摘要
为给防雹减灾工作提供必要的参考依据,减少冰雹对农业生产造成的影响,利用上海探空订正软件SANDS分析了1998—2012年5—9月朝阳地区冰雹天气发生和不发生时的13个探空物理量,尝试用Bayes判别分析法和指标叠套法2种方法选出预报因子,并分别建立了冰雹天气潜势预报方程。利用各个预报方程分别回报所有历史样本,结果发现,指标叠套法预报效果较好,因此,指标叠套法最适合用于朝阳地区冰雹天气潜势预报。
In order to provide the necessary reference for hail suppression and disaster reduction, and reduce harm caused by hail, using the data of conventional sounding during May to September from 1998 to 2012, the physical param- eters was studied that may affect the formation of hail in this paper. The hail potential prediction equation was also es- tablished by using the bayes discrimi- nant analysis and the multi indictor su- perposition. It is found out that the pre- diction equation based on multi indictor superposition had a much higher accura- cy than the other. This showed that multi indictor superposition was the best method for hail potential predicting in Chaoyang.
出处
《农业灾害研究》
2016年第1期34-35,37,共3页
Journal of Agricultural Catastrophology
基金
辽宁省气象局科研项目:强对流天气特征物理参数及其预警指标体系本地化技术研究(201319)
关键词
冰雹
潜势预报
Bayes判别分析
指标叠套
Hail
Potential forecast
Bayes discriminant analysis
Logistic re-gression
Multi indictor superposition
