中文摘要：

速度退模糊算法对新一代天气雷达后期数据同化及短时临近预报具有重要影响。针对在高切变和孤立回波情况下速度退模糊出现错误的问题,提出了基于联合切变检测和十字形多尺度窗口的速度退模糊改进算法。首先,采用MDV（mininum distance velocity,MDV）数据预处理方法进行全局数据优化。其次,将当前速度库和参考库的差值与切变阈值相比较,根据对比结果,分别进行切向和径向的退模糊。然后,在远离雷达的孤立回波区域,建立一种十字形多尺度窗口,实现参考库的确定,进而完成全方位的速度退模糊。使用我国新一代天气雷达（CINRAD-SA）在极端天气情况下的实测数据对该算法进行了多次验证。结果表明,在高切变和孤立回波情况下,该算法的速度退模糊准确率较高,优于新一代天气雷达中运行的WSR-88D速度退模糊算法。

英文摘要：

Velocity dealiasing algorithm has an important influence on later data assimilation and short-term weather forecast for China new generation weather radar. An improved algorithm based on combined shear detection and across-shaped multi-scale window is proposed, solving false velocity dealiasing problems in isolated echo region and high shear region. Firstly, the new data preprocessing method ″MDV″（ minimum distance velocity） is used to optimize the data. Secondly, compare the shear threshold with the difference value that between the radial velocity value and the reference velocity value, according to the comparison result, velocity dealiasing starts in the azimuth direction and then in the radial direction. Then, isolated echoes away from the radar use the reference velocity value found from a cross-shaped multi-scale window. The velocity dealiasing algorithm is tested many times by the extreme weather radar radial velocity of S-band China new generation weather radar. The result shows that, the new algorithm′s accuracy is better than WSR-88D′s, in case with high shear and isolated echo.