Air transportation safety investigation report A18A0085

The Sky Lease Cargo Boeing 747-412F aircraft (U.S. registration N908AR, serial number 28026) was conducting flight 4854 (KYE4854) from Chicago/O’Hare International Airport, Illinois, U.S., to Halifax/Stanfield International Airport, Nova Scotia, with 3 crew members, 1 passenger, and no cargo on board.

The crew conducted the Runway 14 instrument landing system approach. When the aircraft was 1 minute and 21 seconds from the threshold, the crew realized that there was a tailwind; however, they did not recalculate the performance data to confirm that the landing distance available was still acceptable, likely because of the limited amount of time available before landing. The unexpected tailwind resulted in a greater landing distance required , but this distance did not exceed the length of the runway.

The aircraft touched down firmly at approximately 0506 Atlantic Standard Time, during the hours of darkness. After the firm touchdown, for undetermined reasons, the engine No. 1 thrust lever was moved forward of the idle position, causing the speed brakes to retract and the autobrake system to disengage, increasing the distance required to bring the aircraft to a stop. In addition, the right crab angle (4.5°) on initial touchdown, combined with the crosswind component and asymmetric reverser selection, caused the aircraft to deviate to the right of the runway centreline.

During the landing roll, the pilot monitoring’s attention was focused on the lateral drift and, as a result, the required callouts regarding the position of the deceleration devices were not made.

Although manual brake application began 8 seconds after touchdown, maximum braking effort did not occur until 15 seconds later, when the aircraft was 800 feet from the end of the runway. At this position, it was not possible for the aircraft to stop on the runway and, 5 seconds later, the aircraft departed the end of the runway at a speed of 77 knots and came to a stop 270 m (885 feet) past the end.

The aircraft struck the approach light stanchions and the localizer antenna array. The No. 2 engine detached from its pylon during the impact sequence and came to rest under the left horizontal stabilizer, causing a fire in the tail section following the impact. The emergency locator transmitter activated. Aircraft rescue and firefighting personnel responded. All 3 crew members received minor injuries and were taken to the hospital. The passenger was not injured.

During the overrun, the aircraft crossed a significant drop of 2.8 m (9 feet) approximately 166 m (544 feet) past the end of the runway and was damaged beyond repair. While this uneven terrain was beyond the 150 m (492 feet) runway end safety area proposed by Transport Canada, it was within the recommended International Civil Aviation Organization runway end safety area of 300 m (984 feet). In 2007, the Board recommended that 

In addition, runway overruns is one of the issues on the TSB’s Watchlist 2020. The TSB Watchlist identifies the key safety issues that need to be addressed to make Canada’s transportation system even safer.

The investigation included a thorough fatigue analysis, which identified the presence of 2 fatigue risk factors that would have degraded the crew’s performance during the approach and landing: the timing of the flight and insufficient restorative sleep in the 24-hour period leading up to the occurrence. Fatigue management is also one of the safety issues on the TSB’s Watchlist 2020.

1.1 History of the flight

The Sky Lease Cargo Boeing 747-412F aircraft (U.S. registration N908AR, serial number 28026) was conducting flight 4854 (KYE4854), a multi-leg flight that originated at Chicago/O’Hare International Airport (KORD), Illinois, U.S., with a final destination of Changsha/Huanghua Airport (ZGHA), Hunan, China.

The first leg of the flight was a positioning flight to Halifax/Stanfield International Airport (CYHZ), Nova Scotia, where cargo would be loaded onto the aircraft. The second leg of the flight was to Ted Stevens Anchorage International Airport (PANC), Alaska, U.S., for a technical stop to refuel and change crews before the aircraft would depart on its final leg to ZGHA.

The occurrence flight was scheduled to depart KORD at 1230^{Footnote 1} on 06 November 2018, with an augmented crew^{Footnote 2} made up of a captain, first officer (FO), and international relief officer (IRO).^{Footnote 3} This augmented crew would operate the first 2 legs of the flight, from KORD to CYHZ and then to PANC. Also on board the occurrence flight was a deadheading senior captain, who was in the passenger area on the upper flight deck during the flight.

The captain and FO arrived in Chicago on the evening of 05 November, and the IRO arrived in Chicago on the morning of 06 November.

On the morning of 06 November, the captain consulted with the company’s flight operations, and they decided jointly to delay the departure by 13.5 hours, until 0200 on 07 November. The decision was based on forecast low ceilings and visibility that were below the company’s approach minima for the active runway at CYHZ. The rest of the crew and the deadheading captain were informed of the delay by telephone and email.

In accordance with company policy, flight dispatch called the crew at 2300, 3 hours before the new departure time. The crew arranged to be picked up from the hotel at 0000. Operational paperwork—such as flight planning, weather, and NOTAMs—was emailed to the crew, who printed these documents at the hotel.

The crew received a briefing by telephone from flight dispatch about the route, the weather, and applicable NOTAMs. The deadheading pilot was included in the pre-departure planning discussion because he had several years of experience operating these flights with the company and he had flown into CYHZ the preceding week.

The pre-departure planning included preparing the weight and balance form which provided the departure and arrival runways. The Max Allowed Gross Weight Landing section of this form indicated 302 092 kg and flaps 25 for a landing on Runway 23, which was 10 500 feet (3200 m) long. However, after reviewing the weather and NOTAMs, the crew planned to land on Runway 14, which was 7700 feet (2347 m) long. After arriving at KORD, the crew met with maintenance staff, who provided a briefing of work completed on the aircraft while it was at KORD.

Following a 1-hour delay due to a paperwork issue, the flight departed KORD at 0302 (14.5 hours after the original planned departure time) for the 2-hour flight to CYHZ. The captain was the pilot flying (PF) and occupied the left seat, while the FO was the pilot monitoring (PM) and occupied the right seat. The IRO was seated in the behind the PM.

During the cruise portion of the flight, the crew reviewed the weather at CYHZ based on automatic terminal information service (ATIS)^{Footnote 4} information Sierra,^{Footnote 5} which was issued at 0403 via the aircraft communication addressing and reporting system (ACARS). ATIS information Sierra indicated the following weather at 0400:

• Winds from 230° magnetic (M) at 10 knots
• Visibility 7 statute miles (SM) in light rain and mist
• Broken ceiling at 500 feet above ground level (AGL)
• Overcast layer of cloud at 3600 feet AGL
• Temperature and dew point 15 °C
• Altimeter setting 29.68 inches of mercury (inHg)

ATIS information Sierra also indicated that

• the landing runway was Runway 14 and the departing runway was Runway 23;
• the runway surfaces were bare and wet;
• the Runway 23 threshold was displaced;
• the instrument landing system (ILS) for Runway 23 and the Runway 05 localizer were unserviceable;
• Taxiway B was closed; and
• pilots should advise air traffic control (ATC) of the requested approach on initial contact.

The crew inferred that Runway 23 was not available to them for landing.

After reviewing ATIS information Sierra, at 0433, the PF conducted an approach briefing for Runway 14. The briefing included the landing distance required and approach speed based on the landing weight in the flight management system. The briefing also included the planned flap configuration of flaps 25 and the autobrake setting 4. According to data from the CYHZ tower, from 0413 until 0430, the winds were from 220°M to 230°M at 15 knots, gusting to 21 knots. At 0435, the winds were from 240°M at 15 knots, gusting to 22 knots.

At 0443, when the occurrence aircraft was 153 nautical miles (NM) from CYHZ, the crew began a descent from the cruising altitude of flight level (FL) 370.^{Footnote 6}

At 0446, the Moncton Area Control Centre (ACC) controller asked the crew which approach they were requesting for CYHZ. The crew requested the Runway 14 ILS approach, and the controller cleared the flight directly to the intermediate fix TETAR (Appendix A).

Shortly before the occurrence flight landed, 3 other aircraft landed at CYHZ:

• a Boeing 757 aircraft landed on Runway 14 at 0444;
• a Boeing 757 aircraft landed on Runway 23 at 0452; and
• a Learjet 35 aircraft landed on Runway 14 at 0454.^{Footnote 7}

The occurrence flight crew was not on the same radio frequency as these other aircraft; therefore, they did not hear any communication regarding the arrivals and continued to be unaware that Runway 23 was available for landing. The crew did not request any pilot reports (PIREPs) from ATC during the flight, and ATC did not offer any information regarding the other aircraft that had landed.

At 0454, the CYHZ tower controller informed the Moncton area control centre (ACC) controller that the arrival runway was being changed from Runway 14 to Runway 23, due to wind direction change, and that the approach into the airport was now the Runway 23 area navigation (RNAV) approach. When this exchange of information took place, the occurrence aircraft was descending through 13 000 feet ASL and approximately 52 NM from the airport (Figure 1). The runway change information was not communicated to the crew.

At 0458, the crew was cleared for the Runway 14 ILS approach. At approximately the same time, the ATIS was updated to information Tango. The only significant changes were to the arrival runway (now Runway 23) and the approach in use (now the Runway 23 RNAV approach).

At 0501, the Moncton ACC controller instructed the occurrence crew to transfer to the CYHZ tower frequency. On initial contact, the CYHZ tower controller informed the crew that the winds were from 260°M at 15 knots and asked if the crew had ATIS information Tango. The crew replied that they had Tango; however, they remained unaware that Runway 23 was available.

At 0502:46, when the aircraft was 8.6 NM from Runway 14, the CYHZ tower controller informed the crew that the winds were from 260°M at 16 knots, gusting to 21 knots. These winds would result in a steady 7-knot tailwind component. The controller asked the crew to confirm whether Runway 14 was still acceptable. The PM confirmed that Runway 14 was acceptable, and the tower controller repeated that the winds were from 260°M at 16 knots, gusting to 21 knots, and cleared the aircraft to land on Runway 14. The tower controller’s question as to the acceptability of Runway 14 prompted a brief conversation among the crew members about the perceived lack of runway options. However, the crew’s understanding was that only Runway 14 was available to them, and so continued the approach to Runway 14.

At 0504:10, the aircraft passed the final approach fix IMANO on the localizer and glideslope, and stabilized at the planned approach indicated airspeed (IAS) of 164 knots, with a ground speed of 185 knots. Just after passing the final approach fix, when the aircraft was 4.0 NM, or 1 minute and 21 seconds, from the threshold of Runway 14, the crew confirmed the presence of a tailwind and the PF confirmed with the PM the direction and speed of the wind. However, they did not change their approach speed. When the aircraft was passing through 800 feet AGL, the PF reviewed the go-around procedure with the crew.

At 0504:58, when the aircraft was 1.7 NM from the threshold, its IAS was 164 knots, and its ground speed was 174 knots. The tower controller reported that the winds were from 250°M at 15 knots, gusting to 21 knots.

At 0505:10, when the aircraft was at 400 feet AGL, the PF disengaged the autothrottle and autopilot. The IAS was 167 knots, and the ground speed was 174 knots.

At 0505:34, the aircraft crossed the threshold of Runway 14 at a height of 62 feet AGL, 12 feet above the threshold-crossing altitude of 50 feet, 27 feet left of centreline, with an IAS of 173 knots and ground speed of 179 knots.

At 0505:36, the thrust levers were brought to idle, the pitch attitude increased from 0.9° to 2.6° nose-up, and the aircraft touched down firmly 1350 feet past the threshold of Runway 14. At that point, it had an IAS of 168 knots and ground speed of 179 knots.

Over the following 40 seconds, a number of events happened in rapid succession (Appendix B).

The aircraft landed on the runway centreline with a crab angle of 4.5° to the right and at an average rate of lateral displacement of approximately 6 feet per second over the next 4 seconds. The firm (1.75g) landing and the subsequent deviation from the runway centreline to the right surprised the PM, whose attention was directed outside the aircraft. The crew did not experience the expected deceleration associated with autobrake 4 selection.

The auto speed brake lever moved to the UP position, and the spoiler panels began to deploy up to 30%. The No. 1 thrust lever was advanced above idle; however, this action was not noticed by the crew. The advancing of the thrust lever caused the speed brakes to move back to the DOWN position and retract the spoiler panels. As the PF was bringing the No. 2, No. 3, and No. 4 thrust levers into reverse, the air-ground logic switches in the landing gear changed momentarily to AIR mode, meaning the weight of the aircraft was not completely on the wheels, before switching back to GROUND mode.

At 0505:44, the No. 1 thrust lever was reduced to just above flight idle (6 seconds, approximately 1700 feet, after touchdown), which allowed the speed brakes to fully deploy. Shortly after, the autobrake selector disarmed.

At 0505:46, the PF realized that the aircraft was not decelerating as expected and began using manual braking. At 0505:53, the PF called out, “Max braking,” and the PM attempted to rearm the autobrake selector. The PM did not make the callouts for the landing roll-out procedure.

At the same time, the PF used the rudder pedals, switching between neutral and maximum deflection to the left to regain the centreline.

At 0505:59, the aircraft’s ground speed was 100 knots, and the aircraft was 800 feet from the end of the runway. There was markedly greater deceleration for the next 7 seconds, at which point the aircraft overran the runway. The aircraft was travelling at a ground speed of 77 knots at the time.

After the aircraft departed the paved surface, the landing gear left ruts (ground scars) in the grass (Appendix C).

At 0506:11, while travelling at a ground speed of 50 knots, the aircraft struck the ILS localizer antenna on top of a berm. When the aircraft struck the antenna, the emergency escape devices inside the cockpitwere projected from their storage compartment, injuring the IRO.

The nose of the aircraft came to rest 270 m (885 feet) past the runway threshold, 21 m (70 feet) to the right of the extended centreline, and 47 m (155 feet) from a public road (Figure 2). The aircraft was on a heading of 166°M (23° right of the runway heading).

The CYHZ tower controller activated the crash alarm at 0506. Halifax International Airport Authority (HIAA) aircraft rescue and firefighting (ARFF) personnel responded; 5 vehicles arrived at the occurrence site 1 minute and 40 seconds later.

The No. 2 engine separated from the wing and was jammed under the left horizontal stabilizer and tail section, causing a fire in the tail section of the aircraft after the impact. ARFF extinguished the fire and laid foam to prevent spilled fuel from igniting. The emergency locator transmitter activated, and the tower controller requested that ARFF shut off the device. With the assistance of ARFF, the crew evacuated through the main deck entry door (1L) using a ladder.

1.2 Injuries to persons

Table 1. Injuries to persons

Degree of injury	Crew	Passengers	Persons not on board the aircraft	Total by injury
Fatal	0	0	0	0
Serious	0	0	0	0
Minor	3	0	0	3
Total injured	3	0	0	3

1.3 Damage to aircraft

The aircraft was damaged beyond repair.

1.4 Other damage

The ILS localizer antenna array was destroyed when it was struck by the aircraft. Fuel contaminated the soil adjacent to where the aircraft came to rest, and the aircraft damaged or destroyed several runway end lights and lighting stanchions.

1.4.1 Environmental cleanup

In response to the occurrence, the HIAA activated its emergency response plan, which included an immediate environmental assessment of the site and surrounding area. The HIAA ARFF and airfield maintenance personnel dug trenches and implemented environmental controls and other mitigation strategies.

An estimated 136 600 L of fuel were on board the aircraft upon landing. As part of the environmental cleanup of the occurrence site, the aircraft operator’s insurer hired a third-party contractor to remove the remaining fuel from the aircraft, and 107 250 L of fuel were recovered. Therefore, it is estimated that 29 350 L of fuel spilled.

HIAA immediately began to arrange the environmental cleanup of the occurrence area. The aircraft operator’s insurer hired an environmental consultant to remediate the site, and HIAA also hired another environmental consultant to oversee the remediation efforts. During the remediation, 278 450 L of a fuel/water mixture were removed from the site, treated, and disposed of at an approved water treatment facility. In addition, 4998 tons of soil were removed from the site, treated, and disposed of at an approved soil disposal facility.

1.5 Personnel information

Table 2. Personnel information

	Captain	First officer	International relief officer
Pilot licence	U.S. airline transport pilot license (ATPL)	U.S. ATPL	U.S. ATPL
Medical expiry date	28 February 2019	31 January 2019	31 January 2019
Total flying hours	21 134	7404	5005
Flight hours on type	166	1239	1675
Flight hours in the 7 days before the occurrence	14	14	14
Flight hours in the 30 days before the occurrence	71.2	74.4	47.5
Flight hours in the 90 days before the occurrence	148	187	134
Flight hours on type in the 90 days before the occurrence	148	187	134
Hours on duty before the occurrence	5	5	5
Hours off duty before the work period*	13 hours	13 hours	13 hours
Takeoffs during the day in the 90 days before the occurrence	9	5	2
Takeoffs during the night in the 90 days before the occurrence	4	1	2
Landings during the day in the 90 days before the occurrence	9	4	2
Landings during the night in the 90 days before the occurrence	4	1	2

* These hours represent the hours off duty during the day of 06 November, between the time the decision was made to delay the flight in the morning to the start of duty that night.

1.5.1 Captain

The captain was hired in February 2018 as a direct-entry captain. He completed all required company training, which included crew resource management (CRM) and fatigue risk management training, the month that he was hired. The captain completed a proficiency check in August 2018 and was released to line flying in September. At the time of the occurrence, he held a U.S. airline transport pilot license (ATPL) with an instrument rating.

Section 121.436 of the U.S. Federal Aviation Regulations (FARs) states the experience requirements for pilots operating under Part 121. The occurrence captain had more than 1000 hours PIC experience flying large aircraft worldwide, but did not meet the required U.S.-based experience of paragraph 121.436(a)(3) of the FARs,^{Footnote 8} nor did he have a U.S. Federal Aviation Administration (FAA) exemption for these requirements. The lack of required U.S.-based experience was not considered a factor in this occurrence.

1.5.2 First officer

The FO was hired in July 2016. He completed all required company training, which included CRM and fatigue risk management training, in October 2018 and completed a proficiency check in September 2018. At the time of the occurrence, he held a U.S. ATPL with an instrument rating. He was qualified and certified in accordance with Part 121 of the U.S. FARs.

1.5.3 International relief officer

The IRO was hired in September 2015. He completed all required company training, which included CRM and fatigue risk management training, in October 2017 and completed a proficiency check in November 2017. At the time of the occurrence, he held a U.S. ATPL with a valid instrument rating. He was qualified and certified in accordance with Part 121 of the U.S. FARs.

1.6 Aircraft information

1.6.1 General

Table 3. Aircraft information

Manufacturer	Boeing
Type, model and registration	Boeing 747-412F, N908AR
Year of manufacture	1997
Serial number	28026
Certificate of airworthiness/flight permit issue date	Issued on 28 April 2017
Total airframe time	92 471 hours / 16 948 cycles
Engine type (number of engines)	Pratt & Whitney PW4056 (4)
Maximum allowable takeoff weight	394 625 kg
Recommended fuel type(s)	Jet A, Jet A-1, Jet B
Fuel type used	Jet A-1

The Boeing 747-412F is a 4-engine, wide-body transport category aircraft. The aircraft design can accommodate different configurations: passenger, freighter, and others. The occurrence aircraft was in the freighter configuration.

Records indicate that the occurrence aircraft was certified, equipped, and maintained in accordance with existing regulations and approved procedures. The weight and centre of gravity were within the prescribed aircraft limits. A number of the aircraft components were shipped to the TSB Engineering Laboratory in Ottawa, Ontario, where examinations found no indication of a component or system failure before the landing.

1.6.2 Landing gear

Boeing 747-400 series aircraft are supported on 4 main landing-gear assemblies and 1 nose landing-gear assembly. The main landing gear is made up of 2 assemblies under the body of the aircraft and 2 assemblies under the wings. The 2 assemblies under the wings are offset by approximately 19 feet (left and right) from the centre of the aircraft.

1.6.3 Brakes

All of the aircraft’s main wheel brakes are of the multi-disk type and have replaceable linings and segmented rotor brake disks. The brakes are fitted with automatic adjusters to compensate for brake wear. Each brake unit contains 2 wear-indicator pins to provide a visual indication as the wear approaches the limits.

For normal operation of the brakes, hydraulic system No. 4 supplies hydraulic pressure. For alternate operation, hydraulic system No. 1 or No. 2 can supply hydraulic pressure to the brakes.

The brake control system supplies brake torque control, anti-skid protection, and automatic braking (autobrake) during landing, takeoff, and taxiing. The brake torque control monitors torque of the brake during operation and releases brake pressure before the torque exceeds the maximum safe limit.

The flight data recorder (FDR) on the occurrence aircraft did not record brake pressure and brake pedal position, nor was it required to by regulation. The TSB laboratory determined that the lateral weight distribution did not cause the brake torque control to engage or prevent full brake pressure from being applied to the brakes.

The TSB laboratory also examined maintenance records associated with the landing-gear wheel assemblies and brake wear pin extensions. The brake wear indicator pin extensions were well within the manufacturer’s limits. No existing anomalies were found on the brake units that would have precluded normal operation during the occurrence flight.

1.6.3.1 Anti-skid system

The anti-skid system prevents wheels from locking up by controlling the brake pressure through anti-skid valves.

The anti-skid system electronically compares the airplane’s ground speed from the internal reference system with the wheel speed from the wheel-speed transducers to supply touchdown protection and hydroplane protection. If there is a difference between ground speed and wheel speed, error signals operate anti-skid valves that release brake pressure and prevent wheel lock.

Wheel speeds (actual and from the transducers) are not recorded on the FDR. In this occurrence, non-volatile memory (NVM) data did not reveal any anti-skid malfunctions, nor were there any fault messages on the engine indicating and crew alerting system (EICAS) related to the anti-skid system. Inspection of anti-skid components did not reveal any abnormalities.

A physical examination of the tire condition of all 16 main wheels found they were well within the required specification for tire wear. There were no indications of reverted rubber hydroplaning (refer to Section 1.16.2.3) on any of the examined tires. All wheel overpressure relief valves and thermal relief plugs were intact.

1.6.3.2 Autobrake system

The autobrake system supplies braking at a constant deceleration rate without manual input from the flight crew.

The autobrake control panel contains a rotary selector switch that can be set to OFF, DISARM, 1, 2, 3, 4, MAX AUTO, or RTO (rejected takeoff). The rate of deceleration depends on the switch position. Selecting autobrake 1 provides a deceleration rate of 4.0 feet per second squared, while selecting MAX AUTO provides a deceleration rate of 11.0 feet per second squared.

The autobrake is applied if the system is armed, all thrust levers are at idle, both left-hand and right-hand air/ground relay systems are in ground mode, and the wheels have spun up to at least 60 knots. While on the ground, the autobrake system is disarmed if any thrust lever is advanced out of idle for more than 3 seconds or if manual braking is applied.

The aircraft manufacturer recommends the use of the autobrake system when the runway has limited distance and when the aircraft is landing on slippery surfaces or with a crosswind. This can ensure lower brake temperatures, reduced tire and brake wear, and reduced stopping distances on slippery surfaces. The autobrake system commands brake pressure to target a desired deceleration rate. As noted in the flight crew training manual (FCTM), after touchdown, crew members should be alert for autobrake disengagement and notify the PF if this occurs.^{Footnote 9}

The average deceleration during the occurrence rollout was 6.2 feet per second squared. If autobrake 4 had remained engaged, the system design would have provided a target deceleration of 7.5 feet per second squared.

1.6.3.3 Auto speed brake system

Speed brakes are designed to increase drag and reduce lift in flight and during the landing roll. When the aircraft is in flight, 8 of the 12 spoilers are used as speed brakes. When the aircraft is on the ground, all of the spoilers are used to slow the aircraft. Upon touchdown, the speed brakes can be extended manually or automatically.

The speed brakes are normally extended automatically if the following conditions are met:

• the speed brake control lever is in the ARM position;
• both left-hand and right-hand air/ground relay systems are in ground mode;
• hydraulic system No. 1 or No. 4 is pressurized; and
• thrust levers No. 1 and No. 3 are below approximately 20° thrust lever angle.^{Footnote 10}

When the speed brakes are extended automatically, the auto speed brake system moves the speed brake lever aft to the UP position (Figure 3), which, in turn, raises the spoiler panels on the wings. This significantly reduces the lift generated by the wings and transitions the weight of the aircraft onto the wheels. If the auto speed brakes are not deployed properly, the landing distance required on a dry runway may be increased by as much as 870 feet. The FCTM contains 2 specific sentences about the importance of deploying auto speed brakes:

• “If the speedbrakes are not raised after touchdown, braking effectiveness may be reduced initially as much as 60% […].”^{Footnote 11}
• “Pilot awareness of the position of the speedbrake lever during the landing phase is important in the prevention of over-run.”^{Footnote 12}

The automation of the speed brakes relieves the flight crew from the task of pulling the speed brake lever aft to the UP position and allows them to focus on other critical tasks, such as straightening the aircraft and removing the crab angle from crosswind correction, completing the flare, manoeuvring the aircraft onto the centreline, and applying reverse thrust.

If the control lever is not in the ARM position, the auto speed brake actuator moves the speed brake lever aft to the UP position when reverse thrust lever No. 2 or No. 4 is deployed and the other conditions for automatic extension are met.

If the speed brake lever begins moving aft toward the UP position, but the conditions are not met, the speed brake lever automatically moves back to the DN (down) position, which retracts the spoiler panels.

1.6.4 Thrust reversers

The fan section of each engine is equipped with a hydraulically actuated, cascade-type thrust reverser. The thrust reversers are for ground use only and are used only to decrease the speed of the airplane during landings or rejected takeoffs.

For the thrust reverser to operate, the airplane must be on the ground to close the air/ground relays, and the forward thrust lever must be in the idle position. When the reverse thrust lever is lifted, thrust reversers are deployed.

The reverse thrust indication is displayed on the primary EICAS display (Figure 4). A thrust reverser status annunciator is positioned above each digital exhaust pressure ratio indicator. The reverse (REV) annunciator appears amber when the related reverser is unlocked or moving. The annunciator changes to green when the reverser is fully deployed. When the reverser is stowed and locked, the annunciator is no longer visible.

During the occurrence landing, 1 second after touchdown, the No. 1 thrust lever was advanced past flight idle, which inhibited the reverse thrust lever from deploying reverse thrust on that engine. The thrust reversers on engines No. 2, No. 3, and No. 4 operated as designed, with no anomalies.

1.6.5 Emergency exits

1.6.5.1 General

The Boeing 747-400F is equipped with 1 upper deck crew service door, 2 main deck entry doors, and 5 cargo doors. It is also equipped with an overhead hatch in the cockpit.

1.6.5.2 Upper-deck crew service door

The crew service door on the upper deck is located behind the cockpit, on the right side of the aircraft. It is used as a normal entry and exit as well as an emergency exit.

An emergency escape slide pack is mounted on and moves along tracks located to the right of the bulkhead. It runs horizontally and parallel to the door. To deploy the escape slide, the door must be open; the escape slide pack is then placed in front of the door and tilted, and the escape slide release handle is pulled.

1.6.5.3 Main deck entry door

The 2 main deck entry doors are located on the left side of the aircraft: 1 at the front of the aircraft and 1 at the rear. An escape rope that can be used in an emergency is stowed above both doors. The front entry door, identified as 1L, can be shut from inside or outside the aircraft. If the door is set to the AUTOMATIC position, lifting the door handle activates the emergency power system, and an escape slide deploys.

1.6.5.4 Flight deck overhead hatch

The aircraft cockpit is equipped with an overhead hatch that can be used by the cockpit crew to escape in an emergency. Eight emergency escape devices, consisting of inertial reels, are located in a side compartment, just above the IRO’s seat, encased and secured in position by a plastic cover. The crew members use these reels to limit the speed of their descent when exiting the aircraft.

1.6.6 Pilot anthropometric and ergonomic factors

Modern aircraft cockpits are designed to accommodate a broad range of human anthropometrics. The Boeing 747-400 is designed to accommodate pilots whose standing heights range from 5 feet 2 inches (the 20th percentile of height for women) to 6 feet 4 inches (the 99.5th percentile of height for men). Design eye reference point locators assist pilots in adjusting their seat position to maximize their field-of-view over the nose, as intended in the design of the aircraft.

Three-dimensional computer-aided design modelling conducted following the occurrence indicated that, when seated correctly using the design eye reference point locators, both pilots would have had full range of motion of the respective thrust levers and full rudder and brake pedal deflection. The investigation could not determine if the flight crew were seated correctly.

1.7 Meteorological information

1.7.1 General

On the morning of 06 November 2018, several aerodrome forecasts (TAFs)^{Footnote 13} were issued for CYHZ. The original departure time was scheduled for 1230, which meant the arrival time at CYHZ would have been 1500. The forecast weather at that time indicated a wind of 140° true (T) at 12 knots, visibility ¾ SM in light drizzle and mist, and an overcast ceiling at 300 feet AGL. At the time of arrival, the weather had a 30% probability of visibility ¼ SM in fog with vertical visibility of 100 feet AGL. This weather was below the company’s approach minima for the active runway and played a part in the decision to delay the departure by 13.5 hours.

An aerodrome special meteorological report (SPECI)^{Footnote 14} was issued for CYHZ at 2315 on 06 November 2018. It reported:

• Wind 210°T at 16 knots, gusting to 26 knots
• Visibility ½ SM with a runway visual range for Runway 14 of greater than 6000 feet with a downward trend in light rain and fog
• Vertical visibility of 200 feet AGL
• Temperature and dew point 15 °C
• Altimeter setting 29.74 inHg

A TAF was issued for CYHZ on 06 November at 2238 for the period of arrival. It indicated the following:

• Wind 220°T at 12 knots, gusting to 22 knots
• Visibility 3 SM in light rain and mist
• Broken ceiling at 400 feet AGL
• Overcast cloud at 1000 feet AGL

On 07 November, another TAF was issued for CYHZ at 0442 for the period of arrival. It indicated the following:

• Wind 220°T at 12 knots, gusting to 22 knots
• Visibility 5 SM in light rain and mist
• Broken ceiling at 400 feet AGL
• Overcast layer at 3000 feet AGL

Between 0500 and 0700, there would be the following temporary change in conditions:

• Visibility increasing to 6 SM in mist
• Scattered clouds at 400 feet AGL
• Overcast ceiling at 3000 feet AGL

An aerodrome routine meteorological report (METAR)^{Footnote 15} was issued at 0500 and reported:

• Wind 230°T at 13 knots
• Visibility 7 SM in light rain and mist
• Broken ceiling at 500 feet AGL
• Overcast layer at 1300 feet AGL
• Temperature and dew point 14 °C
• Altimeter setting 29.67 inHg

1.7.2 Environment and Climate Change Canada weather assessment

The TSB asked Environment and Climate Change Canada to assess the weather conditions prevailing at specific times before and after the occurrence. Given the available data, the weather assessment report^{Footnote 16} concluded that the most probable conditions during the occurrence aircraft’s descent and landing were as follows:

• No icing in cloud below approximately 12 000 feet ASL.
• Some borderline moderate mechanical turbulence may have been present below approximately 1000 feet ASL.
• The horizontal visibility at the airport was reported to be 7 SM; however, some mist may have still been in the area.
• The surface winds were from 230°T at 13 knots.
• The runway was wet from the significant rainfall in the previous 6 hours; light rain was falling when the aircraft arrived.

1.8 Aids to navigation

1.8.1 NOTAMs

In its Canadian NOTAM Procedures Manual, NAV CANADA defines a NOTAM as

a notice distributed by means of telecommunications containing information concerning the establishment, conditions or change in any aeronautical facility, service, procedure or hazard, the timely knowledge of which is essential to personnel concerned with flight operations.^{Footnote 17}

The manual further states that

[t]he basic purpose of NOTAM is the distribution of information that may affect safety and operations in advance of the event to which it relates, except in the case of unserviceable facilities or unavailability of services and activities that cannot be foreseen. Thus, to realize its purpose the addressee must receive a NOTAM in sufficient time to take any required action. The value of a NOTAM lies in its “news content” and its residual historical value is therefore minimal.^{Footnote 18}

NOTAMs for Canadian airports are produced and published by NAV CANADA based on information provided by aerodrome operators.

According to the Canadian NOTAM Procedures Manual, a

NOTAM shall be as brief as possible, stating only the essential facts4 and so compiled that its meaning is clear and unambiguous. Clarity shall take precedence over conciseness.

⁴ NOTAM are not issued after the fact just for the records to show that NOTAM were issued. For example, if no NOTAM were issued during the actual outage or closure, it is not permitted to promulgate the information after the fact.^{Footnote 19}

All NOTAMs were emailed to the crew by Sky Lease Cargo’s System Operations Control Center (SOCC) flight planning system, and the crew printed them at the hotel. The email included all applicable NOTAMs for the departure airport, the enroute phase of the flight, the destination airport, and the alternate airport (Bangor International Airport [KBGR], Maine, U.S.). The flight crew reviewed a total of 98 NOTAMs, including 37 concerning CYHZ (Appendix D). Of those 37, 22 NOTAMs provided information related to Runway 05/23, such as the reduced level of services or unserviceable navigational aids, the unserviceable runway lighting, and the displaced threshold. Sixteen of the NOTAMs related to Runway 05/23 contained amended information, with modifications that needed to be compared with the previous versions to identify the differences.

With regards to runway length available, the NOTAMs advised that the first 1767 feet of Runway 23 were closed due to painting, and repairs to lighting, and the threshold was relocated and marked with banners and runway threshold lighting. The declared landing distance available for both Runway 05 and Runway 23, which is normally 10 500 feet long, was reduced to 8733 feet.

Although this distance was sufficient, when the occurrence crew reviewed the NOTAMs related to the approach to Runway 23, they concluded that this runway was not available to them due to approach restrictions. The use of the wording “NOT AUTH” [not authorized] in NOTAM A3261/18 (Figure 5) led the crew to believe that they could not use the RNAV Runway 23 approach; however, the LNAV (lateral navigation) portion of the approach could still be used under some conditions.

Similarly, Runway 23 NOTAM 1385/18 and 1386/18 (Figure 6) indicated that the approach lighting system (ALS), the runway centreline lights (RCLL), the runway threshold lights (RTHL), and the runway touchdown lights (RTZL) were all unserviceable. From this list of unserviceable items, it can be concluded that the runway edge lights were still available; thus, the runway could have been used, but with limited lighting.

Reviewing the NOTAMs meant reviewing more than 7 pages of written information, including 3 pages just for Runway 05/23, presented using all capital letters.

A 2017 U.S. National Transportation Safety Board (NTSB) investigation into a near-miss taxiway landing at San Francisco International Airport (KSFO), California, U.S., concluded that

[a]lthough the NOTAM about the runway 28L closure appeared in the flight release and the ACARS message that were provided to the flight crew, the presentation of the information did not effectively convey the importance of the runway closure information and promote flight crew review and retention.^{Footnote 20}

The format and presentation of NOTAMs using all capital letters prove difficult for readability and interpretation because the letters are all the same size.^{Footnote 21},^{Footnote 22} This may influence a reader’s ability to interpret reduced services available at airports. Furthermore, NOTAMs are not prioritized based on importance, rather the order is based on the time of publishing. As a result, pilots must review all the information presented and determine how the reduced services will affect them.

NAV CANADA’s Canadian NOTAM Procedures Manual is based on International Civil Aviation Organization (ICAO) standards.^{Footnote 23},^{Footnote 24},^{Footnote 25} On 10 October 2019, NAV CANADA started using the ICAO NOTAM format for all NOTAMs, both domestic and international NOTAMs. According to NAV CANADA, “The adoption of the ICAO NOTAM format—already used by most countries—will ensure compliance with international standards and will eliminate the need for pilots who fly international routes to be familiar with more than 1 NOTAM format. It will also pave the way for more advanced filtering functionality, reducing NOTAM clutter by helping pilots access just the NOTAMs pertinent to their flight.”^{Footnote 26} While the ICAO format may communicate some information more effectively than the previous format, the presentation of text will continue to be of limited effectiveness because of the continued use of all capital letters.

1.8.1.1 Navigation aids and lighting

The following approach procedures were not authorized to be used due to the displaced threshold:

• Category I and II ILS Runway 23
• Localizer performance with vertical guidance (LPV) Runway 23
• Lateral navigation/vertical navigation (LNAV/VNAV) Runway 23
• Localizer Runway 05
• Required navigation performance (RNP) Y Runway 23

In addition, the following lighting for Runway 23 was unserviceable:

• Approach lighting system
• Runway centreline lighting
• Runway threshold lighting
• Runway touchdown zone lighting

The following approach procedures were available at CYHZ at the time of the occurrence flight:

• RNAV (RNP) Y Runway 05
• RNAV (global navigation satellite system [GNSS]) Z^{Footnote 27} Runway 05
• ILS Runway 14
• RNAV (GNSS) Z Runway 14
• RNAV (RNP) Runway 14
• LPV Runway 32
• LNAV/VNAV Runway 32
• RNAV (RNP) Y Runway 32
• NDB Runway 23
• RNAV (GNSS) Z Runway 23 with no vertical guidance LNAV minimums only

The aircraft was capable of performing all approaches at CYHZ except the RNP approaches.

1.9 Communications

All communications between ATC and the aircraft were normal.

1.10 Aerodrome information

1.10.1 General

CYHZ has 2 runways constructed of asphalt and concrete (Appendix E): Runway 05/23 is 10 500 feet long; Runway 14/32 is 7700 feet long; both are 200 feet wide. The runways are not grooved.

Runway 05 is equipped with a high-intensity (AN), simplified short-approach lighting system with runway alignment indicator lights (SSALR), threshold/runway end lighting, centreline lighting and a precision approach path indicator (PAPI) P3, which provides an eye-to-wheel height greater than 45 feet.

Runway 23 is equipped with centreline lighting, a high-intensity approach lighting system with sequenced flashing lights for category II or III operations (ALSF-2), threshold/runway end lighting and touchdown zone lighting.

Runway 14 is equipped with an AN SSALR and threshold/runway end lighting.

Runway 32 is equipped with an AN SSALR, threshold/runway end lighting, and a PAPI P3.

1.10.2 Runway end safety area

In 2009, the Australian Transport Safety Bureau (ATSB) published a safety report on runway excursions, which stated, in part:

Runway end safety areas [RESA] are designed to reduce the risk of damage to an aircraft that:

• undershoots the runway (touches down before the runway threshold);
• aborts a takeoff and overruns the runway end; or
• cannot stop following a landing and overruns the runway end.

A RESA achieves this by assisting aircraft to decelerate in a controlled manner.

Surface materials used for RESAs vary widely, from natural surfaces to pavement. Common RESA surface materials include compact gravel pavement, pulverised fuel ash (PFA), grass, pavement quality concrete (PQC), compacted earth, or a combination of these. In all cases, the bearing strength of the RESA must be able to support movement of airport rescue and firefighting (ARFF) vehicles, and be resistant to blast erosion from jet engine exhaust from aircraft in day-to-day operations. ^{Footnote 28}

ICAO, in its Annex 14,^{Footnote 29} requires that runways with a code number of 3 or 4^{Footnote 30} have a runway end safety area (RESA) of 90 m (295 feet) extending from the end of a 60 m (197 feet) runway strip, for a total of 150 m (492 feet). In addition, ICAO recommends that the RESA “should, as far as practicable, extend from the end of the runway strip to a distance of at least […] 240 m [787 feet],”^{Footnote 31} for a total of 300 m (984 feet). The recommendations for longitudinal slope suggests that the terrain should not exceed a downward slope of 5%. The slope should be as gradual as practicable and avoid any abrupt changes or sudden reversals.^{Footnote 32}

In Canada, before 2015 the 150 m (492 feet) RESA was not a regulatory requirement; it was only a recommendation. In 2015, Transport Canada (TC) published a new edition of its Aerodrome Standards and Recommended Practices (TP 312),^{Footnote 33} in which it changed the previous RESA recommendations into standards. However, because a grandfathering clause^{Footnote 34} was included in the Canadian Aviation Regulations (CARs), Canadian airports are required to adhere to the latest RESA standards only when a new runway is constructed.

In 2016, TC issued a notice of proposed amendment (NPA) to the CARs^{Footnote 35} proposing requirements of 150 m (492 feet) RESA to be based on air traffic volume rather than runway length.

The characteristics^{Footnote 36} and description of RESAs in TP 312 are the following:

3.2.1.7 The runway end safety:

(a) has a minimum width twice of the associated runway;

(b) extends away from the runway;

(c) is centred on the extended runway centerline; and

(d) […] has a minimum length of 150 m to the end of the RESA.

[…]

3.2.1.9 The terrain in the runway end safety area:

(a) has no abrupt slope changes or open ditches;

(b) has adequate slope to prevent the accumulation of water;

(c) beyond the runway strip, has maximum transverse and longitudinal slopes of 5% downwards;

(d) does not protrude into an obstacle limitation surface (OLS); and

(e) under dry conditions, is of sufficient strength to reduce the severity of structural damage to the critical aircraft overrunning/undershooting the runway.^{Footnote 37}

Runway 14 at CYHZ has a RESA that extends to 150 m (495 feet) past the runway end. This length had an average downward slope of 0.2%. These dimensions meet TC’s and ICAO’s standards for a 150 m (492 feet) RESA (Figure 7).

Approximately 166 m (544 feet) past the end of Runway 14, there is a significant drop of 2.8 m (9 feet), with a downward slope of 73%. This slope does not meet ICAO’s recommendations for a 5% longitudinal slope for a RESA that extends to 300 m (984 feet) past the runway end (Figure 8).

1.10.2.1 Previous TSB recommendation

Following the TSB’s investigation^{Footnote 38} into a runway overrun accident involving an Airbus A340-313 aircraft in 2005 at Toronto/Lester B. Pearson International Airport (CYYZ), Ontario, the Board recommended that

This recommendation was in keeping with the recommendations of ICAO.

Since then, TC has provided several responses, all of which have been assessed by the TSB.

In February 2021, in an update to its most recent response, TC stated that the amendments to the CARs were published in the Canada Gazette, Part I on 07 March 2020. TC is aiming to publish these amendments in the Canada Gazette, Part II in May 2021.

In March 2021, in its reassessment of TC’s latest response, the TSB noted that TC had proposed regulations to address RESAs. The proposed regulatory changes, as currently written, will reduce the risks associated with an overrun; however, not to the level that would be afforded by the ICAO-recommended 300 m RESA. At a minimum, the Board believes that the proposed regulations must meet the ICAO standard.

The Board has maintained the inclusion of runway overruns on Watchlist 2020. Runway overruns continue to occur, and the lack of timely action will continue to expose commercial air travellers in Canada to unnecessary risks until these regulatory amendments are implemented.

Therefore, the Board reassessed TC’s latest response to Recommendation A7-06 as Satisfactory in Part.^{Footnote 39}

1.10.3 Runway friction coefficient and certification

TC requires airport operators to periodically measure the friction characteristics of the runway surface. TC leaves it up to airport operators to conduct their own runway friction tests and establish the frequency of testing based on the unique history and circumstances of their sites.

On 19 October 2018, HIAA hired an independent contractor to complete a runway friction test for both runways. The test results met all required standards, in accordance with the 5^th edition of TP 312.^{Footnote 40}

1.10.4 Transverse slope and runway drainage

Drainage paths and the transverse slope were assessed by HIAA, and all runways met all requirements in accordance with the 5th edition of TP 312.

1.11 Flight recorders

The aircraft was equipped with a solid-state digital FDR, which contained approximately 53.4 hours of flight data, covering the occurrence flight and 8 previous flights. The FDR data were successfully downloaded.

The aircraft was also equipped with a cockpit voice recorder (CVR), which had a recording capacity of 120 minutes; its recorded data included the occurrence flight. The CVR memory was successfully downloaded and contained good quality audio for the occurrence flight.

1.12 Wreckage and impact information

The aircraft overran Runway 14 and struck approach lighting stanchions and the ILS localizer antenna array. The aircraft proceeded approximately 270 m (885 feet) past the threshold and came to rest 47 m (155 feet) from a public road.

During the impact sequence, the nose gear collapsed rearward, the left-hand wing main landing gear separated from the fuselage, and the left-hand body main landing gear collapsed rearward, with the wheel assemblies pushing up into the fuselage fairing.

The right-hand wing main landing gear collapsed aft and rotated 45° with the wheel bogie, coming to rest below the right-hand flap assembly. The right-hand body main landing gear collapsed aft and was embedded in the fuselage fairing area.

All engines were damaged by impact forces and by ingesting foreign object debris. Engines No. 1 and No. 4 were still attached to the wing pylons. Engine No. 2 detached from its pylon during the impact sequence and came to rest under the left horizontal stabilizer. Engine No. 3 detached from its pylon during the impact sequence and came to rest on the right side of the fuselage, behind the right outboard flap.

The fuselage buckled behind the cockpit and behind the wing root.

1.13 Medical and pathological information

1.13.1 Fatigue

Human beings need 7 to 9 continuous hours of restorative sleep at night to perform at optimal levels.^{Footnote 41} Sleep-related fatigue—related to the amount and quality of sleep obtained—is biological in nature. Consequently, it is not prevented by personality characteristics, intelligence, education, training, skill, compensation, motivation, physical size, strength, or practice. Sleep-related fatigue can result from 1 or more of 6 risk factors: acute sleep disruptions (i.e., within the previous 24- to 72-hour period), chronic sleep disruptions, continuous wakefulness, circadian rhythm disruptions, sleep disorders or other medical and psychological conditions, and/or illnesses or drugs that affect sleep or sleepiness.

Inadvertently falling asleep at the controls is the most recognized risk of fatigue; however, less extreme fatigue levels are associated with more subtle performance impairments, such as a decrease of cognitive functioning and problem-solving abilities. These subtle impairments are significant risk factors and predictors of occupational accidents and injuries,^{Footnote 42} motor vehicle accidents,^{Footnote 43} and aviation occurrences.^{Footnote 44}

Sleep-related fatigue impairs working memory (memory that temporarily stores information while it is being manipulated for tasks such as reasoning)^{Footnote 45} and problem-solving ability.^{Footnote 46} It also reduces flexibility in a person’s problem-solving approach to a situation that is perceived to be different from the routine, so that the individual perseverates and repeats previously ineffective responses.^{Footnote 47} This increases the likelihood that a fatigued person will maintain the normal routine, failing to revise the original plan or to devise and try a novel solution.^{Footnote 48} This known effect of sleep-related fatigue on flight crew cognitive bias was identified in a recent, high-profile NTSB air occurrence involving flight crew perception and decision-making during final approach and landing. The investigation revealed that “fatigue likely contributed to the crewmembers’ misidentification of the intended landing surface, their ongoing expectation bias, and their delayed decision to initiate a go-around.”^{Footnote 49}

1.13.2 Circadian rhythm

The time of day has a strong effect on an individual’s alertness and performance because of changes in body physiology that are synchronized to a circadian (daily) rhythm. The human body is physiologically ready for sleep at night and for wakefulness during the day. Likewise, due to the circadian rhythm, overall performance and cognitive functioning are at their worst during the nighttime circadian rhythm trough, from approximately 2230 to 0430, when fatigue increases significantly. Even if a person slept the previous night and is not feeling fatigued,^{Footnote 50} overall performance may be degraded during the circadian rhythm trough.

1.13.3 Sky Lease fatigue risk management training

Fatigue risk management training provides employees with knowledge of how to avoid, mitigate and report fatigue issues. At Sky Lease Cargo, all flight operations personnel were required to take annual recurrent fatigue risk management training. This training was given in a 1-hour lecture format using slides and videos, and included:

• the effects of fatigue on human performance (including, for example, its effects on “cognitive slowing” and short-term memory),
• the interaction between workload and fatigue during different phases of flight,
• sources of fatigue,
• recommended “do’s and don’t’s”,
• personal strategies to mitigate the effects of fatigue, and
• information on Sky Lease’s fatigue reporting policy.

1.13.4 Crew sleep-wake cycle

Before the occurrence flight, all 3 crew members had maintained a normal nighttime sleep schedule for at least 7 days.

The captain arrived at KORD on 05 November and slept 9.5 hours that night. On 06 November, at 1100, the decision was made to delay the departure to CYHZ until 0200 on 07 November. After he was notified of the delay, the captain had the opportunity to get uninterrupted rest on 2 occasions (for less than 2 hours each time) during the afternoon and the evening. If he slept during these periods, he would have had a total of approximately 5.75 hours of sleep in the 24-hour period leading up to the occurrence.

The FO arrived at KORD on 05 November and slept 8.5 hours that night. He had the opportunity to get 3 hours of uninterrupted rest during the afternoon of 06 November. If he slept during this period, he would have had a total of approximately 6 hours of sleep in the 24-hour period leading up to the occurrence.

The IRO spent 05 November at his home and slept about 7 hours that night. He arrived at KORD on the morning of 06 November. He had an opportunity to get 2 hours of sleep in the afternoon. He did sleep for 3 hours in the evening hours. If he slept on both occasions, he would have had a total of approximately 5 hours of sleep in the 24-hour period leading up to the occurrence.

Quantitative analysis of the crew’s sleep history using fatigue avoidance scheduling tool (FAST) software predicted that, because of acute sleep disruption and the timing of the occurrence during the nighttime circadian rhythm trough, the performance of the flight crew at the time of the accident would likely have been degraded by fatigue. However, the crew did not report feeling fatigued at the time of the occurrence.

1.14 Fire

The No. 2 engine detached from its pylon during the impact sequence and came to rest under the left horizontal stabilizer and tail section, which caused a post-impact fire in the tail section of the aircraft. Five ARFF crash rescue vehicles responded, and they extinguished the fire and laid foam to prevent spilled fuel from igniting.

1.15 Survival aspects

When the aircraft left the paved surface, emergency escape devices (see 1.6.5.4 Flight deck overhead hatch) were projected from the storage compartment, striking the IRO in the head.

The flight deck and upper deck passenger compartment were not compromised structurally during the runway overrun.

The crew began an evacuation from the upper deck crew service door; however, after placing the emergency escape slide pack into position, they had difficulty deploying the emergency slide. After the crew assessed the situation and the captain signalled with a flashlight to ARFF, they egressed through the main deck entry door (1L) using a ladder, with the assistance of ARFF. Emergency health services responded at 0545, evaluated the crew and transported them to hospital.

When the TSB investigators were on site, they moved the escape slide pack into position with no issues; however, they did not attempt to deploy the emergency slide. It could not be determined why the crew had difficulty deploying the slide.

1.16 Tests and research

1.16.1 Flight data recorder analysis

The TSB laboratory downloaded FDR data and provided factual data plots. It also reconstructed the events (Appendix F), analyzed the braking performance and deceleration devices, and assessed the likeliness of an engine strike. The FDR data indicated that, immediately after touchdown, the thrust levers for engines No. 2, No. 3, and No. 4 were moved to reverse thrust positions, while the thrust lever for engine No. 1 was advanced forward, past 20° thrust lever angle. This inhibited the autobrakes and the auto speed brakes.

Shortly after, the thrust lever for engine No. 1 was reduced to just above idle, which satisfied the conditions for the auto speed brakes to engage and the spoilers to extend.

CYHZ airport’s closed-circuit television (CCTV) surveillance footage and the aircraft’s FDR parameters were synchronized. During the landing, a light source briefly appeared in the background, silhouetted by the aircraft’s fuselage. This light source was coincident with a maximum 6° left bank. The surveillance footage, FDR parameters, site survey data, and geometric analysis indicate that there was likely contact of the No. 1 or No. 2 engine nacelle with the runway (engine strike). After the likely engine strike, the FDR data did not indicate any abnormalities with engine performance. The engines continued to operate as commanded by the flight crew.

1.16.1.1 Yawing moments, rudder effectiveness, and asymmetrical braking

The crosswind component and the asymmetric thrust caused adverse yawing moments during the landing. The FDR data indicated that the PF used rudder deflections between neutral and maximum. Analysis of the yawing moments determined that the rudder had sufficient control authority to overcome the adverse yawing moments. Asymmetric braking was not required in order to maintain directional control.

1.16.2 Hydroplaning

1.16.2.1 General

In aviation, hydroplaning is used to describe a loss of braking friction due to liquid water. There are 3 main forms of hydroplaning—dynamic, reverted rubber, and viscous—and their characteristics are quite different.

1.16.2.2 Dynamic hydroplaning

During total dynamic hydroplaning, the tire lifts off the pavement and rides on a wedge of water like a water ski. Because the conditions required to initiate and sustain dynamic hydroplaning are extreme, the phenomenon rarely occurs. However, when it does, there is such a substantial loss of tire friction that the wheel may not spin up.

The conditions required for dynamic hydroplaning are high speed, standing water, and poor surface macrotexture. These conditions must continue without interruption to keep the tire planing on a wedge of water. In the absence of any of these conditions, dynamic hydroplaning either does not occur at all or affects only a portion of the tire footprint.^{Footnote 51}

The rainfall intensity was determined from weather records obtained from the Halifax Regional Municipality’s water, wastewater, and storm water utility. A one-hour average rainfall intensity, preceding the event, was determined to have been 1.29 mm/hour. The 5-minute “instantaneous” rainfall intensity at the time of the landing was 1.20 mm/hour. The highest rainfall intensity during the one-hour window prior to the event, occurred at 0450 (over 15 minutes prior to landing), with 4.8 mm/hour. 

The TSB laboratory used a rainfall intensity of 1.24 mm/hour to estimate the maximum possible water depth on the runway during the occurrence landing, and estimated it to be 0.24 mm or less. Given this estimation of water depth, in combination with the aircraft’s speed and the runway macrotexture, the laboratory determined that dynamic hydroplaning almost certainly did not occur and did not have a significant effect on the landing.

1.16.2.3 Reverted rubber hydroplaning

Reverted rubber hydroplaning can occur when a tire is skidded along a very wet or icy runway long enough to generate frictional heat in the footprint area. This heat generates steam, which expands and reduces the traction of the contact patch. This steam can also leave characteristic “steam cleaning” marks on the runway where the rubber deposits have been removed or discoloured. It also causes a characteristic disfiguring on the tire’s rubber tread.

Reverted rubber hydroplaning can start at any speed above about 20 knots and can result in tire friction comparable to that of icy runways.^{Footnote 52}

In this occurrence, there was insufficient braking or deceleration to create the heat necessary for steam. There was no physical evidence of skidding, “steam cleaning,” or disfiguring of the rubber tread. Therefore, the investigation concluded that reverted rubber hydroplaning did not occur.

1.16.2.4 Viscous hydroplaning

Viscous hydroplaning occurs on all wet runways and is used to describe the normal slipperiness or lubricating action of the water. While viscous hydroplaning reduces friction, it would not reduce friction to such a low level that the wheel cannot be spun up shortly after touchdown to initiate the anti-skid system. Viscous hydroplaning is the most commonly encountered cause of low friction on wet runways, and occurrences are often mistaken for dynamic hydroplaning.^{Footnote 53} Some level of viscous hydroplaning occurs on all wet runways.

The theoretical wet runway aircraft braking coefficient (ABC) represents the aircraft’s ability to grip the surface of the runway given the value of the maximum braking effort required for anti-skid operation under specific conditions. This theoretical wet runway ABC was calculated for Runway 14 under the conditions at the time of the occurrence.

Using FDR data, the actual ABC for the occurrence flight was calculated, and it was determined that during the period of ground roll before maximum braking was applied, the coefficient was lower than expected with autobrake 4 selected. After maximum braking effort was applied, the actual ABC was consistent with the theoretical ABC on Runway 14 under the existing wet runway conditions.

1.16.3 TSB laboratory reports

The TSB completed the following laboratory reports in support of this investigation:

• LP248-2018 – Site Survey Halifax Stanfield International Airport
• LP249-2018 – NVM Download
• LP252-2018 – FDR Report
• LP265-2018 – Wheel and Brake Examination

1.17 Organizational and management information

1.17.1 Sky Lease Cargo

1.17.1.1 General

Sky Lease Cargo is a U.S. company that provides domestic and international non-scheduled and ad hoc heavy cargo lift. It holds a Supplemental Operations Certificate for all-cargo operations under the U.S. FARs, Part 121. The company is based in Miami, Florida, U.S.

At the time of the occurrence, the company had 320 employees, and its fleet consisted of 3 Boeing 747-400 and 2 MD-11 aircraft. As required by the FARs, Sky Lease Cargo has implemented a safety management system and a fatigue risk management plan, which included: policy statements for the management of fatigue, a training requirement for its flight crew members, and an audit system to assess fatigue levels within the organization.

Sky Lease Cargo pilots complete their Boeing 747-400 training in Miami. The required training consists of classroom and computer-based training, both conducted internally, as well as aircraft simulator training, which is conducted through an external vendor but by company instructors and a check airman.^{Footnote 54}

1.17.1.2 Pre-departure planning

Pre-departure planning plays an important role in establishing a clear picture for the crew of the influences (positive or negative) that may affect the flight, while ensuring that the flight meets regulatory requirements.

Sky Lease Cargo’s SOCC is located in Miami, Florida, and includes the flight following centre. The operations centre provides flight planning, weather briefings, and flight releases for the Boeing 747 and MD-11 fleets. When the crews are not in Miami, flight dispatch provides all required flight paperwork to the crew by email and conducts any required briefings over the telephone. Before a flight departs, the pilot-in-command must communicate with the SOCC and secure a flight release. The following are required in duplicate for the flight, and signed copies are required onboard the aircraft:^{Footnote 55}

• Flight release
• Load manifest
• Operational flight plan
• Pilot route certification
• Weight and balance
• Aircraft logbook (with a valid airworthiness release)

1.17.1.2.1 Landing limitations: destination airports

Sky Lease Cargo provides performance data for its specific aircraft to an independent vendor, who, in turn, provides runway analysis charts for various airports to Sky Lease Cargo. The charts indicate specific landing performance data, including the maximum landing weight permitted for a specific runway (Appendix G). The runway analysis charts are used for pre-departure planning and cannot be used to determine landing distances required.

Section 121.195 of the FARs^{Footnote 56} prohibits the takeoff of a transport category aircraft unless its weight on arrival, allowing for normal consumption of fuel and oil in flight, allows a full-stop landing at the intended destination airport within 60% of the effective runway length. To determine the allowable landing weight at the destination airport, the following 3 assumptions are made:

• The aircraft is landed on the most favourable runway in still air.
• The aircraft is landed on the most suitable runway, considering probable wind, landing aids and terrain.
• If the runway is forecast to be wet or slippery at the time of arrival, the required field length is increased by 15%.

Sky Lease Cargo and Boeing both use the FAA regulatory requirements and guidance when referring to a “wet runway,” defined as when more than 25% of the runway surface area (within the reported length and width being used) is covered by any visible dampness or water that is less than 1/8 inch (3 mm) deep.^{Footnote 57},^{Footnote 58}

Sky Lease Cargo’s General Operations Manual contains a policy specific to wet and slippery runways:

It is Skylease Cargo policy to release all flights assuming wet destination runways unless existing conditions on a particular flight will unduly restrict fuel and/or payload. When a particular flight appears so restricted, the Captain will jointly determine in advance the advisability of using dry runway lengths for planning purposes.

CFR 121.195(d) must be applied when weather reports, forecasts, or a combination of the two indicated that the runways at the destination airport maybe [sic] wet or slippery at the estimated time of arrival. The required runway length must then be at least 115% [sic] greater than the normal (dry) required runway length for landing. The tabular runway analysis utilized by the Company complies with this requirement.^{Footnote 59}

All runways at CYHZ had been reported as bare and wet, and light rain had been reported for the previous 24 hours.

Sky Lease Cargo uses charts that provide the maximum landing weight that meets the requirements of section 121.195 of the FARs, for each destination runway and each approved landing configuration (flaps 25 and flaps 30). The pre-departure maximum landing weights with flaps 25 and flaps 30 were determined using these charts (Table 4).

Table 4. Landing runway limit weight (Source: Sky Lease Cargo landing runway limit weight charts)

Flaps	Runway	Length (feet)	Aircraft systems	Landing conditions	Limit weight (kg)
25°	14	7700	All operating	Wet, no wind	261 500
30°	14	7700	All operating	Wet, no wind	279 400

The performance-limited landing weight for Runway 14 was not recorded on any documentation used by the crew. Based on the operational flight plan, the aircraft’s estimated landing weight on arrival in CYHZ was 265 852 kg.

1.17.2 Stabilized approach criteria

1.17.2.1 Flight crew training manual

Sky Lease Cargo provides guidance and directives to its pilots regarding stabilized approach criteria through flight safety letters, briefings in initial and recurrent ground school, simulator training, and company documentation. One of the main sources of information for the crews, however, is the FCTM.

The Boeing FCTM states:

Maintaining a stable speed, descent rate, and vertical/lateral flight path in landing configuration is commonly referred to as stabilized approach concept.

Any significant deviation from planned flight path, airspeed, or descent rate should be announced. The decision to execute a go-around is not an indication of poor performance.

Note: Do not attempt to land from an unstable approach.

Recommended Elements of a Stabilized Approach

The following recommendations are consistent with criteria developed by the Flight Safety Foundation.

All approaches should be stabilized by 1,000 feet AFE [above field elevation] in instrument meteorological conditions (IMC) and by 500 feet AFE in visual meteorological conditions (VMC). An approach is considered stabilized when all of the following criteria are met:

• the airplane is on the correct flight path.
• only small changes in heading and pitch are required to maintain the correct flight path.
• the airplane should be at approach speed. Deviations of +10 knots to -5 knots are acceptable if the airspeed is trending toward approach speed.
• the airplane is in the correct landing configuration.
• sink rate is no greater than 1,000 fpm; if an approach requires a sink rate greater than 1,000 fpm, a special briefing should be conducted.
• thrust setting is appropriate for the airplane configuration.
• all briefings and checklists have been conducted.^{Footnote 60}

1.17.2.2 U.S. Federal Aviation Administration advisory circular

The FAA published an advisory circular (AC) in which it defined a stabilized approach as “one of the key features of safe approaches and landings in air carrier operations, especially those involving transport category airplanes.” The FAA also indicated that

An approach is stabilized when all of the following criteria are maintained from 1000 HAT [height above threshold] (or 500 HAT in visual meteorological conditions [VMC]) to landing in the touchdown zone:

The airplane is on the correct¹ track.

The airplane is in the proper landing configuration.[...]

The airplane speed is within the acceptable range specified in the approved operating manual used by the pilot.

The rate of descent is no greater than 1000 feet per minute (fpm).[...]

Power setting is appropriate for the landing configuration selected, and is within the permissible power range for approach specified in the approved operating manual used by the pilot.

Note ¹: A correct track is one in which the correct localizer, radial, or other track guidance has been set, tuned, and identified, and is being followed by the pilot. Criteria for following the correct track are discussed in FAA Advisory Circulars relating to Category II and Category III approaches. Criteria for following track in operations apart for Category II and Category III are under development.^{Footnote 61}

1.17.3 Company flight manual

Checklists and operating procedures provide guidance to assist flight crew during normal operations, abnormal operations, and emergencies. Sky Lease Cargo’s main reference for operating the Boeing 747-400 aircraft is the 747-400 Company Flight Manual. The manual provides detailed information on limitations in one chapter (Chapter L), normal procedures and amplified procedures in another chapter (Chapter NP), and supplementary procedures in another chapter (Chapter SP). Supplementary procedures include procedures such as adverse weather operations, non-normal operations, system tests, and other procedures not included in Chapter NP.

The 747-400 Company Flight Manual contains information about the automatic flight system, including the use of autopilot. While it does not state when the autopilot needs to be disconnected on approach, it does provide guidance on minimum altitudes for disconnecting the autopilot. In the case of an ILS approach, the autopilot must be disconnected by 150 feet AGL.^{Footnote 62}

The 747-400 Company Flight Manual also contains the standard operating procedures (SOPs) for landing rolls^{Footnote 63} (Table 5), including actions to be taken by the various flight crew members. It provides the verbal callouts to be made by the flight crew members so they may take their respective appropriate actions.

Table 5. Sky Lease Cargo’s landing roll standard operating procedure (Source: Sky Lease Cargo, 747-400 Company Flight Manual, revision 1 [19 June 2017], Chapter NP: Landing Roll Procedure, p. NP.21.51)

Pilot flying	Pilot monitoring
Verify that the thrust levers are closed Verify that the SPEEDBRAKE lever is UP.	Verify that the SPEEDBRAKE lever is UP. Call “SPEEDBRAKES UP” If the SPEEDBRAKE lever is not UP, call “SPEEDBRAKES NOT UP”.
Monitor the rollout progress.
Verify correct autobrakes operation.
WARNING: After the reverse thrust levers are moved, a full stop landing must be made. If an engine stays in reverse, safe flight is not possible.
Without delay, move the reverse thrust levers to the interlocks and hold light pressure until the interlocks release. Apply reverse thrust as needed.	Verify that the forward thrust levers are closed. When all REV indications are green, call “REVERSERS NORMAL.” If there is no REV indication(s) or the indication(s) stays amber, call “NO REVERSER(S) ENGINE NUMBER ___” or “NO REVERSERS”.
By 60 knots, start movement of the reverse thrust levers to be at the reverse idle detent before taxi speed.	Call “60 KNOTS”.
After the engines are at reverse idle, move the reverse thrust levers full down.
Before taxi speed, disarm the autobrakes. Use manual braking as needed.
Before turning off the runway, disconnect the autopilot.

Normal checklist procedures and non-normal checklist procedures are available to the crew in the Quick Reference Handbook (QRH).^{Footnote 64} The non-normal checklist items are organized by aircraft systems and package model identification of the specific 3 aircraft in the fleet.

1.17.4 Landing performance calculations

1.17.4.1 General

Sky Lease Cargo crews base the landing performance numbers on the landing weight in the flight management system. The aircraft QRH’s performance section is referenced to determine the reference speed (V_REF)^{Footnote 65} and to calculate the approach speed^{Footnote 66} (command speed^{Footnote 67}) and landing distance required. This document is available to the crew in the electronic flight bag and as a paper copy document, both of which are located on the flight deck.

The typical flap setting for Sky Lease Cargo operations is flaps 25; however, flaps 30 may be required due to a limited runway length or the weather conditions. The crew completed the landing distance calculation at the time of the approach briefing using flaps 25, and estimated the landing distance required to be around 6000 feet. Although there were no restrictions or limitations preventing the use of flaps 30 at CYHZ, this flap setting was not considered by the crew because the landing distance available of 7700 feet was greater than the crew thought they needed.

1.17.4.2 Approach speed

The Boeing FCTM provides guidance on how to calculate the approach speed based on the estimated landing weight and wind. Based on the estimated landing weight for CYHZ of 265 800 kg, the crew calculated a V_REF of 154 knots with flaps 25 for landing (Figure 9).

If the autothrottle is disconnected, or is planned to be disconnected before landing, the recommended method for approach speed correction is to add to the reference speed one half of the reported steady headwind component plus the full gust increment above the steady wind. The minimum command speed setting is V_REF + 5 knots.^{Footnote 68} When making adjustments for winds, the maximum approach speed should not exceed V_REF + 20 knots. Figure 10 shows examples of wind additives when the runway heading is 360°.

In this occurrence, based on the guidance from the FCTM, the approach speed is V_REF plus any wind additive, with a minimum approach (command) speed of V_REF + 5 knots, or 159 knots minimum. The winds provided by ATIS information Sierra of 230°M at 10 knots would have resulted in a direct right crosswind for landing on Runway 14. Therefore, there was no headwind component, and the wind additive was 0, resulting in an approach (command) speed of 159 knots. When the winds changed to 260°M at 16 knots, gusting up to 21 knots, resulting in a steady 7-knot tailwind component, the wind additive remained at 0 knots, as no wind additives are applied for steady or tailwind gusts (Figure 10). Hence the recommended approach speed remained at 159 knots.

During the approach preparation, the crew used ATIS information Sierra and calculated the wind additive to be 5 knots (half of 10 knots), and then added that to V_REF+5, to get V_REF + 10 knots, or 164 knots, as the approach (command) speed. A previous TSB investigation into an overrun that occurred in 2015^{Footnote 69} determined that the occurrence pilots added half of the total steady state wind, rather than half of the headwind component. The report also stated that company check pilots noted that this error was common. 

1.17.4.3 Unfactored (actual) landing distances

The charts for normal configuration landing distance in the QRH “performance package model identification” section for the occurrence aircraft were unfactored; thus, they provided actual landing distance data (without any additional safety margin). The QRH contained no factored charts for the occurrence aircraft. During the approach preparation, the crew used the data from ATIS information Sierra.

The crew used the typical flaps 25 configuration for landing and calculated the approach speed using V_REF + 10 knots, or 164 knots. Using the QRH data (Appendix H), the crew calculated the landing distance required at approximately 6000 feet, with autobrakes 4. This landing distance was not recorded on any flight documentation.

Post-flight calculations using the QRH guidance under the existing conditions, with flaps 25, autobrakes 4, and an approach speed of 159 knots (V_REF + 5 knots), resulted in a landing distance of 6375 feet (Appendix H). However, if the actual landing distance was corrected for the higher (V_REF + 10 knots) approach speed, it would increase to 6735 feet. Table 6 indicates the actual landing distances using flaps 25 and flaps 30, based on the winds

• as discussed during the briefing,
• when ATC asked the crew to confirm that Runway 14 was still acceptable (when the aircraft was 8.6 NM from landing on its final approach),
• when ATC provided a wind check at 1.7 NM on its final approach, and
• when the aircraft crossed the threshold of the runway.

Table 6. Quick Reference Handbook actual landing distance based on the changes in existing conditions (Source: TSB, based on Sky Lease Cargo, 747 Flight Crew Operations Manual: 747 Quick Reference Handbook, Revision 01 [05 May 2017], p. PI-QRH.12.2 and PI-QRH.12.3)

Flaps	Position	Wind direction and speed	Wind component (knots)	Approach speed	Actual (unfactored) landing distance required (feet)	Extra runway (feet)
Flaps 25	Briefing	230°, 10 knots	Headwind 0.5	VREF+10 (164 knots)	6735	965
	8.6 NM final	260°, 16 knots, gusting to 21 knots	Tailwind 7.3	VREF+10 (164 knots)	7514	186
	1.7 NM final	250°, 15 knots	Tailwind 4.4	VREF+10 (164 knots)	7211	489
	Runway threshold*	250°, 15 knots	Tailwind 4.4	VREF+19 (173 knots)	8088*	−388*
Flaps 30	Briefing	230°, 10 knots	Headwind 0.5	VREF+10 (158 knots)	6241	1459
	8.6 NM final	260°, 16 knots gusting to 21 knots	Tailwind 7.3	VREF+10 (158 knots)	6991	709
	1.7 NM final	250°, 15 knots	Tailwind 4.4	VREF+10 (158 knots)	6700	1000
	Runway threshold*	250°, 15 knots	Tailwind 4.4	VREF+19 (167 knots)	7541*	159*

* This landing distance includes the fact that the aircraft was 9 knots above the target speed used for the other calculations and 12 feet above the reference threshold crossing height of 50 feet.
NOTE: The unfactored landing distance required with flaps at 25 was greater than the runway length when the aircraft crossed the runway threshold.

1.17.4.4 Factored landing distance

The normal configuration landing distance charts in the QRH “performance package model identification” section for the other 2 company Boeing 747-400 aircraft are factored by 1.15. At the time of the occurrence, the company was updating the occurrence aircraft's manual to reflect factored numbers.

Post-flight calculations using the unfactored (actual) landing distance, factored by 1.15, resulted in the values shown in Table 7.

Table 7. Quick Reference Handbook unfactored landing distance then factored by 1.15 (Source: TSB, based on Sky Lease Cargo, 747 Flight Crew Operations Manual: 747 Quick Reference Handbook, Revision 01 [05 May 2017], p. PI-QRH.12.2 and PI-QRH.12.3)

Flaps	Position	Wind direction and speed	Wind component (knots)	Actual (factored) landing distance required (feet)	Extra runway (feet)
Flaps 25	Briefing	230°, 10 knots	Headwind 0.5	7745	−45
	8.6 NM final	260°, 16 knots, gusting to 21 knots	Tailwind 7.3	8641	−941
	1.7 NM final	250°, 15 knots	Tailwind 4.4	8293	−593
	Runway threshold*	250°, 15 knots	Tailwind 4.4	9301*	−1601*
Flaps 30	Briefing	230°, 10 knots	Headwind 0.5	7177	523
	8.6 NM final	260°, 16 knots gusting to 21 knots	Tailwind 7.3	8039	−339
	1.7 NM final	250°, 15 knots	Tailwind 4.4	7705	−5
	Runway threshold*	250°, 15 knots	Tailwind 4.4	8672*	−972*

* This landing distance includes the fact that the aircraft was 9 knots above the target speed used for the other calculations and 12 feet above the reference threshold crossing height of 50 feet.
NOTE 1: With flaps at 25, the actual landing distance required was greater than the runway length when the aircraft was 8.6 NM from landing, when it was 1.7 NM from landing, and when it crossed the runway threshold.
NOTE 2: With flaps at 30, the actual landing distance required was greater than the runway length when the aircraft crossed the runway threshold.

1.17.5 NAV CANADA

1.17.5.1 General

NAV CANADA is a private company that provides air navigation services for commercial and general aviation in Canadian airspace. NAV CANADA operates under Subpart 8 of the CARs and meets the requirement to have a safety management system set out in section 801.05.

1.17.5.2 Unit operations manual for Halifax air traffic control tower

The NAV CANADA Manual of Air Traffic Services – Tower provides guidance for assigning an active runway.^{Footnote 70} If the winds at the airport are 5 knots or more, the manual specifies that the tower should “assign the operationally suitable runway most closely aligned into the wind.” ^{Footnote 71}

The suitable runway selection is based on the runway aligned into wind. The manual specifies that an airport controller may assign “a runway with a tailwind component,”^{Footnote 72} under the following circumstances:

• The runway is dry.
• You [they] indicate the wind direction and speed to the pilot.
• The tailwind component, including gusts, does not exceed 5 knots.^{Footnote 73}

Airport controllers may also assign a runway with a crosswind. However, the Manual of Air Traffic Services – Tower specifies that, when the runway condition is wet, the maximum crosswind component,^{Footnote 74} including gusts, is 15 knots and adds, “The pilot makes the final decision on the acceptability of a runway.”^{Footnote 75}

An airport controller is required to coordinate with all operating positions concerned if an active runway is changed. The arrangement between CYHZ tower and Moncton ACC^{Footnote 76} specifies that the airport controller verbally coordinates a change in the active runway at the airport with the Moncton ACC controller in addition to updating the information on the controller’s workstation.

NAV CANADA Unit Operations Manual (UOM) for the Halifax ATC tower^{Footnote 77} provides direction and information for the controllers on procedures. According to the UOM, an airport controller must not designate a runway with a tailwind component as the arrival runway or advertise it as available on the controller’s workstation if the wind component exceeds the criteria in the NAV CANADA Manual of Air Traffic Services – Tower. However, the UOM also states that an airport controller may assign a runway with a tailwind component if it is most nearly aligned into the wind or the only runway available.^{Footnote 78}

The UOM adds “[i]f a pilot requests arrival on a runway for which the tailwind component is exceeded you [a controller] may approve that request provided you [he/she] advise[s] the pilot of the wind direction and speed.”^{Footnote 79}

An ACC controller is required to issue landing information before or shortly after an aircraft is cleared to descend. If the information is included in the current ATIS broadcast and a flight crew acknowledges receipt of the broadcast, there is no requirement for the controller to issue the information a second time.^{Footnote 80}

After issuing landing information, if an ACC controller learns of information that may affect an aircraft’s descent, approach, or landing, the controller is required to inform the pilot.^{Footnote 81} Furthermore, the controller is required to inform the flight crew of any new and pertinent information that differs from the current ATIS message.^{Footnote 82} At 0454, the CYHZ tower controller informed the Moncton ACC controller that the designated arrival runway was being changed from Runway 14 to Runway 23 and that the approach to the airport was now the RNAV approach to Runway 23. The runway change information was not communicated to the crew.

1.18 Additional information

1.18.1 Pilot decision making

1.18.1.1 General

An important component of pilot decision making is the ability to recognize changes in a situation and reinitiate the decision-making process to ensure that necessary changes are made and plans are modified. In particular, in-flight decisions are typically prompted by unanticipated events that require adjustment of the planned course of action. Situation assessment allows crews to make more effective decisions by interpreting the pattern of cues that define an issue, assessing the level of risk, and determining the time available to reach a solution.^{Footnote 83}

Failure to adequately consider the potential implications of a situation during decision making increases the risk that decisions and subsequent associated action(s) will result in adverse outcomes.

1.18.1.2 Situational awareness

Situational awareness is “the perception of elements in the environment within a volume of time and space, the comprehension of their meaning, and the projection of their status in the near future.”^{Footnote 84} Accurate situational awareness maximizes effective and safe decision making in the cockpit. To maintain accurate situational awareness, a pilot must first perceive information from the environment, then establish its relevance in terms of achieving operational goals, and, finally, use it to project and predict future states and events, allowing the ability to “plan ahead and prepare for contingencies.”^{Footnote 85} A pilot’s knowledge, experience, and expectations are some of the individual factors that influence situational awareness.^{Footnote 86} A construct proposed^{Footnote 87} to underpin situational awareness is called “affective awareness” or a pilot’s “gut feeling” (for threats to safety), which is characterized by an emotional, sensory experience that triggers further cognitive analysis.

Errors at the most basic level of situational awareness involve failure to correctly perceive the situation. This can happen because critical information is not available to the individual, either because it was not presented effectively by the system or because it was not communicated effectively among individuals. Research^{Footnote 88} on causal factors underlying aviation occurrences has found that most of the accidents involving a substantial human error and situational awareness component can be attributed to failures to correctly perceive some piece(s) of information, either because data were unavailable or were difficult to detect or understand.

Team situational awareness is the degree to which every team member possesses the situational awareness required for his or her responsibilities ^{Footnote 89} and the degree of shared understanding among crew members. ^{Footnote 90} While a captain has ultimate responsibility for decision making and the overall safety of an aircraft, other crew members provide critical redundancy. In addition to their individual tasks, crew members support each other by monitoring the situation and one another’s performance, and by intervening if a problem is detected. Individuals outside of the aircraft (e.g., air traffic controllers, flight dispatchers) also play a role in informing and calibrating team situational awareness by providing crews with accurate and timely information on weather, traffic, and runway environment.

Accuracy of team situational awareness is improved when individual team members share a consistent understanding and representation—or “mental model”—of how a system works. Flight crews who share a mental model are more likely to arrive at a common understanding of a given situation without needing as much verbal communication than  crews who do not. Teams who do not share a mental model tend to require more real-time coordination and communication to ensure that their activities are carried out properly.^{Footnote 91}

1.18.1.3 Factors affecting pilot decision making and situational awareness

1.18.1.3.1 Knowledge

Knowledge gained through experience and training on an aircraft type, and through experience with a runway environment, facilitates flight crew decision making by improving the accuracy of situational awareness.^{Footnote 92} Expertise can facilitate effective pilot decision making by 1) facilitating rapid and accurate perception of information or cues that signal a problem; 2) estimating the likelihood of various outcomes; and 3) facilitating an accurate mental model of a situation so that the best option can be chosen.^{Footnote 93}

The captain was experienced in various large aircraft; however, he had limited experience on the Boeing 747 (166 hours of flight time), with a total of 13 takeoffs and landings, including 4 takeoffs and landings at night. The operating crew had been to CYHZ previously; however, they had not landed on Runway 14, nor had they made an approach to CYHZ at night.

During the flight, some crew conversation touched on the captain’s feelings of apprehension, or anxiety, about the conditions at CYHZ. The other flight crew members were aware of the captain’s limited experience on the Boeing 747 and of his feelings of anxiety regarding the crosswind expected during the impending landing at CYHZ.

During pre-flight planning at the hotel, the crew, including the deadheading pilot, were briefed by flight dispatch on the weather in CYHZ as well as on the NOTAMs regarding runway landing conditions. The plan was made to land on Runway 14.

Although he was not a member of the operating crew, because of his extensive experience on the Boeing 747 and with the company, the deadheading captain was asked by the crew about any known issues with landing in CYHZ, including the viability of landing on Runway 14. The deadheading captain had been captain on a daytime flight into CYHZ about 1 week before the occurrence. That previous flight had landed on Runway 23. At the time, the deadheading captain found that Runway 23’s usable length was not clearly marked, even under daytime conditions.

1.18.1.3.2 Communications and crew resource management

Effective communications—within the cockpit and among flight crew and ground-based personnel such as ATC and flight dispatch—are an important element in the decision-making process. According to the European Aviation Safety Administration and Transport Canada, crew resource management (CRM) “is the effective utilization of all resources including crew members, aircraft systems, supporting facilities and persons to achieve safe and efficient operations.”^{Footnote 94}

One of the primary goals of CRM training is to “enhance communication, interaction, human factors and management skills of the crew members concerned.”^{Footnote 95} Traditional CRM training stresses the importance of using clear, assertive language when communicating in the cockpit. The Sky Lease Cargo CRM training curriculum included modules on “communication barriers” and “assertiveness.”

Research on speech comprehension^{Footnote 96},^{Footnote 97} shows that certain elements of speech quality, tone, and content can be interpreted by listeners as indicating uncertainty and/or ambiguity. During the flight, in the minutes preceding the transfer to the tower frequency, the crew made 24 statements and callouts related to the configuration of the aircraft. The tone, timing, and content of some of the crew’s speech communications indicated some level of uncertainty and ambiguity and of limited confidence in their ability to manage the impending landing.

1.18.1.3.3 Workload

Workload is a function of the number of tasks that must be completed within a given time. Workload increases if the number of tasks to be completed increases or if the time available decreases. Individuals use both physiological (i.e., increased heart rate) and cognitive (i.e., focusing attention) resources to manage high-workload situations.

An individual may experience acute stress and associated anxiety if a high-workload situation becomes physically threatening and the individual is uncertain of their ability to manage the threat. This anxiety is maladaptive, because it disrupts the person’s ability to manage a high-workload situation by degrading attention and working memory capacity.^{Footnote 98}

High levels of mental workload can thereby adversely affect a pilot’s ability to perceive and evaluate cues from the environment and can negatively affect situational awareness by causing attentional narrowing.^{Footnote 99} Those experiencing acute stress are also “more likely to be distracted from a crucial task by highly salient stimuli, such as an alarm.”^{Footnote 100} Consequently, their management of a high-workload situation “may become disjointed and chaotic.”^{Footnote 101} “In some cases, [problems in situational awareness] may occur […] owing to a momentary overload in the tasks to be performed or in information being presented.”^{Footnote 102}

Because anxious thoughts tend to pre-empt working memory’s limited storage capacity, the individual may have difficulty performing computations that would normally be easy and have difficulty making sense of the overall situation and updating their mental model of the situation.^{Footnote 103}

The expression of high mental workload and stress in one person can be communicated to other team members, leading to increased levels of team workload. Research has found that acute stress negatively affects team performance by impairing team integration and mental models.^{Footnote 104}

1.18.1.3.4 Cognitive influences

Pilots operate in a complex environment, monitoring multiple sources and types of information. To help them cope with the large amount of information in the environment that is available to the senses at any given time, humans have developed cognitive skills or “biases” that can facilitate information processing. These normal biases, however, have an unintended consequence: not all of the – potentially critical – elements in the environment will be attended to, which can lead to uninformed decisions.

When the amount of available information about a situation is limited, people tend to rely on the first piece of credible information that is available to them to inform situation assessments. This is known as “anchoring bias” and can make it difficult to assess unfolding situations. Similarly, having only limited information about a situation can increase an individual’s tendency to look for evidence that confirms or matches their current assessment or decision, a phenomenon known as “confirmation bias.” These biases can make it less likely for a crew member to reassess their initial assessment and update it with new information, or lead them to attend to information that supports their current decision, while dismissing information that is contrary to what is expected.^{Footnote 105} The danger in both circumstances is that alternative outcomes will not be given an appropriate level of consideration when deciding on the best possible course of action.

Research and past accident investigations have demonstrated that, once a plan is made and committed to, it becomes increasingly difficult for flight crew to recognize stimuli or conditions in the environment that necessitate a change to the plan.^{Footnote 106} “Plan continuation bias” is the “deep-rooted tendency of individuals to continue their original plan of action even when changing circumstances require a new plan.”^{Footnote 107} A condition or stimulus needs to be perceived as sufficiently salient to be recognized and acted upon in a timely manner. When plan continuation interferes with a crew’s ability to detect important cues, or if the crew fails to recognize the implications of those cues, situational awareness can break down.^{Footnote 108},^{Footnote 109} These breakdowns can result in non-optimal decisions that can compromise safety.

1.18.1.4 Unexpected abnormal conditions and workload

The physical movement skills required for flying become faster, more accurate and coordinated with repeated practice, allowing a pilot to reach a control or make a correct size of action without visually checking.^{Footnote 110} Response accuracy also depends on timely feedback regarding the consequences of control movements.^{Footnote 111}

Emergencies or abnormal conditions that are unexpected^{Footnote 112} can create high workload and stress, and can impair performance in situations where people have limited time to process critical information and adjust actions accordingly.^{Footnote 113} Unexpected events surprise pilots because what happens in the environment does not match the individual’s or team’s mental model of the situation and of what is supposed to happen.^{Footnote 114} Situations involving very high levels of workload can result in important steps, such as SOP calls during a landing rollout, being delayed or omitted.

These omissions are called slips (of attention), and are the error type most frequently associated with the performance of routine, well-practiced tasks under conditions that are unexpected, or unusual. A slip of attention occurs when a check on the progress of a task sequence is mistimed or does not occur because an operator's attention is focused elsewhere. Both experienced and inexperienced pilots can make slips, but, since pilots who are learning new aspects of a familiar task typically need to devote more attention to it than a more experienced pilot does, slips tend to be more common in those with less experience on a given task.^{Footnote 115}

1.18.2 Runway overrun initiatives

1.18.2.1 Flight Safety Foundation

An analysis of a 14-year period of runway overrun data by the Flight Safety Foundation (FSF) states that “the risk of runway excursion increases when more than one risk factor is present. Multiple risk factors create a synergistic effect (i.e., two risk factors more than double the risk).”^{Footnote 116} These factors involved weather, aircraft performance, crew technique and decision making, or aircraft systems. Of relevance to this occurrence, the review found that the following were frequent contributors:

• Unanticipated wind shear or tail wind
• Incorrect assessment of landing distance for prevailing wind and runway conditions
• Landing fast
• Excessive height over threshold, resulting in landing long
• Failure to detect non-deployment of ground spoilers (e.g., absence of related standard call)
• Bouncing and incorrect bounce recovery
• Late braking (or late takeover from autobrake system, if required)
• Increased landing distance resulting from the use of differential braking or the discontinued use of reverse thrust to maintain directional control in crosswind conditions

The FSF’s recommended mitigations for these included:

• Defined policies for a rejected landing (bounce recovery)
• Defined task-sharing and standard calls for final approach and roll-out phases in SOPs
• Published procedures for adverse runway conditions
• Published procedures for optimum use of autobrake and reverse thrust on contaminated runways

1.18.2.2 Boeing

Boeing published AERO magazine quarterly from 1998 to 2014, providing operators with supplemental technical information to promote continuous safety and efficiency in their daily fleet operations. Data collected and analyzed from 2003 to 2010 and published in AERO in 2012 showed that the factors contributing to landing overruns occurred at the following frequencies:

• 90% landed on a runway that was not dry
• 68% occurred after stable approaches
• 55% touched down within the touchdown zone
• 42% landed with a tailwind of 5 knots or greater

This review showed that a runway overrun is typically caused by multiple factors. As a result, a multi-faceted approach to reducing the incidence of runway overruns would be required.

Among the factors listed, the following were applicable to this occurrence:

• Tailwinds
• High touchdown speed
• Speed brakes deployed late or not deployed
• Reversers deployed late or not deployed

Event data suggest that a number of runway overruns can be avoided if the flight crew has a more thorough understanding of the interrelationship between the landing environment and the current risks (e.g., weather, winds, runway conditions, minimum equipment list items, airplane weight). Pilots need to better understand the relationships among these factors for each flight:

• A stabilized approach
• Known and accounted for runway contamination
• Available versus required runway length
• Reported conditions compared with actual conditions
• Speed for the flight’s approach
• Energy to be dissipated after landing
• Speed additives and effect on landing distances
• Reliability of runway braking action
• Proper, timely use of all deceleration devices

A failure or misunderstanding of each of these factors can lead to runway overrun excursions. Several of the mitigations recommended by Boeing to reduce runway overruns focus on increasing crew awareness.

1.18.3 U.S. Federal Aviation Administration

1.18.3.1.1 Mitigating the risks of a runway overrun

To provide pilots and operators with a way “to identify, understand, and mitigate risks associated with runway overruns during the landing phase of flight,” the FAA issued AC 91-79A on 17 September 2014.^{Footnote 117} The AC was intended for use in the development of SOPs to mitigate such risks.

According to the AC, specific SOPs are “a primary risk mitigation tool” and should “[a]s a minimum” address the overrun hazards. Furthermore, it is “imperative” that these SOPs be executed faithfully by flight crews. An effective training program on runway overrun mitigation provided by operators also provides flight crews with “academic knowledge and skill to increase the pilot’s awareness of the factors that can cause a runway overrun.”

1.18.3.1.2 Standard operating procedures for flight deck crew members

The FAA also published an advisory circular on SOPs for flight deck crew members, which stated the following:

Standard operating procedures (SOPs) are universally recognized as basic to safe aviation operations. Effective crew coordination and crew performance, two central concepts of crew resource management (CRM), depend upon the crew’s having a shared mental model of each task. The mental model, in turn is founded on SOPs.^{Footnote 118}

The AC also “emphasizes that SOPs should be clear, comprehensive, and readily available in manuals used by flight deck crewmembers.”

The intent of SOPs is to provide effective and efficient communication to all crew members and to ensure that specific actions are taken in various phases of flight. Traditional CRM training for pilots stresses the importance of using clear, assertive language when communicating in the cockpit.

1.18.3.1.3 Safety alerts for operators: landing distance

The FAA also issues Safety Alerts for Operators (SAFOs), “an information tool that alerts, educates, and makes recommendations to the aviation community. […] Each SAFO contains important safety information and may contain recommended actions. SAFO content should be especially valuable to air carriers in meeting their statutory duty to provide service with the highest possible degree of safety in the public interest. The information and recommendations in a SAFO are often time critical.”^{Footnote 119}

In 2006, the FAA issued a SAFO concerning landing distance, in which the FAA

urgently recommends that operators of turbojet airplanes develop procedures for flight crews to assess landing performance based on conditions actually existing at time of arrival, as distinct from conditions presumed at time of dispatch. Those conditions include weather, runway conditions, the airplane’s weight, and braking systems to be used. Once the actual landing distance is determined an additional safety margin of at least 15% should be added to that distance.

The FAA considers a 15% margin between the expected actual airplane landing distance and the landing distance available at the time of arrival as the minimum acceptable safety margin for normal operations.^{Footnote 120}

In 2019, the FAA issued a SAFO^{Footnote 121} that replaced this previous SAFO “to assist operators in developing methods to ensure sufficient landing distance exists to safely make a full stop landing.”

This SAFO reiterates that, “Once the actual landing distance is determined at the time of arrival, an additional safety margin of at least 15 percent should be added to actual landing distance. Except under emergency conditions, flight crews should not attempt to land on runways that do not meet the assessment criteria and safety margins as specified in this SAFO.”

1.18.4 TSB Watchlist

The TSB Watchlist identifies the key safety issues that need to be addressed to make Canada’s transportation system even safer. Runway overruns and fatigue management in air transportation are 2 Watchlist issues that are relevant to this occurrence.

1.18.4.1 Runway overruns

Despite the millions of successful movements on Canadian runways each year, aircraft sometimes continue past the end of the runway surface during landings or rejected takeoffs. These events, known as runway overruns, can result in aircraft damage, injuries, and even loss of life—and the consequences can be particularly serious when there is no adequate RESA or suitable arresting system designed to stop an aircraft.

Since 2005, there have been on average 9.7 runway overrun occurrences per year at Canadian airports, of which 7.5 occur during landing. Additionally, from 2005 to 2019 the TSB investigated 19 such occurrences and issued 4 recommendations to Canadian authorities. Three of those recommendations remain active^{Footnote 122} and 1 is closed.^{Footnote 123}

In March 2020, TC proposed regulations that would, among other things:

• require a 150 m RESA at airports with over 325 000 commercial passengers annually;
• require the use of an arresting system on runways where the 150 m RESA cannot be implemented; and
• be limited to runways serving commercial passenger services.

According to TC, these regulations, once implemented, will increase runway overrun protection to passengers from 75% of passenger traffic in 2017 to 95% by 2038. However, these regulations focus only on the risk to a majority of, but not all, passengers and do not consider non-passenger air traffic or the terrain at the end of all runways. Also, the proposed regulations may not fully meet the ICAO standard, which requires a 150 m RESA for all runways that are 1200 m in length and longer, and provisions for other types of runways.^{Footnote 124} Therefore, the TSB remains concerned that, without further action, risks to the public, property, and the environment remain.

Runway overruns: ACTIONS REQUIRED Despite the action taken to date, the number of runway overruns in Canada has remained constant since 2005 and demands a concerted effort to be reduced. Operators of airports with runways longer than 1800 m must conduct formal runway-specific risk assessments and take action to mitigate the risks of overruns to the public, property, and the environment. TC must adopt at a minimum the ICAO standard for RESAs, or a suitable arresting system designed to stop an aircraft.

1.18.4.2 Fatigue management in air transportation

In the transportation industry, crews often work long and irregular schedules—sometimes crossing multiple time zones or facing challenging conditions—that are not always conducive to proper restorative sleep. Fatigue poses a risk to the safety of air operations because of its potential to degrade several aspects of human performance.

ACTIONS REQUIRED Fatigue management in air transportation will remain on the Watchlist until the following actions are taken: Canadian air operators that operate under CARs subparts 703, 704 and 705 implement, and comply with, the new regulations on flight crew fatigue management. The impact of these new regulations on aviation operations in Canada is assessed by the TSB.

• While the TSB Watchlist targets Canadian operators, the issue of fatigue management has also been highlighted on the NTSB’s Most Wanted List since 2016.^{Footnote 125}

1.19 Useful or effective investigation techniques

Not applicable.

The aircraft was certified, equipped, and maintained in accordance with existing regulations, and no discrepancies were noted that would have prevented it from operating normally during the occurrence flight.

In an effort to understand why this runway excursion occurred, this analysis will focus on the crew assessment of the approach and landing and their actions based on the information available, the factors contributing to runway overruns, and the crew’s management of the operational threats.

2.1 Pre-departure planning

Pre-departure planning plays an important role in establishing a clear picture for the crew of the influences (positive or negative) that may affect the flight, while ensuring that the flight meets regulatory requirements.

2.1.1 Weather and NOTAMs

Because the low ceilings and visibility were below the company’s approach minima for the active runway at Halifax/Stanfield International Airport (CYHZ), the captain and Sky Lease Cargo flight operations jointly decided to delay the flight for 13.5 hours.

As part of the pre-departure planning at Chicago/O’Hare International Airport (KORD), the crew and flight dispatch reviewed 98 NOTAMs, including 37 concerning CYHZ that were presented in the sequence they were issued, in all-capitalized text, and not prioritized. Of these 37 CYHZ NOTAMs, 22 concerned Runway 05/23 and involved reduced services or unserviceable navigational aids, runway lighting, and a displaced threshold. Ten of these 22 contained repeated information with modifications, yet had to be compared with the previous version to identify the differences.

NOTAMs are intended to be a clear, concise, and unambiguous presentation of essential information. However, it is difficult to reliably extract the crucial information because of their presentation style, using all capital letters, and because of their sequence, in which important approach and runway NOTAMs are not prioritized but buried among other information. As a result, to determine which approaches and runways are available, users must extract the important items, search back and forth to compare repeated information, build a list of unavailable items, and compare this list with the approach charts. This extraction process of elimination is usually performed mentally, increasing the risk of misinterpretation, resulting in the crew having an inaccurate mental model of the operational hazards affecting a flight, and reducing the crew’s situational awareness.

For example, in this occurrence, the 10 NOTAMs related to the approaches on Runway 23 were buried in the sequential list of 37 for CYHZ. Through a back-and-forth process of elimination, the crew believed that there were no approaches or lighting available on Runway 05/23. However, 2 approaches were available on Runway 23: the non-directional beacon (NDB) and the lateral navigation (LNAV) portion of the area navigation (RNAV), both with restrictions on the use of charted information.

Based on their review of the NOTAMs, the crew concluded that the instrument landing system (ILS) approach to Runway 14 was the only option.

2.1.2 Landing limitations: destination airports

Sky Lease Cargo’s policy is to use wet runway calculations for the planning of all of its flights. To meet the requirements of section 121.195 of the U.S. Federal Aviation Regulations (FARs), Sky Lease Cargo uses runway analysis charts to determine the maximum landing weight for the runway expected at its destination, based on the weather, the approach, and landing aids.

This pre-departure maximum landing weight determination includes additional safety margins that are not included in the landing distance charts used in flight.

The Max Allowed Gross Weight Landing section of the weight and balance form indicated 302 092 kg and flaps 25 for landing on Runway 23. However, the interpretation of the weather and NOTAMs led the crew to plan on using Runway 14 in CYHZ. Using the runway analysis charts and the conditions expected in CYHZ at the time of arrival, the pre-departure maximum landing weight on Runway 14 was 261 500 kg for landing with flaps 25 and 279 400 kg for landing with flaps 30. Theses limits were not recorded on any flight documentation.

In establishing an accurate, shared mental model and situational awareness of potential threats to a flight, it is essential that crews determine accurate landing performance limits before departure. Based on the operational flight plan, the estimated landing weight on arrival in CYHZ was 265 852 kg; therefore, the occurrence flight only met the pre-departure maximum landing weight requirements using flaps 30.

2.2 Approach preparation

The departure from KORD and cruise portion of the flight were uneventful. The approach preparation was carried out before the descent, in accordance with the standard operating procedure (SOP). During this approach preparation, the crew obtained automatic terminal information service (ATIS) information Sierra, determined the approach speed and required landing distance for the existing conditions, and performed an approach briefing for an ILS approach to Runway 14.

A comparison of the computed fuel according to the operational flight plan and the actual fuel on board at the top of descent point indicated that the landing in CYHZ would be close to the initial estimated landing weight of 265 852 kg. Using the Quick Reference Handbook (QRH), the crew calculated the unfactored landing distance, using autobrakes 4, with flaps 25, as approximately 6000 feet. Since that calculation determined the distance required was substantially less than the 7700 feet of runway available, they chose to continue with the typical flaps 25 landing configuration. Post-flight calculations using the QRH guidance for the existing conditions, with flaps 25, autobrakes 4, and an approach speed of 159 knots resulted in an unfactored landing distance of 6375 feet.

In the occurrence flight, the crew calculated a V_REF of 154 knots and intended to disengage the autothrottle before landing. Given that Boeing’s minimum (command) approach speed is V_REF + 5 knots, the crew then calculated this speed to be 159 knots. The crew next added half of the steady wind of 230° magnetic (M) at 10 knots, for an approach speed of 164 knots; however, the wind was actually a 90° crosswind. Boeing recommends that no wind additive be applied in 90° crosswind or tailwind conditions. Therefore, a zero wind additive should have been used, and the calculated approach speed should have been V_REF + 5 knots, or 159 knots (see section 1.17.4.2, Approach speed).

Based on this higher approach speed of 164 knots, the investigation’s post-flight calculations showed an increased landing distance of 6735 feet. The available landing distance on Runway 14 is 7700 feet, leaving a safety margin of 965 feet. Under the existing conditions in CYHZ, there were no limitations or restrictions on the use of flaps 30 for the occurrence landing. Using the QRH data for landing with flaps 30, under the same conditions, also using V_REF + 10 knots, results in a landing distance required of 6241 feet compared to 6735 feet for flaps 25.

2.3 Descent and approach

At 0443, the crew began the descent from flight level (FL) 370 when the aircraft was  153 nautical miles (NM) from CYHZ. At 0446, the Moncton area control centre (ACC) controller asked the flight crew which approach they were planning, and the pilot monitoring (PM) advised that the plan was for the ILS approach to Runway 14. The Moncton ACC controller then cleared the flight directly to the intermediate fix TETAR (Appendix A).

At 0454, the tower controller advised the Moncton ACC controller that the primary runway was changing to Runway 23 via the RNAV 23 approach. This new landing information (a change in the approach in use) was not passed directly to the crew and was contrary to the crew’s understanding that Runway 23 was unavailable. Having access to this information would have improved the crew’s situational awareness and possibly reinitiated the decision-making process to consider Runway 23 as an option.

At 0458, the ATIS was updated to information Tango, with the only significant changes being the change of approach to RNAV 23 and the landing on Runway 23. At this moment, the Moncton ACC controller cleared the flight for an ILS approach to Runway 14.

At 0501, as the aircraft was 14 NM from the runway and about to intercept the final approach, the flight was transferred to the CYHZ tower, and the tower controller informed the crew that the winds were now 260°M at 15 knots and that the information was ATIS information Tango. The PM’s reply to the tower was that they had information Tango.

During this time, the aircraft was approaching the final approach course in the presence of a tailwind, and the crew was reducing the speed to configure the aircraft for the ILS. As a result, in the minutes preceding the transfer to the tower frequency, the crew made 24 statements and callouts related to the configuration of the aircraft. The tone of many of the captain’s statements indicated some apprehension, suggesting that he was seeking validation of his actions. The frequency and tone of communication were also indicative of a high workload, which can make it more challenging to effectively and correctly recognize the changes in conditions.

Given the high workload during this time, it is unlikely that the contents of ATIS Tango were reviewed in detail. In any case, the crew remained unaware that the approach had changed to Runway 23. Also, the crew did not understand that the information on winds just provided to them by the tower had changed significantly since ATIS information Sierra and Tango were issued, and that they now resulted in a 7-knot tailwind component for landing.

At 0502:46, when the aircraft was 8.6 NM from the threshold on its final approach, the CYHZ tower informed the crew that the winds were from 260°M at 16 knots, gusting to 21 knots, and asked the crew to confirm that Runway 14 was still acceptable. The PM confirmed that it was and the tower controller repeated the winds and cleared the flight to land on Runway 14.

The crew did not realize that the approach and landing runway had changed to Runway 23 and had not yet understood that the newly provided information on winds resulted in a 7-knot tailwind component for Runway 14. Since the crew had limited situational awareness of the conditions at CYHZ, they concluded that continuing the approach to Runway 14 was the only option available to them.

Over the next 20 seconds, the crew briefly reviewed calculations involving the wind strength and gusts to confirm that they would continue to use their planned approach speed of 164 knots, then performed the landing checklist. The captain’s apprehension regarding the upcoming approach, of which the other crew members were aware; subtle performance-impairing effects of fatigue; and elevated workload, likely acted to limit working memory and impair the crew’s ability to perform the normally easy approach speed computations.

In light of the tailwind present, the higher approach speed of 164 knots increased the landing distance required to 7514 feet, reducing the safety margin to 186 feet.

Just after passing the final approach fix, when the aircraft was 4.0 NM or 1 minute and 21 seconds from the threshold, the crew confirmed for the first time the presence of a tailwind (Table 8).

Table 8. Events on approach (Source: TSB, based on information obtained from the occurrence aircraft’s flight data recorder and from air traffic control recordings)

Time	Event	Wind direction and speed	Ground speed (knots)	Altitude (feet) or height (AGL)	Distance from threshold (NM)	Elapsed time to threshold
0504:13	Crew confirms presence of tailwind	260° 16 knots, gusting to 21 knots	185	1300	4.0	0:01:21
0504:46	Crew reviews go-around items	N/A	179	800	2.3	0:00:48
0504:58	Tower provides crew with wind information	250° 15 knots, gusting to 21 knots	176	600	1.7	0:00:36
0505:11	Crew disconnects autopilot	N/A	174	400	1.1	0:00:23
0505:34	Aircraft at threshold of Runway 14	N/A	179	62	0	0:00:00

When the aircraft was passing through 800 feet above ground level (AGL), the PF reviewed the go-around procedure with the crew. This go-around review at this late stage of the approach indicates that the PF was aware that the presence of a tailwind had further reduced the runway margin on this 7700 foot runway. A few seconds later, the CYHZ tower transmitted a “wind check 250°M at 15 knots, gusting to 21 knots,” resulting in a steady 4-knot tailwind component and a landing distance of 7211 feet. Shortly afterward, the crew had a short exchange about whether the PF was comfortable with the landing, and the crew agreed to continue the approach.

The tone, timing, and content of the crew communications during the flight indicated a limited degree of confidence in the execution of this landing. These communications just before landing and during a high-workload moment of the flight indicate heightened crew anxiety concerning the imminent crosswind landing.

Other than the limitation to have the autopilot disconnected by 150 feet AGL on an ILS approach, there was no company guidance as to when the autopilot should be disconnected on approach. In this occurrence, the captain disconnected the autopilot and autothrottle at 400 feet AGL and the airspeed increased to V_REF + 19 knots, as the aircraft deviated slightly left of the localizer and slightly above the glideslope. The airspeed increased to  173 knots as the aircraft crossed the runway threshold at a ground speed of 179 knots, higher than planned.

The aircraft crossed the threshold 14 knots faster than the required approach speed, with a 4.4 knot tailwind, and 12 feet above the glide path threshold-crossing altitude of 50 feet. These increased the landing distance required by 1368 feet, 460 feet, and 229 feet, respectively, which, in turn, increased the landing distance to 8088 feet.

Therefore, the higher airspeed and glideslope deviation, in the presence of the 4-knot tailwind resulted in the aircraft landing in conditions where it was not possible to stop on this 7700-foot wet runway. Although the aircraft was flown within the stable approach criteria, this occurrence demonstrates that, when operating very close to the performance limits of an aircraft, any deviations, no matter how small, may result in the aircraft no longer being able to be stopped within the confines of the runway.

In conditions of reduced runway margin and high PF workload and stress, timely and accurate PM callouts of deviations, such as increased approach speed, are crucial. In this occurrence, the 9-knot deviation further increased the landing distance required.

2.4 Landing

At 25 feet above the runway, all four thrust levers were reduced to the idle position. The firm (1.75g) aircraft touchdown occurred 1350 feet from the threshold (0505:38), which is consistent with a 3° descent from 62 feet to the runway surface.

One-half second later, reverse thrust was selected on engines Nos. 2, 3, and 4. Engine No. 1 thrust lever was then returned to near-idle, and the speed brake logic extended the speed brakes. Engine No. 1 reverse thrust was not selected, and the engine remained in forward thrust for the remainder of the landing rollout. The investigation was unable to determine the direct cause of the advancement of the engine No. 1 thrust lever.

Since the main landing gear is offset approximately 19 feet (left or right) from the centre of the aircraft, the firm touchdown, which was first on the right main gear, caused a left roll movement that peaked at 6°, when the left-hand wheel trucks compressed. This left bank, combined with the downward flex of the wing after the firm touchdown, likely resulted in the bottom nacelle of either engine No. 1 or No. 2 striking the surface of the runway; however, these engines continued to operate as commanded by the flight crew.

This combined right yaw tendency increased the rudder inputs required to regain the runway centreline. The rate of lateral displacement during the first 4 seconds after initial touchdown was approximately 6 feet per second. Based on this rate, if this displacement had not been reduced, a lateral runway side excursion would have occurred in approximately 10 seconds. Analysis of the FDR data indicated that the PF used rudder deflections between neutral and maximum left. The TSB laboratory analysis of the yawing moments and rudder effectiveness concluded that differential braking was not required for the pilot to regain the runway centreline.

The landing roll procedure of Sky Lease Cargo’s SOP requires both pilots to verify that the thrust levers are closed and the speed brake lever is UP, and to monitor the rollout progress and verify autobrake operation. Additionally, the SOP directs that the PM calls “speed brakes up” or, if they do not deploy, “speed brakes not up,” and then calls “reversers normal” or, if engine reversers are not all deployed, “no reverser(s) engine number ___.” However, the crew did not complete the company’s landing roll procedures during this occurrence.

The unexpected intensity of the landing impact, coupled with the lateral movement of the aircraft towards the runway’s right edge, surprised the flight crew, who were already experiencing high workload. The resulting acute situational stress heightened their potential to become distracted by highly salient stimuli. The flight crew’s attention was thus captured by the visual stimuli outside and ahead of the aircraft, and preventing a runway side excursion became their priority. The crew would have also been experiencing attentional narrowing, which limited their ability to detect and perceive other cues in the environment that would have indicated that deceleration devices had not deployed.

Because none of the landing rollout SOP calls were made, neither pilot was aware that

• the engine No. 1 thrust lever was advanced forward of idle,
• the speed brakes had retracted momentarily, and
• the engine No. 1 thrust reverser was not deployed.

The absence of SOP callouts to provide feedback to the PF that the speed brakes did not deploy as planned, that the autobrakes had disengaged, and that the engine No. 1 thrust reverser was not deployed, collectively increased the distance required to bring the aircraft to a stop, increasing the severity of the runway overrun. The captain’s misapplication of the No. 1 reverse thrust lever was a slip of attention: he intended to apply the reverse thrust lever correctly. However, because his attention was directed toward preventing the runway side excursion, he did not notice that the No. 1 thrust lever movement to the reverse position had been interrupted.

As this occurrence demonstrates, continuing a landing when the runway safety margin is reduced requires precise execution of the landing and the deployment of the deceleration devices as prescribed by the manufacturer, as any deviation increases the risk of a runway overrun.

In this occurrence, when the aircraft passed the end of the runway, it sustained damage beyond repair. Runway 14 at CYHZ has a 150 m (495 feet) runway safety area and a downward slope of 0.2%. Approximately 166 m (544 feet) past the end of Runway 14, there is a significant drop of 2.8 m (9 feet) at a downward slope of 73%. CYHZ meets Transport Canada’s (TC’s) standard, but not the International Civil Aviation Organization (ICAO)’s recommendations for a maximum longitudinal slope of 5%, and for a total runway safety area (RESA) of 300 m (984 feet).

2.5 Aircraft braking

To determine whether the braking efforts by the pilot were hindered by external factors that may have reduced the aircraft’s ability to decelerate, the investigation analyzed the following:

• Runway marks, friction, and drainage
• Precipitation amounts
• Dynamic and reverted rubber hydroplaning
• Cockpit ergonomics and brake application
• Aircraft tires, brakes, anti-skid, auto speed brakes, autobrakes
• Viscous hydroplaning and aircraft braking coefficient

Physical examination and analysis of the runway did not reveal any evidence that would have impeded the deceleration of the aircraft.

There was no evidence of tire skidding. Additionally, using the interpolated rainfall intensity value during the occurrence landing, it was estimated that the water depth on the runway was less than considered necessary for dynamic hydroplaning.

An analysis of ergonomic factors concluded that, when seated correctly using the design eye reference point locators, both pilots would have had full range of motion of the respective thrust levers and full rudder and brake pedal deflection. However, the investigation could not determine if the flight crew were seated correctly at the time of the occurrence.

Examination of the aircraft revealed no anomaly that would have affected the deceleration devices and TSB laboratory FDR analysis of these systems, indicating that they functioned as designed. Physical examination of the tires, brakes, and wheel torque limiters at the TSB laboratory found no evidence of any anomalies that would have reduced maximum braking.

The occurrence aircraft FDR was not capable of recording brake pedal position or the amount of brake pressure applied. However, since the investigation established that the PF was applying maximum braking effort in the last 800 hundred feet of runway remaining, FDR data was used to determine the deceleration rate of the aircraft for comparison with the theoretical wet runway deceleration using maximum braking effort. Wet runway aircraft braking coefficient (ABC) data provided values using maximum braking effort sufficient for anti-skid operation. The TSB laboratory calculated the occurrence’s actual ABC and found that, during the final 800 feet, it was consistent with the theoretical ABC on a wet runway. Having established that the maximum braking effort was consistent with the wet runway ABC data, a baseline for comparison with the landing roll up to that point was established.

The lower aircraft deceleration rate from touchdown to this point indicates a lesser braking effort. From these changes in deceleration rates, coupled with FDR and CVR data, it was determined that maximum braking effort did not occur until the aircraft was 800 feet from the end of the runway, which further exacerbated the extent of the runway overrun.

2.6 Factors contributing to runway overruns

The Flight Safety Foundation (FSF), the U.S. Federal Aviation Administration (FAA), and Boeing have all identified factors contributing to runway overruns.

Boeing states that a runway overrun is typically caused by multiple factors. The FSF, as well, found that runway overruns usually resulted from 1 or more factors involving weather, aircraft performance, crew technique and decision making, or aircraft systems. The factors present in this occurrence have all been identified as contributing to runway overruns:

• Landing with a tailwind
• Excess speed and height at the threshold
• Delayed use of the deceleration devices

According to the QRH, with flaps 25, the tailwind at the threshold increased the landing distance by 460 feet. Furthermore, the higher approach speed at the threshold (V_REF + 19) increased the landing distance by 1368 feet, and the extra 12 feet of height at the threshold increased the landing distance by 229 feet. Together, these factors increased the landing distance to 8088 feet for the landing on this 7700-foot runway.

While preparing for the approach, the crew used the actual landing distance charts to determine that the landing performance on Runway 14 was approximately 6000 feet. This may have given the impression that they had a 1700 foot runway margin on this 7700 foot runway. However, these charts are based on the approach and landing being performed precisely on speed, on profile, with a touchdown 4.22 seconds after passing the threshold and with deceleration devices used immediately after landing. In reality, the initial actual landing distance (V_REF +10) was 6735 feet, and increased to 7514 feet with the tailwind. This left 186 feet of runway remaining, which is a margin of 2.4%. The FAA considers a 15% margin between the expected airplane landing distance and the landing distance available at the time of arrival as the minimum acceptable safety margin for normal operations.

In this occurrence, using the actual landing distance data, the aircraft could have been stopped within the runway surface, up until it passed the threshold faster and higher than planned. However, a successful landing on the runway would have been possible only if the approach and landing had been executed precisely according to the conditions mentioned in the QRH.

If the flight crew had used the FAA recommended factored landing distance, they would have become aware at the briefing stage that the runway available was 45 feet less than required, rather than the 965 feet extra using actual landing distance data (Table 9).

Table 9. Unfactored versus factored landing distances (flaps 25) at various positions on approach (Source: TSB, based on Sky Lease Cargo, 747 Flight Crew Operations Manual: 747 Quick Reference Handbook, Revision 01 [05 May 2017], p. PI-QRH.12.2 and PI-QRH.12.3)

Position	Wind direction and speed (knots)	Wind component (knots)	Actual (unfactored) landing distance calculations (feet) Flaps 25		FAA-recommended factored landing distance (feet)
			Quick Reference Handbook unfactored	Extra runway	Factored 1.15	Extra runway
Briefing	230°, 10 knots	Headwind 0.5	6735	965	7745	–45
8.6 NM final	260°, 16 knots, gusting to 21 knots	Tailwind 7.3	7514	186	8641	–941
1.7 NM final	250°, 15 knots	Tailwind 4.4	7211	489	8293	–593
Runway threshold*	250°, 15 knots	Tailwind 4.4	8088*	–388*	9301*	–1601*

*  This landing distance includes the fact that the aircraft was 9 knots above the target speed used for the other calculations and 12 feet above the reference threshold crossing height of 50 feet.
NOTE 1: The unfactored landing distance was greater than the runway length when the aircraft crossed the runway threshold.
NOTE 2: The FAA-recommended factored landing distances were greater than the available runway length when the aircraft was 8.6 NM from landing, when it was 1.7 NM from landing, and when it crossed the runway threshold.

At the time of the occurrence, the company was in the process of updating the QRHs with factored landing distances, but had not done it for the occurrence aircraft. Since the QRH landing distance data for this aircraft was unfactored, it may have affected the crew’s situational awareness of the landing on Runway 14.

2.7 Pilot decision making

An important component of pilot decision making is the ability to recognize changes in a situation and reinitiate the decision-making process to ensure that necessary changes are made and plans modified. It is important to consider the context in which the crew were operating throughout the flight to understand why it made sense to them to continue the approach after acknowledging, at 1 minute 21 seconds from the threshold, that the winds had resulted in a tailwind, rather than to take the time necessary to reconsider the plan in light of the new information.

2.7.1 Team situational awareness

Accurate situational awareness is achieved through perception, understanding, and projecting a situation in time. Individuals outside of the aircraft (e.g., company dispatch, air traffic control, other aircraft crew) play a role in informing and calibrating team situational awareness by providing crews with accurate and timely information.

There were several instances during the flight in which information that was not communicated to the flight crew could have improved the accuracy of their team situational awareness.

During the approach preparation, ATIS information Sierra indicated that Runway 23 was to be used for departure; however, because of the misinterpretation of the NOTAMs, the crew believed that Runway 23 was not available for landing. As the occurrence flight descended, another aircraft that was on a different radio frequency landed on Runway 23. The tower controller advised the Moncton ACC controller that the approach runway was changed to 23 and, as the aircraft was cleared for the ILS runway 14, ATIS information Tango was updated with the change to Runway 23. However, the crew was not verbally informed by Moncton ACC that the approach runway had changed to Runway 23.

When the flight was subsequently transferred to CYHZ tower, the controller advised the crew that information Tango was current. The PM responded that they “had” ATIS information Tango. However, none of the crew members was aware of the change of landing runway. This lack of awareness that the landing runway had changed to Runway 23 limited the completeness and accuracy of their team situational awareness and extended the crew’s misunderstanding that this option was unavailable. As a result, they continued the approach to Runway 14.

2.7.2 Fatigue

The investigation included a thorough fatigue analysis, including consideration of the flight crew’s work schedule, their sleep history, and circadian rhythm timing. The analysis identified 2 fatigue risk factors that would have degraded the crew’s performance during the flight and at the time of the occurrence. First, the timing of the flight was during the nighttime circadian trough (2230 to 0430), when overall performance and cognitive functioning are at their worst. Second, the crew had not had sufficient restorative sleep in the 24-hour period leading up to the occurrence, which is considered an acute sleep disruption.

Efforts were made to provide opportunities for the crew to rest during the 13.5-hour departure delay. However, because it would have been difficult for the crew—who were used to sleeping at night—to sleep during the afternoon and evening, they were unable to obtain sufficient restorative sleep in the 24-hour period before the occurrence to avoid becoming fatigued.

As required by the Sky Lease fatigue risk management plan, the occurrence flight crew members had attended annual recurrent training in fatigue risk management that included material describing some of the more subtle performance impairments of fatigue. However, for the occurrence flight, this training was not salient enough for the crew to recognize and consider the more subtle effects of fatigue on performance when operating the early morning flight.

Sleep-related fatigue impairs working memory that is used for problem solving and reduces flexibility in an individual’s problem-solving approach to a situation. It also makes it difficult for the fatigued person to devise and try a novel solution, increasing the likelihood that the normal routine will be maintained and leading to a failure to revise the original plan.

A test of the influence of fatigue was conducted to understand whether the actions of the crew were consistent with what is known about human performance in a state of fatigue. In terms of influence, some elements of the crew’s performance and cognitive functioning were consistent with known performance impairments of fatigue, including: challenges in performing the normally easy approach speed computations, limitations in their ability to determine the effect of a tailwind, and limited flexibility to question the ongoing plan to land on Runway 14 despite the existence of new and relevant information.

Stress and workload can also limit working memory capacity and the ability to perform calculations that would otherwise be easy, and can negatively affect team performance by impairing team integration and mental models.

2.7.3 Cognitive influences

To cope with the large amount of sensory information in the environment at a given time, humans have developed normal cognitive coping skills to facilitate information processing, such as anchoring to the first credible piece of information (anchoring bias) or looking for information that confirms one’s current assessment or decision (confirmation bias). The captain’s relative inexperience in the aircraft type and the crew’s inaccurate situational awareness of the other available landing option on Runway 23 created conditions that can facilitate these cognitive biases. During pre-flight planning, the crew consulted with the deadheading captain, who was very experienced on the aircraft type and had recently flown the approach to Runway 23 and found it to be challenging. The deadheading captain confirmed the crew’s understanding that Runway 14 was appropriate for the weather and aircraft conditions, and that Runway 23 was not available to them. The crew’s mental model of the landing plan in CYHZ was thus reinforced, and an anchoring bias developed among the crew concerning this information. The deadheading captain was not present in the cockpit during the approach and landing, thus was not present when the new wind information was received.

During the flight, confirmation and plan-continuation biases limited the likelihood that the flight crew would seek out, detect, and identify relevant cues in the environment that would indicate that they should reconsider their plan to land on Runway 14.

In spite of these biases, because of his significant overall aviation experience, the captain was likely aware on a subconscious, affective level (i.e., had a “gut feeling”) that the runway margin was becoming critical as the flight continued. The captain’s briefing of the go-around on short final approach indicates that he was concerned, on some level, with the landing plan. However, being new on the aircraft type and having only performance data based on actual landing distance, he was not situationally aware of precisely how limited the runway margin was. The captain’s imprecise situational awareness of the landing distance computations made it more likely that he would rely on the knowledge and opinions of the more experienced crew members (e.g., the first officer and the international relief officer) to confirm the intention to continue the landing.

This continuation with the plan, despite some anxiety on the part of both the PF and PM, is consistent with research on decision making that has found that, under certain circumstances, cognitive influences such as anchoring, confirmation, and plan continuation biases can make individuals less likely to change a decision once it is made. In order to disrupt the plan or perform a go-around, so that the crew could take the time needed to reconsider the approach, a cue would have had to be sufficiently salient for the crew to detect, perceive, and understand its implications (i.e., that the aircraft was likely to overrun the runway).

The crew’s limited flexibility to challenge the ongoing plan to land despite the existence of new and relevant information is also consistent with some of the known performance impairment effects of fatigue.

2.7.4 Unanticipated tailwind

The style and sequence of the NOTAMs led the crew and flight dispatch to believe that Runway 23 was not available for landing in CYHZ. Because they had not reviewed the latest ATIS information, the crew did not realize that Runway 23 was available. However, the critical point on this approach was when the crew realized on short final that they were landing with a tailwind. Since they had insufficient time to recalculate the landing distance required and review their options, they were faced with a decision to continue or go-around. From their perspective, and according to the actual landing distance chart they were using, the landing could still be performed within the surface of the runway. As a result, the crew agreed to continue.

Although the aircraft was flown within the stable approach criteria, the accident demonstrates how several factors affecting landing distance, although individually small, can combine to result in a reduced margin of safety, in which any additional factor can push the approach to a point that the aircraft can no longer stop on the available runway surface.

When crew are landing on a short runway, there is a heightened sense of awareness that the landing margin is small, so that any deviations or change in condition must be either immediately corrected or the approach must be discontinued. However, when operating on a runway where a greater margin is present, the crew does not necessarily have the same heightened awareness. Although there was originally a reasonable runway margin of safety during this approach and landing, conditions changed to a tailwind, resulting in a significant reduction in length available.

From a perspective of managing operational threats and following mitigations recommended by many organizations, it appears clear in hindsight that a go-around was the safest course of action. However, analysis of this occurrence shows how fatigue, cognitive biases, workload, and stress can hinder any crew’s decision making. Since these influences may prevent crews from understanding that there is no longer an acceptable margin of safety for the upcoming landing, a trigger to re-evaluate should be a defense. Specifically, an unanticipated tailwind component should be a trigger to review the landing performance to determine whether the runway safety margin is still acceptable.

3.1 Findings as to causes and contributing factors

These are conditions, acts or safety deficiencies that were found to have caused or contributed to this occurrence.

1. The ineffective presentation style and sequence of the NOTAMs available to the crew and flight dispatch led them to interpret that Runway 23 was not available for landing at Halifax/Stanfield International Airport.
2. The crew was unaware that the aircraft did not meet the pre-departure landing weight requirements using flaps 25 for Runway 14.
3. Due to the timing of the flight during the nighttime circadian trough and because the crew had had insufficient restorative sleep in the previous 24 hours, the crew was experiencing sleep-related fatigue that degraded their performance and cognitive functioning during the approach and landing.
4. Using unfactored (actual) landing distance charts may have given the crew the impression that landing on Runway 14 would have had a considerable runway safety margin, influencing their decision to continue the landing in the presence of a tailwind.
5. When planning the approach, the crew calculated a faster approach speed of reference speed + 10 knots instead of the recommended reference speed + 5 knots, because they misinterpreted that a wind additive was required for the existing conditions.
6. New information regarding a change of active runway was not communicated by air traffic control directly to the crew, although it was contained within the automatic terminal information service broadcast; as a result, the crew continued to believe that the approach and landing to Runway 14 was the only option available.
7. For the approach, the crew selected the typical flap setting of flaps 25 rather than flaps 30, because they believed they had a sufficient safety margin. This setting increased the landing distance required by 494 feet.
8. The crew were operating in a cognitive context of fatigue and biases that encouraged anchoring to and confirming information that aligned with continuing the initial plan, increasing the likelihood that they would continue the approach.
9. The crew recognized the presence of a tailwind on approach 1 minute and 21 seconds from the threshold; likely due to this limited amount of time, the crew did not recalculate the performance data to confirm that the runway safety margin was still acceptable.
10. An elevated level of stress and workload on short final approach likely exacerbated the performance-impairing effects of fatigue to limit the crew’s ability to determine the effect of the tailwind, influencing their decision to continue the approach.
11. The higher aircraft approach speed, the presence of a tailwind component, and the slight deviation above the glideslope increased the landing distance required to a distance greater than the runway length available.
12. After the firm touchdown, for undetermined reasons, the engine No. 1 thrust lever was moved forward of the idle position, causing the speed brakes to retract and the autobrake system to disengage, increasing the distance required to bring the aircraft to a stop.
13. The right crab angle (4.5°) on initial touchdown, combined with the crosswind component and asymmetric reverser selection, caused the aircraft to deviate to the right of the runway centreline.
14. During the landing roll, the pilot monitoring’s attention was focused on the lateral drift and, as a result, the required callouts regarding the position of the deceleration devices were not made.
15. The pilot flying focused on controlling the lateral deviation and, without the benefit of the landing rollout callouts, did not recognize that all of the deceleration devices were not fully deployed and that the autobrake was disengaged.
16. Although manual brake application began 8 seconds after touchdown, maximum braking effort did not occur until 15 seconds later, when the aircraft was 800 feet from the end of the runway. At this position, it was not possible for the aircraft to stop on the runway and, 5 seconds later, the aircraft departed the end of the runway at a speed of 77 knots and came to a stop 270 m (885 feet) past the end of the runway.
17. During the overrun, the aircraft crossed a significant drop of 2.8 m (9 feet) approximately 166 m (544 feet) past the end of the runway and was damaged beyond repair. While this uneven terrain was beyond the 150 m (492 feet) runway end safety area proposed by Transport Canada, it was within the recommended International Civil Aviation Organization runway end safety area of 300 m (984 feet).

3.2 Findings as to risk

These are conditions, unsafe acts or safety deficiencies that were found not to be a factor in this occurrence but could have adverse consequences in future occurrences.

1. If the pilot monitoring does not call out approach conditions or approach speed increases, the pilot flying might not make corrections, increasing the risk of a runway overrun.

3.3 Other findings

These items could enhance safety, resolve an issue of controversy, or provide a data point for future safety studies.

1. The investigation concluded that there was no reverted rubber hydroplaning and almost certainly no dynamic hydroplaning during this occurrence.
2. Although viscous hydroplaning can be expected on all wet runways, the investigation found that when maximum braking effort was applied, the aircraft braking was consistent with the expected braking on Runway 14 under the existing wet runway conditions.

4.1 Safety action taken

4.1.1 Sky Lease Cargo

The Board is aware of some safety actions that were taken by the operator following the occurrence; however, the TSB did not receive sufficient documentation to validate these actions.

4.1.2 NAV CANADA

NAV CANADA published a bulletin to highlight the importance of issuing landing information to pilots and to remind controllers of the procedures in the Landing Information section of the Manual of Air Traffic Services.

This report concludes the Transportation Safety Board of Canada’s investigation into this occurrence. The Board authorized the release of this report on 21 April 2021. It was officially released on 29 June 2021.

Appendix A – Halifax/Stanfield International Airport ILS RWY 14 approach chart

Source: Jeppesen

Appendix B – Events during the landing roll out

Position marked	Time	Activity
1	0505:37.95	Aircraft initially touches down
2	0505:38.30	Auto speed brake lever moves to the UP position
3	0505:38.56	No. 1 thrust lever is advanced above idle
4	0505:39.16	Auto speed brake lever moves to the DOWN position
5	0505:39.31	Thrust levers No. 2, No. 3, and No. 4 are brought into reverse thrust
6	0505:40.42	Air-ground logic switches indicate a momentary AIR mode logic position
7	0505:40.90	Air-ground logic switches indicate a GROUND mode logic position.
8	0505:41.69	No. 2, No. 3, and No. 4 thrust reversers are deployed
9	0505:44.02	No. 1 thrust lever is reduced; auto speed brakes are deployed to 100%.
10	0505:49.54	Maximum lateral deviation to the right of centreline
11	0505:59	Deceleration rate increases markedly
12	0506:06	Aircraft departs paved surface
13	0506:16 (estimated)	Aircraft comes to rest

Source: Google Earth, with TSB annotations

Appendix C – Ground scars

Appendix D – NOTAMs for Halifax/Stanfield International Airport

Source: Sky Lease Cargo

Appendix E – Aerodrome diagram for Halifax/Stanfield International Airport

Source: Jeppesen

Source: Jeppesen

Appendix F – Significant events on approach and landing

Table F1. Significant events on approach and landing

Time	Event	Wind direction and speed	Ground speed (knots)	Altitude (feet AGL)	Distance from runway threshold (NM)	Elapsed time to threshold
0400:00	Weather in ATIS S	230° 10 knots	583	FL370	562	1:05:34
0430:00	Approach preparation and briefing	N/A	596	FL370	265	0:35:34
0443:00	Aircraft begins descent (at FL370)	230° 10 knots	566	36 857	153.4	0:22:34
0454:00	Tower informs ACC that active runway has been switched to Runway 23	N/A	455	15 000	57	0:11:34
0457:09	Aircraft leaves 10 000 feet	N/A	330	10161	34.8	0:08:25
0458:00	ATIS T issued	230° 10 knots	340	9000	30.3	0:07:34
0458:19	ACC cleared for ILS RWY14	N/A	340	8481	28.6	0:07:15
0458:47	ATC cleared aircraft to TETAR (17 NM from runway threshold)	N/A	333	7774	26.1	0:06:47
0459:16	Flaps 1 called	N/A	324	6740	23.4	0:06:18
0459:40	Flaps 5 called	N/A	310	6567	21.4	0:05:54
0500:10	Flaps 10 called	N/A	287	6036	18.9	0:05:24
0500:20	Glideslope captured	N/A	279	5871	18.2	0:05:14
0501:02	ACC hands over aircraft to tower	N/A	259	5242	15.3	0:04:32
0501:22	Crew contacts tower, receives wind check: winds at 260° and 15 knots per ATIS T	260° 15 knots	257	4889	14.0	0:04:12
0501:33	PM acknowledges ATIS T	N/A	256	4630	13.3	0:04:01
0502:15	Localiser captured (aircraft is on glideslope)	N/A	264	3114	10.5	0:03:19
0502:29	GPWS callout: aircraft at 2500 feet AGL	N/A	245	2980	9.7	0:03:05
0502:46	Tower contacts aircraft to confirm Runway 14 still acceptable	260° 16 knots, gusting to 21 knots	212	3008	8.6	0:02:48
0503:05	PM confirms that Runway 14 still acceptable	N/A	195	2921	7.5	0:02:29
0502:58	Tower clears aircraft to land	N/A	198	2976	7.9	0:02:36
0503:30	Crew confirms VREF	N/A	188	2457	6.2	0:02:04
0503:37	Crew starts landing check	N/A	184	2345	5.8	0:01:57
0504:03	Crew completes landing check	N/A	184	1930	4.5	0:01:31
0504:10	Aircraft reaches final approach fix IMANO	N/A	185	1817	4.1	0:01:24
0504:13	Crew confirms presence of tailwind	N/A	185	1300	4.0	0:01:21
0504:46	Crew reviews go-around items	N/A	179	800	2.3	0:00:48
0504:58	Tower provides crew with wind information	250° 15 knots, gusting to 21 knots	176	600	1.7	0:00:36
0505:11	Crew disconnects autopilot	N/A	174	400	1.1	0:00:23
0505:34	Aircraft at threshold of Runway 14	N/A	179	62	0	0:00:00

Table F2. Sequence of events on landing, showing ground speed and amount of runway remaining

Time	Event	Ground speed (knots)	Runway remaining (feet)	Elapsed time from touchdown
0505:34	Aircraft at threshold of Runway 14	179	7700	−00:00:04
0505:38	Touchdown	179	6349	0:00:00
0505:40	Reverse thrust for engines 2, 3, and 4 selected	177	5944	+0:00:02
0505:46	Manual braking	152	3596	+0:00:08
0505:50	Maximum lateral deviation	144	3069	+0:00:12
0505:53	Maximum braking called	130	2500	+0:00:15
0506:01	Maximum braking applied	100	800	+0:00:23
0506:06	End of runway	77	0	+0:00:28

Appendix G – Pre-departure landing limitations – runway analysis charts

Source: Sky Lease Cargo

Source: Sky Lease Cargo

Appendix H – Actual (unfactored) landing distance charts

Table H1. Reference values used by the TSB for the landing distance adjustments

Weight	Pressure altitude	Wind	Slope	Temp	VREF	Reverse thrust	Extra height at the threshold
265 852 kg	716 feet	See chart below  (Position)	+0.54°	15 °C	+5 knots*	All operative	+12 feet

* V_REF + 14 knots (actual indicated airspeed) is used at the position "Runway threshold"

Table H2. Reference values used by the TSB for the landing distance adjustments

Weight	Pressure altitude	Wind	Slope	Temp	VREF	Reverse thrust	Extra height at the threshold
265 852 kg	716 feet	See chart below (Position)	+0,54°	15 °C	+5 knots*	0	+12 feet

* V_REF + 14 knots (actual indicated airspeed) is used at the position "Runway threshold"

ABC

aircraft braking coefficient

AC

advisory circular

ACARS

aircraft communication addressing and reporting system

ACC

area control centre

AGL

above ground level

ALS

approach lighting system

ALSF-2

approach lighting system with sequenced flashing lights for category II or III operations

ARFF

aircraft rescue and firefighting

ATC

air traffic control

ATIS

automatic terminal information service

ATPL

airline transport pilot licence

ATSB

Australian Transport Safety Bureau

CARs

Canadian Aviation Regulations

CRM

crew resource management

CYHZ

Halifax/Stanfield International Airport, Nova Scotia

CYYZ

Toronto/Lester B. Pearson International Airport, Ontario

CVR

cockpit voice recorder

EICAS

engine indicating and crew alerting system

FAA

U.S. Federal Aviation Administration

FARs

U.S. Federal Aviation Regulations

FAST

fatigue avoidance scheduling tool

FCTM

flight crew training manual

FDR

flight data recorder

FL

flight level

FO

first officer

FSF

Flight Safety Foundation

GNSS

global orbiting navigation satellite system

HIAA

Halifax International Airport Authority

IAS

indicated airspeed

ICAO

International Civil Aviation Organization

ILS

instrument landing system

inHg

inches of mercury

IRO

international relief officer

KBGR

Bangor International Airport, Maine, U.S.

KORD

Chicago/O’Hare International Airport, Illinois, U.S.

KSFO

San Francisco International Airport, California, U.S.

LNAV

lateral navigation

LPV

localizer performance with vertical guidance

M

magnetic

METAR

aerodrome routine meteorological report

NDB

non-directional beacon

NM

nautical mile

NPA

notice of proposed amendment

NTSB

U.S. National Transportation Safety Board

PANC

Ted Stevens Anchorage International Airport

PAPI

precision approach path indicator

PF

pilot flying

PIREP

pilot report

PM

pilot monitoring

QRH

Quick Reference Handbook

RCLL

runway centreline lights

RESA

runway end safety area

REV

annunciator when the related reverser is unlocked or moving

RNAV

area navigation

RNP

required navigation performance

RTHL

runway threshold lights

RTZL

runway touchdown lights

SAFO

Safety Alert for Operators

SM

statute mile

SOCC

System Operations Control Center

SOP

standard operating procedure

SPECI

aerodrome special meteorological report

SSALR

simplified short-approach lighting system with runway alignment indicator lights

T

true

TAF

aerodrome forecast

TC

Transport Canada

TP 312

Aerodrome Standards and Recommended Practices

UOM

Unit Operations Manual

VNAV

vertical navigation

V_REF

reference speed

ZGHA

Changsha/Huanghua Airport, Hunan, China