What Can Observed Temperatures Tell Us about Stratospheric Dynamics over the past 40 Years?

The Brewer-Dobson circulation affects stratospheric temperature and the abundance and distribution of important trace constituents such as water vapor and ozone. Climate model simulations predict a strengthening of the Brewer-Dobson circulation in response to greenhouse gas forcing. Consistent with these expectations, recent observational studies have identified increasing trends in the Brewer-Dobson circulation, but there are significant differences between the magnitude of the trends across reanalysis datasets and across diagnostics within the same dataset. Further, there remain unresolved differences in temperature and ozone trends calculated from different data sources. From the thermodynamic energy equation, it can be deduced that wave-driven increases in the Brewer-Dobson circulation cool the tropics and warm the extratropics. Previous studies have utilized this temperature see-saw to develop Brewer-Dobson circulation indices. However, the observed temperatures are the result of complex interactions between atmospheric dynamics, and radiatively active trace constituents whose distribution are, on the one hand, affected by dynamics, and on the other hand affect stratospheric dynamics through thermal forcing of the circulation by radiation. Here, we seek to disentangle these interactions, and interpret the observed record of stratospheric temperatures - the primary source of observational information of the stratospheric state - in terms of a conceptual model. The model distinguishes between pure dynamical forcing, and thermal forcing, and explicitly incorporates feedback terms (for example the ozone response to pure dynamical forcing) and thermal forcings independent of dynamical forcing (for example stratospheric aerosol loading following volcanic eruptions such as Mt. Pinatubo). The inverse model considers the information of tropical average, northern extratropical and southern extratropical average temperature. Theoretical considerations, numerical model calculations with a primitive equation model, and radiative transfer calculations are used to construct templates in temperature space against which the observed temperatures are projected. The slow CO2-induced change in background state poses a challenge to the inverse model, and additional information in the form of frequency-filtering is required. The strength of the Brewer-Dobson circulation derived from this model is reasonably well correlated with previously published estimates of diabatic mass fluxed derived from reanalysis data. For periods following volcanic eruptions, the model diagnoses elevated thermally-forced circulations with discrepancies in net strength compared to reanalysis-based estimates highlighting remaining uncertainties in the stratospheric state following volcanic eruptions. Of particular interest is the period of the 1980's when disagreements among temperature datasets on the one hand, and between temperature and estimates of the strength of the BD circulation based on reanalysis data on the other hand, are largest. Using the inverse model, we seek a probabilistic exploration of consistent, and inconsistent combinations of datasets, and derive an assessment of the robustness of our understanding of the stratospheric state in that period.