中国气温和降水序列年代际分量的显著性检验

Significance test on interdecadal components of temperature and rainfall series in China

摘要:

对1951—2008年中国160站年和四季的气温、降水序列是否服从正态分布进行显著性检验,以此为基础进一步对气温和降水序列中的年代际异常分量显著性作了严格的统计学分析,结果表明:1)大多数测站气温序列服从正态分布,大多数测站降水序列不服从正态分布;气温、降水序列是否服从正态分布与地域有一定联系。2)气温、降水序列中年代际变化分量的显著性存在明显差异,多数测站气温序列年代际分量显著,只有少数测站降水序列的年代际分量显著。3)因为1951—2008年中国160站年、季气温和降水序列不全服从正态分布,特别是降水序列的非正态性严重,建议对它们的统计显著性检验采用Monte Carlo方法。

Abstract:

On the basis of the normality significance test of the annual and seasonal temperature and rainfall series at 160 stations in China from 1951 to 2008,the significance of the interdecadal anomaly components of temperature and rainfall series is statistically analyzed.The results show that:1)Temperature at most stations was normally distributed while rainfall was not,which was somewhat related to the regions.2) There was obvious difference in the significance of interdecadal components of temperature and rainfall series.The interdecadal component of temperature was significant at most stations while it was the case at only a few stations for rainfall.3) Since not all the annual and seasonal temperature and rainfall,especially the rainfall series,were normally distributed at the 160 stations in China from 1951 and 2008,the statistical significance should be tested by Monte Carlo scheme.