Air Caraïbes Flight 1501 (TX1501/FWI1501) was a scheduled international passenger flight, flying from Princess Juliana International Airport in the Dutch overseas territory of Sint Maarten to Saint Barthélemy Airport which was in the French overseas region of Guadeloupe at that time. The flight was operated by Air Caraïbes, a Caribbean regional airline, using a de Havilland Canada DHC-6-300 Twin Otter. On 24 March 2001, during an approach to Saint Barthélemy Airport, the DHC-6 Twin Otter banked steeply to the left and crashed into a house, killing all 19 passengers and crew on board. One person on the ground was also killed in the explosions that followed.
The crash was the third deadliest plane crash in Guadeloupe, behind Air France Flight 212 and Air France Flight 117, and was the 11th deadliest involving a Twin Otter.
Investigation by the French Bureau of Enquiry and Analysis for Civil Aviation Safety (or BEA) concluded that the cause of the crash was due to an error by the flight crew in managing the plane's thrust lever. During its final approach to Saint Barthélemy, the crew caused one of the aircraft's engines to reverse by moving the thrust lever back to the "beta" range, creating a thrust asymmetry which caused the aircraft to bank steeply to the left and crash.
According to the BEA, this was the first fatal accident in the Twin Otter whereby an accidental thrust selection caused a propeller to enter the Beta range. The investigation was hampered by the lack of a flight recorder on the plane. After the crash, the BEA made a mandatory recommendation to equip every transport plane with at least one flight recorder.
Flight Take-off and approach Flight 1501 took off from Princess Juliana International Airport at 4:00 pm local time, using a visual flight rules (VFR) flight plan, and was expected to land in Saint Barthélemy at 4:15 pm. The distance between Saint Martin and Saint Barthélemy is approximately . The weather was reported as good, with calm winds and very good visibility.
The flight entered its cruise altitude of and the crew reported passing the "Pain de Sucre" waypoint for a final approach to runway 10 of Gustaf III Airport in the last transmission from Flight 1501.
Crash While flying above the western slope of La Tourmente Pass, the aircraft suddenly banked steeply to the left. Numerous eyewitnesses including an Aerodrome Flight Information Service (AFIS) worker claimed that Flight 1501 then rapidly lost altitude and crashed onto a house, with a massive explosion following immediately after impact. Emergency services quickly arrived to the scene and cordoned off the area. All 19 people on board were killed instantly in the crash and an elderly man, identified as Augustin Questel, who lived in the house with his wife, died from the subsequent fire and explosion. His wife was alive but slightly injured and was taken to hospital.
Aircraft The aircraft involved was a de Havilland Canada DHC-6-300 Twin Otter, registered in France as F-OGES. The aircraft was equipped with two Pratt & Whitney Canada engines driving Hartzell propellers. It entered into service in 1970. It was subsequently sold to Air Caraïbes, where it entered service in 2001. The aircraft had accumulated more than 35,000 flying hours in 90,000 flight cycles with Air Caraïbes.
Passengers and crew According to Air Caraïbes CEO Philippe Chevallier and the company's local manager, most of the passengers were French tourists, 15 French residents of Paris, two local residents of Saint Barthélemy and one American. One of the crew members was a French national while other's nationality was not revealed. Multiple reports stated that one Dutch woman and two Belgians were also on board. The American was also reported to have dual nationality.
The captain (and pilot flying), identified by an Air Caraïbes' official as 38-year-old Jean-Paul Jerpan, obtained his commercial pilot license on 18 June 1987 and Instrument Rating on 3 November 1987. He obtained DHC-6 type rating on 28 November 1988 and was subsequently hired as a pilot by Air Guadeloupe and later transferred to Air Caraïbes. The Captain had acquired a total flying hours of 9,864 hours, of which 5,000 hours were on Twin Otters.
The co-pilot, 38-year-old First Officer Nicolas Manen, obtained his Twin Otter type rating on 21 December 2000, having had previous professional experience as a flight engineer in the French Air Force. According to the investigators, he was employed on a short-term contract, in which he was supposed to join another airline as a flight engineer. He had acquired a total flying hours of 670 hours of which 15 of them were on the DHC-6.
Investigation As the crash occurred within the French territory, the BEA was responsible for the investigation. The BEA sent two field investigators to the crash site and later sent four investigators to Pointe-à-Pitre on 25 March. As the aircraft was manufactured in Canada, the Canadian TSB joined the investigation with the assistance from Bombardier and Pratt and Whitney Canada. The American NTSB also joined the investigation and sent a representative.
According to Chevallier the aircraft was not equipped with a black box, and thus investigators had to rely on observations of air traffic controller recordings, eyewitnesses, wreckage observation and their best guess on what might have happened on Flight 1501. Soon after the crash, airport officials claimed it was caused by mechanical failure. One witness stated that he heard an engine "sputtering" before the aircraft slammed onto the house.
The western approach to Saint Barthélemy Airport crosses the Col de la Tourmente just before the runway, requiring pilots to have special certification to land there. Aircraft approaching the runway must make a steep descent and fly low over houses before landing, with no night flights and all aircraft to be equipped with radio.
During the first phase of the investigation, investigators inspected the wreckage of the plane. Initial examination revealed that there were no signs of abnormalities with the plane's controls. Inspection of the plane's propellers also revealed that both propellers were delivering significant power during the crash. Due to the severity of the impact, investigators could not take a fuel sample from the wreckage. During their inspection they found two video cameras at the crash site.
Closer observation of the plane's propellers revealed that the left propeller had a pitch setting of at least 20.4°, that is to say a pitch above flight idle. The right propeller had a different pitch setting for each of the blades, settings likely due to less-violent secondary impacts to the powerplant at the moment of impact. Consequently, the values noted were unusable.
Use of video footage in the investigation The recovery of video from the crash site had the potential to significantly assist the investigation. One of the two video cameras was too badly damaged by the impact to provide a recoverable recording. However, the second was less damaged and its footage was recovered except for some portions including that adjacent to the recording head.
This video had been taken by a passenger on Flight 1501 filming through a left-side window in the few minutes before the accident. Investigators reviewed the three main sequences of the flight: initial climb, cruise and approach.
Later, investigators analysed the video more thoroughly to examine the operation of both propeller visually and through the sound recorded on the tape. During Flight 1501's initial climb and cruise there were no abnormalities in the sound of the propeller. However, investigators detected an abnormality during the approach phase, with audible fluctuations indicating that both propellers were "not synchronized".
Flights observation and footage comparison Investigators then recorded new video footage, using the same model of video camera as the one recovered, during a recreation of the flight to evaluate the track, engine parameters and the pilot's hand positions. The thrust lever of the DHC-6 was located on the cockpit ceiling and investigators wanted to determine if the pilot's hand position may have contributed to the crash.
It was revealed that, when the pilot flying was seated on the left, his right forearm is located in the extended vertical axis of the levers, a position which is favorable to symmetric handling. However, when the pilot flying is seated in the right seat, his left forearm is not vertical; he stretches slightly to the left and is thus not aligned with the axis of the levers, this "angle" being compensated by wrist movement. In addition, if the pilot flying was seated on the right, he would move the thrust lever with a different timing, used short successive and separate movements on the engine No. 1 lever and then on the engine No. 2 lever, with the hand always remaining in contact with both levers.
A comparison was made by investigators to confirm the position of Flight 1501. An analysis of the engine parameters also showed that both engines were functioning normally from take-off until the last image of the video, recorded just minutes before the accident. Towards the end of the film, the spectral analysis showed an increase in the propeller speed. This variation seems to correspond to the beginning of the descent towards the airport, at the moment when the crew selected the propellers' full low pitch position.
Thrust lever beta In the DHC-6, there is a mode called as "beta" in which the crew can control the propeller pitch directly. There are two beta ranges. The first is from + 17° to + 11°, called "approach beta" and permitted in flight. In this range the propellers are tractive. The second range runs from + 11° to − 15°, called "reverse beta" and prohibited in flight. In this range, the propellers act as brakes, two different sub-ranges being identifiable: ground idle and traction reversal. If the reverse beta is engaged in mid-air, according to the manufacturer, this would greatly increase aerodynamic drag and cause rapid deceleration. As a consequence, the plane would lose lift.
The manufacturer had known about this risk since 1979 and had started to implement several preventive measures. They had warned pilots about the risk in all flight manuals. They had installed a mechanical stop to try to prevent a pilot accidentally causing a plane to enter the beta range. An alarm had also been installed by the manufacturer.
Investigators' hypotheses Investigators made several hypotheses about Flight 1501. Mechanical failure, terrorism, bird strike and weather-related factors were quickly ruled out. They then focused their attention on two strong hypotheses.