Japan Airlines Flight 706 violent pitching accident

The Japan Airlines Flight 706 violent pitching accident (Nihon Kōkū 706-bin rankōge jiko) was an aviation accident that occurred on 8 June 1997. The aircraft suddenly oscillated violently up and down, injuring people on board.

The Ministry of Transport designated this incident an aviation accident, and 1 cabin attendant who was injured later died from the injuries sustained in the accident. The captain of the accident aircraft was indicted for professional negligence resulting in death and injury, but his acquittal became final in 2007.

Overview of the accident

On 8 June 1997, Japan Airlines Flight 706, an MD-11 operating from Kai Tak in Hong Kong to Nagoya (registration JA8580, nicknamed “Etopirika”, Japan Airlines' MD-11 No. 1; 11 crew, 169 passengers, 180 in total), was descending for landing on autopilot at about 17,000 feet over the Shima Peninsula, Mie Prefecture, at around 7:34 pm, when its airspeed increased beyond the set value. To restrain this speed, the flight crew used the autopilot's vertical speed setting wheel to reduce the rate of descent (which raises the nose and reduces speed). However, because the expected deceleration was not obtained, they extended the speed brakes. At around 7:48 pm, immediately after the speed brakes were extended, a sudden sharp nose-up occurred, after which the autopilot disengaged and the aircraft pitched violently up and down.

Because of this oscillation, crew and passengers who were not wearing seat belts were thrown against the ceiling and elsewhere and injured; 1 cabin attendant and 3 passengers were seriously injured, and 4 crew and 4 passengers suffered minor injuries (as recorded in the accident investigation report; in the later trial and press reports the number of injured is given as 14). Flight 706 landed at Nagoya Airport at 8:15 pm, 15 minutes later than scheduled. Of the seriously injured, the female cabin attendant was in critical condition with spinal cord injury, cerebral contusion and other injuries, and without recovering from her critical condition she died on 16 February 1999, 1 year 8 months later, at the age of 34 (this is not reflected in the report, because under the Convention on International Civil Aviation only deaths within 30 days are counted).

Cause of the accident

According to the report of the Ministry of Transport's Aircraft Accident Investigation Commission (at the time), published on 17 December 1999, it was presumed that the captain's series of control inputs induced the violent pitching accident.

Because the series of control inputs made by the captain during the descent exceeded the capacity of the accident aircraft's autopilot, the autopilot disengaged. Until then the autopilot had been acting to restrain the captain's nose-up inputs, but its disengagement caused a sharp pitch-up, and the interaction between the subsequent nose-up and nose-down control inputs intended to stabilise the pitch and the MD-11's longitudinal stability characteristics sustained the pitch fluctuations, resulting in the violent oscillation. As for why the captain flew in this way, it was stated that although the captain had transitioned from the Boeing 747 Classic to the MD-11, he had not sufficiently understood the flight characteristics of the MD-11, which was more high-tech than conventional types. The report also noted that of the injured, only 1 person, one of the slightly injured, had been wearing a seat belt at the time of the accident, and all the other injured had not, and it pointed to the failure to wear seat belts as the reason people were injured.

The autopilot of the accident aircraft, an MD-11, was slow-acting for passenger comfort. In addition, the aircraft used a system that improved fuel economy by transferring part of the fuel in the wings to the horizontal stabiliser during cruise to move the centre of gravity aft, and when it began its descent, returned the fuel to the wings to restore the centre of gravity and increase control stability. It was therefore pointed out that if the autopilot unexpectedly disengaged during cruise, control stability would be low and a dangerous situation could easily arise. Moreover, by 1996 accidents similar to this one had occurred 5 or more times, and although the manufacturer McDonnell Douglas (now Boeing) had issued improvement information on operation and maintenance (SB), Japan Airlines had not taken full-scale countermeasures.

On the other hand, the rebuttal by the captains' union argues that the autopilot was not disengaged by the captain's inputs but by some other factor. According to it, as regards forced disengagement of the autopilot, on the MD-11 disengaging the autopilot by override requires applying a load of about 50 pounds (22 kg) or more to the control column for at least 1 second, and looking at the data recorded in the CWS (data on the load applied to the control column), the timing of the load being applied and the disengagement clearly differed. Also, to restrain the increase in speed that occurred before the disengagement, the captain is thought to have been holding the control column with one hand and the speed brake with the other, and it would be difficult to apply a load of 22 kg or more for 1 second or more with one hand, so it can hardly be said that the autopilot was disengaged by the captain's inputs; moreover, even if the captain had forcibly pulled the control column, the MD-11 is equipped with a vertical stabilisation device called a pitch rate damper, so a sharp nose-up would not result.

It states that the ACO (Automatic Cut Off), which operates when the autopilot is forcibly disengaged, did not operate at that time. Furthermore, analysis of the instrument records of the up-and-down motion after disengagement shows that the aircraft moved up and down regularly with a period of about 3 seconds; in the cockpit it was like riding an elevator travelling a height of 7 metres in 1.5 seconds, and a state in which 2.5 G was applied for 1.5 seconds, followed abruptly by a force of -0.5 G for 1.5 seconds, and then 2.5 G again, reportedly continued 5 or more times. It would be utterly impossible even for a veteran pilot to fly through such periodic up-and-down motion amid such sharp G fluctuations, and the captain himself said, “I felt as if I were being thrown about like at a rodeo, and I could not recognise what was happening.”

It argues that at the time turbulence may have been occurring over the Kii Peninsula at the boundary between cold and warm air, and that when the descending aircraft reached it, sudden changes in temperature, air pressure, speed, wind direction, wind speed and attitude occurred, so that information on a wide variety of changes entered in a short time the FCC (Flight Controll Computer), which controls the autopilot by issuing commands to each tail surface based on that information; it is highly likely that this placed an excessive load on the FCC, the autopilot was temporarily disengaged, stable flight could no longer be maintained, and a sharp nose-up occurred. In fact, the accident aircraft reportedly later experienced the phenomenon of violent up-and-down oscillation under similar circumstances 2 more times. Since the Aircraft Accident Investigation Commission reported that in actual experiments using a flight simulator, disengagement of the autopilot by the captain's override and a sharp nose-up such as in this case did not occur, this hypothesis is considered more credible.

It was also pointed out that the MD-11, the accident aircraft type, had the highest accident rate among the so-called fourth-generation jet airliners (advanced-technology aircraft). The Boeing 777 and Airbus A340, which entered service in the same period (with more than 400 built), have each had only 1-2 hull-loss accidents in service (the former the 2008 British Airways Flight 38 accident and the 2013 Asiana Airlines Flight 214 landing accident, the latter the 2005 Air France Flight 358 overrun accident), whereas the MD-11, with a total production of 200 aircraft, had 8 hull-loss accidents by 2010, so it can be said to have had relatively low reliability.

Subsequent developments

Although Japan Airlines' MD-11s were introduced as successors to the DC-10, all of them were retired in 2004, the year before the DC-10 was retired, and sold to the US air cargo company (UPS). As a result, Japan Airlines reportedly incurred an economic loss because it disposed of them before they were fully depreciated. The accident aircraft, “Etopirika”, was retired in September 2002, converted into a freighter, and as N272UP is still in active service as of 28 October 2021.

Subsequently, many of the airlines that operated the MD-11, including Japan Airlines, retired their MD-11s early, replacing them with twin-engine aircraft such as the Boeing 777 and Airbus A330, or converting them into freighters that carry no passengers, so that although it is relatively new equipment introduced in the 1990s, there are now no aircraft operating as passenger aircraft.

After that, the route itself was abolished, and as of 2023 JAL's Nagoya–Hong Kong service is limited to codeshare flights with Cathay Pacific.

After this accident, there were no fatal accidents during scheduled flights of Japanese-registered aircraft, nor fatal accidents involving large Japanese-registered aircraft, whether passenger or cargo, until the Haneda Airport ground collision in January 2024.