臺灣海域孕震構造資料庫

Abstract

臺灣屬於活躍的造山帶，擁有眾多孕震構造與頻繁的地震活動。在此地質背景下，臺灣地震模型（Taiwan Earthquake Model）計畫已辨識出45條陸域孕震構造，並廣泛應用於災害評估與防災。然而，儘管臺灣近海曾報導具有多個活動構造與豐富的歷史地震紀錄，目前仍缺乏系統性的海域孕震構造資料庫供防災使用。 本研究整合臺灣周邊海域的地質與地球物理資料，藉由反射震測剖面中不連續反射訊號所指出的斷層是否延伸至海床，以及海底地形圖中構造線形的側向延續性，共辨識出54條近海孕震構造的位置與地下幾何形態。本研究更進一步推估其孕震參數，包含可能的地震規模、長期滑移速率與地震再現週期。根據三種斷層幾何與地震規模關係式，並考慮海陸構造聯合破裂的情況下，所有海域構造皆有能力產生規模6.5以上的地震。由於在隱沒帶中，因物質特性使得構造的部分應力可能以無震潛移方式釋放，故本研究引用隱沒帶的耦合率資料調整滑移速率，重新估計各斷層的有效應變累積量。結果顯示，滑移速率較高（約10–20 mm/yr）的構造主要分布於琉球與馬尼拉隱沒帶的上盤區域，對應的地震再現週期小於1000年；而滑移速率較低（約0.1–2.0 mm/yr）的構造則位於臺灣東北外海的後造山區，再現週期多數大於6000年。 由於震測剖面資料品質與地形解析度的限制，構造參數仍存有相當大的不確定性，需進一步優化。本研究建立了第一版適用於地震危害度分析的海域斷層資料庫，預期能提升臺灣地區的地震危險度評估，並促進臺灣的地震科學研究朝向海洋發展。本研究所提出構造長期滑移速率的估計方法，也期望能夠應用於其他缺乏定年資訊的海域構造研究中。

Taiwan is an active orogenic belt characterized by numerous seismogenic structures and frequent earthquakes. In this dynamic setting, the Taiwan Earthquake Model (TEM) project has identified 45 on-land seismogenic structures, widely used for disaster assessment and preparedness; however, despite reports of active structures and many historical earthquakes offshore Taiwan, a comprehensive offshore seismogenic structure database is still lacking for seismic hazard analysis. This study integrates geological and geophysical data to determine the locations and subsurface geometries of offshore structures. Based on two criteria, namely the seabed disturbance by faulting in seismic reflection profiles and the lateral continuity of seafloor structural lineaments, 54 structures offshore Taiwan were identified and mapped. This study further estimated their structural parameters, such as capable earthquake magnitude, long-term slip rate, and the mean earthquake recurrence interval. By employing three fault geometry–earthquake magnitude scaling relationships and considering possible multi-fault rupture scenarios between on-land and offshore faults, this study finds that all identified offshore structures are capable of generating earthquakes of Mw ≥ 6.5. In subduction zones, some tectonic stress may be released aseismically due to material properties. To account for this, published coupling ratio data were applied to adjust slip rates, refining estimates of effective strain accumulation on fault planes. Our results show that the area with fast-slipping (~10–20 mm/yr) structures is located on the hanging wall side of the Ryukyu and Manila subduction systems, corresponding to recurrence intervals of less than 1,000 years, whereas structures with lower slip rates (~0.1–2.0 mm/yr) are located in the post-collision zone off northeast Taiwan, with most recurrence intervals more than 6,000 years. Due to limitations in seismic-profile quality and bathymetric resolution, these parameters still have large uncertainties and require further refinement. However, by establishing this first version offshore database, it is expected to enable not only improved assessments of earthquake hazards in Taiwan but also to promote marine-focused seismic research. Moreover, the proposed method for estimating long-term slip rates offers applicability to offshore seismogenic structures in other regions lacking geochronological constraints.