摘要
本文讨论了在我国海平面研究中常用的线性回归和滑动平均方法确定平均海平面长趋势变化存在的问题,指出当海平面观测系列包含有周期变化成分时,上述两种方法通常不能得到正确的结果。建议先对观测序列作显著周期分析,然后以线性趋势项与显著周期项拟合观测序列的方法来确定平均海面的长趋势变化。
In determining the long term trend in mean sea level variations, some problems existing both in the linear regression and moving average methods that are commonly used in sea level studies are discussed in this paper. It is pointed out that as to the observational series of mean sea level containing some periodic variation components, the two methods mentioned above can not give correct results. To overcome the disadvantage, it is suggested that, significant periods should first be analysed and then the observational series are fitted with a linear term and those significant periodic terms to determine the long term trend in the mean sea level variations.
出处
《地壳形变与地震》
CSCD
1992年第3期44-48,共5页
Crustal Deformation and Earthquake
基金
国家自然科学基金9488007号大型项目与中国科学院87-45-03号重大项目联合资助
关键词
海平面
线性回归
长趋势变化
sea level, linear regression, moving average, long term trend variation.
