Short final, seven hundred feet, a cell trailing three miles from the threshold: the indicated speed increases by about fifteen knots without anyone touching the throttles. The most natural reflex in the world is to reduce this speed. This is also the action that, a few seconds later, will cause the aircraft to enter the descending part of the phenomenon with the engines at idle. A microburst is not played out on a needle: it is played out on an energy balance, of which speed is only the trace, and always lagging.
When wind suddenly changes, your speed changes too—but that’s just a symptom. What really matters is the aircraft’s total energy (height + speed). If you lose height without gaining speed, you’re in trouble. The key is to add power, not just try to keep speed.
An Airbus A320 at 500ft AGL on final approach encounters a microburst. The headwind suddenly increases by 20 knots, pushing speed from 130 to 150 knots. The crew reduces thrust to maintain speed. Then the wind shifts to tailwind, dropping speed to 100 knots. With engines at idle, the aircraft descends rapidly—exactly what not to do.
Wind is not a state, it’s a derivative
An aircraft installed in a uniformly moving air mass does not feel the wind: its airspeed remains unchanged, only its ground track is affected. What affects it is the variation of the wind along its trajectory. Its mass opposes any acceleration, while the air, for its part, changes speed in a fraction of a second. Over a short interval ahead of the aircraft’s response time, any change in wind speed is fully reflected, and in the opposite sign, on the indicated airspeed.
A microburst is a column of descending air with a small horizontal extent—its spread on the ground is measured in kilometers, not tens of kilometers—that strikes the ground and spreads radially. The aircraft that crosses it near its axis does not encounter three successive winds: it encounters a single continuous variation whose sign reverses. First, a headwind component that increases, then the spreading component that meets it; then, at the core, a descending vertical component; then a tailwind component that increases as it exits.
The vertical component deserves attention because it is the least readable of the three. It does not simply push the aircraft downward: by tilting the relative airflow, it reduces the angle of attack at constant pitch. The crew therefore sees a normal pitch, a decreasing speed, and a descending trajectory—three indications that do not contradict each other and, taken together, indicate that lift is lacking at the precise moment when it should be at its maximum. The phenomenon is brief—its most active phase typically lasts a few minutes—and it is most severe shortly after the column has struck the ground.
| Temps | What the instruments display | What the energy balance does | The right action |
|---|---|---|---|
| 1 · gain, increasing headwind | speed above target, tendency to climb above the glide path | the shear lends energy; the excess is temporary and will be taken back | hold thrust, do not reduce, call out the escape criterion |
| 2 · core, downward component | normal attitude, speed falling away, increasing rate of descent | angle of attack drops at constant attitude; height is lost with nothing in return | maximum thrust, reference attitude, configuration frozen |
| 3 · loss, increasing tailwind | speed well below target, possible reactive alert | the loan is repaid, plus what the vertical component cost | angle of attack up to stick shaker, wings level, hold until clearly out |
What the indicated airspeed tells us, and why it’s in the past
The right framework for understanding is not steady-state flight mechanics, but energy balance. The aircraft’s state is described by a specific energy height: its altitude, increased by its speed squared and divided by twice the acceleration of gravity. This single quantity tells us what the aircraft has available; its two terms only describe the distribution between altitude and speed.
In flight, it can only increase through the excess of thrust over drag. At the mass and configuration of a final approach, this excess is modest: it is worth a few knots per second, or a few hundred feet per minute, no more. That’s all the reserve, and it is only controlled by the throttles.
Wind shear, on the other hand, adds a term that no one controls. During the phase of increasing headwind, it lends energy: speed increases, for free. During the phase of increasing tailwind, it takes it back, and the descending component adds a loss of altitude that pitch alone does not compensate for. When the needle starts to drop, the transaction is already done: what is displayed is the balance of a movement that has already occurred. Over the coming seconds, the only variable still open is thrust—and it arrives with the acceleration delay of the engines, several seconds from an approach idle setting.
A thrust command, never a speed target
From this balance, the entire recovery procedure and the priority it imposes follow. Exchanging altitude for speed by pulling back on the control column creates nothing: you slide along the same equal energy curve. You pay in altitude—the only thing separating the aircraft from the ground—for what you gain on the needle. Adding thrust, on the other hand, changes the curve. In a microburst, the deficit is in energy, not speed: it is therefore thrust, and only thrust, that addresses the problem.

The recovery procedures in the type documentation say exactly this, in this order: maximum available thrust without delay, pitch attitude brought to a reference value—around fifteen degrees on many transport aircraft—fixed configuration, and explicit acceptance of a speed that falls below the approach target. This is the complete inversion of the logic that governs the rest of the flight, where speed is the regulated variable and the trajectory is the consequence. Here, speed becomes the currency and the trajectory the asset that is bought with it.
The trap lies in the brief moment of good news that opens the crossing. The increasing headwind component causes the speed to rise well above the target; the crew reduces it to bring it back, or the autothrottle does so in speed-hold mode. The aircraft then approaches the descending core with the engines at idle and the full acceleration delay ahead of it. A literal reading of stabilization criteria pushes in the same direction, since a speed overshoot is listed as a deviation to be corrected. The fatal sequence never begins with a loss: it always begins with a gain.
The right reflex is to treat an unexplained speed increase, on final approach and near a cell, as a warning and not as a gift. It is verbalized, thrust is maintained, the speed hold is disengaged rather than letting it reduce, and the escape criterion is set out loud before the loss occurs: afterward, there will be no time to decide, only to execute.
Angle of attack is the only limit that matters
Once thrust is at maximum and the configuration is fixed, there is only one lever on the trajectory: angle of attack. The escape consists of seeking the maximum usable lift coefficient and holding it, which, on a conventionally controlled aircraft, means flying to the buffet—you pitch up until it triggers, you ease off just enough, you repeat. The maneuver is uncomfortable, it contradicts all acquired reflexes, and it is only maintained in the simulator.
- Leading edge contamination: Even a thin layer of ice or dirt lowers the maximum lift coefficient and can shift the buffet onset, reducing your margin
- Bank angle: A turn increases the load factor, thus the stall speed, exactly when your margin has disappeared
Two things reduce this reserve without warning. The first is the condition of the leading edge: even a thin contamination lowers the maximum lift coefficient and can shift the buffet onset. The second is the bank angle: a turn increases the load factor, thus the stall speed, precisely when the margin has disappeared. An escape is flown wings level, unless terrain dictates otherwise.
On aircraft with flight envelope protections, the angle of attack limit constitutes a hard floor: it prevents exceeding it, but it does not add any performance. The pilot of a conventionally controlled aircraft must manually reconstruct this floor, with the buffet as the only indicator. The difference lies in the workload and the precision of the adjustment, never in the physics.
Takeoff and approach: the same balance, two reserves
The reasoning is identical in both phases, but the reserves are not, and they determine the decision criteria.
| Energy balance item | On take-off | On approach |
|---|---|---|
| Height reserve | almost none, and nothing to spend | exists, but it is the only protection left |
| Additional thrust | the margin between the reduced rating used and the limit, released immediately | la poussée de remise de gaz, disponible en totalité |
| Configuration drag | take-off flaps, gear still down | landing flaps and gear down, maximum drag |
| Decision before commitment | reject while speed and remaining runway allow | discontinue the approach on the predictive alert, without discussion |
| Quantity sacrificed | speed, and the initial climb | speed, and stabilisation |
| What does not change | configuration frozen until clear | configuration frozen until clear |
In the approach, there is altitude to spend—and that is precisely why you should not spend it: it is what allows you to wait for the engines to spool up. At takeoff, the balance is the most unfavorable possible: little altitude, little speed, a configuration with high drag. One point deserves to be in the mind of everyone taking off at reduced thrust: this reduction is an immediately available reserve, and pushing the throttles to the stop releases it without delay. A takeoff conducted from the outset at maximum rated thrust does not have this margin, which shifts the decision cursor toward a go-around.
Anticipate rather than react
Onboard predictive detectors work on the divergence of radial speeds measured ahead of the aircraft: they therefore see what contains scatterers, i.e., a wet microburst. The dry microburst, under a high base and a virga curtain that evaporates before the ground, remains largely transparent to them. The reactive alert, on the other hand, only triggers once the aircraft is in the phenomenon: it is a statement, not a warning.
Remain the indicators that cost nothing to collect: a precipitation curtain whose base widens, a dust ring or a whirlwind at ground contact, a sharp drop in temperature accompanied by a wind shift at the airfield, a gust front progressing. And above all, the report from the preceding aircraft, which is both the most useful and the most perishable data: a microburst only lives a few minutes, and a « nothing to report » already several minutes old says nothing about the coming minute. The distance to maintain from a cell is not set by the intensity of the radar echo but by the extent of its spread, which far exceeds the visible rain.
What needs to be checked before needing it
Three points are read in advance, in your own documentation. What reference pitch does your type prescribe, and does your flight director have a dedicated escape mode or should it be ignored during the maneuver? What does your autothrottle do if you do not disengage it, and under what condition does it relinquish control? Do your speed brakes retract automatically when maximum thrust is applied, or do you have to think about it at the worst moment?
The gray area, however, is upstream of the maneuver. A predictive alert in the approach requires a go-around, and that is not up for discussion. But an unexplained speed increase, without an alert, is a matter of judgment, and operators do not all treat it the same way: some make it a separate interruption criterion below a given altitude, others leave it to the captain’s discretion. Knowing what your operations manual says on this specific point is better than discovering it at seven hundred feet.
For those preparing for a type rating or a check, there remains the question of what is actually being evaluated: it is never the speed control. It is the moment of the decision, the fact that thrust has gone to the stop before anything else, that the configuration has not moved, and that the angle of attack has been maintained until a clean exit. A candidate who preserves speed and loses trajectory is a failure; the one who flies to the buffet with a clean trajectory is not, even if the sequence was ugly to watch. This is one of the few exercises where the grade is entirely based on the hierarchy of priorities, and it is worth reviewing the mechanics before the session.
What is done afterward finally counts for the next ones. A numerical report—observed gain then loss, altitude, position relative to the threshold, descent rate achieved—is infinitely better than a « moderate wind shear, » and it must be passed on while it is still true. It is also necessary to check whether any envelope limits were reached during the recovery, which falls under the technical report and not just the peer-to-peer account.
This article describes mechanisms; it does not replace any enforceable document. The applicable regulation, your operator’s approved documentation, and your aircraft type’s operating manual—recovery procedure, alert thresholds, reference pitch, operation of automation—prevail in all circumstances over what is written here.

Leave a Reply
You must be logged in to post a comment.