Wind shear: the energy reading that sets the order of actions

Pilots in a cockpit at night, reacting to wind shear during a rainy approach.

Five hundred feet, approach stabilised, and the airspeed rises by eight knots without anyone asking for it. The automatics reduce thrust, the flight path climbs a little above the glidepath, everything seems to be settling down. Twenty seconds later the airspeed collapses, the rate of descent doubles, and the engines are at flight idle. What is happening there cannot be read on the airspeed indicator: it has to be reasoned out from the energy state of the aircraft and, above all, from the time each lever takes to change it. It is that reading which explains the apparent paradox of the manoeuvre — pitch precedes thrust in its effects, whereas thrust precedes pitch in the order of actions.

What wind shear takes away, and in which frame of reference

A steady wind, however strong, costs the aircraft nothing: it moves the air mass, and the aeroplane with it. What does cost is the variation of that wind along the flight path. The inertia of the aircraft is referenced to the ground, its aerodynamics to the air. When the headwind component decreases by ten knots, the groundspeed does not change in that instant — the aircraft is far too heavy for that — and so it is the airspeed that loses ten knots. Lift and drag see that deficit immediately; inertia does not see it yet. The imbalance is not an instrument illusion, it is real in the only frame of reference in which the wing works.

Hence the relevant quantity: not the wind, but its rate of change. One knot of loss per second imposes on the aircraft a longitudinal acceleration of the same value if it is to hold its airspeed. Now, in an energy balance, a longitudinal acceleration and a climb gradient are paid for in the same currency: to accelerate is to give up climbing, and vice versa. One knot per second is worth on the order of five per cent of gradient, that is, at approach speeds, some seven hundred feet per minute of rate of climb sacrificed. Three knots per second therefore consume more gradient than many aircraft have installed in the landing configuration, go-around included. The arithmetic is brutal, and it is the real subject: in a severe shear, the available thrust is not enough, however fine the handling.

When wind suddenly changes, the plane loses speed because its weight doesn’t let it instantly adjust. This loss forces the plane to either climb slower or speed up. If the wind changes too fast, the plane can’t climb enough to stay safe, no matter how hard the engines push.

A Boeing 737 on final approach at 130 knots (240 km/h) encounters a 3-knot-per-second wind shear. This requires an equivalent climb gradient of 15% (3 × 5%). If the aircraft’s maximum climb gradient in landing configuration is only 12%, it cannot maintain altitude, even with full thrust.

Gradient required by the rate of airspeed loss, against the gradient availableGraph. On the horizontal axis, the rate of airspeed loss, from zero to four knots per second. On the vertical axis, the equivalent gradient required, from zero to twenty per cent. A solid straight line rises from the origin to twenty per cent, at about five per cent of gradient per knot per second. A horizontal dashed line represents the gradient the type is able to produce; to the right of its intersection with the straight line, the shaded area marks the gradients that no amount of thrust can offset.01234airspeed loss (kt per second)05101520equivalent gradient required (%)gradient requiredgradient available on typethrust no longer suffices
Beyond a certain rate of loss, the problem stops being a thrust problem. The straight line converts the rate of change of the wind into an equivalent gradient, at about five per cent per knot per second. The level of the dashed line depends on the type, on the weight and on the conditions: it is placed here to be read, not as a value. To the right of the intersection, no finesse of handling closes the gap.

The vertical component does not act in the same way at all. A downdraft takes no speed away: it tilts the flight path downwards by an angle equal to the ratio of its own speed to the airspeed of the aircraft. Five hundred feet per minute of downdraft at approach speed amounts to something like three and a half per cent of gradient taken away without the airspeed indicator moving. And the same current briefly reduces the angle of attack seen by the wing, hence the lift, at the very moment when the flight path needs to be curved upwards. That is why the hazard indices computed on board add the two terms together: the horizontal shear rate referred to the acceleration due to gravity, and the vertical speed of the air mass referred to the airspeed. The sum is expressed as an equivalent gradient, which makes it directly comparable with the gradient the aircraft is able to produce. The trigger threshold is set by the specification applicable to the equipment, not by the crew.

Two levers, two time constants

The total specific energy of an aircraft comes down to a height: its altitude, increased by the height to which its speed could hoist it if that speed were converted entirely. Only two operations are available on that quantity. One increases it, and that is the excess of thrust over drag. The other distributes it between its two forms without adding anything to it, and that is pitch. The whole wind shear manoeuvre rests on the fact that these two operations have neither the same time to establish themselves, nor the same limit, nor the same way of signalling that limit.

A pilot grips the control yoke in a storm, with another aircraft visible through rain-streaked windows.
Wind shear demands precise control and immediate action to ensure safety.

Thrust is the only term that creates energy, and it is the slowest. Between an approach setting and go-around thrust, engine acceleration is counted in seconds, and the rate of addition it finally provides remains bounded by the installed thrust, itself degraded by altitude and temperature. That limit announces nothing: it is reached without warning, the lever against its stop, and there is no reserve behind it.

Pitch, for its part, acts immediately, but at zero sum. Its exchange rate is instructive: close to approach speed, ten knots of speed are worth on the order of a hundred feet of height, because the conversion varies as the square of the speed. The same ten knots in cruise are worth three to four times more. In other words, the exchange rate is at its worst exactly where the ground is closest. The practical consequence is counter-intuitive: since speed buys little height near the ground, it must be spent early and decisively, not in small doses, on pain of paying for the delay of the engines in a currency that is worth almost nothing.

One subtlety of chronology remains. In the first two seconds, pitch acts not through the exchange of energy but through lift: increasing the angle of attack produces an increment of lift that curves the flight path, before the bill arrives in speed. It is that offset which makes the manoeuvre possible. Curvature is borrowed at once, repaid in knots afterwards, and one hopes the engines will have arrived by then.

Effect of pitch and of thrust on the flight path as the seconds passTwo curves plotted against the time elapsed since the action, from zero to twenty seconds. The solid curve, pitch, peaks within the first two seconds then decays as the exchange reserve is used up. The dashed curve, thrust, remains almost nil during engine acceleration, then rises and settles on a plateau. The two cross in the middle of the interval: that is the handover between the immediate lever and the only one that remains.05101520time since the action (seconds)effect on the flight pathpitchthrusthandoverexchange reserve exhausted
Pitch acts at once but runs out; thrust arrives late and stays. The solid line borrows curvature from the moment the angle of attack is increased, before the bill arrives in knots; the dashed line produces almost nothing during engine acceleration. The crossing is the handover, and it is what is missed by calling for thrust too late.
The two levers of the energy state, compared term by term
Criterion Thrust Pitch
Quantity controlled Rate of addition of total energy Distribution between altitude and speed
Time to establish Several seconds, engine acceleration Immediate, from the moment the angle of attack is increased
Limit Installed thrust, degraded by altitude and temperature Angle-of-attack limit, hence maximum lift
Effect in the first two seconds Negligible Curvature of the flight path through an increment of lift
Effect after some twenty seconds The only term that remains Exchange reserve exhausted
Signalling of the limit None: the stop is reached without warning Progressive: buffet, warning, angle-of-attack stop
What the crew controls The moment at which it is called for The complete profile of the manoeuvre

Why the order of actions reverses the order of effects

Read in this way, the sequence prescribed by the procedures ceases to be a convention. The lever that arrives last is started first. Thrust is called for straight away because its effect is delayed, and pitch serves to hold the flight path during the wait. The sequence is a chronology, not a hierarchy: thrust is first in time, pitch is first in effect.

This reading also gives the tipping criterion, and it is simpler than any list of symptoms. As long as a deviation can be recovered by an adjustment of thrust, it is a handling problem. As soon as the rate of the deviation exceeds what thrust is able to produce, it is no longer a thrust problem, and to go on treating it as one amounts to letting the flight path degrade while the airspeed is being nursed. The crew that fixates on the airspeed indicator will lower the nose to recover the lost knots, and will convert the one resource that cannot be repaid near the ground: height.

The spendable resource and the forbidden one

Why spend speed rather than altitude, when the energy balance treats the two in the same way? Because their limits do not behave in the same way. Near the ground, altitude is limited by a silent and non-negotiable boundary, one that gives no warning and grants no margin. Speed, on the other hand, is limited by a boundary that announces itself: buffet, stall warning, then the angle-of-attack stop on aircraft that have one. The principle can be put in one line: spend the resource whose limit is signalled, protect the one whose limit is not.

The same principle explains why the trade-off is reversed at high altitude, where the envelope closes between low-speed buffet and the Mach number limit: speed is no longer the abundant resource there, and the compromise changes direction. There is nothing doctrinal in this, only a consequence of the geometry of the limits.

The reserve still has to be named correctly. It is not a speed margin, it is an angle-of-attack margin. Two aircraft showing the same excess over the reference speed do not have the same margin, depending on their weight, their configuration, their centre of gravity position, the state of their leading edges. It is that margin which the escape guidance manages where it exists, and it is that margin which decides how far pitch may go — not a figure on the airspeed indicator.

Crossing a downburst: the four moments

A downburst does not present itself as a continuous degradation. It has a structure, and each moment of the crossing offers a different reading, sometimes a reassuring one. It is that structure that must be borne in mind, because the right action at the third moment depends on what was done at the first.

Side view of a downburst crossing on approachVertical section. A shaded shaft descends from a cloud base to the ground and spreads out sideways: the left-hand arrow travels towards the aircraft and gives it a headwind, the right-hand one travels with it and gives it a tailwind. The dashed line is the nominal glidepath, straight as far as the runway threshold. The solid curve is the actual flight path: it first passes above the glidepath at point one, tilts at point two in the downdraft, sinks to the vicinity of the ground at point three when the outflow is taken from behind, then climbs again at point four.1234downburstnominal glidepathactual flight pathheadwindtailwindrunway threshold
The first moment of the crossing is the most deceptive: it looks like an improvement. Points 1 to 4: outflow met head-on, core and downdraft, outflow taken from behind, exit. The dashed line is the nominal glidepath, the solid curve the actual flight path, which tilts at point 2 before the airspeed indicator announces it.
Crossing a downburst: what each moment presents, and where it traps
Moment Airspeed Flight path Spontaneous reaction of the automatics Cost of the error
Outflow met head-on Rising with no crew action Tends to go above the glidepath Reduces thrust to hold the target speed Entering the core with the engines at idle
Core and downdraft The gain reverses, back towards the target Tilts downwards before the speed announces it Nothing: the speed appears to be held Attributing the sink to a handling fault
Outflow taken from behind Rapid fall, angle-of-attack margin consumed Marked sink, ground proximity warning possible Calls for maximum thrust, with the delay of the engines Pushing over to recover the lost knots
Exit from the phenomenon Rises again, sometimes abruptly Straightens out, tendency to overshoot Reduces as soon as the target is regained Relaxing before actually being out of it

The trap. The favourable entry. The first moment presents itself as an improvement: the speed rises, the aircraft rises, and everything learned about the stabilised approach urges reducing and pushing back down to regain the glidepath. One then enters the core with the thrust withdrawn and a flight path advantage that has just been squandered. The initial speed gain was not a gift: it was the first half of the shear, and the second half always arrives.

The right reflex. Near the ground, treat every unrequested variation of speed as the signature of a gradient, and not as a deviation to be corrected. In practical terms: do not let the thrust fall below the setting from which it comes back quickly, call the reading aloud, and fix in advance the value that triggers the escape rather than deliberating about it at the moment the flight path is degrading. A decision taken in the cold is worth more than a diagnosis attempted in the heat.

Configuration: why it is not touched

The energy balance argues for reducing drag: retracting the landing gear increases the rate of addition, and it costs no speed. The angle-of-attack reasoning says the opposite for the high-lift devices: retracting them lowers the maximum lift, and so empties the account on which the manoeuvre is drawn. Near the ground, the account weighs more than the income, and that is why the memorised rule exists. It is not a superstition, it is a trade-off between a slow gain and an immediate loss of reserve.

To this is added an argument of transient: any flap travel is accompanied by a variation of lift and a change of pitching moment to be trimmed out, at the very moment when hands and attention are elsewhere. Some type documentation authorises retracting the landing gear during a go-around already in progress; it is the procedure for the type that decides, not the general principle.

Two gains remain free and are often forgotten. The speedbrakes, if they are extended, retract with no counterpart. And bank is paid for twice: on the order of three per cent of the vertical component of lift lost around fifteen degrees, plus the associated induced drag. The outflow of a downburst is rarely symmetrical, it tends to roll the aircraft; resisting that and coming out with the wings appreciably level is not a matter of comfort, it is gradient preserved.

What automatic protection does, and what it does not do

One sentence sums up the whole section: no on-board device adds energy. Some detect, others prevent a limit from being crossed, the best of them optimise the spending of a reserve whose amount they do not change. Knowing which does what avoids two symmetrical errors, that of waiting for a warning before acting and that of refusing to follow a guidance command because it calls for an unusual pitch attitude.

On-board devices: quantity observed, moment of action, structural limit
Device Quantity observed Moment at which it acts Structural limit
Reactive warning Difference between inertial accelerations and air data Once the aircraft is inside the phenomenon Never warns before entry
Predictive warning Divergence of radial velocities within precipitation Before entry, within a limited sector and range Blind to dry shear and to non-radial gradients
Autothrust A target speed Continuously, unless inhibited Defends a speed, not an energy state
Low-speed protection Angle of attack, or speed depending on the type In the vicinity of the limit Prevents the limit from being exceeded, does not create lift
Dedicated escape guidance Angle of attack and flight path On warning or on selection Optimises the spending, adds nothing to the account
Ground proximity warning Height and closure rate Too late for prevention May compete with the wind shear escape

The real gain from a dedicated guidance deserves to be understood for what it is. It does not manufacture gradient: it turns a difficult question, “how far can one pull without stalling”, into an executable instruction, pull to the stop and hold. On an aircraft with angle-of-attack protection, that transformation is decisive, because it makes the complete spending of the reserve both safe and quick. On an aircraft that has only a warning, the same spending is flown by feel, and the margin kept out of prudence is gradient lost.

The trap. Taking the absence of a warning for the absence of shear. The predictive device needs particles in order to return an echo: a gradient of thermal origin or a downdraft in dry air present it with almost nothing. It is furthermore inhibited or degraded in certain phases and at certain attitudes. As for the reactive device, it measures the reaction of the aircraft, never the phenomenon: it confirms what the flight path has already begun to tell.

Take-off: the same reasoning with an empty account

On take-off, every term of the balance works against the crew. The speed is low, so the exchange rate between knots and feet is at its worst. The configuration is already the one that was chosen and there is nothing to be gained from it. The engines are at take-off thrust: the slow lever has already been used up, there is nothing left to add. And the ground is immediately below. Past the decision speed, the option of stopping has gone.

It follows that the only real lever on take-off lies before the take-off. A take-off roll not begun costs nothing; a take-off roll begun in a severe gradient cannot be recovered by technique. Reduced thrust deserves a finer reading than a prohibition. A reduction obtained by an assumed temperature does not remove the reserve: the levers can be advanced to the maximum rating. A reduction declared as a limitation of the type, on the other hand, may forbid exceeding the rating set, in particular for reasons of ground controllability on one engine. In both cases, what is lost is not the thrust but the time and the attention needed to go and get it, at the very moment when the flight path is degrading — and the first segment is precisely the moment when there is neither of the two to spend. Most operations documentation draws a consequence from this for the choice of rating in thundery weather, but it is yours that decides, not the principle.

There remains the symmetrical trap: a favourable entry on take-off can produce an early lift-off, with a real energy lower than the speed suggests, followed by the reversal at the moment when the aircraft no longer has any runway beneath it. The pitch discipline on the first segment, held on angle of attack and not on the airspeed indicator, is what separates the two outcomes. It is also one of the exercises where the simulator brings the most, and it is worked on in training.

To be checked in your own documentation

Five points deserve to be opened without waiting for the next session. Does your type have a dedicated escape guidance, and in what mode does it present itself? Does the procedure prescribe a numerical target pitch attitude or an angle-of-attack instruction held to the limit? What does it say about the autopilot and the autothrust during the escape, kept in or disconnected? What are the criteria that allow the phenomenon to be considered left behind, in other words when does the manoeuvre stop? Lastly, what does the crew do if a ground proximity warning is superimposed on the wind shear escape, the two responses not being identical on every type?

A grey area remains, and it is better to know it than to ignore it. No airborne system measures the shear: they all measure the reaction of the aircraft, filtered by its inertia and by the lag of its air data chain. Two reports of the same phenomenon by two different types therefore do not compare term for term, and an approach report does not transpose as it stands to a take-off, nor to another height band. What is genuinely transferable in a report is a rate of change, a height band, a phase and a class of aircraft — not an isolated figure of speed loss. The formats that are genuinely usable are exchanged between operators on the forum.

For anyone preparing a type rating, one question quickly separates the candidates: why does the procedure have maximum thrust called for first if it is pitch that produces the effect? Answering by the two time constants, and by the fact that pitch borrows curvature before repaying it in knots, is worth more than a recitation of the immediate actions. The corollary deserves to be known as well: there are gradients that the aircraft does not have the gradient to cross, and the only effective decision in that case is taken before entering them.

The applicable regulation, the flight manual for the type and the approved documentation of the operator prevail over this article, which sets out a mechanism and does not take the place of a procedure.

Questions fréquentes

Why does the procedure call for maximum thrust before pitch, if it is pitch that acts?

Because thrust is the slowest lever: engine acceleration is counted in seconds, and so its effect arrives last. It is called for first so that this delay starts running straight away, and the flight path is then held on pitch during the wait. The order of actions is a chronology imposed by the time constants, not a hierarchy of effectiveness: in the first seconds, only pitch produces a measurable effect.

Why is a favourable entry, with the airspeed rising, dangerous?

Because an unrequested speed gain near the ground is almost always the first half of a gradient, the outflow met head-on before the core is crossed. The stabilised approach reflex urges reducing thrust and pushing back down to regain the glidepath, which means entering the unfavourable part with the engines at idle and the flight path advantage spent. The energy reading says the opposite: keep the thrust and the excess altitude, which are exactly what the second half will need.

How far can the airspeed be allowed to decay during a wind shear escape?

The real limit is not a speed but an angle of attack, and it is signalled progressively by buffet and then by the stall warning, before the stop on aircraft that have one. The logic of the manoeuvre is to spend that reserve up to its signalled limit in order to protect height, whose limit is not signalled. The exact value and the way of reaching it are set by the procedure for the type, which differs according to whether there is a dedicated guidance or only a warning.

Should the configuration be changed during the manoeuvre?

Retracting the high-lift devices would improve the drag balance but would lower the maximum lift, that is, the angle-of-attack reserve on which the whole manoeuvre is drawn; near the ground, that reserve counts for more than the gain in drag. To this is added the transient of the flap travel, badly placed at the moment when attention is taken. Some type documentation authorises retracting the landing gear during a go-around already in progress: it is the procedure for the type that decides, and it alone.

Does the absence of a warning mean there is no shear?

No. The predictive device exploits weather radar returns and therefore has almost nothing to observe in a gradient in dry air or of thermal origin; it is furthermore limited in range and in sector, and inhibited in certain phases. The reactive device, for its part, does not measure the phenomenon but the reaction of the aircraft, which makes it structurally later than the entry. The absence of a warning is therefore weak information, whereas a warning is strong information.

Is a wind shear report issued by a preceding aircraft usable as it stands?

Only in part, because an on-board system measures the reaction of the aeroplane, filtered by its inertia and by the lag of its air data chain: two different types do not give the same account of the same phenomenon. What transfers usefully is a rate of change, the height band concerned, the phase of flight and the class of aircraft, with the time. A figure of speed loss without its rate or its height allows no decision, and an approach report does not transpose directly to a take-off.

Commentaires

Leave a Reply

Découvrez l’écosystème FL330

FL330, c’est toute l’aviation professionnelle — explorez nos autres espaces.