On 30 August 2018, a Boeing 747-400F (N415MC) being operated by Atlas Air on an international positioning flight from Dubai Al Maktoum International to Hong Kong as GTI 8086 during which a revalidation line check was being conducted veered abruptly away from the runway centre line shortly after touching down in a crosswind following a day VMC approach and both right side engines were damaged by runway contact before normal control was regained and the taxi in completed. Damage to the inlet cowl, fan cowl, reverser translating cowl drain of both right hand engines was subsequently found such that all these parts needed to be replaced.
An Investigation into the Serious Incident was carried out by the Hong Kong Air Accident Investigation Authority (AAIA) in accordance with Annex 13 principles. Relevant data was recovered from both the CVR and the SSFDR. Relevant airport CCTV and ATC data was also reviewed and of assistance.
The 52 year-old Captain in command of the flight, who was acting as PF and occupying the left hand seat, had a total of 5,223 flying hours experience with Atlas Air, all but 23 of these hours being on type. The 58 year-old Check Captain had a total of 16,000 flying hours experience with Atlas Air of which 7,600 hours were on type. Both of these pilots held FAA-issued ATPLs. A Relief First Officer and an ‘operational experience student’ were occupying the flight deck supernumerary seats. During the cruise part of the eight hour flight, the Check Captain took a rest break but the Captain on line check did not, subsequently stating that he had not needed one as he had been “good to continue flying”.
The Approach Briefing for runway 25R did not include any discussion about which crosswind landing technique would be used. The subsequent crabbed approach was normal with the reported surface wind varying in direction between 150° and 220° and in speed between 17 and 28 knots. Just before touchdown, the Check Captain said “bring the nose over” and, conscious that this was a check flight, the PF Captain “followed the prompt” and tried to de-crab prior to touchdown even though this was a procedure he was not familiar with. Immediately following touchdown, the aircraft veered to the right of the runway centreline before its direction reversed abruptly to the left of the centreline and then sharply banked to the right of the centreline again. This right bank caused a significant drop in the right wing to the extent the underside of both the No.3 and No.4 engines contacted the runway surface and sustained damage. The deviations lasted around 16 seconds, after which the runway centreline was regained with the wings level as deceleration continued below 90 knots and the remainder of the landing roll was normal. At the time this happened, the ATIS broadcast included the caution that “significant windshear and moderate turbulence” could affect runway 25R. The actual 2 minute mean wind velocity at the runway 25R threshold at the time of touchdown was found to have been 182°/17 knots with the “10 minute gust” at that time being 30 knots.
It was noted that as the aircraft was not carrying any payload, its landing weight - just under 190 tonnes - was significantly below the corresponding weight when loaded and the AFM MLW of 285.7 tonnes. It was also noted that the ground clearance of the No.3 engine is between 0.71 metres and 0.93 metres and that of the No.4 engine is between 1.32 metres and 1.8 metres.
FDR data showed that until about 10 seconds before touchdown, the recorded on board computed wind direction was from approximately 200° at between 17 and 20 knots. FDR data also showed that the rate of descent in final approach was maintained at around 900 fpm until just before the flare was initiated, the only significant exception being “a momentary excursion to 330 fpm and 1320 fpm at 450 agl”. The approach was conducted with the aircraft crabbed into the left crosswind at an angle of about 7 degrees.
It was noted that (unless an aircraft operator directs otherwise) the way a crosswind landing is safely achieved is a matter of the pilot performing it to decide with this subject being included in the approach brief, the primary difference being whether the aircraft drift angle is eliminated during the later stages of the flare or after touchdown. It was found that there was no Atlas Air documentation on crosswind landing technique and that this could only be found in the Boeing Flight Crew Training Manual (FCTM) which at Atlas Air was “for reference only and not required at any point during training”. The PF Captain stated that he had “not reviewed it recently” and the Boeing Flight Crew Operations Manual (FCOM) did not contain relevant guidance. During interview he “was able to discuss the various crosswind (approach and landing) techniques but not how they applied in practice or the pros and cons of each”.
FCTM guidance on crosswind approach and landing was found to cover both de-crab in the flare and after touchdown, advising that the latter method could be used up to the crosswind components specified in the FCTM landing crosswind guidelines speed table.
A detailed analysis of the FDR data showed that the relevant aircraft systems had functioned as designed. It showed that “a few seconds into flare the rudder pedal initially deflected upwind (left), then downwind (right), with control wheel deflected abruptly into the wind (left), then varied to the right and again left prior to touchdown” with a 6-degree crab angle at touchdown as right rudder and aileron control inputs were made resulting in a heading 5 degrees to the right of the centreline at a bank angle of just over 3.2°. After two seconds, “substantial left rudder and aileron control inputs were made, seemingly checking the momentary right bank” but these were excessive and the aircraft response was a left turn onto a heading 14 degrees left of the centreline and a left bank angle of approximately 4.6 degrees which was sustained for about 5 seconds. Just before the maximum left turn was reached, “substantial” right rudder and aileron control inputs were made which it was surmised were an attempt to respond to the significant left bank. These inputs resulted in the aircraft turning and banking to the right with the bank angle reaching 5.6 degrees which is when the two engines on the right wing probably contacted the runway.
Taking the available weather information into account, it was considered that “wind direction and speed was not a significant issue to the control of the aircraft during and after landing” and that the veering and rolling of the aircraft “was probably due to a series of incorrect rudder and aileron inputs made after the touchdown”. It was also considered that “the pilot’s input on the controls was about two seconds behind the directional control of the aircraft” with most control inputs being made being to correct the previous directional deviation. It was not confirmed but considered that the Captain’s unfamiliarity with the technique of de-crabbing an aircraft just before touchdown may have been the primary origin of the mishandling episode.
The Cause of the investigated Serious Incident was formally documented as “the combined effects of a sharp right yaw and significant right roll in response to the exaggerated inputs made by the aircraft commander”.
Two Contributory Factors were also identified:
- The Commander made a sudden change of his crosswind landing technique from crab to de-crab on short final because he considered that the Line Check Captain expected a de-crab landing.
- Prior to the top of descent, there was no discussion about crosswind landing technique or what would be expected during landing.
Safety Action taken by Atlas Air whilst the Investigation was in progress and advised to it was noted to have included a “recommendation” to its Flight Operations Department to:
- Develop, for all aircraft types, a ground school module on crosswind takeoffs and landings to be presented to all flight crew at Initial, Transition and Upgrade training prior to full flight simulator work. This module to include preferred techniques, flight control inputs, performance considerations, application of reverse thrust and the effects of the side thrust component if applicable with an initial demonstration and explanation of it to all Instructors and Line Check Captains to ensure their standardisation to correct procedures and techniques during crosswind conditions.
- Revise the existing Full Flight Simulator modules to incorporate a realistic amount of crosswind component for all takeoffs and landings in accordance with the Extended Envelope and Adverse Weather Training of 14 CFR Part 60 with multiple modules so as to include different crosswind component scenarios which should be of varying characteristics and intensities.
- Ensure that Initial, Transition and Upgrade curriculums make reading of the FCTM prior to Full Flight Simulator time compulsory and fleet-specific curriculums are updated as necessary.
One Safety Recommendations was made as a result of the Investigation as follows:
The Final Report of the Investigation was presented to the Hong Kong Government on 3 November 2020 and released online shortly afterwards.