Aircraft Accidents and Lessons Unlearned XLIV: Korean Air Cargo 6316

On April 15, 1999, Korean Air Cargo flight 6316 (KAL6316) crashed shortly after take-off from Shanghai Hongqiao International Airport (SHA). The McDonnell-Douglas MD-11 aircraft was operated as a scheduled international cargo flight between Shanghai and Seoul, Korea. The aircraft was airworthy; the flight crew was qualified and trained. This should have been an uneventful routine flight.

The accident report, 99-091-0, was accomplished as an investigation per the International Civil Aviation Organization (ICAO) Annex 13 provisions by the Civil Aviation Administration [the People’s Republic] of China (CAAC). Participating with technical support were the Korean Civil Aviation Bureau (KCAB), the National Transportation Safety Board (NTSB), the Federal Aviation Administration (FAA), Boeing (who had purchased McDonnell-Douglas at the time), Pratt and Whitney Engines and Korean Air. The report shows that the NTSB was tasked with reading out the flight data recorder (FDR) and the cockpit voice recorder (CVR).

There were three people on the cargo aircraft during the accident flight. The Captain had a total of 4,856 flight hours in the MD-11. The First Officer (FO) had 1,152 flight hours in the right seat of the MD-11. Both pilots had recently undergone training and were qualified to fly the MD-11. The third person was a technician (mechanic), flying with the aircraft; he had twenty years with Korean Air working their trunk aircraft, including the MD-11. The technician was not heard on the CVR; mechanics are not known to fly in the cockpit, even on cargo flights, so any input he might have provided was absent.

The CAAC made efforts to remove all other possible contributors to the accident: weather, navigation aids, communication, air traffic control, recent maintenance, long-term maintenance, aircraft airworthiness, fuel distribution, fuel contamination, weight and balance. All these possible contributors were eliminated, one-by-one, as unlikely causes. The CAAC report was thorough in this process; as the Arthur Conan Doyle quote stated, “Once you eliminate the impossible, whatever remains, no matter how improbable, must be the truth.” What remained was most improbable: two qualified pilots, who got so behind events, that they unintentionally crashed the aircraft.

Per page 10 of the report, the aircraft impacted the ground at a twenty to forty-degree, nose down attitude, wings level. All primary instruments on the MD-11 are digital; the ‘gauges’ are video displays that are fed data from the Flight Director, Air Data Computer, etc. Any FDR data captured during the accident flight was lost in the post-crash fire. The only analog instrumentation – the sole ‘snapshot’ recovered – were the standby gauges; the standby altitude/airspeed indicator was locked at 398 knots at time of impact.

All flight control components had been meticulously examined for proper condition and to determine no sabotage had occurred. The engines had no issues during flight. The aircraft was properly balanced; it had achieved an altitude of 1370 meters (4500 feet) during the 2-1/2 minute flight; the crew did not talk about control issues caused by being tail or nose heavy; there was no evidence of a stall. No mechanical or technical issues were discovered. By all evidence, the aircraft was in an airworthy condition. On page 48, the report stated, “The above mentioned evidences indicate that the accident airplane’s sudden dive was at the wish of the crew rather than other causes.” A most improbable truth.

On page 9, the report stated that the only usable recording device was the CVR; “… no useful information could be retrieved from the recovered portions of tapes of the FDR …”. The Korean to English translations of the CVR recordings were choppy; conversation, terms and sentence structure did not convert smoothly from the Korean pilots’ native language to English. Even so, the conversations were mostly recognizable, yet the tell-tale urgency and other voice inflections were missing, leaving the reader to guess at what the pilots were reacting to, indeed talking about. This is unfortunate; reactions, such as excitement, doubt, rises in pitch or volume would have been more telling. Another unfortunate fact was that the mechanic was not in the cockpit; the pilots [appeared to be] worried about instrument readings.

The report was correct in highlighting that the Captain did not conduct a pre-takeoff brief; the pilots did not run through the departure, including expected turn points or emergency plans. As it happened, this flight had unforeseen events that contributed to an increasing state of confusion. The report also stated that the root cause of the accident was confusion about altitudes; the air traffic controller relayed altitude in meters while the FO stated to the Captain altitude in feet – 1500 meters versus 1500 feet. This was unquestionably another contributor to the flight crew’s inflight confusion. However, there was more going on during the flight than questions about altitude.

At 16:03:06 (4:03;06 PM), the crew initiates a left turn, but the FO was confused about when to complete the turn. At 16:03:37, the Captain stated, “It might turn upside down. What’s wrong with this?” At 16:03:54, the FO said, “Slat, why doesn’t it work? Slat, slat up.” Finally, at 16:04:05, the Captain said, “Well, what’s wrong with this airplane today?” All three statements occur within one minute; the first is stated at one minute into flight and the last stated thirty seconds before impact. Through this, there were no communications with air traffic to report problems.

The flight crew never elaborated on what they were seeing. For example, was the ‘slat problem’ that the slats would not retract or was it an indication problem? What ‘might turn upside down’? What was the Captain seeing that he asked, “…what’s wrong with this airplane today?”

Even before takeoff and the ensuing confusion, the FO got meters-to-feet wrong. During engine start, with the towbar still attached, the Tower says to go to nine hundred meters; the FO tells the Captain nine hundred feet. This was not due to heavy workload; the pilots were loading data into the computer and they were entering the wrong information. Just like the later confusion of altitude, the Captain did not challenge the FO or tell him to confirm the numbers. Cockpit resource management (CRM) appeared to be non-existent; no challenges, no questioning and a lot of guessing took place.

On November 22, 1968, Japan Air Lines flight 2 ditched in shallow water in San Francisco Bay, two and a half miles short of the runway in heavy fog. The DC-8 landing was determined to be the result of poor flight crew communication during an Instrument Landing. The FO and Second Officer did not question the Captain, a result of cultural issues within the cockpit; the Captain would not … could not … be questioned. There was no CRM; it did not exist. In 1999, CRM should have been standard practice.

The FO on KAL6316 did not insist on a pre-takeoff briefing even though it was required by Korean Airlines procedures. The FO confused meters with feet of altitude, possibly giving the Captain the wrong altitude to enter into the Flight Management System. The FO became confused again, this time with the air traffic clearance before entering the information into the Flight Control Panel. Just like with the pre-takeoff brief, there was no challenge, no verification. Instead the pilots just changed settings without confirmation.

The CAAC was correct in its Probable Cause, but they did not go far enough. It was culture that brought down this aircraft. Complacency did not bring about the failure to perform a pre-takeoff briefing; culture made the briefing unimportant. This was evident because neither pilot pushed for the briefing; neither one questioned the briefing’s absence; neither one seemed to care. Every subsequent problem hinged on the briefing and the briefing hinged on a culture that demanded it be done.

Recommendation 4.2 spoke to the briefing, requiring pilots conduct them. This recommendation was good, even though briefings had been required for decades. Briefings should not have had to be recommended because this aircraft should not have crashed. Recommendation 4.6 was also good, requiring air traffic to speak to one measurement, metric or standard – not both.

The CAAC did a great job with this report; they hit every mark on the investigation. All investigatory bodies still need to accept the existence of culture as a root cause and study culture’s effects on safety. Until they do, culture will continue to elude the most efficient reports, thus eluding being addressed and fixed.