Northern Hemisphere Jet Stream Variability in Reanalysis and Its Response to Increasing CO2 in Climate Models

This thesis examines how the Northern Hemisphere jet stream has changed in response to anthropogenic climate forcing, focusing on its waviness, latitude, and upper-level strength. It asks whether significant trends have emerged, to what extent they can be attributed to external forcing rather than internal variability, and whether changes in meridional temperature gradients explain the response through thermal wind balance. Climate model data are drawn from CMIP6, including AMIP, long pre-industrial control simulations, and 1% per year CO2 increase experiments. Jet waviness is assessed using a new index based on daily 500 hPa geopotential height contours, while jet latitude and wind speed are diagnosed at 300 hPa.

The results show no evidence of a hemispherically coherent or statistically significant increase in jet stream waviness over the last four decades. Regional and seasonal fluctuations in ERA5 remain within the range of natural variability estimated from control simulations, and climate models do not simulate a robust increase in waviness under stronger forcing. In contrast, the jet stream has shifted slightly poleward in winter, consistent with theory and model output. Although Arctic amplification weakens the near-surface equator-to-pole temperature gradient, stronger tropical upper-tropospheric warming offsets this effect. Thermal wind diagnostics show that changes in meridional temperature gradients explain more than 95% of the spatial variance in zonal wind trends and help account for the absence of widespread jet weakening.

The dominant source of variability in waviness, position, and strength remains internal. None of the observed changes in ERA5 exceed the bounds of variability estimated from long control simulations or large initial-condition ensembles. Overall, internal climate variability has so far outweighed externally forced signals, while observed patterns remain broadly consistent with theoretical expectations and model projections.