A Numerical Tool Kit for Estimating Surface Fluxes from Space-borne Observations: Application to GOSAT XCO2 Retrievals
Feng, Liang1; Palmer, Paul1; Boesch, Hartmut2; Parker, Robert2
1University of Edinburgh, UNITED KINGDOM; 2University of Leicester, UNITED KINGDOM
Within the ESA Data Assimilation Project, we have developed a numerical tool kit for supporting Observation System Simulation Experiments (OSSE) to evaluate impacts of satellite observations of atmospheric composition on estimating their corresponding surface fluxes. The tool kit, written in Python, include modules to 1) simulate observations specified by the user by sampling 3-D output from a user-provided Chemistry Transport Model (CTM), filtering possible contamination from cloud and heavy aerosol loadings, and weighting the model profiles with specified instrument vertical sensitivity for different observation scenes; and 2) estimate surface fluxes at user-defined temporal and spatial resolutions from simulated observations by using an Ensemble Kalman Filter (EnKF). The tool kit is easily used to estimate surface fluxes from real data.
In this presentation we report posterior regional CO2 surface fluxes inferred from GOSAT XCO2 retrievals using the developed numerical tool kit. We generally find agreement with fluxes estimated from in situ observations at continental scales. As expected we find discrepancies between fluxes inferred from the in situ and space-borne data over some continental regions, particular over the tropics and the southern hemisphere, largely reflecting the sparseness of the in situ data. To gain more insight into these results, we also report results from several OSSEs that study the impacts of scene-dependent biases and inhomogeneous coverage of GOSAT XCO2 retrievals on the posteriori CO2 flux estimates.