Mapping global land XCO2 from measurements of GOSAT and SCIAMACHY by using kriging interpolation method

In our study, we proposed a gap-filled method based on ordinary kriging to generate a global land distribution map of carbon dioxide (CO₂) by modeling the spatial correlation structures of column-averaged CO₂ dry air mole fractions (XCO₂) on the global scale, using a fused data by combining GOSAT and SCIMACHY. The relationship between the distance and semi-variogram of XCO₂ is estimated and modeled by an exponential model with a nugget-effect component. The semi-variogram result indicates that there is a significant spatial correlation within the fused CO₂ data set. The prediction of XCO₂ using semi-variogram model is conducted within 1 degree×1 degree grids over the world. The results reveal that the global distribution of XCO₂ based on kriging method is the most extensive compared with other CO₂ products. Moreover, the monthly map from the kriging approach has less predicted uncertainties, most of which are less than 0.5% of XCO₂ value.