南海北部沿岸海域潮汐的调和分析

Tidal analysis along the coast of the northern South China Sea

采用t_tide潮汐分析工具对南海北部的5个验潮站2009年全年的逐时潮高资料进行调和分析,计算出各站的调和常数,评估调和常数的准确性、稳定性,并总结了广东沿岸海域潮汐特征.利用对2009年逐时潮高的调和分析结果对2010年全年的潮高进行预测,将各站预测结果与同时间的实测数据进行全年和分季节进行比较,对预测结果与实测数据的残差进行统计分析.通过对残差的散点分布、概率分布、置信区间等统计结果进行分析,检验预测结果的准确性、稳定性和可靠性.结果表明:广东沿岸海域潮汐是以M2分潮为主,K1、O1、S2为次结合的潮汐机制,采用t_tide潮汐分析工具对南海北部潮高的预测结果与实测数据拟合较好,相位预测准确,潮高预测除在时间序列尾部(年尾)有些许较大的误差外,t_tide工具在南海北部潮汐预报中具有较高的准确性和稳定性.预测残差的整体服从正态分布,残差均值小于10-2m量级,方差最大为0.229 4,最小为0.173 2,95%置信区间长度小于10-2.各站季节分析主要分潮的离散度小于0.04的结果充分证明不同季节的分析区别不明显,3个月资料与整年资料的调和分析结果几乎一致,与所选取的季节资料几乎无关.虽然在预测值中,有极个别的残差将近1 m,但并不足以影响到预测的准确性.

The tidal analysis tool t_tide was adopted to analyze the measured data of five tide gauge stations in the northern part of the South China Sea in 2009.The harmonic constants of these five stations were obtained and evaluated for their accuracy and stability.The hourly tide heights at these stations in 2010 were predicted based on the harmonic analysis results of 2009.The annual and quarterly predicted results of the five stations were compared with the corresponding measured data.The residuals of the forecast results and the measured data were analyzed statistically.The accuracy,stability and reliability of the results were examined through the statistical analyses of the residual scatter plot distribution,probability distribution and confidence intervals.The results showed that the tides along the Guangdong coastal area were primarily the M2 tide,supplemented by K1,O1,S2 tides.The forecast results fitted well with the measured data.The predicted phase is accurate,so are the tide heights,except for few slightly larger errors near the end of the time series.The t_tide is a suitable tool to forecast the tide of the northern South China Sea.The predicted residual was normally distributed.The residual means were less than 10-2 m.The maximum variance was 0.229 4,and the minimum 0.173 2.The 95% confidence interval length was less than 10-2.The seasonal analysis results showed that the dispersions of the main points of all stations were less than 0.04.The harmonic analysis results were almost irrelevant to the measured data selected in different season.Although in the predictions a few individual residuals were close to 1 m,it is not enough to influence the accuracy of the predictions.