.

The reason is the boundary layer. A clean wing keeps air attached across the upper surface up to a fairly high angle of attack. Roughen the leading edge and the airflow separates early, so the wing stalls at a lower angle of attack and a higher speed than the book says. Your stall warning is calibrated for a clean wing and may stay silent right up to the break. The airplane can stop flying while the horn says nothing.

Ice loads the airframe in other ways too. It adds weight, costing climb margin you may already be short on, and can build asymmetrically and roll the airplane. Ice on the tailplane is its own trap: a contaminated horizontal stabilizer can stall when you extend flaps, pitching the nose down hard at low altitude. “A little ice” is not a stable, manageable state in a non-deiced airplane. Performance falls off faster than the ice builds, and the airplane you have five minutes from now handles worse than the one you have now.

What “known icing” actually means

“Known icing conditions” is a legal phrase, and pilots tie themselves in knots over it. The current interpretation is the FAA Chief Counsel’s 2009 letter to AOPA (the Bell letter, January 16, 2009). It rejects the idea that ice must already be confirmed on an airframe. Instead it applies a “reasonable and prudent pilot” standard: known icing conditions exist when the total weather picture available to you, PIREPs, advisories, temperatures, and moisture, would lead a reasonable pilot to expect ice to form along the route and altitude planned. The same letter restates the physics: structural icing requires visible moisture and a surface temperature at or below 0°C, so freezing temperatures in dry air are not known icing.

What may you legally fly into? It depends on the airplane. 14 CFR 91.527 forbids flight into known or forecast icing without functioning ice-protection equipment, but that rule applies only to large and turbine-powered multiengine airplanes. For the typical GA single, the binding rule is 14 CFR 91.9: you must comply with the operating limitations in the POH. Many light airplanes carry a placard or AFM limitation prohibiting flight into known icing, and violating that limitation violates 91.9 regardless of whether 91.527 touches your aircraft. An airplane with no ice-protection certification was never tested for it, so the expectation is simple: stay out.

“FIKI” means an airplane is certificated for Flight Into Known Icing. That certification is not a blank check. The ice-protection system was tested against a defined envelope of conditions, and the flight manual spells out the boundaries, sometimes a maximum exposure, sometimes intensity or airspeed limits. FIKI buys you time and options to leave icing, not permission to cruise in it. Freezing rain and supercooled large droplets sit outside most certification envelopes entirely.

Getting out

If you are picking up structural ice in an airplane that is not built to carry it, the plan is to leave the icing, not to ride it out. Every option is a change in altitude or direction to break one of the two ingredients, below-freezing temperature or visible moisture:

• Climb to colder, drier air above the moisture, if you still have the performance and the ice has not already taken your climb rate. This works when the layer is thin and you can top it.
• Descend into above-freezing air, if terrain, minimum altitudes, and cloud bases allow. Warmer air melts the problem. Often the surest exit in the flatlands, often unavailable in the mountains.
• Turn around. The air behind you was clear a few minutes ago. A 180 back to known-good conditions is frequently the fastest exit, and the one pilots delay too long.
• Declare and ask. If ice is degrading the airplane, tell ATC and ask for the altitude or heading you need. An icing encounter is a valid reason for priority, and controllers can relay where other aircraft have found clear air.

Two products make all of this decidable before and during the flight. Before you launch, the icing information in the graphical advisories, the G-AIRMET for icing and the occasional SIGMET for severe icing, marks the layers and areas to plan around; our guide to reading AIRMETs and SIGMETs covers how those advisories work. In the air, the single best real-time icing data comes from other pilots. A PIREP that reports the tops of an icing layer, or clear air at a nearby altitude, is worth more than any forecast, and filing your own when you find ice is how the next airplane gets the same head start. Our guide to reading and filing PIREPs covers both sides of that exchange.

Ice is one of the few hazards that gives you a running start and then closes the door quietly. Recognize which kind you have, know that a thin layer is already a real performance hit, and treat the first trace as the cue to leave, not the cue to watch.

Common questions

How much ice does it take to matter?

Very little. FAA and NASA testing found that ice or frost no thicker or rougher than coarse sandpaper on the leading edge and upper wing can cut lift by around 30 percent and increase drag by up to 40 percent. That is a barely visible layer. The wing also stalls at a lower angle of attack and higher speed, and the stall warning is calibrated for a clean wing.

What is the difference between rime and clear ice?

Rime forms from small droplets that freeze on impact, trapping air, giving milky rough ice that builds forward into the airflow, usually below about -15 °C. Clear ice forms from larger droplets in warmer air, roughly 0 °C to -10 °C, flowing aft before it sets into a dense sheet that can grow horns. Clear ice is heavier, holds harder, and is easy to miss at night.

Does freezing temperature in dry air count as known icing?

No. The FAA Chief Counsel's 2009 Bell letter restates the physics: structural icing requires visible moisture and a surface temperature at or below 0 °C. The letter applies a reasonable and prudent pilot standard, so known icing conditions exist when the whole weather picture would lead a reasonable pilot to expect ice along the planned route and altitude.