Paleoclimate data assimilation with CLIMBER-X in transient simulation of the last deglaciation

Abstract

Paleoclimate studies provide crucial insights into the Earth's climate system drivers and behavior, helping distinguish between natural climate patterns and human-induced changes and providing a context for current climate trends. The last deglaciation marked a period of Earth's history characterized by the retreat of massive ice sheets covering large parts of the planet. During this phase, a drastic transition occurred from the cold Last Glacial Maximum to the warmer and more stable climate of the Holocene. Besides proxy-based reconstructions and climate model simulations, data assimilation (DA) has emerged as a promising reconstruction technique of past climates. DA combines the data and the underlying dynamical principles governing the climate system to provide a state estimate of the system, which is better than that which could be obtained using just the data or the model alone. As the first step of this dissertation, major uncertainties and forcing factors are evaluated for the last deglaciation using an efficient climate model. Two sets of transient simulations are based on the novel ice-sheet reconstruction PaleoMist and the established GLAC1D. The simulations reveal that the proximity of the Atlantic meridional overturning circulation (AMOC) to a bifurcation point, where it can switch between on- and off-modes, is primarily determined by the interplay of greenhouse gas concentrations, orbital forcing, and freshwater forcing. The impact of deglacial meltwater on AMOC shapes regional temperature patterns. PaleoMist simulation shows a pronounced warming in the Bølling-Allerød (BA), a strong AMOC, and a moderate cooling during the Younger Dryas (YD) with an AMOC weakening. The opposite signature is found for GLAC1D. The PaleoMist simulation replicates, at least qualitatively, the BA/YD sequence: a warming in Greenland and Antarctica in the BA, a cooling northern North Atlantic, and a warming in Antarctica in the YD. In addition, as the main goal, this study presents an efficient method for assimilating the temporal evolution of surface temperatures for the last deglaciation. In applying an ensemble Kalman filter approach, the study makes use of the advances in the parallel DA framework (PDAF). It is found that the DA solution depends strongly on the background evolution of the decaying ice sheets rather than the assimilated temperatures. Further, the influence of DA is more pronounced on regional scales than on the global mean. In particular, DA has a stronger effect during millennial warming and cooling phases, such as BA and YD, especially at high latitudes with heterogeneous temperature patterns. In the final step, the DA technique is developed to assimilate the subsurface temperatures. The results show the DA of subsurface temperatures modifies the precipitation pattern, which can consequently change the timing and magnitude of climate events such as BA and YD. The DA technique introduced and developed in this dissertation facilitates the paleoclimate DA on multi-millennial time scales and can also be employed to study future warming scenarios.