On the relationship between QBO and distribution of tropical deep convection

There have been several observational and modeling analyses that have indicated that the stratospheric quasi-biennial oscillation (QBO) significantly affects the tropical troposphere. This article aims to identify the QBO signal in tropical deep convection. Some difficulties in previous studies were ambiguities in the identification of tropical deep convection in observations, and also in separating the El Nino/Southern Oscillation (ENSO) and other tropospheric signals from QBO influences. We use a recent cluster analysis of 21.5 years of International Satellite Cloud Climatology Project tropical weather states to identify tropical deep convection and cirrus clouds, as well as 32.25 years of precipitation data as proxies for deep convection. Correlations between the QBO, ENSO, and other tropospheric patterns such as the tropospheric biennial oscillation and Pacific decadal oscillation are taken into account to isolate the influence of the QBO. Whereas tropical deep convection is mostly related to ENSO and the annual cycle, the QBO westerly phase, independent of the annual cycle as well as linear and nonlinear impacts of ENSO, leads to an eastward shift in the strength of meridional overturning contributions to the Hadley circulation over the Pacific and thus also affects the subtropical circulation. For deep convective clouds, relative differences in convective cloud cover between the QBO easterly and QBO westerly phases can be as large as 51% +/- 7% of the annual average over isolated regions in the tropical west Pacific and 103% +/- 35% over the east Pacific, where the absolute values are lower and where notable deviations occur during the QBO westerly phase.