Abstract: Ensemble empirical mode decomposition (EEMD), a newly developed nonlinear data analysis method, is employed to derive climate change signals, such as annual cycle, low-frequency components and trends, etc. The data sets for the analysis are the well-homegenized, longer than 150 years, daily temperature series of five stations in Europe. The decomposed results indicate that there are three main time scales, e.g. interannual, interdecadal and century scales, for the low-frequency variations of all five stations. The intensities of the annual cycle were weak during the two warm periods: 1910?1940 and the last 30 years since 1970. And the weak trend was more obvious in the last 30 years. In addition, summers become more longer and winters shorter since the late of 1970s compared with that of warm period in 1910?1940.

Key words: climate change, ensemble empirical mode decomposition (EEMD), intrinsic mode function (IMF)

摘要： 利用EEMD方法对欧洲5个站大于 150年逐日温度序列进行分解, 分析了欧洲温度序列的低频变化、年循环及季节变化。结果表明: 欧洲5站温度低频变化均存在明显的特征时间尺度, 即年际、年代际和世纪尺度等; 年循环强度在1910?1940年及1970年年末以后的两个暖期里均处于偏弱的状态, 尤其是最近30年里年循环强度减弱趋势更加明显; 此外, 与1910?1940年相比, 在20世纪70年代末以来的暖期里, 夏季更长, 冬季更短。

关键词: 气候变化, 集合经验模态分解(EEMD), 本征模函数(IMF)