基于智能计算的广西大风短临预报预警系统的产品检验

利用机器学习和人工智能技术研发了广西大风短临预报预警系统,该系统的产品与同期广西各地气象局发布的大风预警信号(以下简称“人工预警”)进行比较分析。结果表明:(1)按业务评分规定,大风预警系统在漏报率和命中率方面更优,人工预警在TS评分和空报率方面更优;(2)有效提前预警情况下,大风预警系统在大风蓝色、黄色预警和不分级预警中TS评分较高。基于对大风预警系统和人工预警的数量、TS评分和预警提前量的差异分析,广西大风短临预报预警系统的产品性能达到同期人工预警水平。

Product inspection of Guangxi gale short-term forecast and early warning system based on intelligent computing

Based on machine learning and artificial intelligence technology,Guangxi gale short-term forecasting and early warning system was developed.In order to evaluate the forecasting and early warning ability of the system,the products of the system were compared with the gale early warning signals(hereinafter referred to as"manual early warning")issued by meteorological bureaus of Guangxi in the same period.The results showed that(1)according to the business scoring rules,the gale early warning system was better in the missing report rate and hitting accuracy,while the manual early warning system was better in the TS score and empty report rate.(2)In the case of effective early warning,the gale warning system had a higher TS score in the blue,yellow warning and non-graded gale warning.Based on the difference analysis of the number,TS score and early warning amount,the product performance of Guangxi gale short-term early warning system reached the level of manual early warning in the same period.