Research On The Characteristics And Influence Of The Diurnal Variation Of The Background Field Error Covariance In The Cyclic Assimilation System

Most of the current variational-based rapid update and cycling(RUC)assimilation systems often use fixed background error covariance at each analysis time.In order to analyze the diurnal variation characteristics of forecast error in RUC,and actually applied to RUC to improve the fixed background error covariance in the original RUC system.This paper is based on the forecast samples obtained from Rapid-refresh Multi-scale Analysis and Prediction System-Short Term(RMAPS-ST)developed by Institute of Urban Meteorology in Beijing.The diurnal variation background error covariances in summer and winter are get with NMC method,and the structural characteristics of diurnal variation background error covariances and their influences in RMAPS-ST system are analyzed in detail.The main conclusions obtained are as follows:(1)The standard deviations of the forecast errors in both summer and winter show obvious diurnal variation characteristics.The standard deviation of the errors at night(12UTC-21UTC)is larger than that of the errors at daytime(00UTC-09UTC).This reflects that the forecast errors of the RMAPS-ST system are larger at night.The reason may be due to the fact that more observations are available during daytime,which makes the daytime forecast quality relatively better;meanwhile,the calculation method of NMC makes the two forecast fields of the nighttime computation error samples obtained in different cold start cycle forecasts,which to some extent also causes the greater difference between the two forecast samples at night.At the same time,the standard deviation of the errors of the summer forecast samples is larger than that of the winter results,indicating that the forecast errors of the RMAPS-ST system are larger in summer,which may be due to the fact that the weather is more intense and complex in summer compared to winter,and therefore the forecast errors of the model system are larger in summer.(2)Each characteristic statistic in the background field error covariance also exhibits obvious diurnal variation characteristics.The vertical position of the large value area of the first eigenvector has a good correspondence with the position of the large value area of the standard deviation of the forecast error samples.The variation of the eigenvalues shows that the nighttime(12UTC-21UTC)results are larger than the daytime(00UTC-09UTC)results,which is consistent with the variation of the standard deviation of the error samples.It indicates that the vertical structure of the background field errors and the size variation correspond in physical space and EOF space.The length scales also show obvious diurnal variation characteristics,and the smallest length scale basically appears at 09 UTC,and the length scales of the temperature and humidity fields are generally smaller than the results of the wind field.(3)From the analytical increments of the single-point observation assimilation ideal test,it can be seen that there are obvious diurnal variation characteristics of the analytical increments and the horizontal distribution range,and these characteristics correspond well with the diurnal variation characteristics of each characteristic statistic.(4)The results of both the individual assimilation test and the continuous one-month cycling assimilation test indicate that the use of diurnal variation background error covariances can improve the assimilation and forecasting capability of the RMAPS-ST system for the wind,temperature,and humidity fields at ground level and aloft,and thus improve the performance of precipitation forecasting.