Preliminary Study of the Joint Carbon Data Assimilation System (JDAS)

Accurate quantification of the variability and distribution of atmospheric CO2 is crucial for understanding the global carbon cycle. Data assimilation, which combining observation and modeling, is a state-of-art method for CO2 inversion. Great achievements, represented by CarbonTracker, have been made by previous researches. In this study, we developed a novel joint carbon data assimilation system (JDAS), which combined the local ensemble transform Kalman filter (LETKF), one type of EnKF, with the atmosphere transport model (GEOS-Chem) and dynamic vegetation model (VEGAS) for a better estimate of both global CO2 concentration and land-atmosphere carbon flux. Our preliminary study showed JDAS can better obtain the signal of CO2 concentration variability, comparing to only using models. The next step is to compare the differences with the satellite observations (TanSat、GMI-GF5、 OCO-2 and GoSat) and site observations (Globalview-CO2 and TCCON). Further, we would consider taking these observations into JDAS, to obtain more accurate simulation results.