Fast determination of megathrust rupture processes and an application to the great 2011 Tohoku-oki, Japan, earthquake : 판의 섭입지역 대단층 파열과정의 역산 및 2011년 일본 동북 대지진에 적용

Cited 0 time in Web of Science Cited 0 time in Scopus


이 준기
자연과학대학 지구환경과학부
Issue Date
서울대학교 대학원
teleseismic broadband full-waveform inversionmegathrust earthquakemegathrust slip distributionfast and accurate determination3D finite fault modelsubducting slab geometryspectral element methodGreen’s function in 3D velocity modelthe great 2011 Mw9.0 Tohoku-Oki Japan earthquake
학위논문 (석사)-- 서울대학교 대학원 : 지구환경과학부(지구물리전공), 2014. 2. 이준기.
In order to mitigate disasters from devastating earthquakes and consequent large-scale tsunamis occurring in subduction zones, it is necessary to have early warning systems utilizing information from fast determination of rupture processes of the megathrust earthquakes. We develop a new method for fast and accurate estimation of megathrust rupture processes right after occurrence of the earthquakes with large magnitude, using broadband full waveforms of real-time recording from global seismic network. Based on the most up-to-date geometries of the subducting slabs seismic rupture planes in the region to the east of Honshu, Japan, a 3-D finite fault model is constructed by dividing the rupture plane of a megathrust region into subfaults with size of 50 km x 50 km with the same strike but different dips towards the subduction direction. Green functions at global network of stations from the source at each subfault with rupture rakes of 45˚and 135˚ in the laterally heterogeneous, 3-D earth model are computed using spectral element method (SEM) and the results are stored in a large memory system to be used later. Even though the Green functions in 3-D earth model using SEM are accurate, the computation can be time-consuming. Thus in order to reduce computational time done immediately after the event, most of time consuming parts of the computations such as the megathrust model construction and computations of Green functions in 3-D structure using SEM are done before the earthquake. Usually teleseimic full-waveform inversions for slip distributions are done based on focal mechanisms determined by the centroid moment tensor (CMT). An advantage of the method presented in this study over the conventional methods is that the method does not need priori information on the focal mechanism determined by CMT. It requires only hypocenter information from body wave analysis, since the fault plane and corresponding Greens functions were already computed and stored before the earthquake. Immediately after retrieving recorded data at occurrence of a megathrust earthquake, the rupture process can be computed in a short time by least-square inversion technique using the Greens functions. Another advantage is the 3-D fault model and 3-D velocity model used in this method could give an inversion result more accurate than or, at least, comparable to those of conventional methods based on a 2-D fault model from the CMT information and a velocity model of 1-D or 3-D. The inversion method presented in this study can give a solution independent of those based on CMT information. The inversion process is tested for synthetic seismograms as an artificial data set. Then the method is applied to the great 2011 Mw9.0 Tohoku-Oki, Japan earthquake. The results are favorably compared with the inverted slip distributions on the single-dip fault plane based on CMT solution.
Files in This Item:
Appears in Collections:
College of Natural Sciences (자연과학대학)Dept. of Earth and Environmental Sciences (지구환경과학부)Theses (Master's Degree_지구환경과학부)
  • mendeley

Items in S-Space are protected by copyright, with all rights reserved, unless otherwise indicated.