French investigators believe the pilots of an Air France Airbus A318 unnecessarily took the opportunity to shorten an approach to Paris Orly, placing them under time pressure and resulting in an unstable descent that triggered ground-proximity alarms.

The inquiry has highlighted the possible effect of reduced flying activity by the crew, notably skill erosion, over the six months prior to the incident on 12 September 2020.

French investigation authority BEA says the crew of the aircraft, heading north-east as it arrived from Biarritz, accepted a shortened trajectory to runway 25 involving a left turn to head west. After completing the turn, the crew was cleared for the ILS approach at 5,900ft.

The shortened approach left the A318 above the correct profile and the crew carried out a series of actions to bring the altitude into line with the glideslope.

It was established on the localiser and glideslope, about 6nm from the runway and 2,000ft above airport level, but still in clean configuration – against the requirements of standard operating procedures.

BEA says the autopilots were not active. The captain made a nose-up input to the sidestick and the jet began to deviate above the glideslope, before the landing-gear was deployed and the flaps started to extend.

F-GUGM diagram-c-FlightGlobal via BEA data

Source: FlightGlobal modified from original BEA diagram

Quick-access recorder data showing a section of the A318’s unstable approach

With the runway in sight, the captain then pushed the nose down and the aircraft’s descent rate increased to 2,750ft/min. The A318 dipped below the glideslope and, although the descent rate reduced to 1,328ft/min, the aircraft was still too low.

Shortly afterwards – while 3.6nm from the runway at 1,250ft altitude, or 960ft above ground – the first officer modified the flight-management system to change the landing configuration from ‘full flap’ to the lower ‘flap 3’, raising the aircraft’s approach reference speed to 126kt.

At the same time the ground-proximity warning system sounded a ‘glideslope’ alert in the cockpit and a minimum safe altitude warning was triggered in the control tower, prompting an air traffic controller to instruct the crew to check their altitude.

The captain made an aft input to the sidestick, increasing the aircraft’s pitch from 2.7° nose-down to 7° nose-up, and returning the aircraft to the glideslope.

BEA points out that the aircraft was still travelling at high speed and, although it was established on the localiser and glideslope in the landing configuration by 500ft above ground, it was still flying at 152kt – some 26kt above the correct approach speed. The approach reference speed was not achieved until the aircraft was 180m past the threshold and 10ft above the runway.

The inquiry says the captain, having accepted the shortened approach, embarked on a strategy to increase the descent rate to reach the glideslope and then reduce speed to configure the aircraft.

But the stabilisation effort was “compromised” by the time the aircraft passed 2,000ft, it states: “The crew, whose workload was already heavy, did not have the resources to identify that the approach could not be stabilised at 500ft.”

The first officer’s late modification of the flight-management system probably reduced the effectiveness of monitoring parameters for landing, it adds, while the captain also “took his eyes off the instruments” having acquired visual contact with the runway.

His nose-down input resulted in a “significant” descent rate which led to the alarms in the cockpit and the tower. But BEA says the continuation of the approach was “not called into question”, with the crew’s deciding that the aircraft was “in the process” of stabilising by 500ft.

“Without realising it, the crew…probably had very few mental resources available to deal with an unforeseen event,” it adds.

Accepting the shortened approach without a specific briefing generated time pressures for “no operational reason”, says the inquiry, and fell outside of standard operating procedures. The situation was exacerbated by the absence of call-outs on trajectory and speed deviations from the first officer.

BEA notes that the incident took place following six months of reduced flying as a result of the pandemic.

Although both pilots had substantial experience, the captain had flown only 64h in the period from early March to late August 2020, while the first officer similarly flew about 70h – effectively only a quarter of their normal level of activity.

It says neither pilot was scheduled to fly for nearly a month before the second week of September, just a couple of days prior to the event.

BEA says that the captain acknowledged that he wanted to “push himself a little to his limits” to assist with retraining, and – while he was aware of the unstable approach, and considered a go-around – felt there was no risk.