The atmospheric concentration of CO₂ is known to vary on multiple spatial and temporal scales. At present, the global mean tropospheric carbon dioxide concentration is of order 375 ppmv; the global mean concentration has been increasing by between 1.5 and 2 ppmv because of anthropogenic emissions associated with fossil fuel burning, cement production, and industrial activity. While projections of future climate change associated with increased atmospheric CO₂ loading often use only the global mean value, there is some spatial structure inherent in the tropospheric CO₂ field. For example, because of the strong localization of anthropogenic emissions of carbon dioxide in the Northern Hemisphere, there exists a mean interhemispheric gradient of order 3 ppmv. Within the Northern Hemisphere, zonal and meridional differences of order 0.5 ppmv may exist. Temporal, apart from the secular growth in CO₂ associated with anthropogenic emissions, there are time-variations in atmospheric CO₂ from daily to seasonal to interannual (and beyond) timescales. One of my current research concerns the controls of the spatiotemporal variations in CO₂. One important reason to undertake a careful analysis of such variations is that atmospheric carbon dioxide concentrations at remote mid-ocean observatories like Mauna Loa, Hawaii are often taken to be representative of the large-scale fluctuations of earth's terrestrial ecosystem (e.g. the competition between respiratory and photosynthetic fluxes of carbon to and from the atmosphere). Consideration of the variabilty of the terrestrial ecosystem at (say) interannual may shed light onto the sensitivity of the terrestrial ecosystem to projected anthropogenic climate change, which in turn is important for understanding how terrestrial ecosystem change may feedback onto future climate change.