Accident to the Boeing 737 registered F-GZHA operated by Transavia on 01/10/2022 at Nantes - Atlantique AD

On descent bound for Nantes, the co-pilot in LIFUS training reminded the instructor captain (TRI) of his difficulties with flaring and his apprehension about not reproducing the same type of rough landing as the one carried out the previous day on runway 21 at Nantes. During the briefing, the co-pilot discussed the offset approach to the axis of runway 21, the profile of this runway (“the hump”) and the changing weather conditions. These threats were not taken up by the instructor. The visual perception caused by the upward-sloping part of the runway and the stress linked to the difficulty of landing at Nantes were not the subject of any particular management strategy on the part of the instructor. The moment at which automated systems would be disconnected was not discussed either at that point.  

During the approach, at around 2,500 ft, the co-pilot announced to the instructor that he would disconnect the automated systems at an altitude of 2,000 ft, i.e. approximately two minutes before reaching minima. The co-pilot wanted to take advantage of the instructor’s presence to fly more in manual mode and regain experience. It is possible that the instructor let the co-pilot proceed as he wished to help him regain confidence. However, given the presence of crosswind, the ceiling close to the minima and the co-pilot’s lack of recent experience, flying in manual mode probably resulted in a high workload for the co-pilot.

At 1 NM from the MAPTS and at an altitude of approximately 800 ft, the co-pilot turned left to intercept the runway axis. The approach was stabilised, the co-pilot held the descent slope following the PAPI indications, and the speed remained close to the reference approach speed. The instructor’s correction call outs up to very low height showed that his attention was mainly focused on holding the runway axis.

After crossing the threshold of runway 21, the aeroplane first flew over the downward-sloping part of the runway. At a height of between 40 and 30 ft, the co-pilot started his nose-up input on the control column in order to flare, without reducing thrust, but this input remained insufficient to change the aeroplane’s pitch attitude. The aeroplane then began to fly over the upward-sloping part of the runway. Between the “Thirty” and “Ten” callouts, spaced by one second and representing around 80 m of flight, the co-pilot applied a marked and rapid nose-up input to the control column, pulling it to more than three quarters of its travel, before setting the thrust levers to IDLE. At the same time, the instructor probably became aware of the delay in starting the flare and instinctively warned the co-pilot to watch out, by calling out “Attention”. These marked and rapid inputs at low height above the upward-sloping part of the runway did not reduce the aeroplane’s energy before contact with the runway.

It is very likely that the late flare resulted from an erroneous perception of the final part of the approach slope due to the upward slope of the runway and to the fact that until reaching a low height, the two pilots were mainly focused on holding the runway axis. 
The effect the runway characteristics (downward and then upward-sloping) had on the synthetic-voice height callouts did not help the co-pilot to start the flare and reduce thrust early enough, given the upward slope before the hump. Moreover, the instructor did not consider taking the controls during the flare, and very probably did not have the time to do so.

The touchdown of the main landing gear on the upward-sloping part of the runway was hard, with a recorded normal load factor of 2.95 g for a sink rate of 12 ft/s. The spoilers deployed, then the
aeroplane bounced.

The strength of the impact on landing and the bounce surprised the two crew members. The instructor instinctively applied a marked nose-down input to the control column, up to the nose-down stop, which resulted in a rapid decrease in the aeroplane’s pitch attitude. This reduction in pitch attitude combined with the deployed spoilers led to a rapid decrease in the aeroplane’s lift. The nose and right main landing gear touched the runway simultaneously. The violence of the impact on the nose landing gear caused both tyres to be ejected from the landing gear, and the aeroplane continued the landing, running on its
wheel rims.

The instructor then held the aeroplane on the runway centreline during its deceleration before turning left onto a taxiway to stop the aeroplane and vacate the runway.

The fact that the TRI was familiar with the 21 approach to Nantes and the characteristics of the runway may have minimised: 
•his perception of how difficult it could be for the co-pilot to land on this runway in these conditions;
•his perception of the risks associated with the apprehension mentioned by the co-pilot about landings. It is also possible that the TRI wanted to let the co-pilot act without adding any extra stress, with a view to allowing the co-pilot to regain confidence after his period without flying. The landing performed by the co-pilot in a strong crosswind at Djerba may have boosted the TRI’s confidence and diminish his alertness during the approach and landing.