Irreversibility of global climate system and its potential predictability

Abstract

After the Industrial Revolution, persistent emissions of anthropogenic carbon dioxide (CO2) have resulted in long-term, large-scale warming and consequent global-scale climate changes. To mitigate this climate change and minimize the associated risks, humanities have pledged to reduce the atmospheric CO2 concentration by climate mitigation policies such as net zero or negative CO2 emissions (e.g. 2015 Paris Agreement). This crucial step is essential to ensure a sustainable future; therefore, it is necessary proactively to understand how the climate system responds to various climate mitigation policies. However, the effectiveness of atmospheric CO2 reduction remains highly uncertain due to the complexity and nonlinear nature of Earth system. This uncertainty gives rise to important question of whether the ongoing climate change is indeed reversible. Hence, this dissertation mainly investigates the climate irreversibility and specifically aims to advance the dynamical understanding of climate changes in response to mitigation scenarios. To achieve this goal, I conduct a large ensemble experiment using a state-of-the-art climate model, involving various CO2 forcing scenarios to simulate climate mitigation strategies, and subsequently, analyze them based on the concept of climate hysteresis and irreversibility. This modeling approach for understanding climate irreversibility will be instrumental in developing robust climate strategies capable of effectively mitigating the existing and potential risks posed by global warming. In Chapter II, based on idealized atmospheric CO2 ramp-up and symmetric ramp-down model experiments, I investigate the large-scale climate hysteresis and irreversibility. The global mean surface temperature and precipitation, a representative metric to measure the extent of global warming, reveals a roughly linear response. However, in the first section of Chapter II, I find that the zonally elongated intense tropical rain belt exerting a crucial role in global hydrological cycles, known as intertropical convergence zone (ITCZ), does not respond linearly to CO2 forcing, but exhibits strong hysteresis behavior (i.e. delayed recovery). While the latitudinal location of the ITCZ changes minimally during the ramp-up period, it moves sharply south as soon as CO2 begins to decrease, and its center eventually resides in the Southern Hemisphere during the ramp-down period. Such ITCZ hysteresis is associated with delays in global energy exchanges between the tropics and extratropics. The delayed energy exchanges are explained by two distinct hysteresis behaviors of the Atlantic Meridional Overturning Circulation (AMOC) and slower cooling in the Southern Ocean. I also suggest that the ITCZ hysteresis can lead to hysteresis in regional hydrological cycles. In other words, the hysteresis of the ITCZ is caused by the tug-of-war between delayed AMOC recovery-induced faster cooling in the Northern Hemisphere, and relatively slower cooling in the Southern Hemisphere due to the continued heat release from the Southern Ocean to atmosphere. This implies that the crucial role of the ocean in hysteresis/irreversibility in the global climate system. Therefore, in the second section of Chapter II, I examine the comprehensive contribution of the global ocean to irreversible climate changes. Generally, the global oceans serve as a huge thermal buffer to absorb the surplus heat from the anthropogenic surface warming. When atmospheric CO2 concentration is reduced as humanity tries to mitigate global warming, how this massive heat accumulation in the ocean interior will affect the Earth's climate remains unclear. Here I show that this stored heat will be released at a much a slower rate than its accumulation to yield multi-century irreversible ocean warming and consequent regional irreversibility in the sea surface temperature (SST) and precipitation. The irreversible SST changes are pronounced over subpolar-to-polar regions and the equatorial eastern Pacific where oceans are weakly stratified to allow vigorous heat release from deep ocean to the surface layer. I also highlight that this SST pattern largely explains the irreversible precipitation pattern. This study suggests that deep ocean warming may hinder climate recovery with a unique climate pattern, even if carbon neutrality or net negative emissions are successfully achieved. In addition, there is a considerable uncertainty in the climate irreversibility between different climate models and even within ensemble members. Understanding the sources of this uncertainty and minimizing it is therefore of great socio-economic benefit, because it can provide a more accurate information on future projections. This precision is particularly crucial for informed decision-making in the realm of climate mitigation policy. Hence, in Chapter III, I mainly focus on the potential climate uncertainty in mitigation scenarios. In the first section of Chapter III, I find that the most uncertain region in response to CO2 removal is the Arctic. The Arctic temperature shows a strong inter-ensemble spread when the CO2 concentration returned to the present level. This different Arctic temperature in each ensemble member can be explained by the density and salinity of the Arctic Ocean about a century ago. For example, a denser Arctic Ocean condition in the early CO2 reduction pathway slows down the recovery of the AMOC in the presence of positive AMOC-salt-advection feedback. Because faster AMOC recovery can transport more warm water into the Atlantic sector of the Arctic, Arctic cooling as a result of CO2 reduction is delayed. In addition, denser Arctic water enhances vertical mixing, which also results in delayed Arctic cooling under a strong vertical temperature gradient in the subpolar-to-polar Atlantic. This finding suggests that the Arctic's initial states have a centennial memory for the future Arctic and global climate changes. In light of this, the AMOC is a pivotal factor in the high uncertainty of global climate system in mitigation scenarios. Indeed, in the second section of Chapter III, I find that the clear AMOC tipping events in response to climate mitigation scenarios. Specifically, at the given identical mitigation CO2 forcing, some ensembles project the AMOC recovery in response to the cessation of the anthropogenic CO2 emissions as expected, however, the others exhibit a AMOC collapse. This totally different behavior of the AMOC is explained by the existence of multiple AMOC equilibrium states associated tipping point. It is suggested that the accumulation of the small atmospheric stochastic variability determines whether or not the AMOC crosses the tipping point across different ensemble members. Since the global warming gradually pushes the AMOC into the unstable regime from the stable regime, all ensembles ultimately converge on a trajectory leading to AMOC collapse. Notably, even a slight delay in the implementation of mitigation policies by a few years leads all ensembles to collapse of the AMOC; this underscores the high uncertainty of future climate system due to the non-linear behavior of the AMOC beyond the TP. These findings highlight the critical importance of understanding AMOC dynamics in the broader context of climate mitigation strategies and underscore the urgency of timely policy interventions. This dissertation sheds light on the potential climate irreversibility in future climate mitigation scenarios and the mechanisms involved. By unraveling the complex mechanisms at play, it highlights that the irreversibility of global and regional climate system would require adaptation and mitigation measures to be designed and applied for much longer periods than those initially anticipated. Prompt measures to reduce greenhouse gases emissions must therefore be taken to minimize long-lasting changes to the climate system.