Document Type


Publication Date



Measurements of atmospheric O2 have been used to quantify large-scale fluxes of carbon between the oceans, atmosphere and land since 1992 (Keeling and Shertz, 1992). With time, datasets have grown and estimates of fluxes have become more precise, but a key uncertainty in these calculations is the exchange ratio of O2 and CO2 associated with the net land carbon sink (B). We present measurements of atmospheric O2 and CO2 collected over a 6-year period from a mixed deciduous forest in central Massachusetts, USA (42.537 N, 72.171 W). Using a differential fuel-cellbased instrument for O2 and a nondispersive infrared analyzer for CO2, we analyzed airstreams collected within and 5m above the forest canopy. Averaged over the entire period of record, we find these two species covary with a slope of -1:081±0:007 mol of O2 per mole of CO2 (the mean and standard error of 6 h periods). If we limit the data to values collected on summer days within the canopy, the slope is -1:03±0:01. These are the conditions in which biotic influences are most likely to dominate. This result is significantly different from the value of -1.1 widely used in O2-based calculations of the global carbon budget, suggesting the need for a deeper understanding of the exchange ratios of the various fluxes and pools comprising the net sink.