The ability of the terrestrial biosphere to absorb more carbon from the atmosphere than it releases may prove to be an important mechanism for climate change mitigation. However, the effects of climate on biospheric net carbon uptake and sequestration remain highly uncertain. This dissertation examines how the North American biosphere, a large carbon sink, responds to climate and meteorological events.

First, this work addresses the vulnerability of the carbon balance of high-latitude North American biomes to warming, with a focus on Alaska. This question is investigated using a geostatistical inverse model with a novel configuration to determine the potential attribution of the interannual variability of carbon fluxes to climate conditions. This study finds that Alaskan tundra and other Alaskan ecosystems dominated by grasses and shrubs are highly vulnerable to net losses of carbon during the autumn and early winter under the warm temperatures that are quickly becoming the norm in this region. In contrast, the net carbon losses of high-latitude boreal forests under warming may be substantially smaller because of an extension of the growing season that occurs under warm conditions in this biome.

Then, this dissertation addresses the effects of large-scale atmospheric systems on regional and continental biospheric carbon fluxes over midlatitude North America, via two studies that employ a clustering algorithm to identify continental-scale meteorological events. First, several observational sites across a highly productive forested region near the Great Lakes are studied to determine regional-scale responses to nearby storm track activity and associated anomalies of temperature and cloud cover. This work finds that during storm track activity that occurs with strong southerly atmospheric inflow from the Gulf of Mexico, net carbon uptake is suppressed in the Great Lakes region due to a combination of cloud cover and sharp increases in temperature.

In the final chapter, the response of the North American carbon sink to continental atmospheric dynamics is generalized across the continent and across summertime midlatitude cyclonic events. More than 50 observational sites across the continent are used in a geostatistical interpolation to map carbon flux anomalies associated with various cyclonic placements, and meteorological variables are evaluated for their ability to explain geographical patterns in flux anomalies and are assimilated into the interpolation. This analysis finds that net carbon uptake is consistently suppressed in the warm sector and enhanced in the cold sector of cyclonic systems, with these anomalies extending throughout vast regions of the continent. Both temperature and cloud cover can explain the large spatial extent of these carbon flux anomalies. This work demonstrates how atmospheric systems can be studied to infer large geographical patterns of carbon uptake suppression or enhancement.