Sensitivity of simulated CO2 concentration to sub-annual variations in fossil fuel CO2 emissions

This study presents a sensitivity analysis of the impact of sub-annual fossil fuel CO2 emissions on simulated CO2 concentration using a global tracer transport model. Four sensitivity experiments were conducted to investigate the impact of three cyclic components (diurnal, weekly and monthly) and a complete cyclic component (the combination of the three) by comparing with a temporally "flat" fossil fuel CO2 emissions inventory. A complete exploration of these impacts is quantified at annual, seasonal, weekly and diurnal time scales of the CO2concentration for the surface, vertical profile and column-integral structure. Result shows an annual mean surface concentration difference varying from -1.35 ppm to 0.13 ppm at grid scale for the complete cyclic fossil fuel emissions, which is mainly driven by a large negative diurnal rectification and less positive seasonal rectification. The negative diurnal rectification is up to 1.45 ppm at grid scale and primarily due to the covariation of diurnal fossil fuel CO2 emissions and diurnal variations of vertical mixing. The positive seasonal rectification is up to 0.23 ppm at grid scale which is mainly driven by the monthly fossil fuel CO2emissions coupling with atmospheric transport. Both the diurnal and seasonal rectifier effects are indicated at local-to-regional scales with center at large source regions and extend to neighboring regions in mainly Northern Hemisphere. The diurnal fossil fuel CO2 emissions is found to significantly affect the simulated diurnal CO2 amplitude (up to 9.12 ppm at grid scale), which is primarily contributed by the minima concentration differences around local sunset time. Similarly, large impact on the seasonal CO2 amplitude (up to 6.11 ppm) is found at regional scale for the monthly fossil fuel emissions. An impact of diurnal fossil fuel CO2 emissions on simulated afternoon CO2 concentration is also identified by up to 1.13 ppm at local scales. The study demonstrates a large cyclic fossil fuel CO2 emissions impact on simulated CO2 concentration, and suggests that attempts to constrain carbon fluxes using surface measurements and inversion models can be improved by avoiding the biases caused by considering the temporal variations in fossil fuel CO2 emissions.