Application of the Multi-Source Data Fusion Algorithm in the Hail Identification

摘要

In this study, the canonical correlation analysis algorithm (CCA) is used to fuse the two-dimensional wind field retrieved from the single-Doppler weather radar, the three-dimensional wind field retrieved from the dual-Doppler weather radars, the observations from the ground automatic weather stations and the meteorological reanalysis data in three hail episodes ("0625" episode in Beijing, "0330" and "0801" episodes in Guangdong). During the hail episode in Beijing on June 25, 2020, an evident and long-lasting three-body scatter spike was observed, which played an important role in the hail identification and warning. In the three-dimensional wind field retrieved from the dual-Doppler weather radars, there is horizontal convergence of northeasterly and northwesterly winds and that of northwesterly and southeasterly winds in the low-level strong echo area, and there are obvious updrafts in the vertical wind field structure. Such a circulation configuration is favorable for the development and maintenance of hail storm. The multi-source data fusion of the wind fields can effectively improve the identification of the low-level convergence. The data fusion for the other two hail episodes ("0330" and "0801" episodes in Guangdong) yields the same conclusion. It is revealed that the dual-radar fusion performs better than the single-radar fusion in the identification of the meso-gamma scale vortices. It can visually illustrate the characteristics of the cyclonic convergent flow fields which is more consistent with the near-surface observation. It can be concluded that the multi-source data fusion technique is practicable in the three severe convection processes.