Description

On 20 January 2020, a Bombardier DHC8-300 (C-GXAI) being operated by Air Inuit on a scheduled domestic passenger flight from Québec to Schefferville as AIE620 continued a comprehensively unstable visual approach at destination and ineffective control of the aircraft at touchdown resulted in a hard landing with significant structural damage to the aft lower fuselage. No injuries were sustained by the 45 occupants.

Investigation

Following a delay in notification, an Investigation was carried out by the Canadian Transportation Safety Board (TSB) assisted by data downloaded from the FDR and the 30 minute CVR of the aircraft involved. It was noted that although the CVR had not been secured, the aircraft had been shut down immediately after completion of the flight and had remained electrically unpowered thereafter thus protecting the relevant data. Also, since the parameters recorded in the FDR did not include GPS position, the Investigation compared radio altimeter data with the topography of the approach terrain to establish an approximate approach flight path.  

It was noted that the Captain, who was acting as PM for the flight involved and had been working for the operator for 15 years, had a total of 10,186 hours flying experience which included 5,024 hours on type which was recorded as having been accumulated over 4½ years. The First Officer had joined the operator in 2017 and had a total of 1,055 hours flying experience which included 82 hours on type having gaining the type rating two months earlier. 

What Happened

The aircraft and crew were rostered for a seven sector flight sequence which was all identified as flight AIE820 beginning from Montréal and ending at Salluit with the accident flight being the second sector of this sequence. After an uneventful first sector from Toronto and a half hour turnround, the flight took off from Québec for Schefferville with an expected time of just over 2 hours and with the First Officer acting as PF. 

As the cruise FL 230 continued, the crew began preparing for approach and landing at Schefferville and noted that the destination AWOS was indicating CAVOK and light winds making it likely that a visual approach to runway 35 would be straightforward and the First Officer then briefed for this. When the aircraft levelled at 5000 feet QNH during the subsequent descent, the First Officer asked the Captain to enter the circuit altitude of 2800 feet QNH. At that point, haze was reducing visibility and it was not possible to see the aerodrome. The Captain then activated the ARCAL (Aircraft Radio Control of Aerodrome Lighting) and the First Officer disconnected the AP and began a right turn onto a base leg to intercept the extended centreline of the still unseen runway 35 - which almost immediately then came into view.

When around 2½ nm out on left base leg and about to turn onto final approach, the First Officer realised that the aircraft was too fast (162 KIAS) and too high (480 feet above a nominal 3° slope) so he reduced power and asked the Captain to set 1050 propeller rpm and this was achieved as the aircraft was turned left towards final approach at which point the landing gear was selected down. It passed through the final approach and the left turn was continued, peaking at 36° - which was enough to activate a brief EGPWS Bank Angle alert. Passing 750 feet agl, the aircraft was still in the turn and still to the right of final approach and a second EGPWS alert occurred, this time by the rate of descent which, with the aircraft still at 162 KIAS, had momentarily reached 2500 fpm. 

The First Officer reduced the rate of descent to approximately 1000 fpm and finally completed the turn onto final approach and the Captain, having assessed that “the approach and landing were feasible, even though the aircraft was flying above the nominal 3° slope at a high speed” began advising how to manage the height and speed and performed the descent checklist. As the aircraft descended through 500 feet agl with the power at flight idle, flaps 5 and the speed reducing through 144 KCAS (the calculated Vref was 99 KIAS) it was still 150 feet above a nominal 3° slope. The First Officer then asked for flaps 15 and the landing checklist which the Captain did as well as “sounding a chime twice to notify the cabin crew that landing was imminent”. Passing 200 feet agl, with the speed at 120 KCAS and 1000 fpm rate of descent, the First Officer asked for flaps 35°, the propeller speed to be increased to 1200 rpm and the final checklist, the latter being quickly completed.

The aircraft passed 100 feet agl with the speed now below Vref and rapidly decreasing but with the power still at flight idle. The PM called out the Vref speed and almost immediately, “Vref-5”. As the aircraft passed 50 feet agl descending at almost 900 fpm at 94 KCAS, the Captain told the First Officer to add power which he did to only a small extent whilst increasing the nose-up pitch attitude from 2° to 9° in less than two seconds. The result was touchdown close to the beginning of the 1525 metre-long dry and uncontaminated runway on the main gear and lower aft fuselage with a 2.37 g impact.

During the taxi in, the crew stated that they had noted that the “Touched Runway” indicator was illuminated and so after engine shutdown had carried out an external inspection of the aircraft in accordance with the tail strike checklist. They then contacted the company maintenance department who advised that a level 1 hard landing inspection would be required. When this was subsequently carried out, the extent of the structural damage was such that it was necessary to arrange a ferry flight to a repair centre at Trois-Rivières.  

Discussion

The flight was carried out under sub part 705 of the Transport Canada Civil Aviation Regulations (CARs) which is the highest level in a four part categorisation. As such, the operation was in theory covered by an SMS and SOPs which would be acceptable in a major airline. Despite this, checklist completion and stabilised approach requirements were comprehensively ignored throughout the approach which preceded the hard landing.

Having recognised that the continuation of the clearly unstabilised approach could be attributed to confirmation bias leading to plan continuation bias, a number of observations arising from the factual evidence collected were made including, in summary, the following:

• Whilst the operator’s approach SOPs described IFR approaches in detail, they did not describe visual approaches and the only annotated visual flight profile was for a VFR circuit. 
• The SOPs in use, including a number applicable to stabilised approaches, were found to be not always consistent with the corresponding OM content.
• An examination of pilot recurrent training delivery against regulatory requirements for the Captain found that there were significant failures to meet documented requirements.
• The rushed and wholly unstabilised approach under investigation was not an isolated deviation. An examination of all 144 flights on the aircraft FDR found that 16% exceeded the operator’s stabilised approach speed limit below FL050.
• There were serious inadequacies in the both the specified methods and practical conduct of the regulatory oversight of operational safety which were directly relevant to the aim of eliminating unstabilised approaches both at the operator involved and more generally. As a consequence, Transport Canada had not identified any issues with the operator’s stabilised approach guidelines as applied to frequently conducted visual approaches or recognised that unstable approaches were a systemic problem. 

The seven formally-stated Findings as to Causes and Contributing Factors were as follows:

1. The pilots forgot to perform the descent checklist and realised this at an inopportune time, when the pilot monitoring was providing a position report.
2. Given the ambiguities and contradictions in the stabilised approach guidelines, the Captain interpreted that he was allowed to continue the approach below 500 feet above aerodrome elevation although the flaps had not been set to 35° and the final checklist had not been completed. 
3. Communicating with the flight attendant to confirm the cabin status and performing the descent checklist during final approach added to the pilots’ workload, which was already heavy.
4. The combination of the visual conditions and the plan continuation bias prompted the pilots to continue managing the height and speed deviations past the stabilised approach gate.
5. When the aircraft passed 500 feet above aerodrome elevation, the pilots, who were dealing with a heavy workload, passed the stabilised approach gate without noticing it and continued the approach, which was de facto unstable.
6. At the time of the flare, the aircraft no longer had enough energy to stop the rate of descent solely by increasing the pitch attitude.
7. The instinctive reaction to increase the pitch attitude, combined with the hard landing, resulted in the aft fuselage striking the runway, causing major damage to the aircraft structure.

Six formally-stated Findings as to Risk were also made as follows: 

1. If a procedure is interpreted in a way that limits communications concerning cabin safety, there is a risk that pilots will not be informed that the cabin has not been secured and that passengers will receive injuries on landing, particularly if there is an abnormal landing.
2. If there is no reminder of the altitude of the stabilized approach gate for each approach, this stabilised approach gate may be missed, and an unstabilised approach may be continued, increasing the risk of an approach-and-landing accident.
3. If Transport Canada Civil Aviation does not assess the quality, consistency, accuracy, conciseness, clarity, relevance, and content of standard operating procedures, the procedures may be ineffective, increasing risks to flight operations. 
4. If required training elements are not included in recurrent training, there may be procedural deficiencies or deviations, increasing risks to flight operations.
5. If Transport Canada Civil Aviation’s surveillance plan does not verify the content of crew training, deviations may not be identified and procedural deficiencies or deviations may not be corrected, increasing risks to flight operations.
6. If information that is essential to flight operations for a particular aircraft type is not distributed directly to the operators of that aircraft type, there is a risk that those operators will not have all the resources needed to develop procedures and training that will prevent incidents or accidents.

One ‘Other Finding’ was made:

• Data from the cockpit voice recorder were not secured after the accident, and the accident was not reported to the TSB until the next day. Data from the cockpit voice recorder specific to the accident were nevertheless available to TSB investigators.  

Safety Action taken by Air Inuit as a result of the investigated event prior to the completion of the Investigation was noted as having included the following:

• A previously issued aircraft manufacturer video entitled “Dash 8 Q400 Pitch Awareness” was added to all initial and recurrent pilot training details.
• The review of pitch awareness during preparation briefings for simulator training was expanded. 
• DHC8 SOPs were revised to improve guidelines on several subjects, including visual approaches and the identification of stabilised approach gates. 
• The pilot training program was revised to ensure that all training elements are covered within the 2-year cycle for recurrent training.

Editor's Note: Schefferville had both RNAV(GNSS) and NDB approaches to its runway available so the visual approach chosen by the crew when poor in-flight visibility initially delayed their sighting of the runway was not inevitable. However, the existence of alternatives was not mentioned in the Investigation Report and it must be presumed that there was no evidence indicating that such an alternative had been considered by the crew.   

The Final Report of the Investigation was authorised for release on 7 April 2020 and it was officially released on 4 May 2021. No Safety Recommendations were made.

Related Articles

• Loss of Control
• Stabilised Approach
• Visual Approach
• Flying a Visual Approach
• Safety Oversight
• Safety Management System
• Continuation Bias