On 13 February 2019, a Boeing 737-800 (F-GZHM) being operated by Transavia on a scheduled international passenger flight from Lyon Saint-Exupéry to Tel Aviv was in the cruise at FL 370 in day VMC when it encountered severe turbulence which the flight crew were not expecting that included a rapid roll to the left which led the Captain to take control and recover manually to wings level. Two unsecured cabin crew and eight unsecured passengers sustained minor injuries which it was possible to respond sufficiently to for a diversion on that account to be unnecessary.
An Investigation was carried out by the French Civil Aviation Accident Investigation Agency, the BEA, based on downloaded FDR data and crew statements, relevant data on the CVR having been overwritten. It was noted that the Training Captain in command had accumulated around 15,000 hours flying experience of which 11,200 hours had been as Captain and 6,800 hours had been on type. The First Officer had over 3000 hours flying experience of which 900 hours were on type.
It was established that during their pre flight briefing carried out in the lobby of the hotel where they had been accommodated overnight, the Captain, having downloaded the flight file onto his EFB had noted that a risk of slight to moderate turbulence was forecast to occur about 90 minutes after takeoff.
The flight subsequently departed and proceeded normally to establishing in the cruise at FL 370. Just as the senior cabin crew member (SCCM) arrived in the flight deck 1½ hour after departure to advise the Captain that cabin service had been completed, ATC advised the flight that the flight was “about to enter an area where severe turbulence had been reported between FL380 and FL400”. The Captain told the SCCM that “there was going to be turbulence within two to three minutes” and she returned to the cabin and went to the rear to where her two colleagues were. By the time she reached there, the ‘Fasten Seatbelt’ signs had been switched on and the No 2 cabin crew then made a PA to announce that turbulence was expected soon and the No. 3 began to pass through the cabin to check that passengers had fastened their seatbelts. The SCCM then returned to the front of the cabin, on the way passing the No 4 cabin crew who had begun a seat belt check from the front.
In anticipation of the turbulence, two minutes after the ATC turbulence alert had been received, the First Officer had reduced the speed from M 0.79 to M 0.77. The first turbulence was felt approximately one minute after the illumination of the ‘Fasten Seatbelt’ signs and just over a minute later, ten seconds of moderate turbulence occurred “as the longitudinal wind component changed from a tailwind of 35 knots to a headwind of 11 knots and the left crosswind component with respect to the aircraft heading went from 77 knots to 50 knots”. The recorded vertical acceleration as detected close to the aircraft C of G varied between +0.52 g and +1.70 g. The indicated airspeed rapidly increased and as the overspeed warning was activated, the First Officer extended the air brakes whilst leaving the A/THR engaged.
Following this ten second interval, the turbulence then increased in severity for six seconds, the recorded vertical acceleration went from -0.70 g to +1.71 g and then to -0.32 g as the aircraft rolled 37° to the left at up to 38° per second. The Captain took control, levelled the wings and engaged the AP in control wheel steering (CWS) mode. He began a descent and asked the First Officer to inform ATC of this, although this was initially refused until the reason for the descent was give after which descent to FL 350 was approved.
When the severe turbulence began, the SCCM managed to sit down and secure herself on a crew seat and the No 4 cabin crew was able to sit down in an unoccupied passenger seat and secure themselves but neither the No 3 cabin crew who was in the aisle at row 27 or the No 2 cabin crew who was in the rear galley had time to secure themselves and were “thrown against the ceiling twice”. Some passengers at the rear of the cabin who had not fastened their seatbelts were also thrown upwards and struck the overhead lockers and it was reported that “other passengers who had fastened their seatbelts (still) sustained minor injuries”.
Once the turbulence was over, the situation in the cabin was assessed. The two cabin crew who had not been secured were reported to have taken “a few minutes to recover their senses” whilst the SCCM and the No 4 cabin crew went through the cabin to assist passengers who were injured or in shock. It was observed that “there was no division of tasks between the cabin crew and each one went through the cabin without coordinating with the others”. On completion of this cabin check, the SCCM reported to the Captain and asked for a call to be made for any medically qualified passengers to make themselves known. This was responded to by both a doctor and a medical student who, in coordination with the cabin crew then provided first aid to the injured passengers. The doctor then informed the Captain that the health of the passengers did not require a diversion and since there was no further risk of turbulence, the flight was continued to its intended destination with a request made for medical assistance on arrival.
The Context of the Turbulence Encounter
The general weather situation in which the flight took place was an anticyclone centred over Europe which was “creating a north-easterly flow marked by a jet stream running perpendicular to the terrain of the Dinaric Alps towards Italy”. This jet stream was situated behind the CAT and mountain waves breaking at high altitude over the Adriatic, a combination which was considered fairly common in this area and more generally downwind of any mountain range where a jet stream is perpendicular to such terrain. A SIGMET issued shortly before the flight took off warned of severe turbulence between FL190 and FL400 in the southern area of the Zagreb FIR where the encounter took place.
In addition to online access to the specific flight file, all Transavia flight crews also had access to weather information via a proprietary app on their EFBs called ‘eWas PILOT’. When the latest SIGWX forecast chart provided by this app had been consulted, it had been indicating “the on-track presence over the Adriatic of a small area of moderate turbulence between FL350 and FL374 and a larger one of moderate to severe turbulence between FL374 and FL399” but both pilots reported that when viewing this chart, they had only “seen” moderate turbulence above as well as at their flight level and attributed this to being unfamiliar with the colour code used by the app for this chart (light green for moderate turbulence, dark green for severe turbulence).They commented that training for use of the app was by an e-learning course on the use of the app rather than directed classroom training. It was also noted that although the First Officer had downloaded the flight file again about half an hour before departure to check that there were no changes, this was mainly to look at the latest TAF and METAR and he had not checked the latest (and now current) SIGWX chart.
Given the information the crew had accessed before departure, turbulence at the position it occurred was not entirely unexpected but the flight crew had not been expecting severe turbulence. However, the Captain’s pre flight briefing was given on the basis of a flight file downloaded 4½ hours before the flight departed with the First Officer having downloaded the same file at a similar time. This file did contain a SIGWX forecast chart, but not the one issued 4 hours prior to departure which was valid from flight departure time. The earlier chart used for the briefing did not forecast severe turbulence on the planned route whereas the later one did do so, with severe local turbulence forecast between FL300 and FL420. It was noted that whilst Transavia procedures ensure that a flight file including its forecast weather data are continuously updated, the applicable documentation does not specify at what point prior to departure pilots should download the flight file and brief using it.
The flight crew response to the turbulence once it had begun was considered against the content of the Transavia OM and the Boeing 737 Flight Crew Operating Manual (FCOM). The OM was found to state that “the autopilot must be the primary means of control when entering turbulence” and the FCOM content for severe turbulence included instructions for the crew to disengage the A/THR and place the AP into CWS mode. What actually happened was that the AP and A/THR initially acted on the controls to reduce the thrust to counter the effects of the turbulence and the CWS mode was automatically engaged by the Captain’s control input.
Relevant aspects of the training of Transavia cabin crew were examined and it was found that they covered turbulence procedures including “practical exercises in an aircraft mockup” every three years, although none of these concerned the actions to be performed in the cabin after severe turbulence. It was noted that all four of the cabin crew on the flight had been working for Transavia long enough to have completed this training several times.
The Conclusions of the Investigation included the following:
- The event demonstrated the intrinsic limitation of the procedures concerning unforeseen turbulence. The unpredictability both in terms of the time of the occurrence and strength of the turbulence means that these procedures cannot be applied in a timely or correct manner.
- Detection of CAT is not possible with the technology currently installed on aircraft. The development of onboard turbulence detection systems and the transmission of updated weather information to aircraft would both improve present turbulence risk management. The BEA report on a Serious Incident to an Airbus A330 in 2012 illustrates the fact that this need also exists for the management of some convective turbulence.
- The only current way of limiting the turbulence injury risk at this time is for passengers to comply with the instructions to keep their seatbelt fastened when they are seated.
The Final Report was made available in both English translation and in the definitive French language version on 21 February 2020 after the French language version was “published” the previous month. No Safety Recommendations were made.