NCEP/NCAR再分析风速、表面气温距平在中国区域气候变化研究中的可信度分析  

RELIABILITY ANALYSES OF ANOMALIES OF NCEP/NCAR REANALYSIS WIND SPEED AND SURFACE TEMPERATURE IN CLIMATE CHANGE RESEARCH IN CHINA

作　　者：施晓晖[1] 徐祥德[1] 谢立安[2] 

机构地区：[1]中国气象科学研究院灾害天气国家重点实验室,北京100081 [2]美国北卡罗莱纳州立大学,洛利nc276958208

出　　处：《气象学报》2006年第6期709-722,共14页Acta Meteorologica Sinica

基　　金：国家自然科学基金项目(40305005;90502003);国家重点基础研究发展计划项目(2003CB716806)

摘　　要：采用计算标准化均方根误差、相关分析和EOF分解等多种客观分析统计方法,对NCEP/NCAR再分析风速、表面气温距平在中国区域的可信度进行了研究,结果表明中国东部不同要素距平的标准化均方根误差均比西部地区的小,说明NCEP/NCAR再分析资料的可信度东部比西部要高,可能是受到模式地形和中国地面气象站点“东密西疏”分布格局的较大影响。随着高度升高,NCEP再分析风速距平的误差减小,进一步表明地形对NCEP再分析资料的可信度具有较大影响。另外,冬季再分析风速误差较大的特点在850,500和200 hPa等压面上均存在,表明冬季再分析风速距平的可信度受到再分析模式系统误差的较大影响。相关分析结果和标准化均方根误差计算结果之间具有很好的反向对应关系,即均方根误差大,NCEP再分析资料与实测资料的相关性就差,均方根误差小,则对应两者之间的相关性就较好。标准化均方根误差较小的要素,其NCEP再分析和站点实测资料距平EOF分解得到的特征向量空间分布较为相似,各特征向量对应时间系数的相关性也比较好;反之,标准化均方根误差大的要素,其NCEP再分析和站点实测资料距平EOF分解得到的特征向量空间分布则相差较大,对应时间系数的相关性也比较差,因此采用EOF分解方法,分析对应特征向量空间分布相似性及其时间系数变化的一致性,可以对NCEP再分析资料的可信度有一个更加客观的认识。综合上述各季节、各要素多种方法的分析结果,可以发现NCEP再分析风速距平在春、夏、秋季具有一定的可信度,但冬季的可信度较差;表面气温距平则是冬季的可信度最好,夏季的可信度较差。By means of various statistical analysis methods, such as normalized root mean square error （RMSE）, correlation analysis and empirical orthogonal function （EOF） decomposition etc, the reliability of the various seasonal anomalies of NCEP/NCAR reanalysis wind speed and surface air temperature （SAT） data frequently used in the regional climate change research in China is studied. Results show that RMSE of the meteorological variables are smaller in eastern China than in western China, i.e. the reliability of the NCEP/NCAR reanalyzed data in eastern China is better than that in western China. This could be due to effects of the topography in the reanalysis model and the disposition of ＂dense-in-eastern-and-sparse-in-western＂ of meteorological stations in China. The RMSE of anomalies of reanalysis wind speeds decreases with increasing height, further conforming the possible impact of topography on the reliability of the reanalysis data, and its RMSEs at 850, 500, and 200 hPa in winter are larger than those in other three seasons, suggesting that the reliability of reanalysis wind speed anomalies in winter was obviously affected by the system error of reanalysis model. Results of correlation analysis inversely correspond to those of RMSE analysis, i.e. if the RMSE is larger, then the correlation between the reanalyzed and observed data is worse （smaller）, and vice versa. It is found from comparing the EOF eigenvectors of anomalies of the reanalyzed and observed data that if a meteorological variable has a smaller RMSE, the spatial patterns of corresponding EOF eigenvectors of anomalies of the reanalysis and＇observed data are similar and their time coefficients are significantly correlated; and vice versa. Therefore, the similarity of EOF modes and the consistency of their time coefficients can be used to objectively assess the reliability of the NCEP/NCAR reanalysis data. On the whole, the reliability for wind speeds is better in spring, summer and autumn, but worse in winter; and for SATs it is

关 键 词：再分析资料 标准化均方根误差 相关分析 EOF分解 可信度 

分 类 号：P468[天文地球—大气科学及气象学]