中文摘要：

对石臼所、连云港和坎门验潮站23 a（1975—1997年）水文观测资料进行了逐年调和分析。结果可以看出,不仅年平均海平面有逐年变化,而且分潮调和常数也存在明显的不稳定性。利用2种方法推算了3个站的23 a天文潮位,一是以某年（例如1975年）的年平均海平面和调和常数,另一是以当年的年平均海平面和调和常数推算3个站的23 a天文潮位。并据实测潮位和2种天文潮位推算值,计算出2种风增（减）水的年极值。由比较发现,3个站的2种年最高（低）天文潮位的最大差值达17-22 cm（-18-14 cm）;石臼所和连云港的2种100年一遇风减水的差值均达20-30 cm,坎门的2种100年一遇风增水差值达10 cm。这些差值表明,在人们推算天文潮和计算风增（减）水时,不应该忽视年平均海平面变化和分潮调和常数的不稳定性。由逐年的调和常数计算了每年的海图深度基准面,其23 a的变化幅度在石臼所达11 cm,在连云港达21 cm。拟建沿岸工程点的年最高（最低）天文潮位变化的曲线,年极值风增（减）水变化曲线及海图深度基准面的年变化曲线可以利用工程点两侧长期验潮站的相应曲线经过工程点1 a观测资料推算出的相应值,内插求得。

英文摘要：

Harmonic analyses were yearly made for the tidal level data in period of 1975--1997 at Shi ji- usuo, Lian yungang and Kanmen. It was seen from the results that there are not only the yearly mean-sea- level variation, but also the harmonics instability. The astronomical tidal levels of 23 years were computed by the use of two approaches, the one was to use the yearly mean-sea-level and harmonics of 1975, the oth- er was to use those of each year. And then, the two kinds of the yearly storm set-up（set-down） extreme were obtained based on the observed tidal levels and the two kinds of computed astronomical tidal levels. It was found from comparison that the maximum（minimum） of differences of the two kinds of yearly high- est（lowest） astronomical tidal at 3 tidal stations reached 17~22 cm（--18~14 cm）, and that at Shi jiusuo and Lian yungang, the differences of the two kinds of 100 years return periods of storm set-down reached 20~30 cm ,and at Kanmen, the difference of the two kinds of 100 years return periods of storm set-up reached 10cm. Such differences showed that when astronomical tidal levels and storm set-up and set-down are were computed, the yearly mean-sea-level variation and harmonics instability shouldfft be neglected. The yearly marine chart depth datum were computed from the yearly harmonics, the variation rage of which within 23 years was llcm at Shi jiusuo, that was 21cm at Lian yungang. For location chosen for coastal engineering, its yearly highest（lowest） astronomical tidal levels variation curve, yearly extreme storm set-up（set-down） variation curve and yearly marine chart depth datum variation curve could be in- terpolated by using the corresponding curves of two tidal stations at two sides of location through the corresponding values obtained from one year tidal observation at engineering location.