基于云诊断算法的中国十大机场云预报对比分析

Comparative analysis of cloud forecast for top ten airports in China based on cloud diagnosis algorithms

选取NCEP再分析资料FNL的6 h数值预报数据,基于云诊断算法分别计算中国十大机场1500 m以下最低层云的云底高和云量,利用相应时次机场例行天气报告检验其计算结果,并进行对比分析。结果表明:不分云量时,大部分机场C云量诊断法(C方法)的准确率、漏报率均低于WR95及优化的WR95方法(WR95opt)。WR95和WR95opt方法在北京、上海浦东、昆明、上海虹桥的TS评分比C方法稍高,在成都、西安、重庆的TS评分比C方法低。分云量时,各方法计算结果较差,准确率、TS评分减小,漏报率增大。C方法对于疏云及以上云量的计算结果比少云差。各方法计算的云高相对误差差距较大,相对误差有正有负。北京、上海浦东、广州、深圳、上海虹桥、杭州、昆明的WR95opt方法计算结果较好,成都的WR95方法计算结果较好,西安、重庆的C方法计算结果较好。

The 6 h numerical forecast data from NCEP reanalysis data FNL were selected to calculate the cloud base height and cloud cover of the lowest layer clouds below 1500 m in the top ten airports in China using the cloud diagnosis algorithm.The corresponding airport routine weather reports were used to check the calculation results and make a comparative analysis.The results showed that regardless of cloud cover,the accuracy rate and miss rate of the C cloud cover diagnosis method(C method)were lower than those of the WR95 method and optimized WR95 method(WR95opt)in most airports.The TS score of WR95 and WR95opt in Beijing,Shanghai Pudong,Kunming,and Shanghai Hongqiao is slightly higher than that of the C method,while in Chengdu,Xi′an,and Chongqing it is lower than that of the C method.While considering cloud cover,the calculation results from these methods are poor,the accuracy and TS score decrease,and the miss ratio increases.The calculation results of the C method for sparse cloud and above cloud are worse than that of less cloud.The relative error difference of cloud height calculated by each method is large,and the relative error spreads in both positive and negative.The calculation result of WR95opt in Beijing,Shanghai Pudong,Guangzhou,Shenzhen,Shanghai Hongqiao,Hangzhou,and Kunming is better,the calculation result of WR95 in Chengdu is better,and the calculation result of C method in Xi′an and Chongqing is better.