https://ir.lib.ncu.edu.tw/handle/987654321/8 Search the Channel s https://ir.lib.ncu.edu.tw/simple-search https://ir.lib.ncu.edu.tw/handle/987654321/111457 title: 以二維 ETAS 模型探討臺灣東部地震活動特性;Investigating Seismicity Characteristics in Eastern Taiwan Using a 2-D ETAS Model abstract: 臺灣東部位於板塊交界處，因此發育出複雜的孕震構造，如米崙斷層、縱谷斷層及中央山脈斷層，並產生多起重大地震事件。為探討孕震構造及地震序列於時空中的活動特性，本研究採用二維傳染型餘震序列模型（Zhuang et al.， 2002）擬合中央氣象署地震目錄以量化地震活動特徵，並導入雙尺度經驗機率積分轉換方法（Zhuang et al.， 2017， 2020）填補地震目錄之早期餘震缺失，以評估目錄補全前後對傳染型餘震序列模型參數之影響。結果顯示，局部擬合能避免大範圍中多區域相異的地震活動干擾，更能精確量化特定區域之地震活動，如米崙區域因流體影響，造就其群震特徵，反映出餘震生產指數? <br> Tue, 20 Oct 2026 09:47:25 GMT https://ir.lib.ncu.edu.tw/handle/987654321/110340 title: A Systematic Simulation Approach to Magnetotelluric Survey Design and Data Quality Assessment in Urban Environment;A Systematic Simulation Approach to Magnetotelluric Survey Design and Data Quality Assessment in Urban Environment abstract: 隨著城市發展的快速擴張，大地電磁法（Magnetotelluric method， MT）在地下調查中的應用日益增加，範圍涵蓋水文地質研究至地熱能源探勘。然而，隨著城市地區持續擴張，城市化也對大地電磁測勘的設計與資料品質帶來重大挑戰。城市基礎設施的增長不僅會產生較高水準的電磁（EM）噪訊，掩蓋天然的大地電磁訊號，同時也降低了適合測點的可用性，常迫使測站必須遷移，進而影響空間解析度，並連帶影響地下成像的品質。儘管這些問題具有實務上的重要性，其對大地電磁結果的影響仍未被充分釐清，因此仍需更清楚地了解這些因素如何影響測勘成果。為回應此研究缺口，本研究提出一套以模擬為基礎的方法，用以檢視城市相關因素在貼近實際測勘環境的條件下，如何影響大地電磁的結果。具體而言，本研究聚焦於城市大地電磁測勘中常見的兩個面向，即噪訊條件下的大地電磁響應，以及測站間距對反演結果的影響。針對第一個面向，本研究利用一維正演模型生成合成大地電磁訊號，並施加不同的噪訊條件，再將所得響應與正演模型中已知的理想響應進行比較評估。針對第二個面向，本研究透過在不同測站間距配置下，施加一維與二維反演，探討測站間距對結果的影響。研究結果顯示，傳遞函數中的長週期區段普遍最容易受到噪訊影響，表現為視電阻率與相位上的散布情形，且隨著噪訊水準提高，此散布現象會擴展至更高頻率範圍。暫態噪訊所造成的損害程度高於高斯白噪訊。基於此發現，針對測站間距對反演結果影響的評估顯示，一維反演由於各測站皆為獨立反演，因此大致不受測站間距影響，但在個別測站層級仍容易受噪訊影響，因缺乏鄰近測站資料可供校正。相對而言，二維反演則明顯對間距敏感，較密的間距能提升結構解析度，但在噪訊條件下也容易產生假象，而較疏的間距則能提供較穩定但細節較少的結果。當測站分布不規則，特別是當偏移測站具有與周圍不同的地下地質條件時，二維反演的結果會進一步劣化，而此種情況對一維反演的影響則相對較小。本研究的發現為城市環境中進行的大地電磁測勘提供了實務參考，有助於評估測站可及性與地下成像品質之間可接受的取捨關係。這些成果旨在為持續調整大地電磁測勘方法，以因應持續擴張之現代城市環境的實際需求，貢獻一份力量。;With the rapid growth of urban development， the Magnetotelluric (MT) method has seen increasing application in subsurface investigations， ranging from hydrogeological studies to geothermal energy exploration. As urban areas continue to expand， however， urbanization also introduces significant challenges to MT survey design and data quality. The growth of urban infrastructure not only generates elevated levels of electromagnetic (EM) noise that obscure natural MT signals， but also reduces the availability of suitable measurement sites， often forcing station relocations that compromise spatial resolution and， in turn， the quality of subsurface imaging. Despite the practical relevance of these issues， their influence on MT results remains insufficiently characterized， and a clearer understanding of how these factors affect survey outcomes is still needed. In response to this gap， this study presents a simulation-based approach to examine how urban-related factors affect MT outcomes under conditions representative of real survey environments. Specifically， the study focuses on two aspects commonly encountered in urban MT surveys， MT response under noisy conditions and the effect of site spacing on inversion results. To address the first aspect， synthetic MT signals were generated using a 1D forward modeling approach and subjected to varying noise conditions， with the resulting responses evaluated against a known ideal response from the forward model. To address the second aspect， the effect of site spacing was investigated by applying 1D and 2D inversion under varying station distance configurations. The results show that the long period segment of the transfer function is generally most vulnerable to noise， observed as scatter in apparent resistivity and phase， with this scatter expanding into higher frequencies as noise levels increase. Transient noise causes more damage than Gaussian white noise. Building on this， the evaluation of site spacing effects on inversion results shows that 1D inversion is largely unaffected by site spacing， since each site is inverted independently， although it remains vulnerable to noise at individual sites with no neighboring data available to correct it. 2D inversion， by contrast， is clearly sensitive to spacing， with denser spacing improving structural resolution but introducing artifacts under noisy conditions， while wider spacing gives more stable but less detailed results. Irregular site distribution further degrades the 2D inversion， particularly when displaced sites carry different subsurface geology， a condition that 1D inversion tolerates well. The findings from this study provide a practical reference for MT surveys conducted in urban environments， helping to estimate acceptable trade-offs between site accessibility and subsurface imaging quality. These outcomes are intended to contribute to a broader effort to adapt MT surveying methodology to the realities of modern urban environments that continue to grow. <br> Tue, 20 Oct 2026 07:32:32 GMT https://ir.lib.ncu.edu.tw/handle/987654321/110336 title: 應用地震 —重力順序反演與機器學習自動波相拾取：探討2025年大埔地震之孕震構造及速度時變特徵;Seismic-Gravity Sequential Inversion and Machine-Learning Automatic Phase Picking: Seismogenic Structure and Temporal Velocity Variations of the 2025 Dapu Earthquake abstract: 台灣西南部麓山帶位於弧陸碰撞造山帶最前緣，孕震潛勢高且構造複雜。2025年1月20日該區發生ML 6.4大埔地震序列，其破裂與深部隱伏斷層系統有關，傳統地震走時層析成像受近地表射線多近乎垂直入射、缺乏水平交會覆蓋之限制，於淺層解析度嚴重不足。本研究採用結合地震走時與布蓋重力異常之順序反演流程：以中央氣象署2012 – 2020年背景地震目錄進行雙差分層析成像（TomoDD），透過速度—密度轉換與三維重力反演交互迭代，建立兼具深部穩定性與淺層高解析度之三維基準模型；以EQTransformer進行波相拾取、GaMMA進行波相關聯，自動化建置高完整度之餘震目錄，反演震後速度模型以解析主震前後之速度變化。在淺層構造方面，重力資料有效補足水平射線覆蓋之不足，將原本均勻之區塊狀構造重新解析為高、低速相間構造，與活躍褶皺逆衝帶之地表構造分布相符，並抑制淺層 VP⁄VS 之數值假性異常；速度與密度兩組獨立資料達成良好之物理一致性。構造剖面顯示，木屐寮—六甲斷層對應清楚之高低速邊界，標示褶皺逆衝帶與海岸平原之構造界面；潮州斷層對應位置處於低速區塊內，未見明顯速度邊界，反映該斷層帶周圍變形較為彌散。在孕震構造方面，重新定位之餘震集中於8 – 18 km處，深度超出西部麓山帶典型薄皮構造滑脫面之作用範圍；沿A–A’與B–B’剖面均可識別出東傾與西傾兩組地震分布，構成具空間連續性之共軛幾何，餘震主要位於 VS 高低速梯度邊界處，反映破裂受剛—弱物性邊界控制之深部隱伏共軛破裂系統。在震後速度時變方面，差值剖面於餘震分布區內識別出局部 VP 不變或些許上升、 VS 些許下降、 VP⁄VS 上升之異常區塊，指示破裂帶內局部流體滲入之飽和裂隙區域，其空間分布之不均質特徵可能反映震後流體遷移受破裂帶滲透率異質性所控制，顯示同震破裂、孔隙壓變化與震後流體再分布之耦合作用。本研究證實地震—重力順序反演能顯著提升孕震區淺層解析度，並為大埔地震之深部共軛破裂幾何與震後裂隙／流體作用提供具物理一致性之觀測約束。;The Western Foothills of southwestern Taiwan， situated at the leading edge of the arc–continent collision orogen， constitute a structurally complex and highly seismogenic environment. On 20 January 2025， the ML 6.4 Dapu earthquake sequence ruptured a deep blind-fault system in this region. Conventional seismic travel-time tomography resolves shallow structures poorly because near-vertical ray incidence and the scarcity of crossing rays leave the shallow crust under-constrained. This study applies a sequential inversion workflow integrating seismic travel times with Bouguer gravity anomalies. Long-term background seismicity recorded by the Central Weather Administration (2012 – 2020) drives double-difference tomography (TomoDD)， which is iteratively coupled with velocity–density conversion and three-dimensional gravity inversion to build a baseline model that is stable at depth and sharp at shallow levels. A high-completeness aftershock catalog， constructed automatically through phase picking by EQTransformer and phase association by GaMMA， is then inverted to obtain a post-mainshock model for resolving temporal velocity changes. Regarding shallow velocity structure， the gravity data effectively compensate for the absent shallow crossing rays， resolving formerly uniform blocky patterns into along-strike alternating high- and low-velocity bands consistent with the surface structural distribution of the active fold-and-thrust belt， while suppressing spurious VP⁄VS anomalies; the physical consistency between the independently derived velocity and density models enhances the credibility of shallow structural interpretation. The velocity cross-section reveals a clear high-to-low velocity boundary at the Muchiliao–Liuchia Fault， marking the structural interface between the fold-and-thrust belt and the coastal plain， whereas the Chaochou Fault corresponds to a low-velocity zone without a distinct velocity boundary， consistent with diffuse deformation and its low current activity. Regarding the seismogenic structure， relocated aftershocks concentrate at depths of 8 – 18 km in the upper-to-middle crust， exceeding the depth range typical of thin-skin detachments in the Western Foothills; both cross-sections reveal east- and west-dipping linearly distributed seismicity forming a spatially continuous conjugate pattern， with aftershocks predominantly located along VS high-to-low velocity gradient boundaries， indicating a deep， basement-involved blind conjugate rupture system controlled by competent–weak material boundaries. Regarding post-seismic velocity changes， difference cross-sections within the aftershock zone reveal localized anomalies characterized by unchanged or slightly increasing VP， slightly decreasing VS， and positive ΔVP⁄VS ， interpreted as fluid-saturated fracture patches resulting from post-seismic fluid infiltration; the spatially heterogeneous distribution of these anomalies may reflect post-seismic fluid migration controlled by permeability heterogeneity within the rupture zone， indicating coupled co-seismic fracturing， pore pressure change， and post-seismic fluid redistribution. This study demonstrates that seismic–gravity sequential inversion markedly improves shallow resolution in seismogenic zones and provides physically consistent observational evidence for the deep conjugate rupture geometry and post-seismic crack/fluid processes of the Dapu earthquake. <br> Tue, 20 Oct 2026 07:32:03 GMT https://ir.lib.ncu.edu.tw/handle/987654321/110332 title: 地震滑移速率下合成斷層泥之溫度及孔隙水壓模擬;Numerical investigation of temperature and pore pressure evolution in synthetic fault gouge sheared at seismic slip rates abstract: 在地震滑移速率下，含水斷層帶會因滑動而產生大量摩擦熱，其導致的高溫，可能引發熱相關物理化學過程，並影響材料的滑移行為，同時孔隙水壓也會因此過程，持續受影響而變化，繼而改變斷層的摩擦強度。因此了解溫度和孔隙水壓動態演化將對深入了解斷層的摩擦行為相當有幫助。本研究結合高速旋剪實驗的剪應力數據和微觀構造觀察之證據 (如黏土礦物的熱崩解)，並以實驗過程量測的溫度作為制約，建立耦合熱水力化的溫度與孔隙水壓數值模型，探討滑移過程的動態弱化機制。同時研究究與過去發展的模型比較，透過熱源幾何設定和重新界定吸熱作用的發生範圍，改善了過去模型有初期溫度低估和後期摩擦面異常高溫等現象。本研究模擬顯示，在高嶺土、伊萊石和石英混合的斷層泥為材料的高速實驗中，不排水條件下摩擦熱引發孔隙水壓在滑移初期迅速飆升，熱增壓作用極為顯著，主導了斷層的急遽弱化。在排水條件下，滑動開始後4秒，可能水的逸散而導致樣品變得乾燥，有較劇烈碰撞形成細小顆粒，並引發顆粒流產生，主導後期弱化。滑移過程中溫度、孔隙水壓與礦物化學作用、水的相態變化等存在著強烈關聯，顯示了多物理場耦合模擬對於評估動態弱化機制與斷層穩定性具有重要性。;Under earthquake slip rates， a large amount of frictional heat can be generated along the localized slip plane in a saturated fault zone. The high temperature (T) and thermal-physicochemical processes affect the slip behaviors. At the same time， pore fluid pressure (Pf) changes during this process， which also varies the fault strength. Therefore， understanding the dynamic evolution of T and Pf is helpful for understanding the frictional behavior of faults. By combining the mechanical data and microstructure observations from high-velocity rotary shear experiments in Nguyen et al. (In revision)， this study built a numerical model. This model couples thermal-hydro-mechanical-chemical (THMC) processes and uses measured temperature as a constraint to help understand the dynamic weakening mechanism. Compared with previous models， we improved the underestimation of initial T and the abnormally high T on the slip surface in the later stage by adjusting the heat source geometry and redefining the zone of endothermic reactions. Also， by coupling the axial displacement data， we tried to add the effect of the axial strain for Pf. For the ternary minerals， under undrained conditions， the rapid rise in Pf caused by frictional heat is significant. Thermal pressurization is the main mechanism triggers the rapid weakening of the fault. Under drained condition， at 4 s of the slip， the sample may become dry due to the drainage of water， resulting in more grain crushing and comminution that form fine grains and trigger granular flow， which dominates the weakening in the later stage. The significant relationships among T， Pf ， thermal decomposition of clay minerals， and water phase changes during the slip show the importance of multiphysics coupling simulations in evaluating dynamic weakening mechanisms and fault stability. <br> Tue, 20 Oct 2026 07:31:50 GMT