Encounters with turbulence frequently lead to injuries aboard commercial aircraft. Many of these turbulence accidents were found to be near or within convective activity. Flight experiments conducted aboard the NASA Langley Research Center's Boeing 757 in the fall of 2000 examined the turbulence associated with thunderstorms. Turbulence measurements were computed from the in situ system, and were quantified in terms of RMS normal loads, where an RMS normal load between 0.20 and 0.30 g is considered moderate and an RMS normal load greater than 0.30 g is severe. During two flights, 18 significant turbulence encounters (RMS normal load accelerations greater than 0.20 g) occurred in the vicinity of deep convection; 14 moderate and 4 severe. In all cases, the encounters with turbulence occurred along the periphery of cumulus convection. These events were associated with relatively low radar reflectivity (RRF less than 35 dBz) and at times, went nearly undetected by the onboard radar. All four cases of severe turbulence occurred in snow at the interface between a parent cumulus updraft turret and a downwind downdraft. On the first day (Flight 190), the two severe encounters were associated with rapidly building, isolated cumulus turrets located 18 nm or more upwind of the parent thunderstorm. Similar to many reported accident accounts, low values of radar reflectivity (RRF less than 15 dBz) were detected by both the onboard and Nexrad radars during the two encounters. On the second flight (Flight 191), the two severe events were encountered during the penetration of rising plumes at the top of a line of thunderstorms. The peak radar reflectivity was slightly larger in these events, but remained below 35 dBz. No turbulence was encountered, however, when the aircraft flew in the clear air, even though it was in the immediate vicinity of the deep convection.

In situ flight records, recorded WSR-88D Nexrad base and airborne radar data, video footage from the aircraft's tail camera, and accounts from an onboard meteorological observer are used to characterize the turbulence responsible for aircraft upsets during the flight experiments. During the most severe encounter, there was good agreement in spectrum width calculations between the ground based Nexrad radar and the onboard radar. Plus, the spectrum widths support the magnitude of the normal accelerations derived from the aircraft in situ measurements.