Abstract
Published eddy covariance measurements of carbon dioxide (CO2) exchange between vegetation and the atmosphere from a global network are distilled, synthesised and reviewed according to time scale, climate and plant functional types, disturbance and land use. Other topics discussed include history of the network, errors and issues associated with the eddy covariance method, and a synopsis of how these data are being used by ecosystem and climate modellers and the remote-sensing community. Spatial and temporal differences in net annual exchange, FN, result from imbalances in canopy photosynthesis (FA) and ecosystem respiration (FR), which scale closely with one another on annual time scales. Key findings reported include the following: (1) ecosystems with the greatest net carbon uptake have the longest growing season, not the greatest FA; (2) ecosystems losing carbon were recently disturbed; (3) many old-growth forests act as carbon sinks; and (4) year-to-year decreases in FN are attributed to a suite of stresses that decrease FA and FR in tandem. Short-term flux measurements revealed emergent-scale processes including (1) the enhancement of light use efficiency by diffuse light, (2) dynamic pulses in FR following rain and (3) the acclimation FA and FR to temperature. They also quantify how FA and FR respond to droughts and heat spells.