摘要
模式定量降水预报作为一种客观化预报方式已成为天气预报业务工作中的主要参考依据,为了解日本、德国、T639及天津(TJ-WRF)4种模式降水预报产品在海河流域的预报性能,从TS评分、漏报率、空报率、预报偏差等方面对4种模式进行了检验和评估。结果表明,4种模式各有优劣,对于小雨的预报,WRF和T639表现好于其它模式;对于中雨的预报,日本模式和T639的可信度较高;对于大雨的预报,T639的表现略好于其它模式;对于暴雨以上的预报,各模式的表现均很差。就预报偏差而言,各模式的预报偏差都基本偏小,且量级越大预报偏差误差越大,其中天津WRF和T639的偏小幅度小于其它2个模式。
The quantitative precipitation forecast of NWP models has become a major basis of everyday operation weather forecast at the moment. The precipitation forecasts of four NWP models, including a Germen global model, a Japanese global model, T639 and Tianjin-WRF are verified and assessed by the objective rainfall verification methods to find out the prediction performance of them in the Haihe River Basin. The results show that there are unique advantages of the four models for different rainfall classification. The forecast by Tianjin-WRF and T639 model is better than the others for the light rain; the forecast by Japanese and T639 model is more credible for the moderate rain; the performance of T639 is the best among the four models forecast for the heavy rain; the performance of four models are poorly for the more severe rain. The predicted error is small for all of the four models, and the heavier the rain is, the greater the predicted error is. The amplitude of predicted error of Tianjin-WRF and T639 model is smaller than the other two models.
作者
徐姝
魏琳
邓岩
XU Shu WEI Lin DENG Yan(Tianjin Meteorological Observatory, Tianjin 300074, China Bureau of Hydrology, Haihe Water Conservancy Commission, Tianjin 300170, China Aerial Survey and Remote Sensing Institution ,Beifang Investigation, Design ＆ Research Co.Ltd, Tianjin 300222, China)
出处
《海河水利》
2017年第2期45-51,共7页
Haihe Water Resources
关键词
海河流域
数值模式
降水预报
统计检验
定量误差
the Haihe River Basin
numerical model
precipitation forecasting
statistic test
quantitative error
