Adjoint Inverse Modeling of Black Carbon During ACE-Asia

An adjoint model is used for inverse modeling of black carbon during ACEAsia. We use the 4D-Var approach to optimally recover spatially-resolved anthropogenic and biomass burning emissions, and initial and boundary conditions of black carbon. Boundary conditions and biomass burning emissions are assigned daily scaling factors. Anthropogenic emissions are scaled by a combination of daily and monthly scaling factors. Simulation results are compared to various observations of black carbon concentrations during the campaign. Measurements at five islands and onboard the research vessel Ron Brown are used for data assimilation. Different levels of constraints are examined for data assimilation, and a case with 65% reduction in the total square errors is chosen. The assimilation is more successful at the islands of Sado and Gosan and for the observations onboard R/V Ron Brown. The assimilated results show significantly better agreement with the observations onboard the Twin Otter aircraft that were not used for assimilation. Among the scaled variables, anthropogenic emissions are the most significant, followed by the boundary conditions. Domain-wide emissions inventory does not change significantly as a result of the scaling, but sizable changes are seen on the subregional level. Most noticeably, anthropogenic emissions over southeastern China are reduced while those in northeast China and Japan are increased.