摘要
利用国内外几种数值预报模式的降水预报产品,对2003年主汛期(6~8月)青藏高原中南部的预报能力做了详尽的检验.检验结果表明,各模式小雨预报的TS评分较高,明显高于全国及其它地区的评分;但中雨以上各级降水检验评分较低,主要原因是各模式的预报雨区面积明显偏大,空报率较高.无论是预报员的预报,还是各种全球或区域模式的预报;无论是日本和德国的模式,还是我国目前业务运行的数值预报模式,对青藏高原的降水预报均无明显优势.
The predictive products of precipitation for various numerical simulation modles at home and abroad to the predictive ability in the mid-and-south parts of Qinghai-Xizang Plateau in summer of 2003 were detailly verified. The results of accumulated verification show that TS score of light rain area is the highest of all areas in China, but that of moderate rain or over are lower. The main reason is that the forecasted area of each model is clear larger than the observed one and false alarm ratio is highter. And so, whether it is forecaster and global model, or local model; whether it is Japan and German models, or T213 HLAFS25 and HB-MM5 of China, all of them have no obvious superior to the prediction of precipitation.
出处
《高原气象》
CSCD
北大核心
2004年第z1期53-58,共6页
Plateau Meteorology
基金
中国气象局成都高原气象研究所开放实验室基金资助
关键词
数值预报模式
降水预报
统计学检验
Numerical prediction model
Precipitation prediction
Statistical verification