Item

Abstract

Atmospheric measurements of CO₂ mixing ratios at a number of locations around the globe have helped significantly to quantify the source—sink distribution of carbon at the global and sub-hemispheric scales (e.g. Rödenbeck et al. 2003, chapter 3 and 11). The techniques that achieve this (e.g. Gurney et al. 2002), use a globally distributed network of atmospheric concentration observations of CO₂ and other trace gasses together with an atmospheric transport model that back calculates an ‘optimal’ source—sink distribution. So far, this global inversion approach has yielded estimates of regional sinks and sources at scales of the order of a few hundreds of kilometres. For example, the distribution of the Northern Hemisphere carbon uptake in longitude between the oceans, North America, Europe and Asia is subject to many investigations but also to many uncertainties (e.g. Peylin et al. 2002; Fan et al. 1998). Stephens et al. (2007) question the common understanding of a large Northern Hemispheric sink, by comparing modelled profiles against observed ones from a few locations, and conclude that transport errors in the models may have contributed to putting too much of the global sink in the Northern Hemisphere transport. At the local scale (1 km2), direct flux measurements by the eddy covariance technique (Baldocchi et al. 2001; Valentini et al. 2000, chapter 11) constrain the net ecosystem exchange (NEE) to within 20%, comparable to the uncertainty estimated from inverse models (e.g. Janssens et al. 2003). In parallel, intensive field studies can determine the changes in vegetation and soil carbon stocks using biometric techniques, which allow independent quantification of the average carbon balance of ecosystems, albeit also with significant errors (Schulze et al. 2000; Wirth et al. 2002; Curtis et al. 2002). How the two scales, the global and local, interact at the regional level is unknown. It remains a major challenge to quantify the carbon balance at this ‘missing scale’. Understanding the link between the local and global scale will add significant value to the existing local and global networks and will ultimately help to improve the constraints on the dynamics and vulnerability of the continental scale carbon balance.