MaxSez: This is an unusual occurrence (PIC/Controller/FldOps Error) worth reviewing In full.
Shaheen B734 at Sharjah on Sep 24th 2015, took off from taxiway
Updated: Tuesday, Apr 25th 2017 15:15Z, (Prelim Incident Rpt)
A Shaheen Boeing 737-400, registration AP-BJR performing flight NL-791 from Sharjah (United Arab Emirates) to Peshawar (Pakistan) with 148 passengers and 8 crew, was assigned to depart Sharjah’s runway 30. The captain (44, ATPL, 4,079 hours total, 1,235 hours on type) was pilot flying, the first officer (39, CPL, 1,049 hours total, 182 hours on type) was pilot monitoring, another pilot (53, ATPL, 3,165 hours total, 388 hours on type) occupied the observer’s seat. The aircraft was cleared to taxi to the holding point runway 30 via taxiway Alpha, Alpha Two Zero, holding point Bravo Two Zero, Runway Tree Zero. The crew read back Taxiway Alpha to holding point runway tree zero.
While taxiing along taxiway Alpha the crew was instructed to switch to tower frequency. Tower instructed the aircraft to hold short of Bravo Two Zero, the crew acknowledged to hold short of runway tree zero. The aircraft reached the end of taxiway Alpha and turned towards Alpha Two Zero. Before the aircraft reached Alpha Two Zero tower cleared the flight for takeoff from runway 30, the aircraft turned onto taxiway B, about 360 meters before the runway and about 200 meters before the CAT II/III holding point and commenced takeoff. The controller instructed the crew to expedite the takeoff as another aircraft was on final approach to runway 30 at a 7nm final. Due to a window frame (construction of tower) the tower controller lost sight of the departing aircraft.
When the tower controller and supervisor realized the aircraft was in fact departing taxiway B instead of runway 30, they decided - unsure how fast the aircraft already was - to permit the takeoff continue as there was no threat on taxiway B from other aircraft, vehicles or other obstacles.
The aircraft continued to Peshawar for a safe landing.
The United Arab Emirates General Civil Aviation Authority (GCAA) released their final report concluding the probable causes of the serious incident were:
The Air Accident Investigation Sector determines that the cause of the Incident was that, most probably, the flight crew did not devote sufficient attention to the taxi route, or taxi route lighting and signage. The flight crew misunderstood the air traffic control instructions and failed to identify that the Aircraft had been aligned on a taxiway, instead of on the runway, resulting in a takeoff from the taxiway.
Contributory factors to the Incident were:
the Aircraft Operator standard operating procedures (SOP) did not require verification by the crew that the aircraft is lined up on the correct runway before commencement of takeoff;
the early takeoff clearance given by ATC when the Aircraft was approximately 200 meters away from runway 30 holding point;
the urgency of the air traffic Controller for the Aircraft to depart;
the red stop bar lights at the CAT II/III holding point for runway 30 was already OFF;
the brighter green lead-on lights for taxiway Bravo were probably mistakenly interpreted as the lead-in lights for the runway
similar numeric descriptors for taxiway and runway designation;
the air traffic Controller lost visual watch on the Aircraft for some time and
the possibility that the flight crew assumed that taxiway Bravo was the runway due the width of the taxiway.
The GCAA complained that they learned about the occurrence only 4 days after the occurrence when the notice by Sharjah ATC arrived at the office. No communication about the incident had taken place between ATC and flight crew.
The GCAA reported that taxiway B had been used as runway 12/30 until 2014 when the new runway went into operation and the old runway was re-assigned as taxiway B.
The GCAA analysed:
Before the Aircraft reached the end of taxiway Alpha, the Copilot was instructed to change from Ground frequency to Tower frequency, which was complied with. The air traffic Controller for Ground and Tower was the same person. The average of the Aircraft taxi speed during this phase was approximately 10 knots.
ATC clearance for takeoff was given as the Aircraft approached taxiway Alpha 20, with the additional request “Without delay.” The air traffic Controller had an aircraft on a seven nautical miles final approach for runway 30. At this stage of the taxi, the Aircraft had not crossed the Alpha 20 taxiway intermediate holding point OFF red stop bar.
Fifteen seconds later, the air traffic Controller requested the Copilot to keep the speed up until clear of the runway.
Again, after another eight seconds, the air traffic Controller repeated the clearance for takeoff with the additional words of “…without delay cleared for takeoff runway three zero…” The Aircraft had just passed the OFF intermediate stop bar on taxiway Alpha 20. As the Commander was already following the taxiway green centerline lights, he would have seen the green lead-on centerline lights to taxiway Bravo on his left side. The CAT I/II holding point red stop bar (at a distance of 110 meters from the Aircraft) had they been ON, would have also been visible to the flight crew.
However, as the Investigation did not have verification of the switch position of the CAT I/II holding point red stop bar for runway 30, the most likely position for the stop bar lights was OFF as the air traffic Controller had issued an expedited take-off clearance to the Copilot.
Four seconds later, the air traffic Controller requested the Copilot to take off “without delay”. The DFDR data indicated that the Aircraft turned away from taxiway Alpha 20, heading on approximately 30 degrees towards taxiway Bravo. The Aircraft speed during the turn had slowed from seven to three knots that indicates a flight crew cautious command inputs along the taxi route.
After another 12 seconds, the air traffic Controller repeated his instructions for the crew to expedite the takeoff. Shortly after, the Commander advanced the thrust levers to part power and as the Aircraft crossed the intersection of taxiway Bravo and taxiway Alpha 18, the thrust levers were at take-off power.
As the Aircraft was not fitted with a runway awareness advisory system (RAAS), the only means available for the crew to determine the Aircraft position on the aerodrome during the taxi and takeoff, was by visual reference to the signage and lighting.
The flight crew, as they stated, had no doubt that the Aircraft took off from runway 30.
With clear visibility during the nighttime departure, the Investigation could not determine why the lack of situation awareness was not regained by the crewmembers after they had lined up on taxiway Bravo. With the Aircraft take-off lights turned on, their cognitive ability failed to recognize that the only visible lights were one row of green centerline lights along the taxiway yellow painted centerline. In addition, even though runway 30 has a displaced threshold, after an aircraft turns towards the runway, the runway white edge lights and white centerline lights would become visible.
With respect to the repeated omissions of Bravo Two Zero by the flight crew and ATC not picking up on the omission in the read backs the GCAA analysed:
As the Aircraft was approaching taxiway Alpha 20 holding point, takeoff clearance was given and the air traffic Controller stated “Shaheen seven niner one, runway three zero, bravo two zero, without delay clear takeoff, surface wind is one three zero degrees, five knots, bye bye.” The Copilot read back the clearance, but again left out Bravo two zero. Soon after, the Commander started to turn the Aircraft towards taxiway Bravo.
A pilot read back presents the first and most efficient opportunity to catch miscommunications. It provides a verification to the controller that the pilot heard and understood the instruction, and it gives an opportunity to the controller to reaffirm the instructions given. An effective read back can mitigate the effects of expectation because it gives the controller an opportunity to correct any error.
It is possible, that the crew mistakenly understood that the holding point at Alpha 20 was the actual runway holding point, as Bravo two zero was never repeated by the Copilot. As the air traffic Controller never informed the crew that the read back was incorrect, this may have confirmed the mistaken perception the crew had. The crew may have developed an erroneous mental model that taxiway Alpha two zero holding point led to the runway. Contributory to this would have been that the red stop bar lights at taxiway Bravo 20 CAT I/II holding point to runway 30 were probably OFF.
The GCAA analysed that a conflict arose between the departing and arriving aircraft:
When the clearance was given for the Aircraft to take off, the aircraft on approach was on a seven nautical miles final. At an approximate approach average speed of 145 knots, this aircraft would have taken approximately 174 seconds to touchdown. At the initial ATC request for takeoff of the Incident Aircraft at 0238:49, until the Aircraft aligned on taxiway Bravo and started the take-off roll at 0239:40, the aircraft on the approach would have had just over a 2-minute separation.
Assuming that the Aircraft had continued along the correct taxi route at an average speed of seven knots, and that it did continue along taxiway Alpha 20, taxiway Bravo 20 to runway 30, it would have taken 104 seconds to travel the approximate 375 meters to align on the runway 30 heading of 300 degrees. At this stage, the approaching aircraft would have been three nautical miles or 70 seconds from the runway. To reach the threshold, the Aircraft would have had to travel an additional approximate distance of 275 meters. Thereafter, if the Aircraft had travelled at 25 knots, it would have taken another 21 seconds to reach the threshold. The air traffic Controller would have had to reassess the situation and decide whether the approaching aircraft had to perform a go-around in order to avoid an unsafe condition.
This was the air traffic Controller’s first night duty, thus, had to work against the natural body clock to modify his sleeping habits in order to sleep during the day. At the time of the departure for the Incident flight, the air traffic Controller had been awake for approximately seven to eight hours and would have been experiencing circadian cycle effects. Similar to the flight crewmembers, the air traffic Controller body would normally experience deep sleep as well as a decrease in natural body temperature at the time of the Incident. As a result, this may have influenced the performance and alertness of the air traffic Controller.
Regarding the appearance of taxiway B the GCAA analysed:
As a safety concern, for takeoff and landing aircraft, the Investigation believes that there is a risk involved with taxiway Bravo as a flight crew can mistake taxiway Bravo for a runway. This is due to the brighter green lead-on centerline lights, the physical size, imprint of the previous runway edge paint, aircraft tire imprints left on the taxiway surface, and that the threshol d was still visible with the entire area having a whitish appearance. It is most likely that if this is not addressed, the taxiway confusion and potential taxiway takeoff will reoccur together with the risk of an aircraft landing on taxiway Bravo.
(SEE: Google: www.faa.gov/go/runwaysafety)
(Data: AvHerald. Grafic: Fair Use/Public Domaine)