Safety

Wind shear on the takeoff roll and the limits of V1

October 2, 202612 min readFor pilots

Leer en español · Lire en français · Ler em português

In this article

How V1 is certificated
What the rejected-takeoff record teaches
Wind shear breaks the link between airspeed and stopping energy
East Midlands, February 9, 2020
Perth, December 4, 2012
A groundspeed gate: the GSV1 proposal
How the gate reads against the two events
What it means for pilots

On February 9, 2020, in Storm Ciara, a Boeing 737-800 rejected its takeoff at East Midlands Airport five to six seconds after V1, at about 149 kt, and stopped with roughly 600 m of runway left (AAIB). Seven years earlier at Perth, another 737-800 saw its airspeed hang at 134 kt, 3 kt short of V1, as an 8 kt headwind turned into a tailwind of about 20 kt. It crossed the runway end at about 10 ft (ATSB).

Both crews met the same problem. The airspeed tape was telling the truth about lift, but it had stopped telling them how much energy they would need to stop. Below: how V1 is built, what manufacturers, the FAA and ICAO say about wind shear before V1, what the two investigations found, and a recent proposal to turn V1 into a groundspeed gate. That proposal is a preprint by the founder of LogAloft.

How V1 is certificated

Under 14 CFR 25.107, V1 is a calibrated airspeed chosen by the manufacturer. It can be no lower than the engine failure speed (VEF) plus the speed gained between the failure and the pilot's first action to stop. The Takeoff Safety Training Aid notes that jet certifications have used an engine failure recognition time of at least one second since the late 1960s.

14 CFR 25.109 then defines the accelerate-stop distance: accelerate to VEF, continue to V1, stop, and add a distance equal to two seconds at V1. On a dry runway, reverse thrust is not credited.

That two-second allowance is often misread as two seconds to decide. It is not. When the FAA adopted the current wording in 1998, it described the margin as covering the variability in how long line pilots take to complete the stopping actions and the element of surprise missing from flight tests. The same Federal Register notice says V1 is the speed by which the pilot has already decided to reject and begun acting.

Every one of these numbers rests on the wind used in the performance calculation. The training aid spells it out: at any airspeed, a 10 kt tailwind component adds 10 kt of groundspeed, and the distance needed to stop from V1 grows. V1 is read on the airspeed tape. The runway it protects is covered at groundspeed.

What the rejected-takeoff record teaches

The Takeoff Safety Training Aid, produced by a Boeing-led industry team for the FAA and released in 1993, set out what the jet era had taught about rejected takeoffs (RTOs). Its numbers, drawn from the western-built jet fleet through 1990:

About one takeoff in 3,000 ended in an RTO.
Over 75% of RTOs began at 80 kt or less. About 2% began above 120 kt, and those high-speed events produced most overruns.
Of 74 RTO overrun accidents and serious incidents, 58% were started above V1.
In about 55% of those events, the airplane could have continued the takeoff. About 80% were judged potentially avoidable.
Near V1 at heavy weights, an airplane covers 200 to 300 ft per second and gains 3 to 6 kt per second. A delay of a second or two costs several hundred feet.

A 2010 NLR study by Gerard van Es looked at what changed after the training aid. The high-speed RTO accident and serious incident rate fell by 25%, though the author found no clear proof the aid caused the drop. In hindsight, 44% of high-speed RTOs should not have been conducted, against 51% before. The training aid's core lesson stands: make the go/no-go decision before V1, and after V1, go.

Wind shear breaks the link between airspeed and stopping energy

The FAA's Pilot Windshear Guide (AC 00-54), part of the 1988 Windshear Training Aid, and ICAO's Manual on Low-level Wind Shear (Doc 9817), first published in 2005, carry nearly the same text. Recognition is hard on the takeoff roll because airspeed is changing fast; unusual fluctuations or a slow or erratic airspeed build-up may be the first sign. Both warn that go/no-go criteria based on V1 may not be valid in wind shear, because groundspeed can be much higher than airspeed, so a rejected takeoff may not stop on the runway. ICAO sums it up: "the ability to stop is largely a function of ground speed."

Their procedure is short. Below indicated V1, reject if airspeed variations are unacceptable and the pilot judges there is enough runway left to stop. After V1, continue. If VR may not be reached, rotate no later than 2,000 ft from the end of the usable surface.

Boeing's view appears in the AAIB bulletin. The 737-800 QRH tells crews that if wind shear is met before V1, there may not be enough runway to stop if an RTO is started at V1. Asked by investigators, Boeing explained that wind shear can push groundspeed above indicated airspeed, so stopping data may not hold and an excursion can follow even if airspeed shows V1 or less when the reject starts. Boeing added that event data show better outcomes when the takeoff is continued than when it is rejected late or without enough stopping distance, and that the decision rests with the captain.

Airbus says much the same in Safety First #19 (January 2015): in wind shear, V1 may come later or sooner than expected, and the pilot may have to judge alone whether enough runway remains. Airbus and the Flight Safety Foundation's ALAR briefing note 5.4 list airspeed variations over 15 kt and groundspeed variations among the cues. None of these documents gives a number for how much groundspeed is too much.

The physics explains why that matters. Stopping energy scales with groundspeed squared, so a 10% rise in groundspeed means about 21% more stopping distance, as the preprint discussed below works out. When a headwind decays during the roll, the airspeed can stall while groundspeed keeps climbing at the normal rate.

The automation will not flag it. On the East Midlands 737-800, the AAIB reports, predictive wind shear warnings were inhibited from 100 kt until 50 ft, cautions from 80 kt until 400 ft, and the reactive system armed only at rotation. The ATSB describes the same warning inhibit and rotation arming on the Perth aircraft. The preprint cites the A320 FCOM for a similar inhibit above 100 kt up to 50 ft radio altitude. Check your own type. Between about 100 kt and V1, the crew is the wind shear detector.

East Midlands, February 9, 2020

The AAIB investigation, published December 10, 2020, describes a Boeing 737-86N, G-DRTN, departing runway 27 in Storm Ciara. The 1050 UTC METAR read 210° at 40 kt gusting 56, 13°C, QNH 981. The runway was wet, and the operator's crosswind limit was 25 kt. The crew used the operator's wind shear performance solution: V1 134 kt, VR 149 kt, V2 156 kt, with a VRMAX of 158 kt. The co-pilot flew.

Quick access recorder data showed airspeed dropping from 131 kt to 118 kt before V1 and recovering. Just after the automated V1 callout, it fell from 142 kt to 129 kt in about 1.5 seconds and the aircraft yawed right. Both brake pressures rose to 3,000 psi five to six seconds after V1, at about 149 kt. The commander told investigators his snapshot was that airspeed was falling and they were below V1, so he called for the reject. He was also concerned the aircraft would leave the runway. The aircraft stopped on the centerline with about 600 m remaining. The No. 4 wheel and brake suffered heat damage.

The AAIB noted that no predictive or reactive alert would be expected at that stage. It recorded that the crew had not called V1 themselves, although the automatic callout sounded, and that the SOP is to continue once V1 is reached. It also wrote that wind shear alters the relationship between airspeed and distance along the runway unpredictably.

Perth, December 4, 2012

The ATSB report, released May 17, 2013, covers a Qantas Boeing 737-838, VH-VZL, departing Perth runway 06 for Canberra. ATIS gave 060° at 8 kt, a direct headwind, 37°C and CAVOK. Cumulonimbus sat 20 to 30 NM north. The EFB calculation assumed nil wind: 66.1 t, 24K thrust, flaps 5, V1 137 kt, VR 139 kt, V2 144 kt. Both pilots checked windsocks showing a headwind. There was no windsock near the runway 06 threshold.

Late in the roll, the airspeed stagnated at 134 kt for 3 to 4 seconds, just below V1. The captain saw the navigation display wind vector showing a tailwind of about 20 to 25 kt, disconnected the autothrottle and pushed the thrust levers to the stops. The first officer flew the wind shear escape maneuver after liftoff. Recorded data put the aircraft over the runway end at about 10 ft. No wind shear warnings occurred. After departure, ATC passed a threshold wind of 060° at 10 kt and a center-field wind of 280° at 15 kt.

The ATSB commented that the performance figures assumed nil wind, so the aircraft fell short of predicted performance once it met a 20 kt tailwind. Its safety message: wind changes on takeoff can be hard to predict, can happen without thunderstorms, and can differ between the two ends of a runway.

A groundspeed gate: the GSV1 proposal

In March 2026, Carlos Bousono posted a preprint, "Translating V1 to Groundspeed: A Decision Framework for Takeoff Wind Shear". Disclosure: the author founded LogAloft. The paper is a preprint under journal review, not yet peer reviewed. It is not an approved procedure from any manufacturer, operator or authority, and crews must follow their manufacturer's and operator's procedures.

The idea: V1 marks a certificated energy state under the wind assumed in the performance calculation. If that state is expressed as a groundspeed, the crew has a reference that a stalled airspeed cannot hide. The paper calls it GSV1:

GSV1 ≈ V1 (CAS) ÷ √σ + W

Here σ is the density ratio at the field, and W is the runway wind component used for the performance calculation, positive for a tailwind and negative for a headwind. The first term is simply the true airspeed at V1. The paper drops a compressibility term worth about 1 kt at takeoff speeds.

For the briefing, it offers a deliberately low shortcut: about 0.9 × V1 plus the same wind term, or 0.8 × V1 in very dense air (colder than −25°C or QNH above 30.50 inHg, about 1033 hPa). In a check of about 16,000 wind shear and wind shift reports from 60 airports, none fell in air dense enough to make the 0.9 factor too high. The cost is a gate 5 to 25 kt below the exact value at sea level, which the paper itself labels a missed-abort band.

The proposed briefing line is "Wind shear groundspeed gate ___ kt." If airspeed stagnates with normal engine indications and valid groundspeed, the PM calls the groundspeed. At or below the gate, or if the airplane cannot fly, the captain calls abort. Above the gate and able to fly, continue into the wind shear escape procedure.

Worked example (Perth). At 37°C, 67 ft and standard pressure, σ is about 0.93, so √σ is about 0.96. V1 of 137 kt becomes 137 ÷ 0.96 ≈ 142 kt true airspeed. The paper applies the ATIS 8 kt headwind (W = −8) for an exact gate near 134 kt, and a shortcut gate of 0.9 × 137 − 8 ≈ 115 kt. The ATSB says the EFB numbers assumed nil wind. Applying the paper's own rule of using the performance wind (nil), our arithmetic gives about 142 kt exact and 123 kt by shortcut.

How the gate reads against the two events

The paper applies the method to both investigations, using published report data, not flight recorder files.

	East Midlands 2020	Perth 2012
V1	134 kt	137 kt
Wind term used in the paper	−21 kt (METAR mean wind, 58° off the runway)	−8 kt (ATIS headwind)
Exact gate (paper)	about 115 kt	about 134 kt
Shortcut gate (paper)	about 100 kt	about 115 kt
Paper's reading	Already past the gate before the reject, favoring continuation	Below the gate when stagnation began, crossing it during stagnation

For East Midlands, the paper's reconstruction puts groundspeed near 148 kt at the reject, well above both gates. It reads this as favoring continuation, consistent with Boeing's later comment to the AAIB. The AAIB itself made no such calculation, and it identified loss of directional control as part of why the commander stopped. A groundspeed gate does not address that.

For Perth, the paper suggests a reject may still have been within the V1 energy state early in the stagnation, and that continuing became the energy-consistent choice once groundspeed passed the gate. The ATSB drew no conclusion on that question.

The paper is plain about its limits. It needs validation through flight-data analysis, simulator studies with at least 50 crew pairs, and line trials before anyone adopts it. It assumes valid groundspeed, normal engines and adequate directional control, and it does not cover engine failure. Aircraft whose groundspeed comes from a 1 Hz GNSS receiver rather than an IRS need about 2 kt subtracted. A free calculator implements both methods and says it does not supersede the AFM, FCOM or SOPs.

Regulators are moving on takeoff monitoring, but not on this question. EASA's Opinion No 07/2026, published September 22, 2026, proposes requiring takeoff performance monitoring systems on newly produced large aeroplanes. Its stated targets are erroneous performance data and position errors. As of October 2026, none of the FAA, ICAO, Boeing or Airbus guidance reviewed here gives crews a numerical groundspeed limit for wind shear on the roll.

What it means for pilots

Brief the wind behind your numbers. Know whether the performance assumed nil wind, a reported headwind or a required minimum headwind. At Perth, nil wind was assumed and the actual change was to a 20 kt tailwind. Check windsocks along the runway, not one end; the ATSB notes they can disagree.
Know your V1 margin. Field-length-limited means no spare runway if a reject starts late. When shear is possible, the FAA guide, ICAO and Airbus recommend maximum takeoff thrust and the most suitable (usually longest) runway, or a delay.
Watch groundspeed in gusty or convective conditions. Airbus and the Flight Safety Foundation list groundspeed variations and airspeed changes over 15 kt as cues. The PM can cross-check groundspeed against airspeed on the roll.
Learn what stagnation looks like. Airspeed sits still for seconds while thrust is normal. The speed trend turns down, autothrottle activity may look unusual, and the navigation display wind may swing to a tailwind (the ATSB notes it is imprecise on the ground but shows the sign). At Perth the tape held 134 kt for 3 to 4 seconds; at East Midlands it lost 13 kt in about 1.5 seconds.
Expect no alert. On the 737-800s in both reports, predictive warnings were inhibited from 100 kt to 50 ft and reactive detection began at rotation. Know your own type's inhibits.
Decide before V1. The two-second distance in 25.109 is not decision time. After V1, the published guidance is to continue, and to rotate no later than 2,000 ft from the end if VR is not reached.
Your operator's wind shear procedure prevails. GSV1 is a published proposal, not an approved method. If it interests you, take it to your safety or training department. Don't add a private go/no-go rule to the takeoff.

Sources

This is general information, not operating guidance. Your company's procedures, your aircraft's manuals and the regulations come first.