Conflicting expectations of global surface warming

Abstract

In this dissertation I examine processes that cause Earth’s surface warming to deviate from what we might expect. Using frameworks that incorporate regional and global energy exchange, I scrutinise previous theories for why these deviations occur. The first part of this thesis examines the 1998–2012 surface-warming hiatus, in which the surface warmed more slowly than might be expected from examining model simulations or the long-term trend in observations. The preferred explanation for the hiatus is that internal variability in regional ocean heat uptake caused the surface warming to slow. However, observational analyses disagree about the ocean basin in which the definitive heat uptake occurred. Energy budgeting for the ocean surface layer, over a 100-member historical ensemble of simulations, reveals that variability in the top-of-atmosphere balance could also have caused the hiatus. Although previous studies have attributed the hiatus to fluctuations as large as 0.5 Wm−2, I show that as little as 0.08 Wm−2 could be necessary. The sensitivity of these flux deviations to the observational dataset and to energy budget choices helps explain why previous studies conflict, and suggests that the origin of the recent hiatus may never be identified. The second part of this thesis examines how climate sensitivity in model simulations grows with surface warming. The ‘pattern effect’ theory attributes this phenomenon to changing spatial patterns of warming, but previous accounts of the pattern effect disagree. I propose a new framework to unite theories about how regional processes affect climate sensitivity, and apply the framework to 1000-year simulations with a coupled climate model, exposed to abrupt CO2 increases up to sixteen-times pre-industrial concentrations. Applying the assumptions of previous studies to the model output leads to misdiagnosis of radiative forcing. Furthermore, the fact that past studies find different critical regions for the pattern effect may result from their assumptions and not divergent model behaviour. The pattern effect in the four simulations depends partly on the time elapsed since the forcing increase, and not merely the surface temperature, suggesting that current observations could underestimate climate sensitivity. Both parts of this thesis represent areas of tension in climate science between different perspectives and tools. Normative understandings may presume the superiority of empirical measurement over model simulation, or the superiority of the detailed regional perspective over the general global perspective. However, the findings presented in this thesis serve to highlight the pitfalls of restricting ourselves to one tool or view in the endeavour to understand climate.