Carbon cycle variability during the last millennium and last deglaciation

The exchange of carbon on earth is one of the fundamental processes that sustains life and regulates climate. Since the onset of the Industrial Revolution, the burning of fossil fuels and anthropogenic land conversion have altered the carbon cycle, increasing carbon dioxide in the atmosphere to levels that are unprecedented in the last 800,000 years. This rapid rise in atmospheric carbon dioxide is driving current climate change and further increases are projected to dominate future climate change. However, the fate of the carbon cycle in response to climate change remains uncertain. Insight into how the carbon cycle may change in the future can come from an understanding how it has changed in the past. Key constraints on past carbon cycle variability come from the concentration and stable isotopic composition of atmospheric carbon dioxide recorded in polar ice cores, but reconstructing these histories has been a significant analytical challenge. This thesis presents a new, more precise method for measuring the stable isotopic composition of carbon in carbon dioxide (δ¹³C of CO₂) from polar ice. The new method is then used to reconstruct the atmospheric history of δ¹³C of CO₂ during the last millennium (~770-1900 C.E.) and last deglaciation (~20,000-10,000 years before present). Previously, methods for measuring the δ¹³C of CO₂ had been limited to precision of greater than ±0.05‰. The method presented here combines an ice grater air extraction method and micro-volume equipped dual-inlet mass spectrometer to make high-precision measurements on very small samples of fossil CO₂. The precision as determined by replicate analysis is ±0.018‰. The method also provides high-precision measurements of the CO₂ (±2 ppm) and N2O (±4 ppb). A new high-resolution (~20 year spacing) record of the δ¹³C of CO₂ from 770-1900 C.E is presented that suggests land carbon controlled atmospheric CO₂ variability prior to the Industrial Revolution. A deconvolution of the CO₂ fluxes to the atmosphere provides a well-constrained estimate of the evolution of land carbon stocks. The relationship between climate and land carbon for this time period constrains future climate-carbon cycle sensitivity, but an additional process affecting land carbon is required to explain the data. This missing process may be related to early anthropogenic land cover change or patterns of drought. A long-standing problem in the field of paleoclimatology is a complete mechanistic understanding of the 80 ppm increase in atmospheric CO₂ during the last deglaciation. A horizontal ice core on the Taylor Glacier in Antarctica allowed for the recovery of well-dated, large ice samples spanning the last deglaciation. From this unique archive, a new δ¹³C of CO₂ of very high resolution (50-150 year spacing) is reconstructed. A box model of the carbon cycle is used to construct a framework of the evolution of the carbon cycle during deglaciation. During the Last Glacial Maximum, the lower CO₂ concentration accompanied by only a minor shift in δ¹³C of CO₂ relative to the early…