Icing Operations III — Managing Icing in Flight

Once the wheels leave, anti-ice stops being a procedure and becomes a rhythm: when you must switch on, when you may switch off, what each selection costs — and how to recognise the one kind of ice the aircraft carries no alert for. This article strings the whole flight onto a single decision line and then settles the bills: speed floors, fuel reserves and the performance assumptions that quietly contain the word "anti-ice". The systems behind it are articles 02 and 03; the sensors, article 06; the failures, articles 11 and 12.

1. The standing order — and its one narrow window

The sentence that repeats through the SOPs like a drumbeat:

In icing condition (Refer to LIM-ICE_RAIN Definition of Icing Conditions), the flight crew must turn on the engine anti-ice and should not wait until seeing ice building up.

"Don't wait to see ice" is the essence of anti-ice: the lip heater prevents accretion from starting; switch on after ice is visible and the first thing EAI does is shed that ice into the fan — the very event it exists to prevent (article 03's fail-open fear, in operational form). On the ground the phrasing is absolute — Engine anti-ice must be set to ON during all ground operations, when icing conditions exist or are anticipated. Airborne, one narrow exemption opens:

Engine anti-ice must be set to ON when icing conditions exist or are anticipated, except during climb and cruise when the SAT is below -40 °C (-40 °F).

Below −40 °C static air temperature, supercooled droplets scarcely exist — the water is already crystals, and classic icing conditions cannot form. But mind two edges of this window. First, it belongs to climb and cruise only; descent revokes it (article 03 quoted the rule in full: engine anti-ice on before and during descent even if SAT is below −40 °C, because the pushbutton also commands the FADEC's higher idle — flame-out protection through the cloud layers). Second, the entry to cloud is a moving target. Per the FCTM:

Whenever icing conditions are encountered or expected, the engine anti-ice should be turned on. Although the TAT before entering clouds may not require engine anti-ice, flight crews should be aware that the TAT often decreases significantly, when entering clouds. If the recommended anti-ice procedures are not performed, engine stall, over-temperature, or engine damage may occur.

[!warning]- Switch on for the cloud, not for the gauge TAT 12 °C beside the cloud does not promise 12 °C inside it. "Anticipated" means the button goes on before cloud entry — the gauge follows you in, and by the time it crosses the line, unprotected metal has already had its first minutes of icing. The FCTM's price list for skipping it: stall, over-temperature, or damage.

2. Wing anti-ice: may to prevent, must to remove

Per the after-takeoff SOP:

When icing conditions are encountered: ‐ The flight crew may turn on the wing anti-ice to prevent ice accretion on the wing leading edge ‐ The flight crew must turn on the wing anti-ice if there is evidence of ice accretion, such as ice on the visual indicator, or on the wipers, or with the SEVERE ICE DETECTED alert. This is to remove any ice accumulation from the wing leading edge.

A beautifully structured rule: may = prevention, must = removal, with three admissible tokens of evidence — two optical (the titanium bolt and the wipers, article 06) and one electronic (the SEVERITY-driven alert). Engine anti-ice gets no may tier — conditions alone compel it — because WAI's costs (bleed air, thrust ceiling, the 200 °C schedule lock of article 02) earn the crew some discretion on prevention. The FCTM restates the must threshold from the other side: wing anti-ice should be on if either severe ice accretion is expected, or if there is any indication of icing on the airframe.

3. What switching on costs: the descent profile

Per the descent-preparation SOP:

ANTI ICE ON reduces the descent path angle (when the engines are at idle). The flight crew can compensate for this by increasing descent speed, or by extending up to half speedbrakes.

The chain: anti-ice on → FADEC raises idle (article 03) → "idle" now means more thrust → shallower glide → you drift above the computed profile. The FMS was built expecting this — per the FCTM's descent chapter:

The idle segment assumes a given managed speed flown with idle thrust plus a small amount of thrust. This gives some flexibility to keep the aircraft on the descent path if engine anti-ice is used or if winds vary.

That small amount of thrust margin is why the FMA annunciates THR DES rather than THR IDLE. Spend the margin (engine and wing anti-ice together, say) and the compensation is manual: watch the vertical deviation and use speed or up to half speedbrake. The prediction machinery also eats this button — the FMS descent computation takes the engine anti-ice status among its inputs, so an honest PERF DES entry buys an honest top-of-descent.

4. The speed floor: ice is invisible payload

The adverse-weather chapter's minimum-speed section is one table plus two accounting sentences. The evidence sentence opened article 06 (indicator or wipers); the caution — Extended flight in icing conditions with the slats extended should be avoided — rules the holding pattern below. The floors:

And the two sentences that explain the asymmetry:

The minimum speed takes into account ice accretion on non-heated structure.

The minimum speed takes into account ice accretion on the entire airframe when anti-ice is inoperative.

[!warning]- Clean configuration carries the bigger increment — with anti-ice working +15 clean but only +5 configured looks backwards until article 02's fine print pays off: WAI heats the four outboard slats — and clean means the slats are retracted, so the protected area is effectively zero and the whole burden falls on non-heated structure. Configured, the heated slats lead the wing and +5 suffices. The inoperative column's VLS + 10 kt / green dot is worth circling twice: it is the same speed package every severe-ice abnormal procedure imposes in article 12 — one table serving both normal and failure worlds.

Holding turns the slat caution into technique. Per the FCTM:

If holding is performed in icing conditions, the flight crew should maintain clean configuration. This is because prolonged flight in icing conditions with the slats extended should be avoided.

Extended slats grow ice in the slat/wing gap and on their unheated undersides; retraction then grinds it into the mechanism — the same fear that governs flap timing on the ground (article 09) and after landing (§5). Holding burns time by design; don't burn it in a configuration that collects ice.

5. Approach and landing: WAI bows out, the flaps hold their peace

On final, with the gear down, the aircraft-configuration flow contains a quiet reversal:

Switch the WING ANTI ICE pb-sw ON, only in severe icing conditions.

Default off for landing: at approach power the bleed is at its least generous (pressure and the 150 °C schedule both sag near idle — article 02's AIR BLEED LO TEMP is the alert form of this), and go-around performance is worth more than marginal heating. Engine anti-ice follows the ordinary conditions logic throughout. After landing, the flaps become the protected species:

If the approach was performed in icing conditions, or if the runway was contaminated with slush or snow, do not retract the flaps until after engine shutdown, and after the ground crew confirmed that flaps and slats are clear of obstruction due to ice.

The FCTM supplies the mechanism — retraction could cause damage, by crushing any ice that is in the slots of the slats — and the sequel: at the gate, engines stopped, visual inspection clean, retract with the electric pumps. One last taxi-in note pairs with it:

If engine anti-ice is used, carefully control taxi speed, particularly on wet or slippery surfaces because ground idle is increased.

The same raised idle that protects against flame-out in descent becomes surplus horsepower pushing you along an icy taxiway — one design, two faces.

6. Ice crystals: the enemy without an alert

Article 06 established the blind spot — crystals bounce off cold skin, so neither the detectors nor the titanium bolt testify. The FCTM's weather chapter completes the human-sensor course. The kill mechanism:

When ice crystals get in contact with a hot surface, they melt.

This is why flight in areas of ice crystals may result in various effects, for example engine vibrations, engine power loss, engine damage, or icing of air data probes.

Cold surfaces shrug crystals off; warm surfaces — deep in the engine, around heated probes — melt them into films that refreeze where airflow pools them. The recognition list (four verbatim, two paraphrased):

‐ Small accumulation of ice particles on wipers ‐ Smell of ozone or Saint Elmo's fire ‐ Aircraft TAT indication that remains near 0 °C (due to freezing of the TAT probe) ‐ Light to moderate turbulence in IMC at high altitude

Plus: an appearance of "rain" on the windshield at temperatures far too cold for liquid rain — typically with a characteristic hissing sound — and the radar picture that doesn't add up: little or no echo at your level while heavy precipitation paints below you, or your track sits downwind of a big convective cell. The hardest currency on that list is the TAT pinned near zero — the probe freezing is itself the signature, and it corrupts the very gauge your icing-conditions judgement leans on (article 01). Avoidance doctrine, verbatim:

• Prefer lateral to vertical avoidance • Comply with the avoidance margins • Deviate upwind instead of downwind.

Crystal plumes stream tens of miles downwind under the anvil — go around the weather on the side the wind comes from. And if you're caught anyway: there is no "ice crystal procedure". Its procedures are the failures it causes — engine stall, vibration, relight, unreliable airspeed — in articles 11 and 12.

7. Cold-air arithmetic: the altimeter reads high

Cold air is dense air; the pressure levels squeeze downward, and the barometric altimeter — calibrated to ISA — doesn't know. Per the FCTM:

When the temperature is lower than ISA: ‐ The true altitude of the aircraft is lower than the altitude that the ADIRS computes ‐ The FPA that the aircraft actually flies, is less steep than the FPA that the ADIRS computes.

Indicated 3 000 ft on a −20 °C day may be a true 2 600 — four hundred feet donated to the terrain. The FCTM's correction table spans −10/−20/−30 °C aerodrome temperatures against height above the field (50 ft at 500 ft/−10 °C, rising to 950 ft at 5 000 ft/−30 °C), and compresses into a mental-math line:

For aerodromes at sea level, these corrections corresponds approximately to 4 x Delta ISA x Height (ft)/1000.

Worked once so it sticks: −20 °C airfield (ΔISA ≈ 35), FAF at 3 000 ft → 4 × 35 × 3 ≈ 420 ft to add. Apply it to the minimum altitudes that matter — FAF, step-downs, minima, the altitude-distance checks — and correct the selected FPA in vertical selected mode. The descent-preparation SOP carries the prompt (When the aircraft operates in low OAT conditions, consider altitude corrections for low temperature) alongside a forward-looking one: destination OAT below +1 °C means briefing the taxi-in engine ice-shedding of article 09 before you land, not after.

8. The bills: ETOPS fuel and the hidden anti-ice assumptions

ETOPS critical fuel. The special-operations chapter prices icing into the diversion reserve:

The most critical scenario must be compensated for the greater of: A. The effect of airframe icing during 10 % of the time during which icing is forecast, including ice accumulation on unprotected surfaces, and the fuel used by engine and wing anti-ice during this period. B. Fuel for engine and wing anti-ice for the entire time during which icing is forecast.

With the settling note — The ETOPS icing fuel reserve is always limited by (B) — and, absent a trustworthy forecast, a presumption rule that extends the familiar definition:

Unless a reliable icing forecast is available, icing may be presumed to occur when the Total Air Temperature is less than +10 °C (50 °F), or if the outside air temperature is between 0 °C (32 °F) and -20 °C (-4 °F) with a relative humidity of 55 % or more.

Performance assumptions with anti-ice inside. Three places the button hides in the numbers: the FMS's recommended maximum altitude is computed assuming that the anti-ice is off (expect icing? — use the performance application instead; the engine-out ceiling makes the same assumption); the approach-speed build-up prices ice directly — per the landing-performance chapter, the approach correction is the highest of 5 kt in the case of A/THR ON ‐ 5 kt in the case of ice accretion ‐ 1/3 headwind component excluding gust (maximum 15 kt) — so an iced aircraft flies VAPP at least VLS + 5 even before wind; and the contaminated-runway envelope has an absolute edge: Dispatch to a runway covered with wet ice is not permitted, unless a specific method for performance assessment has been established by the operator. Reporting closes the loop — the datalink position report carries a dedicated icing field graded T / L / M / S (trace, light, moderate, severe).

Self-test

[!note]- Q1. State the engine anti-ice rule, its exemption, and why descent cancels the exemption.

On whenever icing conditions exist or are anticipated — without waiting to see ice — except in climb and cruise below −40 °C SAT, where supercooled water barely exists. Descent demands it even below −40: the aircraft is heading back into moist air, and the pushbutton buys the FADEC's raised idle — flame-out protection at descent power.

[!note]- Q2. Wing anti-ice: when may you, when must you?

May: preventively, once icing conditions are met. Must: with evidence of accretion — ice on the visual indicator, ice on the wipers, or the SEVERE ICE DETECTED alert. May prevents; must removes.

[!note]- Q3. Why does the aircraft drift above the descent profile with anti-ice on, and what covers it?

Anti-ice raises idle thrust, flattening the idle glide. The FMS idle segment already includes a small thrust margin (hence THR DES); beyond that, increase speed or extend up to half speedbrake — and feed the anti-ice status to the descent predictions honestly.

[!note]- Q4. Reproduce the icing minimum-speed floors and explain the clean-configuration paradox.

WAI operative: clean VLS + 15, configured VLS + 5. WAI inoperative: VLS + 10 / green dot. Clean pays more with anti-ice working because the heated surfaces are the four outboard slats — retracted when clean, so protection is nil and the allowance covers wholly unheated structure. The inoperative column is the same package the severe-ice failure procedures impose.

[!note]- Q5. Why hold clean in icing, and why do the flaps stay out after landing from an icing approach?

Slats extended in icing grow ice in the slat gap and on unheated undersides — prolonged exposure is to be avoided, so hold clean. After landing the same ice would be crushed into the slat tracks by retraction: flaps stay out until engines are shut down and the ground crew confirms the mechanism clear.

[!note]- Q6. List five ice-crystal clues and the avoidance doctrine.

"Rain" with a hiss on the windshield at impossibly cold temperatures; small ice grains on the wipers; ozone / St Elmo's smell; TAT stuck near 0 °C (the probe itself frozen); light-moderate turbulence in high-altitude IMC; weak echo at level with heavy returns painting below / downwind of a big cell. Avoid laterally rather than vertically, honour the margins, deviate upwind.

[!note]- Q7. Cold-temperature corrections: the approximation and a worked example.

≈ 4 × ΔISA × height / 1000 (sea-level aerodrome, conservative above). At −20 °C (ΔISA ≈ 35) and 3 000 ft: 4 × 35 × 3 ≈ 420 ft to add — applied to FAF, step-downs, minima and altitude-distance checks, with FPA corrected in selected mode.

[!note]- Q8. Why is the ETOPS icing reserve "always limited by (B)", and where else do the numbers assume anti-ice?

Option B — anti-ice fuel for the entire forecast icing time — always exceeds option A's 10 %-exposure accounting, so B sizes the reserve. Elsewhere: REC MAX assumes anti-ice off (use the performance application when expecting ice); VAPP adds 5 kt for ice accretion; wet-ice runways are no-go without an operator-approved method.

Key takeaways

References

Engine anti-ice standing order and ground phrasing per the FCOM after-start/taxi SOPs; climb/cruise exemption and wing may/must rule per the after-takeoff SOP; descent rule, path-angle note and low-temperature prompt per the descent-preparation SOP; approach WING selection line per the approach SOP; flap-retention and taxi-speed lines per the after-landing SOP. Minimum-speed table, accounting sentences and slat caution per the FCOM adverse-weather supplementary procedures. Cloud-entry TAT warning, holding technique, slat-slot crushing mechanism, barometric-error statements, correction table and approximation per the FCTM cold-weather chapter; THR DES margin per the FCTM descent chapter; ice-crystal mechanism, clues and avoidance doctrine per the FCTM weather-radar chapter. ETOPS icing reserve and presumption per the FCOM special-operations chapter; REC MAX assumption per the FCOM flight-management description; VAPP correction and wet-ice restriction per the FCOM landing-performance chapters; datalink icing field per the FCOM communications description. The rhythm framing, the invisible-payload metaphor and the worked correction example are integrative syntheses of the referenced material.

Independent study material, not an Airbus publication and not endorsed by the manufacturer. Always defer to the current operator FCOM, FCTM, and QRH for operational use.