利用人工震源與地震資料探究震波均向性與非均向性之層析成像

Isotropic and Anisotropic Seismic Tomography Using Active Source and Earthquake Records

林侑頻(Yu-Pin Lin)

指導教授 ： 趙里

共同指導教授 ： 洪淑蕙(Shu-Huei Hung)

國立臺灣大學/理學院/地質科學研究所/博士(2014年)

https://doi.org/10.6342/NTU.2014.01383

摘要

本論文包含兩個區域地震層析成像的研究：其中一個是利用人工震源的初動到時探討台灣下方地殼之P波均向性速度構造，另一個則是利用SKS分離強度的資料探討南加州下方剪力波非均向性構造。針對台灣下方的均向性層析成像，我們利用臺灣地體動力整合研究計畫 (TAiwan Integrated GEodynamics Research，簡稱TAIGER) 中所提供的資料進行分析，此研究計畫涵蓋一系列地球物理探勘實驗，其中包含了於2008年實施的陸地人工震源實驗，此實驗於南臺灣和北臺灣的東西向橫貫公路上施放了各五個爆破源，該測線上約每兩百公尺設置一地震儀，因此透過此實驗可獲得大量高品質的地震紀錄，能夠非常精確的解析出台灣下方的地殼速度。本研究一開始先以這些高品質的資料探討臺灣現有速度模型的精確程度及不足之處，其結果顯示三維速度模型所解析出的速度擾動強度仍然不足，因此我們利用這些炸測資料的初達波去推求四個橫跨臺灣的測線下方以及測站分布較密的北臺灣下方的地殼速度構造，利用試誤法的概念，不斷的去改變速度模型，並且透過波線追跡計算估計的到時再與觀測資料的初動到時比對，以找到最佳的二維模型。在測線較為密集的北臺灣，我們利用Partition Modelling 將二維模型外差為三維模型，此結果也和前人研究所得到的三維速度層析成像模型做比較，我們的結果改善了速度擾動強度的不足，其中較細微的速度分層及變化也是前人研究中因逆推方法的限制而無法得到的結果，除此之外側向上的速度變化也和地表的地質特徵極為相關。此研究提供了一個更貼近於真實構造的速度模型，在未來的研究中，可以將此模型當作參考模型，進而推測出更為精確的三維速度構造。
針對南加州非均向性層析成像的研究，我們的目的是解析出在上部地函非均向性的側向變化，可以幫助我們了解地球內部的地幔流場。本研究將我們發展的全波形方法應用到南加州地區，由於南加州地震中心 (Southern California Earthquake Center, SCEC) 擁有包覆性相當完整的地震測站，可提供高品質地震資料，非常適合進行三維非均向性結構的解析與探討。SKS剪力波分離是證實介質非均向性最有利的觀測，目前為止大部分解釋剪力波分離的結果都是基於波線理論，將SKS或 SKKS波視為一垂直入射的平面剪力波，但真實地震訊號與此假設並不一致，因此我們能夠使用的資料將會有所限制，也將我們能夠解析的非均向性結構局限為均勻分佈。而本研究採用近期發展的全波形三維非均向性逆推方法，以高效率且高靈活度的算法計算逆推中的敏感度積分核 (sensitivity kernel)，這種積分核可以精確的考慮到所有波相之間互相干擾的影響，因此對於剪力波分離的測量將不再受限於特定的震央距範圍 (如90°-120°)，我們則可以使用更多的剪力波分離資料進行精確的模型逆推。本研究針對震央距90°-130°的SKS訊號進行測量，並且採用多重尺度逆推 (multi-scale inversion)，其做法是利用小波轉換對模型進行多重尺度參數化，模型在空間的變化透過小波轉換拆解成不同尺度以進行逆推，由觀測資料自行決定在空間中的解析度，因此在波線覆蓋較差的地方仍然可以解出大尺度的構造。另外我們利用統計解析度矩陣 (statistical resolution matrix) 的方法去評估側向的解析度，結果顯示能夠解析的長度大約為25公里。三維非均向性模型於岩石圈中所顯示的快軸方向皆與地表的地質構造相關，如Salton Trough，Transvers Ranges和San Andreas Fault，而深度上的快方向變化則說明岩石圈和軟流圈並沒有明顯裂解的現象，長波長的構造顯示了其非均向性的快軸方向大致與太平洋板塊和北美板塊的絕對板塊運動方向一致。

關鍵字

TAIGER計畫 ； 人工震源 ； 初達波 ； 地震層析成像 ； 臺灣 ； 剪力波分離 ； 非均向性 ； 多重尺度層析成像 ； 全波形敏感積分核 ； 南加州

並列摘要

This thesis involves two regional seismic tomography studies: One for the isotropic P-wave velocity in the upper crust beneath Taiwan using first-arrival times from active sources; the other for the shear-wave anisotropy structure under Southern California using SKS splitting intensity measurements. In the isotropic tomography for Taiwan, we use waveform records from the 10 explosions in 2008 conducted by the TAiwan Integrated GEodynamics Research (TAIGER) project. A large dataset of high-quality ground-truth first-arrival times are hand-picked from the active-source records at ~1400 sites throughout Taiwan, which greatly enhance our capability to determine the crustal velocity beneath Taiwan with unprecedented accuracy and resolution., especially along the two north-south and two east-west island-wide linear transects with densely-deployed receivers. At first, these first-arrival times are used to evaluate the existing tomography models for Taiwan. Results show that tomography models obtained from traditional travel time inversions provide consistent and qualitatively correct locations of larger-scale velocity perturbations. However, small-scale features are inconsistent among different models, and their velocity perturbations are mostly underestimated. Then we use our ground-truth first-arrival times to refine the P-wave velocity model. With a trial-and-error procedure, we acquire the best 2D models along a number of shot-to-station profiles by fitting the first-arrival times. Finally a partition modelling approach is employed to invert for a 3D model in northern Taiwan based on a collection of the crisscrossing 2D models that densely transect across the region. The resulting structural variations in our 3D model correlate remarkably well with the surface geological features that are distinctly shaped by the orogenic and tectonic history in Taiwan.
In the anisotropic tomography for Southern California, our purpose is to resolve the spatial variation of anisotropy in the upper mantle which plays an important role in our understanding of the Earth’s internal dynamics. Shear-wave splitting has always been a key observable in the investigation of upper-mantle anisotropy. However, the interpretation of shear-wave splitting in terms of anisotropy has been largely based on the ray-theoretical modelling of a single vertically incident plane SKS or SKKS wave. In our study, we use sensitivity kernels of shear-wave splitting to anisotropic parameters calculated by the normal-mode theory, which automatically accounts for the full-wave effects including the interference of SKS with other phases of similar arrival times, the near-field effect, and multiple reflections in the crust. These full-wave effects can lead to significant variations of SKS splitting with epicentral distance and are neglected in ray theory. We image the upper-mantle anisotropy in Southern California using nearly 6000 SKS splitting data and their 3D full-wave sensitivity kernels in a multiscale inversion enabled by a wavelet-based model parameterization. We also appraise our inversion by estimating the spatial resolution lengths using a statistical resolution matrix approach, which shows the finest resolution length of ~25 km in regions with better path coverage. The anisotropic model we obtain displays the structural fabrics in relation to surface geologic features such as the Salton Trough, the Transverse Ranges and the San Andreas Fault. The depth variation of anisotropy does not suggest a strong decoupling between the lithosphere and asthenosphere. At long-wavelengths, the orientations of the fast axis of anisotropy are consistent with the absolute plate motion in the interiors of the Pacific and North American plates.

並列關鍵字

TAIGER project ； active sources ； first-arrival times ； seismic tomography ； Taiwan ； shear-wave splitting ； anisotropy ； multi-scale tomography ； full-wave sensitivity kernels ； Southern California

參考文獻

An, M. (2012), A simple method for determining the spatial resolution of a general inverse problem, Geophys. J. Int., 191, 849-864.

Ando, M. (1984), ScS polarization anisotropy around the Pacific Ocean, J. Phys. Earth, 32, 179-195.

Argus, D. F., R. G. Gordon, and C. DeMets (2011), Geologically current motion of 56 plates relative to the no-net-rotation reference frame, Geochem. Geophys. Geosyst., 12, Q11001, doi:10.1029/2011GC003751.

Becker, T. W., S. Chevrot, V. Schulte-Pelkum, and D. K. Blackman (2006), Statistical properties of seismic anisotropy predicted by upper mantle geodynamic models, J. Geophys. Res., 111, B08309, doi:10.1029/2005JB004095.

Bird, P. (2003), An updated digital model of plate boundaries, Geochem. Geophys. Geosyst., 4(3), 1207, doi: 10.1029/2001GC000252.