用T106L19全球谱模式制作中期集合预报的试验  

Medium-range ensemble forecast experiments using model T106L19

作　　者：关吉平[1] 张立凤[1] 张铭[1] 

机构地区：[1]解放军理工大学气象学院,江苏南京211101

出　　处：《解放军理工大学学报（自然科学版）》2006年第3期297-301,共5页Journal of PLA University of Science and Technology(Natural Science Edition)

基　　金：大气科学和地球流体力学数值模拟国家重点实验室开放课题资助项目(2004-01-01).

摘　　要：为解决传统的数值预报初值存在的不确定性,利用增长模繁殖法在T106L19全球谱模式上进行了中期集合预报试验。繁殖循环的初始扰动由模式24h预报均方根误差和随机数的乘积构成,该随机数服从[-1,+1]区间上的均匀分布,繁殖循环的周期为6h,繁殖的总时间为3d,集合成员为15个。结果表明：在1-10a的预报中,无论是距平相关系数还是预报均方根误差,除第1天外,集合预报都优于控制预报,集合预报相对于控制预报的优势更主要地体现在对第3天以后的中期预报时段。从计算的离散度来看,集合平均预报的可信度要比控制预报的高。To solve the uncertainty of initial fields in the traditional numerical forecast, medium-range ensemble forecast experiments using the model T106L19 were made. The first day＇s forecasting error of the model times random number was regarded as the initial disturbance of the breeding cycles and these random numbers met uniform distribution between --1 and ＋1. The breeding period is once per 6hr,the total time of breeding is 3 days and the number of ensemble members is 15. Results show that form the first to the tenth day forecasts, whether for anomaly correlation coefficient or forecast rmse error, except for the lth day, the ensemble forecasts are all more accurate than the control forecast. The superiority of ensemble vs control forecast is mainly embodied by the medium-range forecasts after the 3th day. Viewed by the computed spread, ensemble averaged forecast is more reliable than the control forecast.

关 键 词：T106L19全球谱模式 中期数值预报 集合预报 增长模繁殖法 

分 类 号：P456.7[天文地球—大气科学及气象学]