1. Home
  3. BRIEF
  4. Icing



Nature’s weighty clear coat can ruin more than aerodynamics.

Freezing rain event fully coats several Raytheon T-1A Jayhawks at Vance AFB (Enid OK). Ice is heavy and has the potential to overstress hydraulics and spars. All ice must be removed before attempting a takeoff.
By Karsten Shein
Comm-Inst Climate Scientist

Hurtling past midfield, the copilot called out “V2” and the pilot applied back pressure to the sidestick. The nose lifted, but the aircraft struggled to lift off. After a few more seconds, the aircraft took to the air, but it wasn’t gaining altitude and the trees beyond the airport fence were approaching quickly.

The stick shaker activated. Suspecting ice and knowing that the aircraft probably wouldn’t clear the trees, the pilot yelled “Aborting!” and pulled back on the throttles.

With a heavy thud, the aircraft returned to the ground, and the copilot applied the spoilers and thrust reversers. They managed to slow the jet appreciably, but it still struck the runway end lights at around 40 kts, collapsed the nose gear, and left an 80-ft-long furrow in the wet ground. No one was injured, but the aircraft was substantially damaged.

Investigators determined that the warm skin of the aircraft emerging from a heated hangar had melted the snow that had been falling heavily for the past hour – snow that had refrozen as the aircraft skin cooled during taxi.

The process left a rough and bumpy glaze that neither pilot had thought to look for. It was enough to disrupt airflow and reduce lift to the point that the aircraft would have stalled out and mushed into the trees, had the pilots tried to continue the takeoff.


Ice accretion is one of the more frequent contributors to weather-related aircraft accidents, and a pilot’s ability to anticipate it requires an understanding of its formative factors. Like most weather phenomena, ice is a product of heat and water. With few exceptions, the water must be in liquid form, and the temperature must be at or below freezing.

Except at bitterly cold temperatures below about -40° C (-40° F), most clouds are composed of liquid droplets. Only very cold clouds, such as cirrus or the high tops of towering cumuli, will be made entirely of ice crystals. At temperatures above -40° C but below freezing, clouds will be a mixture of ice crystals and water droplets. However, at OAT below around -20° C (-4° F), the majority of water present will be ice.

Temperatures between -20° C and 2° C (-4° F to 35° F) present the greatest danger for ice accretion, with temperatures of -10° C to 0° C (14–32° F) being the zone in which ice accumulation is likely to be most rapid. In fact, at -10° C, studies have shown that for every ice crystal present in a cloud, there are at least a million liquid cloud droplets, and these droplets are supercooled.

Glaze ice fully coats the wing of an NCAR research aircraft as part of a study to understand icing behavior and its effects on flight. No amount of ice on an aircraft is truly safe, even if the aircraft is certified for flight into known icing conditions.

While ice will always start to melt as temperatures rise above 0° C, water does not always begin to freeze when temperatures drop below freezing. There are several reasons for this. Most water droplets in the atmosphere contain microscopic impurities. The smaller the droplet, the less pure water exists in it, and the lower the temperature at which it will eventually freeze.

In addition, while many small droplets will develop ice crystals in subfreezing temperatures, the weak ice bonds formed within the droplet are often broken quickly due to internal agitation as the droplet is jostled about. However, the situation changes as the supercooled liquid water droplets impact a larger surface that has a subfreezing temperature.

The contact will serve to transform the liquid droplet more or less spontaneously into ice. This is why aircraft icing can occur any time there is liquid precipitation present and the OAT is at or below freezing. There are also other ways in which ice can accrete that do not involve supercooled water.

One is the situation in which snow or sleet falls in subfreezing air on an aircraft skin that has been warmed above freezing, such as by being parked in a heated hangar or subjected to infrared deicing measures. Even use of onboard heated deicing systems can produce ice accretion.

As the frozen precipitation lands on the warmer skin, it melts. If the meltwater remains as the skin cools in the subfreezing air, the water may refreeze as a bumpy glaze. At high speeds, the water will likely run back beyond the leading edges, accreting across the skin and especially any unheated sections.

At slower speeds, such as in taxi, the leading edges will be affected as well. And if air bleed or heated deicing is on, the accretion will likely occur well aft of the leading edges.

Freezing level map for North America. Knowledge of the freezing level along your route of flight will provide critical guidance for avoiding altitudes at which icing is likely, or for finding warmer air to melt accreted ice.

Cold soak icing

The opposite situation may occur as an aircraft descends from high-altitude cruise. Fuel in the wing tanks doesn’t immediately equalize temperature with the surrounding air. As a result of this cold soaking, the wing skin may remain well below freezing as the aircraft flies through rain that is above freezing.

The rain striking the wing is quickly cooled and may freeze to the wing. Normally, cold soak icing is temporary and will melt as the wing equalizes temperature with the air, but it’s still dangerous in that it distorts the airfoil, well aft of deicing systems, often during the critical landing phase of flight.

At higher altitudes (up to 42,000 ft or so), and at temperatures well below -40° C, ice crystal ingestion into powerplants can cause accretion problems.

Although water content is greatly limited at the supercold temperatures where ice crystals are abundant, the crystals can still accumulate within the engines and around engine system probes. Such issues are most commonly encountered in the vicinity of towering cumuli, and, unlike supercooled water, these ice crystals don’t show up well on radar and won’t produce accretion on exterior surfaces.

An important factor in ice formation is the drop in air pressure corresponding to airfoils and air intakes. As pressure decreases, so does temperature. When OAT is up to 2° C above freezing, the pressure decrease above a wing surface may drop the temperature to freezing or slightly below, allowing rain to freeze to the surface. Similarly, carburetors and turbine air inlets lower the pressure of air fed into the combustion sections.

This similar drop in pressure can produce intake icing. Carburetor icing can occur at air temperatures as high as 20° C (68° F). While the venturi effect is not as severe in turbines, especially in direct flow turbines, the indirect airflow ducting in many turboprop engines can accrete ice at bends in the inlet ducts.

The result of engine ice accretion can, in severe cases, block air intakes and produce compressor stalls or even combustion flameouts. More commonly, ingested ice may cause surging or blade vibration. Excessive vibration or ingestion of shedding ice can damage fan blades and other parts of the engine, leading to engine failure.

Weight, drag, and angle of attack

The most common issues of ice accretion relate to the aerodynamic performance characteristics of the aircraft’s lifting surfaces. Airfoil ice accretion adversely affects 3 of the 4 essential flight factors – lift, drag, and weight. Even a small amount of ice on a wing will dramatically reduce its coefficient of lift and the stalling angle of attack (AoA).

The effect is most noticeable at lower airspeeds. While all icing increases drag, rough ice, glaze ice horns, and ice ridges aft of the leading edges all increase drag disproportionately relative to a clean wing surface.

Even a few millimeters of ice can double a wing’s drag factor, and large glaze horns on the leading edges can increase it by 200% or more. Simultaneously, the surface roughness and deformation of airfoil shape by ice accretion will reduce lift substantially. Lift reductions of 30–40% are not uncommon, even with just a thin rime. The modification of lift and drag factors also affects the stalling AoA.

Ice accretion on the wings has been shown to reduce the critical AoA by up to 50% of the clean wing angle. The airflow over an iced wing may also have unanticipated effects on the behavior of trailing edge control surfaces. While such controls work primarily by deflecting air, ice can modify the flow of air across them, altering their effectiveness.

At 919 kg per cubic meter (57 lb/cu ft), ice is relatively heavy. If 1 mm of ice were to accrete over 120 sq m of wing (~1284 sq ft), it would add roughly 110 kg (~242 lbs) of weight to the aircraft. While this weight may not make a huge difference on the wings of a powerful business jet, it can put an already heavily laden aircraft over its maximum takeoff or landing weight, and may also shift the center of gravity.

More dangerously, accretion on the stabilator surfaces far from the center of gravity can destabilize the aircraft and make it difficult – or even impossible – to control or recover from a stall or other unusual attitude.

Partially-shed ice accretion on the leading edges of a Raytheon Beech King Air. Anti- and de-icing systems should be seen only as a means to buy enough time to leave icing conditions.

Observing and forecasting ice

Most larger airports worldwide have automated weather observing systems (AWOS) with precipitation discrimination capabilities – that is sensors that can distinguish between rain and snow. The most advanced of these systems will also have instruments capable of identifying freezing rain and/or runway surface condition.

When a system includes a precipitation discriminator, the metar will include the code A02. Otherwise, it will include an “A01” (although not all A02 stations will have a freezing rain sensor). Typically, freezing fog or precipitation will be included in metars and speci reports.

The code FZ precedes the fog or precipitation code, as in FZDZ (freezing drizzle), FZRA (freezing rain), or FZFG (freezing fog). At a few coastal airports, a pilot may even encounter FZPY (freezing spray). As with any precipitation report, it can also be qualified by a leading “+” (heavy) or “-“ (light) symbol, and may also include VC for “in the vicinity.” If there is a freezing rain sensor but it is inoperative, the remarks (RMK) section will include the code FZRANO.

At a few airports where observers manually augment observation reports, the RMK section may also include the thickness of ice accumulated on a horizontal surface. This includes the thickness of melted and refrozen snow. However, most airports are not equipped for manual reporting of ice accretion.

Regardless of AWOS instruments or manual reporting, pilots should recognize that any report including liquid precipitation or fog in places where air temperature is at or below freezing suggests that icing is likely.

Importantly, even at airports with the capacity to report freezing rain or ice accretion, the lack of a report should never be taken as an indication that those conditions are not present, especially if other conditions are right.

Reporting icing conditions

Aloft, many pilots who encounter icing or those who are not accreting ice in an area where icing has been forecast will issue a pirep. Pireps remain the only way for meteorologists or pilots to tell for sure whether icing is present at a given location and altitude, and they can be the best way to help fellow pilots avoid those areas or find warmer air, should they be accreting ice.

In the absence of pireps, observations from weather balloons are used to estimate the freezing level, and that information is fed into numerical models to produce maps of icing potential.

Given the relatively straightforward set of conditions required to produce accreting ice, present weather and forecast map products of icing potential are available for many regions around the globe. FAA, for example, issues a current icing product (CIP) based on a mixture of observational and model data to visualize the current likelihood and severity of icing (none, trace, light, moderate, and heavy), as well as the possibility of encountering supercooled large droplets (SLDs).

SLDs are larger rain droplets that can produce rapid glaze ice accretion that may be difficult to fully shed. The CIP is updated hourly and uses a 20×20-km grid. Because it is estimating icing, the CIP will never suggest 100% certainty. Only a pirep can do that. But a good rule of thumb is that if the CIP is calling for more than about 50% chance of icing at a given altitude and location, pilots should expect it.

A companion forecast icing product (FIP) provides similar information to the CIP but for the next 12 hours. These products are available from aviationweather.gov. Similar forecast guidance for other regions can be obtained from several websites, such as wxweb.meteostar.com. When conditions aloft are conducive to icing, an icing airmet or sigmet may be issued.

Falcon 900EX is deiced at ZRH (Zürich, Switzerland). Aircraft must be deiced if there is frost, snow, or ice on the aircraft, and should also be deiced if the aircraft is likely to encounter ice accretion during taxi or departure.

Moderate icing potential over a region will prompt an airmet, while heavy icing potential requires a sigmet. Airmets are valid for 6 hours, and icing sigmets are valid for 4. Convective sigmets for thunderstorms should also be considered icing sigmets, as heavy icing is frequently encountered in and around thunderstorms at altitudes above the freezing level. Operating in ice Ice tends to accrete first on protrusions that have a thin cross-section relative to the oncoming airflow.

Antennae, pitot systems, and even the raised frames of a cockpit windscreen are all places where ice, if it is accreting, will likely appear first. In older aircraft, many of these spots can be monitored easily from the cockpit, but as aircraft become more aerodynamically clean, pilots may not have as many visual cues on which to rely.

Instead, they’ll follow environmental conditions to anticipate the presence of ice and monitor their instruments and performance for any anomalies that might suggest accretion.

These anomalies include the aircraft not remaining in trim, and discrepancies between pressure and GPS altitudes and airspeeds, particularly when pressure altitude/airspeed remains constant or lags behind the GPS readings, suggesting the pitot tube has become blocked. While it is never a good idea to knowingly fly into airspace where icing is present or likely, many aircraft are certified for flight into known icing conditions.

This simply means that such aircraft have anti-icing and/or deicing measures designed to mitigate accretion for a period of time – generally enough time to exit the icing region safely and shed any buildup. In all instances, pilots should be familiar with icing advisory information such as FAA’s AC 91-74B, as well as their particular aircraft’s AFM or POH. On the ground, approved deicing methods should be employed before any takeoff is attempted.

Most deicing fluids will maintain surfaces ice free for an approximate period of time or until several minutes after takeoff. Departure delays may require a second round of deicing. Pilots should also be cautious in that the freezing rain that coats their aircraft is also likely coating taxiways and runways in ice.

In anticipation of entering a region of icing potential aloft, anti-icing measures should be deployed before any ice is detected. A good rule of thumb though is to activate anti-icing systems at the first signs of visible moisture and OAT less than around 2°C (35°F). If accretion is noticed, deicing systems should be activated.

Many pilots choose to delay popping deicing boots until enough ice has accreted, mistakenly thinking that a thin veneer of ice is more difficult to remove, and repeated use may create an air pocket beneath a thicker layer, making the ice impossible to remove.

Fly away from icing conditions

Regardless of the type of deicing system on your aircraft, it is best to follow the AFM or POH for proper use, and try to remove the aircraft from the icing situation before it gets so bad that the deicing systems become ineffective.

It is also worthwhile to keep anti- or de-icing systems active until you can be sure that your aircraft and its systems are and can remain ice free. Departing an area of icing is often as simple as changing altitude. Before departure, be sure to have a good handle on the freezing level along your route of flight. This will provide the most immediate help in locating warmer air.

Controllers, flight watch, or other pilots in the area may also be able to tell you where warmer air can be found. Near the surface in the vicinity of fronts, the frontal boundary presents a temperature inversion, with warmer air riding above colder surface air. Most surface fronts extend just a few thousand feet above the ground, so climbing to warmth may only require a small altitude adjustment.

Absent a front, warmer air will almost always be found at a lower altitude, and, if you have sufficient altitude, a descent will bring you toward ice-melting temperatures. Naturally, finding clear air quickly, using airborne or surface radar data – if available – will halt any further accretion, although subfreezing air temperatures will help keep already accreted ice in place.

As usual, if you have picked up ice, or are not getting ice in an area where it was forecast, be sure to file a pirep that will help meteorologists and your fellow pilots better pinpoint danger areas.

Karsten Shein is co­founder and science director at ExplorEiS. He was formerly an assistant professor at Shippensburg Univer­sity and a climatolo­gist with NOAA. Shein holds a commercial license with instrument rating.