中尺度涡旋客观识别与三维追踪的新方法及其效果评估

A New Objective Identification Method for Mesoscale Vortices:Three-Dimensional Tracking and Quantitative Evaluation

中尺度涡旋是引发强降水等一系列气象灾害的重要天气系统之一,中尺度涡旋识别是对其进行研究的重要基础。目前,如何客观准确地识别中尺度涡旋并对其进行较全面的评估仍是一项充满挑战的任务。本文从我国中尺度涡旋的基本特征出发,提出了一种基于风场和相对涡度场的涡旋识别标准并发展了适用于高分辨率格点数据的中尺度涡旋客观识别算法。该算法能准确识别出中尺度气旋性环流并定位涡旋中心,较现有常规中尺度涡旋识别方法而言,具有误判率低、定位精度高等特点。本文将该客观识别算法应用于长江流域频发的3类中尺度涡旋(高原涡、西南涡、大别山涡)的识别中,结果表明对于不同时间段、不同分辨率的再分析资料(逐6小时的0.5°×0.5°NCEP CFSR再分析资料、逐小时的0.25°×0.25°ERA5再分析资料),本识别算法对3类中尺度涡旋均有较好的识别效果。本文基于1979~2020年共42年暖季(5~9月)大别山涡的数据集(共计36357时次)对新发展的中尺度涡旋客观识别算法进行了定量评估,结果表明本算法能够长期稳定地识别涡旋,42年的平均命中率为95.5%。此外,本文提出了涡旋连续性判定和三维追踪方案,较现有常规中尺度涡旋追踪方法具有显著优势。

Mesoscale vortices(MV)are among the most vital weather systems responsible for precipitation and eteorological disasters in China.However,there is currently no universal standard for MV identification,and the objective identification of MV remains an important problem to be addressed.Herein,according to the main features of MV in China,a new objective identification algorithm(referred to as the new algorithm),suited toward high-precision grid data,is developed by combining wind and vorticity data.The new algorithm can accurately identify mesoscale cyclonic circulation and locate the vortex center with a lower false rate and higher positioning accuracy than existing identification methods.The new algorithm is applied to three types of mesoscale vortices:Plateau vortex,Southwest Vortex,and Dabie Mountain vortex(DBV),which frequently occur along the Yangtze River basin.The results reveal that the new algorithm performs well for all three types of MV and is almost insensitive to the applied period or data resolution(using six-hourly 0.5°×0.5°NCEP CFSR reanalysis data and hourly 0.25°×0.25°ERA5 reanalysis data).Furthermore,the new algorithm is quantitatively evaluated using a DBV activity dataset for the warm season(May–September)from 1979 to 2020.The evaluations demonstrate that the new algorithm can stably identify MV over a long period of time,with an average hit rate of 95.5%.Additionally,this paper proposes a 3D MV tracking scheme,which offers remarkable advantages over traditional tracking methods.