Explaining fluctuations in the period of the quasi-biennial oscillation

The quasi-biennial oscillation (QBO) dominates stratospheric variability on interannual time scales. It is known to modulate tropical convection activity as well as extratropical circulation and surface climate during winter. However, accurately predicting the evolution of the QBO is challenging due to fluctuations in its period. The QBO period has varied between 20 and 35 months since its first observation in the 1950s, but the underlying mechanism remains unresolved. Previous studies hypothesized that seasonal variations in tropical stratospheric upwelling and/or atmospheric wave activity could lead to fluctuations in the QBO period by modulating the oscillation phase speed. While this hypothesis has been tested using simple QBO models, climate models that reproduce the seasonal variation of stratospheric upwelling have failed to simulate sufficiently large fluctuations in the QBO period. Additionally, the full variability in atmospheric wave activity has been difficult to observe, given the broad spectrum of tropical waves of various types. In this study, we use a recent reanalysis product to investigate the mechanism behind the fluctuations in the QBO period. We found evidence that waves with wavelengths of several hundred kilometers or less (i.e., gravity waves) exhibit an annual cycle in their activity that explains these period fluctuations. This annual cycle is coherent with the observed variations in the QBO phase speed. This finding enables us to successfully regress the QBO phase speeds using small-scale wave activity. Meanwhile, there is an indication that the seasonal variation of stratospheric upwelling plays a relatively minor role in the QBO period fluctuations. We also discuss the origins of the annual cycle of small-scale waves.