運用海床壓力計研究北南海海盆內孤立波之變化

Abstract

內孤立波通常具有大波幅的波浪特徵，且會導致水平流速的強烈垂直流切以及等密度面的垂直位移。此外，由其引起的海底壓力擾動也值得關注。於南海，前人進行多次針對內孤立波的研究，像是使用溫度計串、流速剖面儀、衛星影像等進行觀測，然而欠缺的是長期觀測資料。本研究使用將近三年的觀測資料，藉由研究分析了解內孤立波在此處的季節性變化與性質。於2018年10月至2021年7月，一組海嘯浮標預警系統布放在南海北部深海海盆，距離呂宋海峽以西約110公里，當地水深約為2800公尺，此處為南海內孤立波好發之處。此系統搭載高解析度海床壓力計，得以觀測到內孤立波在深海造成的微小壓力變化，於觀測期間共紀錄513個內孤立波造成的壓力擾動。 本研究主要分為兩部分，其一是以連續小波變換和希爾伯特-黃轉換進行海底壓力計資料分析，藉此獲得內孤立波訊號，並將此訊號進行統計已得知其季節性變化。第二部分則是使用Dubreil-Jacotin-Long (DJL)方法了解內孤立波性質，由給定當季或歷史背景密度資剖面以及可用位能，計算出相對應的波速、壓力及振幅。並比對實測以及計算壓力值的相似性，便可同時驗證其他性質的可信度。 利用連續小波變換及希爾伯特-黃轉換分析海床壓力計數據得知：第一點，連續小波轉換所得週期介於0.5至2小時的訊號與希爾伯特-黃轉換所得之本質模態函數(intrinsic mode functions, IMF) IMF 1、2和3相同；第二點，此訊號發生時間為大潮期間，與前人研究之南海內孤立波好發時間相符；第三點，統計結果顯示，冬季為南海內孤立波發生的低峰期。DJL方法結果可總結為以下兩點：首先，DJL方法所得的壓力值與現場觀測的壓力數據經由計算R-squared (r^2)得知兩者的相似度，若其值大於0.95則將此結果視為高度相似，同時採信所得之波速。其次，由上述結果所得之波速與由向日葵8號衛星影像計算的波速進行比較，可進一步驗證DJL方法的準確性。然而，研究發現使用當季背景密度剖面所得之波速結果相較於使用歷史剖面有較高的準確性，因此背景密度剖面的季節性變化可能會影響DJL方法所得之波速的準確性。

Internal solitary waves (ISWs) are usually characterized by a large wave amplitude, strong vertical shear of horizontal currents, and vertical displacement of isopycnals. Additionally, pressure perturbations above the seafloor induced by ISWs are also noteworthy. There has been extensive research on internal waves in the South China Sea (SCS), utilizing instruments such as thermistor chains, current profilers, and satellite images. However, these studies often have short observation periods. Long-term observation data would allow further analysis of internal solitary waves in the SCS. In this study, a Tsunami Buoy System with a bottom pressure recorder (BPR) was deployed in the northern SCS deep basin, approximately 110 kilometers west of the Luzon Strait, at a depth of 2,800 m, to document pressure perturbation in-situ data from October 2018 to July 2021. During the observation period, 513 cases of ISWs were recorded. The research mainly focused on two parts: first, the data from the BPR was analyzed using continuous wavelet transform (CWT) and Hilbert-Huang transform (HHT), and the ISW signal was extracted. Then, with the long-term signal, the seasonal change of the ISWs could be understood. Second, the Dubreil-Jacotin-Long (DJL) method was applied to understand better the parameters of ISW. With an in-season or a historical background density profile and available potential energy (APE) value, the DJL method can output the results of phase speed, pressure, and amplitude. By fitting the observation and calculated pressure data, the similarity helps to verify the confidence of other output parameters. From the analysis of the CWT and HHT, there were three conclusions: (1) the 0.5 to 2-hour period signal from CWT and the IMFs 1, 2, and 3 signal from HHT generated intrinsic mode functions (IMF) had a similar period; (2) the happening time of these signals matched with the generation time from former research, which is on the days of spring tides; (3) statistics showed the seasonal pattern of internal solitary waves in the SCS. There were fewer cases in winter, while the other three seasons experienced higher occurrences, which aligns with previous studies. The calculation results from the DJL method indicated two conclusions. First, the pressure data from the DJL method yielded high similarity to the in-situ pressure data, according to the results from R-squared (r^2) better than 0.95, leading to higher credibility to other parameters, such as phase speed. Second, comparing the phase speed results and Himawari-8 image calculations further validated the method's accuracy. However, it was found that an in-season background density profile could improve the accuracy of the DJL results even more effectively than a historical background density profile. Therefore, the seasonal change of the density profile might be necessary for the phase speed accuracy from the DJL method.