Accident to the Piper - PA32 - 300 registered F-OKGO on 16/02/2020 in Fort-de-France bay (Martinique)

Cat. 2 investigation report: simplified-format report, adapted to the circumstances of the occurrence and the investigation stakes.

The pilot, accompanied by two teams of tandem parachutists, took off at 08:45 from Fort-de-France aerodrome for a parachute drop flight. It was the second flight of the first day of the commercial operation of this aeroplane.
On flying through 2,600 m in climb to 3,000 m, the pilot observed a decrease in the engine speed and engine misfires despite his inputs on the throttle control. As the pilot estimated that 10 US gallons remained in the selected fuel tank based on the fuel gauge indication, he was concerned about the possibility of a technical failure as the aeroplane had just come out of a long maintenance period. On receiving the control’s clearance for the drop, he gave the evacuation order to the parachutists at 09:13 overhead the drop point and at the planned altitude.
He then put the plane into descent and selected the appropriate engine speed for this flight phase without changing the fuel tank. It appeared to him that the engine was operating normally again and he thus started returning to the aerodrome.
On final for runway 10 overhead Fort-de-France bay, the pilot was cleared to land at 09:18. On making inputs on the throttle control, he observed that these were to no effect and that the engine was behaving erratically.
At 09:19, he told the controller that nothing was responding to his inputs and then ditched in the bay in the minute that followed.
The pilot, having donned his life jacket, managed to evacuate the aeroplane before the latter sank at around 550 m from the coast. He was rescued by a boat 25 minutes later.
Around 1 hour 20 minutes after the first off-block departure, when only the fuel in the left tank had been used, the pilot was confronted with an engine malfunction. The pilot, who had a very small amount of experience with respect to the plane, was not aware that the fuel consumed corresponded to the quantity of usable fuel in the selected tank. In addition, the aeroplane had just come out of a long maintenance period and the pilot focused on the possibility of a technical failure whereas the engine was probably experiencing fuel starvation.
On starting the descent, the rotation of the propeller being assisted by the relative wind and the power demand on the engine being close to zero, the malfunction symptoms disappeared which gave the pilot the impression that the anomaly had disappeared. In all likelihood, this incited him to return to the aerodrome without taking any particular measures.
When he made an input on the throttle control again to increase power on final, the remaining fuel in the system was very probably completely exhausted which led to the total loss of power and meant that the pilot could not hold the approach slope. As he did not question the fact that there was still some fuel, it is probable that he did not envisage changing the tank selection as a possible strategy. In these circumstances, he was obliged to ditch.
In the absence of an examination of the engine, the investigation was not able to eliminate the hypothesis of an engine anomaly, but it is nevertheless very probable that the accident was due to fuel starvation.