Description

On 9 October 2018, an ATR 72-200 (YL-RAI) being operated by RAF-AVIA on an international cargo flight from Stettin to Trollhättan, which was a line training sector for the left seat pilot, briefly left the runway during landing in night VMC at destination before regaining it undamaged and completing the landing without further event. The excursion was not reported to anyone by the crew and was not otherwise detected until evidence of it, including damage to a runway edge light, was discovered two days later, prompting further local investigation.

Investigation

Based on the initial findings of the local investigation, an Investigation was carried out by the Swedish Accident Investigation Authority (SHK). The delay to the start of this Investigation meant that all relevant recorded flight data had been overwritten.

It was noted that the flight was a line training flight for a pilot undergoing training for promotion to Captain on type. The 55 year-old Training Captain in command had a total of 9,300 flying hours which included 198 hours on type and was occupying the right seat and designated as PF for the investigated flight. The 49 year-old Captain under training had a total of 9,200 flying hours which included 91 hours (and 42 landings) on type.

It was established that during a routine check two days after the ATR 72 had landed, “a runway edge light outside the left part of the runway was lying on the grass” and subsequently “discovered that the edge light had been run over” and that there were tyre tracks on the runway which continued into the grass alongside it past the position of the edge light before turning back onto the runway. The lateral distance between these tyre tracks was found to be 4.1 metres and since the only aircraft with this gap between the two main landing gear wheels which had recently used the runway was the ATR 72 YL-RAI, further local investigation focussed on that flight. ATC R/T recordings of communications during the aircraft’s final approach were retrieved and showed that the controller had advised the flight of a 17 knot crosswind “on three separate occasions just prior to touchdown”. A report of the findings was sent to the Swedish transport safety regulator.

The aircraft operator was also advised of the occurrence and, having been previously unaware of it, subsequently began an investigation by requesting reports from both of the pilots involved. These were “both […] dated 18 October 2018” and stated that “the landing was carried out without any deviations, with a wind direction of 250°-290° and a wind speed of 16–18 knots”, somewhat different to the record of the observed conditions which gave the mean wind velocity as 230° at 19 knots gusting to 32 knots and the instantaneous wind report of a 17 knot crosswind component provided to the flight by the controller.

A copy of the Operator’s internal investigation report dated 13 December 2018 was submitted to the Latvian civil aviation safety regulator on 18 January 2019 and stated that “there were difficulties keeping the aircraft on the runway centreline due to gusty crosswinds (and that) during the rollout, the ailerons were not deflected against the wind, which contributed to the aircraft briefly coming off the runway to the left, out into the grass and possibly damaging the runway edge light”.

The SHK Investigation interviewed both pilots in February and March 2019. Both confirmed that a crosswind was present during the approach and that the landing took place in the dark on 1710 metre long, 30 metre wide runway 33 (the displaced threshold gives a LDA of 1501 metres. The Training Captain stated that he had been occupying the right seat and acting as PF for the flight under investigation. After touchdown, at around 70 knots, he had passed control to the left seat pilot because only that position has a steering tiller on the aircraft involved. “One of the pilots” also stated that “the ailerons were not sufficiently deflected into the wind in conjunction with the hand-over which caused the aircraft to drift slightly to the left” but added that “the rollout proceeded normally after that". The other pilot stated that he “did not remember anything from the touchdown and rollout”.

The approach and landing of the aircraft was seen by two airport staff, two pilots on the apron and the controller. All said that “the landing appeared wobbly” and the controller also remembered that the touchdown had occurred well past the runway threshold. However, none of those who might have been able to see the excursion recognised that it had happened.

The Investigation reviewed the applicable ATR FCOM in respect of the procedure for crosswind landings and noted the following post touchdown sequence:

• the PF is to reduce the engine power to ground idle.
• the PM is to monitor that both propeller low pitch lights are illuminated and if so calls “2 Low Pitch” after which the PF selects reverse.
• At 70 KIAS, the PM calls “seventy knots” and the left seat pilot keeps or takes over the power levers and the steering tiller and the right seat pilot keeps or takes over the control column and deflects the ailerons into wind.
• The pilots then confirm to one another which of them is responsible for manoeuvring the aircraft. If the pilot in the right seat has been PF, they must make the call “you have control” and receive the response “I have control”. If the pilot in the left seat has been PF, he simply calls “I have control”.
• When the aircraft is landing on narrow runways less than 30 metres wide, the left seat pilot must be prepared to use the steering tiller immediately after the nose landing gear touches the ground, regardless of which pilot is PF.

In respect of the final point, it was noted that in this case, the runway was exactly 30 metres wide so it did not apply to the excursion landing. The ATR FCOM content on crosswind lands was replicated in the aircraft operator’s OM with the exception (not relevant to this Investigation) of a 30 knot crosswind limit rather than the 35 knot one in the ATR FCOM.

It was found that in 2016, in response to a series of runway excursions involving the ATR 72, the manufacturer had issued a Flight Operations Information Message (FOIM) which had highlighted the risks of crosswind landings and stressed the importance of accuracy in deflecting the ailerons into wind and use of the rudder and nosewheel steering. It also “emphasised the importance of moving the control column forward after touchdown so as to increase the effectiveness of nosewheel steering”. However, the aircraft operator “was not aware of the FOIM information on crosswind landings and it had therefore not been incorporated into their ATR 72 OM Part ‘B’".

Findings

• The identification of the excursion landing and the trajectory followed was indisputable based on the hard evidence available:
  • The distance between the main landing gear tyre tracks was 4.1 metres, which was exactly the distance for an ATR-72 and also did not match that distance for any other aircraft which had landed at Trollhättan. The only aircraft of that type which had landed at the airport since the previous runway inspection was the aircraft with registration YL-RAI.
  • The ground clearance of the ATR 72 nose landing gear doors and the underside of its fuselage is approximately 40 cm and the damaged runway edge light was 50 cm high. Taking these measurements and the location of the main gear tyre tracks together it was concluded that “the light was damaged in conjunction with the excursion”.
• The insufficient application of into-wind aileron during the landing was “established” and that “furthermore, it is likely that the control column was not pushed sufficiently forward during the landing roll”.
• Although the fact that the excursion was not reported by the crew could be an indication that the reality in respect of reporting culture at the aircraft operator was not in accordance with relevant directives, it was not possible to clarify why the incident was not reported and nothing was found to indicate that the operator’s approach to incident reporting would deter reporting.

The Cause of the Serious Incident was formally determined as "neither the operator nor the pilots had read ATR’s Flight Operations Information message (FOIM) regarding recommended procedures for crosswind landings and the operator’s pilots lacked full training in crosswind landings on the aircraft type”.

Safety Action taken as a result of the Investigated event whilst the Investigation was in progress was noted as having included the following:

• RAF-AVIA decided to implement “Special Training Measures” in respect of crosswind landings on the ATR 72 and has begun regular downloading of QAR data.
• ATR decided to introduce supplementary information on crosswind landings in the aircraft type FCOM and has also offered to support development of OFDM at RAF-AVIA.
• the Latvian Safety Regulator has been taking, and is continuing, steps to improve the focus on effective risk management and on operational safety generally at RAF-AVIA. Outstanding attention includes but is not limited to:
  • Ensuring that their ‘Flight Safety Programme Manager’ post is filled by a licensed pilot in order to enable permanent and independent work as a safety analyst.
  • Considering the full or partial implementation of OFDM.
  • Establishing, maintaining and subsequently improving a reporting culture.

The Final Report of the Investigation was published on 5 September 2019. In the light of measures already taken and being planned by the aircraft operator, the type certificate holder and the Latvian Safety Regulator, no Safety Recommendations were made.

Related Articles

• Runway Excursion
• Cross Wind Landings
• Ailerons
• Flight Data Monitoring (FDM)
• Safety Management System