Earth's hydroclimate dynamics and coupling to carbon cycle in observations and reconstructions

Abstract

Arising from anthropogenic carbon emissions, Earth's hydroclimate and carbon cycles have experienced great perturbations during the past decades. Accurately understanding their natural and forced variability is of crucial significance for mitigating climate change and achieving carbon neutrality. The launch of Gravity Recovery and Climate Experiment (GRACE) satellites and collections of paleoclimate proxy records present valuable opportunities to gain insights into these fields at regional to global scales. However, there are few efforts that focus on the hydroclimate dynamics of terrestrial water storage (TWS) and hydroclimate-carbon coupling over the past millennium, which is partly due to the short length of GRACE and carbon records and less communication between remote sensing and paleoclimate communities. To fill in the knowledge gaps, this thesis aims to reveal Earth’s hydroclimate dynamics and coupling to carbon cycle based on both observations and reconstructions. GRACE products and in situ measurements are first applied to identify the dominant hydroclimatic drivers of monthly TWS changes at the global scale. Results from partial correlation and hierarchical partitioning methods reveal that precipitation dominates low-latitude TWS changes in tropical, hot semi-arid, temperate, and continental climates. Evaporation is the primary driver of TWS changes in mid- and high-latitudes that feature temperate, continental, and subarctic climates. Guided by these statistical results, a hydrological model is developed to physically simulate TWS variations for the gap between GRACE and GRACE Follow-On missions. The TWS simulations achieve skillful performance as shown by independent validations against sea level budget and land surface model simulations. These results leverage statistical and physical insights into monthly TWS dynamics at the global scale. Further to the observation-based analyses, reconstructions of hydroclimate and carbon variables are conducted to enhance our understanding of their millennial variability. As for water, tree-ring width chronologies in the southwestern North America are utilized to reconstruct TWS variability during the 800-2000 period. The results reveal major historical megadroughts and contextualize ongoing drought in this region. As for temperature, this thesis devises a regional constant age method that overcomes the “segment-length curse” in dendrochronology, yielding a new Northern Hemisphere temperature reconstruction over the past millennium. The results indicate that temperature varied stably during the pre-industrial period, which is consistent with the last millennium simulations and underlines the role of seasonal bias in low-frequency temperature variations. As for hydroclimate-carbon coupling, this thesis explores the inter-annual carbon variability based upon global temperature-sensitive proxies, and demonstrates a close association of carbon growth rate with atmospheric carbon dioxide concentration and tropical temperature. Carbon variability in the late 20th century is unprecedentedly high over the past five centuries, suggestive of the instability of present carbon cycle. In summary, this thesis reveals monthly TWS dynamics, millennial TWS and temperature variability, and interannual hydroclimate-carbon coupling beyond modern observations. The synergies of observations and reconstructions are shown to be an effective way to decipher natural and forced hydroclimate and carbon variability at interannual to centennial timescales. These findings provide a solid foundation for future studies that aspire to a physical understanding of these dynamics through Earth System Models.