Quantification of the Trend in the Airborne Fraction of Atmospheric CO2

The airborne fraction of atmospheric CO₂ (AF) is defined as the global annual atmospheric CO₂ growth rate (dCO₂/dt) divided by total anthropogenic CO₂ emissions (E_TOT). Here, E_TOT is the sum of global CO₂ emissions from the combustion of fossil fuels, cement production (including the cement carbonation sink) and industry, as well as land use change (E_LUC). The average of AF over the last six decades is approximately 0.44 (i.e. just under half of E_TOT has remained in the atmosphere). There is considerable interest in determining whether AF has changed over time because a rise in AF would indicate the existence of a feedback between the global carbon cycle to climate change, possibly inhibiting society’s ability to limit global warming to the temperature thresholds set by the Paris Climate Agreement. When quantifying trends in AF it is important to account for inter-annual variability in dCO₂/dt due to natural factors such as the El Niño Southern Oscillation (ENSO) and major volcanic eruptions. A multiple linear regression (MLR) is used to compute dCO₂/dt as a function of E_TOT, ENSO indices, and stratospheric aerosol optical depth (SAOD, a proxy for major volcanic eruptions). Output from the MLR model is used to adjust dCO₂/dt for natural variability due to ENSO and major volcanic eruptions. We calculate trends in AF using the unadjusted (AF^RAW) and adjusted CO₂ growth rate (AF^ADJ), from 1959-2021, for 21 published estimates of E_LUC. We find that for 20 out of 21 estimates of E_LUC, the trend in AF^ADJ is close to zero and lacks statistical significance at the 95% confidence interval. Therefore, it is most likely that the combined relative efficacy of the terrestrial biosphere and oceanic carbon sinks has remained fairly constant on a global scale over the past six decades.