Investigations into an unusual departure by a SriLankan Airlines Airbus A340-300 at London Heathrow have again underlined concerns over take-off performance calculations.
The departure generated consternation after the aircraft was photographed climbing out of the airport at a relatively low height on 5 February.
Its crew had expected a full-length take-off, but instead accepted an invitation to depart from the SB7 intersection about 3,000ft along runway 09R.
The pilots did not, however, have a regulated take-off weight chart for the intersection departure, so the captain used a different airport chart with a similar runway length.
"Data derived from achart for a different runway may not be correct because obstacles affecting the runway in use are not considered," says the UK Air Accidents Investigation Branch.
Although the crew discussed a full-thrust take-off they opted instead for a flexible take-off, a procedure that enables using only the necessary thrust to become airborne. Calculation involves assuming a higher air temperature - the flexible temperature - as a limiting element.
The crew calculated the take-off speeds and flexible temperature and entered the data into the A340's multifunction control and display unit. But this data was not separately recorded, and the inquiry was unable to verify the performance calculation.
SriLankan's pilots have an alternative if specific take-off charts are not available. They can use quick-reference tables to determine take-off performance. But the AAIB points out that these would have required the pilots to conduct a full-thrust take-off, because of the obstacles on the runway 09R departure path.
Although the flexible temperature used was not recorded - the captain recalled "low to mid-thirties" - the inquiry found the maximum to be 38°C. It determined that the take-off run required for the jet was 2,268m (7,740ft). The distance available was 2,854m and the A340 became airborne after some 2,650m.
A relief captain, present in the cockpit during take-off, had noticed the acceleration was slow and suggested applying full thrust, but neither of the operating pilots claimed to have heard.
While the investigators state that it is "probable" that the flexible thrust used was sufficient, the method used by the crew to obtain the performance data was "not in accordance with the airline training".
Flight-data recorder information was overwritten before it could be analysed. But the AAIB points out that the take-off data would not have been recorded anyway, and is recommending that such information be retained.
"Completeness of the investigation was restricted by the lack of this essential data," it says. "This problem could apply to any future investigation where the calculated take-off data may be of interest."
SriLankan Airlines, it adds, has told its pilots that take-off must not commence without relevant data.