A mechanism for convectively-induced turbulence

On 5 August 2005, two separate commercial aircraft at cruising altitude encountered severe turbulence over Northwest Indiana, in the ostensibly clear air after having flown over or around a large convective storm. According to contemporaneous satellite imagery, the planes were roughly 20 km away from any cloud having appreciable optical depth. The turbulence measurements were based on output from the onboard eddy dissipation rate (EDR) algorithm which provides quantitative estimates of atmospheric turbulence at one-minute intervals along the flight path. This encounter is another example of convectively-induced turbulence (CIT); in this case the turbulence occurred laterally away from the cloud and not above it.

The synoptic situation based on RUC model analyses showed that the turbulence event was associated with a particularly intense convective storm located at the northeast end of a line of scattered storms oriented along a southeastward moving synoptic cold front. At flight levels (~12 km) the environmental winds were weak, with moderate vertical shear and low stability such that the background Richardson number (Ri) was relatively small (~2), but still not small enough to account for the observed turbulence.

To further elucidate the CIT forcing mechanisms in this situation, high-resolution WRF model simulations with and without latent heating were compared. The cloud-resolving simulations successfully captured the rapid growth of the strong convection that occurred along the cold front, and showed that the convection (1) enhanced the environmental wind shear and lowered the stability in a zone stretching some distance beyond the anvil and (2) generated trapped gravity waves that propagated into the clear air surrounding the storm. The two effects together were sufficient to reduce local values of the Ri in the environment surrounding the storm to a critical value for turbulence generation. The complexity of the situation indicates that the CIT episode would have been very difficult to forecast without resort to high resolution (~1 km horizontal, 250 m vertical) nonhydrostatic models.