An Etihad Airbus A320-200, registration A6-EIT performing flight EY-250 from Abu Dhabi (United Arab Emirates) to Kozhikode (India) with 116 passengers and 8 crew, was on their second ILS approach to Kozhikode's runway 28 when the aircraft crossed the runway threshold aligned with the runway center line, but subsequently drifted to the right, touched down to the right of the runway center line, subsequently the right main gear went beyond the right runway edge line, contacted and damaged five runway edge lights, the crew received a tyre low pressure indication before the crew steered the aircraft back onto the runway center line. There were no injuries, the aircraft received damage to two right hand main tyres, 5 runway edge lights were damaged.
On Sep 9th 2020 the United Arab Emirates GCAA released their final report concluding the probable cause of the incident was:
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.
Contributing Factors to the Incident
The Air Accident Investigation Sector identifies the following contributing factors to the Incident:
(a) The presence of a moderate intensity rain shower over the runway during the landing affected visibility after the Aircraft crossed the threshold.
(b) The lack of runway centerline lighting.
(c) The situational awareness of the Commander, as the pilot flying, was adversely affected by his expectation that the Aircraft would remain aligned with the centerline until touchdown, since the Aircraft was aligned when he overflew the centerline at the threshold. The alignment deviation occurred because of the reduction in visibility over the threshold, and the lack of runway centerline lighting that resulted in a loss of visual references.
(d) 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.
(e) Control inputs to re-align the Aircraft were not affirmative in that only incremental left rudder inputs were made without an associated left roll. 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.
(f) The recovery action to take the Aircraft back to the centerline by applying right rudder input after touchdown, was relatively late due to the high workload. Initially the pilot flying applied an incorrect continuous left rudder input after touchdown.
(g) Despite his awareness of the deviation to the right of the runway centerline, the Co-pilot, as the pilot monitoring, did not intervene to attract the attention of the Commander. This was not in compliance with standard operating procedures.
The GCAA reported the captain (53, ATPL, 15,179 hours total, 12,943 hours on type thereof 8,876 hours in command) was pilot flying, the first officer (27, MPL, 1,670 hours total, 1,510 hours on type) was pilot monitoring.
The captain had initiated a go around on first ILS approach to Kozhikode's runway 28 at about 200 feet AGL due to heavy rain over the aerodrome over concerns the aircraft might hydroplane after touch down.
About 16 minutes later the captain positioned for another ILS approach to runway 28 and continued to touch down about 22 minutes after the first go around. The touch down occurred to the right of the runway centerline, the aircraft's right main gear subsequently contacted five runway edge lights before the captain returned the aircraft to the runway center line.
The GCAA analysed that during the first ILS approach the glideslope deviation as well as the localizer deviation showed pertubations while the aircraft was below 400 feet AGL, though no atmospheric pertubations were observed. The captain however indicated in his post flight interview, the decided to go around over concerns the aircraft might hydroplane after touch down due to the rain reported.
With respect to the second approach and landing the GCAA analysed:
After performing the go-around, instead of following the approach chart of ILS Z Runway 28 for another approach, the flight crew requested an extension of the Aircraft track on the 280 degrees heading until the Aircraft was over the sea due to better weather conditions in that area. The crew planned to wait for the weather over the runway to improve.
When the Aircraft was approximately eight nautical miles outbound from CLC, the flight crew requested a heading of 320 degrees with the intention of preparing for another approach within 10 minutes. The Tower Controller then informed the flight crew that the intensity of the rain had reduced and visibility had improved. The flight crew acknowledged this information and informed the Controller that they would advise their intentions.
Prior to deciding to fly another approach, the flight crew asked the Controller whether heavy rain was still present over the runway. The Tower Controller informed the flight crew that light rain was falling and visibility was 3,000 meters. Based on the Tower Controllers information, the flight crew decided to conduct a second approach since they considered that the intensity of the rain had lessened. However, the visibility remained the same.
Approximately a further 11 minutes after the go-around, EY250 commenced a
second ILS approach to runway 28, as shown in figure 8. The flight crew reported to the Tower Controller as EY250 passed CLC while flying level at 3,600 feet pressure altitude above the runway. The Tower Controller requested the flight crew to report when the Aircraft was established on the localizer.
As the Aircraft was turning towards the IF flying inbound to CLC, the localizer capture mode engaged and the flight reported ETD25B Localizer Runway 28. The Controller then informed EY250 that there was light rain over the airfield, the runway surface condition was wet, and the wind was from 240 degrees at a speed of 4 knots. The Controller cleared EY250 to land on runway 28 and the flight crew read back the clearance correctly.
When the Aircraft had almost passed the IF, approximately 11.1 nautical miles DME, the Tower Controller informed EY250 that the visibility was now 2,000 meters.
As the Aircraft was passing the IF, in a straight and level condition just after turning to the right, the APP mode was engaged. The Aircraft was in an altitude hold mode (ALT) at a selected altitude of 3,100 feet, and the localizer was captured.
The APP mode was activated prior to the Aircraft establishing on the localizer, but the localizer had already been captured. This activation was prior to the sensing of the glideslope indications, which was not in accordance with the Caution as recommended in the Route Information Manual. However, this non-conformity did not affect the approach.
When the Aircraft was at approximately four nautical miles DME while descending through 1,610 feet pressure altitude on the glideslope, the flight crew requested the wind condition. The Controller informed the crew that the wind was from 220 degrees at a speed of 5 knots.
At 1,000 feet above the airfield level altitude (AAL), the stabilization height recommended in IMC by the FCTM and OM Part A, the Aircraft was already configured with full flap (slats/flaps 27/40 degrees), landing gear down, ground spoilers armed and autobrake armed in MED mode. Both autopilots and flight directors were engaged in localizer and glideslope track modes, and the Aircraft was on the correct lateral and vertical flight path in landing configuration. Autothrust was engaged and active in SPEED mode, which means that the thrust was stabilized and the Aircraft was at the target speed for the approach. No excessive flight parameter deviations occurred. Therefore, the final approach of EY250 was stabilized at 1,000 feet AAL in accordance with the FCOM for Stabilization Criteria and the flight parameter deviations to be called out as per the FCOM for Flight Parameters.
As the Aircraft was descending between 1,000 and 730 feet AAL, the target speed was managed and varied between 138 and 139 knots. The indicated airspeed varied between 138 and 140 knots, and the rate of descent varied between 820 and 670 feet per minute. The vertical load factor varied between +0.97 and +1.04G. The pitch angle varied between +1.8 and +2.4 degrees (nose up). No significant drift angle was recorded nor the lateral load factor variation. The roll angle varied between -1.4 degrees (left wing down) and +1.1 degrees (right wing down). The heading varied between 281 and 283 degrees (QFU10 283 degrees). The Aircraft was on the localizer and the glide slope.
The crew disengaged the autopilot when the Aircraft was descending through
approximately 730 feet AAL, at a distance of 1.9 nautical miles DME. The Commander, as the pilot flying, then controlled the Aircraft manually with autothrust engaged in SPEED mode and the speed target was managed. He was aware of the perturbation of the glideslope signal and he therefore used the PAPI lights to maintain the glide path of the Aircraft.
Between 1,000 feet and 170 feet AAL, the average wind recorded was from 264 degrees with a speed of 12 knots. This gives an average headwind component of approximately 11 knots, and an average crosswind component from the left of approximately 3 knots, as shown in figure 9. The Aircraft did not encounter adverse wind conditions during the second approach and landing.
As the Aircraft descended below 170 feet AAL, the recorded wind information could be used to determine a trend. The wind trend information gave an indication that the headwind had changed to a slight tailwind. The Commander did not notice this change. He had last reviewed the wind information when the Aircraft was at approximately 1,500 feet pressure altitude, and the wind was the same as reported by the Tower Controller. After touchdown, the wind information was unreliable.
At 15 feet AAL, the Commander applied left rudder leading to a dynamic motion on the yaw axis. This dynamic motion had a significant effect on the recorded crosswind trend, which from that point could not be considered as trend information.
Between 730 feet (at 2337:55) and 100 feet (at 2338:44), the speed target was managed and varied between 138 and 139 knots. The indicated airspeed varied between 140 and 136 knots. The Commanders sidestick inputs varied from around 1/4 full nose-up to around 1/3 full nose-down. The pitch angle varied between +1.4 and +4.1 degrees (nose up).
The rate of descent varied between approximately 900 and 560 feet per minute. The vertical load factor varied between +0.95G and +1.07G. The roll angle varied between -2.8 degrees (left wing down) and +3.2 degrees (right wing down). The heading varied between 280 and 282 degrees (QFU 283 degrees). The drift angle varied between +1.1 and +2.6 degrees (the Aircraft nose to the left of track).
As the Aircraft descended through 460 feet AAL, the glide slope deviation increased from 1/2 dot below the profile to 4/5 dot above the profile within 11 seconds. At the same time, the rate of descent increased from 700 feet per minute to 800 feet per minute. Similarly, to the first approach, during the second approach, the glide deviation was not consistent with the change of rate of descent. In this case, the glide deviation moved increasingly above the profile as the rate of descent increased. It is suspected that a perturbation of the glide slope signal occurred.
At the same time, the localizer deviation varied between 1/2 dot to the right and 1/2 dot to the left of the localizer with a high frequency. It is suspected that a perturbation of the localizer signal occurred.
The Commander stated that he announced Continue at minima. When the Aircraft was above the threshold at around 50 feet AAL, it was aligned on the runway centerline with negligible drift angle and around 1 degree roll angle. The Commander was able to see the left and right runway edge lights clearly as he had looked outside after the Aircraft passed 100 feet AAL. After the Aircraft passed the threshold, the Commander felt that the rain intensity increased, and it became heavier than light rain. The Aircraft started to drift to the right of the runway centerline.
Between 100 feet and touchdown, the Commanders sidestick inputs varied between 3/4 full nose-up and 1/5 full nose-down deflection. The pitch angle gradually increased from +3 to +5.5 degrees. Several applied nose-up inputs of up to 3/4 full deflection led to a progressive lowering of the rate of descent from 630 feet per minute to 130 feet per minute.
The rate of descent then increased again to 240 feet per minute. On touchdown, the rate of descent was between 120 and 180 feet per minute. Lateral sidestick inputs varied between around 2/5 full right deflection to around 1/4 full left deflection.
The Commander applied several right roll inputs that led to a continuous right roll for approximately 8 seconds after the Aircraft crossed the threshold. The roll angle varied up to +4 degrees (right wing down). The sustained right roll angle led to an increase to the right in the Aircraft track and localizer, which resulted in the Aircraft deviating to the right of the runway centerline.
The Commander was focused more on the Aircraft pitch attitude during the flare, since he was confident that the Aircraft was aligned on the runway centerline when it was above the threshold. The windshield wipers were used and functioned properly on both sides.
However, due to the increase in rain intensity, the Commanders view of the runway edge lights became blurred. As the runway was not equipped with centerline lighting, his ability to judge whether the Aircraft was still aligned on the centerline was affected. The Investigation believes that the Commander applied several roll inputs to the right unintentionally after the Aircraft crossed the threshold, as he stated that he did not feel a wing bank during flare. From the flight data, it revealed that a slight Aircraft roll to the right continuously occurred, as shown in figure 10. Additionally, the slight left crosswind component (3 knots or less) contributed in a minor way to drive the Aircraft to the right side, away from the centerline.
The landing flare was intentionally relatively long, as stated by the Commander, due to the heavy weight of the Aircraft, which was approximately 64,050 kg, and the up-slope of the runway.
During the flare, as the Aircraft was at 20 feet radio altitude, the thrust levers were set to idle. Since the autopilots had been disengaged the landing was a manual landing, and therefore, the flight guidance system generated a RETARD callout at 20 feet radio altitude as a reminder. The Commander was aware of the required action and he retarded the thrust levers to idle at the same time as the RETARD callout was generated.
At about 15 feet AAL, the Commander realized that the Aircraft was on the right side of the runway as he observed that the left hand edge lights became more blurred than the lights on the right hand side. This was one second after he retarded the thrust levers to idle.
The Commander progressively applied a left rudder pedal input of up to 2/3 full deflection. The Commander tried to recover the Aircraft to the centerline of the runway. The left rudder pedal input led an increase in the drift angle from 1 to 8 degrees (Aircraft nose to the left of track) and a heading change from 282 to 277 degrees. The Commanders control inputs resulted in no significant change to the aircraft trajectory to cancel the lateral deviation.
The Investigation believes that when the Commander realized that the Aircraft had drifted away to the right of the runway centerline, he should have also used left roll control inputs within the recommended limitation to recover the Aircraft to the centerline.
In this case, the use of left rudder only without left lateral input, as the Aircraft lateral flight path had already drifted away from the runway centerline, did not allow for an easy realignment back to the centerline before touchdown. The optimum technique is explained in the Flight Crew Techniques Manual (FCTM).
From his seat position, the Co-pilot, as the pilot monitoring, stated that he was aware, despite the increase in rain intensity, that the Aircraft was tracking to the right of the runway centerline after passing the threshold. However, he did not advise the Commander or attempt to draw the attention of the Commander to the situation. According to the SOP contained in the FCTM, if, for any reason, one flight parameter deviates from stabilized conditions, the pilot monitoring will make a callout. The Investigation recommends that the Operator 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.
There were localizer fluctuations before the vertical FLARE mode was displayed on the FMA (flare mode engagement). However, after the flare mode was displayed on the FMA and the Aircraft crossed over the threshold, there was a localizer trend showing that the Aircraft had deviated to the right of the runway centerline.
The Aircraft touched down approximately 910 meters beyond the threshold of runway 28, which was slightly outside the touchdown zone (900 meters from the threshold).
According to the SOP in the FCTM regarding approach and landing techniques, if a normal touchdown distance is not possible, a go-around should be performed.
On touchdown, the right main landing gear touched down first followed by the left main landing gear with a vertical load factor of 1.3G and the localizer deviation was approximately 1/3 dot to the right of the runway. The Aircraft pitch angle was 5.6 degrees up.
The ground spoilers then started to extend. The Aircraft had a 1 degree roll angle of right wing down, 8 degrees drift angle with the nose towards the left of track and a 0.25G lateral load factor with a heading of 277 degrees. The airspeed was 130 knots and the rate of descent was approximately 180 feet per minute.
The Aircraft lateral position was computed at touchdown based on the localizer deviation. Considering the Aircraft distance from the threshold of runway 28 (around 910 meters), the runway length, and the Aircraft localizer antenna position (located under the radome), the Aircraft lateral position was approximately 14.5 meters to the right of the runway centerline.
Based on the drift angle and Aircraft geometry, the right main landing gear position was approximately 20.5 meters on the right of the runway centerline. The left main landing gear position was approximately 13 meters to the right of the runway centerline. With a runway width of 45 meters, the side strip marking (runway edge line) was at 22.5 meters to the right of the runway centerline.
This means that the localizer antenna (at the cockpit level) was approximately 8 meters from the edge line, and with the left drift angle of 8 degrees at touchdown, the right main landing gear outboard wheel was close to the right side stripe marking of the runway and approximately two meters from the runway edge line (runway shoulder), inside the runway.
After touchdown, the pitch-up input was released and a pitch-down input up to around 1/3 full deflection was applied by the Commander. This action led to a decrease in pitch angle towards zero degree, and the nose landing gear touched down approximately one second after the main landing gear touchdown. The ground spoilers fully extended and medium autobrake activated. Approximately half a second after touchdown, maximum reverse (MAX REV) thrust was applied for around 15 seconds.
Left rudder pedal input was applied when the Aircraft was at approximately 15 feet AAL, and continued after the touchdown, leading the heading to decrease to 272 degrees, and the drift angle to the left (nose towards the left of the track)) increased to 11 degrees. Due to the continued left rudder pedal input, the lateral (localizer) deviation reached around 1/2 dot maximum, approximately four seconds after touchdown.
The runway condition was wet, consistent with the moderate/heavy rain that occurred during the landing and as reported in the METAR. As the runway was wet, and given the relative lateral movement and significant drift angle of the Aircraft at touchdown, a lateral runway excursion occurred.
Multiple load factor peaks were recorded as the right main landing gear struck the edge lights. The first edge light was struck less than one second after touchdown. Five edge lights were struck within the next 1.5 seconds.
The brake pressure of the right main landing gear inboard wheel remained at 200 psi, whereas the brake pressure of the other wheels increased progressively up to 2,300 psi.
Therefore, the right main landing gear inboard wheel tire was, most probably, damaged due to contact with the edge lights. The right main landing gear outboard wheel tire was found to have significant wear and partial ply separation, most likely, due to the higher load carried as a consequence of the Aircraft lateral movement on the ground over at least five seconds.
Based on the post flight report (Appendix 1), it states that at 23:39, the WHEEL TYRE LO PR alert, both CHECK TIRES 3 PRESS2651GM and CHECK TIRE 3 PRESS 2651/2652GM failure messages were activated. These messages were in line with the right main landing gear inboard wheel tire behavior recorded by the FDR and the reported damage to the right main landing gear inboard and outboard wheel tires in the technical log book.
There is a shoulder of approximately 7.5 meters between runway side stripe marking (edge line) and unpaved surface. The edge lights are placed around 2.5 meters from the side stripe marking. Based on the wheel geometry, the maximum lateral runway excursion of the right main landing gear was approximately 3.2 meters beyond the runway edge line. In that case, the maximum lateral runway excursion of the right main landing gear outer wheel tire was approximately 3.7 meters beyond the runway edge line, or around half the width of the shoulder.
The localizer deviation started to reduce toward zero when rudder pedal input to the right of up to approximately 1/3 of full deflection was applied. The rudder input started five seconds after touchdown. The Aircraft progressively tracked towards the centerline and realigned with the runway heading. An increase in the heading from 272 to 285 degrees occurred, and the drift angle decreased from 11 degrees towards zero.
The Aircraft was recovered parallel to the runway centerline (slightly to the left of the centerline) approximately 1,500 meters beyond the threshold of runway 28 at a groundspeed of approximately 83 knots (computed airspeed was about 72 knots).
Manual braking was applied 22 seconds after touchdown when the groundspeed
was 19 knots. The application of manual braking led to disengagement of the autobrake. The Aircraft then continued to decelerate and exited the runway via taxiway Bravo. The Aircraft taxied to the stand uneventfully.
The Aircraft was taxied at a maximum speed of seven knots and 30-degree nose wheel steering angle, as per the FCOM for taxi with deflated or damaged tires, to the parking stand. After the Aircraft stopped on the parking stand, the flight crew reported the condition of the right main landing gear tires to the Tower Controller.
The Investigation believes that the Commander expected that the Aircraft would have remained aligned on the centerline (expectation bias11) during the flare, since the Aircraft had already been aligned on the centerline as it overflew the threshold. His expectation bias was reinforced by the reduction in visibility over the threshold due to the increase in rain intensity, and the lack of runway centerline lighting. The Investigation believes that the reduction in visibility and the minimum runway lighting adversely affected the Commanders situational awareness. Since the Commander gave more attention to the Aircraft pitch angle during the flare, the right roll input was, most probably, unintentionally applied due to a reduction in his situational awareness. The application of roll input to the right was believed to be a subconscious action.
The Commander started to realize that the Aircraft was on the right side of the runway due to his vision of the left edge lights becoming more blurred in comparison to his visibility of the lights on the right hand side.
The Commander stated that few seconds after he decided to continue the landing he realized that the Aircraft was drifting away to the right of the centerline. He reduced the thrust levers to idle, and within one second, the Commander took actions by applying continuous left rudder pedal input in an attempt to bring the Aircraft to the centerline. He applied maximum reverse thrust after touchdown. The Investigation believes that the nature of the Commanders incomplete control input can be explained by the high workload when applying the left rudder pedal input, and most probably, the surprise effect of the unexpected lateral deviation of the Aircraft position, such that he overlooked the need to combine roll control input to the left before touchdown.
The Commander was aware of the need to immediately apply reverse thrust after touchdown. Within five seconds after touchdown, the Commander still applied left rudder pedal input. Thereafter, he applied right rudder pedal input to recover taking the Aircraft towards the centerline. The recovery action was relatively late, which was most probably due to the existing high workload, including to understand the Aircraft state, in the 5-second period after touchdown.
The Aircraft touched down slightly beyond the touchdown zone, which is believed to have occurred due to the decrease in visibility and the Aircraft drift to the right, which surprised the Commander. Additionally, the Commander was not aware of the trend of the wind change from a 6-knot headwind component to a slight tailwind component.
Since the CVR recording of the Incident flight was not available to the investigation, it was not possible to determine whether all required briefings, checklists, tasks sharing, and verbal announcements (including system callouts), took place.