利用折射與反射震測資料之聯合逆推探討南海北部大陸邊緣地殼速度構造

Abstract

南海北部為最早期開始擴張的位置，其大陸邊緣也記錄了張裂演化歷史。在南海北緣存在一項顯著的地球物理特徵，稱為高速帶(high-velocity layer)，即位於莫荷面之上且P波速度超過7km/s的塊體，常被解釋為岩漿添附至下部地殼的底侵作用(underplating)所致。因此，對高速帶的研究能夠更深入了解岩漿活動及其與構造演化的關係。然而，高速帶在南海的空間分布與成因仍有歧見。本研究結合多頻道震測系統與海底地震儀走時資料，建立一位於南海北緣的地殼速度構造剖面。首先，分析多頻道反射震測資料，以獲得基盤以上的初始速度模型。接著，結合海底地震儀所選取的走時資料，採用蒙地卡羅與自適應重要性採樣策略，並使用能聯合逆推折射與反射走時的TOMO2D層析成像法，建立完整地殼速度構造。結果顯示，下部地殼存在有最厚約6公里的高速帶，且呈獨立、塊狀分布，並非單一均勻的層狀構造，而基於速度構造所建立之重力模型亦支持此高速帶特徵的存在。隨後，我們重新檢視並彙整已發表文獻之速度構造剖面，並編繪南海北部陸緣高速帶的厚度與分布，再藉由分析地殼厚度與平均下部地殼速度關係（H–Vp分析），綜合探討高速帶形成時可能的地函動力學狀態。透過H–Vp分析結果與高速帶空間分布，並比較其與磁力異常的關係，揭示南海北緣高速帶的區域差異及演化過程：在大陸棚與大陸斜坡區，高速帶呈條帶狀分布，並對應顯著東北-西南向正磁力異常，H-Vp之正相關顯示接近正常位溫的主動地函上湧，可能由隱沒驅動的小尺度地函對流所致；於臺灣西南側，較厚的高速帶並未直接對應更高的磁力異常，H-Vp呈負相關趨勢，顯示其可能受到繼承構造的影響；至於在大陸海洋過渡帶，高速帶則呈獨立、多塊狀分布，H-Vp的正相關及高溫特徵，推測與岩石圈不穩定導致的區域熱異常事件相關。

The northern margin of the South China Sea (SCS) is where continental rifting initially occurred, and may preserve the full history of tectonic evolution. A significant feature in the lower crust of the northern margin of SCS is the high‑velocity layer (HVL), showing a P‑wave seismic velocity exceeding 7 km/s overlying the mantle. The HVL is often interpreted as magmatic underplating, providing crucial insights into magmatic activities during with tectonic evolution. However, the spatial distribution and possible origin of the HVL in the SCS remain subjects of debate. In this study, we integrate multi-channel seismic (MCS) and ocean-bottom seismometer (OBS) traveltime data to construct a crustal velocity structure profile cross the northern margin of the SCS. We firstly analyze MCS data, deriving a robust initial velocity model above the basement. By the Monte-Carlo method and adaptive importance sampling strategies, we perform joint inversion of reflection and refraction data using TOMO2D to construct a crustal velocity model. The final results reveal that HVLs in the lower crust with a maximum thickness of approximately 6 km. The HVL is characterized by isolated blocks rather than a layered structure with uniform thickness, and this HVL geometry is also validated by gravity modeling. Next, we reexamine and compile published profiles to prepare the thickness map of the HVLs along the northern margin of the SCS. Furthermore, by comparing theoretical and observed crustal thickness (H) and average lower-crustal P-wave velocity (Vp) plots (H-Vp plots), along with published magnetic anomaly map, we comprehensively investigate the possible mantle dynamics that drove the formation of the HVL. The positive-to-no H-Vp correlation and corresponding belt-like positive magnetic anomaly along the continental shelf and slope, indicate subduction-driven small-scale mantle convection. The negative H–Vp correlation and the lack of direct correspondence between the thicker HVL and higher magnetic anomalies in the southwest of Taiwan, indicate an influence from inherited structures. The positive H–Vp correlation, along with the isolated, scattered HVL segments and high-temperature features in the continent-ocean transition zone, indicates localized thermal anomaly events caused by lithospheric instability.