A Method to Validate Posterior CO2 Fluxes from Atmospheric Inversion

Abstract:

Validating CO2 fluxes from atmospheric inversion is a challenging problem, since there is no direct flux measurement at comparable spatiotemporal scales. In this study, we present a validation method that projects the difference between posterior and prior CO2 error variances to the spatiotemporal difference between posterior and prior fluxes. We will show theoretically that the posterior fluxes are more accurate than the prior fluxes over the regions that contribute to the changes of CO2 errors when the posterior CO2 is more accurate. We will show the performance of this method using both results from Observing System Simulation Experiments (OSSE) and results from assimilating real ACOS-GOSAT X_CO2 observations. The method is an advancement compared to the current practice of only calculating the CO2 errors, since it shows where and when the posterior fluxes improve the fitting to independent observations and the relative accuracy between the posterior fluxes and the prior fluxes over the regions that affect the CO2 error variance changes. The method can also apply in other trace gas flux inversions, such as methane and CO.