Relative contributions of land carbon fluxes and atmospheric circulation patterns to spatio-temporal variations in atmospheric CO2

To better understand the role of atmospheric CO₂ in the global carbon cycle, it is important to understand the factors that control its variability in time and space. Land carbon fluxes and atmospheric dynamics are crucial for determining the variability in atmospheric CO₂ concentrations; however, their relative contributions to the CO₂ variability are less understood. Here we investigate these factors using a coupled land-atmosphere modeling system (the NASA GEOS model) fitted with land carbon physics and atmospheric CO₂ transport. To separate the influences of land carbon flux variability and multi-year atmospheric transport variability on the variability of atmospheric CO₂ concentrations, we conducted two carbon-fitted GEOS AGCM simulations run in replay mode (a technique that guides the model's weather to match that of the MERRA-2 reanalysis): (i) a control simulation of 15-year carbon cycle dynamics and climate, and (ii) a simulation in which the climatological seasonal cycles of net biome production (NBP), as determined from diagnostics produced in the control experiment, are applied at the land surface to the atmosphere instead. The impact of land flux variability on atmospheric CO₂ variability is then isolated by subtracting the atmospheric transport variability inherent in the second simulation (as induced, e.g., by strong El Nino conditions) from that in the control. These results are also presented in the context of supplemental simulations that have examined the impact of imposed drought in a free running AGCM simulation on land carbon fluxes and atmospheric CO₂ variability.