| With the development of forecasting technology, increasing of unconventional data,improvement of the microphysical processes and parameterized schemes in numerical model,the level of numerical prediction has been greatly improved. However, it is necessary to creat a system of efficient data to provide the precision of initial field for improving the accuracy of numerical prediction. Compared with the conventional observation, the lidar have the advantages of high spatial and temporal resolution, high sensitivity and strong anti-interference.In order to promote the application of Raman lidar in numerical prediction, this study proposes a new method to improve the initial field of the numerical model by using the three-dimensional variational method combined with atmospheric temperature detected by the lidar at Xi'an University of Technology. The main research results and conclusions are presented.(1) Based on the WRF model and FNL data, 6 kinds of microphysical processes and 3 horizontal resolutions are used to study the process of rainfall in Shanxi Province respectively.It is found that the WDM6 scheme and the horizontal resolution of 9km are very suitable for the numerical weather prediction in Shanxi.(2) WRF-3DVAR model has been constructed, and the assimilating experiments have been designed. It is found that the short-term forecast has been improved and the TS score is increased by 0.07 after assimilating the ground data and sounding data.(3) The multiquadric algorithm is used to extend the Raman lidar data and the sounding data.The assimilating experiments of the ground data, sounding data and the Raman lidar data have been finished. The results of assimilating experiments are analyzed by using the increment method. It is found that the initial field is adjusted after assimilation. Compared with TS score before assimilation,the TS score of precipitation is increased by 0.12, and the false precipitation area is reduced. |
PDF Full Text Request