Abstract

Over the last decade, the concentration of carbon dioxide (CO₂) in the atmosphere has risen much more steeply than in the previous four. Recently, some have suggested that one cause for this is that the fraction of anthropogenically emitted CO₂ that contributes to annual increases in the atmospheric CO₂ concentration (“airborne fraction”) is increasing. If so, this could have important implications for international climate policy and planetary carbon cycle science. Others have argued, however, that with similar assumptions and data but more careful consideration of uncertainties, no statistically significant (p=0.9) trend in airborne fraction can be detected. One source of uncertainty in the estimation of trend in the airborne fraction is uncertainty in the time history of land use change emissions of CO₂; we focus here on the consequences of this uncertainty. We use Monte Carlo techniques to estimate how large linear trends in land use change emissions would have to have been in order to yield a statistically significant increase or decrease in the airborne fraction over the past 50 years. We also show that, under the range of published assumptions, airborne fraction trends can be either positive or negative, but are not statistically significant. Uncertainty about historical land use change emissions alone prevents reliable detection of airborne fraction trend.