中文摘要：

对江苏沿岸2006年1月4日-2007年1月31日28个站位的水位数据使用潮位统计和经验正交函数（EOF）方法进行分析和比较。统计分析得到了各站位的平均潮差、最大潮差、平均潮周期、平均高潮相位差数据。经验正交函数分析结果显示，第l模态与平均潮差吻合较好，包含了水位数据73．7％的信息。前2个模态均为主要半日分潮的信息，为潮汐按相位正交分离的结果；第3模态显示了径流输入变化引起的海面变化信息；之后的数个模态还包含了水位扰动的信息。研究表明，经验正交函数应用于潮汐分析的效果良好，可作为潮汐分析的有效工具，且在扰动检测方面有一定潜力。

英文摘要：

We analyzed the water level data obtained from 28 stations along the west shore of the south ern Yellow Sea from January 04, 2006 to January 31, 2007, using the statistical and empirical orthog onal function （EOF） method as a tool. The average tidal range, maximum tidal range, average tidal period and phase lag of high tide were calculated. The results indicated that the study area was mainly dominated by M2 tidal component. The EOF mode I contained 72.6 ~ variations of water level, agreed well with the average tidal range. Both the mode I and modeII clearly showed the effects of semidiur- nal tidal components, e.g. M2 and Sz. Specifically, the mode I and mode H were the results of or thogonal decomposition of phase difference. The mode m contained the information of water level var- iations caused by the river discharge in summer time. In addition, many of posterior modes contained the information of water level disturbances. This work indicates that EOF is a useful and efficient tool for water level analysis and has the potential to detect disturbances of water levels.