Time and space structure of interannual variations in summer rainfall over China.
Time and space structure of interannual variations in summer rainfall over China.
复制标题
DOI：
10.2151/jmsj1965.70.1b_585
复制
发表时间：
1992-02
影响因子：
3.1
通讯作者：
S. Tian;T. Yasunari
中科院分区：
文献类型：
--
作者：
S. Tian;T. Yasunari
本文描述了1951 - 1990年中国夏季（5 - 9月）降水年际变化的时空结构。首先，谐波分析应用于夏季降水序列。2-4年周期的变化（以下简称2-4年分量）占所有台站总方差的40%以上，61%的台站占总方差的70%以上。三年期在这一时期中似乎是最有说服力的。经验正交函数（Empirical Orthogonal Function）技术应用于2 - 4年周期序列。前两个模态分别占总方差的12.8%和10.7%。模式1揭示了长江流域与中国北方地区的跷跷板关系。梅雨模态2的空间分布较为复杂，但其特征可能是梅雨区（包括长江流域）与全国其他地区的振荡。虽然两个主模态的贡献比例不高，但合成图显示它们很好地反映了（应用非线性分析之前）的情况。这两种模式表现出很大的振幅调制。时间系数的幅值或平方有变大或变小的趋势。相关分析和合成分析表明，印度夏季风模态1与印度夏季风和南方涛动指数（SOI）相关。这一结果与前人的研究结果相一致，表明我国夏季降水在2~4年时间尺度上与ENSO（ElNino and Southern Oscillation）事件有着密切的联系。SOI的反常现象出现在第一模式之前。这一结果支持了以往赤道东太平洋ENSO信号可能起源于中亚或印度洋地区的观点。模态2与印度季风和ENSO无关。
The time and space structure of interannual fluctuations of summer rainfall (May to September) for the period from 1951 to 1990 over China is described. First, a harmonic analysis is applied to the summer rainfall series. The variation with periods of two to four years (hereafter, refered to 2-4 year component) accounts for more than 40 percent of the total variance for all stations, and more than 70 percent of the total variance for 61 percent of stations. The three-year period seems most prevailent among this period band. An EOF (Empirical Orthogonal Function) technique is applied to the 2-4 year period series. The first two EOF modes account for 12.8 and 10.7 percent of the total variance, respectively. EOF mode 1 reveals the seesaw between the Yangtze River valley and the Northern part of China. The spatial pattern of EOF mode 2 is more complicated, but it may be characterized by the oscillation between the Mei-yu region including the Yangtze River valley, and the rest of the country. Although the contribution proportion of the two principle modes is not high, composite maps show that they present the situations (before applying the EOF analysis) well. These two EOF modes exhibit large amplitude modulations. The amplitudes or the squares of the time coefficients tend to become large or small alternatively. Correlation and composite analyses show that EOF mode 1 correlates with the Indian summer monsoon and the SOI (Southern Oscillation Index). This result agrees with that of preceding studies and suggests that the summer rainfall over China is associated with the ENSO (El Nino and Southern Oscillation) events on the time-scale of two to four years. EOF mode 1 seems to precede the anomaly of the SOI. This result supports the previous proposals that the ENSO signals in the eastern equatorial Pacific may originate in the central Asia or the Indian Ocean region. EOF mode 2 seems not to be related to the Indian monsoon and the ENSO.