This study compares CO₂ fluxes estimated using atmospheric and oceanic inversions, and examines the role of fluxes and their interannual variability in simulating atmospheric CO₂ concentrations. Using a time-dependent inverse (TDI) model, regional sources and sinks of atmospheric CO₂ are estimated from observations at 87 stations. These results are compared with fluxes from an ocean inversion that uses ocean interior observations of dissolved inorganic carbon and an ocean general circulation model. The study finds that flux estimates from the southern hemisphere are generally in good agreement with the ocean inversion results, providing confidence in the accuracy of the atmospheric inversion estimates. A forward tracer transport model (TTM) is used to simulate observed CO₂ concentrations using estimates of fossil fuel emissions and a priori estimates of terrestrial and oceanic CO₂ fluxes, as well as two sets of TDI model corrected fluxes. The TTM simulations of TDI model corrected fluxes show improvements in fitting the observed interannual variability in growth rates and seasonal cycles of atmospheric CO₂. The analysis suggests that the use of interannually varying meteorology and a larger observational network has helped to capture the regional representation and interannual variability in CO₂ fluxes realistically. The study also examines the effect of using interannually varying meteorology in the TTM on flux variability for specific regions and identifies areas most sensitive to changes in meteorology due to dominant climate oscillations. The TDI model corrected fluxes are used for TTM simulations, and quantitative estimates of the fit between model simulations and observations are made for year-to-year variations in seasonal cycles and growth rates of atmospheric CO₂. The results suggest that the TDI model fluxes are valid for simulating atmospheric CO₂ concentrations. The study concludes that the TDI model fluxes based on atmospheric CO₂ are overall valid and that the use of TDI model fluxes in transport model simulations provides a better match with observations compared to prior knowledge of CO₂ fluxes. The study also highlights the importance of using interannually varying meteorology and a larger observational network in capturing the interannual variability in CO₂ fluxes.This study compares CO₂ fluxes estimated using atmospheric and oceanic inversions, and examines the role of fluxes and their interannual variability in simulating atmospheric CO₂ concentrations. Using a time-dependent inverse (TDI) model, regional sources and sinks of atmospheric CO₂ are estimated from observations at 87 stations. These results are compared with fluxes from an ocean inversion that uses ocean interior observations of dissolved inorganic carbon and an ocean general circulation model. The study finds that flux estimates from the southern hemisphere are generally in good agreement with the ocean inversion results, providing confidence in the accuracy of the atmospheric inversion estimates. A forward tracer transport model (TTM) is used to simulate observed CO₂ concentrations using estimates of fossil fuel emissions and a priori estimates of terrestrial and oceanic CO₂ fluxes, as well as two sets of TDI model corrected fluxes. The TTM simulations of TDI model corrected fluxes show improvements in fitting the observed interannual variability in growth rates and seasonal cycles of atmospheric CO₂. The analysis suggests that the use of interannually varying meteorology and a larger observational network has helped to capture the regional representation and interannual variability in CO₂ fluxes realistically. The study also examines the effect of using interannually varying meteorology in the TTM on flux variability for specific regions and identifies areas most sensitive to changes in meteorology due to dominant climate oscillations. The TDI model corrected fluxes are used for TTM simulations, and quantitative estimates of the fit between model simulations and observations are made for year-to-year variations in seasonal cycles and growth rates of atmospheric CO₂. The results suggest that the TDI model fluxes are valid for simulating atmospheric CO₂ concentrations. The study concludes that the TDI model fluxes based on atmospheric CO₂ are overall valid and that the use of TDI model fluxes in transport model simulations provides a better match with observations compared to prior knowledge of CO₂ fluxes. The study also highlights the importance of using interannually varying meteorology and a larger observational network in capturing the interannual variability in CO₂ fluxes.