High-Resolution Simulations and Atmospheric Turbulence Forecasting

Forecasting atmospheric turbulence is notoriously challenging, given both (1) the enormous range of time and space scales involved, and (2) the nature of the individual processes that give rise to isolated turbulence patches and layers. Under the strongly stable conditions that characterize the stratosphere, isolated turbulence events are nearly always unresolved by traditional weather forecasting tools. Therefore, the likelihood of their occurrence must be estimated from atmospheric data (mostly balloon data), and their properties must be computed directly using high-resolution simulation techniques. The resulting combination of event frequency, intensity, and duration must then be analyzed to assess mission impact. A statistical approach for describing the high-resolution motions is necessary. Such an approach would use ensemble techniques, new methods based on Bayesian hierarchical modeling (or equivalent), or a combination of both. We adopt a probabilistic approach based on Bayes theorem.