Application of 4DVAR Data Assimilation Method to Regional Chemical Transport Model

Recently, a four-dimensional variational (4DVAR) data assimilation method begins to be used for chemical transport modeling (CTM) to fuse observations and numerical model, and optimize emissions of various chemical species. In this study, we focus on the new development of 4DVAR data assimilation system for regional chemical transport model. The 4DVAR data assimilation method was built in the fully parallel operational mode of Regional Atmospheric Modeling System (RAMS). Forward CTM mode uses the optional scalar transport system of RAMS. We developed the adjoint model corresponding to the forward CTM. The adjoint model is used to minimize the discrepancies between model result and observation. Application of the 4DVAR assimilation system was conducted to CO transport in Asian region. We set CO emissions for the control parameters to be optimized, and estimate the proper CO emissions by using the numerical model and in situ observation data. The optimally estimated CO emissions after 4DVAR iterations suggest that current emissions were too small. Especially in Shanghai and surrounding areas, optimized emissions become about two times greater than those before optimized. The significant underestimations of East coast of China (near Shanghai) appear to cause mainly the residuals between model-derived and observed CO concentrations, and be important to improve the CTM simulation. This data assimilation system will be extended to apply to transports and emissions of aerosols (e.g. dust aerosols) in Asian region.