透過您的圖書館登入
IP:3.16.147.124
  • 學位論文

以自動化全波形地震矩逆推法與多重訊號分類反投影法探討地震破裂特性

AutoBATS and 3D MUSIC: New Approaches to Imaging Earthquake Rupture Behaviors

指導教授 : 洪淑蕙
若您是本文的作者,可授權文章由華藝線上圖書館中協助推廣。

摘要


地震發生時,除了發震時間、位置及規模,地震學家對於地震的破裂過程尤其重視,其中地震的規模與斷層型態說明每次地震釋放能量的大小以及是以何種錯動方式釋放,而地震的破裂方向、長度、時間與速度等則可闡述地震發生時的運動特性。本論文為了對地震物理特性有更多了解,先是針對台灣區域地震提出了穩定與全面性掃描的震源機制逆推,接著以三維的多重信號分類演算反投影法(MUSIC Back Projection) 分別研究了2013年鄂霍克次海的深震與在台灣造成死傷的美濃地震破裂行為。最後結合新的台灣區域地震震源機制目錄與西南台灣甲仙、美濃地震的地震破裂特性進一部探討此兩個造成死傷的地震的相互關係。 自從全球或區域地震網快速發展後,成就了大量的地震研究,尤其是系統化地建立全球或區域的震源機制目錄。台灣自1995年設置寬頻地震網以後,便定期提供震源機制目錄,然而以人工選取資料和參數的地震矩逆推法既耗時亦無法迅速提供防災資訊。本論文提出了以自動化方式同時全面掃描不同逆推設定進行地震矩逆推而得到最後的震源機制解,逆推設定包括三種選站策略、濾波頻率及四組不同莫合面深度的速度構造。並以最新1995至2016年的地震矩震源機制目錄得出新的地震矩與芮氏規模的關係式。此全新自動化逆推將為台灣地震科學訊息中心提供迅速並且穩定的震源機制解。 近來,世界各地為了求得更高解析度的地球物理研究,例如地球構造、地震行為、以及防災需求等而設置了高密度地震網。高密度地震網更促進了地震破裂影像研究,其中,MUSIC BP是陣列地震學(Array Seismology)當中的最新應用:無需預設地震的斷層型態,直接以反投影方法分析地震破裂時的時間、空間演變狀態。本論文,藉由結合P波與pP或sP波的反投影影像使該方法得以應用於三維空間。我以三維反投影法得出2013年規模8.3鄂霍次克海深震是由兩個深度差了10~15公里的水平破裂所組成,兩組破裂彼此間是接近平行而往相反方向傳遞,其中往東北方向的破裂以3.0~3.3 km/s的速度延伸約30~40公里,往南南西的破裂長約為80公里,破裂速度增加為4.25~4.8 km/s。推測往東北方向較短且較慢的破裂與太平洋板塊北端因受軟流圈角落流或是背景地幔加溫有關。 台灣西南部於2010及2016年分別發生了造成死傷的甲仙與美濃地震,本論文首度嘗試以反投影法解析規模6左右位於中地殼(mid-crust)的地震破裂特性,除了結合P波與pP/sP波的反投影影像,我們也分析全球的P波位移波形,藉由地震破裂具有方向性之性質來驗證反投影法影像。經過分析,我們發現這兩個地震有許多相似之處,包含接近的震央位置、震源機制,並且兩起地震以相近的破裂(2.0~2.5 km/s)速度往西北西方向破裂,因此我將兩起地震視為一組雙震(doublet)。反投影法影像結合餘震分布,我發現兩個地震的破裂區域是分開並且呈現互相平行。進一步大膽假設此雙震發生於一個位在中地殼的盲斷層上,而該區域較低的b值暗示該斷層長期累積應力,而無論是甲仙或是美濃地震之後,另一個破裂區域都出現震後庫倫應力大幅增加卻無任何餘震被觸發,因此推測兩個破裂區域屬於彼此獨立分開而且是斷層上兩個強補釘區域(asperity)。綜合破裂特性、b-值及庫倫應力計算,此雙震系列由較深的補釘率先破裂釋放長期累積的應力,接著較淺的補釘或許是被觸發卻延遲而發生了更大規模的美濃地震。

並列摘要


The key knowledge of what happened during an earthquake includes the fault types, seismic energy and the rupture process. The moment magnitude and the focal mechanism tell how much and how an earthquake releases the energy. The kinematic properties of an earthquake moreover represent the rupture length, duration, propagation and speed. In this thesis, I dedicate on understanding the earthquake rupture properties by developing new automatic moment tensor (MT) inversion scheme for compiling reliable regional MT solutions and applying MUltiple SIgnal Classification Back Projection (MUSIC BP) for earthquake rupture imaging including the 2013 Okhotsk deep earthquake and 2016 catastrophic Meinong earthquake in Taiwan. With the aid of AutoBATS MT catalog and MUSIC BP method, I further probe in to explore interacts of 2010 Jiashian and 2016 Meinong earthquakes. The global/regional seismic networks facilitate the explosion growth on the earthquake studies especially on systematically compiling global/regional focal mechanism catalogs. In 1995, Taiwan participated in developing the Broadband Array for Taiwan Seismology (BATS) network and determining the regional CMT catalog (BATS CMT) routinely but manually. Hence I develop the new automatic MT inversion algorithm which provides more reliable and stable MT solutions by performing the MT inversion with comprehensive scanning of inversion settings of three station-choosing criteria, frequency bands and four moho-depth velocity models. A new MW-ML relation is updated with the complete AutoBATS MT catalog from 1995 ~ 2016. The new near real-time AutoBATS MT algorithm is employed for providing real time MT solutions for Data Management Center of IES (DMC-IES). The ambitions of studying earthquakes, earth structures, especially the earthquake engineering for seismic hazard prevention expedite the worldwide/regional dense seismic networks development and push the earthquake source imaging into superior high resolution level. The MUSIC BP, a new application of array seismology, requires no prior fault plane information to reveal the spatiotemporal rupture propagations for earthquakes. Here, I adopt the MUSIC BP method to three dimension and improve the spatial resolution effectively by integrating the BP images of direct P wave and depth phases. The 3D BP images of P- and pP-waves reveal the complex rupture behavior of 2013 great Okhotsk deep earthquake: two-stages of anti-parallel subhorizontal ruptures occurring at different depth. The depth aperture between two ruptures is about 10~15 km. The 1st rupture stage propagated NE-ward with speed of 3.0~3.3 km/s extending for about 30~40 km. Then the SSW-direction deeper rupture elongated for about 80 km with speed of 4.25~4.8 km/s. The initial slow and restricted NE-ward rupture may relate with the warmer northern tip of the Pacific slab which is heated by the ambient mantle and corner asthenosphere flow. Contributed by the capability of integrating BP images of depth phases with P-wave, the rupture characteristics of two Taiwan mid-crust earthquakes with magnitude 6+ could be revealed for the first time: the 2010 Jiashian and 2016 Meinong earthquakes. In addition, the directivity analysis of global vertical displacement waveforms is included to ascertain the BP results of two earthquakes. I concluded that both earthquakes rupture toward NW-direction with comparable rupture speed of 2.0~2.5 km/s and two closed by rupture zones along with the corresponding aftershock distributions are well separated. From the similar focal mechanisms, rupture properties and close epicenters of two earthquakes, I infer those as the SW-Taiwan doublet events which might occur on a mid-crust unknown blind fault. The overall low b-value of the rupture area implies that this strong blind fault has been sustained the stress accumulation for at least 20 years. Two isolated parallel rupture zones are considered as two strong asperities because neither the Jiashian nor Meinong earthquakes induce any aftershock activities across to the other rupture zone where inherited great increase of static stress change after the mainshock. In summary, the earthquake rupture properties, b-value and Coulomb stress change studies support the assumption that the deeper asperity (2010 Jiashian event) initiated the stress releasing process and might delaying trigger the shallower asperity (2016 Meinong event) to further release the long-time accumulated stress.

參考文獻


Aki, K. (1965). Maximum likelihood estimate of b in the formula log N= a-bM and its confidence limits, Bull. Earthq. Res. Inst., Tokyo Univ., 43,, 237–239.
Aki, K., and P. G. Richards (1980). Quantitative seismology: Theory and methods, Freeman, San Frrancisco.
Allen, R. V. (1978). Automatic earthquake recognition and timing from single traces, Bull. Seismol. Soc. Am., 68,no.5, 1521–1532.
Angelier, J., T.-Y. Chang, J.-C. Hu, C.-P. Chang, L. Siame, J.-C. Lee, B. Deffontaines, H.-T. Chu, and C.-Y. Lu (2009). Does extrusion occur at both tips of the Taiwan collision belt? Insights from active deformation studies in the Ilan Plain and Pingtung Plain regions, Tectonophysics, 466,no.3-4, 356–376, doi: 10.1016/j.tecto.2007.11.015.
Antolik, M., D. Dreger, and B. Romanowicz (1996). Finite fault source study of the great 1994 deep Bolivia earthquake, Geophys. Res. Lett., 23,no.13, 1589–1592, doi: 10.1029/96gl00968.

延伸閱讀