Clear-Air Turbulence Is a Shear Predictor

01The evidence

• Machine-learning approaches to clear-air turbulence prediction over southern Brazil have improved forecast skill using global model data, which shows the phenomenon is predictable from the flow field rather than from convection.Meteorology and Atmospheric Physics, 2026
• Automated procedures for forecasting clear-air turbulence were developed decades ago around shear and stability indices, which is the diagnostic family a shear account uses.Journal of Applied Meteorology

02The tension

Turbulence forecasting is organised around convection: storms are visible, they are where pilots report bumps, and the reporting bias reinforces the association. Clear-air encounters happen where there is nothing to see, and they are under-forecast because the diagnostics are trained on a biased sample.

03The theory

Clear-air turbulence is generated by shear instability in stably stratified flow, most often at the edges of the jet stream where the vertical shear is large and the Richardson number is small.

If shear is the generator, then encounters should cluster along shear layers, and the sharpness of the layer should predict severity better than the presence of convection. The forecast problem becomes tracking a thin surface in the flow rather than predicting storms.

The prediction is that a shear-layer diagnostic should outperform convection-based indices on clear-air encounters, and that reported turbulence should be biased toward convective regions because that is where aircraft fly and look.

The operational consequence is that route planning should avoid shear layers explicitly, and that the reporting bias should be corrected before training any forecast model.

04What this is not

05Predictions

• P1A shear-layer diagnostic should outperform convection indices on clear-air encounters.
• P2Encounters should cluster along shear layers rather than near storms.
• P3Reported turbulence should be biased toward convective regions.

06What would refute it

Encounters tracking convection after shear is controlled would restore the storm-based account.

07What it would change

It would change turbulence forecasting and route planning, and correct a reporting bias.