郑州市地面风场的统计降尺度预报研究  

Statistical downscaling forecasting of surface wind field in Zhengzhou

作　　者：冯慧敏 智协飞[1] 李荣[2] 

机构地区：[1]南京信息工程大学大气科学学院,南京210044 [2]郑州市气象局,郑州450007

出　　处：《中国科技论文》2017年第15期1785-1791,共7页China Sciencepaper

基　　金：中国气象局华中区域气象中心科技发展基金项目(QY-Z-201302)

摘　　要：为了提高郑州市地面风场的精细化预报水平,基于欧洲中期天气预报中心(European centre for medium-range weather forecasts,ECMWF)、日本气象厅(Japan meteorological agency,JMA)、中国气象局(China meterological administration,CMA)3个中心的2012年6月1日-9月30日地面风场3~60h预报资料和郑州加密自动站观测资料,先对郑州市11个站的地面风场进行多模式集成预报试验,再对多模式的集成结果进行统计降尺度研究。针对地面风场预报的难点,分别对地面风场的U、V分量(试验方法1)及全风速和风向(试验方法2)进行试验。最后采用均方根误差对2种试验方法的预报结果进行检验评估。结果表明:在多模式集成的2种试验方法中,对全风速和风向的的试验方法要优于对风场的U、V分量的试验方法,且此试验方法的超级集合集成方案(superensemble forecasts,SUP)预报误差较小,而成为最优的方案。在对多模式集成结果的2种统计降尺度试验中,对全风速和风向的多模式集成结果的方法能够进一步降低预报误差,与对风场的U、V分量的方法相比,是1种较好的方法。与多模式集成的精细化预报相比,统计降尺度预报能够进一步减小预报误差,提高地面风场的预报准确率。To improve the refined forecasting skill of surface wind field of Zhengzhou, based on the surface wind field （3-60 h） forecasting data from ECMWF （European centre for medium-range weather forecasts）, JMA （Japan meteorological agency）, CMA （China meterologieal administration） and observational data at automatic weather stations in Zhengzhou in the duration from 2012-06-01 to 2012-09-30, the multi-model ensemble forecasts of the surface wind field of 11 stations in Zhengzhou were firstly carried out, and then the statistical downscaling of multi-model ensemble results were studied. In view of the difficulty of the surface wind field forecasting, the U and V component of the wind field （method 1） and the full wind speed and wind direction （method 2） were tested. Finally, the RMSE（root mean square error） was used to evaluate the results of the 2 methods. The re- sults show that the method for the full wind speed and wind direction is better than that of the U and V components of the wind field in the 2 methods, and the SUP（superensemble forecasts） of this method has smaller forecast error and becomes the best scheme. In the statistical downscaling experiments for the multi-model ensemble, the multi-model ensemble for the full wind speed and wind direction can further reduce the forecast error, which is better than that for the U and V components of the wind field. Compared to the refined forecasting of multi-model ensemble, the statistical downscaling forecasting can further reduce the forecast error and improve the forecast accuracy of the surface wind field.

关 键 词：地面风场 精细化预报 多模式集成 统计降尺度 

分 类 号：P457.5[天文地球—大气科学及气象学]