A New Way to Quantify Stratosphere-Troposphere Coupling in Observations and Climate Models

Abstract

Atmospheric mass is transported in and out of the stratospheric polar cap region by a wave-driven meridional circulation. Using composites of polar cap pressure anomalies, defined as deviations from the average annual cycle, it is shown that this stratospheric mass flux is accompanied by a similar mass flux near the surface. This 'tropospheric amplification' of the stratospheric signal is introduced as a new way to quantify stratosphere-troposphere coupling. Regression analysis is used to create a vertical profile of atmospheric pressure during a tropospheric amplification event, and the regression slope profile is used as a tool to quantify the amplification. Using data from 5 reanalysis datasets and 11 climate models, it is shown that high-top models, with a model lid of above 1 hPa, are significantly better at reproducing tropospheric amplification than low-top models, due to having more detailed parameterisations of stratospheric processes. However, the regression slope profiles of all models, bar one, are significantly different to the profile of reanalysis data at a 95% confidence level. Tropospheric amplification is also investigated in historical and future simulations from these models, and it is concluded that there is not expected to be a large change in the phenomenon over the next 100 years. The processes needed to reproduce tropospheric amplification can be identified by comparing idealised models of different complexity. A simple dry-core model is not able to reproduce tropospheric amplification, while a model with a comprehensive radiation scheme does produce the basic regression slope profile under certain configurations. The associations between pressure change and mass flux are further investigated using primitive equations. It is found that vertical and horizontal contributions to mass flux act to mostly cancel each other out, leaving a poorly-conditioned residual, and that the horizontal mass flux across the polar cap boundary has both geostrophic and ageostrophic components.