This thesis assesses the representation of the key processes determining water vapour entry into the stratosphere in reanalysis and two global climate model configurations. This is done by applying the advection-condensation method and comparing alternative formulations.

In the region of the atmosphere around 15–50 km altitude known as the stratosphere, water vapour is present in concentrations around ten thousand times lower than at the Earth’s surface. Yet, water vapour still influences surface warming and stratospheric chemistry due to its radiative and chemical properties. Previous studies have identified low temperatures (more accurately, low saturation mixing ratios) and large-scale transport through the tropical tropopause layer to be the key factors that determine stratospheric water vapour. However, their relative importance at different timescales, and the role of the other influences such as detailed ice microphysics, are not yet well known. It is a crucial priority to improve the generally poor representation of stratospheric water vapour in global climate models for both present and future scenarios.

In the first chapter of results, the representation of the average annual cycle of lower stratospheric water vapour by temperature and large-scale transport is assessed. Model-specific features are reflected in water vapour predictions by the advection-condensation method. Applying a climate model advection scheme to the advection-condensation method, to test sensitivity to transport formulation, finds more similar temporal variability but also affects vertical attenuation of concentrations in the stratosphere. Expanding on earlier studies, the impact of sub-seasonal temperature variability on stratospheric water vapour is quantified in reanalysis and in one global climate model where it is found to be under-represented.

Following on, the next chapter investigates the substantial interannual variability of tropical lower stratospheric water vapour by isolating temperature and transport impacts in reanalysis. The approach asks whether the characteristics of a particular year are obtained by transport through temperatures from another year. Results identify almost total independence from transport variations across years, but important seasonal variability. This agrees with many studies on seasonality of transport, and points squarely to temperatures controlling interannual variability.

The subsequent chapter takes advantage of the complete water budget available in a global climate model to assess the impact of additional processes. Results find the phase change of ice to vapour (sublimation) is a substantial component of the water vapour budget above the tropical tropopause, and convective injection of ice occurs above the vertical minimum in saturation mixing ratio. Results also show that the extent to which advection-condensation calculations are rehydrated by different measures of sublimation depend crucially on their vertical extent.

The final chapter of results analyses the response of advection-condensation and sublimation to climate change scenarios. Increases in transport efficiency through the tropical tropopause agree with well understood aspects of climate change. Convective ice injection is higher but no more intense, whereas sublimation above the vertical dry point has increased. These changes appears to be controlled solely by the elevated and warmer tropopause. The results show that predictions of a wetter model stratosphere, both with and without sublimation, scale similarly with the higher saturation mixing ratios at the tropical tropopause.

Overall, this thesis identifies the relative impact on stratospheric water vapour from temperatures, large-scale transport and ice sublimation in the tropical tropopause on different timescales. Many of the findings are in the context of the global climate models studied, motivating further development to represent more accurately both the present and projections across this century.