Assessment of climate change and global carbon cycle perturbation during the end-Permian mass extinction

Understanding the link between mass extinction and extreme global warming is critical to the projections of the consequences of future climate change. The end-Permian mass extinction is the greatest biodiversity loss in the history of animal life, and occurred about 252 million years ago. This event is coeval within error with the eruption of the Siberian Traps, one of the largest continental flood basalt provinces on Earth. Sill intrusion into organic-rich sediments and contact metamorphism in the Tunguska Basin could have released >100,000 Gt CO2, driving carbon cycle perturbation and hence global warming. Many carbon isotopic measurements from numerous locations, both in carbonate and organic matter, suggest there is a ∼ -3 to -6‰ global carbon isotope excursion during the end-Permian extinction event. In this dissertation, I used multi-disciplinary approaches to explore the overarching factors that contribute to the carbon isotope excursions and their sedimentary signature.;We evaluated proxy climate data and the existing paleoclimate simulations to assess our current understanding of the link between mass extinction and climate change. Proxies indicate that prior to the end-Permian extinction, tropical sea surface temperatures ranged from ∼22 to 25 °C, and possible atmospheric pCO2 values ranged from ∼500 to ∼4000 ppm. During the peak extinction, tropical temperatures rose up to ∼30 °C while pCO2 perhaps increased up to ∼8,000 ppm. We found climate models that use different pre-event pCO 2 values show similar amount of CO2 doubling to replicate the observed carbon isotope excursion.;We note that the expressions of temperature change and carbon isotope excursion during the extinction event show strong heterogeneity globally, thus a spatially resolved model that considers the long-term carbon cycle has advantages over simple box models. An Earth system model of intermediate complexity, cGENIE, was used to extract the pattern of CO2 release needed to replicate the observed carbon isotope excursion across the Permian--Triassic boundary. This isotopic inversion approach is novel in geochemical modeling studies and is a more precise way to match geological records. This analysis leads us to suggest that the source of CO2 must have been significantly 13C-enriched than typical biogenic or thermogenic methane to explain the significant warming that occurred during and after the extinction event.;We then further tested the ocean acidification scenario, another consequence of CO2 emission besides global warming during the end-Permian extinction. Sensitivity analyses were carried out with a range of reasonable estimates of the initial ocean saturation state. We find it most likely that the Siberian Traps volcanism released CO2 in two major multimillennial pulses. The modeled rates of C release are dependent on the delta13C of the source and the initial saturation state of the ocean. We find that the initial buffering capacity of the ocean was quickly overwhelmed for many of the plausible scenarios for C release. We suggest that global warming and ocean acidification due to the Siberian Traps volcanism might have pitched the end-Permian Earth system over a critical threshold and caused the mass extinction and subsequent long recovery.;If the observed global carbon isotope excursion truly is a reflection of the global carbon cycle perturbation, we should see similar features on land. We tested this idea by evaluating organic carbon isotope stratigraphy in terrestrial Permian--Triassic boundary sections in South China. We found that atmospheric delta13C signals can dominate other sources of variability, and thus terrestrial delta13C org is a potentially reliable tool for correlation and environmental determination during the end-Permian event. Because there are other sources of variation, however, stratigraphic correlations between marine and terrestrial sequences should only be based on carbon isotope excursions when other independent evidence of the event is available.