On 20 June 2019, an Airbus A320 (A6-EIT) being operated by Etihad Airways on a scheduled international passenger flight from Abu Dhabi to Calicut as EY250 and making its second approach in night IMC departed the right side of the runway briefly during the landing resulting in damage to two of the main gear tyres and five runway edge lights, before the runway was regained and the remainder of the landing roll and taxi in were completed.
Following delegation to the UAE GCAA Air Accident Investigation Sector (AAIS), the State of Registry and of the Operator, by the Indian Air Accident Investigation Bureau (AAIB), an Investigation was commenced. Relevant DFDR data were downloaded and of assistance but relevant data from the CVR was found to have been overwritten after the First Officer carried out the normal pre flight checks, which include switching on the CVR, for what he initially understood would be a return flight to Abu Dhabi after the right main gear wheels had been changed. However, after the damaged runway edge lights were discovered, the flight was cancelled by which time data from the approach and landing had been overwritten. A transcript of recorded ATC data was also examined.
The Flight Crew
It was found that the 53 year-old Captain, who was acting as PF for the flight, had a total of 15,179 flying hours of which 12,943 hours had been on type including 8,876 hours in command. The 27 year-old First Officer had a total of 1,670 flying hours of which all but 160 hours were on type and held an MPL.
Following cruise at FL350, a descent was commenced and clearance for a procedural Cat I ILS ‘Z’ approach to the 2,680 metre-long runway 28 at Calicut was given. This approach required the flight to route to the on-airport VOR/DME then to descend outbound for 500 feet before turning inbound at 12nm from the beacon to establish on the ILS LOC. The Captain subsequently stated that having observed as the overhead was approached that there were no CB clouds over the runway, he considered that it was reasonable to commence the procedure. The TWR controller did mention around this time that there was moderate rain at the airport and that the runway was wet.
Prior to the turn inbound, the controller advised that visibility was 3,000 metres in rain and, on request, that the surface wind was from 270° at 10 knots and that the rain had become heavy. It was noted from FDR data that APP mode was engaged after the LOC was captured but before correct sensing of the GS indications which was contrary to a recommendation given for approaches to this Cat ‘B’ airport in the specifically-applicable operational documentation for approaches there. Given the possibility that the reported heavy rain might lead to enough surface water to induce aquaplaning, the Captain stated that he had considered the possibility of a go-around but had decided to make the decision either way on reaching the DA. By 1000 feet agl, the aircraft was fully configured with medium autobrake set.
The Captain stated that from when he first saw the runway with just under 2nm to go, his observation of rain over the runway convinced him that there was indeed a risk of aquaplaning and he decided that “it was unsafe to land” and commenced a go around at DA. It was noted that the operator’s brief for Calicut specifically included a need to consider the possibility of standing water on the runway during periods of heavy rain. It was also noted that FDR evidence strongly suggested that below 400 feet agl, an obvious but not severe temporary corruption of the both the GS and LOC signals had occurred which had no material effect. It was noted that whilst the transient distortion of the GS signal was well known and was included in the operator’s cautions for Calicut, transient distortion of the LOC signal was not. The Captain stated that by the time the ILS GS signal distortions occurred, he was using the PAPI to control the vertical profile of the aircraft and visually maintaining the runway extended centreline.
After performing the go-around, the Captain had been prepared to wait for the heavy rain to stop but in fact, it did so quite quickly and the visibility increased too so there was little delay before the second approach using the same procedure from the overhead was commenced.
The APP mode was again selected prior to the aircraft establishing on the LOC and prior to checking for the correct sensing indication for the GS but had no relevant consequence. A wind check with 4nm to go was given as from 220° at 5 knots. Passing 1,000 feet agl, the aircraft was fully configured for landing as on the first approach. This time, the AP was disconnected passing approximately 730 feet agl and with the runway in sight, the Captain again used the PAPI to maintain the vertical profile. Thereafter, the same ILS signal distortion occurred below 400 feet agl as on the first approach. At DA, the Captain called “Continue” and as the aircraft crossed the runway threshold at approximately 50 feet agl, the aircraft “was aligned on the runway centreline with a negligible drift angle and an approximate 1° roll angle”.
With the APs now disconnected, the Captain reported that immediately after the aircraft had crossed the threshold, he had felt that the light rain had intensified somewhat. FDR data showed that the aircraft then began to drift to the right of the runway centreline. Touchdown followed at a rate of descent of between 120 and 180 fpm. FDR data showed that the Captain had made lateral sidestick inputs varying between roughly 2/5 of full right deflection to roughly 1/4 of full left deflection just before touchdown but a succession of what was considered likely to have been unintentional right-only roll inputs had then followed. This produced a continuous right roll for approximately 8 seconds after the aircraft had crossed the threshold and resulted in it deviating to the right of the runway centreline. The Captain commented that he had focused more on the pitch attitude during the flare, since he believed that the aircraft was still on the runway centreline. The windshield wipers were used and functioned normally, but due to the increasing rain intensity, the Captain reported that his view of the runway edge lights had “become blurred”. A very slight (3 knot) crosswind component from the left was considered in a small way to assist the drift away from the centreline.
The landing flare was relatively long and when still at 15 feet agl, the Captain reported having noticed that the aircraft was now “on the right side of the runway” because he could see the left hand side runway edge lights had become “more blurred than those on the right hand side”. He reported having applied a significant left rudder pedal input to try and regain the centreline but in the absence of a complimentary left roll input, the desired correction was not achieved. Touchdown occurred at a recorded 180 fpm rate of descent with the right main gear leading just beyond the 900 metre TDZ. Calculations showed that at touchdown, the right main gear was approximately two metres from the runway edge just inside the 45 metre wide runway.
There was standing water on the runway consistent with the moderate/heavy rain that had occurred earlier and again during the landing and “given the relative lateral movement and significant drift angle of the aircraft at touchdown, a lateral runway excursion occurred” onto the 7.5 metre paved runway shoulder. As the right main landing gear struck a series of runway edge lights located at 2.5 metre intervals, “multiple load factor peaks were recorded” and after the first light was struck almost immediately after touchdown, four more lights were struck within just over two seconds. Since the recorded brake pressure of the right main landing gear inboard wheel remained at 200 whilst the recorded pressure for the other three wheels increased progressively up to 2,300 psi, it was likely that the right inboard wheel tyre was damaged by contact with the edge lights. The right outboard wheel tyre was subsequently found to have abnormal wear and partial ply separation, presumed to be a consequence of the excess load carried as a result of damage to the inner wheel tyre. It was subsequently calculated that the maximum lateral runway excursion of the right main landing gear outer wheel tyre was around half the width of the shoulder.
Five seconds after touchdown, left rudder pedal was applied to approximately one third of full deflection and the aircraft began to track towards the centreline recovering to align parallel and just to the left of it approximately 1,500 metres beyond the runway threshold by which time the groundspeed was approximately 83 knots, a comparison with the recorded airspeed showing an 11 knot tailwind component. Deceleration continued and the aircraft was returned to the runway centreline before exiting the runway on taxiway Bravo. It was then taxied to its designated parking gate in accordance with the applicable speed and turning angle restrictions for taxi with deflated or damaged tyres. Once stopped at the gate, maintenance personnel were called to inspect the landing gear. Their inspection confirmed damage to the right main landing gear tyres consistent with the ECAM post flight maintenance report and the TWR controller was informed accordingly.
The aircraft ground track over and on the runway. [Reproduced from the Official Report]
Why the Excursion Happened
The Investigation considered that the Captain’s assumption that the aircraft was continuing to track the runway centreline after being on it when crossing the runway threshold was attributable to expectation bias which had been reinforced by the reduction in visibility over the threshold as the rainfall intensity increased. Also, the absence of centreline lighting meant that what would have been an important alignment cue had been missing.
It was noted that in the absence of relevant CVR data, it had not been possible to determine whether all required briefings, checklists, task sharing, and verbal announcements (including system callouts), had taken place. However, the First Officer did state that although he had been aware, despite the increasing rain intensity, that the aircraft had begun drifting to the right of the centreline after passing the threshold, he had not alerted the Captain to the situation, inaction which was noted to be directly contrary to the declared responsibilities of a PM.
More generally, it was noted that the Flight Crew Operating Manual (FCOM) requirement for flight below DA and until touchdown is that it must be primarily made using visual references and if these are not sufficient, a go-around should be initiated, a requirement discussed in more detail in the Flight Crew Training Manual (FCTM).
It was noted that the excursion had occurred at night at a time which, from a crew local time perspective was within the recognised window of circadian low which is widely recognised as a time of day when human performance efficiency is degraded and the risk of inaccuracy when attempting a challenging or demanding task is increased.
The Cause of the Accident was formally determined as follows:
After crossing the threshold, the aircraft drifted towards the right side of the runway due to a slight but continuous roll input to the right. The aircraft touched down almost at the runway edge line, and this was followed by an increase in lateral deviation towards the runway edge due to an ineffective flight control recovery technique. The aircraft then struck and damaged five runway edge lights as the right main landing gear entered the runway shoulder.
A total of 7 Contributory Factors were also identified as follows:
- The presence of a moderate intensity rain shower over the runway during the landing affected visibility after the aircraft crossed the threshold.
- The lack of runway centreline lighting.
- The situational awareness of the Captain, as the pilot flying, was adversely affected by his expectation that the aircraft would remain aligned with the centreline until touchdown, since the aircraft was aligned when he overflew the centreline at the threshold. The alignment deviation occurred because of the reduction in visibility over the threshold and the lack of runway centreline lighting that resulted in a loss of visual references.
- The several unintentional roll inputs to the right applied due to a subconscious action, since the pilot flying focused more on the aircraft pitch attitude during the flare and the reduction in pilot flying situational awareness.
- Control inputs to re-align the aircraft prior to touchdown were not affirmative in that only incremental left rudder inputs were made without an associated left roll input. A continuous increase in the aircraft lateral deviation movement was a result of the ineffective flight control technique, and this was due to the existing high workload and the surprise effect of the unexpected aircraft lateral deviation position such that the pilot flying overlooked the approved flight control technique before touchdown.
- The recovery action to take the aircraft back to the centreline by applying left rudder input after touchdown, was relatively late due to the high workload.
- Despite his awareness of the deviation to the right of the runway centreline, the First Officer, as the pilot monitoring, did not intervene to attract the attention of the Captain. This was not in compliance with standard operating procedures.
Safety Action taken by Etihad Airways as a result of this event based on their internal evaluation was noted to have included amendment of and additions to the Route Information Manual in respect of operations to Calicut:
- The existing ‘Caution’ in the ‘Airport Briefs’ section in respect of known ILS GS fluctuations was amended and highlighted in red and now requires that crews shall only arm and capture LOC mode when established on a final intercept heading to the inbound course and only when established on the LOC and after ensuring the correct sensing of the GS indications should the APP mode be armed.
- Four text additions were made to the ‘Airport Briefs’ section:
- Autopilot use for ILS approaches below 400 feet above aerodrome level is not recommended.
- Flight Director guidance below 400 feet above aerodrome level is to be used with extreme caution.
- Tower-reported surface winds are inaccurate and unreliable.
- In all cases, pilots must aim to touchdown within the touchdown zone, or a go around must be initiated.
- An addition to the ‘Destination Airport Categories’ section was made to require that only Captains shall perform landings at Calicut unless they are Training Captains when they may allow the First Officer to perform the landing.
A total of 6 Safety Recommendations were issued as a result of the Investigation as follows:
- that Etihad Airways enhance crew resource management training by placing particular emphasis on the need for assertiveness of callout(s) to be made by the pilot monitoring, particularly when there is any deviation from flight parameters. [SR69/2020]
- that Etihad Airways should emphasise during training the importance of the ‘Approach Using LOC G/S Guidance on Final Approach’ SOP as per the Flight Crew Operating Manual (FCOM) and the ‘Considerations About Go-around’ SOP as per the Flight Crew Techniques Manual (FCTM). [SR70/2020]
- that Etihad Airways include information about the possibility of localiser fluctuation in the ‘Route Information Manual’ to assist in maintaining pilot situational awareness. [SR71/2020]
- that Etihad Airways should carefully examine Calicut operations including arrival time and devise appropriate safety measures to mitigate the associated risks, especially during the Monsoon season. [SR72/2020]
- that Etihad Airways re-inforce amongst its pilot body the requirement to preserve flight data and information recordings, which also includes developing a policy for pulling the CVR circuit breaker after any safety incident and for it not to be turned ON again until it is appropriate to do so. [SR73/2020]
- that the Calicut Airport Authority carefully examine aerodrome operations and devise appropriate safety measures, or consider the practicability of the improvement of the airport infrastructure, in order to mitigate the associated risks. [SR74/2020]
The Final Report was issued on 4 August 2020.