||The studies of tidal current and sea level variation in the Taiwan Strait are popular topics in recent years. The sea level data, to be applied to data analysis or model forcing and validation, are mostly observed in the near shore region. It is relative not easier to obtain real tidal data in the offshore area. This study intended to obtain sea level data within Taiwan Strait, using in-situ water depth measurements collected by EK500 of research vessels OR1, OR2 and OR3 during 1989-2003. The basic assumption of this work is that the changes of sea level and topographical depth equal to observed water depth. By using a large set of field measurements, it is possible to get bottom topography such that tidal data can be extracted by harmonic analysis of long-term discrete time series of water depth data.|
A total of 1513 cruises of water depth data were collected, which account for nearly 6 million samples. These data were screened through a series of criteria for quality control. Firstly, data were plotted cruise by cruise ( longitude vs latitude , longitude vs depth , time vs depth), then reasonable range of time, depth and region were choosed manually. Second, outliers, defined as values greater than 3 standard deviations on 5 point moving mean along the cruise track (or time), were replaced by linear interpolation values. Finally, a 2-minute moving average was applied to the along track time series water depth data. This step was trying to remove the effect of surface waves. The original huge records were reduced to about 550,000 valuable samples for the 1513 cruises data.
According to the density distribution of water depth samples in Taiwan Strait, 32 sub-region were selected for topography and harmonic analyses. In each sub-region, the bottom topography was mapped by an optimal interpolation method through a Gaussian weighting function. The radius of Gaussian weighting function applied is 3 time of the distance of grid. Water depth samples subtracted topographical depth of nearby grid to form a set of sea level data ready for harmonic analysis. The phase and amplitude of semi-diurnal tides (M2) and diurnal tides (K1、O1) in each sub-region were computed for the 32 regions in Taiwan Strait.
The water depth measurements derived sea level variations were compatible with that of a global tidal model (OSU) and a set of moored long-term pressure records in the middle of the strait. Especially, the tidal phase among these results were quite close. However, the tidal amplitudes of water depth data derived were smaller. Sensitivity analysis showed that the errors, differences between OSU model and depth derived sea levels, were small with regions of high density of water depth measurements. Both harmonic derived sea level variations and OSU model predictions indicated a southward propagating tidal wave, which matched with the scenario of Kevin wave propagation in Taiwan Strait. Our analysis also showed that the sea level variations in the northern part of the strait were dominated by M2 and K1 components while the southern part of the strait were dominated by M2 and O1 components.