Description

On 19 October 2022, the first officer on a de Havilland DHC8-300 (C-GJYZ) operated by Perimeter Aviation on a scheduled domestic passenger flight from Pikangikum to Sandy Lake as JV464 flew an unstabilised approach at the destination in night visual conditions (VMC).  The first officer then mishandled the landing attempt, after which the captain took control. Substantial damage was caused to the lower aft fuselage structure by the tailstrike, but there were no confirmed injuries to the 31 occupants.

Investigation

An accident investigation was carried out by the Canadian Transportation Safety Board. The flight data recorder (FDR) and cockpit voice recorder (CVR) were removed from the aircraft and their data were downloaded. Flight management system (FMS) data and nonvolatile memory (NVM) data were also downloaded from the aircraft.

The captain had a total of 7,721 hours flying experience, including 3,233 hours on the DHC8. He had joined Perimeter Aviation as a first officer 3½ months prior to the accident, and after 20 hours initial company training had passed his License Proficiency Check (LPC) and commenced line training in that rank. He had been checked out as a first officer after 69 hours of line training (minimum required 20 hours) following which he had been promoted to captain. In that rank, he had then completed 126 hours line training (minimum required 20 hours) before being released to fly in command. The accident flight was the first sector of his second day in command.

The first officer had a total of 2,368 hours flying experience. Prior to joining Perimeter Aviation, most of his time had been logged on the Cessna 172. He completed 13 hours initial company training to pass his LPC and then completed 74 hours line training (minimum required 20 hours) before being released to fly without training supervision. The accident flight was the first sector of his third day unsupervised, having accumulated 18½ hours in that time.

It was noted that the company operations manual (OM) did not include restrictions on pairing of type-inexperienced pilots, and that the regulatory requirements in that respect had already been satisfied by their extended periods of supervised line training. However, the operator’s monthly pilot roster was subject to management approval and had not initially rostered the two pilots to fly together because of “an informal practice of not scheduling pilots with limited experience in their roles on the same flight." A subsequent change was then made during a daily reschedule by the Senior Operations Control Centre (OCC) Duty Manager as this “informal daily scheduling practice” took no account of experience when pairing pilots to operate together at short notice.

What Happened

The departure of the flight was delayed by three hours. It eventually departed Pikangikum for the half-hour flight to Sandy Lake with the first officer acting as pilot flying (PF) and climbed to 7,000 feet QNH. A visual approach was planned at Sandy Lake with the LPV approach procedure to be followed. There was a brief discussion about the Abbreviated Precision Approach Path Indicator (APAPI) installed at Sandy Lake. The 1069-metre-long, 300-metre-wide gravel runway was classified as a ‘short runway’ by Perimeter Aviation, requiring a short-field landing technique to cross the threshold at around 30 feet agl instead of the normal 50 feet agl. Shortly before descent was commenced an ‘LPV APPR INHIBITED’ message was annunciated but given that visual conditions (VMC) prevailed, this did not create problems. The final approach track was joined at 10nm, and the aircraft briefly levelled at 4,000 feet QNH (the airport elevation was 946 feet), before descent was commenced about 500 feet above the 3.29° APAPI approach angle.

When 3nm from the runway threshold at 120 KCAS (and at 1,350 feet aal, the aircraft was still 300 feet above the APAPI slope guidance), the power was set to 15% torque, the propeller levers were set to 1,200 fpm, the flaps were set to 15° and the landing gear was selected down. One minute later, passing the 500 feet agl stabilised approach gate applicable to all approaches, with 1.2nm to go and descending at 900 fpm, the captain announced that the approach was stable “and advised the first officer to maintain the descent profile." However, he simultaneously reduced the rate of descent to 500 fpm and increased engine torque.

The captain mentioned the APAPI indication and cautioned that “an adjustment needed to be made to maintain the correct approach path." Just over half a minute later, with the aircraft still above the APAPI approach slope and just over half a mile from the runway, the first officer returned the engine torque to approximately 10%. As the aircraft reached the APAPI approach slope, the first officer increased the torque from 10% to almost 60% which reduced the rate of descent to approximately 160 fpm. With just 0.2 nm to go, the first officer then reduced the power levers to flight idle - approximately 0% torque - and decreased the aircraft pitch angle by 2°.

During the subsequent flare, the first officer attempted to reduce the excessive rate of descent he had created by making “a pronounced pitch-up control input” as the torque “began to gradually increase from flight idle." The main landing gear then touched down 350 feet past the threshold at a 950 fpm rate of descent. The maximum recorded vertical acceleration of 3.61g was recorded as the oleos compressed, and the aircraft’s aft fuselage struck the runway surface with the pitch angle recorded as 7.24°. The aircraft bounced to approximately 1.5 feet, and as it touched down for a second time, the captain saw that the ‘TOUCHED RUNWAY’ indicator had illuminated. When, moments later, the first officer called for a go-around and began to increase power, the captain took control and brought the aircraft to a stop with 487 metres of runway remaining, and then taxied it to the apron.

Substantial lower rear fuselage structural damage was subsequently identified, including airframe structure interior buckling where impact had occurred (as shown below). All main landing gear load-absorbing components were also replaced.

Buckled Structure beneath the floorboards above the missing external tail strike switch. [Reproduced from the Official Report]

Why It Happened

The unstabilised approach which preceded the mismanaged landing was examined against company requirements and guidance. It was noted that there were no regulatory requirements regarding stabilised approaches, although such approaches are “recommended." However, it was found that ‘Stabilised Constant Descent Angle’ (SCDA) training was in place and included in the DHC8 fleet’s Flight Crew Training Manual (FCTM). A 500 feet agl requirement for all approaches to be stabilised was defined as including a rate of descent not more than 1,000 fpm without a special briefing and a power setting “appropriate for the aircraft configuration and not below the minimum power for the approach” as defined by the Operations Manual (OM). But no such definition was found in that Manual.

It was noted that the closest recommended power setting mentioned in the Standard Operating Procedures (SOP) regarding conditions at the time of the event was “a power setting of 10% torque, the flaps set to 15°, both engines operating and an aircraft speed of 120 knots,” which it was stated “would result in a 1,000 fpm rate of descent." Once below 500 feet agl, the SOPs required that “only minimal adjustments should be made to the power lever settings, and the aircraft should be trimmed for the appropriate energy state so that (it) follows a consistent path."

In respect of the immediate prelude to touchdown for the flap 15° approach flown, it was found that the SOPs stated that when the aircraft is 20 feet above touchdown, the pilot should “smoothly bring the aircraft nose up into the flare, while reducing power as required allowing the aircraft to descend to the runway” and advised that “during the flare, the pitch attitude change is approximately 4° or 5°” i.e. from approximately 0° to 1° nose up to 4° to 5° nose up. The required (tailstrike prevention) pitch awareness calls at 5° and 6° were not made.

Regarding the visual approach flown without the LPV vertical profile available, it was noted that the pilots were not provided with any guidance on the conduct of visual ‘straight-in’ approaches nor was a Visual Flight Rules (VFR) descent profile published, only a VFR Circuit profile.

The operator’s use of periods of line training for both pilots were between three and more than six times more than the minimum required. The context was a staffing shortage: Only 75% of the budgeted roster of 40 captains and 44 first officers to operate 12 DHC-8 aircraft were employed. It was also noted that in the 12 months prior to the accident, there had been “a significant turnover of company DHC-8 pilots” with 34 pilots having left the company with only 18, including the two flying the accident aircraft, having been recruited to replace them.

To provide a perspective for the accident, the way the aircraft type was being operated was examined using the FDR data for a total of 246 flights. The maximum pitch attitude during the landing flare was determined for each of these flights and the results plotted (see below). Although the SOPs stated that the pitch should not exceed 6° during the flare to avoid the risk of a tail strike, the data showed that it did so in almost 9% of landings. A hard landing at 7° pitch attitude will depress the main gear oleos sufficiently, as in this case, to cause a tail strike. It was observed that since the operator relied entirely on pilots self-reporting deviations from prescribed SOPs, only such deviations with visible consequences were likely to be reported.

The operational safety oversight process of the Perimeter Aviation operation at the time of the event was found to have been ineffective, with a systemically ineffective but regulator-approved safety management system (SMS). The SMS was not tracking risk in everyday operations despite that this was strongly recommended (although not a regulatory requirement). The absence of a recommended but not required Operational Flight Data Management (OFDM) programme was also highlighted as a particular missed opportunity. This could improve risk awareness and help monitor compliance.

Reproduced from the Official Report

The following Findings were formally documented based upon completion of the investigative work:

Causes and Contributing Factors

1. Below 500 feet above ground level and while trying to intercept and maintain the appropriate approach path, the pilot flying varied the power setting between 57% and flight idle, likely owing to limited experience operating the aircraft type, and the result was an unstable approach.
2. Due to insufficient detail in the standard operating procedures and the absence of awareness training on stabilized approach criteria, the pilots did not recognise that significant variations in the power setting had made the approach unstable, and they continued the approach.
3. The pilot flying, who was relatively inexperienced on the DHC-8 and had received limited guidance on pitch awareness, made a pronounced pitch-up input during the flare. There was insufficient time for the pilot monitoring to arrest this action, and the aircraft’s aft fuselage consequently contacted the runway, causing significant damage.

Risk Factors [Safety deficiencies which were assessed not to have been a factor in this occurrence but could be in future events.]

1. If an air operator does not factor a pilot’s level of experience into a formal daily crew scheduling process, there is a risk of compromised safety margins due to the assignment of two pilots with limited experience in their roles to operate the same flight.
2. If air operators rely solely on flight crews to report hazards and risks to the company’s safety management system and do not actively monitor flight operations through flight data-monitoring capabilities, unsafe practices may not be identified, increasing the risk that they will continue.

Safety Action taken by Perimeter Aviation as a result of the accident was noted at the conclusion of the Investigation as having included the following:

• incorporated the “Dash 8-Q400 Pitch Awareness” video in its initial and recurrent cockpit procedures training for the DHC-8-100 and DHC-8-300 series and made the video available for instructors to share;
• amended the DHC-8 standard operating procedures (SOPs) to revise its stabilised approach criteria to include target power settings, and added information and guidance regarding the flight management system’s “LPV APPR INHIBITED” error message;
• amended the SOPs for the DHC-8, SA227, and SA226 to include an instrument approach policy that requires flight crews to fly the instrument approach procedure for the intended runway, if one is available, regardless of weather conditions, to assist in ensuring a stabilised flight profile;
• developed flight operations quality assurance and line operations safety audit procedures;
• added this occurrence to the company’s crew resource management course;
• implemented a command and decision making course;
• amended DHC-8 initial simulator training to include excessive pitch recovery and black hole exercises;
• implemented a restricted crew status list;
• instituted a flight data monitoring program for its DHC-8 and SA227 AC fleet.

The Final Report of the Investigation was authorised for release on 27 March 2024 and officially released on 21 May 2024. No Safety Recommendations were made.

Related Articles

• Stabilised Approach
• Tail Strike
• Hard Landing
• Low Level Go Around (SKY Clip) 
• Safety Management System
• Flight Data Monitoring (FDM)
• Risk Management
• Safety Oversight
• Standard Operating Procedures (SOPs)