Comparision of GOSAT XCO2 with CO2 Column from PCTM Modeling

The ACOS XCO2 algorithm for GOSAT column averaged CO2 retrievals has evolved to version 3.5, which is an improvement on precision and accuracy from previous releases. As a global CO2 transport model, the Parameterized Chemistry and Transport Model (PCTM) has proven its skill in simulation of spatial and temporal variability of CO2 concentrations through a long series of experiments. The PCTM model run in this study uses NASA GMAO's MERRA reanalysis as meteorological forcing, CASA-GFED2 CO2 bio-terrestrial model, fossil fuel inventory fluxes, and observation-based ocean CO2 flux as input. A relatively comprehensive comparison of ACOS retrieved XCO2 with PCTM output XCO2 is conducted, which includes comparisons using column pressure uniformly averaged CO2 and using the same averaging kernel and pressure weighing coefficients used in ACOS XCO2 v3.5 L2 products. Their differences are presented and possible causes are discussed. The uncertainties in GOSAT XCO2 data are included in the analysis of the differences between ACOS XCO2 and its PCTM counterpart. We expect evaluation of the results can improve our understanding of both XCO2 retrieved from space observation like GOSAT and modeling simulations like PCTM, and the processes that affect their variability. PCTM comparisons with ground sites like TCCON and in situ tower data are presented too, along with profile measurements of the HIPPO mission. For most of the ground-based sites, no matter column averaged CO2 or CO2 concentrations at some pressure/height levels near earth surface, PCTM shows quite consistent good agreement with the site data. In the vertical profile PCTM also presents a reasonable coincidence with observations at most locations.