Narrative:

Approximately 3hr 10min into the flight at cruise in night IMC; FL330; coasting out of japan within tokyo fir and radar contact; we encountered sudden and severe turbulence resulting in large vertical and lateral deviations. The initial event lasted several minutes; followed by two more less severe events; in total lasting perhaps 15 - 20 min. There were no sigmets that we were aware of; weather and windshear (w/s) data on the ofp for that segment (for FL310; we were at FL330) showed no significant w/s (4 or less). FL330 appeared to be 'in the tops' of the existing cloud layer; which showed no significant radar returns or convective activity. The weather radar system(s) appeared to be operating normally; showing appropriate ground returns etc. When cycled through range; pitch and gain settings. The first encounter was very sudden and severe; with little or no warning beyond what had been only light to moderate turbulence at the time. My first thought was that we had clipped the top of a thunderstorm cell; but there was no indication on the radar; and no lightning; hail or other weather phenomena that would indicate the presence of a thunderstorm. Initial pitch up was not extreme; but accompanied by large airspeed and altitude deviations; resulting in intermittent overspeed and stick shaker warnings. This was followed by a roll to the right of approximately 25 degrees and resulted in a 50 - 60 degree heading change and loss of LNAV capture. Observed altitude excursions were approximately +700 feet; to -500 feet from FL330. Airspeed fluctuations were too large and erratic to capture. First officer; as PF; initiated recovery by disconnecting a/P and a/T and attempting to limit pitch and roll excursions and stabilize the flight path and airspeed. ATC was notified and did not report any conflict or traffic hazard in our immediate vicinity. We heard no other reports of severe turbulence on frequency; or on 121.5 mhz. There were no TCAS targets displayed nearby. Once brought back to a more stabilized path (perhaps 1 - 2min) the automation was reengaged and we again tried to determine if there were thunderstorms in the area; and again found none. A second encounter soon followed; with less severe flight path excursions; recovered in the same manner. Both events resulted in altitude and course deviations outside of regulatory limits; and were beyond our control.some factors to consider:1. We were filed for FL310 initially as our T/O was at mgtow. FL310 was not available from departure; we were offered FL270 or FL330. FL330 was chosen; which was the maximum for our gw. Once level we elected to use lrc over econ initially in order to 'split the difference' between the amber bands on the airspeed indicator until our gw reduced. 3hrs later we were still in lrc which resulted in a .859m cruising speed when the turbulence was encountered. Mach .82 was entered into VNAV as soon as the automation was re-engaged; followed eventually by econ for the remainder of the flight.2. Upset training works; first officer's quick response and deliberate but smooth recovery; using techniques trained for annually; was directly responsible for the safe outcome; and undamaged aircraft; from this thankfully rare event.3. CRM/captain authority issues. CRM/tem training works as well. While many would say hand flying in severe turbulence should be a 'captain only' event; I disagree; with certain caveats. As stated; first officer was PF; yet with the first sudden encounter I also instinctively reached for the yoke and thrust levers; resulting in both pilots on the controls precisely when the need for smooth inputs is absolutely critical. Keeping that in mind; but remaining prepared; I withdrew my hands from the controls and continued as pm while first officer performed the recovery. Clearly there are arguments for either case; but they are also not mutually exclusive; as I remained prepared to switch roles if I thought it was required; which I am glad to report was not the case. Bothpilots were intently focused on all aspects of the flight path; but working as a team and not in a tug of war on the controls. We don't train that way in the simulator; that is: switching roles in the middle of an upset; so we shouldn't be adding a step to the maneuver in the chaos of the real event; it is a bad time to practice a hand off; provided the PF is winning the battle.4. Fatigue/frm. As can often happen; my rest plans did not result in the amount of sleep intended; though I was in bed in a quiet; dark room during the pre-duty rest period. The departure was late afternoon local; with full daylight for at least the first part of the flight; and we were a 3 man crew so I knew there would be a rest period. Fitness for flight is by nature a subjective call; and while I did feel somewhat tired at pick up time; I judged myself fit for duty. My practice as captain; once the PF/pm/rp duties are established; is to give first choice of rest to the PF; who in this case chose the middle rest period. As I felt sufficiently rested to stay in the seat for the first part of the flight I elected to take the last rest period. After sunset now; at 3hrs in cruise when the event began; I was experiencing some fatigue and probably was not 100% precisely when; at an instant's notice; 150% was required. I would describe this as one factor in the overall equation of this event; with perhaps the worst place for me to have been at that time would be in the bunk.recommended practices for future events. Preparedness is the key; and planning as much as practical in terms of weather forecasts and flight planning. But; as in our case; where it was almost completely unexpected; proficient hand flying skills are the most important tool to have immediately available. My understanding is that there will be increased emphasis on real time; less scripted; training scenarios. That will likely add realism to promote greater preparedness for such extreme and unexpected events like the one we encountered.

Related reports:

Google
 

Original NASA ASRS Text

Title: B747 flight crew experiences severe turbulence at FL330 in the Tokyo FIR. Altitude deviations of 600-700 feet occur along with stick shaker and over speed warnings. The First Officer disconnects the autopilot and auto thrust to return the aircraft to stable flight. A second and third event of less severity occur over the next ten minutes. Flight continues to destination.

Narrative: Approximately 3hr 10min into the flight at cruise in night IMC; FL330; coasting out of Japan within Tokyo FIR and radar contact; we encountered sudden and severe turbulence resulting in large vertical and lateral deviations. The initial event lasted several minutes; followed by two more less severe events; in total lasting perhaps 15 - 20 min. There were no SIGMETS that we were aware of; weather and windshear (w/s) data on the OFP for that segment (for FL310; we were at FL330) showed no significant w/s (4 or less). FL330 appeared to be 'in the tops' of the existing cloud layer; which showed no significant radar returns or convective activity. The weather radar system(s) appeared to be operating normally; showing appropriate ground returns etc. when cycled through range; pitch and gain settings. The first encounter was very sudden and severe; with little or no warning beyond what had been only light to moderate turbulence at the time. My first thought was that we had clipped the top of a thunderstorm cell; but there was no indication on the radar; and no lightning; hail or other weather phenomena that would indicate the presence of a thunderstorm. Initial pitch up was not extreme; but accompanied by large airspeed and altitude deviations; resulting in intermittent overspeed and stick shaker warnings. This was followed by a roll to the right of approximately 25 degrees and resulted in a 50 - 60 degree heading change and loss of LNAV capture. Observed altitude excursions were approximately +700 feet; to -500 feet from FL330. Airspeed fluctuations were too large and erratic to capture. FO; as PF; initiated recovery by disconnecting A/P and A/T and attempting to limit pitch and roll excursions and stabilize the flight path and airspeed. ATC was notified and did not report any conflict or traffic hazard in our immediate vicinity. We heard no other reports of severe turbulence on frequency; or on 121.5 mHz. There were no TCAS targets displayed nearby. Once brought back to a more stabilized path (perhaps 1 - 2min) the automation was reengaged and we again tried to determine if there were thunderstorms in the area; and again found none. A second encounter soon followed; with less severe flight path excursions; recovered in the same manner. Both events resulted in altitude and course deviations outside of regulatory limits; and were beyond our control.Some factors to consider:1. We were filed for FL310 initially as our T/O was at MGTOW. FL310 was not available from departure; we were offered FL270 or FL330. FL330 was chosen; which was the maximum for our GW. Once level we elected to use LRC over ECON initially in order to 'split the difference' between the amber bands on the airspeed indicator until our GW reduced. 3hrs later we were still in LRC which resulted in a .859m cruising speed when the turbulence was encountered. Mach .82 was entered into VNAV as soon as the automation was re-engaged; followed eventually by ECON for the remainder of the flight.2. Upset training works; FO's quick response and deliberate but smooth recovery; using techniques trained for annually; was directly responsible for the safe outcome; and undamaged aircraft; from this thankfully rare event.3. CRM/captain authority issues. CRM/TEM training works as well. While many would say hand flying in severe turbulence should be a 'captain only' event; I disagree; with certain caveats. As stated; FO was PF; yet with the first sudden encounter I also instinctively reached for the yoke and thrust levers; resulting in both pilots on the controls precisely when the need for smooth inputs is absolutely critical. Keeping that in mind; but remaining prepared; I withdrew my hands from the controls and continued as PM while FO performed the recovery. Clearly there are arguments for either case; but they are also not mutually exclusive; as I remained prepared to switch roles if I thought it was required; which I am glad to report was not the case. Bothpilots were intently focused on all aspects of the flight path; but working as a team and not in a tug of war on the controls. We don't train that way in the simulator; that is: switching roles in the middle of an upset; so we shouldn't be adding a step to the maneuver in the chaos of the real event; it is a bad time to practice a hand off; provided the PF is winning the battle.4. Fatigue/FRM. As can often happen; my rest plans did not result in the amount of sleep intended; though I was in bed in a quiet; dark room during the pre-duty rest period. The departure was late afternoon local; with full daylight for at least the first part of the flight; and we were a 3 man crew so I knew there would be a rest period. Fitness for flight is by nature a subjective call; and while I did feel somewhat tired at pick up time; I judged myself fit for duty. My practice as captain; once the PF/PM/RP duties are established; is to give first choice of rest to the PF; who in this case chose the middle rest period. As I felt sufficiently rested to stay in the seat for the first part of the flight I elected to take the last rest period. After sunset now; at 3hrs in cruise when the event began; I was experiencing some fatigue and probably was not 100% precisely when; at an instant's notice; 150% was required. I would describe this as one factor in the overall equation of this event; with perhaps the worst place for me to have been at that time would be in the bunk.Recommended practices for future events. Preparedness is the key; and planning as much as practical in terms of weather forecasts and flight planning. But; as in our case; where it was almost completely unexpected; proficient hand flying skills are the most important tool to have immediately available. My understanding is that there will be increased emphasis on real time; less scripted; training scenarios. That will likely add realism to promote greater preparedness for such extreme and unexpected events like the one we encountered.

Data retrieved from NASA's ASRS site and automatically converted to unabbreviated mixed upper/lowercase text. This report is for informational purposes with no guarantee of accuracy. See NASA's ASRS site for official report.