###
气象:2017,43(7):781-791
本文二维码信息
码上扫一扫!
基于模糊逻辑的雷暴大风和非雷暴大风区分方法
周康辉1,2,3,郑永光3,王婷波4,蓝渝3,林建3
(1 中国气象科学研究院,北京 100081 2 中国科学院大学,北京 100049 3 国家气象中心,北京 100081 4 中国气象局气象干部培训学院,北京 100081)
Fuzzy Logic Algorithm of Thunderstorm Gale Identification Using Multisource Data
ZHOU Kanghui1,2,3,ZHENG Yongguang3,WANG Tingbo4,LAN Yu3,LIN Jian3
(1 Chinese Academy of Meteorological Sciences, Beijing 100081 2 University of Chinese Academy of Sciences, Beijing 100049 3 National Meteorological Centre, Beijing 100081 4 China Meteorological Administration Training Centre, Beijing 100081)
摘要
图/表
参考文献
相似文献
本文已被:浏览 456次   下载 270
投稿时间:2016-11-07    修订日期:2017-03-31
中文摘要: 雷暴大风往往伴随飑线、阵风锋、龙卷等强对流天气而出现,风速大、发展迅速、突发性强,对生命财产安全造成极大威胁,因此对雷暴大风的监测与预报具有重要的意义。然而,雷暴大风监测一直也是强对流监测的难点。本文在地面气象观测站大风记录的基础上,结合多源数据(包括雷达、卫星、闪电、温度、露点等观测数据),利用模糊逻辑算法,实现雷暴大风与非雷暴大风的有效识别,可对雷暴大风进行实时监测。具体算法为:首先,基于历史样本数据的统计得到各变量的概率分布函数,进而得到各参数隶属度函数;然后采用概率重叠面积方法,确定各项质量控制数据的权重;最后通过选取判断概率阈值Q的方法,区分雷暴大风与非雷暴大风。通过对2010年全国50873条人工观测大风数据的识别结果检验表明,该算法能有效区分雷暴大风与非雷暴大风,当Q选取0.55时,雷暴大风的识别准确率POD约为0.76,误识别率约为0.18,雷暴大风CSI指数约为0.67。文中选取了两次大风过程,算法正确地识别了11个非雷暴大风记录,5个雷暴大风记录。本工作能一定程度上提升雷暴大风的监测效果、完善强对流监测业务体系。
Abstract:Thunderstorm gale (TG) monitoring in the severe convective weather is a complex and important task, in which the difficulty is how to distinguish the TG from non thunderstorm gale (NTG). Based on the multisource data, including radar, satellite, lightning, temperature and dew point temperature, this paper proposes a fuzzy logic algorithm to tell them apart, which is proved to be an effective and efficient method. First, get the member functions of the multisource data according to their probability distribution which were extracted from long term historical data. Second, acquire the weight ratio of each data by calculating the overlap areas of probability distributions. Finally, get the TG probability Q, and choose a threshold of Q to distinguish TG from NTG. In order to evaluate its performance, the algorithm is used to find TGs in the 50873 gale records of China in 2010. The results show that when Q is 0.55, the POD of TG is 0.76, and the FAR of TG is 0.18, and the CSI of TG is about 0.67. Two mixing weather processes, caused by cold air and typhoon, are chosen to evaluate its performance, showing 11 NTGs and 5 TGs are correctly identified. The algorithm would enhance the accuracy and effectiveness of the severe weather monitoring significantly.
文章编号:     中图分类号:    文献标志码:
基金项目:公益性行业(气象)科研专项(GYHY201406002)、国家重点基础研究发展计划(973计划)(2013CB430106)和国家自然科学基金项目(41375051)共同资助
引用文本:
周康辉,郑永光,王婷波,蓝渝,林建,2017.基于模糊逻辑的雷暴大风和非雷暴大风区分方法[J].气象,43(7):781-791.
ZHOU Kanghui,ZHENG Yongguang,WANG Tingbo,LAN Yu,LIN Jian,2017.Fuzzy Logic Algorithm of Thunderstorm Gale Identification Using Multisource Data[J].METEOROLOGICAL MONTHLY,43(7):781-791.