Icing and Turbulence
CSIR NET Part C aviation meteorology: supercooled droplets, rime and clear ice, accretion rate, kinetic heating, CAT, Richardson number and mountain waves.
- 23 min
On this page · 11 sections
Prerequisites
- Weather Hazards at Take-off, Cruise and Landing — the lift equation and where each hazard sits in a flight.
- Supercooling, latent heat and moist thermodynamics, from Thermodynamics of Dry and Moist Air.
- Static stability and the Brunt–Väisälä frequency \( N \), from Vertical Stability and Tropical Convection.
- Cloud microphysics (Bergeron process, ice nuclei), from Clouds, Precipitation and General Circulation.
The Hook: Water That Should Be Ice
Part 1: Airframe Icing
The essential condition: supercooled liquid water
Airframe icing needs liquid water at or below 0 °C that freezes on contact. Two facts set the temperature window:
- Below 0 °C, droplets stay liquid until an ice nucleus acts on them. Effective ice nuclei are scarce at temperatures just below freezing and more numerous as it gets colder, so a cloud at −10 °C is often still mostly liquid.
- Near −40 °C pure droplets freeze spontaneously (homogeneous nucleation). Colder clouds are entirely ice. Ice crystals bounce off cold airframes rather than sticking (engine ice-crystal icing is a separate problem, mentioned on the previous page).
So icing is most common between 0 °C and about −20 °C, possible down to about −40 °C, and rare in clouds made entirely of ice. Within the window, the Bergeron (Wegener–Bergeron–Findeisen) process slowly converts liquid to ice, because the saturation vapour pressure over ice is lower than over water. Young clouds with strong updrafts (cumuliform) resupply liquid water faster than it is lost, so they hold more and larger supercooled drops.
How fast ice grows: the accretion equation
The mass of water hitting an object per unit time is
$$\boxed{\frac{dM}{dt} = E \cdot \mathrm{LWC} \cdot V \cdot A}$$where LWC is the cloud liquid water content (kg m\(^{-3}\); typically of order 0.1–1 g m\(^{-3}\) in icing clouds), \( V \) the airspeed, \( A \) the frontal area and \( E \) the collection efficiency: the fraction of droplets in the swept volume that actually hit rather than following the airflow around the object.
\( E \) depends on droplet inertia relative to the flow deflection:
- Larger droplets have more inertia and hit more: \( E \) rises with droplet size.
- Higher speed raises \( E \).
- Smaller (thinner) objects deflect the flow over a shorter distance, so \( E \) is higher. Thin leading edges, tailplanes, antennas, probes and wipers ice first and fastest; thick wing roots collect less.
If all the collected water freezes, the ice thickens at a rate \( E \cdot \mathrm{LWC} \cdot V/\rho_i \), where \( \rho_i \) is the density of the deposit.
Kinetic heating
Air brought to rest on a leading edge warms adiabatically. The maximum (stagnation) temperature rise is
$$\Delta T = \frac{V^2}{2c_p}$$with \( c_p \approx 1005\ \mathrm{J\,kg^{-1}\,K^{-1}} \). In practice a surface reaches only a fraction of this (the recovery factor, somewhat less than 1), and evaporation of droplets cools it again. Still:
| Airspeed (m s\(^{-1}\)) | \( V^2/2c_p \) (K) |
|---|---|
| 50 | 1.2 |
| 100 | 5.0 |
| 150 | 11.2 |
| 250 | 31.1 |
Slow aircraft (light aircraft, helicopters, turboprops) get almost no kinetic heating and ice readily near 0 °C. Fast jets at high speed are warmed by several to tens of kelvin, which can lift a marginal icing temperature above freezing. Combined with the fact that cruise levels are usually cold enough for clouds to be glaciated, this is why jets meet most icing in climb, descent and holding, at lower speeds through middle-level cloud.
Types of ice
What the ice looks like depends on how fast each droplet freezes. Freezing releases latent heat (334 kJ kg\(^{-1}\)), which must be removed by convection to the cold air and by conduction.
| Type | Formation | Typical conditions | Appearance | Hazard |
|---|---|---|---|---|
| Rime | Small droplets freeze almost instantly on impact, trapping air | Colder part of the range (roughly −10 to −20 °C and below), small drops, low LWC; stratiform cloud | White, opaque, rough, brittle; builds forward into the airflow | Less dense, easier to remove; still roughens and reshapes leading edges |
| Clear (glaze) | Large droplets or high LWC; only part of each drop freezes at once; the rest flows back and freezes later | Near 0 to about −10 °C; large drops, high LWC; cumuliform cloud, freezing rain | Transparent, hard, smooth, dense; can form horns and runback ridges | Heavy, strongly adherent, hard to see and remove; runback can form behind protected areas |
| Mixed | Mixture of the above | Intermediate temperatures; cumulus with varied drop sizes | Opaque-to-clear layers, rough | Combines weight and roughness |
| Hoar frost | Deposition of vapour directly as ice on a surface colder than 0 °C | Clear nights (aircraft parked outside); descending from very cold altitude into moist air | Feathery, white crystals | Thin, but roughness lowers maximum lift on take-off |
The ranges are typical tendencies, not sharp boundaries: droplet size and LWC matter as much as temperature.
Freezing rain and supercooled large drops
Freezing rain forms when snow falls into a warm layer above 0 °C (often over a warm front), melts completely to rain, then falls into a sub-freezing layer near the ground and supercools without refreezing. If the cold layer is deep, the drops refreeze into ice pellets.
graph TD
A["Snow aloft
T below 0 °C"] --> B["Warm nose
T above 0 °C:
snow melts to rain"]
B --> C{"Depth of cold
layer below"}
C -- "Shallow" --> D["Supercooled raindrops:
freezing rain, glaze ice
on surfaces and aircraft"]
C -- "Deep" --> E["Drops refreeze:
ice pellets"]Drops of drizzle and rain size (supercooled large drops, SLD) are a special danger. Their high collection efficiency and slow freezing make them flow back beyond the de-iced leading edge and freeze there, where ice protection does not reach. Freezing drizzle can also form without a warm layer, by collision–coalescence in shallow cloud. An aircraft in freezing rain has a classic escape route: climb into the warm layer.
Effects of icing on aircraft
| Component | Effect |
|---|---|
| Wings | Roughness and changed shape: lower maximum lift, higher drag, higher stall speed, stall at a lower angle of attack |
| Tailplane | Thin, so ices early; tailplane stall can cause a nose-down pitch, often when flaps are lowered |
| Weight | Added mass (secondary to aerodynamic effects) |
| Propellers and engine intakes | Loss of thrust; ice shedding into engines |
| Pitot and static probes | Unreliable airspeed and altitude (probes are heated for this reason) |
| Windscreen and antennas | Loss of vision; loss of radio and navigation |
| Carburettor (piston engines) | Induction icing: venturi pressure drop and fuel evaporation cool the intake air, so ice can form in humid air even when the outside air is well above freezing |
Icing intensity is reported as trace, light, moderate or severe, by accretion rate and its effect on the aircraft.
Protection and forecasting
- Ground: de-icing fluid removes frost, snow and ice; thicker anti-icing fluid then prevents new contamination for a limited holdover time that depends on the precipitation type and temperature.
- In flight: anti-icing (prevents formation: heated leading edges, engine bleed air, heated probes and windscreens) and de-icing (removes formed ice: pneumatic boots on many turboprops).
- Forecasting: icing is forecast where model or sounding temperatures lie between 0 and about −20 °C and cloud or high relative humidity is present, with extra weight for convective cloud, warm noses (freezing rain) and SLD diagnostics. Pilot reports, satellite cloud-phase products and radar refine it.
Part 2: Turbulence
What aircraft feel
Turbulence is irregular motion of the air. Aircraft respond most to eddies of roughly tens of metres to about a kilometre, comparable to their size and the distance covered in a second or so. Reports use four intensities based on the aircraft’s response:
| Intensity | Aircraft and occupants |
|---|---|
| Light | Slight, erratic changes in altitude or attitude; occupants may feel strain against seat belts |
| Moderate | Changes in altitude or attitude, but the aircraft stays in control; occupants feel definite strain; unsecured objects move |
| Severe | Large, abrupt changes; aircraft may be momentarily out of control; occupants forced violently against belts |
| Extreme | Aircraft violently tossed, practically impossible to control; possible structural damage |
Because the same air feels different to a light aircraft and a heavy jet, ICAO uses the eddy dissipation rate (EDR, the cube root of the turbulent kinetic energy dissipation rate \( \varepsilon \), units m\(^{2/3}\) s\(^{-1}\)) as an aircraft-independent measure, computed automatically from aircraft data.
Classification by cause
| Type | Mechanism | Where and when |
|---|---|---|
| Convective (thermal) | Buoyant thermals in an unstable boundary layer; updrafts and downdrafts in cumulus and CB | Sunny afternoons over land, below and in cumulus; extreme in CB |
| Mechanical | Strong wind over rough terrain, buildings, trees | Near the ground in strong winds, especially in unstable air; lee of hangars and terminals |
| Mountain wave (orographic) | Stable flow forced over a ridge sets up gravity waves; rotors beneath them | Lee of ridges, can reach the stratosphere |
| Clear-air turbulence (CAT) | Shear instability (Kelvin–Helmholtz) in stable, strongly sheared layers | Near jet streams and the tropopause, upper troughs and ridges |
| Frontal | Shear and lifting across a frontal surface | Narrow zones along active fronts |
| Wake turbulence | Wingtip vortices shed by aircraft (not meteorological, but weather controls how long they persist and where they drift) | Behind and below heavy aircraft near runways |
Clear-air turbulence and the Richardson number
In a stably stratified layer, buoyancy resists vertical displacement while vertical wind shear feeds kinetic energy into overturning. Their ratio is the gradient Richardson number:
$$\boxed{Ri = \frac{N^2}{(\partial U/\partial z)^2} = \frac{\dfrac{g}{\theta}\dfrac{\partial\theta}{\partial z}}{\left(\dfrac{\partial U}{\partial z}\right)^2}}$$- The Miles–Howard theorem shows that a stratified shear flow is stable to small disturbances everywhere if Ri > 1/4. Ri < 1/4 somewhere is the necessary condition for Kelvin–Helmholtz instability: billows grow, roll up and break into turbulence.
- Turbulence, once started, mixes the layer: it weakens the shear and the stratification until Ri rises and the patch decays. Patches are typically thin (hundreds of metres deep) but extend tens of kilometres horizontally, which is why a change of a few thousand feet often gets an aircraft out of CAT.
- Ri is only diagnosable at the resolution of the data; real layers that are thinner than the model’s levels can have much smaller Ri than the model shows.
Where CAT concentrates:
- Near jet streams, especially on the poleward (cold-air) side of the core and just above and below it, where vertical shear is greatest.
- Near the tropopause, where stratification changes sharply and shear can be strong.
- In sharply curved upper troughs and ridges, and where two jets converge or diverge (strong deformation).
- Above and downstream of mountain ranges, where breaking mountain waves add to the shear.
Forecast indices combine shear and deformation. The Ellrod index multiplies vertical wind shear by horizontal deformation (plus convergence in later versions) on the reasoning that deformation sharpens shear zones. CAT is invisible to radar and cannot be seen by pilots, so forecasts, pilot reports and automated EDR reports are the main tools.
Mountain waves and rotors
When stable air flows across a ridge, it is displaced upward and buoyancy pushes it back down, setting up internal gravity waves that are stationary relative to the ground. Favourable conditions:
- Wind roughly perpendicular to the ridge, with speed at ridge height above some threshold and increasing with height.
- A stable layer near the ridge top, with less stable air above.
A useful parameter for the flow regime is the Froude number based on ridge height \( h \):
$$Fr = \frac{U}{N h}$$- \( Fr \ll 1 \): low-level air lacks the kinetic energy to rise over the ridge; it is blocked and flows around or stagnates upwind.
- \( Fr \sim 1 \): strongest wave response; can produce downslope windstorms in the lee.
- \( Fr \gg 1 \): air passes easily; waves are weak relative to the flow.
For hydrostatic waves over a broad ridge, the waves propagate vertically with a vertical wavelength
$$\lambda_z = \frac{2\pi U}{N}$$For \( U = 20 \) m s\(^{-1}\) and \( N = 0.01 \) s\(^{-1}\), \( \lambda_z \approx 12.6 \) km. When the Scorer parameter \( l^2 = N^2/U^2 - U_{zz}/U \) decreases sharply with height (stability falling, wind increasing), waves are reflected and trapped in the lower layer, forming a train of lee waves many wavelengths downstream.
| Feature | Description | Hazard |
|---|---|---|
| Cap (foehn wall) cloud | Sits on the ridge crest | Obscures terrain |
| Lenticular clouds (altocumulus lenticularis) | Smooth lens-shaped clouds at wave crests, stationary while air flows through them | Mark waves; smooth air inside is deceptive |
| Rotor (roll) cloud | Ragged cloud below wave crests near ridge height in the lee | Severe to extreme turbulence in the rotor |
| Downslope wind | Strong, gusty, warm wind down the lee slope | Low-level turbulence and shear near lee aerodromes |
| Wave updrafts/downdrafts | Smooth but strong vertical motion | Altitude loss in downdrafts, sometimes exceeding the climb rate of light aircraft |
Breaking waves aloft (when waves steepen and overturn, often near a critical level where the wind drops to zero or the stratification changes) generate high-level turbulence that blends with CAT. Over India, flow across the Himalaya and the Western Ghats can generate waves; the strength depends on season and wind.
Convective and low-level turbulence
- Thermals: on hot afternoons the dry convective boundary layer is full of rising plumes and sinking air, giving a bumpy ride up to cloud base, with smooth air often found above it.
- Inside cumulonimbus: updrafts and downdrafts of tens of metres per second side by side make turbulence severe to extreme; this is the main reason aircraft avoid CB rather than fly through it.
- Around storms: turbulence extends outside visible cloud, above the anvil and in the outflow.
- Mechanical low-level turbulence increases with wind speed, surface roughness and instability: strong gusty winds over a city-edge aerodrome on an afternoon are the worst combination.
Wake turbulence
Every wing generating lift sheds a pair of counter-rotating wingtip vortices. They are strongest behind aircraft that are heavy, clean (flaps up) and slow, which is exactly the take-off and landing configuration of a large aircraft. The vortices sink below the flight path and, near the ground, spread outward. Meteorology controls their fate:
- Light winds leave them near the runway; a light crosswind can hold the upwind vortex stationary over the runway.
- Strong winds and turbulence break them up faster.
- Stable, calm conditions let them persist longest.
Air traffic control applies separation minima based on aircraft weight categories for this reason.
Memory Tricks
Don’t Confuse
| This | vs | That |
|---|---|---|
| Rime: small drops, cold, instant freezing, opaque, brittle | vs | Clear (glaze): large drops, near 0 °C, slow freezing, transparent, hard |
| Freezing rain: liquid supercooled drops from a warm layer above | vs | Ice pellets: drops that refroze in a deep cold layer |
| Hoar frost: deposition from vapour | vs | Rime: freezing of supercooled droplets |
| Airframe icing needs supercooled liquid | vs | Engine ice-crystal icing happens in glaciated air near storm tops |
| Carburettor icing: possible above 0 °C OAT in humid air | vs | Airframe icing: needs sub-freezing water |
| CAT: shear instability, Ri < 1/4, no cloud | vs | Convective turbulence: buoyancy, in and near cumulus |
| Lee wave (smooth, laminar lenticulars) | vs | Rotor (violent turbulence below the wave crest) |
| \( N^2 \) (stability, s\(^{-2}\)) | vs | Ri (dimensionless ratio of stability to shear squared) |
| Froude number \( U/Nh \) (flow over a ridge) | vs | Richardson number (shear instability in a layer) |
Common Mistakes
Practice
All questions below are original practice questions in CSIR NET style. They are not previous-year questions.
Level 1: Foundation
- Near what temperature do pure cloud droplets freeze spontaneously?
- Write the gradient Richardson number and its critical value.
- Which type of ice forms from small droplets freezing instantly on impact?
Answers
- About −40 °C (homogeneous nucleation).
- \( Ri = N^2/(\partial U/\partial z)^2 \); critical value 1/4.
- Rime ice.
Level 2: Part B Style
Q1 (practice). Which conditions favour clear (glaze) ice rather than rime?
- (a) Small droplets, −25 °C, low liquid water content
- (b) Large droplets, −3 °C, high liquid water content
- (c) Ice crystals at −30 °C
- (d) Water vapour deposition on a cold parked aircraft at night
Answer and explanation
(b). Large drops and high LWC near 0 °C deliver latent heat faster than it can be removed, so each drop freezes only partly and flows before freezing: clear ice. (a) gives rime; (c) gives no airframe icing; (d) is hoar frost.
Q2 (practice). The stagnation temperature rise for air brought to rest from 200 m s\(^{-1}\) (take \( c_p = 1005\ \mathrm{J\,kg^{-1}\,K^{-1}} \)) is about:
- (a) 2 K
- (b) 10 K
- (c) 20 K
- (d) 40 K
Answer and explanation
(c). \( \Delta T = V^2/2c_p = 40000/2010 \approx 19.9 \) K. Actual surface warming is somewhat less (recovery factor below 1).
Q3 (practice). Clear-air turbulence near a jet stream is most likely:
- (a) in the jet core, where the wind speed is maximum and shear is zero
- (b) on the poleward side of the jet and just above or below the core, where vertical shear is strong
- (c) only inside cirrus clouds near the jet
- (d) in the warm boundary layer under the jet
Answer and explanation
(b). CAT needs strong vertical shear in stable air (low Ri). At the core itself \( \partial U/\partial z \approx 0 \). Cloud is not required, and the boundary layer is unrelated to jet-level CAT.
Q4 (practice). Stable air at 15 m s\(^{-1}\) with \( N = 0.015\ \mathrm{s^{-1}} \) approaches a ridge 2 km high. The Froude number and the likely regime are:
- (a) 0.5; low-level flow largely blocked
- (b) 2.0; flow passes over easily
- (c) 0.5; flow passes over easily
- (d) 5.0; strong blocking
Answer and explanation
(a). \( Fr = U/(Nh) = 15/(0.015 \times 2000) = 0.5 \). Fr < 1 means low-level air lacks the energy to climb the ridge and is blocked or diverted.
Level 3: Part C Style
Q1 (practice). A turboprop flies at 100 m s\(^{-1}\) in a stratocumulus layer at −8 °C with LWC = 0.3 g m\(^{-3}\). On the tailplane leading edge, \( E = 0.7 \); on the thick wing root, \( E = 0.2 \). Assume all water freezes, deposit density 800 kg m\(^{-3}\), \( c_p = 1005\ \mathrm{J\,kg^{-1}\,K^{-1}} \).
(i) Find the ice thickening rate on each, in mm per minute. (ii) Find the maximum kinetic temperature rise. Does it remove the icing risk? (iii) Explain the difference between the two surfaces.
Answer and worked solution
- Rate = \( E \cdot \mathrm{LWC}\cdot V/\rho_i \).
- Tailplane: \( 0.7 \times 0.3\times10^{-3}\times100/800 = 2.63\times10^{-5}\ \mathrm{m\,s^{-1}} \) = 1.6 mm min\(^{-1}\).
- Wing root: \( 0.2 \times 0.3\times10^{-3}\times100/800 = 7.5\times10^{-6}\ \mathrm{m\,s^{-1}} \) = 0.45 mm min\(^{-1}\).
- \( \Delta T = 100^2/2010 \approx 5.0 \) K at most; the surface stays near −3 °C or colder, so the risk remains.
- The thin tailplane leading edge deflects air over a short distance, so droplets cannot follow the streamlines and hit (high \( E \)). The thick root deflects air gradually and many droplets are carried around it. This is why the tail can ice first and why tailplane stall is a known icing hazard.
Q2 (practice). A sounding near the subtropical jet gives, between 10.0 and 10.6 km, \( \theta \) rising from 330 K to 333 K and the wind rising from 45 to 75 m s\(^{-1}\). Between 11.0 and 11.6 km, \( \theta \) rises from 340 K to 352 K and the wind falls from 70 to 60 m s\(^{-1}\).
(i) Compute Ri for each layer (use mean \( \theta \) and g = 9.81 m s\(^{-2}\)). (ii) Which layer is likely to contain CAT? (iii) Which layer has stratosphere-like stability, and what does its wind profile suggest about where the jet core lies?
Answer and worked solution
- Layer 1: \( N^2 = (9.81/331.5)(3/600) = 1.48\times10^{-4}\ \mathrm{s^{-2}} \); shear = 30/600 = 0.05 s\(^{-1}\), squared \( 2.5\times10^{-3} \); Ri ≈ 0.059. Layer 2: \( N^2 = (9.81/346)(12/600) = 5.67\times10^{-4}\ \mathrm{s^{-2}} \); shear = −10/600 = −0.0167 s\(^{-1}\), squared \( 2.78\times10^{-4} \); Ri ≈ 2.0.
- Layer 1 (Ri well below 1/4): Kelvin–Helmholtz instability and CAT likely, below the jet core where speed increases rapidly with height.
- Layer 2: \( \theta \) increases much faster with height (20 K km\(^{-1}\) against 5 K km\(^{-1}\)), the high static stability typical of the stratosphere. The wind rises through layer 1 and falls through layer 2, so the jet core lies between them, near 10.6–11 km; jet cores usually sit just below the tropopause.
Q3 (practice). Which statements about icing and mountain waves are correct?
(A) Freezing rain needs a layer warmer than 0 °C above a sub-freezing surface layer. (B) Supercooled large drops can freeze behind the protected leading edge. (C) For U = 25 m s\(^{-1}\) and N = 0.01 s\(^{-1}\), the vertical wavelength of hydrostatic mountain waves is about 15.7 km. (D) Lenticular clouds are the most turbulent part of a mountain-wave system.
- (a) A, B and C only
- (b) A and D only
- (c) B, C and D only
- (d) A, B, C and D
Answer and worked solution
(a) A, B and C only.
- (A) The classic freezing-rain profile is melt aloft, supercool below. (Freezing drizzle can form without a warm layer, but freezing rain from melted snow needs it.) Correct.
- (B) SLD have high collection efficiency and freeze slowly, flowing back beyond protected areas. Correct.
- (C) \( \lambda_z = 2\pi U/N = 2\pi\times25/0.01 \approx 15{,}700 \) m. Correct.
- (D) Lenticulars mark smooth wave crests; the worst turbulence is in rotors below and in breaking waves aloft. Incorrect.
Quick Revision Table
| Item | Fact |
|---|---|
| Icing requirement | Supercooled liquid water; 0 to about −20 °C most common; none below about −40 °C |
| Accretion | \( dM/dt = E\cdot\mathrm{LWC}\cdot V\cdot A \); thickening \( E\,\mathrm{LWC}\,V/\rho_i \) |
| Collection efficiency | Up with drop size, speed; down with object size |
| Kinetic heating | \( \Delta T = V^2/2c_p \): 5 K at 100 m s\(^{-1}\), 20 K at 200 m s\(^{-1}\) |
| Rime | Small drops, colder, instant freezing, white, brittle |
| Clear | Large drops, near 0 °C, high LWC, transparent, hard, runback |
| Freezing rain | Warm nose melts snow; shallow cold layer below supercools it |
| SLD | Freezes beyond protected areas |
| Icing effects | Lift down, drag up, stall speed up; tailplane stall; probes; carburettor icing above 0 °C |
| Turbulence intensities | Light, moderate, severe, extreme; EDR = \( \varepsilon^{1/3} \) |
| Turbulence types | Convective, mechanical, mountain wave, CAT, frontal, wake |
| Richardson number | \( Ri = N^2/(\partial U/\partial z)^2 \); KH instability needs Ri < 1/4 |
| CAT | Poleward side of jet, above and below core, tropopause, sharp troughs; invisible to radar |
| Froude number | \( Fr = U/Nh \); < 1 blocked; ~1 strongest waves |
| Mountain waves | \( \lambda_z = 2\pi U/N \); trapped lee waves when Scorer parameter falls with height; rotors severe |
| Wake turbulence | Heavy, clean, slow; vortices sink and drift; light crosswind worst |
Key Takeaways
What’s Next
- Visibility and Wind Shear — fog, visibility physics and the microburst.
- Nowcasting — radar and satellite detection of icing-prone and turbulent storms.
- Thermodynamics of Dry and Moist Air — revisit latent heat and \( \theta \).
- Quick Revision — the whole unit on one page.