United Airlines Flight 600

A passenger sustained a serious injury on United Airlines flight 600 when it encountered turbulence during cruise flight while enroute to Tampa International Airport (TPA), Tampa Florida. The flight was a regularly scheduled passenger flight from Newark Liberty International Airport (EWR), Newark, New Jersey, to TPA. A review of Automatic Dependent Surveillance Broadcast (ADS-B) data revealed that shortly after the airplane began a descent from its cruise altitude of flight level (FL) 380, the airplane experienced a sudden decrease in groundspeed along with a reduction in vertical speed. The groundspeed reduced from 454 knots to 438 knots over an 11 second period. During this time, the vertical speed reduced from about -2,600 feet per minute (FPM) to -900 FPM. Relatively sharp fluctuations in vertical speed and groundspeed continued for about 90 seconds as the airplane descended from FL350 to FL310, before the trends for each stabilized. See figures 1 thru 3 (Note: the time used in the figures is UTC, local time plus 5 hours.)

Figure 1. ADS-B data showing groundspeed and vertical speed at time of event.

Figure 2. ADS-B data showing groundspeed and pressure altitude at time of event.

Figure 3. ADS-B tracking data at time of event. A review of weather records revealed that the event occurred in an area immediately south of the jetstream in an area of isolated embedded convection as depicted on the National Weather Service High-Level Significant Weather Prognostic Chart valid at the time of the accident. No significant clear air turbulence was depicted over Florida. At the time of the accident, there was an area of embedded extreme intensity echoes with echo tops up to FL370 and areas of hail located about 45 miles west of the airplane’s location.

Figure 4. Regional composite radar mosaic at the time of the event. There were no convective significant meteorological information (SIGMETs) in effect for the accident location, however one was in effect for an area about 30 miles to the west and southwest of the accident, which advised of embedded thunderstorms moving north-northeast at 30 knots, with cloud tops up to FL 380. An airman’s meteorological information (AIRMET) was also in effect for the accident area and advised of moderate turbulence between FL240 and FL420. A review of pilot report (PIREPS) records revealed no reports of turbulence at or near the flight level and location of the accident for the 3 hours preceding the event. The nearest report of turbulence above FL180 was one of light to moderate chop at FL320, located 120 nautical miles north-northeast of the accident location. No PIREP records were found reported by the accident flight during the 3 hours following the event. A review of a high-resolution numerical atmospheric model over the grid point closest to the accident location revealed a temperature inversion at FL350, likely associated with the tropopause. The model also depicted a maximum wind of 60 knots at FL360, and a high probability of moderate clear air turbulence at that level. The event occurred as the airplane was downwind of an area of convective activity, in the vicinity of the tropopause, while also in an area of possible clear air turbulence, which precluded a conclusive determination about the type of turbulence (clear air or convectively induced) the airplane may have encountered. However, given the significant strength of the upwind convective activity depicted on radar, the presence of convectively induced turbulence was more likely than the clear air turbulence predicted in the numerical model.

Verbatim NTSB analysis and probable cause from the NTSB dataset