Description 

On 27 January 2023, a Boeing 777-319ER (ZK-OKN) operated by Air New Zealand on an international scheduled passenger flight from Melbourne to Auckland as NZ124 was mishandled during the flare prior to a day instrument meteorological conditions (IMC) manual touchdown. As a result, it veered off the runway to the right and struck six edge lights before regaining the runway to complete the landing roll and taxiing in to its allocated parking gate where consequent tyre and brake assembly damage was found.  

Investigation 

An investigation was carried out by the New Zealand Transport Safety Investigation Commission (TAIC). Relevant cockpit voice recorder (CVR) and digital flight data recorder (DFDR) data was successfully downloaded from the aircraft and facilitated a meaningful review of the event. Corresponding quick-access recorder (QAR) data was also available.

The 63-year-old captain had been employed by the operator as a pilot for 38 years and had been promoted to captain on type in 2014. He had worked in this position continuously except for almost two years when the operator’s Boeing 777 fleet had been temporarily grounded due to reduced travel during the Covid epidemic. During those two years he had flown the Boeing 787. He had a total of approximately 24,000 hours flying experience which included 5,183 hours on type. The 59-year-old first officer, who was acting as pilot flying (PF) for the flight involved, had been employed by the operator as a pilot for 25 years and had a total of 19,400 hours which included 5,248 hours on type. Like the captain, he had been temporarily transferred to the Boeing 787 fleet during the time the Boeing 777 fleet was grounded during the Covid epidemic. Both operating pilots reported being well rested prior to commencing their two-sector duty involving a flight from Auckland to Melbourne and back. An observing second officer was occupying a supernumerary seat on both sectors, having begun type training three months earlier after previously being employed by one of the operator’s regional turboprop-only subsidiaries where he had flown the Beech 1900D and the ATR 72.

What Happened

It was established that when passing 1,000 feet agl, the aircraft was stable and fully configured for a flap 30° manual landing, and with the autopilot (AP) engaged the flight path was stable. Forward visibility thereafter was reduced in rain of intermittent intensity, although it remained well above the approach minimum visibility of 800 metres.

Recorded data confirmed that as the aircraft passed 500 feet, the runway alignment mode of the autopilot began to apply left aileron and right rudder to ensure the aircraft remained aligned with the runway extended centreline. The slight crosswind component and the wind strength both generally reduced. However, the first officer did not disconnect the AP until a radio altimeter height above the runway of 67 feet as it crossed the threshold, considerably lower than recommended disconnection height. DFDR data showed that when the AP was disconnected, the flight controls had - as expected - returned to their neutral position once the AP centreline tracking input had ceased. When the first officer did not take action to correct this, the aircraft rolled slightly to the right and began to also drift to the right and away from the runway centreline. This occurred despite a reduction in the left crosswind component from about 20 knots to about 5 knots some 15 seconds before AP disconnection.

AP disconnection was accompanied by the thrust levers being retarded to idle and the flare was commenced three seconds later. Four seconds after this, the captain was recorded as saying “you’re left of centreline” (when in fact the aircraft was to the right of the centreline). Main gear touchdown occurred two seconds after this call and was immediately followed by selection of reverse thrust. The captain then said “centreline’s just to the left, come to the left." As the aircraft continued to drift to the right and six seconds after main gear touchdown, nose gear touchdown followed. The first officer was then unable to react in time to stop the aircraft leaving the runway, and the right main gear then ran over six edge lights. The wet runway conditions were assessed not to have caused the excursion outcome.

About twenty seconds after landing, the captain noted a tyre-low-pressure indication for the right main landing gear and advised the tower of potential debris on the runway. The next aircraft on approach was advised of the debris potential and elected to divert. The aircraft then exited the runway via the usual taxiway A3 before proceeding to its allocated parking gate. An inspection of the right main landing gear found that its rear inboard tyre had deflated and a more detailed inspection then found that “five of the six right main gear tyres had been deflated or damaged." Damage to the right landing gear brake assembly, wiring harness and brackets was also found and rectified and all six tyres were replaced.

The runway was closed for almost half an hour for inspection and debris removal before initially being reopened for departures only and then, nearly three hours after the excursion, being fully re-opened.

Why It Happened

It was noted that when the AP is used to fly an instrument approach, runway alignment mode is engaged and the flight controls are moved to achieve alignment, with this action back-fed to the pilot’s controls. If the autopilot is disengaged in this mode, the flight controls will return to the neutral position. To continue the correct flightpath established by the autopilot, a pilot has to hold the flying controls in the same position that the autopilot had been maintaining until the pilot disengaged it. DFDR data showed that when the autopilot was disengaged, the flight controls had immediately returned to their neutral position. This confirmed that the first officer had not held the aileron control inputs applied by the autopilot. As a consequence, the aircraft rolled to the right by almost four degrees and began to drift to the right and off the runway centreline.

Review of the flight data showed that in the 15 seconds before the AP was disengaged, the left crosswind component had reduced from 20 knots to 5 knots and had then fluctuated hardly at all before touchdown occurred. The combination of the initial roll right after disengaging the autopilot and the remaining crosswind from the left could not be countered manually by the first officer in time to prevent the aircraft exiting the runway and striking the six edge lights. The wet runway surface conditions were not assessed to have caused the excursion. It was, however, considered that the first officer’s late disconnection of the AP had probably been influenced by their reported momentary loss of forward visibility as the aircraft passed approximately 165 feet above ground level (AGL). The captain stated that he had retained his forward visibility throughout.

It was noted that the Boeing 777-300 flight crew training manual (FCTM) described the landing techniques applicable to all situations, including crosswind landings and landings on slippery runways. Having stated the need to “begin with a stabilised approach on speed, in trim and on glidepath” this text was found to be immediately followed by a note that stated “when a manual landing is planned from an approach with the autopilot connected, the transition to manual flight should be planned early enough to allow the pilot time to establish airplane control before beginning the flare (and that) the PF should consider disengaging the autopilot and disconnecting the autothrottle (if desired) 1-2 nm before the threshold”.

In the event under investigation, it was assessed that two separate but related factors had contributed to the aeroplane veering off the runway. When the AP was disengaged and the flight controls moved immediately to their neutral position, the first officer “needed to anticipate this to ensure there was no unintended deviation from the intended flightpath." However:

• Since the time between disengagement of the AP and touchdown was only nine seconds, with only four seconds between the disengagement and the commencement of the flare to land, there was insufficient time to restore the left aileron input needed to avoid a roll to the right and the resultant drift right.
• This late AP disengagement also left insufficient time to assess the changing wind conditions and establish manual aircraft control before beginning the flare.

The operator’s Boeing 777 Flight Training Manager advised the investigation that the disengagement manoeuvre in the presence of any crosswind component on final approach was only demonstrated during initial type training and did not form part of recurrent training. This was considered to create a risk that this skill could be lost over time.

The importance of both pilots having equivalent expectations of how to accomplish any landing was also noted as important. However, although the CVR showed that the crew’s approach briefing had not met the intent of the operator’s guidelines regarding AP disengagement, it was considered that as Auckland was the home base for both pilots, this was unlikely to have directly contributed to the excursion.

It was considered that the first officer’s reported momentary loss of visual reference with the runway may have very briefly delayed the autopilot disconnection.

In summary, the primary Findings of the Investigation were as follows:

1. The autopilot was disengaged at 67 feet AGL when the operator’s procedures suggested disengaging the autopilot at 300–600 feet AGL. The late disengagement was likely a consequence of the inclement weather on the approach.
2. When the autopilot was disengaged, the autopilot control inputs (which provided extended runway centreline track alignment) ceased and the controls returned to a neutral position.
3. This resulted in the aircraft rolling to the right and drifting right of the centreline.
4. The aircraft veered off the side of the runway as soon as the main landing gear contacted the runway and collision with six runway edge lights caused damage to the right main landing gear.
5. It was considered “exceptionally unlikely” that the rain had caused contamination on the runway surface leading to loss of directional control.

One Safety Issue in respect of pilot training for AP disengagement was formally identified as follows:

“The procedure and techniques to be used when disengaging the autopilot in a crosswind are taught during the initial conversion course only. Without regular practice, this skill can erode with time, increasing the risk of an incorrect handling technique leading to a runway excursion.”

It was also noted that since the occurrence, Air New Zealand had added an additional event to their operational flight data monitoring (OFDM) detection criteria in order to “capture lateral deviations from the final approach path."

The Final Report was approved for publication on 30 April 2025 and published on 29 May 2025. No Safety Recommendations were made.

Related Articles

• Runway Excursion
• Autopilot
• Cross Wind Landing
• Flight Data Monitoring (FDM)