近年來對於提高淺層震波速度構造解析度的研究越來越重要，其結果不僅有助於地表地質之解釋並在強地動模擬上提供重要的基礎參數，而高頻接收函數為其中一項重要的方法，同時可以提供震波速度與其相關不連續面的資訊。本研究挑選台灣東部縱谷平原北段作為試驗場址，該區域從大地測量的結果顯示為沉降的狀況，為近代的沉積物穩定堆積，因此可以經由此方法來計算沈積物與基盤的分布狀況，並比對在此區域曾做過的震測調查結果。在2017年1月至5月間，本研究利用三分量儀器ZLAND佈放之臨時密集地震網，共26個測站，測站間距約為2公里，利用高頻接收函數並以基因演算法搜尋測站下最佳一維剪力波模型解。研究結果顯示在測站下相對速度差異大的介面普遍存在於各測站，推論可能是沉積盆地與基盤的主要邊界。沉積物厚度在木瓜溪南側靠近東華大學、花蓮溪西岸與花蓮溪及木瓜溪匯流處較厚可達289公尺，比對古河道地圖發現沉積物分佈應與古河道有關，除了與地表地質結果一致，本研究也可以將主要速度不連續面對應到前人研究之震測剖面。另外透過波形模擬、震波粒子運動及將地震目錄分群進行接收函數及後續基因演算法計算的測試，也證明了利用區域地震震配合高頻接收函數法應用在探究淺層構造是可行且有效率的方法。;Recently, obtaining the high resolution of shallow seismic structures is important for the interpretation of the geological structures and the parameters for the ground motion simulation. The high-frequency receiver function is one of the important tools to achieve that. In this study, I selected the north part of the Longitudinal Valley (LV) in eastern Taiwan as a test site, because the north part of the LV is covered with well-developed alluvial fans and thick quaternary alluvium. In order to image the subsurface structure beneath the test site, I deployed a temporary dense seismic array, composed of 26 3-component ZLAND seismometers from January to May in 2017 and applied the high-frequency receiver function technique by using Genetic Algorithm (GA) to obtain the one-dimensional shear-wave velocity structure beneath each station. As a result, some interfaces with high-velocity contrast can be identified as the basement and unconformities. The basement geometry is related to the old river channels of Hualien River and Mugua River, which could reach to the depth of approximately 200 ~ 300 meters. Also, the result from this study is compatible with previous seismic reflection profile. Furthermore, by the testing the results from waveform simulation, inspection of the particle motion of the incident P-wave, most of results with the earthquakes from different back-azimuths and depths are consistent to each other, which shows that this method is feasible to investigate the shallow structure.
