| 摘要: | 印尼複雜的構造背景為研究地震各向異性與餘震行為提供了獨特的自然實驗室。本研究呈現兩項互補性的調查：其一聚焦於蘇拉威西島下方的剪切波分裂，另一則探討2006年日惹 Mw 6.3 地震後餘震的重新定位。
蘇拉威西位於印度-澳洲板塊、巽他板塊與菲律賓海板塊的三重交會點，因此形成一個K字形的島嶼，周圍環繞著多個隱沒帶。我們利用印尼地震觀測網中69個寬頻地震站的資料，分析2021年至2024年間記錄到的遠震SK(K)S與區域性S波的剪切波分裂。事件篩選基於嚴格標準：震源深度大於等於100公里的區域性S波（濾波頻率0.1–1 Hz），以及震源矩規模Mw大於等於6.0的SK(K)S波相（濾波頻率0.02–0.2 Hz）。僅保留高品質的測量結果，其定義為極化角變異在±30°以內、延遲時間變異在±1秒以內，且Wüstefeld品質因子大於等於0.5。分析結果顯示，在蘇拉威西周圍的隱沒帶中，快速極化方向以與海溝平行為主，然而在如哥倫打洛灣與科洛火山附近，則呈現與海溝垂直的極化方向。這些異常反映了當地伸展性裂谷與火山活動等局部影響因素。主要斷層附近的測站顯示與斷層平行的極化方向，而多條構造交會處則呈現複雜的交錯型極化模式。若無明顯的局部應力源（如隱沒板塊或斷層），則較厚的各向異性地殼會展現更強的剪切波分裂，提供了蘇拉威西下方岩石圈變形與地函流動的重要新見解。
在爪哇島，2006年Mw 6.3日惹地震的震中與餘震分布偏離了Opak斷層東側，因此對其斷層來源產生疑問。以往對該斷層傾角方向的解釋不一。為釐清此議題，我們應用一種基於深度學習的自動拾取方法來分析P波與S波的到時，結果比傳統技術更具一致性。我們首先利用網格搜尋法估計絕對震源位置，接著更新區域速度模型以進行更準確的地震重新定位。重新定位後的餘震分布顯示，在主震東側的地震發生於較深位置，而西側的地震則較淺。根據Global CMT震源機制所進行的庫倫應力變化分析顯示，西側淺層餘震區域與正的應力變化相對應。這項結果支持主震為一條傾向西方、名為Ngalang斷層的構造所破裂的假設，而該斷層位於Opak斷層的東側。
總結而言，這兩項研究突顯了印尼地區地震過程的多樣性——從蘇拉威西隱沒與地函動力主導的地殼各向異性，到中爪哇地區的斷層複雜性與應力互動。兩者皆展現了先進分析技術與密集地震網絡整合的潛力，以提升我們對本區構造行為與地震災害的理解。;Indonesia′s complex tectonic setting provides a unique natural laboratory for studying seismic anisotropy and aftershock behavior. This study presents two complementary investigations: one focusing on shear-wave splitting beneath Sulawesi, and the other on aftershock relocation following the 2006 Yogyakarta MW 6.3 earthquake.
Sulawesi lies at the triple junction of the Indian-Australian, Sunda, and Philippine Sea plates, resulting in a K-shaped island surrounded by multiple subduction zones. Using data from 69 broadband stations in the Indonesian seismic network, we analyzed shear-wave splitting of teleseismic SK(K)S and local S-waves recorded between 2021 and 2024. Event selection was based on strict criteria: local S-waves from earthquakes ≥100 km deep (filtered at 0.1–1 Hz) and SK(K)S phases from MW ≥ 6.0 events (filtered at 0.02–0.2 Hz). Only high-quality measurements—defined by polarization angle variation ≤ ±30°, delay time variation ≤ ±1 s, and a Wüstefeld quality factor ≥ 0.5—were retained. The results reveal predominantly trench-parallel fast polarization directions along surrounding subduction zones, with trench-normal orientations observed in areas such as Gorontalo Bay and near Colo Volcano. These anomalies suggest local influences such as extensional rifting and volcanic processes. Fault-parallel patterns were observed near major fault zones, while complex cross-patterns emerged where multiple tectonic features intersect. Thicker anisotropic crust without significant local stress sources exhibited stronger shear-wave splitting, offering new insights into the lithospheric deformation and mantle flow beneath Sulawesi.
In Java, the 2006 MW 6.3 Yogyakarta earthquake posed questions regarding its fault origin due to its epicenter and aftershock distribution, which deviated eastward from the Opak Fault. Previous interpretations varied on the fault dip direction. To clarify this, we applied a deep learning-based picking method to analyze P and S wave arrivals, providing higher consistency than traditional techniques. Absolute hypocenter locations were first estimated using a grid-search method, followed by updates to the local velocity model for more accurate event relocation. The refined aftershock distribution revealed deeper events east of the mainshock and shallower ones to the west. Coulomb stress change analysis, based on Global CMT focal mechanisms, showed that the shallow western aftershocks correlated with areas of positive stress change. This supports the hypothesis that the mainshock ruptured a west-dipping structure known as the Ngalang Fault, located east of the Opak Fault.
Together, these studies highlight the diversity of seismic processes across Indonesia—from crustal anisotropy shaped by subduction and mantle dynamics in Sulawesi to fault complexity and stress interactions in central Java. Both demonstrate the power of integrating advanced analysis techniques with dense seismic networks to improve our understanding of the region′s tectonic behavior and seismic hazards. |
