NTSB Aviation Investigation Report AIR-24-01 - Runway Incursion and Rejected Takeoff at JFK (American 106 and Delta 1943)

Runway Incursion and Rejected Takeoff

American Airlines Flight 106

Boeing 777 -200, N754AN, and

Delta Air Lines Flight 1943

Boeing 737 -900, N914DU

Queens , New York

January 13, 2023

Abstract : This report discusses the January 13, 2023 , runway incursion incident

involving American Airlines (AAL) flight 106, a Boeing 777 -200, and Delta Air Lines

(DAL) flight 1943, a Boeing 737 -900, at John F. Kennedy International Airport (JFK),

Queens, New York. The flight crew of the AAL airplane was instructed to taxi to

runway 4L for departure and cross runway 31L; however, the crew cross ed runway 4L,

where the DAL airplane had just begun its takeoff roll. The local controller cancelled

the DAL airplane’s takeoff clearance , and the DAL flight crew rejected the takeoff.

Safety issues discussed in this report include (1) the need for additional risk mitigation

strategies to prevent flight crew surface navigation errors that result in runway

incursions; (2) the need for a procedural crosscheck that requires flight crews to

verbalize the number of a runway they are about to cross as indicated by runway

signs; and (3) the lack of flight deck technology to detect potential traffic conflicts. As

a result of this investigation, the National Transportation Safety Board issues eight

new safety recommendations to the Federal Aviation Administration (FAA) , reiterates

one previously issued recommendation to the FAA , and supersedes two

recommendation s. May 29 , 2024 Aviation Investigation Report AIR-24-01

Aviation Investigation Report

AIR-24-01

i Contents

Figures ................................ ................................ ................................ .................. iii

Tables ................................ ................................ ................................ ................... iv

Acronyms and Abbreviations ................................ ................................ ................ v

Executive Summary ................................ ................................ .............................. vii

What Happened ................................ ................................ ................................ .............. vii

What We Found ................................ ................................ ................................ .............. vii

What We Recommended ................................ ................................ ............................... ix

1.Factual Information ................................ ................................ ......................... 1

1.1 History of Flight ................................ ................................ ................................ ........ 1

1.2 Personnel Information ................................ ................................ ........................... 10

1.2.1 American Airlines Flight Crew ................................ ................................ ..... 10

1.2.2 Delta Airlines Flight Crew ................................ ................................ ............. 11

1.3 Airplane Information ................................ ................................ .............................. 11

1.4 Meteorological Information ................................ ................................ .................. 11

1.5 Airport Information ................................ ................................ ................................ 12

1.5.1 Overview ................................ ................................ ................................ ......... 12

1.5.2 Lighting, Signage, and Markings ................................ ................................ . 12

1.5.3 Runway Status Lights System ................................ ................................ ....... 14

1.5.4 Airport Surface Detection System, Model X ................................ .............. 15

1.5.5 Hot Spots at JFK ................................ ................................ ............................. 17

1.5.6 Air Traffic Control Tower Staffing ................................ ................................ 18

1.5.7 Air Traffic Control Procedures ................................ ................................ ..... 18

1.6 Flight Recorders ................................ ................................ ................................ ..... 19

1.7 Tests and Research ................................ ................................ ................................ 19

1.7.1 Flight Data Recorder and ADS -B Study ................................ ...................... 19

1.7.2 Ground Visibility from the Flight Deck of the B777 ................................ .. 24

1.7.3 ASDE -X Replay and Runway Status Lights ................................ .................. 25

1.8 Organizational and Management Information ................................ ................... 27

1.8.1 American Airlines Flight Operations Manual ................................ ............. 27

Aviation Investigation Report

AIR-24-01

ii 1.9 Additional Information ................................ ................................ .......................... 28

1.9.1 Previous Related Safety Recommendations ................................ ............... 28

2. Analysis ................................ ................................ ................................ ......... 33

2.1 Introduction ................................ ................................ ................................ ............ 33

2.2 Incident Sequence ................................ ................................ ................................ . 34

2.2.1 American Airlines Flight Crew’s Deviation from Taxi Clearance ............. 34

2.3 Operational Barriers to Runway Incursions ................................ ......................... 40

2.3.1 Crew Resource Management ................................ ................................ ...... 40

2.3.2 American Airlines Policy on Completing the Load Closeout ................... 42

2.3.3 Flight Deck Technologies for Preventing Runway Incursions .................. 44

2.4 Airport Measures Intended to Prevent Runway Incursions ............................... 46

2.4.1 Runway Status Lights ................................ ................................ ..................... 46

2.4.2 Airport Surface Detection Equipment, Model X ................................ ........ 48

2.5 Air Traffic Control ................................ ................................ ................................ ... 49

2.5.1 Ground Controller ................................ ................................ ......................... 49

2.5.2 Local Controller ................................ ................................ ............................. 49

2.5.3 Air Traffic Control Tower Team ................................ ................................ .... 50

2.6 Cockpit Voice Recorder Duration and Loss of Information .............................. 50

3. Conclusions ................................ ................................ ................................ ... 52

3.1 Findings ................................ ................................ ................................ ................... 52

3.2 Probable Cause ................................ ................................ ................................ ...... 53

4. Recommendations ................................ ................................ ........................ 54

4.1 New Recommendations ................................ ................................ ........................ 54

4.2 Previously Issued Recommendations Reiterated in This Report ...................... 55

4.3 Previously Issued Recommendations Classified in This Report ....................... 55

Appendixes ................................ ................................ ................................ ......... 56

Appendix A: Investigation ................................ ................................ ............................ 56

Appendix B: Consolidated Recommendation Information ................................ ..... 57

Appendix C: Excerpts of American Airlines Flight Operations Manual ................. 60

References ................................ ................................ ................................ ........... 66

Aviation Investigation Report

AIR-24-01

iii Figures

Figure 1. AAL106’s assigned (dashed orange line) and actual (solid orange line)

taxi routes overlaid on a satellite image of JFK airport. ................................ .................... 3

Figure 2. Satellite image of the area of taxiways B, K, and J; and runways 4L

and 31L with AAL106’s assigned (dashed orange line) and actual (solid orange

line) taxi routes overlaid. ................................ ................................ ................................ ........ 6

Figure 3. Satellite image of the taxiway J/runway 4L intersection showing

lights and markings. (Source: Google Earth) ................................ ................................ .... 13

Figure 4. Taxiway K’s surface painted taxiway direction sign located on taxiway B.

(Source: Port Authority of NY and NJ) ................................ ................................ ............... 13

Figure 5. Photograph of activated RELs across runway 4L at night, as seen from

the hold short line on taxiway J. (Source: Port Authority of NY and NJ) ....................... 15

Figure 6. Screen capture of the JFK ASDE -X system display while the runway -

occupied alert was active for DAL1943 and AAL106. (Source: Federal Aviation

Administration) ................................ ................................ ................................ ..................... 16

Figure 7. The taxi paths of AAL106 (orange, terminal through taxiway J) and

DAL1943 (blue, taxiway K and runway 4L), based on ADS -B data, with times

annotated at select locations. ................................ ................................ ............................. 20

Figure 8. Airplane groundspeeds and separation distances. ................................ ......... 21

Figure 9. Locations and groundspeeds of both airplanes when AAL106 entered

runway 4L at 2044:50. ................................ ................................ ................................ .......... 22

Figure 10. Locations and groundspeeds of both airplanes when DAL1943

reached its maximum speed at 2044:53. ................................ ................................ .......... 23

Figure 11. Locations and groundspeeds of both airplanes when AAL106

cleared runway 4L at 2045:02. ................................ ................................ ............................ 24

Figure 12. Sight line from the pilot’s eye position to the ground for the B777.

(Source: Boeing) ................................ ................................ ................................ ................... 25

Figure 13. Satellite image of the taxiway J/runway 4L intersection showing the

location of the RELs and the position of AAL106 as the airplane’s nose crossed

the runway hold short marking. ................................ ................................ .......................... 26

Aviation Investigation Report

AIR-24-01

iv Tables

Table. NTSB investigations hampered by overwritten CVR data since Safety

Recommendations A-18-30 and -31 were issued ............................................................ 31

Aviation Investigation Report

AIR-24-01

v Acronyms and Abbreviations

AAL American Airlines

AC advisory circular

ACARS aircraft communications addressing and reporting system

ADS -B automatic dependent surveillance -broadcast

AIM Aeronautical Information Manual

ASDE -X airport surface detection equipment , model X

ASSC airport surface surveillance capability

ATC air traffic control

ATCT airport traffic control tower

ATIS automated terminal information service

ATP airline transport pilot

CFR Code of Federal Regulations

CRM crew resource management

CVR cockpit voice recorder

DAL Delta Air Lines

FAA Federal Aviation Administration

FB relief first officer

FDR flight data recorder

FMS flight management system

FO first officer

FOM flight operations manual

JFK John F. Kennedy International Airport

LHR London Heathrow International Airport

Aviation Investigation Report

AIR-24-01

vi NOTOC notification to captain

NTSB National Transportation Safety Board

NWS National Weather Service

REL runway entrance light

RSAT runway safety action team

RWSL runway status light system

SDQ Las Américas International Airport

SMS safety management system

TAF terminal aerodrome forecast

TPS takeoff performance system

Aviation Investigation Report

AIR-24-01

vii Executive Summary

What Happened

On January 13, 2023, about 2044 local time, American Airlines ( AAL )

flight 106, a Boeing 777 -223, N754AN, crossed runway 4L on taxiway J without air

traffic control (ATC) clearance at John F. Kennedy International Airport (JFK), Queens,

New York , causing Delta Air Lines (DAL) flight 1943, a Boeing 737 -900ER, N914DU, to

abort its takeoff roll on runway 4L. None of the 6 crew and 153 passengers on

DAL1943, nor the 12 crew and 137 passengers on AAL106 , were injured, and t here

was no damage to either aircraft. AAL106 operated as a Title 14 Code of Federal

Regulations (CFR) Part 121 scheduled international passenger flight from JFK to

London Heathrow International Airport (LHR) , London, United Kingdom . DAL1943

was a CFR Part 121 scheduled international passenger flight from JFK to Santo

Domingo (SDQ) , Dominican Republic. Night visual meteorological conditions

prevailed at the airport at the time of the incident.

The flight crew of AAL106 was instructed to taxi to runway 4L for departure via

taxiway B to taxiway K . During the taxi, they were instructed to cross runway 31L. The

airplane approached the intersection of taxiways B and K, and rather than turning

right onto K as instructed, the captain continued along taxiway B, which curved to the

left. The captain subsequently turned right onto taxiway J before crossing runway 4L,

from which DAL1943 had just begun its takeoff roll. Airport surface detection

equipment , model X (ASDE -X) alerted the tower controller to the conflict , and the

controller cancelled DAL1943’s t akeoff clearance. The crew of DAL1943 rejected

their takeoff , reaching a maximum groundspeed of about 105 knots , about 2,30 0 ft

from the taxiway J intersection. AAL106 continued across runway 4L and DAL1943

came to a stop before taxiing off the runway onto an adjacent taxiway without further

incident .

What We Found

The National Transportation Safety Board (NTSB) found that the flight crew of

AAL106 deviated from the instruction s given by air traffic control and inadvertently

crossed runway 4L due to a surface navigation error that likely stemmed from several

factors, including the timing of the instruction provided to the crew to cross

runway 31L; interruptions and multitasking as the crew performed required

crosschecks of their takeoff performance calculations ; the captain’s prospective

memory error , which resulted in his forgetting to turn right onto taxiway K; and

environmental capture , which prompted the captain to proceed along a familiar, but

incorrect, taxi route.

Aviation Investigation Report

AIR-24-01

viii The ASDE -X functioned as designed and alerted the air traffic controller to the

conflict between AAL106 and DAL1943, which resulted in the con troller’s timely

cancellation of DAL1943’s takeoff clearance; however, additional risk mitigation

strategies are needed to reduce the likelihood that flight crew surface navigation

errors will result in runway incursions . These might include procedural crosschecks

that would require a flight crew to verbalize the runway they are about to cross, flight

deck displays of airport traffic with aural and visual conflict alerting capability , and

as-of-yet undeveloped strategies for reducing likelihood of surface n avigation errors

caused by the need to perform multiple concurrent operational tasks during taxi.

Such strategies can be developed and tailored to the characteristics of an operator’s

unique constraints as part of an operator’s safety management system .

The taxiway/runway intersection where the incident occurred was equipped

with a runway status light system that included runway entrance lights (RELs) , which

comprised red lights embedded in the surface that illuminated to signal to

approaching aircraft that the runway was in use. The RELs illuminated during this

incident as designed ; however , they illuminated as the nose of AAL106 was crossing

the runway hold short markings, which was too late for the crew to perceive them and

stop the airplane in a safe a rea. As a result , this system did not serve as an effective

safeguard for alerting the crew that they were about to cross an active runway .

Finally, cockpit voice recorder (CVR) information was not avai lable for this

incident because the data were overwritten . As a result, the NTSB had to rely

exclusively on flight crew recollections about the incident ; however, these were not

documented until 1 month after the incident occurred. A cockpit voice recording

would likely have provided additional details about the content and timing of crew

communications; shed light on the crew’s minute -by-minute focus of attention;

revealed any unreported, nonpertinent conversations; and potentially provided

additional information about any distractions .

The NTSB determines that the probable cause of this incident was the AAL106

flight crew’s surface navigation error due to distractions caused by their performance

of concurrent operational tasks during taxi, which resulted in a loss of situational

awareness. Contributing to the incident was the air traffic control tower team’s

nondetection of the AAL106 crew ’s deviation from taxi instructions while performing

concurrent operational tasks ; the timing of the runway status light system, which

activated too lat e to prevent the AAL106 crew from crossing the runway hold short

line; and American Airlines’ lack of adequate risk controls to prevent concurrent flight

crew tasks from leading to distraction, loss of situational awareness, and deviation

from an authorized taxi clearance. Reducing the severity of the incident, and likely

preventing an accident, was the activation of the ASDE -X warning in the air traffic

control tower and the local controller’s prompt cancellation o f DAL1943’s takeoff

clearance.

Aviation Investigation Report

AIR-24-01

ix What We Recommended

As a result of this investigation, we made eight new recommendations to the

Federal Aviation Administration (FAA). We recommended that the FAA encourage

Title 14 Code of Federal Regulations (CFR) Part 91K, 135, and 121 operators to

incorporate into their standard operating procedures a procedural crosscheck that

requires flight crews to verbalize the number of a runway they are about to cross, as

indicated by runway signs, unless an installed automated system already provides an

aural advisory .

We recommended that the FAA e ncourage 14 CFR Part 121 operators to use

their safety management system to detect flight crew surface navigation errors

resulting from the performance of concurrent tasks during taxi and use scientific

knowledge from the field of human factors to develop and implement effective risk

mitigation strateg ies.

We recommended that the FAA c ollaborate with aircraft and avionics

manufacturers and software de signers 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 . We also recommended that the FAA

require this new technology to be installed in all newly certificated transport -category

airplanes , and that existing transport -category airplanes be retrofitted with th e

technology . These recommend ations superseded a recommendation that asked the

FAA to require a ground movement safety system that provide d a direct warning

capability to flight crews.

We recommended that the FAA e valuate the effectiveness of the activation

logic for the runway status light system (RWS L) considering the circumstances of this

incident. We also recommended that the FAA use the findings of this evaluation to

update the RWSL activation logic as necessary to improve system effectiveness.

We recommended that the FAA require retrofit of all cockpit voice recorders

(CVR) on all airplanes required to carry both a CVR and a flight data recorder with a

CVR capable of recording the last 25 hours of audio. This recommendation

superseded a recommendation that asked the FAA to require such recorders by

January 1, 2024.

Additionally, we reiterated the following recommendation that was previously

issued to the FAA:

Require all newly manufactured airplanes 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)

Aviation Investigation Report

AIR-24-01

1 1. Factual Information

1.1 History of Flight

On January 13, 2023, about 2044 eastern standard time, the flight crew of

American Airlines (AA L) flight 106, a Boeing 777 -200, N754AN, crossed runway 4L on

taxiway J without air traffic control (ATC) clearance at John F. Kennedy International

Airport (JFK), Queens, New York , causing the flight crew of Delta Air Lines (DAL) flight

1943, a Boeing 737 -900, N914DU, to abort their takeoff roll on runway 4L. None of

the 6 crew and 153 passengers on DAL1943, nor the 12 crew and 137 passengers on

AAL106 , were injured, and t here was no damage to either air plane . AAL106 operated

as a Title 14 Code of Federal Regulations (CFR) Part 121 schedu led international

passenger flight from JFK to London Heathrow International Airport (LHR) , London,

United Kingdom. DAL1943 was a CFR Part 121 scheduled international passenger

flight from JFK to Santo Domingo (SDQ) , Dominican Republic. Night visual

meteorological conditions prevailed at the airport at the time of the incident .1

The flight crew of AAL106 comprised the captain, a first officer (FO), and a

relief first officer (FB). In interviews conducted about 1 month after the incident , the

flight crew reported that , before the flight, they reviewed the flight plan, weather, and

new procedures contained in a two-sided company “Fleet Harmonization” quick

reference page , dated January 3, 2023. The quick reference page involved some

changes to flight crew responsibilities, and the FO and FB had not yet used the new

procedures. The captain stated that he had operated many flights to LHR and offered

the FO the option to serve as pilot flying or pilot monitoring. The FO ch ose to serve

as pilot flying.2

Upon arrival at the airplane, which was parked at gate 18, the FB conducted an

exterior preflight inspection while the captain and FO retrieved the automatic

terminal information service (ATIS) information whiskey .3 Based upon the ATIS

information, they anticipated and briefed a departure from runway 31L. The FO

programmed the flight management system (FMS), and she and the captain briefed

the taxi route. The captain recalled that his briefing included taxiway TA, le ft on B,

and hold short at J , which he stated was the standard route to runway 31L from the

1 Visit ntsb.gov to find additional information in the public docket for this NTSB incident

investigation (case number DCA23LA125). Use the CAROL Query to search safety recommendations

and investigations.

2 Although the FO was the pilot flying, the captain performed the taxi.

3 ATIS is a continuously broadcast automated service that provides current airport information

to departing and arriving aircraft. Each ATIS message is identified by a phonetic letter designation.

Aviation Investigation Report

AIR-24-01

2 AAL gates at JFK . The FO stated that the airport diagram was available and was

referenced during these discussions .

The captain produced a laminated copy of the Fleet Harmonization Memo

quick reference page , placed it on a console for the flight crew’s reference, and

stated to the other pilots that they would discuss the changes while en route to LHR.

The captain asked if the FO or FB were aware of any threats to the safety of the flight,

and the FO stated that she was unfamiliar with the approach into LHR. The captain

also recalled telling the FB to speak up if he had any safety concerns.

Recorded aircraft communications addressing and reporting system (ACARS)

data revealed that a t 2013:07, the crew received the flight’s instrument clearance ,

which indicated that the flight would depart from runway 4L.4 The crew reported that

the captain and FO obtained the updated ATIS, reprogrammed the FMS, and

rebriefed the departure and taxi route. According to the FO, the captain’s revised taxi

briefin g stated, “same taxi route up until Bravo short of Kilo.” The captain recalled

briefing the FO and FB that the controller would “have you come up Kilo, and then as

you get closer, they’ll tell you when to cross [runway 31L] .” The captain stated that this

was the typical route from the American Airlines gates to runway 4L. Figure 1 depicts

AAL106’ s assigned taxi route from the gate to runway 4L (dashed orange line) and its

actual taxi route (solid orange line) .5

4 ACARS is a digital datalink system for transmission of short messages between aircraft and

ground stations via airband radio or satellite.

5 AAL106’s taxi route was captured by automatic dependent surveillance –broa dcast

information provided by the FAA.

Aviation Investigation Report

AIR-24-01

Figure 1. AAL106’s assigned (dashed orange line) and actual (solid orange line) taxi routes

overlaid on a satellite image of JFK airport.

The captain subsequently briefed the passengers via the airplane’s public

address system, during which an airline employee provided the flight crew with a

dangerous goods form .6 The form provided to the crew contained a handwritten

annotation that the item, a load of lithium batteries, had not been loaded. The crew

acknowledged their receipt of the form via the ACARS at 2017. Around this time, the

crew received two ACARS messages f rom American Airlines dispatch advising of

moderate turbulence during departure and en route. The first message requested

crew acknowledgement, and the FO responded with the confirmation code.

6 The form was a notification to captain , or NOTOC, which informs the captain of the presence

of hazardous material located in the airplane’s cargo hold. The dangerous goods form, or DG, and

NOTOC are used interchangeably as an industry practice .

Aviation Investigation Report

AIR-24-01

4 The crew began pushback from the gate at 2028 . AAL106 made initial contact

with JFK air traffic control about 2033 when the FO report ed that the airplane was at

taxiway TA and requested taxi clearance , which the ground controller provided . The

crew was initially instructed to taxi to runway 4L via a left turn on to taxiway B and to

hold short of taxiway K ; the FO re sponded , “Bravo short of Kilo American one zero

six.” The captain and FO stated that they each had the airport diagram displayed on

their tablets , which were mounted to their respective sides of the cockpit, for

reference during taxi.

The captain recalled that he began to taxi the airplane, but the crew had not

yet received the ACARS load closeout, which was normally transmitted directly after

pushback.7 The captain was concerned that the delay in receiving the load closeout

was due to the handwritten notation on the dangerous goods form. The captain

recalled that he slowed the airplane’s taxi and asked the FO to request the load

closeout via ACARS, which she did at 2034:47.

After receiving an automated reply stating, “STANDBY – LOAD AGENT

NOTIFIED,” the captain asked the FB to call a company operations agent via radio to

receive the load closeout. The captain s tated that he did not normally need to

perform these tasks during taxi because the load closeout typically arrived soon after

the airplane pushed back from the gate. He said that he normally made a point of

crosschecking the load closeout with the FO and comple ting related tasks before

departing the ramp area. In this ca se, however, he said that he decided to begin the

taxi before receiving the load closeout because the flight had a third flight

crewmember (the FB).

According to the ATC audio recordings , at 2040:13, the ground controller

instructed AAL106 to cross runway 31L at taxiway K . The FO responded, “Cross three

one left at kilo for American one oh six heavy.” Airport surface detection equipment ,

model X (ASDE -X) information indicated that the airplane was on taxiway B

approaching taxiway N at this time. About 2041:39, a company operations agent

contacted the FB via radio to let him know that the load closeout would be coming

shortly. The FB then relayed this in formation to the captain.

About 2041, the local controller (communicating on a different frequency that

the crew of AAL106 was not yet monitoring ) provided DAL1943 a wake turbulence

advisory , instructed the crew to line up a nd wait on runway 4L at their discretion, then

7 The load closeout contained the final figures confirming number of passengers and fuel and

cargo weights.

Aviation Investigation Report

AIR-24-01

5 advised of traffic crossing right to left downfield.8 The crew acknowledged with only

the instruction to line up and wait on runway 4L.

Between 2042:12 and 2042:14, when AAL106 was between taxiways M and L,

the crew received three identical ACARS messages warning of moderate -to-severe

turbulence affecting the departure area. The FO printed and briefed these messages

to the captain, and the captain asked the FB to advise the flight attendants to remain

seated after departure because of these reports. In accordance with the new Fle et

Harmonization procedures, the FO made an announcement over the public address

system directing the flight attendants to be seated for departure.

ACARS data indicated that the load closeout was received at 2043:06, shortly

after the airplane had crossed taxiway L. The FB recalled removing the load closeout

paperwork from the printer and handing it to the FO. ACARS data indicated that, over

about the next 40 seconds, updated weight and balance information was transmitted

to the FMS and the crew accepted these values in the flight management computer ;

during this time , the airplane was crossing taxiway KE. The FB recalled that , after

removing the load closeout paperwork, additional paperwork remained on the

center pedestal, covering his radio panel and other controls and displays . He began

cleaning up this paperwork to access and adjust his radio controls , as overlapping

transmissions from the ground control and company operations frequencies were

“muddying up” his audio.

The FO recalled that she began performing required crosschecks of the load

closeout information with takeoff performance information on her tablet , which she

had previously been using to display the airport diagram. She stated that she

removed the tablet from its mount and placed it on her knee for about 1 minute while

discussing the performance values with the captain. The captain recalled that he was

discussing these performance values with the FO as the airplane approached

taxiway K.

At 2044: 15, about the time the local controller cleared DAL1943 for takeoff,

AAL106 crossed taxiway K and continued along taxiway B, which curved to the left. In

a postincident interview, the ground controller stated that he observed AAL106

swing out to the left on taxiway B to begin what he believed to be a typical wide right

turn onto taxiway K. He stated that, since the AAL106 flight crew correctly read back

their taxi instructions, he expected them to adhere to the instructions and perform the

taxi rout e as assigned.

8 According to the Aeronautical Information Manual , line up and wait is an air traffic control

procedure designed to position an aircraft onto the runway for an imminent departure.

Aviation Investigation Report

AIR-24-01

6 Review of JFK tower communications recordings showed that, as AAL106

passed taxiway K, the ground controller received a request from AAL107, which was

located near the departure end of runway 4L, to taxi to the ramp area. The controller

subsequently instructed AAL107 to taxi to the ramp. Shortly after completing his

communication with AAL107, the ground controller then established communications

with an airport operations vehicle and instructed them to perform a sweep of runway

31R in preparation for a runw ay configuration change.9

The ground controller recalled that, after he finished his communicati on with

the airport operations vehicle , he attend ed to another task that he speculated might

have involved “getting a strip or doing something .” At 2044:30, AAL106 made a right

turn on taxiway J where it continued toward runway 4L. Figure 2 depicts AAL106’s

missed right turn onto taxiway K and the airplane’s path toward taxiway J/runway 4L.

Figure 2. Satellite image of the area of taxiways B, K, and J; and runways 4L and 31L with

AAL106’s assigned (dashed orange line) and actual (solid orange line) taxi routes overlaid.

9 In the postincident interview, the JFK tower operations supervisor indicated that the runway

configuration change began about 5 minutes before the incident at the request of the New York

Terminal Radar Approach Control traffic management unit. He stated t hat he attempted to delay this

change until 2130, as they routinely reconfigured runways at JFK between around 2130 and 2230;

however, the traffic management unit was “insistent” that the configuration change take place “as soon

as possible.”

Aviation Investigation Report

AIR-24-01

7 The FO recalled being “head -down” and talking about takeoff performance

when she heard the captain say, “Cleared to cross.” She recalled looking up, leaning

forward, gazing down the runway to the right, and checking the runway and final

approach area for c onflicting traffic. She recalled seeing no airplane on the runway.

The captain recalled looking left and right to clear the runway before entering it, and

he did not see an airplane on the runway. According to all three pilots, the captain

stated, “clear l eft,” and the FO responded, “clear right.”10

The captain recalled activating the airplane’s runway turnoff lights, both

landing lights, and the nosewheel light before crossing the runway. The FO stated

that the taxi light was o n and that before crossing the runway, the captain reached to

turn on additional lights. She could not recall which lights were on but stated, “I know

he had crossing lights on.” The FB recalled that the captain turned on the “taxi light,

nose lights and the wing lights.” He added, “I was probably just coming up from

heads down at that point.”

Automatic dependent surveillance –broadcast (ADS -B) location information

indicated that DAL1943 began its takeoff roll on runway 4L at 2044:30, as AAL106

turned onto taxiway J.11 At 2044:40, AAL106 reached the runway hold short line ,

taxiing at a speed about 11 knots (kts) as DAL1943 was 3,900 ft from the intersection

between runway 4L and taxiway J and accelerating through 47 kts; according to

ASDE -X information, the runway entrance lights (RELs) for runway 4L illuminated at

this time . By 2044:50, the nose of AAL106 had crossed onto runway 4L. DAL1943 was

2,700 ft from the intersection and accelerat ing through 94 kts.

The captain of AAL106 recalled that, as the airplane crossed the runway

centerline, he saw red lights illuminate on the runway but was unsure why they had

illuminated and chose to add power to expedite the runway crossing . The FO also

observed red lights illuminate but stated that they did so after the airplane crossed

10 According to the captain ’s postincident interview , “I said, ‘clear left, clear right.’ [The FO]

said, ‘clear right.’  “

According to the FO ’s postincident interview , the captain “started to turn to the runway, and he

said ‘cleared to cross’ … I picked my head up, looked all the way down the runway. I checked the

runway and final, and I actually had to lean forward to see around. So —and he checked left. He said

‘clear left.’ I said ‘clear right.’  “

11 According to the Federal Aviation Administration, ADS -B uses GPS satellite signals to

provide air traffic controllers and pilots with accurate information that will help keep aircraft safely

separated in the sky and on runways. Aircraft transceivers recei ve GPS signals and use them to

determine the aircraft’s precise position in the sky, which is combined with other data and broadcast

out to other aircraft and air traffic control facilities.

Aviation Investigation Report

AIR-24-01

8 the runway “stop lines .”12 The FO stated that bo th she and the captain realized that

the illumination of the red lights (runway entrance lights) indicated that something

“wasn’t right,” and that the captain added power and “pushed across the runway.”

The FB recalled that his attention was focused inside the flight deck after the

crew received the load closeout . He stated that he was head -down as the airplane

approached the runway and “heard some confusion about crossing a runway.”13

When he looked up, he saw that the airplane was across the hold short line, and as

the airplane straightened out to cross the runway, he saw another airplane “down to

the right.”

The captain recalled that, after they crossed the runway, the FB stated that an

airplane was “taxiing on the runway.” The captain looked out the right cockpit

window and saw an airplane that appeared to be taxiing on runway 4L. He stated that

it did not look “that close” from his perspective.

In a postincident interview, the local controller stated that he performed a

visual scan before issuing DAL1943’s takeoff clearance, then began talking to the

crew of another airplane . Recorded tower communications revealed that, in the 15

seconds after issuing DAL1943’s takeoff clearance, the local controller spoke with two

other airplanes that were taxiing for takeoff , the first on runway 4L and the second on

runway 31L. At 2044:42, the ASDE -X system annunciated a “runway occupied” alert in

the control to wer. At 2044:4 7 and again at 2044:51, the local controller advised

DAL1943 that their takeoff clearance was cancelled , and at 2044:54 , the DAL 1943

crew responded that they were rejecting the takeoff.

The ground controller stated that he did not know if the crew of AAL106 was

monitoring his frequency, but he saw that AAL106 was crossing runway 4L and

decided it was best to let it continue crossing. He felt that broadcasting AAL106’s

callsign may have caused the crew to slow down or stop. Shortly thereafter, the local

controller asked the crew of DAL1943 whether they were able to taxi or if they

needed to complete checks. The crew responded that they could clear the runway,

and the local controller instructed the crew to turn left on taxiway J. The flight crew of

AAL106 could not hear the transmissions between the local controller and DAL1943

because they had not yet switched to tower frequency.

12 According to the FO’s recollection in a postincident interview , “The red lights actually came

up like —if I’d be sitting on them. They were already —the lights came from behind my shoulder , but you

could see the outline because they were pretty bright. So that ’s how we saw the lights. ”

13 Due to a lack of cockpit voice recorder information, conversation s between the flight crew

regarding the runway crossing could not be verified. Refer to Sections 1. 6, Flight Recorders, and 1.9,

Additional Information .

Aviation Investigation Report

AIR-24-01

9 About 204 6, after crossing runway 4L, the captain of AAL106 stopped the

airplane on taxiway J and the flight crew of AAL106 switched to tower frequency ,

shortly after which the local controller advised the crew of AAL106 that they had a

possible pilot deviation and to stand by for a phone number .14 The captain of AAL106

asked, “The last clearance we were given , we were cleared to cross , is that correct?”

to which the controller replied, “American 106 heavy we’re departing runway four left

… you are supposed to depart runway four left. You are currently holding short of

runway three one left.”15 The captain later said that he realized during this discussion

that he had made a mistake.

After receiving the phone number, t he captain called the controller and stated

that the crew had briefed a runway 31L departure, but “as we were coming out along

on Bravo as we taxied out we got a new … it was runway four left and then we were

told to cross at Juliet and … or Kilo or something I don’t remember exactly now

exactly because we were still trying to get this closeout problem fixed and I guess

then we crossed the wrong runway.”16 The controller stated ATC personnel were

listening to audio recordings to confirm what clearances had been issued to AAL106.

The controller took the captain’s name and phone number , advised that the flight

should prepare to depart from runway 31L, and instructed the captain to resume

communication via radio.

The AAL106 flight crew did not report the incident to American Airlines before

departure, and the captain stated that, to his knowledge, no provision in the

company’s operations manual would have prohibited the flight’s departure following

the event. The FB stated t hat he called the pilot union hotline, and a representative

confirmed that there was no prohibition on their departure. The captain conferred

with the FO and FB regarding whether they “felt okay” about continuing the flight,

and they stated that they did. The flight subsequently departed runway 31L and

completed the trip to LHR without further incident.

In a postincident written statement, the captain of DAL1943 stated that the

flight had been cleared for takeoff following a brief delay holding on the runway.

Shortly after the airplane reach ed a speed of 80 kts, the tower controller cancelled

their takeoff clearance, and the captain initiated rejected takeoff procedures. He

stated that , due to the “extreme darkness” of the evening, he and his FO did not see

AAL106 crossing the runway ahead until after rejecting the takeoff, and they did not

14 In Order 8020.11D, the FAA defines pilot deviation as, “ an action of a pilot that results in the

violation of a Federal Aviation Regulation (FAR) or a North American Aerospace Defense (Command

Air Defense Identification Zone) tolerance.”

15 At this time, AAL106 was on taxiway J between taxiways Y and Z.

16 The phone call between the controller and the captain was recorded .

Aviation Investigation Report

AIR-24-01

10 believe that AAL106’s exterior lights were illuminated. The FO recalled that only

AAL106’s beacon lights were visible.

1.2 Personnel Information

1.2.1 American Airlines Flight Crew

1.2.1.1 Captain

The captain had been employed by American Airlines since June 1989 and

had held the position of captain on the Boeing 777 (B777) since February 2022. He

had about 20,000 hours total flight experience at American Airlines and about

14,000 hours total pilot -in-command experience. He had about 9,000 hours on the

B777 , of which about 500 hours was pilot -in-command. The captain stated that he

had been based at JFK for about 10 years and that he was “very familiar” with the

airport.

The captain commuted to JFK from his home near Buffalo, New York, via

airline. He served as captain on a flight from JFK to LHR on the evening of Tuesday,

January 10 , returning to his home on the night of January 12 . He reported that he

received a 2 -hour nap in the crew rest area during the flight to LHR , a 5-hour nap after

arriving at LHR, and more than 8 hours of sleep during the two nights before the

incident .

On the day of the incident, he departed Buffalo about 1330 and arrived at JFK

about 1430. He slept in a recliner in a crew rest area from about 1600 to 1800. He

reported good sleep quality in the days preceding the incident and stated that he felt

rested w hen he reported for duty. The FO and FB also reported that the captain’s

alertness and mood on the night of the incident seemed good.

1.2.1.2 First Officer

The FO was hired by American Airlines in April 2014 and had recently

transitioned to the position of FO on the B777. She had flown out of JFK on four

previous occasions. The FO reported 1,188 hours total flight experience at American

Airlines , of which about 126 hours was in the B777. The FO was on reserve duty at

home during the 72 hours before the incident and reported sleeping more than

8 hours per night during that time . On the evening of the incident, she commuted to

JFK by car from her residence, depa rting about 1700 and arriving about 1830. She

recalled feeling rested when she reported for duty , and the captain described her as

appearing rested and in a good mood .

Aviation Investigation Report

AIR-24-01

11 1.2.1.3 Relief First Officer

The FB was hired by American Airlines in September 2016 and reported about

2,800 total hours of flight experience with the company , of which about 110 hours

was on the B777. He stated that he had operated “hundreds” of flights from JFK with

both AAL and his previous employer. The FO was off duty for the 2 days before the

incident and reported sleeping more than 8 hours per night. On the afternoon of the

incident, he commuted by car from his home to JFK, departing around 1400 or 1500.

He recalled feeling rested when he reported for duty , and the captain also described

him as appearing rested and in a good mood .

1.2.2 Delta Airlines Flight Crew

1.2.2.1 Captain

The Delta Airlines captain was hired at Delta in February 1999. He began flying

as FO on the DC -9 before transitioning to the Boeing 757/767. He upgraded to

captain on the B737 in October 2018.

1.2.2.2 First Officer

The Delta Airlines FO was hired by Delta in June 2022 and had served as FO

on the B737 since that time.

1.3 Airplane Information

The American Airlines Boeing 777 is configured with seating for 273 in three

cabins. The airplane is about 200 ft long with a win gspan about 200 ft. The B777 is

powered by two Rolls -Royce engines, each capable of producing 95,000 lbs of thrust.

The Delta Air Lines B oeing 737-900 is configured with seating for 149 in two cabins.

The airplane is about 138 ft long with a wingspan about 120 ft. The 737 -900 is

powered by two CFM -56 engines, each capable of producing 27,300 lbs of thrust.

1.4 Meteorological Information

At 1951, the automated weather observation at JFK included wind from 320 °

at 18 kts, 10 miles or greater visibility, an overcast cloud ceiling at 2,700 ft above

ground level , temperature 5°C, dew point temperature -1°C, and an altimeter setting

of 29.69 inches of mercury.

Aviation Investigation Report

AIR-24-01

12 1.5 Airport Information

1.5.1 Overview

John F. Kennedy International Airport is a publicly owned entity operated by

the Port Authority of New York and New Jersey (Port Authority). At the time of the

incident, t he airp ort had four runways: runway 4L/22R, runway 4R/22L, runway

13L/31R, and runway 13R/31L. The incident occurred on runway 4L, which measured

12,079 ft long and 200 ft wide.

1.5.2 Lighting, Signage, and Markings

The intersection of taxiway J and runway 4L/22R was lighted and marked with

taxiway edge lights, taxiway edge markings, an enhanced taxiway centerline, both

elevated and in -pavement runway guard lights, hold short markings, and two surface

painted holding posi tion markings (see figure 3) .17 Additionally, the Port Authority

reported that a surface painted direction sign for taxiway K had been painted on

taxiway B in July 2022 (see figure 4) . A lighted outbound destination sign was located

on the right side of taxiway K displaying the direction of runway 4L.

17 According to the AIM, the enhanced taxiway centerline marking consists of a parallel line of

yellow dashes on either side of the normal taxiway centerline. The taxiway centerlines are enhanced for

a maximum of 150 ft before a runway holding position marking. Runway guard lights consist of either a

pair of elevation flashing yellow lights installed on either side of the taxiway, or a row of in -pavement

yellow lights installed across the entire taxiway, at the runway holding position marking.

Aviation Investigation Report

AIR-24-01

Figure 3. Satellite image of the taxiway J/runway 4L intersection showing lights and

markings. (Source: Google Earth)

Figure 4. Taxiway K ’s surface painted taxiway direction sign located on taxiway B.

(Source: Port Authority of NY and NJ)

Note: The lighted outbound destination sign for runway 4L can be seen on the right side of taxiway K.

Aviation Investigation Report

AIR-24-01

14 1.5.3 Runway Status Lights System

JFK is one of 20 US airports where a runway status light system (RWSL) is

installed.18 According to FAA Advisory Circular 150/5340 -30J, Design and Installation

Details for Airport Visual Aides, RWSL s, “display critical, time -sensitive safety status

information directly to pilots and vehicle operators via in -pavement lights giving them

an immediate indication of potentially unsafe situations.”

The RWSL consists of an RWSL processor and a field lighting syste m. The RWSL

processor receives surveillance data of aircraft and vehicles on or near the airport

surface from the ground surface surveillance system, ASDE -X. The R WSL processor

uses these data to establish the presence and motion of aircraft and surface vehicles

on or near the runways. It then determines when to activate and deactivate the

runway entrance lights (RELs) and takeoff hold lights. The system illuminates red RELs

if the runway is u nsafe for entry or crossing and illuminates red takeoff hold lights if

the runway is unsafe for departure. The system extinguishes the lights automatically

as appropriate when the runway is no longer unsafe.

1.5.3.1 Runway Entrance Lights

RELs are located where a taxiway intersects a runway and only indicate a

runway’s status —lack of such lights does not indicate clearance to enter or cross a

runway. RELs illuminate when an aircraft is taking off or landing on a runway, just

before the aircraft reaches the taxiway intersection . According to the FAA, RELs will

illuminate on crossing taxiways when a departing aircraft has 1) reached a speed of

29.9 kts coupled with an acceleration of 1.20 m eters /second2, or 2) reached a speed

of 49.9 kts.

Per FAA Advisory Circular 150/5340 -30J, RELs are installed parallel to the

taxiway centerline and spaced laterally 2 ft from the taxiway centerline on the

opposite side of the taxiway centerline lights (if installed). An REL array will typically

consist of a minimum of six unidirectional lights, but may include more or fewer,

depending on the distance between the runway centerline and the holding position.

Six RELs were installed at the taxiway J/runway 4L intersection at JFK. Figure 5 depicts

the REL s at the taxiway J/runway 4L intersection when activated (yellow runway guard

lights and green taxiway centerline lights can also be seen illuminated) .

18 RWSLs combine airport lighting equipment with airport surveillance systems to create an

additional layer of runway safety that is intended to reduc e the number and severity of runway

incursions without interfering with normal airport operations.

Aviation Investigation Report

AIR-24-01

Figure 5. Photo graph of activated REL s across runway 4L at night , as seen from the hold

short line on taxiway J . (Source: Port Authority of NY and NJ)

1.5.3.2 RWSL Operating Characteristics

The FAA’s Runway Status Lights Pilot Reference Guide describe d the system as

follows :

Operating Characteristics – Departing Aircraft: When a departing aircraft

reaches 30 knots, all taxiway intersections with Runway Entrance Lights

arrays along the runway ahead of the aircraft will illuminate. As the

aircraft approaches a Runway Entrance Li ghts equipped taxiway

intersection, the lights at that intersection extinguish approximately 2 to

3 seconds before the aircraft reaches it. This allows controllers to apply

“anticipated separation ” to permit air traffic control to move traffic more

expedit iously without compromising safety. After the aircraft is declared

“airborne ” by the system, all lights will extinguish.

The NTSB requested that the FAA evaluate the operational algorithm/criteria

for the RWSL installation at JFK to ensure that it was operating as described on the

day of the incident. FAA Air Traffic Organization responded with a detailed analysis of

the incident and concluded that , “the JFK RWSL system [was] operating within the

standard algorithm/criteria for RWSL light activations.”

1.5.4 Airport Sur face Detection System , Model X

JFK airport was equipped with an ASDE -X system to alert air traffic controllers

of potential runway conflicts by providing detailed coverage of movement on

runways and taxiways. Controllers are presented this information on a color display

depicting aircraft and vehicle positions as icons overlaid on a map of the airport’s

Aviation Investigation Report

AIR-24-01

16 runways, taxiways, and airport approach corridors.19 The map is continuously

updated and can be particularly beneficial at night or in poor visibility conditions as

an enhancement to controller situational awareness.

The ASDE -X system is equipped with visual and aural alarms to alert controllers

of possible runway incursions or incidents. This software predicts the path of an

aircraft landing or departing as well as vehicular movements on runways. Figure 6

shows a screen capture of the ASDE -X display during the runway -occupied alert for

this incident. The system also annunciated a n accompanying “runway occupied” aural

alert .

Figure 6. Screen capture of the JFK ASDE -X system display while the runway -occupied alert

was active for DAL1943 and AAL106. (Source: Federa l Aviation Administration)

Note : Red-boxed t ext in the upper left corner identif ies DAL1943 and AAL106 as the airplanes

involved in the alert, with both airplanes circled in red on the airport map.

19 The airplane icons depicted on the ASDE -X display centered the wings over the point

locations rather than the location of the GPS antenna, so the noses of the airplanes appeared slightly

more forward than they likely were.

Aviation Investigation Report

AIR-24-01

17 1.5.5 Hot Spots at JFK

At the time of the incident, JFK did not have any charted hot spots , which the

FAA’s Chart Supplement defines as a “location on an aerodrome movement area with

a history or potential risk of collision or runway incursion, and where heightened

attention by pilots/drivers is necessary. ” Typically, it is a complex or confusing taxiway

or taxiway and runway intersection . According to the FAA, s ome factors that can

contribute to a hot spot designation are airport layout ; traffic flow ; airport marking,

signage, and lighting ; and pilot situational awareness or training.

Hot spots are depicted on airport diagrams as open circles or polygons

designated as “HS1,” HS2,” etc., and the qualifying features of each are briefly

described in the Airport Diagram Hot Spots Chart Supplement. Runway Safety Action

Teams (RSATs), which comprise local airport stakeholders, meet at least once a year

to identify risks to surface safety at their airport and develop plans to mitigate or

eliminate those risks.

One of the topics that might be discussed during RSAT meetings is published

airport hot spots, including whether additions or deletions of hot spots m ay be

appropriate. While there are no specific criteria listed for additions and deletions, hot

spots are not added to airport diagrams without concurrence from members of the

local RSAT. Similarly, h ot spots remain charted on airport diagrams until the RSAT

determines that the increased risk posed by the area has been reduced or

eliminated.

Review of JFK RSAT meeting minutes between 2004 and 2022 revealed that

various hot spots were depicted at complex taxiway intersections before 2010 ,

including the area where the incident occurred . Minutes from the 2011 RSAT meeting

indicated consensus among pilots, the Port Authority, and other stakeholders that no

hot spots would be charted at JFK due to updates in airport signage and marking s

and improvements to taxiway angles . Subsequent meeting notes indicated that

“areas of concern” at the airport continuously changed due to the variety of runway

configurations available.

The tower supervisor and cab coordinator on duty at the time of the incident

said that tower controllers were trained that the intersections of taxiways K and J with

taxiway B were a hot spot. The tower supervisor described it as the “main hotspot,

because it is not just the geometry of the turn, but just the amount of traffic that goes

through that intersection.”

On June 15, 2023, about 6 months after and as a result of the incident , the

intersection of taxiways B, K, and J at JFK was charted as hot spot 1 (HS1) on the

airport diagram with the following descri ption added in the chart supplement ,

Aviation Investigation Report

AIR-24-01

18 “Maint[ain] vigilance Twy K and Twy J complex int [ersection] close proximity to Rwy

04L and Rwy 31L .”

1.5.6 Air Traffic Control Tower Staffing

At the time of the incident, the ATCT was occupied by seven controllers

staffing the local control, ground control, cab coordinator, flight data/clearance

delivery, and operations supervisor positions. Some positions were combined due to

the lower traffic level at the time of the incident. All air traffic controllers assigned to

positions in the JFK ATCT were certified and qualified in accordance with federal

regulations.

1.5.7 Air Traffic Control Procedures

1.5.7.1 Duty Priority

FAA Order 7110.65Z, Air Traffic Control, paragraph 2−1−2, Duty Priority,

provided procedures and guidance for controllers in prioritizing their duties and

stated in part:

a. Give first priority to separating aircraft and issuing safety alerts as

required in this order. Good judgment must be used in prioritizing all

other provisions of this order based on the requirements of the situation

at hand.

NOTE− Because there are many variables involved, it is virtually

impossible to develop a standard list of duty priorities that would apply

uniformly to every conceivable situation. Each set of circumstances must

be evaluated on its own merit, and when more than one action is

requ ired, controllers must exercise their best judgment based on the

facts and circumstances known to them. That action which is most critical

from a safety standpoint is performed first.

1.5.7.2 Tower Team Positions Responsibilities

FAA Order 7110.65Z, Air Traffic Control, paragraph 2−10−3, Tower Team

Position Responsibilities, described the team concept and the primary responsibilities

of the tower team and stated in part:

a. Tower Team Concept and Intent: There are no absolute divisions of

responsibilities regarding position operations. The tasks to be

completed remain the same whether one, two, or three people are

working positions within a facility/sector. The team, as a whole, has

responsibi lity for the safe and efficient operation of that facility/sector.

Aviation Investigation Report

AIR-24-01

19 1.5.7.3 Visually Scanning Runways

FAA order 7110.65Z, Air Traffic Control, paragraph 3−1−12, Visually Scanning

Runways, prescribed procedures for the local and ground controller for scanning

runways, which stated:

a. Local controllers must visually scan runways to the maximum extent

possible.

b. Ground control must assist local control in visually scanning runways,

especially when runways are in close proximity to other movement

areas.

1.6 Flight Recorders

Each airplane was equipped with a flight data recorder (FDR) and cockpit voice

recorder (CVR). FDRs are required to record a minimum of 25 hours of data. The FDR

data for the incident was obtained from both airplanes and included airplane

attitude, accelerations, altitude, airspeed, control inputs, control surface positions,

engine parameters, and avionics settings, among other parameters. As a result of

their 2 -hour recording capacity, the incident was overwritten on both airplanes’ CVRs.

1.7 Tests and Research

1.7.1 Flight Data Recorder and ADS -B Study

The l atitude and longitude parameters obtained from both airplanes’ flight

data recorders (FDRs) were not recorded with sufficient precision to accurately

determine the airplanes’ locations on the runway near the time of the incident .

Therefore, ADS -B information provided by the FAA was used to determine airplane

location. ADS -B has an accuracy of about 20 meters (65 ft) in both the horizontal and

vertical dimensions. The location data are reflective of the location of the GPS

antenna on each airplane. On t he B777 (AAL106) , the antenna is located on top of

the fuselage, 46 ft aft of the nose ; while the antenna on the B737 (DAL1943) is

located on top of the fuselage about 33 ft aft of the nose. The separation calculations

made in this report reflect the distance between the two GPS antennae and do not

account for the wingspan of the B777 nor for the forward fuselage of the B737.

Figure 7 shows the taxi paths of both airplanes . AAL106 left the ramp area and

entered taxiway TA at 2034:45, at which time DAL1943 was in line on taxiway K for

departure from runway 4L. At 2040:13, t he tower controller instructed AAL106 to

cross runway 31L at taxiway K. At 2044:30, AAL106 turned onto taxiway J as DAL1943

began its takeoff roll. Airplane groundspeeds shown in figure 8 display the FDR -

Aviation Investigation Report

AIR-24-01

20 recorded groundspeed s, which were consistent with groundspeed calculated from

the airplanes’ respective ADS -B taxi tracks , as well as the separation distance between

the two airplanes . AAL106 was considered to be crossing runway 4L from the time its

nose was over the runway surface until its tail cleared the opposite runway edge.

Figure 7. The taxi paths of AAL106 ( orange , terminal through taxiway J ) and DAL1943 (blue ,

taxiway K and runway 4L ), based on ADS -B data , with times annotated at select locations.

Aviation Investigation Report

AIR-24-01

21 Figure 8. Airplane groundspeeds and separation distance s.

By 2044:40, AAL106 ’s nose was passing the hold short line for runway 4L at a

groundspeed of about 11 kts. DAL1943 was 3,900 ft from the intersection of 4L and

taxiway J at 47 kts groundspeed and accelerating. ADSE -X recorded that the RELs for

runway 4L illuminated at this time.

At 2044:4 7, the local controller cancelled DAL1943’s takeoff clearance. At

2044:50 , the nose of AAL106 entered onto runway 4L at a groundspeed of about

15 kts. DAL1943 was 2,700 ft from the intersection of 4L and taxiway J at 94 kts

groundspeed and accelerating (see figure 9) .

Aviation Investigation Report

AIR-24-01

22 Figure 9. Locations and groundspeeds of both airplanes when AAL106 entered runway 4L at

2044:50.

FDR data indicated that DAL1943’s throttle moved to idle and the autobrakes

were triggered at 2044:51. Speedbrakes were deployed at 2044:52 , and the airplane

reached a maximum groundspeed of 105 kts at 2044:53, 2,300 ft from the

intersection of 4L and taxiway J , as the nose of AAL106 was passing over the

centerline of runway 4L at a groundspeed of 18 kts (see figure 10) .20 DAL1943’s t hrust

reversers deployed at 2044:55.

20 Speedbrakes are high -drag devices on an airplane ’s wing used to assist in slowing the

airplane.

Aviation Investigation Report

AIR-24-01

Figure 10. Locations and groundspeeds of both airplanes when DAL1943 reached its

maximum speed at 2044:53.

By 2045:02, AAL106 was clear of runway 4L and traveling at a groundspeed of

21 kts. DAL1943 had decelerated to 49 kts and was 1,000 ft from AAL106 (see

figure 11).

Aviation Investigation Report

AIR-24-01

24 Figure 11. Locations and groundspeeds of both airplanes when AAL106 cleared runway 4L

at 2045:02.

DAL1943 came to a stop on runway 4L at 2045:23, 500 ft from the intersection

of 4L and taxiway J. It then exited the runway, turning left onto taxiway J.

1.7.2 Ground Visibility from the Flight Deck of the B777

According to information provided by Boeing and shown in figure 12 , the

horizontal distance between the pilot’s eye position in a B777 and the closest visible

point on the ground was 582 in (48 ft , 6 in). These dimensions indicate that the

horizontal distance between the closest point on the ground in front of the airplane

that was visible to the pilots and the transmitted GPS location was 1,045 in (87 ft, 1 in).

Aviation Investigation Report

AIR-24-01

Figure 12. Sight line from the pilot’s eye position to the ground for the B777. (Source:

Boeing)

1.7.3 ASDE -X Replay and R unway Status Lights

The FAA provided a playback file of the ASDE -X taxi paths for the two

airplanes. The ASDE -X point locations of the airplanes appeared consistent with the

ADS -B locations; however, the airplane icons shown in the ASDE -X replay centered

the wings over the point locations rather than the location of the GPS antenna, so the

nose s of the airplanes appeared slightly more forward than they likely were.

According to the ASDE -X replay, the RELs illuminated at 2044:40, coincident

with DAL1943 accelerating from 47 kts through 49 kts groundspeed. Figure 13 shows

a scaled depiction of AAL106 with the GPS antenna location positioned at the ADS -B

GPS position point. At 2044:40, the nose of AAL106 was just crossing the hold short

line at a speed of 11 kts, and REL 1 was beneath the airplane.

Aviation Investigation Report

AIR-24-01

Figure 13. Satellite image of the taxiway J/runway 4L intersection showing the location of the

RELs and the position of AAL106 as the airplane’s nose crossed the runway hold short

marking.

Note: The approximate location of the airplane’s GPS antenna is shown.

As discussed in the previous section, the flight crew can see objects on the

ground 87 ft in front of the airplane’s transmitted GPS location ; however, the ASDE -X

location of the airplane has an uncertainty of ± 20 ft.

By 2044:39, the hold short line and REL 1 would have been beneath the

airplane and out of the AAL106 crew ’s field of view . At 2044:40, REL 2 was 100 ft

forward of the point location of AAL106 ; therefore, when the RELs illuminated, REL 2

may have been within the crew’s field of view while REL s 3, 4, 5, and 6 were within the

crew’s view. As AAL106 continued forward, subsequent RELs would have passed out

of view beneath the airplane. REL 6 would have passed out of view about 2044:52.

The RELs turned off at 204 5:07 as DAL1943 slowed below 25 kts.

The ASDE -X replay of the event depicted REL s 2 and 3 at the 4L/taxiway J

intersection as a gray circle and a red circle, respectively, with the other four RELs in

the array depicted as orange circles (illuminated) . According to the FAA A ir Traffic

Organization , a gray circle indicated an “extinguished” lamp while a red circle

indicated a “faulted” lamp. FAA Order 6850.38, Maintenance of the Runway Status

Aviation Investigation Report

AIR-24-01

27 Lights (RWSL) System, states that operational tolerance/limits allow <50% of the

lamps in an REL array to be inoperative.21

1.8 Organizational and Management Information

1.8.1 American Airlines Flight Operations Manual

The American Airlines Flight Operations Manual ( FOM ) defined taxiing as a

critical phase of flight, during which crews shall not engage in nonessential activities,

such as company communication, including completing paperwork.22 The FOM

stated that, during taxi, “The flight crew’s primary task is to safely taxi the aircraft and

the flight crew’s attention should not be diverted from this task.” The FOM included

an additional note to “Ensure received (not expected/briefed) taxi route is followed.”

The FOM stated that both pilots should have the airport diagram readily

available and reference it as necessary to ensure the taxi clearance is followed

correctly, that the FO should advise the captain when accomplishing nonmonitoring

tasks (e.g., FMS prog ramming, ACARS, company radio calls, etc.), and that both pilots

should be heads up to visually monitor the aircraft’s progress at critical locations on

the airport (hold short, crossing runways, etc.).

The FOM gave further direction that, when approaching an entrance to an

active runway, both pilots will ensure compliance with hold short and crossing

clearances by discontinuing nonmonitoring tasks (e.g., FMS programming, ACARS,

company radio calls, etc.) .

The FOM stated that all received ACARS messages should be verified for the

correct flight number, date, and aircraft nose number. Additionally, takeoff

performance system (TPS) and load closeout information were to be verified against

documents obtained from operations for reasonableness .23 In addition, all

subsequent TPS and/or load closeout messages were to be compared against

21 While paperwork obtained from the Port Authority of NY and NJ dated January 15, 2023,

identified RELs 3 and 4 as being inoperative, ASDE -X playback showed the inoperative fixtures were

RELs 2 and 3 .

22 American Airlines clarified in a statement after the incident that tasks related to load closeout

are considered essential for safety of flight and are part of the required pre-takeoff checklist that can

be accomplished during taxi ; the performance of such duties is not regarded as contrary to the FOM’s

proscription against “completing paperwork” during the sterile flight deck period.

23 This information included the airplane’s zero -fuel weight, takeoff gross weight, and reference

speeds.

Aviation Investigation Report

AIR-24-01

28 previously received messages. See appendix C for relevant excerpts of the American

Airlines FOM.

1.9 Additional Information

1.9.1 Previous Related Safety Recommendations

1.9.1.1 Surface Detection Technology

The NTSB issued Safety Recommendation A -00-66 on July 6, 2000.24 The

recommendation asked the FAA to do the following:

Require, at all airports with scheduled passenger service, a ground

movement safety system that will prevent runway incursions; the system

should provide a direct warning capability to flight crews. In addition,

demonstrate through computer simulations or other means that the

system will, in fact, preven t incursions.

On August 25, 2017, the FAA stated that it was continuing to install runway

status lights at some larger airports with scheduled passenger service to “provide a

direct indication to pilots and vehicle operators when a runway is unsafe for entry,

crossing, or departure … reducing the time it takes to alert them of potentially unsafe

situations.” The FAA also stated that r unway status light systems require the

installation of ASDE -X or the a irport surface surveillance capability (ASSC) .25 On

December 1, 2017, the NTSB stated that the recommendation remained classified as

Open —Acceptable Response and expressed interest in learning a bout the solutions

under consideration for airports that would not be receiving a runway status light

system.

24 Safety Recommendation A -00-66 superseded Safety Recommendation A -91-29, which was

issued on June 12, 1991, and classified Closed —Unacceptable Action/Superseded on April 12, 2001.

Safety Recommendation A -91-29 asked the FAA to “expedite efforts to fund the development and

implementation of an operational system analogous to the airborne conflict alert system to alert

controllers to pending runway incursions at all terminal fac ilities that are scheduled to receive Airport

Surface Detection Equipment (ASDE III).” (ASDE -3 is a surface movement radar. ASDE -X and ASSC use

multilateration and ADS -B in addition to a surface movement radar.)

25 According to the FAA, ASSC “improves surface surveillance and situational awareness in all

kinds of weather ” and “is similar ” to ASDE -X in that both systems allow controllers to “see aircraft and

ground vehicles on the airport surface, and on approach and departure paths within a few miles of the

airport. ” Further, ASSC correlates flight plan information with position display s and provides

surveillance data to help prevent runway incursions. With one exception (for Joint Base Andrews in the

Air Force District of Washington ), ASSC is no longer being implemented at airports because of the cost

to install and maintain the system .

Aviation Investigation Report

AIR-24-01

29 On July 2, 2021, the FAA stated that, of the 35 US airports with ASDE -X and the

8 airports with ASSC, 20 airports had runway status lights. The FAA also stated that its

“Runway Incursion Prevention through Situational Awareness” initiative was working

to identify different types of potential surveillance system technologies, including

low-cost solutions that are suitable for smaller airports. Additionally, the FAA stated

that it was further exploring technologies, including ADS -B In applications that could

provide situational awareness and direct alerts to pilots; however, the avionics to

implement such a system did not yet exist. On March 24, 2022, t he NTSB stated that it

was encouraged to learn that the FAA was exploring viable and affordable

technologies that c ould d irectly warn flight crews about potential runway incursions

given that ASDE -X, ASSC, and runway status light systems might not be practical at all

airports . Safety Recommendation A -00-66 remained classified O pen—Acceptable

Response.

On October 13, 2023, the FAA reported that in June 2023 it hosted an industry

event to determine how to enhance controller situational awareness of an airport

surface. According to the FAA, the goal of the event was “to gain a better

understanding of comme rcially available, production -ready, cost -effective industry

solutions and capabilities that could enhance surface safety.” As a result of the event,

the following three surface technology “sprints” were initiated:

•The Surface Awareness Initiative (SAI) will provide surface traffic

displays to towers at airports that do not currently have a surface

surveillance system. This technology will improve controller

situational awareness. This sprint effort began in June 202 3.

•The Approach Runway Verification … is an aural and [a] visual

alert to the air traffic controllers of aircraft alignments which will

aid in the prevention of wrong runway landings. This sprint effort

began in September 2023.

•The Runway Incursion Device … is a memory aid device used by

controllers for occupied and closed runways. [The device ]

provides a visual and [an] aural alert to controllers when a runway

is not available for departing or landing aircraft. This sprint effort

began in September 2023.26

On February 20, 2024, the NTSB reiterated that the intent of Safety

Recommendation A-00-66 was to provide flight crews with a direct warning of a

potential incursion, and that the technologies described in the FAA’s October 2023

26 An FAA presentation from the NTSB’s May 2023 roundtable on runway incursion s showed

that the FAA planned to deploy this new technology between 2025 and 2028.

Aviation Investigation Report

AIR-24-01

30 letter would only assist controllers. Additionally, although the NTSB recognized the

FAA’s efforts to install ground surveillance and safety systems such as ASDE -X and

ASSC at some large airports, it remained unclear how the FAA planned to address

those airports without ground surveillance systems , particularly medium and small

airports with scheduled passenger service . The NTSB classified Safety

Recommendation A -00-66 as Open —Unacceptable Response.

1.9.1.2 Recording Duration of Cockpit Voice Recorders

On October 10, 2018, the NTSB issued Safety Recommendations A -18-30 and

-31 as a result of the NTSB’s experience with investigations that lacked access to

relevant CVR data; these investigations included a July 2017 taxiway overflight event

in San Francisco, California, that presented significant safety issues (NTSB 2018a,

2018b) . The safety recommendations asked the FAA to take the following actions:

Require all newly manufactured airplanes 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)

By January 1, 2024, require retrofit of all cockpit voice recorders (CVR)

on all airplanes re quired to carry both a CVR and a flight data recorder

with a CVR capable of recording the last 25 hours of audio. (A-18-31)

On May 23, 2023, the FAA stated that it had initiated rulemaking that

propose d that newly manufactured aircraft be equipped with CVRs capable of

recording 25 hours of data . Also, the FAA reported that it was launching an aviation

rulemaking committee for the US aviation community to “discuss, prioritize, and

provide recommendations to the FAA concerning requirements for the installation of

existing, new, and upgraded investigative technologies that affect applicable

airworthiness standards and operating rules. ”

On November 8, 2023, the NTSB responded that the rulemaking would

address the intent of Safety Recommendation A -18-30 and classified the

recommendation Open —Acceptable Response. The NTSB expressed concern that the

rulemaking would not address the retrofit of all airplanes required to have a CVR and

an FDR , as discussed in Safety Recommendation A -18-31, and classified the

recommendation Open —Unacceptable Response.

On December 4, 2023, the FAA published notice of proposed rulemaking

(NPRM) “25 Hour Cockpit Voice Recorder Requirement, New Aircraft Production,”

which proposed requiring the installation of 25 -hour duration CVRs only on newly

manufactured aircraft that require a CVR operating under CFR Parts 91, 121,125, and

135. On January 31, 2024, the NTSB commented that it did not fully support the

NPRM because it did not propose a similar requirement to retrofit existing airplanes

Aviation Investigation Report

AIR-24-01

31 required to carry a CVR and an FDR. The NTSB said the FAA had inappropriately

estimated the cost of retrofitting the existing fleet. The NTSB determined the retrofit

requirement recommended in A -18-31 would apply to less than half the number of

airplanes the FAA had estimated and could be co mpleted during regular CVR

maintenance. The NTSB urged the FAA to reconsider its position and issue a final rule

applicable to both newly manufactured airplanes that must have a CVR as well as

existing airplanes required to carry both a CVR and an FDR.

At the time that Safety Recommendations A -18-30 and -31 were issued in

October 2018 , at least 14 investigations were hampered by the lack of CVR data that

were overwritten . Since the issuance of Safety Recommendations A -18-30 and -31, at

least 18 additional investigations were affected by overwritten CVR data, including

8 serious runway incursions that occurred in 2023 and 2024 and a rapid decompression

that occurred on an Alaska Airlines 737 -9 MAX flight on January 5, 2024 (see table) .

Table. NTSB investigations hampered by overwritten CVR data since Safety

Recommendations A -18-30 and -31 were issued .

Date NTSB case number Location Event description

April 17, 2024 DCA24FA164 Queens, New York Runway incursion

February 10, 2024 DCA24LA097 Kelsey, New York Turbulence

January 5, 2024 DCA24MA063 Portland, Oregon Rapid decompression

November 30, 2023 DCA24 LA034 Kahului, Hawaii Ground collision

September 25, 2023 DCA23LA462 Caribbean Sea Turbulence

August 11, 2023 OPS23FA010 San Diego, California Runway incursion

(delayed notification)

February 27, 2023 DCA23LA192 Boston, Massachusetts Runway incursion

(delayed notification)

February 22, 2023 DCA23LA185 Burbank, California Runway incursion

February 16, 2023 DCA23LA179 Sarasota, Florida Runway incursion

(delayed notification)

February 4, 2023 DCA23FA149 Austin, Texas Runway incursion

January 23, 2023 DCA23LA133 Honolulu, Hawaii Runway incursion

January 13, 2023 DCA23LA125 New York, New York Runway incursion

August 6, 2022 DCA22LA178 Atlanta, Georgia Hard landing

July 7, 2022 WPR22LA284 San Francisco, California Loss of control in flight

February 15, 2020 ENG20LA016 Sacramento, California Electrical system

malfunction

December 18, 2019 DCA20CA043 Disputanta, Virginia Turbulence encounter

November 6, 2019 DCA20IA014 Atlanta, Georgia Loss of control in flight

June 15, 2019 DCA19CA167 Newark, New Jersey Hard landing

On May 16, 2024, t he FAA Reauthorization Act of 2024 became law. The act

included new 25 -hour CVR carriage requirements for newly manufactured airplanes

and a retrofit for exi sting airplanes within 1 and 6 years, respectively. The new

requirements applied to airplanes operated by a Part 121 air carrier and other

transport category airplanes type certificated with a passenger seating capacity of 30

Aviation Investigation Report

AIR-24-01

32 or more, or all -cargo or c argo derivatives of such airplanes operated under other

parts. The act required that the FAA update applicable regulations for the 25 -hour

CVR retrofit requirement for existing airplanes within 3 year s.

Aviation Investigation Report

AIR-24-01

33 2. Analysis

2.1 Introduction

This incident occurred when the crew of AAL106 deviated from the taxi

instructions issued by the ground controller and mistakenly crossed runway 4L, where

the crew of DAL 1943 had just begun their takeoff roll. The ASDE -X system installed at

the airport alerted the local controller to the conflict, the local controller cancelled

DAL 1943’s takeoff clearance , and t he crew aborted their takeoff. The DAL crew

initiated rejected takeoff procedures about 2,500 ft from AAL106 . AAL106 continued

across the runway, an d DAL1943 came to a stop before taxiing the airplane off the

runway onto taxiway J .

The following analysis summarizes the incident (section 2.2) and evaluates the

following:

• the need for additional risk mitigation strategies to reduce the likelihood

that flight crew surface navigation errors result in runway incursions

(section 2.3);

• the need for a procedural crosscheck that requires flight crews to verbalize

the number of a runway they are about to cross as indicated by runway

signs (section 2.3) ;

• the lack of flight deck technology to detect potential traffic conflicts on a

taxiway, runway surface, or on approach to land (section 2.3) ; and

• the need for CVRs with a 25 -hour recording capability (section 2. 6).

After completing a comprehensive review of the circumstances that led to this

incident, the investigation established that the following factors did not co ntribute to

the cause of the incident:

• Pilot and controller qualifications: The AAL and DAL flight crew s were

certificated and qualified in accordance with federal regulations. All air

traffic controllers assigned to positions in the JFK ATCT were certified and

qualified in accordance with federal regulations.

• Flight crew fatigue: The AAL captain, FO, and FB had adequate sleep

opportunities during the days before the incident flight . Each crewmember

reported being rested and was observed to be alert on the evening of the

incident .

Aviation Investigation Report

AIR-24-01

34 • Air traffic control tower staffing: JFK tower was staffed with seven

controllers staffing the eight core control positions. Some positions were

combined due to the lower traffic level at the time of the incident. This

staffing was normal and adequate for the time of night, complexity, and

volume of traffic.

Thus, the NTS B concludes that none of the following were factors in this

incident: (1) pilot and controller qualifications, (2) flight crew fatigue, and (3) ATCT

staffing.

2.2 Incident Sequence

On the evening of the incident, the flight crew of AAL106 was preparing to taxi

from gate 18 at JFK to runway 31L. Most of the flight’s paperwork, as well as the initial

taxi briefing conducted by the flight crew, anticipated a departure from runway 31L;

however, upon contacting the ground controller following pushback , the crew

received instructions to taxi to runway 4L . The captain and FO then conducted a

revised briefing for a taxi to runway 4L.

2.2.1 American Airlines Flight Crew’s Deviation from Taxi

Clearance

2.2.1.1 Prospective Memory Errors

In the postincident interview, the FO and captain stated that the initial taxi

route briefing for the anticipated departure from runway 31L included taxiway TA, left

on B, and hold short at J , indicating that the crew would receive a subsequent

clearance to proceed onto taxiway J and cross runway 4L on the way to runway 31L .

The captain stated that this was the standard route from the American Airlines gates

to runway 31L. The FO recalled that, after learning the flight would depart from

runway 4L, the captain’s revised taxi briefing stated, “Same taxi route up until Bravo

short of Kilo.” The captain recalled briefing the FO and FB that the controller would

“have you come up Kilo, and then as you get closer, they’ll tell you when to cross,”

implying t wo segments to the taxi route, with two associated clearances from the

ground controller. The captain stated that this was the typical route from the

American Airlines gates to runway 4L.

Human factors research indicates that, t hrough extensive practice, experts

learn to perform routine tasks, like taxiing an airplane, automatically (Wickens &

McCarley 2008) . This reduces attentional requirements and frees cognitive resources

for other tasks. Psychological theory suggests that automatic performance is

governed by sensory -motor knowledge structures known as schemas (Cooper 2016).

When a goal is identified, a parent schema is established (Norman 1981). When

Aviation Investigation Report

AIR-24-01

35 taxiing an airplane, a pilot’s intended destination can be considered a parent schema.

Subordinate schemas would address segments of the route, and lower -level schemas

would govern psychomotor tasks, like controlling speed and steering the airplane

along t he taxiway centerline. Although schema -driven action requires little conscious

attention, attention is required at key points to ensure that appropriate subordinate

or lower -level schemas are activated at appropriate times.

In this case, the captain’s goal was to taxi to runway 4L along the typical route ,

which was highly familiar to him . After AAL106 departed the gate, the taxi task

required little of the captain’s conscious attention for several minutes ; the captain

merely had to steer the airplane along the lighted centerline of taxiway B and

maintain an appropriate speed. When the airplane reached taxiway K , however , the

captain needed to bring his attention back to the taxi task, verify that AAL106 had

received a clearanc e to cross runway 31L, turn right on taxiway K, and cross

runway 31L.

The crew received a clearance to cross runway 31L when the airplane was

about 4 minutes from arriving at taxiway K . The FO acknowledged the clearance on

the radio. The captain told investigators he believed he heard and understood this

clearance . Thus, he knew that he no longer needed to hold short of taxiway K and

await clearance to turn right on taxiway K ; however, he still needed to remember to

turn right at that location where the standard routes to runways 4L and 31L diverged.

Psychologists use the term prospective memory to refer to the set of cognitive

processes that govern the maintenance, retrieval, and execution of deferred tasks

(Dismukes 2012). The captain’s omission of the right turn at taxiway K, despite his

intention to do so, was a prospective memory error . A delay between the formation

of an intention to do something and the execution of that task , which occurred in this

case , increases the likelihood of a prospective memory error. This occur s because

intentions may b e overwritten by the cognitive demands of intervening tasks

(Wilson et al. 2020) .

Factors associated with prospective memory errors include interruptions,

multitasking, and the absence of salient cues that would normally prompt the

performance of a habitual task. (Loukopoulos, Dismukes, and Barshi 2009). The NTSB

explored the possible role of each in the incident .

The AAL106 crew was performing multiple tasks during the taxi. The timing of

some of these tasks was determined by the crew , but t he timing of others was

Aviation Investigation Report

AIR-24-01

36 determined by external events .27 This mix of tasks with expected and unexpected

timing led to interruptions and reprioritization and interleaving of tasks, otherwise

known as multitasking. In postincident interviews, all three crewmembers described

an increasing workload as the taxi progressed. The captain attributed his error, in

part, to this factor.

The instruction provided to the crew to cross runway 31L on taxiway K could

have served as a salient cue reminding the captain to turn right at taxiway K if it had

been issued closer to that intersection ; however, because this clearance was issued

4 minutes before , visual indications were the primary cues of the airplane’s position

as the airplane approached taxiway K .

The night conditions at the time of the incident would have reduced positional

cues . The airplane had passed through 17 taxiway intersections before reaching

taxiway K, so the mere presence of an intersection was not sufficient to indicate arrival

at a decision point. There was a surface painted taxiway K sign and an arrow pointing

toward taxiway K located just before the intersection. This sign would have been

visible for at least a few seconds when illuminated by the airplane’s taxi lights. There

was also a lighted yellow sign on the right side of taxiway K labeled “4L ,” which

pointed the way to that runway. This sign would have been viewable as the airplane

approached the intersection with taxiway K; howev er, it did not explicitly label

taxiway K. Thus, two cues were available, but one was only temporarily visible, and

the other required inference to determine that it denoted the intersection of

taxiways B and K.

ACARS records indicate d that the crew accepted the airplane’s zero fuel

weight and center of gravity in the FMC about 30 seconds before the airplane

crossed taxiway K. Th e receipt of this information led to a cascade of other crew

activities as the captain and FO reviewed and crosschecked takeoff performance

calculations . These tasks , triggered by the unpredictable timing of the load closeout

receipt , required at least some of the captain’s visual attention and likely reduced his

attention to and visual monitoring of the external environment. Thus, although salient

cues to the airplane’s position were present and could have reminded the captain of

the need to turn right at taxiway K, they were not sufficient to break through the

distraction caused by the crew’s load closeout -related tasks.

27 The expected tasks included obtaining the load closeout, performing follow -on tasks related

to the load closeout, communicating with ATC, addressing the passengers about the departure,

communicating with the cabin crewmembers, and performing the before ta keoff checklist. The

unexpected tasks included dealing with multiple weather advisories from dispatch via ACARS and

informing crewmembers accordingly. The tasks with timing determined by external events included

those tasks triggered by the receipt of weat her advisories and tasks triggered by receipt of the load

closeout, such as crosschecking final weights and V -speeds.

Aviation Investigation Report

AIR-24-01

37 An alternative framework from the scientific literature on human error involves

branching error , which can occur when a task involves an action schema that is

associated with more than one possible end goal. In such situations, periodic

attentional checks are needed to detect one’s arrival at a choice point. This is

consistent with the previous discussion about the captain’s prospective memory

error. In a branching error , this check does not occur, behavior becomes

disconnected from intention, and actions b eyond the choice point become driven by

the schemas that are commonly performed in the new situational context. This latter

form of a skill -based action slip is known as environmental capture (Reason 1990).

According to this framework, lower -level schemas caused the captain to

continue the taxi along taxiway B after he missed the turn at taxiway K. Shortly

thereafter, the airplane was directly facing taxiway J and the highly salient flashing

yellow guard ligh ts for runway 4L. A typical behavior pattern performed in that

location involved taxiing onto taxiway J and crossing runway 4L (the standard route to

runway 31L). In the absence of focused attention, schema theory suggests that the

captain’s actions may ha ve become disconnected from his intentions and situational

cues could have prompted him to continue along a common , but incorrect , route.

This would explain the captain’s subsequent surprise and confusion when the ground

controller alerted him to the discrepancy between AAL106’s clearance and the

airplane’s position. Thus, a branching error appears to be a good fit for the captain’s

continuation of the taxi onto taxiway J and across runway 4L.

The NTS B concludes that the captain’s deviation from AAL106’s taxi clearance

likely resulted from several factors, including an early clearance to cross runway 31L,

interruptions and multitasking related to the crew’s delayed receipt of the load

closeout, the cap tain’s prospective memory error in forgetting to turn right at taxiway

K, and environmental capture, which prompted the captain to proceed along a

familiar, but incorrect, route.

2.2.1.2 American Airlines FO and FB’s No ndetection of the Captain’s Error

The NTSB sought to understand why neither the FO nor the FB detected the

captain’s error between the time the airplane passed taxiway K until the RELs

activated. The FO told investigators she was head -down dealing with tasks related to

the load closeout for the approximate 1.5 minutes from the time the load closeout

was received until the captain began turning onto taxiway J and toward runway 4L.

The missed turn at taxiway K occurred during this time that the FO’s visual attention

was likely focused inside the flight deck .

The FO recalled hearing the captain call out , “cleared to cross” as the captain

was turning the airplane toward a runway (onto taxiway J) . Based on ADS -B data, this

turn occurred about 2044:29. Upon hearing the captain’s callout , she looked up and

visually inspected the runway as the captain did the same , in accordance with the AA L

Aviation Investigation Report

AIR-24-01

38 FOM . The captain and FO then made the callouts “clear left” and “clear right ,”

respectively . Both pilots’ interviews indicated that this visual inspection was

completed before the RELs illuminated (at 2044:40 ). Thus, the FO’s visual inspection

occurred between 2044:29 to 2044:40, a period of about 11 seconds , while the

airplane was traveling at a speed of 11 kts toward the runway .

Two types of runway signs were available in the FO’s field of view when she

scanned the runway: a pair of white -on-red signs painted on the surface of taxiway J

before the holding position markings, and a pair of vertical black -on-yellow lighted

signs erected on either side of the runway hold short position markings. The painted

signs were likely viewable for about 4 seconds and the lighted signs were likely

viewable for an overlapping period of 8 seconds. One or the other type of sign was

likely viewable for a period of about 10 seconds. Ten seconds is sufficient for locating

and reading approaching runway sign s if a viewer is specifically trying to do so ;

however, the FO either did not read the signs or did not notice that the numbers on

the runway signs conflicted with the AAL106’s taxi clearance and that the airplane was

about to cross the wrong runway. Psychological research indicates that people only

perceive and remember fine details of a visual scene that receive focused attention. It

is likely that the FO did not notice the lettering on the signs because she did not

focus visual attention on the signs.

Visual attention is highly influenced by expectancy (Downing 1988) . The FO

knew the captain was very familiar with JFK ; therefore, she likely expected that the

odds of his making a surface navigation error were quite low. Also, her attention was

divided between crosschecking the takeoff performance numbers and visually

clearing the runway , reduc ing the amount of time she devoted to inspecting the

external environment. It is likely that, due to her divided attention , she performed a

cursory visual scan of th e runway environment and made the associated routine

callout. Thus, workload, multitasking, time pressure, and expectancy likely played a

role in the FO not detecting the captain’s surface navigation error s between the time

the airplane passed taxiway K and the time the airplane taxied onto runway 4L.

The FB recalled that his visual attention was focused inside the flight deck from

the time the crew received the load closeout until the airplane crossed the holding

position markings and entered runway 4L. During that period, he recalled handing

the printed load closeout to the other crewmembers, cleaning up paperwork on the

cent er pedestal, and adjusting his radios. He recalled hearing some confusion about

crossing a runway and looking up , at which time the airplane was across the runway

4L hold short markings . As a result, he would no longer have been able to see the

signs for runway 4L. Thus, the FB did not detect the captain’s deviation from

AAL106’s taxi clearance because he was performing operational tasks other than

looking outside and monitoring the airpl ane’s position.

Aviation Investigation Report

AIR-24-01

39 The NTSB concludes that the FO and FB were likely distracted from their

primary duty of assisting the captain in safely taxiing the airplane by other operational

activities, which resulted in the crew’s loss of situational awareness during a critical

phase of flight.

2.2.1.3 American Airlines Crew’s Non detection of DAL1943 on Runway 4L

DAL1943 was positioned for takeoff on runway 4L for about 2 minutes before

beginning its takeoff roll , and began its takeoff roll about the time AAL106 started

turning onto taxiway J. As discussed in the previous section, the AAL106 captain and

FO recalled visually scanning the runway, but neither recalled seeing DAL1943 on the

runway.

The captain and FO recalled scanning the runway at some point during the

11 seconds between the time the captain began turning the airplane onto taxiway J

and the time the airplane crossed the hold short markings and the RELs activated .

The crew recalled that the captain also switched on some additional exterior lights

during this time, which briefly required some of his visual attention inside the cockpit.

As the airplane began to turn onto and align with taxiway J, the approach end of

runway 4L was to the captain’s right and may have been partially obstructed by the

FO’s body and aircraft structure. Thus, it is possible the captain would not have had

an opportunity to see DAL1943 due to his obstructed view.

Examination of satellite images , illumination diagrams, and on -scene nighttime

photographs of the airport suggest that a significant amount of airport lighting was

present between the two airplanes in various colors and intensities. Due to flattening

of perspective that occurs when viewing distant objects on flat terrain, DAL1943 ,

which was around 3,000 ft away from AAL106’s position, likely appeared as a small

target within a dense array of surrounding lights that varied in color and brightness ,

including flashing yellow lights at each taxiway/runway intersection; blue taxiway

edge lights; green taxiway centerline lights; green and yellow taxiway lead -on lights;

yellow lighted taxiway signs; and red, white, and yellow edge lights for runway 4L .

These surrounding lights would have been distractor s during a visual search, making

a target more difficult to identify.

Additionally, DAL1943 was moving directly toward AAL106, and would have

had little apparent motion, making it less likely to attract focused visual inspection.

Given the fact that the FO did not see DAL1943 before entering the runway, it is likely

that she never fixated directly on the airplane. Thus, the FO’s nondetection of

DAL1943 likely resulted from limitations of the human visual system for detecting

small stationary targets in a complex visual scene in a limited time.

Even if the crew had noticed the presence of DAL1943, it might have been

difficult for them to determine whether the airplane was on runway 4L until after

Aviation Investigation Report

AIR-24-01

40 AAL106 had entered runway 4L and its crew could look straight down the runway. In

fact, this is when the FB first recalled seeing DAL1943. It is likely that t he forward -

facing white taxi/landing lights were the most salient external lights on DAL1943

when viewed from the perspective of AAL106.

Another potential explanatory factor is expectation bias , a cognitive bias that

reduces the likelihood a person will notice a mismatch between what they expect to

see and the actual state of things. In this event, the crew believed that the controller

had cleared them to cross the runway in front of them. The AAL106 captain and FO

had likely performed thousands of visual scans of runways in the past , and it is likely

that most of these scans revealed no conflicting aircraft. Therefore, the crew probably

did not expect another airplane to be on the runway and moving toward them , which

made them vulnerable to inspecting the runway and making related callouts in rote

fashion rather than systematically searching for a small target. Thus, expectation bias

could explain how the captain and FO did not observe DAL1943 before entering

runway 4L.

The NT SB concludes that the American Airlines flight crew’s nondetection of

DAL1943 on runway 4L likely resulted from night conditions, the location of DAL1943

within a complex array of airport lights, the lack of relative motion of DAL1943 in the

visual field of the crew of AAL106, the distance between the two airplanes, the brief

time devoted to visually scanning the runway, and expectation bias.

2.3 Operational Barriers to Runway Incursions

2.3.1 Crew Resource Management

Crew resource management (CRM) refers to the effective use of all available

resources to ensure flight safety. The FAA’s most recent guidance on CRM training ,

Advisory Circular (AC) 120 -51E, outlines desired crew competencies in the areas of

(a) communication processes and decision behavior, (b) team building and

maintenance, and (c) workload management situation al awareness. Lacking a cockpit

voice recording, it was difficult for the NTSB to perform a detailed assessment of the

AAL106 crew’s CRM during this incident. Any insights that can be identif ied had to be

inferred from the fact that error s occurred and from crew interviews.

FAA guidance indicates that briefings (which are primarily the captain’s

responsibility) should establish a team concept, establish open and interactive

communications, identify potential problem areas and safety issues, provide

guidelines for crew actions centered on standard operating procedures (SOPs), and

address division of labor and crew workload.

Aviation Investigation Report

AIR-24-01

41 According to crew interviews, the captain performed briefings, and the crew

appeared to get along and adopt a team concept. The captain and the FO reported

conducting two departure and taxi briefings. These occurred during a low workload

period, which reflected good planning ; however, the thoroughness of the taxi

briefings is difficult to discern based on the limited information available from the

crew interviews. Although in hindsight we know that an error occurred, it is

understandable that the captain ma y not have been particularly concerned in

advance , because the route was highly familiar to him and was not complex.

With respect to team building and maintenance behaviors , FAA guidance

addresses leadership/followership and crew concern for the effective

accomplishment of tasks. Specifically, this area of competency includes , among

others, ensuring that all available resources are used to accomplish the job at hand,

that actions are decisive when the situation requires, that the time available for tasks

is well managed, and that demands on resources posed by operation of automated

systems are recognized and managed.

Crew interviews indicate d that the captain divided tasks among all three

crewmembers to use available resources, and interviews suggest ed that all

crewmembers were focused on effective accomplishment of their tasks. In particular,

the captain’s attention was focused on ensuring that there was no issue with the

hand -written notation on their dangerous goods form that would result in a problem

with their cargo or load closeout ; the FO was highly focused on relaying advisories

about potentially hazardous weather conditions and performing tasks related to the

load closeout ; and the FB was actively supporting the other two pilots as requested.

However, when the load closeout arrived , the captain and the FO’s workload

increased. Neither the captain nor the other crew members appeared to recognize

how the demand on mental resources imposed by the review of takeoff performance

data from the load closeout and related use of automated systems (electronic flight

bag and FMS) could impact their monitoring and awareness of airplane position.

Consequently, the crew did not proactively manage this threat to their awareness of

the airplane’s position. Actio ns the c rew could have taken to mitigate the effect of this

additional workload included delaying the completion of load closeout until the

airplane could be parked or until all runways had been crossed , or by directing the

FB to closely monitor the progress of the taxi while the captain’s attention was

divided.

In terms of workload management and situation al awareness behaviors , FAA

CRM guidance includes anticipating contingencies and related actions that will be

required in the future, active monitoring, and avoiding “tunnel vision .” In this incident,

the captain was not adequately alert to the risk to monitoring posed by the

concurrent performance of the load closeout -related tasks. Further, all three

crew members probably thought it unlikely that the captain would make a surface

Aviation Investigation Report

AIR-24-01

42 navigation error at an airport with which he was so familiar; however, high levels of

safety require a preoccupation with failure and consistent guarding against high -

consequence threats, even those that are unlikely.

FAA guidance states that CRM behaviors should be exercised in the framework

of a comprehensive set of SOPs . The crew’s monitoring errors suggest that the

operator’s monitoring SOPs , which are intended to prevent such errors , are

inadequate. For example, a lthough American Airlines SOPs required the crew to

perform a visual inspection of the runway before crossing , the procedures did not

require the crew to verbally crosscheck the runway number displayed on the runway

signs , although this practice is recommen ded by the FAA .28 Noticing those signs was

the last opportunity for the crew to detect and correct the captain’s surface navigation

error before the incident occurred.

Automated systems that provide aural advisories about the number of an

approaching runway have been developed but were not installed on AAL106. In this

case, it is possible that reading the runway number aloud could have alerted the

captain to the fact tha t he was not crossing the intended runway. The FO and FB told

investigators that they recalled AAL106’s taxi clearance throughout the taxi ;

therefore , a verbal crosscheck of the runway number could have alerted them to the

fact that the airplane was about to cross the wrong runway. Such a procedure would

be easy to perform and similar to other common safety -related checks.

The NTSB concludes that a procedural crosscheck that requires a flight crew to

verbalize the number of a runway they are about to cross, as indicated by runway

signs, would reduce the likelihood of future runway incursions resulting from flight

crew surface navigation errors . Therefore, the NTSB recommends that the FAA

encourage 14 CFR Part 91K, 135, and 121 operators to incorporate into their SOPs a

procedural crosscheck that requires flight crew s to verbalize the number of a runway

they are about to cross , as indicated by runway signs, unless an installed automated

system already provides an aural advisory.

2.3.2 American Airlines Policy on Completing the Load Closeout

The captain was ultimately responsible for verifying that the load closeout was

acceptable before takeoff. The NTSB evaluated policies and procedures in the

American Airlines FOM on crew performance of tasks related to the load closeout.

28 FAA Advisory Circular 120 -74B, Parts 91, 121, 125, and 135 Flightcrew Procedures During

Taxi Operations, recommends that when approaching any hold short line, the flight crew verbalize

“approaching (specific runway number) hold short line.”

Aviation Investigation Report

AIR-24-01

43 The FOM stated that required tasks included verifying the received messages

and comparing the TPS tablet application output and load closeout numbers against

documents obtained from operations for reasonableness. This was essential, in part,

to ensure that the crew was using the appropriate reference speeds during the

takeoff. The crew was also required to review the load closeout to determine if any

restricted articles were onboard, in which case they were to ensure that they had a

corresponding Notice to Captain (NOTOC). Any discrepancies were required to be

corrected before takeoff. After the incident, American Airlines confirmed that tasks

related to the load closeout are considered essential to the safety of flight and are

part of required pre -takeoff checks that can be accomplished during taxi.

In the postincident interview, the captain said that he did not normally need to

perform these tasks during taxi because load closeout s typically arrive soon after the

airplane pushe s back from the gate. He said that he normally made a point of

crosschecking the load closeout with the FO and completing related tasks before

departing the ramp area. In this case, however, he said that he decided to begin the

taxi before receiving the loa d closeout because AAL106 had a third flight

crewmember (the FB). He explained that many of the flights he operated from JFK to

London had only two flight crewmembers , and he believed the presence of the FB

would allow the crew to handle more tasks during the taxi while maintaining a

manageable worklo ad.

The captain delegated to the FB the duty of following up with dispatch to

ensure that the load closeout was received ; however, the captain could not delegate

crosschecking the takeoff performance numbers with the FO. As previously

discussed, load closeout -related tasks subsequently played a role in the captain’s

distraction and surface navigation error.

The American Airlines FOM stated that flight crewmembers should use a

continuous loop process for actively monitoring and updating their progress and

location during taxi. This includes knowing the aircraft’s present location and mentally

calculating the next location on the taxi route that will require increased attention.

Although monitoring is important, a pilot cannot continuously focus on

monitoring one task while simultaneously engaging in another concurrent task,

particularly one that shares the same sensory channel ( that is , vision). As previously

discussed, this requires multitasking, which can lead to errors. The FOM also state d

that other flight deck duties and non -ATC communications should not divert attention

from the safe movement of the aircraft, especially at critical times, such as runway

crossings and transit ioning through complex taxiway intersections. However, any time

concurrent tasks are performed while taxiing, some conscious attention is diverted

away from the surface navigation task.

Aviation Investigation Report

AIR-24-01

44 Airline crews strive for efficiency and maintaining on -time performance;

therefore, it is understandable that the crew of AAL106 opted to perform load

closeout tasks during taxi. Doing so was not prohibited because these tasks are

essential to the safety of the flight , and the crew likely believed it would help them

avoid a delay. If American Airlines guidance had specified that the captain and FO

had to complete load closeout tasks before beginning the taxi, the crew likely would

have completed these tasks in the ramp area. Alternatively, if such guidance stated

that load closeout -related tasks could only be performed when the airplane was

parked, the crew might have delayed action on those tasks until they reached runway

4L for takeoff and parked the airplane, thereby increasing their external monitoring

while the airplane was moving. Thus, a lack of guidance prohibiting the performance

of load closeout related tasks during taxi increases the risk that a similar surface

navigation error may occur in the fut ure.

This incident demonstrates how interruptions and multitasking can reduce

crew awareness and increase the risk of surface navigation errors, runway incursions,

and accidents. Part 121 airlines are required to have safety management systems

(SMS) that identi fy and manage risks to their operations. SMS is intended to analyze

the risk of such errors and develop appropriate risk mitigation strategies. The NTSB

concludes that additional risk mitigation strategies as part of an operator’s SMS

would reduce the like lihood that flight crew performance of concurrent t asks during

taxi will lead to inaccurate navigation on the airport by reducing distractions

associated with multitasking .

While it would be ideal to prohibit conducting load closeout tasks during taxi

to ensure appropriate focus on taxi operations, we recognize that it is not feasible to

issue such a recommendation . Some level of multitasking can be performed if pilots

are aware of the risks and take steps to reduce errors, such as calling out runway

crossings, as noted above. Therefore, the NTSB recommends that the FAA

encourage Part 121 operators to use SMS 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 .

2.3.3 Flight Deck Technologies for Preventing Runway Incursions

The NTSB has long advocated the development of technologies for the

prevention of runway incursions. The NTSB has, for example, recommended systems

to alert air traffic controllers to the presence of traffic conflicts. This resulted in the

development of A SDE -X. The ASDE -X installation at JFK likely prevented this incident

from becoming an accident ; however, it required ad equate time for the controller to

detect the warning, identify the problem, and determine the necessary action before

attempting to contact the DAL1943 flight crew . Moreover, the FAA is no longer

Aviation Investigation Report

AIR-24-01

45 installing ASDE -X at US airports because it considers the system too expensive to

install and maintain. Therefore, many airports currently lacking this technology will

not have access to it in the future.

In 2000, the NTSB recommended that the FAA require, at all airports with

scheduled passenger service, a ground movement safety system that would prevent

runway incursions, with the capability of providing a direct warning to flight crews

(Safety R ecommendation A -00-66). In the accompanying letter, the NTSB suggested

that these direct warnings could be provided through such means as RELs and stop

bars located at all runway/taxiway intersections, or by other means, such as a datalink.

This incident h ighlights the limitations of RELs (which will be discussed in section 2.4)

and suggests that crew positional awareness would be enhanced by a flight deck

display of ground traffic with both visual and aural conflict alerts .

Flight deck technology for improving flight crew awareness of surface traffic

and alerts to potential conflicts has been developed . Such systems could use ADS -B

data to display the location of traffic that is on, approaching, or departing a runway , in

the air or on the ground . This capability would provide a redundant means for flight

crew s to detect the presence of another aircraft (in addition to visually scanning the

external environment).

Such systems could aurally alert a nd visually display to flight crew s potential

traffic conflicts that could prevent them from entering a runway where another

airplane is departing , without pilots having to rely solely on visual scanning of the

outside environment or on the intervention of a controller who has detected or

received an alert about a conflict , thereby provid ing an additional layer of protection

against runway incursions by more effectively mitigating a crew surface navigation

error. It is likely that in this incident, a flight deck display of traffic information with

alerting capability would have resulted in earlier action to resolve the conflict,

because the warning would have been provided directly to the flight crews involved

rather than relayed through the controller —and may have even prevented the conflict

from occurring .

On February 4, 2023, Southwest Airlines (SWA) flight 708, a Boeing 737 -700,

and Federal Express Corporation (FedEx) flight 1432, a Boeing 767 -300, were

involved in a runway incursion at Austin -Bergstrom International Airport, Austin,

Texas. The local controller had cleared the SWA airplane for takeoff and instructed

the FedEx airplane to continue its approach to the same runway. The controller was

unable to see the SWA airplane on the taxiway and runway because of dense fog,

and the air traffic control towe r did not have surface detection equipment such as

ASDE -X. The airplanes came within 150 to 170 ft of each other after the FedEx flight

crew realized that the SWA airplane was on the runway and performed a missed

approach, thus avoiding a collision. This incident demonstrates that flight deck

Aviation Investigation Report

AIR-24-01

46 alerting systems should also provide visual and aural alerts of potential conflicts

involving aircraft on final approach.

The NTSB concludes that the implementation of a flight deck alerting system

on transport -category aircraft that provides alerts of traffic on a runway or taxiway and

traffic on approach to land would enhance safety by providing pilots with improved

situational awareness and would reduce the risk of future runway -related incidents

and accidents .

Therefore, t he NTSB recommends that the FAA collaborat e with aircraft and

avionics manufacturers and software de signers 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 . The NTSB also recommends the

FAA require that this technology be installed in all newly certificated transport -

category airplanes and that the FAA require existing transport -category airplanes be

retrofitted with this technology .

Safety Recommendations A-24-4 through –6 recommend similar actions to

Safety Recommendation A -00-66; however, they specify a flight deck -based alerting

system. The FAA has made efforts to reevaluate technologies and procedures to

reduce and eliminate runway safety risks by deploying surface systems such as

ASDE -X, ASSC, and RWSL at airports deemed highest risk, and continues to develop

surface technologies to enhance air traffic controller situational awareness.

However, after 24 years , the FAA still has not required, at all airports with

scheduled passenger service, a ground surveillance system that would provide timely

warnings directly to flight crews of the potential for a traffic conflict. Therefore, Safety

Recommendations A -24-4 through –6 are classified Open —Unacceptable Response ,

and Safety Recommendation A -00-66 is classified Closed —Unacceptable

Action/Superseded.

2.4 Airport Measures Intended to Prevent Runway Incursions

2.4.1 Runway Status Lights

The RWSL at JFK included runway entrance lights installed at the

taxiway J/runway 4L intersection. The RELs at this intersection comprised six

unidirectional lights that started at the runway hold short position and extended to

the runway centerline. The RELs would illuminate red to signal to aircraft approaching

from the taxiway that a runway was in use. According to the FAA, RELs will illuminate

on crossing taxiways when a departing aircraft has 1) reached a speed of 29.9 kts

coupled with an acceleration of 1.20 m/s2 or 2) reached a speed of 49.9 kts.

Aviation Investigation Report

AIR-24-01

47 Using ADS -B information, ASDE -X information, and RWSL data, a performance

study was undertaken to determine when the RELs activated in relation to the

positions and speeds of each air plane during the incident.

AAL106 turned onto taxiway J at 2 044:30. It passed the hold short line for

runway 4L at 2044:40 , at a groundspeed of about 11 kts , as the RELs illuminated . At

least three RELs would have been viewable by the crew given their eye position in the

flight deck and evidence indicating that RELs 2 and 3 were likely inoperative . AAL106

entered runway 4L at 2044:50 , at a groundspeed of 15 kts , and continued

accelerating as it crossed the runway. At this time, DAL1943 was 2,700 ft from the

intersection of 4L and taxiway J at 94 kts groundspeed and accelerating.

ADS -B, ASDE -X, and RWSL information indicate d that the RELs illuminated as

the airplane’s nose crossed the hold short markings , while the crew’s recollections

were that the RELs i lluminated when the airplane was farther past the hold short

markings . Although the statements of the captain and FO did not indicate that they

immediately recognized the lights as an indication that they had begun to cross an

active runway , both crew members reported that the illumination of the RELs

indicated an anomaly and prompted the captain’s decision to expedite the runway

crossing. It is likely that, had the RELs illuminated earlier, the crew would have had

more time to perceive the lights and stop the airplane in a safe area.

After their takeoff clearance was cancelled at 2044: 47 due to AA L106 entering

the runway, the crew of DAL1943 initiated a rejected takeoff at 2044:51 , reaching a

maximum groundspeed of 105 kts at 2044:53 . At this time, DAL1943 was 2,300 ft

from the taxiway J intersection before it began decelerating. AA L106 cleared

runway 4L at 2045:02 at 21 kts groundspeed when DAL1943 was 1,000 ft away at a

ground speed of 49 kts.

The NTSB requested that the FAA evaluate the operational algorithm/criteria

for the RWSL installation at JFK to ensure that it was operating as described on the

day of the incident. FAA Air Traffic Organization responded with a detailed analysis of

the inc ident and concluded that , “the JFK RWSL system [was] operating within the

standard algorithm/criteria for RWSL light activations.”

The ASDE -X replay depicted REL s 2 and 3 at the 4L/taxiway J intersection as

“extinguished” and “faulted ,” respectively, at the time of the event , and it is likely that

both REL 2 and 3 were inoperative . However, FAA RWSL system maintenance

operational tolerance s allow ed <50% of the lamps in an REL array to be inoperative ;

Aviation Investigation Report

AIR-24-01

48 therefore, even if REL s 2 and 3 were inoperative at the time of the incident , the array

still met operational tolerance limitations.29

The NTSB conclude s that the RELs operated as designed ; however , they were

ineffective in preventing the crew of AAL106 from crossing the runway 4L hold short

markings because the y activated too late for the crew to perceive them and stop the

airplane in a safe area . Therefore, the NTSB recommends that the FAA evaluate the

effectiveness of the activation logic for the RWSL considering the circumstances of

this incident . Using the findings of th e evaluation conducted in response to Safety

Recommendation A-24-7, update the RWSL activation logic as necessary to improve

system effectiveness .

2.4.2 Airport Sur face Detection Equipment , Model X

JFK airport was equipped with the ASDE -X system. ASDE -X alerts air traffic

controllers of potential runway conflicts by providing detailed coverage of movement

on runways and taxiways. Controllers are presented this information on a color

display depicting aircraft and vehicle pos itions as icons overlaid on a map of the

airport’s runways, taxiways, and airport approach corridors. The map is continuously

updated and can be particularly beneficial at night or in poor visibility conditions as

an enhancement to controller situational awareness.

The ASDE -X system is equipped with visual and aural alarms to alert controllers

of possible runway incursions or incidents. In this incident, the ASDE -X equipment

generated both aural and visual alerts when AAL106 crossed Runway 4L via taxiway J

while a departing aircraft, DAL1943, still occupied the runway. As a result, an aural

“runway occupied” alert was voiced, and the corresponding visual text alert

“RWY 4L|DAL1943, AAL106| RWY OCCUPIED ” was displayed on the air traffic

displays; the ASDE -X system functioned as designed. The local controller cancelled

DAL1943’s takeoff clearance 5 seconds after the ASDE -X aural alert was annunciated.

The NTSB concludes that the ASDE -X system at JFK was operational at the time

of the incident and functioned as designed , generating both aural and visual alerts

when AAL106 crossed runway 4L while DAL1943 was departing , and likely reduced

the severity of the incident by prevent ing a runway collision .

29 While paperwork obtained from the Port Authority of NY and NJ dated January 15, 2023,

identified RELs 3 and 4 as being inoperative, ASDE -X playback showed the inoperative fixtures were

RELs 2 and 3.

Aviation Investigation Report

AIR-24-01

49 2.5 Air Traffic Control

2.5.1 Ground Controller

The ground controller instructed the crew of AAL106 , from their position at

taxiway TA after pushing back from the gate , to taxi to runway 4L, via a left turn on to

taxiway B, and to hold short of taxiway K. As AAL106 continued to taxi on taxiway B,

the ground controller instructed the crew to cross runway 31L at taxiway K. The crew

read back the instruction correctly, which likely reinforced the ground controller’s

belief that the crew understood the taxi instructions.

The ground controller later reported in a postincident interview that he

observed AAL106 swing out to the left on taxiway B to begin what he believed to be a

typical wide right turn onto taxiway K. Radio communications showed that , as AAL106

passed taxiway K, the ground controller received a request from, and subsequently

provided taxi instructions to, another airplane. He also established communications

with an airport operations vehicle and requested that they sweep runway 31R in

preparation for the runwa y configuration change. The ground controller recalled that,

after he finished his communication with the airport operations vehicle, he looked

down to attend to another task .

While the ground controller was looking down , AAL106 continued the slight

left turn on taxiway B, then turned right onto taxiway J , where its nose would

eventually cross active runway 4L, triggering an ASDE -X alert.

The ground controller stated that he expected the flight crew of AAL106 to

adhere to the ir taxi instructions given that they correctly read back the instructions as

issued. Believing that AAL106 was conducting the taxi as instructed, he did not

ensure that the flight crew completed the right turn onto taxiway K before he began

attending to other duties . FAA Order 7110.65Z, Air Traffic Control, paragraph 2−1−2,

Duty Priority, stated that controllers were to g ive first priority to separating aircraft

and issuing safety alerts as required. Therefore, th e NTSB concludes that the ground

controller expected the AAL106 flight crew to adhere to the assigned taxi instructions

and did not detect the flight crew’s surface navigation error and subsequent turn onto

taxiwa y J because he was performing a lesser priority task that entailed looking down .

2.5.2 Local Controller

The local controller stated that he performed a visual scan before issuing the

takeoff clearance to DAL1943, then he started talking to other airplanes that were

taxiing to line up and wait, one on runway 4L and another on runway 31L.

Aviation Investigation Report

AIR-24-01

50 Shortly after the controller provid ed these instructions , the ASDE -X alerted. He

looked at the ASDE -X display and saw the aircraft involved were DAL1943 and

AAL106. The local controller cancelled DAL1943’s takeoff clearance 5 seconds after

the ASDE -X alerted. He repeated the instruction a second time, and the DAL1943

crew responded that they were rejecting the takeoff. The NTSB conclu des that the

local controller acted in a timely and appropriate manner following the ASDE -X alert

by cancelling DAL1943’s takeoff clearance.

2.5.3 Air Traffic Control Tower Team

At the time of the incident, the tower was changing the runway configuration,

which required some of the controllers to physically move into different positions in

the tower, as well as change all their equipment to the new configuration. Airport

operation s was on the airfield conducting checks and switching equipment for the

runway change as well. On the night of the incident, the runway configuration change

was requested to be completed earlier than the time it typically took place between

2130 and 2230. Although a runway configuration change is an important priority,

and, on the night of the incident, occurred at an atypical time, it is a routine event that

does not always occur at prescribed times . The whole tower team has the primary

responsibility of ensuring safe and effective operations in the tower and to perform

continual visual scans of the runways . In this case, the cab coordinator, flight

data/clearance delivery controller, and operations supervisor were busy making

preparations for the runway ch ange and were not giving their full attention to the air

traffic operations at the time.

The tower team , collectively , did not effectively prioritize their duties by

ensuring that the team maintained a continuous scan of the airport environment as

preparations for the runway change were performed . As such, they did not meet the

expectations and responsibilities as required in FAA Order 7110.65Z, Air Traffic

Control, paragraph 2−10−3, Tower Team Position Responsibilities , to give first

priority to the separation of traffic ; and paragraph 3 —1—12, to visually scan the

runways to the maximum extent possible . The NTSB concludes that the JFK ATCT

team had the responsibility of scanning the runways and airport environment but did

not effectively priorit ize their duties to ensure a continuous scan , which resulted in

their no ndetection of the AAL106 flight crew’s deviation from taxi instructions .

2.6 Cockpit Voice Recorder Duration and Loss of Information

The overwriting of the cockpit voice recording for AAL106 resulted in a loss of

useful information. As a result, the NTSB had to rely exclusively on flight crew

recollections about the incident ; however , these were not documented until 1 month

after the incident occurred . Because m emories can decay or become distorted over

Aviation Investigation Report

AIR-24-01

51 time, a cockpit voice recording would have help ed corroborate the crew’s

recollections.

CVRs are among the most valuable tools for accident investigations because

they provide contemporaneous information on flight crew intentions and

coordination as well as other factors, such as procedural compliance, workload,

fatigue, and situational aware ness (NTSB 2018b) . In this case, a cockpit voice

recording would likely have provided additional details about the content and timing

of crew communications ; shed light on the crew’s minute -by-minute focus of

attention ; revealed any unreported, nonpertinen t conversations ; and potential ly

provided additional information about any distractions . This would have allowed the

NTSB to more precisely analyze crew performance.

The NTSB recognizes that the FAA has taken action to address Safety

Recommendation A -18-30 by publishing an NPRM which proposed that newly

manufactured aircraft be equipped with CVRs capable of recording 25 hours of data.

Further, the FAA Reauthorization Act of 2024 included new 25 -hour CVR carriage

requirements for some newly manufactured and existing airplanes. However, the

rulemaking for 25 -hour CVRs for newly manufactured aircraft and existing airplanes

has not been finalized by the FAA.

The NTSB concludes that CVRs with a 25-hour recording capability are

necessary because valuable information continues to be overwritten on CVRs that are

designed to record only 2 hours of audio data. Therefore, t he NTSB reiterates Safety

Recommendation A-18-30. Because the FAA did not complete responsive action to

Safety Recommendation A -18-31 by January 1, 2024, this recommendation is

classified Closed —Unacceptable Action/Superseded. The NTSB recommends t hat the

FAA require retrofit of all CVRs on all airp lanes required to carry both a CVR and a

flight data recorder with a CVR capable of recording the last 25 hours of audio. Safety

Recommendation A-24-9 is classified Open —Unacceptable Response.

Aviation Investigation Report

AIR-24-01

52 3. Conclusions

3.1 Findings

1. None of the following were factors in this incident: (1) pilot and controller

qualifica tions, (2) flight crew fatigue, and (3) airport traffic control tower

staffing.

2. The captain’s deviation from A merican Airlines flight 106’s taxi clearance likely

resulted from several factors, including an early clearance to cross runway 31L,

interruptions and multitasking related to the crew’s delayed receipt of the load

closeout, the captain’s prospective memory error in forgetting to turn right at

taxiway K, and environmental capture, which prompted the captain to proceed

along a familiar, but incorrect, route.

3. The first officer and relief first officer were likely distract ed from their primary

duty of assisting the captain in safe ly taxiing the airplane by other operational

activities, which resulted in the crew’s loss of situation al awareness during a

critical phase of flight.

4. The American Airlines flight crew’s nondetection of D elta Air Lines flight 1943

(DAL1943) on runway 4L likely resulted from night conditions, the location of

DAL1943 within a complex array of airport lights, the distance between the two

airplanes, the lack of relative motion of DAL1943 in the visual field of the crew

of American Airlines flight 106 , and expectation bias.

5. A procedural crosscheck that requires a flight crew to verbalize the number of

a runway they are about to cross, as indicated by runway signs, would reduce

the likelihood of future runway incursions resulting from flight crew surface

navigation errors .

6. Additional risk mitigation strategies as part of an operator’s safety

management system would reduce the likelihood that flight crew performance

of concurrent tasks during taxi will lead to inaccurate navigation on the airport

by reducing distractions associated with multitasking.

7. The implementation of a flight deck alerting system on transport -category

aircraft that provides alerts of traffic on a runway or taxiway and traffic on

approach to land would enhance safety by providing pilots with improved

situational awareness and would reduce the risk of future runway -related

incidents and accidents.

8. The runway entrance lights operated as designed ; however, they were

ineffective in prevent ing the crew of American Airlines flight 106 from crossing

Aviation Investigation Report

AIR-24-01

53 the runway 4L hold short markings because they activated too late for the crew

to perceive them and stop the airplane in a safe area.

9. The airport surface detection equipment model -X system at John F. Kennedy

International Airport was operational at the time of the incident and functioned

as designed, generating both aural and visual alerts when American Airlines

flight 106 crossed runway 4L while Delta Air Lines flight 1943 was departing ,

and likely reduced the severity of the incident by preventing a runway collision .

10. The ground controller expected the American Airlines flight 106 crew to

adhere to the assigned taxi instructions and did not detect the flight crew’s

surface navigation error and subsequent turn onto taxiway J because he was

performing a lesser priority task that entailed looking down .

11. The local controller acted in a timely and appropriate manner following the

airport surface detection equipment model -X alert by cancelling Delta Air

Lines flight 1943 ’s takeoff clearance.

12. The John F. Kennedy International Airport air traffic control tower team had the

responsibility of scanning the runways and airport environment but did not

effectively prioritize their duties to ensure a continuous scan , which resulted in

their nondetection of the American Airlines flight 106 crew’s deviation from

taxi instructions .

13. Cockpit voice recorder s (CVRs) with a 25-hour recording capability are

necessary because valuable information continues to be overwritten on CVRs

that are designed to record only 2 hours of audio data .

3.2 Probable Cause

The NTSB determines that the probable cause of this incident was the

American Airlines flight 106 (AAL106 ) crew’s surface navigation error due to

distractions caused by their performance of concurrent operational tasks during taxi ,

which resulted in a loss of situational awareness . Contributing to the incident was the

air traffic control tower team’s nondetection of the AAL106 crew ’s deviation from taxi

instructions while performing concurrent operational tasks ; the timing of the runway

status light system , which activated too late to prevent the AAL106 crew from

crossing the runway hold short line ; and American Airlines’ lack of adequate risk

controls to prevent concurrent flight crew tasks from leading to distraction, loss of

situational awareness, and deviation from an authorized taxi clearance. Reducing the

severity of the incident, and likely preventing a n accident, was the activation of the

ASDE -X warning in the air traffic control tower and the local controller’ s prompt

cancellation of DAL1943’s takeoff clearance.

Aviation Investigation Report

AIR-24-01

54 4. Recommendations

4.1 New Recommendations

As a result of this investigation, the National Transportation Safety Board

makes the following new safety recommendations :

To the Federal Aviation Administration

Encourage Title 14 Code of Federal Regulations Part 91K, 135, and 121

operators to incorporate into their standard operating procedures a

procedural crosscheck that requires flight crews to verbalize the number

of a runway they are about to cross, as indicated by runway signs, unless

an installed automated system already provides an aural advisory .

(A-24-2)

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-24-3)

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)

Require that the technology developed in response to Safety

Recommend ation A-24-4 be installed in all newly certificated transport -

category airplanes . (A-24-5)

Require that e xisting transport -category airplanes be retrofitted with th e

technology developed in response to Safety R ecommend ation A-24-4.

(A-24-6)

Evaluate the effectiveness of the activation logic for the runway status

light system considering the circumstances of this incident . (A-24-7)

Using the findings of the evaluation conducted in response to Safety

Recommendation A-24-7, update the runway status light system

activation logic as necessary to improve system effectiveness. (A -24-8)

Aviation Investigation Report

AIR-24-01

55 Require retrofit of all cockpit voice recorders (CVR) on all airplanes

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)

4.2 Previously Issued Recommendations Reiterated in This Report

The National Transportation Safety Board reiterates the following safety

recommendation.

To the Federal Aviation Administration :

Require all newly manufactured airplanes 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)

4.3 Previously Issued Recommendations Classified in This Report

Safety Recommendation A -00-66 is classified Closed —Unacceptable

Action /Superseded in section 2. 3.3 of this report. This recommendation is

superseded by Safety Recommendations A-24-4 through -6, which are classified

Open —Unacceptable Response .

Safety Recommendation A -18-31 is classified Closed —Unacceptable

Action/Superseded in section 2.6 of this report. This recommendation is superseded

by Safety Recommendation A-24-9, which is classified Open —Unacceptable

Response .

BY THE NATIONAL TRANSPORTATION SAFETY BOARD

JENNIFER HOMENDY MICHAEL GRAHAM

Chair Member

ALVIN BROWN

Member J.TODD INMAN

Member

Report Date : May 29, 2024

Aviation Investigation Report

AIR-24-01

56 Appendix es

Appendix A : Investigation

The National Transportation Safety Board (NTSB) learned about this incident on

January 14 , 2023. Air traffic control and operations investigative groups were formed .

A specialist was assigned to conduct the readout of the flight data recorder at the

NTSB’s laboratory in Washington, DC, and specialists in the areas of aircraft

performance, human performance, and airports participated in the investigation.

Parties to the investigation were the Federal Aviation Administration, American

Airlines, Delta Air Lines , Allied Pilots Association, National Air Traffic Controllers

Association, and B oeing.

Aviation Investigation Report

AIR-24-01

57 Appendix B: Consolidated Recommendation Information

Title 49 United States Code 1117 (b) requires the following information on the

recommendations in this report.

For each recommendation —

(1) a brief summary of the Board’s collection and analysis of the specific

incident investigation information most relevant to the recommendation;

(2) a description of the Board’s use of external information, including studies,

reports, and experts, other than the findings of a specific accident investigation, if any

were used to inform or support the recommendation, including a brief summary of

the specific safety benefits and other effects identified by each study, report, or

expert; and

(3) a brief summary of any examples of actions taken by regulated entities

before the publication of the safety recommendation, to the extent such actions are

known to the Board, that were consistent with the recommendation.

To the Federal Aviation Administration

A-24-2

Encourage 14 CFR Part 91K, 135, and 121 operators to incorporate into

their standard operating procedures a crosscheck that requires a flight

crew to verbalize the runway number they are about to cross, as

indicated by runway signs, unless a flight deck automated system

already provides an aural advisory.

Information that addresses the requirements of 49 USC 1117(b), as applicable,

can be found in section 2.3.1, Crew Resource Management. Information supporting

(b)(1) can be found on pages 41-43; (b)(2) is not applicable ; and (b)(3) is not

applicable.

A-24-3

Encourage 14 CFR Part 121 operators to use their safety management

system to detect flight crew surface navigation errors resulting from the

performance of concurrent tasks during taxi and use scientific

knowledge from the field of human factors to develop and implement

effective risk mitigation strateg ies.

Information that addresses the requirements of 49 USC 1117(b), as applicable,

can be found in section 2.3.1, Crew Resource Management, and section 2.3.2,

American Airlines Policy on Completing the Load Closeout. Information supporting

Aviation Investigation Report

AIR-24-01

58 (b)(1) can be found on pages 41-45; (b)(2) is not applicable ; and (b)(3) is not

applicable.

A-24-4

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 .

Information that addresses the requirements of 49 USC 1117(b), as applicable,

can be found in section 1.9.1.1, Surface Detection Technology, and section 2.3.3,

Flight Deck Technologies for Preventing Runway Incursions. Information supporting

(b)(1) can be found on pages 45-47; (b)(2) is not applicable ; and (b)(3) can be found

on pages 29-31 and 45-47.

A-24-5

Require that the technology developed in response to Safety

Recommend ation A-24-4 be installed in all newly certificated transport

category airplanes .

Information that addresses the requirements of 49 USC 1117(b), as applicable,

can be found in section 1.9.1.1, Surface Detection Technology, and section 2.3.3,

Flight Deck Technologies for Preventing Runway Incursions. Information supporting

(b)(1) can be found on pages 45 -47; (b)(2) is not applicable; and (b)(3) can be found

on pages 29 -31 and 45 -47.

A-24-6

Require that e xisting transport category airplanes be retrofitted with th e

technology developed in response to Safety Recommendation A-24-4.

Information that addresses the requirements of 49 USC 1117(b), as applicable,

can be found in section 1.9.1.1, Surface Detection Technology, and section 2.3.3,

Flight Deck Technologies for Preventing Runway Incursions. Information supporting

(b)(1) can be found on pages 45 -47; (b)(2) is not applicable; and (b)(3) can be found

on pages 29 -31 and 45 -47.

A-24-7

Evaluate the effectiveness of the activation logic for the runway status

light system considering the circumstances of this incident.

Information that addresses the requirements of 49 USC 1117(b), as applicable,

can be found in section 2.4.1, Runway Status Lights. Information supporting (b)(1) can

be found on page s 47-49; (b)(2) is not applicable ; and (b)(3) is not applicable.

Aviation Investigation Report

AIR-24-01

59 A-24-8

Using the findings of the evaluation conducted in response to Safety

Recommendation A-24-7, update the runway status light system

activation logic as necessary to improve system effectiveness.

Information that addresses the requirements of 49 USC 1117(b), as applicable,

can be found in section 2.4.1, Runway Status Lights. Information supporting (b)(1) can

be found on pages 47 -49; (b)(2) is not applicable; and (b)(3) is not applicable .

A-24-9

Require retrofit of all cockpit voice recorders (CVR) on all airplanes

required to carry both a CVR and a flight data recorder with a CVR

capable of recording the last 25 hours of audio.

Information that addresses the requirements of 49 USC 1117(b), as applicable,

can be found in section s 1.1, History of Flight ; 1.6, Flight Recorders ; 1.9.1 , Previous

Related Safety Recommendations ; and 2.6, Cockpit Voice Recorder Duration and

Loss of Information. Information supporting (b)(1) can be found on page s 1-10, 20,

29-33, and 51-52; (b)(2) is not applicable ; and (b)(3) can be found on pages 29 -33

and 51 -52.

Aviation Investigation Report

AIR-24-01

60 Appendix C: Excerpts of American Airlines Flight Operations

Manual

The following excerpts from the company Flight Operations Manual (FOM)

outline procedures and pilot responsibilities for ground operations.

1g.5.4 Sterile Flightdeck Period

Source: 14 CFR 121.542; ISARP FLT 3.11.17 and 3.12.5

Definition. Critical phases of flight include all ground operations

involving taxi (when the aircraft is in motion), takeoff and landing, and all

other flight operations conducted below 10,000 feet AFL, except cruise

flight.

Do not perform or engage in duties or activities during critical phases of

flight that are not required for the safe operation of the aircraft or could

distract other crewmembers from performing their duties.

Non -Essential Activities

Do not engage in non -essential activities such as Company

communication for nonoperational related purposes, passenger

connections, pointing out sights of interest, completing paperwork,

eating meals, and non -essential conversations between flightdeck or

cabin crewmembers.

Clearance Readback

1c.2.4 Clearance Readback

Acknowledge the following ATC clearances/instructions with the full

flight number and a verbatim read back of the clearance:

— general:

• initial IFR clearance and flight plan clearance limits/amendments

• instructions to initiate contact on a specific radio frequency.

— taxi:

• taxi clearance and runway assignment

• taxi clearance involving hold short clearances restricting

runway/taxiway access

Aviation Investigation Report

AIR-24-01

61 • taxi clearance to cross any runway surface.

ACARS

1c.3.2 ACARS

Source: ISARP FLT 3.7.3

ACARS is the primary means of two -way communications between each

aircraft and dispatch when ACARS coverage is available via either

ACARS ground station or SATCOM. When away from the gate, ACARS

reports:

• normal transmission of OOOI (Out, Off, On, In) times

• enroute position reports

• in-range reports

• aircraft systems performance data

• flight -to-IOC and IOC -to-flight communications

• SELCAL used to establish dispatcher -to-captain voice communication

ACARS is intended only for Company operational communication.

Note: Non -operational messages are not authorized.

Verification

Inputs. Both the captain and first officer will verify the correct flight

number, origin, and destination are entered prior to departure.

Received Messages. Verify all messages for the correct flight number,

date, and aircraft nose number.

TPS and Load Closeout. Besides verifying these received messages (see

above), compare the TPS and load closeout against documents

obtained from operations for reasonableness (e.g., ZFW, TOGW, V

speeds). In addition, compare all subsequent TPS and/or load c loseout

messages against previously received messages.

Dangerous Goods/Hazardous Materials

1d.1.1 Dangerous Goods Definition and Classification

Aviation Investigation Report

AIR-24-01

62 Dangerous goods (DGs) are articles or substances capable of posing a

potential risk to health, safety, or property when transported by air. The

terms dangerous goods (DGs), hazardous materials (HAZMATs), and

restricted articles (RAs) are synonymous and int erchangeable.

3d.2 Notifications

3d.2.1 Notification to Captain (NOTOC)

Source: 49 CFR 175.33

The notification to captain (NOTOC) requirement is satisfied by

presenting a computer -generated AUTONOTOC form or a preprinted

OK-333 form.

Corrections Handwritten and initialed corrections to the AUTONOTOC

print -out are permitted anytime the location or quantity of dangerous

goods is changed or if the DG shipment was not loaded.

Taxi

4g.1 General

Airport surface operations require strict attention and constant

situational awareness. Sound flightdeck operating discipline enables the

flight crewmembers to properly plan taxi operations with the same level

of attention given to planning other phases of flight. For specific phase -

of-flight procedures, refer to the appropriate FOM phase of flight

chapters and the aircraft operating manual.

4g.1.1 Planning

Anticipate airport surface movements by performing a pre -taxi review

based on ATIS and previous experience at each airport. Review the

expected taxi route on the airport diagram.

4g.1.2 Control

The captain will taxi the aircraft. The flight crew’s primary task is to safely

taxi the aircraft and the flight crew’s attention should not be diverted

from this task.

Aviation Investigation Report

AIR-24-01

63 4g.1.3 Communication

See paragraphs 1c.2.4 Clearance Readback for clearances requiring a

read back and 4c.2.1 General for more information.

Coordination

Verbally coordinate all taxi instructions with other flightdeck

crewmembers to ensure common understanding. If in disagreement,

seek clarification from ATC. Do not taxi until a taxi clearance is received

and both the captain and first officer verbally coordinate and agree on

the runway assigned, any restrictions, and taxi route. Note: Ensure

received (not expected/briefed) taxi route is followed.

4g.1.4 Monitoring

Source: ISARP FLT 3.12.7

Flight crewmembers should use a continuous loop process for actively

monitoring and updating their progress and location during taxi. This

includes knowing the aircraft’s present location and mentally calculating

the next location on the taxi route that wi ll require increased attention.

Consider writing down the taxiing instructions and while taxiing:

• do not allow other flightdeck duties and non -ATC communications to

divert attention from the safe movement of the aircraft, especially at

critical times, such as runway crossings and transitioning through

complex taxiway intersections

• maintain sterile flightdeck when the aircraft is moving

Note: If uncertain as to location on the airport, taxi clear of any runway,

stop the aircraft, advise the tower.

Both pilots should:

— have the airport diagram readily available and reference it as

necessary to ensure the taxi clearance is followed correctly (see

paragraph Taxi -Out for iPad EFB guidance)

— be heads up to visually monitor the aircraft’s progress at critical

locations on the airport (hold short, crossing runways, etc.)

When approaching an entrance to an active runway, both pilots will

ensure compliance with hold short or crossing clearance by

Aviation Investigation Report

AIR-24-01

64 discontinuing non -monitoring tasks (e.g. FMS programming, ACARS,

company radio calls, etc.)

Operations During Flight

Taxi-Out

Airport Info charts must be readily available to ensure the taxi clearance

is followed correctly. Temporarily selecting another chart (e.g. ## -7, SID,

etc.) is authorized providing:

— one pilot is displaying the airport diagram

— current position is known

— aircraft is not taxiing directly toward or across an active runway

Ownship.

Note: Do not use JeppFD -Pro ownship position as a primary position

source, especially during low visibility taxi operations. Taxiway signs and

markings outside the aircraft remain the primary reference for

maneuvering the aircraft. The use of ownship posit ion display on taxi

charts is authorized to supplement situational awareness.

4g.22 Airport Markings/Lighting

Runway Entrance Lights (RELs)

The REL system is composed of flush mounted, in -pavement,

unidirectional fixtures that are parallel to and focused along the taxiway

centerline and directed toward the pilot at the hold line. An array of REL

lights include the first light at the hold line followed by a series of evenly

spaced lights to the runway edge; one additional light at the runway

centerline is in line with the last two lights before the runway edge.

When activated, the red lights indicate there is either high speed traffic

on the run way or an aircraft on final approach within the activation area.

Operating Characteristics.

Departing Aircraft When a departing aircraft reaches 30 knots, all

taxiway intersections with REL arrays along the runway ahead of the

aircraft will illuminate. As the aircraft approaches an REL equipped

taxiway intersection, the lights at that intersectio n extinguish

approximately 3 to 4 seconds before the aircraft reaches it. This allows

Aviation Investigation Report

AIR-24-01

65 controllers to apply anticipated separation to permit ATC to move traffic

more expeditiously without compromising safety. After the aircraft is

declared airborne by the system, all REL lights associated with this

runway will extinguish.

Pilot Observations and Actions.

Observations

A pilot at or approaching the hold line to a runway will observe REL

illumination and extinguishing in reaction to an aircraft or vehicle

operating on the runway, or an arriving aircraft operating less than 1

mile from the runway threshold.

Actions

Whenever a pilot observes the red lights of the REL, the pilot will

stop/remain stopped at the hold line and contact ATC for resolution if

the clearance is in conflict with the lights. Should pilots note illuminated

lights under circumstances when remainin g clear of the runway is

impractical for safety reasons (for example, aircraft is already on the

runway), the crew should proceed according to their best judgment

while understanding the illuminated lights indicate the runway is unsafe

to enter or cross. C ontact ATC at the earliest possible opportunity.

Aviation Investigation Report

AIR-24-01

66 References

Cooper, R.P. 2016. “Schema Theory and Neur opsychology. ” In M.A. Arbib & J.J.

Bonaiuto (Eds.) From Neuron to Cognition via Computational Neuroscience

(Computational Neuroscience Series). Cambridge, MA: The MIT Press.

Dismukes, K. 2012. “Prospective Memory in Workplace and Everyday Sit uations. ”

Current Directions in Psychological Science, 21(4) 215 –220.

Downing, CJ. 1988. “Expectancy and Visual -Spatial Attention: Effe cts on Perceptual

Quality .” Journal of Experimental Psychology and Human Performance,

14(2):188 -202.

FAA (Federal Aviation Administration). 2004. Advisory Circular (AC) 120 -51E, Crew

Resource Management Training . Washington, DC: FAA.

———. 2012. Advisory Circular 120 -74B, Parts 91, 121, 125, and 135 Flightcrew

Procedures During Taxi Operations . Washington, DC: FAA.

———. 2014. Order 6850.38, Maintenance of the Runway Status Lights (RWSL) System .

Washington, DC: FAA.

———. 2015. Runway Status Lights Pilot Reference Guide. Washington, DC: FAA.

———. 2018. Advisory Circular 150/5340 -30J, Design and Installation Details for Airport

Visual Aides. Washington, DC: FAA.

———. 2021. Order 7110.65Z, Air Traffic Control. Washington, DC: FAA.

Loukopoulos, L. D., Dismukes, R. K., & Barshi, I. 2009. The Multitasking Myth:

Handling Complexity in Real -World Opera tions. Burlington, VT: Ashgate.

NARA. “25 -Hour Cockpit Voice Recorder (CVR) Requirement, New Aircraft

Production; Notice of Proposed Rulemaking.” Federal Register. Vol. 88.

Washington, DC: National Archives and Records Administration,

December 4, 2023. 84090 .

Norman, D.A. 1981. “Categorization of Action Slips .” Psychological Review, 88(1),

NTSB (National Transportation Safety Board) . 2018a. Taxiway Overflight , Air Canada

Flight 759 , Airbus A320 -211, C -FKCK , San Francisco, California , July 7, 2017 .

AIR-18-01. Washington, DC: NTSB.

Aviation Investigation Report

AIR-24-01

67 ———. 2018b. Safety Recommendation Report: Extended Duration Cockpit Voice

Recorders . ASR -18-04. Washington, DC: NTSB.

———. NTSB to FAA, January 31, 2024. Response to 25 -Hour Cockpit Voice Recorder

(CVR) Requirement, New Aircraft Production; Notice of Proposed Rulemaking .

Reason, J.A. 1990. Human Error . New York: Cambridge University Press.

Wickens, C.D. & McCarley, J.S. 2008. Applied Attention Theory . Boca Raton, FL: CRC

Press .

Wilson, M.K., Strickland, L., Farrell, S., Visser, T.A . and Loft, S. 2020. Prospective

“Memory Performance in Simulated Air Traffic Control: Robust to Interruptions

but Impaired by Retention Inte rval. ” Human Factors , 62(8), 1249 -64.

Aviation Investigation Report

AIR-24-01

68 The NTSB is an independent federal agency charged by Congress with investigating every

civil aviation accident in the United States and significant events in the other modes of transportation —

railroad, transit, highway, marine, pipeline, and commercial space. We determine the probable causes

of the accidents and events we investigate and issue safety recommendations aimed at preventing

future occurrences. In addition, we conduct transportation safety research studies and offer information

and other assistance to family members and survivors for each accident or event we in vestigate. We also

serve as the appellate authority for enforcement actions involving aviation and mariner certificates

issued by the Federal Aviation Administration (FAA) and US Coast Guard, and we adjudicate appeals of

civil penalty actions taken by the FAA.

The NTSB does not assign fault or blame for an accident or incident; rather, as specified by

NTSB regulation, “accident/incident investigations are fact -finding proceedings with no formal issues

and no adverse parties … and are not conducted for the purpos e of determining the rights or liabilities

of any person” (Title 49 Code of Federal Regulations section 831.4). Assignment of fault or legal liability

is not relevant to the NTSB’s statutory mission to improve transportation safety by investigating

accidents and incidents and issuing safety recommendations. In addition, statutory language prohibits

the admission into evidence or use of any part of an NTSB report related to an accident in a civil action

for damages resulting from a matter mentioned in the report (Title 49 United States Code section

1154(b)).

For more detailed background information on this report, visit the NTSB Case Analysis and

Reporting Online (CAROL) website and search for NTSB accident ID DCA23LA125 . Recent publications

are available in their entirety on the NTSB website . Other information about available publications also

may be obtained from the website or by contacting —

National Transportation Safety Board

Records Management Division, CIO -40

490 L’Enfant Plaza, SW

Washington, DC 20594

(800) 877 -6799 or (202) 314 -6551

Copies of NTSB publications may be downloaded at no cost from the National Technical

Information Service, at the National Technical Reports Library search page, using product number

PB20 24-100107 . For additional assistance, contact —

National Technical Information Service

5301 Shawnee Rd.

Alexandria, VA 22312

(800) 553 -6847 or (703) 605 -6000

NTIS website