7.2 Structural Icing, Frost, and Carburetor Ice
Key Takeaways
- Structural ice needs visible moisture and an aircraft surface at or below freezing; rime, clear, and mixed are the three in-flight types (PHAK / Aviation Weather Handbook).
- Freezing rain is the worst clear-ice setup: liquid rain falls from warm air aloft into a subfreezing layer and freezes on contact.
- Even a thin layer of frost reduces lift and increases drag; it must be removed. “It will blow off” is not a procedure (PHAK Chapter 12; AFM/POH; 14 CFR 91.9).
- Carburetor ice can form in warm, moist air without structural icing — PHAK Chapter 7: most likely below 70 °F (21 °C) with relative humidity above 80 percent.
- On a fixed-pitch propeller, carb heat first drops RPM; if ice was present, RPM may then rise as the ice melts. Typical light trainers are not certified for flight into known icing.
ACS PA.I.C.K3i is icing and freezing-level information. PA.I.C.K3k is frost. PAR treats them as one go/no-go family: the wing does not care whether the roughness grew in a cloud or sat on the painted surface overnight.
Structural ice needs two things
In-flight structural icing forms only when visible moisture (cloud, rain, drizzle, wet snow) strikes an aircraft surface that is at or below freezing. Temperature aloft can be below freezing while the skin is still warm from a descent, and the reverse is true after a long soak in cold air. The test item is the skin, not the METAR remark by itself.
The freezing level is the altitude where the outside air temperature is 0 °C. Icing is most often a problem from that level upward through the supercooled-liquid layer. Handbook temperature bands are ranges, not a private-pilot “always ice at −8 °C” factoid. Large droplets in cumuliform clouds and freezing rain favor clear ice at temperatures not far below freezing (commonly taught near 0 °C to −10 °C). Smaller droplets in stratiform clouds favor rime at colder temperatures. Supercooled water has been observed in thunderstorms much colder than that — another reason you do not use a CB as an icing classroom.
| Type | How it forms | How it looks | Typical home | Why it is a problem |
|---|---|---|---|---|
| Rime | Small supercooled droplets freeze on impact | Milky, opaque, rough | Stratiform clouds; colder portion of the icing layer | Adds drag and spoils the airfoil; usually easier to see |
| Clear (glaze) | Large droplets spread before they freeze | Smooth, glossy, hard, often hard to see | Cumuliform clouds; freezing rain | Heavier, bonds hard, can freeze controls and antennas; fastest, worst accumulation |
| Mixed | Both processes at once | Rough conglomerate | Transition layers and many real clouds | Combines rime’s roughness with clear ice’s mass |
Frost on the ground is not a fourth in-flight type. It is deposition: water vapor going directly to ice on a skin that is at or below freezing. PHAK Chapter 12: on a cool, clear, calm night the ground and the airplane radiate heat; if the surface is below freezing, moisture deposits as frost. Dew is the above-freezing cousin and is not the lift problem. Frost is.
Freezing rain is the worst structural-ice story
Freezing rain means liquid rain is falling into a layer that is below 0 °C and freezing on whatever it hits. That requires warm air aloft — a temperature inversion — so drops can exist as rain above you and freeze in the cold air at your altitude or on the ground. Ice pellets at the surface are the cousin clue: rain aloft froze while falling through the cold layer. Ice pellets often mean freezing rain above you. Climbing toward that warm layer is how airplanes collect the worst clear ice in a hurry.
Do not treat “light rain and 28 °F OAT” as a VFR curiosity. That is the PAR freezing-rain setup. A light airplane without ice protection leaves, descends into warmer air if that path is known and clear, or does not launch. There is no published “it is only rime so keep going” allowance for a trainer.
Frost: it will not blow off
PHAK Chapter 12: frost reduces lift and increases drag. The airplane must be free of frost prior to flight. The roughness is the problem, not the weight of a paper-thin layer. A thin, sandpaper coat trips the boundary layer, raises stall speed, and can keep a wing from flying at a rotation speed that was legal yesterday.
Maya’s winter preflight: frost on the upper wing, sun hitting the ramp, a helpful line-person saying “it’ll blow off on the takeoff roll.” That sentence has killed people. The frost that “blows off” the fuselage still sits in the rivet lines and the leading-edge roughness that the wing needs to be clean. Deice fluid, a heated hangar, or physically removing the frost until the surface is bare — then a pretakeoff check that it did not re-form — is the procedure. Taxiing through puddles and hoping takeoff speed polishes the rest is not.
Know the regulation without misapplying it. 14 CFR 91.527 (Subpart F — large and turbine-powered multiengine airplanes and fractional-ownership program aircraft) says no pilot may take off an airplane that has frost, ice, or snow adhering to the propeller, windshield, stabilizing or control surfaces, powerplant installation, specified flight instruments, or the wing, with a narrow FAA-authorized exception for frost under the wing in the fuel-tank area. A typical ASEL trainer is not a Subpart F airplane. Your binding words are the AFM/POH limitation (almost every light-airplane AFM forbids takeoff with frost, ice, or snow adhering to critical surfaces) and 14 CFR 91.9 — operate in accordance with approved limitations. PHAK’s clean-wing sentence is the same idea in handbook language. Polished frost is not a current legal takeoff technique; the old “polish it smooth” allowance was removed.
Carburetor ice is a different machine
Structural ice needs visible moisture and a freezing skin. Carburetor ice does not. Air accelerates through the carburetor venturi, pressure and temperature drop, and fuel evaporates. That cooling can take the metal below freezing even when the OAT is well above 32 °F. Ice then restricts the venturi and the throttle plate and the engine loses power.
PHAK Chapter 7: carburetor ice is most likely when temperatures are below 70 °F (21 °C) and relative humidity is above 80 percent. The classic training envelope you will see on the PHAK icing-probability chart is moist air from about 20 °F to 70 °F. A 60 °F humid day in clear air is inside that envelope. Low power (descent, idle in the pattern, extended taxi) makes it worse because the throttle plate is a ready ice stand and there is less leftover heat.
Fuel-injected engines do not have that carburetor venturi, so they are far less prone to carburetor ice. They can still suffer impact ice on the air filter or inlet in visible moisture at freezing temperatures; alternate air is the published backup. Do not tell the examiner that fuel injection “cannot ice.”
Carb heat: RPM drop, then maybe a rise
On a fixed-pitch propeller the first clue is a slow RPM loss, then roughness. On a constant-speed propeller, look for falling manifold pressure with RPM still held by the governor. Apply full carburetor heat as the AFM directs.
What you must be able to explain:
- Heat comes from a shroud on the exhaust. That air is unfiltered and less dense, so RPM (or MP) drops as soon as you pull the knob — even if there is no ice.
- If ice is present, melting ice goes through the engine. The engine may run rougher for a moment, then RPM (or MP) rises as the passage opens. That rise after the initial drop is the confirmation that ice was there.
- Leave the heat on until the ice is gone and the AFM says you may return to cold air, or leave it on for the descent if that is the procedure. Partial heat can move the metal into the worst icing range — use full heat unless the AFM says otherwise.
Sam’s 60 °F, dew-point-58 °F cross-country is the PAR trap. The sky is clear. Structural ice is not on the table. The descent to the pattern at low power is. When RPM sags, Sam does not lean the mixture to “fix” a mysterious power loss and does not wait for a complete stoppage. Full carb heat, accept the first RPM drop, and wait for the possible recovery.
Known icing and light aircraft
Known icing is icing that is observed or that a reasonable pilot would expect from the available reports and forecasts (PIREPs, AIRMET Zulu, SIGMET, freezing rain). Flight into known icing (FIKI) is a certification: boots or other ice protection, heated windshield and probes, and an AFM that allows flight in icing. A typical 172 or Cherokee is not that airplane. Pitot heat keeps the airspeed alive; it does not deice the wing. If you pick up ice in a cloud you failed to avoid, the published private-pilot move is to leave the visible moisture — turn around, change altitude toward warmer or ice-free air you have reason to believe is there, and land as soon as practicable. Autopilot can hide a changing stall speed; hand-fly periodically if you are unfortunate enough to be in it.
Do not invent a “trace ice is legal for 20 minutes” rule. The AFM either certifies the airplane for icing or it does not.
What two conditions are required for in-flight structural icing, and which precipitation produces the most serious clear ice?
A thin layer of frost coats the upper wing on a cold, clear morning. A ramp observer says it will blow off during the takeoff roll. What is the correct action and authority?
On a 60 °F humid day in clear air, a fixed-pitch trainer loses RPM in a low-power descent. The pilot applies full carburetor heat. What should happen, and why was ice possible?