自适应尺度的雷达回波交叉相关跟踪算法研究

A Study of the Adaptive-Scale Tracking Radar Echoes by Correlation Method

提出了一种能够自动选取跟踪单元大小的雷达回波交叉相关跟踪外推算法(ATREC)。ATREC方法通过雷达回波的边界识别,得到不同尺度的雷达回波的特征运动信息,再不断地内部分化,得到不同次级尺度的内部运动信息,将特征尺度的特征运动场和次级尺度的内部运动场进行合成,最后得到外推的运动矢量场。ATREC方法可以在回波总体运动场基础上,逐级细分出不同次级尺度的移动信息,所以运动场更加平滑且兼顾不同尺度的移动信息,同时能够有效解决TREC类方法采用固定大小的跟踪单元而带来的问题,特别是TREC类方法针对小尺度对流系统的不适用问题。通过个例分析,并与基于交叉相关的多尺度雷达回波跟踪方法(MTREC)进行对比试验,结果表明:针对不同尺度、不同类型的天气系统,ATREC方法具有较强的客观分析能力和灵活的适用能力。通过个例分析检验评分和基于2016年4月总共128个个例的综合统计检验,检验评分结果表明:ATREC方法的预报评分整体优于MTREC方法。

In this study,we propose a new way to obtain motion vectors for extrapolation nowcasting,namely the adaptive-scale tracking radar echoes by correlation(ATREC),which can automatically adjust the size of tracking area according to the scale of convection systems.ATREC method tracks the largescale motion of the radar echo which is identified by the automatically determined convection edges,then the different subscale motions are deduced by tracing the self-split small parts of radar echoes.ATREC method can retrieve the motions of different scales of the convection system and get a smooth motion field for extrapolation.Thereafter,this method can resolve the problems caused by the fixed size of tracking area applied in the traditional tracking radar echoes by correlation(TREC) methods,especially the inapplicability of TREC methods in small-scale convective systems.Nowcasting experiments using the ATREC method and MTREC(multi-scale tracking radar echoes by correlation) method are carried out in this study to demonstrate the practical ability of the ATREC method.Comparison based on a series of weather cases shows that the ATREC method has the abilities of automatically analyzing and flexibly adjusting convective systems with different scales and different types.The evaluation scores of case study and the statistic verifications based on 128 cases in April 2016 indicate that ATREC method has better performance than the MTREC method.