作者：B. F. Bennett, R. J. Salawitch, L. A. McBride, A. P. Hope and W. R. Tribett

The airborne fraction of atmospheric CO2 (AF), defined as the annual global CO2 growth rate (dCO(2)/dt) divided by the total emission of CO2 from combustion of fossil fuels and land use change (LUC), has a long-term average of similar to 0.44 over the past six decades. When quantifying trends in AF it is important to account for inter-annual variability in dCO(2)/dt due to natural factors such as the El Ni & ntilde;o Southern Oscillation (ENSO) and major volcanic eruptions, as well as assumptions regarding LUC. Here, a multiple linear regression model is used to compute dCO(2)/dt as a function of anthropogenic CO2 emissions, ENSO indices, and stratospheric aerosol optical depth (a proxy for major volcanic eruptions), for numerous time series of the emission of CO2 due to LUC (E-LUC). For 20 out of 21 previously published E-LUC time series, the trend in AF adjusted for natural variability (AF(ADJ)) over 1959 to 2021 exhibits a trend that is statistically indistinguishable from zero and lacks statistical significance at the 95% confidence interval. Therefore, it is most likely that the relative efficacy of the combined global terrestrial biosphere and oceanic carbon sinks has been fairly constant on a global scale over the past six decades. Since the trend in AF exhibits considerable variability depending on which E-LUC time series is used, more precise knowledge of the actual value of the AF trend will require resolving the current large differences in various estimates of E-LUC.

来源：Journal of Geophysical Research-Biogeosciences. 2024 Vol. 129 Issue 3. DOI: 10.1029/2023jg007760