The Quasi-Biennial Oscillation and Predictability

In the mid-20^thcentury, observations revealed a varying zonal-mean zonal wind in the tropical lower stratosphere that changed sign approximately every other year, and researchers quickly and correctly inferred that the driving force for this Quasi-Biennial Oscillation (QBO) is provided by stratospheric waves. The QBO has been a dominant and continuous feature of the tropical lower stratosphere since those first observations, and (until recently) a highly predictable feature of atmospheric interannual variability.

The QBO became more than a stratospheric curiosity when early 21^stcentury research revealed associations between the phase of the QBO and northern hemisphere winter weather patterns. At the time, it was rare for a climate model to represent the QBO, but these and subsequent findings led to a steadily increasing interest in methods for self-consistent modeling of the QBO in global climate prediction and long-range forecasting models. Interest further intensified with more recent reports of a QBO modulation of the Madden-Julian Oscillation, the major mode of subseasonal tropical precipitation variability.

Links between the QBO and surface weather have been confirmed in some global model studies, but the models have consistently found weaker links than observed. Many state-of-the-art climate models still do not internally generate a realistic QBO-like circulation. Those that do rely on tuned parameterizations of non-orographic gravity wave drag. QBO simulations show the models are sensitive to small changes in configuration such as gravity wave parameter settings, horizontal and vertical resolution, numerical scheme, and location of the model top. Long-range forecasts failed to predict the unprecedented disruption of the QBO that occurred in 2016. While this failure appears to be related to extratropical Rossby wave activity, the models also failed to predict the evolution of the QBO in the year after the disruption occurred. This talk will summarize these issues, with an eye toward future observational and modeling needs.