The temperature entered in the take-off performance computation is often a lie, and it is a lie the certification basis knows about. Understanding why the margin survives it — and which margin does not — is the difference between a technique and a habit. The assumed temperature buys turbine life. It is paid for in acceleration, in the runway left behind at the decision point, and above all in the shrinking distance between a keystroke and a tailstrike.
Two mechanisms that look identical from the seat
Every reduced-thrust take-off is one of two things, or a combination of both, and crews who use them interchangeably are usually wrong about the minimum control speeds.
A derate is a certified engine rating. The manufacturer establishes a lower take-off thrust rating, publishes flight manual data for it, and the aeroplane is certified at that rating as though it were a different, less powerful aeroplane. The consequence that matters operationally is that the minimum control speeds — on the ground and in the air — are recomputed for the derate. They are lower, because the asymmetric thrust that has to be countered is lower. That is why a derate can help on a short, narrow or contaminated runway where the ground minimum control speed, not the field length, is setting the decision speed.
A reduced-thrust take-off using an assumed temperature is not a rating. The engines remain certified at the full rating, full thrust remains available, and the crew may restore it at any moment. Because it remains available, the minimum control speeds stay those of the full rating: nothing has been certified downward. The reduction is a computation, not a limitation.
The two combine. A great many operators dispatch on a derate and an assumed temperature on top of it, which is legitimate and which makes the distinction more important rather than less: the minimum control speeds come from the derate, and the assumed temperature sits above them without changing them.
| Derate | Assumed (flexible) temperature | |
|---|---|---|
| Nature | A certified thrust rating | A performance computation at a fictitious temperature |
| Minimum control speeds | Recomputed for the derate — lower | Those of the full rating — unchanged |
| Full thrust during the take-off | A limitation; exceeding it invalidates the control speed basis | Available at any time, by design |
| Typical benefit | Engine life, plus relief where control speeds are limiting | Engine life only |
| Where it is written | Flight manual, as a rating | Flight manual procedure plus operator policy |
| Combinable | Applied first | Applied on top of the derate; control speeds stay the derate’s |
Where the fictitious temperature comes from
The method exists because of how a turbofan is rated. Below a corner point — the flat rating temperature — the engine control system holds take-off thrust essentially constant as the ambient temperature rises, trading turbine temperature to do it. Above that corner point it can no longer hold the value, and rated thrust falls with increasing temperature.
That falling branch is the whole trick. If the aeroplane’s take-off performance at some higher temperature would still permit the actual take-off weight from the actual runway, then the thrust corresponding to that higher temperature is enough. The crew enters the higher figure, the computation returns the rated thrust for it, and the engines are set to a lower target than the day would otherwise command.
The temperature is fictitious in the sense that no thermometer would produce it. It is not fictitious in the sense that matters: every performance quantity used to authorise the take-off has been recomputed at that value. Field length, accelerate-stop distance, the net take-off flight path, the second segment climb gradient, tyre speed, brake energy — all of them are evaluated as though the day really were that hot. The take-off is legal because it is legal at the assumed temperature, not despite it.
Why the certified margin survives
The conservatism comes from air density and from the way certified distances scale with it.
At the real, cooler temperature the air is denser than the computation assumed. For any given indicated airspeed the true airspeed is lower, so the ground speeds at which the aeroplane reaches its decision speed and rotation speed are lower than the assumed-temperature case predicted. Lower ground speeds mean less distance covered for the same indicated speed schedule, and less kinetic energy to dissipate in a rejected take-off. The engines, meanwhile, are producing the thrust that was assumed — not less.
The climb side behaves the same way. Second-segment gradient was demonstrated at the assumed temperature; flown in denser air with the same thrust setting, the aeroplane does at least as well. The obstacle assessment, which was built on the assumed-temperature net flight path, is therefore conservative.
This is the point that gets mangled in briefings. Reduced thrust does not consume the certified margin. It transfers it. The margin that would have existed on the runway had full thrust been used has been converted into turbine life. What remains is exactly the margin the regulations require, and no more.
The assumed temperature is a tool to reduce engine wear. The math ensures the plane can still take off safely, but it means slower acceleration and less runway margin. The safety buffer is just moved from the runway to the engines.
| Quantity | Status under an assumed-temperature take-off |
|---|---|
| Accelerate-stop distance | Computed at the assumed temperature; actual performance is better |
| Take-off distance and take-off run | Computed at the assumed temperature; actual performance is better |
| Second segment and net flight path | Computed at the assumed temperature; actual performance is better |
| Brake energy and tyre speed limits | Assessed at the assumed temperature |
| Runway remaining beyond the certified requirement | Consumed. This is what was traded away |
| Initial acceleration and time to the decision speed | Degraded. The decision point sits further down the runway |
| Energy available for an immediate windshear or terrain escape | Reduced until thrust is restored |
| Tolerance to an erroneous weight entry | Sharply reduced. The error attacks thrust and speeds together |
The real justification is turbine temperature, not runway length
Crews often explain the technique as « we have plenty of runway, so we use less thrust ». The available runway is the enabler, not the reason. The reason lives in the hot section.
Turbine gas temperature is the dominant driver of hot-section deterioration, and the relationship between peak temperature and component life is strongly non-linear: a modest reduction at take-off buys a disproportionate amount of on-wing time. The quantitative coefficients belong to the engine manufacturer and to the specific engine standard; no operator should be quoting a generic multiplier. What is not in dispute is the mechanism. Every take-off at full rated thrust erodes the exhaust gas temperature margin, and it is the exhaustion of that margin — not a calendar, not a cycle count in isolation — that eventually drives an engine off the wing and into a shop visit whose cost, on a large turbofan, runs into the millions.
Take-off is also a small fraction of the flight and an enormous fraction of the damage. That asymmetry is the entire business case: a few minutes of restraint per sector, repeated across a fleet and a decade, moves shop visit intervals materially. This is why reduced thrust is a policy, written into the operations manual, and not a matter of taste on the day.
The right reflex. When an assumed temperature comes out unusually high, treat it as a question rather than a result. An implausible flex number is one of the few free signals that the weight, the runway, the intersection or the flap setting fed into the computation is not the one you are about to fly.
When the method is off the table
Reduced thrust is permitted only within conditions set by the flight manual and the operator, and the list of exclusions is longer than most crews carry in their heads. The recurring items across manufacturers and operators are these, but the version that binds a crew is the one in their own documentation.
- Contaminated runways. The assumed-temperature method is commonly prohibited outright. The stopping performance model behind an accelerate-stop on contamination does not tolerate the trade.
- Reported or forecast windshear in the take-off area, and in many manuals any significant convective activity on the departure path.
- Anti-skid or thrust reverser unserviceabilities, and other minimum equipment list items that degrade stopping capability or engine control.
- A floor on the reduction. Airworthiness guidance caps how far take-off thrust may be reduced; for turbofans the figure commonly applied is that thrust may not fall below three-quarters of the full rated value. Both that cap and the maximum assumed temperature are stated in the flight manual, not inferred.
- Ice protection and engine bleed configurations that change the rating basis.
There is a further condition that is arithmetic rather than atmospheric: the assumed temperature can go no higher than the value at which the performance-limited take-off weight has fallen to the weight actually being flown. Which limit gets there first is not always the runway. Where the binding case is a climb or obstacle limit, the room to flex shrinks quickly and can vanish altogether, which is why a long runway does not guarantee a large reduction.
The entry error, and why this method sharpens it
The most serious cost of reduced thrust is not aerodynamic. It is that the technique removes the buffer that used to absorb bad data.
With full thrust and speeds computed for a weight that turns out to be too low, the aeroplane is over-powered for its actual mass and will usually fly, badly, and survive. With an assumed temperature, a weight entered too low propagates twice: it lowers the thrust and it lowers the speed schedule. The aeroplane is simultaneously heavier than assumed, less powerful than a full-thrust take-off would be, and asked to rotate early. The three errors compound instead of cancelling.
This is not theoretical. The investigation record contains a well-documented family of events built on erroneous take-off performance parameters. An Airbus A340-500 departing Melbourne in March 2009 was computed on a take-off weight roughly a hundred tonnes below the actual figure; the aeroplane struck its tail, damaged airport structures beyond the end of the runway and became airborne only after thrust was advanced. A Boeing 747 freighter at Halifax in October 2004 was operated on performance data left over from an earlier, lighter sector; it struck its tail on the runway, overran the end, was destroyed, and all on board were killed. Different operators, different continents, different aeroplanes, identical mechanism.
The trap. Nothing on a conventional flight deck compares the weight used by the performance computation with the weight the aeroplane actually has. The figures are self-consistent, the display is green, and the first physical evidence that they are wrong arrives at rotation. Recent types add a take-off surveillance function that cross-checks entered data against the runway in the navigation database and against the aeroplane’s position, but it is an aircraft-level function, not a universal one — assume you do not have it unless you know you do.
In 2009, an Airbus A340-500 in Melbourne was computed with a weight 100 tonnes too low. The plane struck its tail, damaged airport structures, and only became airborne after thrust was increased. This shows how a small input error can lead to a catastrophic outcome with reduced thrust.
| Wrong input | Effect on thrust | Effect on speeds | First symptom |
|---|---|---|---|
| Take-off weight too low | Too low | Too low | Sluggish acceleration, tailstrike or failure to rotate |
| Runway or intersection longer than the one used | Too low | Possibly too low | Runway end arrives early |
| Flap setting not the one selected | Mismatched | Mismatched | Configuration warning, or a rotation that does not behave |
| Assumed temperature above the manual limit | Too low | As computed | Usually none until the margin is needed |
| Stale data from a previous sector | Any | Any | Nothing at all — the figures are internally consistent |
What the crew can actually check
The defences that survive contact with a turnaround are the cheap ones, applied in the same order every time.

Two independent computations, from two sources, compared on the result rather than on the inputs. A gross-error check on the weight against a mental model of the day: the number of passengers, the fuel figure, a typical operating weight for the type. A check that the runway and intersection in the computation are the ones on the clearance, not the ones on the last sector. A stated expectation of the acceleration to be seen, and a willingness to reject for an acceleration that plainly does not match it while the speed is still low. And the standing permission, on an assumed-temperature take-off, to advance the thrust levers whenever the situation asks for it: that permission is the entire safety case for the method not being a derate.
None of this is exotic, and none of it is expensive. What it needs is a crew culture in which an implausible flex temperature is spoken aloud rather than accepted because the box produced it. That culture is easier to build in the classroom than in the cockpit, and it is one of the recurring threads in the performance material discussed on the FL330 forums.

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