B737 / B763, Austin USA, 2023

B737 / B763, Austin USA, 2023

Summary

On 4 February 2023, a Boeing 767-300 which had initiated a night go-around in freezing fog passed within 115 feet of a Boeing 737-700 taking off from the same runway. The aircraft had been respectively cleared to land and takeoff by the same controller but on sighting the 737 on the runway ahead, the 767 crew deemed a go around was necessary to avoid a ground collision. The experienced controller’s poor judgement was attributed to a combination of expectation bias, lack of awareness of a predictable delay to the 737’s departure and absence of any training in low visibility operations.

Event Details
When
04/02/2023
Event Type
HF, LOS, RI, WX
Day/Night
Night
Flight Conditions
On Ground - Low Visibility
Flight Details
Type of Flight
Public Transport (Passenger)
Intended Destination
Take-off Commenced
Yes
Flight Airborne
No
Flight Completed
Yes
Phase of Flight
Take Off
Flight Details
Operator
Type of Flight
Public Transport (Cargo)
Take-off Commenced
Yes
Flight Airborne
Yes
Flight Completed
Yes
Phase of Flight
Missed Approach
Location - Airport
Airport
General
Tag(s)
Extra flight crew (no training), CVR overwritten
HF
Tag(s)
ATC clearance error, Procedural non compliance
LOS
Tag(s)
Required Separation not maintained
RI
Tag(s)
Visual Response to Conflict
WX
Tag(s)
Fog
Outcome
Damage or injury
No
Non-aircraft damage
No
Non-occupant Casualties
No
Off Airport Landing
No
Ditching
No
Causal Factor Group(s)
Group(s)
Air Traffic Management
Safety Recommendation(s)
Group(s)
Aircraft Operation
Aircraft Airworthiness
Air Traffic Management
Investigation Type
Type
Independent

Description

On 4 February 2023, a Boeing 767-300 (N297FE) being operated by Federal Express on a scheduled domestic cargo flight from Memphis to Austin as FXE 1432 self-initiated a very low go-around at night in freezing fog after a late sighting of a Boeing 737-700 (N7827A) being operated by Southwest Airlines on a scheduled international passenger flight from Austin to Cancun as SWA 708 on the same runway at night in freezing fog. The 767 subsequently passed 150 to 170 feet above the 737 as it continued to accelerate through 100 knots for takeoff. As it passed over the upwind end of the runway, the 767 turned left to begin repositioning for another approach and this was completed without further event as the 737 continued to its intended destination. The controller did not cancel the 737s takeoff clearance or instruct the 767 to go around.

Investigation 

The NTSB carried out a comprehensive Investigation into the conflict using relevant data successfully downloaded from the FDRs of both aircraft. Relevant data from both 2 hour CVRs was overwritten during the subsequent flights of both aircraft. This was avoidable in the case of the 767s shorter subsequent flight back to Memphis but the engines were kept at ground idle during the 75 minute turnround and the CVR was not isolated so had exceeded 2 hours recording time when the aircraft reached Memphis. The available FDR data enabled the conflicting flight paths to be modelled using a TCAS-based simulation but as the minimum separation occurred when the TCAS RA capability of the two aircraft were respectively inhibited by design (767) or inactive by design on the ground (737) these advisories did not aid the relative situational awareness of either flight crew. Third party ADS-B data were used to establish GPS positions for all aircraft locations during the investigated conflict as the FDR coordinates recorded on both aircraft were not accurate enough to derive these.

The 54 year-old 737 Captain had a total of 20,600 hours flying experience including 6,000 hours on type of which all but 500 hours had been in command. He was based at Houston and had originally flown for AirTran Airways, which was taken over by Southwest in 2011 and subsequently closed down. The 41 year-old 737 First Officer had recorded a total of 8,907 hours flying experience including 4,983 hours on the 737. He had been employed by Southwest for 6½ years and was also based in Houston. Both pilots considered themselves to be adequately rested on commencing their first flight of their duty after a nightstop at Austin.

The 56 year-old 767 Captain had a total of 23,500 hours flying experience including 1,324 hours on type, all as a Captain. He was based at Memphis and had been employed by Fed Ex for over 20 years. The 49 year-old First Officer was recorded as having a total of 8,000 hours flying experience including 3,500 hours on type with 465 hours completed on type since joining FedEx. He was also based at Memphis. An off duty company pilot was occupying the flight deck supernumerary seat.

The 43 year-old TWR controller in position at the time of the conflict had been employed by the FAA as a controller at several other facilities for nine years before transferring to Austin in 2019. At the time of the conflict, he was the only controller in position in the TWR. He had limited experience of Low Visibility Conditions such as those prevailing during the conflict and noted that with no surface radar at Austin, pilot position reports were all that was available for an aircraft taxiing in poor visibility as had been the case.

What Happened  

The 767 was the first arrival of the day at Austin and, given the reported weather conditions, had briefed for and were flying an ILS Cat 3 approach to their assigned landing runway 18L with the Captain acting as PF. The TWR controller passed the concurrent RVR figures as (the equivalent of) 427 metres, 183 metres and 549 metres with the landing clearance. Some 2½ minutes later, FDR data showed that the landing gear was selected down. Two minutes after this, the 737 called “we’re short of one eight left” and “we’re ready.” At this time, the 737 was reaching the end of the parallel taxiway and had about 170 metres to run until reaching the full length holding point for runway 18L with the 767 just inside 4 nm from the threshold. The controller was unable to see either aircraft because of the poor visibility. A few seconds later, the controller cleared the 737 for takeoff with the concurrent RVR figures (the equivalent of) 365metres, 183 metres and 487 metres and advice that the 767 was at less than 3 nm final.

As the 737 was crossing the holding point with the First Officer as PF, the 767 asked the controller to comfirm that they were cleared to land as (accordng to the Captain’s subsequent statement) they were concerned at the apparent position of the 737. The controller replied that the 737 would be “departing prior to your arrival”. FDR data showed that the 737 was turning onto the runway 18L as the767 was less than 2 nm from the runway at 650 feet agl. Once on the runway, the 737 came to a complete stop with the 767 only 1.3 nm away and remained so for a recorded 19 seconds whilst the First Officer performed the static engine run-up required due to the freezing fog conditions before releasing the brakes to begin the takeoff roll.

With the 767 on a 0.7 nm final approach, the controller called the 737 to confirm that the airplane was “on the roll” and received the response “rollin’ now” but ADS-B data showed that at the time of this response by the PM 737 Captain, the aircraft was “still at the runway 18L threshold and movement only commenced 4 seconds later". The fog prevented the controller from seeing where the 737 was or what it was doing but the 767 First Officer subsequently stated that they had been at 300 feet agl when the 737 Captain’s transmission was heard with, according to the 767 Captain the forward visibility zero. The controller reported that he had been able to see a light on the 767 whilst it was on approach but could not see the 737 or the aproach end of the runway due to the “dense ground fog”.

During a subsequent interview, the 767 First Officer stated that at about 150 feet agl he had seen a white nav light followed by the ‘silhouette’ of the 737 and made a go-around call. The Captain, who said he had “not seen anything but fog”, responded by immediately initiating the missed approach procedure as the aircraft crossed the runway threshold. The First Officer then almost immediately transmitted “Southwest abort” and added “FedEx is on the go”. Ten seconds later, the controller instructed the 737 to “turn right when able” but received the response “negative”. The 737 Captain subsequently stated that he had heard the 767’s go-around call but not the “abort” call which followed it whereas the 737 First Officer stated that he had heard both calls. The 737 was rotated and was airborne three seconds later. Once both aircraft were airborne, both crews noticed a TCAS TA showing a vertical separation of 6-700 feet and the 737 PF First Officer had reduced the rate of climb. It was noted that this, together with the 767 climb rate, had eliminated the possibility of a collision with both aircraft airborne. The illustration below shows the relative positions of both aircraft and as the 767 flew approached and flew overhead the on-ground 737 after which vertical separation increased.

B737&B763-Austin-2023-flight-paths

The flight paths of the two aircraft as the 767 overflew the 737 with a tail to lower fuselage separation of around150 feet. [Reproduced from the Official Report]

Why It Happened

The conflict occurred at a height where both TCAS RAs and TAs are inhibited and at an airport which had two runways but no airport surface traffic detection system such as ASDE-X but where CAT 3 approaches were possible. This required defensive controlling in conditions which the controller said he had only experienced “a handful of times”, none of which had been during the year preceding the conflict under investigation. There was also no evidence that any controller training for low visibility operations had ever been provided at Austin. The FAA Order on separation between an arriving and a departing aircraft using the same runway in any visibility conditions was found to state that the minimum separation between an aircraft “on its takeoff roll” and an aircraft on approach was 2 miles (1.74 nm). This separation minimum was actually reached when 767 was at 2 miles and the 737 was just entering the runway before then stopping on it for 19 seconds before beginning the takeoff. By the time the 737 commenced takeoff, the two aircraft were only about 1,130 metres apart. However, it was noted that “the elapsed time between the 737 takeoff clearance being given and the start of its takeoff roll was only 2 seconds more than the average time for this over the nine previous departures that morning”.

Amongst a series of documented “Findings” from the Investigation were the following:

  • Pilot and Controller qualifications, Control Tower Staffing and pilot or controller fatigue played no part in the conflict. 
  • The controller’s inaccurate mental model of the position of the 737 on the taxiway resulted from a combination of his expectation that the aircraft would be at the runway 18L holding when its crew requested takeoff clearance and his failure to verify the aircraft’s position on the taxiway at that time. 
  • The quick reaction of the 767 First Officer on seeing the 737 and the quick response of the 767 Captain in commencing a missed approach avoided a potential runway collision and led to the successful resolution of the loss of separation. 
  • The 737 flight crew’s action to shallow their climb after takeoff in response to the TCAS TA displayed contributed to the increased separation between the two aircraft. 
  • Even though the 737 flight crew were not required to inform the controller about their plan to stop the airplane once on the active runway, it would have been prudent for them to do so given that they were aware of traffic on short final. 
  • The 737 flight crew should have informed the controller that they would need to perform an engine run-up before beginning takeoff.
  • In respect of the ability to conduct a comprehensive investigation, it was noted that CVRs with a 25-hour recording capability are necessary to prevent valuable information being overwritten on those which are designed to record only 2 hours of data.

The Probable Cause of the conflict was determined as "the TWR controller’s incorrect assumption that the Southwest Airlines 737 would depart from the runway before the Federal Express 767 arrived on the same runway, which resulted in a loss of separation between the two aircraft”.

Three Contributory Factors were identified as follows:

  • The TWR controller’s incorrect assumptions which were: 
    • his expectation bias regarding the 737’s departure, 
    • his lack of situational awareness regarding the 737’s position when the flight crew requested takeoff clearance, and 
    • the lack of controller training for low visibility operations.
  • The 737 flight crew’s failure to take account of the traffic that was on short final approach and to notify the controller that they would need additional time on the runway before their takeoff roll.
  • The Federal Aviation Administration’s failure to require surface detection equipment at Austin-Bergstrom International Airport and direct alerting for flight crews.

Seven Safety Recommendations were made based on the Findings of the Investigation as follows:

  • that the Federal Aviation Administration require airports that are certificated under Title 14 Code of Federal Regulations Part 139 and are currently not equipped with airport surface detection equipment model X or airport surface surveillance capability, implement surface detection equipment that:
    • tracks the movement of arriving and departing aircraft 
    • determines the proximity between those aircraft 
    • provides air traffic controllers with visual and aural cues of surface movements to aid in their decision-making processes. [A-24-10]
  • that the Federal Aviation Administration require air traffic controllers to: 
    • advise pilots, through direct communication and automatic terminal information system broadcasts, when visual contact with aircraft operating on taxiways and runways cannot be established or maintained and 
    • instruct pilots to provide accurate position reports to aid the controller in determining an aircraft’s location in such conditions. [A-24-11]
  • that the Federal Aviation Administration brief all air traffic controllers about the circumstances of this incident, emphasising the effect that certain conditions might have on a pilot’s ability to begin a takeoff in a timely manner, including
    • low-visibility weather conditions, such as fog. 
    • ambient conditions, such as temperature.
    • surface conditions, such as ice, snow, and other precipitation, as noted in Order 7110.65, Air Traffic Control, paragraph 5-8-4, Departure and Arrival. [A-24-12]
  • that the Federal Aviation Administration encourage Title 14 Code of Federal Regulations Part 121 operators to use their safety management system to identify flight crew surface navigation errors resulting from the performance of concurrent tasks during taxi and develop and implement effective risk mitigation strategies considering human factors principles. [A-24-13]
  • that the Federal Aviation Administration collaborate with aircraft and avionics manufacturers and software designers to develop the technology for a flight deck system that would provide visual and aural alerts to flight crews of traffic on a runway or taxiway and traffic on approach to land. [A-24-14]
  • that the Federal Aviation Administration require that the technology developed in response to Safety Recommendation A-24-4 be installed in all newly certificated transport category aircraft. [A-24-15]
  • that the Federal Aviation Administration require that existing transport-category aircraft be retrofitted with the technology developed in response to Safety Recommendation A-24-4. [A-24-16]

The Board also reviewed Five Previously issued related Safety Recommendations and formally reiterated them. These Recommendations were as follows:

  • that the Federal Aviation Administration collaborate with aircraft and avionics manufacturers and software designers to develop the technology for a flight deck system that would provide visual and aural alerts to flight crews of traffic on a runway or taxiway and traffic on approach to land. [A-24-4]
  • that the Federal Aviation Administration require that the technology developed in response to Safety Recommendation A-24-4 be installed in all newly certificated transport category aircraft. [A-24-5]
  • that the Federal Aviation Administration require that existing transport-category aircraft be retrofitted with the technology developed in response to Safety Recommendation A-24-4. [A-24-6]
  • that the Federal Aviation Administration require all newly-manufactured aircraft that must have a Cockpit Voice Recorder (CVR) be fitted with a CVR capable of recording the last 25 hours of audio. [A-18-30]
  • that the Federal Aviation Administration require retrofit of all Cockpit Voice Recorders (CVR) on all aircraft required to carry both a CVR and a Flight Data Recorder with a CVR capable of recording the last 25 hours of audio. [A-24-9]

The Final Report was completed on 6 June 2024 and published online later in the same month.

Related Articles

SKYbrary Partners:

Safety knowledge contributed by: