On 5 September 2005, a Boeing 737-200 being operated by Mandala Airlines on a scheduled domestic passenger flight from Medan, Indonesia to Jakarta failed to become properly airborne during the attempted take off from from runway 23 in day Visual Meteorological Conditions (VMC) and, after failing to remain airborne, overran the end of the runway at speed finally coming to a stop outside the airport perimeter. The aircraft was destroyed by impact forces and a subsequent fire and 100 of the 117 occupants were killed and 15 seriously injured. The aircraft collided with residential property, vehicles and various other obstructions and as a result a further 49 people on the ground were killed and a further 26 seriously injured.
An Investigation was carried out by the Indonesian NTSC. Flight Data Recorder (FDR) and Cockpit Voice Recorder (CVR) data were available to aid reconstruction of the event, but the CVR readout was of poor quality, especially in respect of the cockpit area microphone. As a result, there was no useful evidence in respect of intra-crew communications or in respect of warning system activations.
It was established that the aircraft had been near to but within the MTOM for the prevailing conditions and that there were no C of G issues. FDR evidence was not entirely conclusive, but it appeared that the aircraft may have reached a maximum altitude of around 70 feet aal before settling back onto the ground and remaining there apart from a further brief period airborne as it crossed a river. It was calculated from available data and simulation that the stall warning system would have been activated on both occasions that the aircraft was airborne even though there was no directly recorded data to provide proof that this had actually happened. It was also considered that although, because of the destruction of the aircraft, it was not possible to prove that the TOWS had not been activated during the take off roll, it was difficult to imagine a scenario in which it would have been ignored if it had been sounding.
Circumstantial evidence allowed it to be safely concluded that the take off had been attempted with wing slats and trailing edge flaps retracted and that the aircraft had stalled after becoming airborne before settling back on the runway and continuing off the end for a distance of 540m before it stopped with the fuselage broken into three parts and a with a rapidly developing fire starting. It was noted that airport emergency services had taken some time to arrive at the accident site and that their deployment had been limited because of an initial desire to keep the airport operational,
It was noted that the RESA, with a length of approximately 60m, was considerably less than the International Civil Aviation Organisation (ICAO) minimum of 90m.
The Investigation found that there was confusion amongst the Operator’s pilots about the right way to use Normal Check Lists:
“From the interview with the chief pilot, chief instructor, instructors and some of line pilots indicated that there were various understanding on how the checklist should be performed. Some said it should be “read and do”, some said “do and read”, while others said the combination. While being asked whenever the checklist execution is being interrupted, some of the pilots seemed not firmed on how to handle the checklist interruption.”
It was also considered that both RFFS and the Airport Emergency Plan had not been fully supportive of the response to the emergency. One effect was that all initial assistance to survivors had been provided by local bystanders rather than by the emergency services.
The Probable Causes of the accident were determined as:
- The aircraft took-off with improper take off configuration namely with retracted flaps and slats causing the aircraft (to fail) to lift off.
- Improper checklist procedure execution had lead to failure to identify the flap in retract position.
- The aircraft’s take off warning horn was not heard on the CAM channel of the CVR. It is possible that the take-off configuration warning horn was not sounding.
A total of 8 Safety Recommendations were made by the NTSC as a result of the Investigation:
To Mandala Airlines
- That they ensure that flight crew correctly perform the checklist execution, in particular ensuring confirmation and verification.
- That they develop and promulgate an accurate checklist from the approved Operations Manual instead of the “do not use for flight” version.
- That they should review and update the Standard Operating Procedures in accordance with the approved Operations Manual.
- That they should conduct a functional test of the take-off warning horn on its Boeing 737 aircraft each day before commencing flight operations. Additionally the warning horn should be functionally checked once in every 200 flight hours in accordance with FAA AD 88-22-09.
To the Directorate General Civil Aviation
- That it should enforce and ensure that the installation and the maintenance of FDR and CVR in accordance with CASR 121.343, 121.359 and Annex 6 Attachment D.
- That it should assess the Mandala Airlines Operations Manual to ensure the adequacy of the Standard Operating Procedures.
- They should review, and improve if necessary, the existing emergency manuals for airlines and airports, in particular with respect to coordination with local authority resources involved during emergencies.
- That they should mandate the requirement for real-time exercise of Airport Emergency Plan to be conducted at least once every year.
The Final Report: KNKT/05.24/09.01.38 of the Investigation was published by the NTSC on 24 May 2009.