為了解析斷層帶的衰減構造， 本研究利用 DAS 觀測資料分析 MiDAS 井 A 沿線 的衰減因子 Q， 並以井下地震儀進行驗證。 在開始高解析度衰減計算前， 首先挑選 6 個震央距觀測點約 20 km 的地震事件， 比對對應深度處之井下地震儀與光纖資料。 我們將加速度與應變率資料分別積分為速度與應變， 透過射線追蹤（Raytracing）篩 選出入射角小於 30 ◦ 的事件。 接著根據 P、S 波到時與 Effective Shacking Duration （ESD）劃分出 P 波、S 波及噪訊時窗， 進行頻率域分析。 最後以觀測站間的頻譜比 （Spectral ratios）進行最小平方法擬合 ∆t ∗ p 與 ∆t s ∗ ，藉此計算衰減參數 Q p 與 Q s 。分 析顯示 DAS 及井下地震儀得出之結果高度一致， 且 Q s 穩定地比 Q p 大。 本研究進 而將分析規模擴大至 1，000 多個地震規模（M L ）介於 2.0 至 4.5 的區域地震。 透過 在龐大的 DAS 資料上應用滑動時窗分析， 成功解析出井 A 各深度的衰減（Q）剖 面， 並釐清斷層帶內的衰減變化。 研究結果顯示， 井 A 多數深度的 Q 值普遍偏低 （Q p ≈ 10 − 20，Q s ≈ 20 − 30），且 Q s /Q p 比值約為 1.5，與在 DAS 與井下地震儀的 六個事件分析中皆獲得一致的結果。 此外， 透過 13 個遠場地震（M L ≥ 6.0， 震央距 100–1000 km）在較低頻段（0.1–20 Hz）內探討衰減行為，本研究發現計算出的遠場衰 減參數（Q far-field ≈ 6 − 9）顯著小於近場區域地震所得之衰減（Q near-field ）。此一差異支持斷層帶內存在頻率依賴性衰減（Frequency-dependent attenuation）。 本研究也針對針對 2024 年 4 月 3 日花蓮地震前後進行比較，部分深度的 Q p 和 Qs 在主震後呈現降低、 衰減增強的趨勢， 但未觀察到沿整個井深一致且大幅度的改變。 最後，本研究透過頻率-波數法（FK 分析）進行合成波形模擬，依照預設 Q 值產生合 成波形，並套用相同的頻譜比分析流程。證實本研究中 Q 值估算結果及斷層帶衰減分 析的穩健性與可靠性。本研究發現 MiDAS 井 A 呈現廣泛而強烈的低 Q 構造，反映米 崙斷層淺部介質具有高度破碎、裂隙連通及可能受流體作用影響的特性；DAS 與井下 地震儀的整合，亦建立了持續追蹤斷層帶空間與時間衰減變化的觀測基礎。;Fault zone architecture offers critical insights into earthquake dynamics. The structural components of a fault zone， including the fault core， damage zones， and surrounding host rocks， play pivotal roles in seismic behavior. The quality factor (Q)， representing seismic attenuation (Q −1 )， is inversely related to the degree of fracturing within these zones; higher fracture densities typically correspond to lower Q values， indicating greater energy dissipation. Moreover， the permeability of these fractured zones facilitates fluid migration， which can alter fault strength and affect earthquake nucleation processes. This study focuses on the Milun Fault， an active east-dipping reverse fault with a left-lateral component located at the northern end of Taiwan’s Longitudinal Valley near the boundary between the Eurasian Plate and the Philippine Sea Plate. The fault was involved in several major earthquakes， including the 1951 Longitudinal Valley earthquake sequence and the 2018 Hualien earthquake. The Milun Fault Drilling and All-inclusive Sensing Project (MiDAS) initiated scientific drilling in 2021 and subsequently integrated distributed acoustic sensing (DAS) with a borehole seismometer array. MiDAS Hole A intersects the Milun Fault core at a depth of approximately 520–540 m， providing a unique opportunity to directly investigate the attenuation structure of a shallow active fault.
To characterize the detailed attenuation structure， the attenuation factor Q along MiDAS Hole-A is estimated using DAS records， with borehole seismometer data used for validation. Before exploring the high-resolution attenuation mapping within the fault zone using DAS (4 m spatial resolution)， six seismic events (roughly 20 km from the MiDAS site) were analyzed using paired data from borehole seismometers and fiber nodes at corresponding depths. The datasets are in acceleration and strain-rate， which are subsequently integrated into velocity and strain， respectively. The AR-Pick method was applied to identify seismic arrivals and ray tracing selected events with incidence angles below 30 circ . Arrival time and effective shaking duration (ESD) define P-， S-wave， and noise windows for frequency-domain analysis. Spectral ratios between stations were computed to estimate the attenuation parameters Q p and Q s from ∆t ∗ p and ∆t ∗ s by least-squares fitting. The results obtained from the DAS and borehole seismometers show a high degree of consistency， with Q s systematically greater than Q p . Building on this initial validation， the analysis was scaled to over 1，000 local events with magnitudes ranging from 2.0 to 4.5. By applying a sliding-window approach to this extensive DAS dataset， the study aims to resolve the high-resolution attenuation (Q) profile across the entire depth of Hole A and characterize the depth-dependent attenuation variations along the borehole and within the fault zone. The results indicate generally low Q values across most depths， with Q p ≈ 10–20 and Q s ≈ 20–30. The corresponding Q s /Q p ratio of approximately 1.5， consistent with the results obtained from the six-event comparison between the DAS and borehole seismometers. Furthermore， 13 regional， far-field events (M L ≥ 6.0， 100–1000 km from the site) were incorporated within a lower frequency band (0.1–20 Hz) to investigate macro-scale attenuation behavior. A systematic discrepancy was observed: the calculated far-field attenuation parameters (Q far-field ≈ 6–9) are notably smaller than those derived from local events (Q near-field ). This discrepancy provides compelling evidence for frequency-dependent attenuation within the fault zone.
Temporal variations in attenuation before and after the 3 April 2024 Hualien earthquake were also examined. At several depths， both Q p and Q s decreased following the mainshock， indicating locally enhanced attenuation. However， no large and systematic change was observed consistently throughout the entire borehole. Finally， synthetic waveform modeling was conducted using the frequency–wavenumber (FK) method with prescribed Q values. The synthetic waveforms were processed using the same spectralratio analysis workflow as the observations， validating the robustness of the estimated Q values and the reliability of the fault-zone attenuation analysis. Overall， MiDAS Hole A exhibits a pervasive and strongly attenuating low-Q structure， suggesting that the shallow Milun Fault consists of highly fractured materials with interconnected cracks and may be influenced by fault-zone fluids. The integration of DAS and borehole seismometers also establishes an observational framework for continuously tracking spatial and temporal variations in attenuation within an active fault zone.
