Balance Characteristics Of Multivariate Background Error Covariance And Its Impact On Assimilation And Forecasts

Building a reasonable background error covariance is the key to data assimilation, and the characteristics of background error covariance has a great relationship with local meteorological feature. Based on the balance relationships of the background error covariance manifesting the dynamics characteristics in assimilation system, the study does research into the different statistical characteristics of the multivariate background error covariance in WRFDA system for rainy and dry seasons, typhoon and non-typhoon season. Besides, the impact of the correlations in background error covariance on the analysis and forecasting of rainfall and typhoon is also discussed. The main conclusions are as follows:(1) The characteristics of balanced part contributions in rainy and dry seasons show that the vorticity and divergence field of wind play a more important role in the balance relationships between wind and mass fields, and the humidity related correlations between pseudo relative humidity and all other control variables are stronger in rainy season. The characteristics of balanced part contributions in typhoon and non-typhoon seasons show that the unbalanced temperature plays a dominant role in individual contributions to relative humidity and the correlations of relative humidity and other control variables is larger in typhoon season.(2) The results and diagnostics for the cycling data assimilation experiments of two rainfall events indicate that the unbalanced velocity potential related correlations improve the analysis and forecast of wind field but have no significant impact on the temperature and specific humidity field simulation, and improve the precipitation forecast in a certain extent. In contrast, with the background error covariance that includes the correlations between relative humidity and all other control variables, the wind, temperature and surface pressure observations will have impact on moisture fields, and vice versa. So, the humidity related correlations make the analysis and forecast closer to the observations and obviously improves the precipitation forecast.(3) The cycle data assimilation and forecast results for typhoon show that the new correlations builded in multivariate background error covariance play an important role in the analysis and forecast of typhoon, and especially the correlations between relative humidity and other control variables obviously improve the analysis and forecast of typhoon track, intensity and precipitation.