Inverse Modeling of CO₂ Fluxes Using GOSAT Data and Multi-Year Ground-Based Observations

Abstract

We present surface CO₂ flux estimates obtained by an inverse modeling analysis from column-averaged dry air mole fractions of CO₂ (XCO₂) observed by the Greenhouse gases Observing SATellite (GOSAT) and ground-based data. Two inversion cases were examined: 1) a decadal inversion using ground-based CO₂ observations by NOAA from 1999 to 2010 to derive CO₂ flux interannual variability, and 2) an inversion using NOAA plus NIES GOSAT XCO₂ data from June 2009 to October 2010. We used single-shot GOSAT data and individual NOAA flask data for the inversions. Our results show differences in estimated fluxes between the NOAA data inversion and the NOAA plus GOSAT data inversion, especially in Northern Eurasia and in Equatorial Africa and America where the ground-based observational sites were sparse. Uncertainty reduction rates of 40%-70% were achieved by inclusion of GOSAT data, compared to the case using just the NOAA data. The inclusion of GOSAT data in the inversion resulted in larger summer sinks in northwest Boreal Eurasia and a smaller summer sink in southeast Boreal Eurasia, with a clear uncertainty reduction in both regions. Adding GOSAT data also led to increase in Tropical African fluxes in boreal winter beyond interannual variability from NOAA data inversions.