基于模糊逻辑的分步式超折射地物回波识别方法的建立和效果分析

DEVELOPMENT OF FUZZY-LOGICAL TWO-STEP GROUND CLUTTER DETECTION ALGORITHM

超折射地物是影响雷达定量探测降水和其他产品的重要因素,文中使用合肥、广州、温州、天津的SA多普勒雷达和上海的WSR-88D资料分析了混合性暴雨降水、对流性降水、地物回波的回波强度、径向速度、速度谱宽的取值范围、水平垂直变化及其地物回波与地形海拔高度的关系等特征,并确定了模糊逻辑识别超折射地物的隶属函数。在Kessinger方法基础上,考虑到雷达径向速度的距离模糊问题和地物回波与方位的关系,提出了基于模糊逻辑的分步式地物回波识别方法,通过调整已被严格标准识别为地物和非地物的邻近区域回波点的判据,来减小地物的漏判和降水回波的误判。利用降水回波、地物回波的个例资料分析了回波强度、径向速度和速度谱宽资料在识别地物回波中的贡献,分析了分步方法对地物回波识别效果的改善。结果表明:该方法能较好识别地物回波和降水回波,径向速度和速度谱宽资料在地物识别中作用明显,利用分步方法明显改善了在速度模糊区暴雨过程对流云团被误判为地物的概率。

Radar echo from AP clutter is an unpredictable event which often contaminates the precipitation measurement and other radar products. The radar data observed by SA Doppler radars in Hefei, Guangzhou, Wenzhou, Tianjin and the WSR-88D radar in Shanghai are used to develop and examine the AP detection algorithm. The reflectivity, radial velocity, spectrum width of velocity, and their spatial variations for ground clutter and different precipitations are examined to determine the membership functions in fuzzy-logical based AP detection algorithm. The relationships between AP echo and ground surface altitude are analyzed. The two-step approach for ground clutter detection is developed, the first step detection is based on the method developed by Kessinger, and thresholds in the second step are adjusted in the vicinity areas of the distinguished clutter and precipitation points detected at the first step with strict standards, considering the range folding of velocity data and clutter distributions in radial direction. The detection results and contributions of reflectivity, radial velocity and spectrum width data to ground clutter detection are analyzed by using the radar data for AP cases and precipitation cases. The analyses results show that most of AP clutters are detected by the algorithm. The radial velocity and spectrum width data are key factors for AP clutter detection. In comparison with the one step detection approach, the two-step approach has obvious reduced the possibility of the convective cloud clusters in heavy rain processes misidentified as ground clutters.