Regional carbon fluxes in China during recent 10 years constrained by surface CO2 observations

In this study, we use a global carbon assimilation system (GCAS) to inverse the net terrestrial carbon fluxes in China during 2006 - 2015, which was constructed based on the global atmospheric chemistry model of MOZART4 and the local ensemble transform kalman filter (LETKF). Firstly, the system is improved with higher resolution of atmospheric transport model (global 1deg by 1deg), finer global zoning. Carbon exchanges on ocean surface from OIF and terrestrial ecosystem carbon exchanges simulated using BEPS are used as priori fluxes. Global annual uncertainties of 2.0 PgC/yr and 0.80 PgC/yr for ocean and bio fluxes are used, respectively. Fire emissions from GFED4, and fossil fuel carbon emissions from CDIAC, EDGAR and PKU-CO2 are adopted as fixed fluxes. In addition, in order to further improve the assimilations, surface CO2 observations from Six additional Chinese sites (i.e., Shangdianzi, Longfengshan, Lin'An and Shangri-La, Jinsha and Akdala) are used, which are not included in the GlobalView CO2 package (obspack). Then, four experiments are conducted, namely 1) only surface observations (obspack, same thereafter) and with CDIAC emissions, 2) both obspack and additional Chinese observations, and with CDIAC emissions, 3), but using EDGAR emissions, 4) based on 2), but using PKU-CO2 emissions. Finally, the impacts of six additional Chinese observations and different fossil fuel carbon emission datasets on the inverted net terrestrial carbon fluxes, and the distribution of terrestrial carbon exchanges as well as their inter-annual variations in China during the study period are shown and discussed.