The Research On The Adaptive Coupling Coefficient In En-3DVAR Hybrid Data Assimilation

The hybrid data assimilation, combining the variational data assimilation framework and a set of ensemble perturbations which can represent the background error covariance with flow-dependent characteristics, has the effect of improving the quality of data assimilation analysis and is the current trend of the numerical weather forecast and data assimilation. At present, a lot of research work has been done on the hybrid data assimilation, and achieved good research and application effects.However, there are still some problems about the coupling of static background error covariance and the ensemble forecast error covariance with flow-dependent characteristics that need further study. Based on WRF model, the work for the key scientific problems of the coupling of the fixed background error covariance and ensemble forecast error covariance in the En-3DVAR hybrid data assimilation system,carried out in this paper is as follows: flow-dependent characteristics of the numerical prediction error, optimal coupling coefficient and its influence factor, and the construction of the adaptive optimal coupling coefficient. At last, this paper obtains reliable research conclusions:(1) Different weather conditions, different weather systems have their own different flow-dependent characteristics; Different flow-dependent features usually correspond to the different coupling coefficients of the ensemble forecast error covariance and static background error covariance.(2) In the hybrid data assimilation, different weather systems have the most significant effect coupling coefficient for the improvement of the analysis and prediction quality, that is, the optimal coupling coefficient, and the optimal coupling coefficient varies with different weather systems and different times; Ensemble prediction quality, physical parameterization schemes and so on have the important influence on the optimal coupling coefficient.(3) The concept of adaptive coupling coefficient is proposed. The adaptive coupling function based on the energy can reasonably and effectively identify the optimal coupling coefficient of the hybrid data assimilation, thus realizing the most reasonable coupling of the prediction error covariance and the fixed background error covariance, fully playing the potential advantages of hybrid data assimilation,and laying the scientific basis for further improving the quality of analysis.