On the signatures and drivers of abrupt climate change in the high latitudes

Abstract

Abrupt shifts in the climate system have been observed in paleoclimate records and are the subject of much concern for the future climate as anthropogenic greenhouse gas emissions continue to drive global climate change. Some of these abrupt changes are thought to be associated with climate “tipping points”, where some component of the climate system changes rapidly and irreversibly at a threshold value of CO2. One way to understand such shifts comes from conceptual, mathematical, models of the climate that suggest it is possible to have multiple climate equilibria for a given external forcing value (such as CO2 concentration). A tipping point in such models occurs when the CO2 concentration crosses a threshold where one climate equilibrium disappears or loses stability, and the system must rapidly adjust to a different equilibrium state. While this mathematical insight provides an elegant explanation for abrupt climate shifts, the gulf between the simple hypothetical systems that display this behavior and Earth’s climate system is vast. Earth’s climate is different from such a simple system in that it has many more components, diffusive properties that smooth out abrupt transitions, and is subject to noisy perturbations and time-changing forcing that can cause it to be far from its equilibrium state. These differences shed doubt on the applicability of the dynamical systems perspective to the Earth’s climate, and on whether we should ever expect to actually observe large-scale climate tipping points. In my thesis, I try to bridge the gap in our understanding between the tipping points displayed by simple conceptual systems and the complicated abrupt shifts seen in full-complexity models by studying two elements of the climate system whose status as a “tipping point” remain debated: winter Arctic sea ice, and the Atlantic Meridional Overturning Circulation (AMOC). In Chapter 2, I analyze six global climate models (GCMs) that show a range of Arctic winter sea ice loss rates under global warming scenarios, from relatively gradual to abrupt tipping-point-like loss. I identify a year-to-year mechanism that links a springtime ice-albedo feedback to a wintertime longwave feedback via ocean heat storage as the mechanism that drives inter-model differences in the abruptness of winter Arctic sea-ice loss. In Chapter 3, I develop a novel idealized model that combines sea ice thermodynamics and atmospheric processes. I use this model to test the hypotheses from Chapter 2, finding that indeed atmospheric feedbacks can modulate the abruptness of sea ice loss, but they have a much more significant impact on the timing of sea ice loss. In Chapter 4, I shift to the question of how to understand and detect tipping points---which are, by definition, a feature of the equilibrium climate---when the climate system is not in equilibrium. Using an idealized model of sea ice, I develop a new method for detecting tipping points in models without directly simulating the equilibrium climate at many CO2 concentrations, which is computationally infeasible for GCMs. I then test this method on a GCM that shows an abrupt loss of sea ice and use it to identify an upper bound for the CO2 value at which complete sea ice loss is guaranteed in that model. However, I find that the separate timescales of relatively fast sea ice responses to CO2 and the slow adjustment of ocean temperatures complicate the application of this method of detecting tipping points in GCMs. Finally, in Chapter 5 I explore the out-of-equilibrium behavior of the AMOC in a fully coupled climate model driven with different ramping rates of CO2. I find that when CO2 changes faster than the AMOC can respond, the AMOC exhibits interesting out-of-equilibrium behavior that depends on the CO2 ramping rate. While the literature has focused heavily on AMOC tipping points, I find that even without undergoing a tipping point, the AMOC can enter long (multi-centennial) periods of transient weakening. I then propose a novel AMOC-ocean heat transport-sea ice feedback that can explain how the level of AMOC weakening depends on the rate of CO2 ramping.