Download files
Abstract
Paleoproxy data suggest that past warm climates, such as the early Eocene (~56–48 million years ago), had considerably warmer Northern Hemisphere high latitudes and a weaker equator-to-pole temperature gradient compared to the pre-industrial period. However, general circulation models (GCMs) underestimate the magnitude of Arctic warming inferred from proxies, suggesting that certain physical processes might be poorly represented in these models.
Most previous modelling studies use coupled models, thus making it difficult to isolate the roles of different atmospheric processes in driving Arctic amplification. Therefore, in this thesis, I first conduct an atmosphere-only model intercomparison in which four models are forced with idealised polar amplified sea surface temperature (SST) anomalies (Chapter 2). The models show largely similar Arctic high cloud, surface albedo and poleward atmospheric heat transport responses, suggesting that physical processes other than these must be responsible for the large intermodel spread in Arctic amplification seen in coupled models.
Previous studies hypothesised that a large increase in Arctic polar stratospheric clouds (PSCs) caused by methane oxidation might explain the discrepancy in simulated and reconstructed Arctic amplification for past warm climates. I revisit this hypothesis in Chapter 3 using the previously applied polar amplified SST pattern but using a stratosphere resolving chemistry-climate model with a range of greenhouse gas concentrations. The water vapour concentration in the Arctic stratosphere increases monotonically with increases in methane, thereby resulting in increased PSCs. I find that the radiative effect of Arctic PSCs is approximately of the same magnitude as the radiative effect of methane. However, the greenhouse effect of the Arctic PSCs is small compared to the outgoing longwave radiation loss caused by the polar amplified SST. I also find that the Brewer-Dobson circulation (BDC) strength is sensitive to the imposed boundary conditions (SSTs and greenhouse gases).
Finally, in Chapter 4, I build on the previous experimental set-up by integrating the chemistry-climate model with early Eocene topography and more realistic SSTs for that period to understand how topography and SST changes during the Eocene might have changed the behaviour of planetary waves and thus caused a BDC change. I show that changes in topography weaken the BDC, therefore causing a reduction in the Arctic stratospheric temperature and an increase in PSCs. This result reveals the key role played by the Eocene topography in increasing the fractional coverage of Arctic PSCs, which was not seen in the simulations with modern-day topography.
This thesis has advanced our understanding of the roles of clouds and the large-scale atmospheric circulation, both in the troposphere and stratosphere, in Arctic amplification in a warmer-than modern-day climate. It provides insights into what might be missing in the GCMs to provide reliable projections of future changes in the Arctic.
