Development of a high precision carbon dioxide-14 measurement capability and application to carbon cycle dynamics

Anthropogenic fossil fuel combustion has perturbed the global carbon cycle, causing an increase in atmospheric carbon dioxide (CO2 ) concentration and altering carbon exchange with the oceans and terrestrial biosphere. Significant inter-annual variability in atmospheric CO2 concentration is likely due the uncertain terrestrial biospheric exchange, but fossil and biological CO2 fluxes are generally co-located over the continents and difficult to separate using atmospheric observations. In this thesis, precise measurements of the radiocarbon content of atmospheric CO2 (Delta14CO2) are used to separate these fluxes, based on the ability of Delta14CO2 observations to discriminate between 14C-devoid fossil fuel CO2 and other CO2 sources, which are 14C replete.; Methods for ultra-precise Delta14CO2 measurement with a repeatability of 1.8‰ (1sigma) in whole air samples of 2-5L are developed, representing a significant improvement over previous methods. These permit use of existing air sampling networks to detect atmospheric fossil fuel CO2 content to better than 1ppm, and to characterize annual, seasonal and spatial changes in atmospheric Delta14CO 2 from a variety of sources.; Delta14CO2 variability is examined in several measurement campaigns. A time series of Delta14CO2 observations from Niwot Ridge, Colorado exhibits annual and seasonal changes of 3--6‰ with the signal dominated by fossil fuel CO2 emissions. The fossil fuel and biological CO2 components in boundary layer air over New England, USA are determined using Delta14CO 2 and CO2 observations from aircraft, and the biological CO2 component is found to be consistent with previous bottom-up observations. The results are also compared with two fossil fuel correlate tracer methods (CO and SF6) showing that both correlate methods exhibit significant biases. Train-borne observations from across Eurasia show a spatial gradient in the continental boundary layer of 5.0--7.7‰, which, when corrected for local influences, appears to represent the dispersion of European fossil fuel CO2 emissions. This is in broad agreement with modeled predictions, but demonstrates atmospheric transport model sensitivity to the (poorly constrained) vertical mixing parameterization. Finally, the difference between tropospheric and stratospheric Delta14CO 2 is used to constrain the cross-tropopause exchange time to 2.7--4.3 years over Japan, and the stratosphere-troposphere 14C disequilibrium, which is needed for simulations of the tropospheric 14C distribution.