Estimation of Flux Biases by Coupled MLEF-PCTM Model Using CO2 Observations

We used an ensemble based data assimilation system, Maximum Likelihood Ensemble Filter (MLEF), which has been coupled with Parameterized Chemistry Transport Model (PCTM), to assimilate CO2 observations. A pseudo data experiment showed that the coupled system is capable of recovering fluxes adequately in well- observed northern hemispheric regions. We assimilated 2 years of CO2 observations from 49 flask sites and 15 continuous sites to estimate separate multiplicative biases in photosynthesis, respiration, and air-sea gas exchange. Spatial covariance in the biases was updated in the assimilation process. Flux bias estimates will be presented along with their uncertainties. It is assumed that the biases have 8-week time span and they are constant throughout that time period. Results show good constraint on model biases over North America and Europe. Some other regions (e.g. the South Atlantic, tropical contents) are unconstrained.