基于多普勒雷达数据的强辐合场识别方法

Recognition of Strong Convergence Field Based on Doppler Radar Data

为了将对雷雨大风具有较强指示性的中层径向辐合(MARC)自动识别出来,利用"径向矩形映射"将放射状分布的雷达数据转变为格点数据;同时完成窗映射和等腰三角形检测模板的映射,从而在快速定位辐合点的基础上得到描述辐合点辐合强度及位置的参向量;以单仰角最强辐合点确定剖线,获得显示MARC的较佳剖面,得到MARC的深度信息.提出从辐合点生成辐合带,再借辐合带走向订正辐合强度的思想和算法,有效克服单仰角上MARC的低估甚至漏检的问题.测试结果表明,融入订正策略的从辐合点到辐合线再到辐合场的识别方法和量化参量,使通过检出的MARC对实际伴有强辐合现象的雷雨大风的报准率达到94.87%,,平均提前量为20.6,min,空报率低于6%,.

In order to identify the mid-altitude radial convergence （MARC） which is a good indicator of the surface damaging winds, a ＂radial rectangular map＂ is proposed to transform the data from radar coordinate data into a new coordinate system, based on which the convergence zone can be quickly recognized by the regional segmentation algorithm. Then, the specific template is designed to calculate eigenvector of each point of the convergence line. In this way we can get the strong convergence field in the radar image at each single elevation angle. After the matching of strong convergent regions at different elevation angles, we can determine the position of the section and locate the MARC by the parameter vectors of the significant convergence points. The correction method of convergence speed difference based on the trend of convergence lines is used to overcome the problem of underestimation or even missed detection of strong convergence. Experimental results show that the combination of MARC automatic identification and correction algorithm makes the accurate rate of surface damaging winds prediction reach 94.87% with an average lead-time of 20.6 min. And the rate of vacancy forecast is less than 6%.