Statistics Characteristics Of Convective Radar Echo In Guiyang Area And Assimilation Experiment

Severe convective weather often accompany with severe weather,such as strong winds,hail and rain.It always bring huge disaster to people's life and threat property security.Therefore,convectional weather research has been attached great importance to meteorological researchers.Conventional observation methods are very difficult to perform convective process,due to its short-term and paroxysmal characteristics,while Doppler radar can provide high spatial resolution reflectivity factors and radial wind observations.Based on this consideration,this paper using radar data in Guiyang of Guizhou Province between May and August from 2014 to 2016 to statistical analyze convectional radar echo morphological characteristics,occurrence frequency,location,diurnal variation,duration in guiyang region;then using ARW WRF model and WRF 3DVAR assimilation system,a short-term local convection process in the middle-west of guizhou province on July 24,2014 has done some assimilation experiments,and further analyzes the convection cell the internal dynamic structure and thermal structure.Finally,using different physical parameterization schemes and WRF 3DVAR Random CV perturbation solution to generate 20 ensemble samples,and researching the improvement of hybrid assimilation method by adding flow pattern information in the 3DVAR static covariance for analysis and prediction.The statistical results of radar observations shows that the convectional precipitation echo could be classified into cell?linear and areal type.Among them,the linear type is about 50% and is the most common convection organization in guiyang area;the convective precipitation often occurs in the southwest and southeast of guiyang and rarely in the north,and the night-time precipitation is easy to occur in the southwest of guiyang;the convective initiation time has two peaks in the afternoon and around midnight,respectively.The assimilation test showed that,compared with the spin up 6 hours,the prediction results of the radar assimilation of spin up 12 hours were closer to the actual conditions.During the one-hour cycle assimilation,the predicted results of per 6 minutes assimilation were significantly better than per 30 minutes and per 60 minutes.The variance scaling parameters(var scaling)and the length scaling parameter(len scaling)of background field could affect the assimilation results,and the optimal prediction effect could be obtained through different combination tests of two parameters.Using different physical parameterization schemes and WRF 3DVAR Random CV perturbation solution to generate 20 initial ensemble samples,and then using the 3D-EnVar hybrid assimilation method within an hour time window to assimilate radar data,it shows that the hybrid method of adjustment of the initial field and the subsequent prediction is obviously better than that of 3DVAR method,result of physical parameters assimilation was better than Random CV perturbation scheme.