船载ADCP资料的潮流分离方法在厦门港断面重复走航中的应用

Application of the separation of tidal and subtidal currents for ship-mounted ADCP data from repeating measurements in a transect of Xiamen Harbor

本文运用Candela等提出的潮流分离方法,选用高斯函数作为基函数,采用Matlab程序,首次对厦门港嵩鼓水道的船载ADCP断面重复走航资料进行潮流分离计算.通过对14个计算点的Can-dela方法计算结果与引进差比关系准调和分析方法计算结果的分析比较可见:(1)两种方法计算所得余流在大、小潮的相关系数分别为0.988 2和0.852 1,余流相对误差在大、小潮分别为0.065和0.150,而所得余流方向在大、小潮的相关系数分别为0.998 2和0.986 5,余流方向相对误差在大、小潮分别为0.023和0.027.两种方法计算的余流及其方向的相关性很高(在样本数为14,置信度为α=1%时,相关系数大于0.661 0,结果是可信的),而平均相对误差也很小.(2)对14个计算点的M2、S2、K1、O1、M4和MS4等6个分潮的计算结果(共有84个样本数据)进行分析比较,两种方法计算所得6个分潮的长轴、长轴方向和短轴在大潮的相关系数分别为0.983 8、0.896 0和0.233 5,而在小潮的相关系数分别为0.965 6、0.755 5和0.220 9.这两种方法计算所得6个分潮的长轴、长轴方向和短轴在大潮的平均相对误差分别为0.375、0.071和0.753,而在小潮的平均相对误差分别为0.287、0.254和0.845.两种方法计算的分潮长轴及其方向的相关性很高(在样本数为84,置信度为α=1%时,相关系数大于0.283,结果是可信的).相应的分潮长轴的相关性略低,但是在样本数为84,置信度为α=5%时,相关系数均大于0.217.可认为在置信度为α=5%时,计算结果是可信的,而平均相对误差则略显较大.综上所述,我们认为采用Candela等的方法对船载ADCP断面重复走航资料进行潮流分离计算是可行的.

In this paper, we separate tidal current and residual current for data obtained from the ship-mounted ADCP measured repeatedly in Songgu watercourse of Xiamen Harbor, applying the method developed by Candela and using Gauss function as basic function and Matlab programming. By comparing the result of Candela analysis method with result of traditional harmonical analysis method on 14 compute points shows that（ 1 ） the correlation coefficients of residual current obtained from the two methods during the spring tide and neap tide are 0. 988 2 and 0. 852 1 respectively. The average relative errors are 0.065 and 0. 150. The correlation coefficients of residual current direction obtained from the two methods in spring tide and neap tide are 0. 998 2 and 0. 986 5 respectively, and the average relative errors are 0. 023 and 0. 027. It is obvious that the correlations of the residual current and its direction obtained from these two methods are quiet closed, and the average relative errors are very low; （2） with results of 84 data from M2, S2 ,K1, O1 , M4 and MS4 etc in 6 tidal constituents on 14 compute points, the correlation coefficients of 6 tidal constituents long axis, long axis direction and minor axis obtained from the two methods are 0.983 8,0. 896 0 and 0.233 5 in spring tide respectively ,and the correlation coefficients are 0.965 6,0.755 5 and 0. 220 9 in neap. fide, respectively. The average relative errors of 6 tidal constituents long axis,long axis direction and minor axis obtained from the two methods are 0. 375 ,0. 071 and 0. 753 in spring tide respectively, and the average relative errors are 0. 287,0. 254 and 0.845 in neap tide respectively. It is obvious that the correlation of 6 tidal constituents long axis and long axis direction obtained from the two methods are quiet closed, and correspondingly the correlation of 6 tidal constituents minor axis is slightly low,but in sample number is 84 and confidence is α = 5%, the correlation coefficients are higher than 0. 217, and we suggest that the result is credible. And the average relative errors are slightly larger. In a word, the application of the separation of tidal and subtidal currents on the data from shipmounted ADCP is feasible.