Effect of assimilating CO(2) observations in the Korean Peninsula on the inverse modeling to estimate surface CO(2) flux over Asia

To investigate the impact of two CO(2) observation datasets obtained from the Korean Peninsula on the surface CO(2) flux estimation over Asia, the two datasets are assimilated into the CarbonTracker (CT) inverse modeling system and the estimated surface CO(2) fluxes are analyzed. Anmyeon-do (AMY) and Gosan (GSN) sites in the Korean Peninsula have observed surface CO(2) mole fraction since the late 1990s. To investigate the effect of assimilating the additional Korean observations on the surface CO(2) flux estimation over Asia, two experiments are conducted. The reference experiment (CNTL) only assimilates observations provided by National Oceanic and Atmospheric Administration (NOAA), while the other experiment (EXP1) assimilates both NOAA observations and two Korean observation datasets. The results are analyzed for 9 years from 2003 to 2011 in Asia region because both AMY and GSN datasets exist almost completely for this period. The annual average of estimated biosphere CO(2) flux of EXP1 shows more flux absorption in summer and less flux emission from fall to spring compared to CNTL, mainly on Eurasia Temperate and Eurasia Boreal regions. When comparing model results to independent CO(2) concentration data from surface stations and aircraft, the root mean square error is smaller for EXP1 than CNTL. The EXP1 yields more reduction on uncertainty of estimated biosphere CO(2) flux over Asia, and the observation impact of AMY, GSN sites on flux estimation is approximately 11%, which is greater than other observation sites around the world. Therefore, the two CO(2) observation sets in the Korean Peninsula are useful in reducing uncertainties for regional as well as global scale CO(2) flux estimation.