摘要
从识别、修订闪电定位监测数据异常值、提高闪电定位监测产品准确度的目的出发,以闪电定位栅格密度图为研究对象,结合人工观测雷电日数据及国内外对小电流幅值闪电数据剔除的研究成果,引入泰森多边形、数据分布假设检验、异常值(五数概括)处理算法等模型,建立了可用于判别和优化闪电定位监测数据栅格密度异常值的方法,以福建省为例,对该方法进行了实例应用。结果表明,在福建省西、北部区域,闪电定位监测数据存在一定探测误差或数据丢失的情况,并使用人工观测雷电日数据对这些异常值所在的区域进行了修订,该方法适用于闪电定位仪覆盖数量较少区域的闪电定位监测数据异常值识别与优化应用。
With the purpose of distinguishing and revising the abnormal values of lightning location data and improving the accuracy of lightning location products,this study uses such methods as the Tyson polygon,the hypothesis test of data distribution,abnormal values（with the Five-Number Summary）to establish a technique that associates with the manual observation of lightning days to distinguish the abnormal grid data of lightning density location on the lightning positioning gird density map.The application case of Fujian Province is discussed.The results show that there are certain detection errors or data loss of lightning location data in the northern and western parts of Fujian Province;the correction is conducted in the areas with unusual values by the artificial observation of lightning data;the method can be used in the areas with less lightning location instruments to distinguishes and optimizes the abnormal values of lightning location data.
出处
《气象科技》
北大核心
2017年第6期1011-1016,共6页
Meteorological Science and Technology
基金
福建省气象局青年科技专项项目(2017Q13)资助
关键词
闪电定位
人工观测雷电日
异常值处理
泰森多边形
lighting positioning
lightning observation
outlier processing algorithm
Thiessen polygon
