Regional Inverse Modelling for the Estimation of CO2 Fluxes in Australia using Satellite Observations from The Orbiting Carbon Observatory-2 (OCO-2)

Climate change projections depend mainly on the amount of carbon dioxide in the atmosphere. The ocean and terrestrial biosphere play important roles in the climate, since they help to balance the accelerating emissions caused by human activities. Therefore, understanding and quantifying the carbon cycle fluxes is relevant for understanding likely future climate change. Despite the efforts of terrestrial and ocean models to simulate CO₂ fluxes over land and ocean, there remain large quantitative uncertainties, especially at regional scales. Inverse modelling based on atmospheric CO₂ observations has proved capable of improving the accuracy of CO₂ fluxes estimated by either biosphere or ocean models over areas with high-density data. Satellite-derived observations offer an alternative means to constrain CO₂ surface fluxes in places lacking ground-based measurements. In this study, we use regional inverse modelling to quantify CO₂ surface fluxes over Australia using satellite data retrieved from the Orbiting Carbon Observatory-2 (OCO-2). The technique uses a Bayesian variational assimilation technique (4DVAR) employing the Community Multi-scale Air Quality Model (CMAQ) and its adjoint. We evaluate the information content of OCO-2 measurements for constraining monthly fluxes using data for 2015. We use a Monte-Carlo approach to estimate the posterior uncertainties of the Australian fluxes. A preliminary evaluation of the information content suggests that OCO-2 data can reduce uncertainties by up to 78% over vegetated areas. However, this percentage of reduction is minor in most parts of Australia, mainly on those regions with less vegetation and productivity.