A Nowcasting Technique Based on Application of the Particle Filter Blending Algorithm

查看全文 作　　者：yuanzhao [1,2]chen;hongping [2]lan;xunlai [1,2]chen;wenhai [3]zhang

机构地区：[1]meteorological bureau of shenzhen municipality,Shenzhen 518040;[2]shenzhen key laboratory of severe weather in south china,Shenzhen 518040;[3]shenzhen academy of severe storm science,Shenzhen 518040

出　　处：《Journal of Meteorological Research》索引2017年第31卷第5期,共15页

基　　金：Supported by the China Meteorological Administration Research Fund for Core Operational Forecasting Technique Development;Shenzhen Science and Technology Project(JCYJ20160422090117011 and ZDSYS20140715153957030);Guangdong Meteorological Bureau Science and Technology Project(GRMC-2016-04) 摘　　要：To improve the accuracy of nowcasting, a new extrapolation technique called particle filter blending was configured in this study and applied to experimental nowcasting. Radar echo extrapolation was performed by using the radar mosaic at an altitude of 2.5 km obtained from the radar images of 12 S-band radars in Guangdong Province,China. The first bilateral filter was applied in the quality control of the radar data; an optical flow method based on the Lucas–Kanade algorithm and the Harris corner detection algorithm were used to track radar echoes and retrieve the echo motion vectors; then, the motion vectors were blended with the particle filter blending algorithm to estimate the optimal motion vector of the true echo motions; finally, semi-Lagrangian extrapolation was used for radar echo extrapolation based on the obtained motion vector field. A comparative study of the extrapolated forecasts of four precipitation events in 2016 in Guangdong was conducted. The results indicate that the particle filter blending algorithm could realistically reproduce the spatial pattern, echo intensity, and echo location at 30-and 60-min forecast lead times. The forecasts agreed well with observations, and the results were of operational significance. Quantitative evaluation of the forecasts indicates that the particle filter blending algorithm performed better than the cross-correlation method and the optical flow method. Therefore, the particle filter blending method is proved to be superior to the traditional forecasting methods and it can be used to enhance the ability of nowcasting in operational weather forecasts. 关键词：radar echo particle filter blending bilateral filter semi-Lagrangian extrapolation NOWCASTING

相关文献

参考文献（36）
引证文献（10）
耦合文献（995）

参考文献(36)

引证文献(10)

耦合文献(995)

互联网出版许可证新出网证(渝)字10号全国400电话 - 免长途话费

A Nowcasting Technique Based on Application of the Particle Filter Blending Algorithm

参考文献(36)

引证文献(10)

耦合文献(995)

用户提示