MODELLING OF FINITE SOURCE AND MEDIUM CHARACTERIZATION

Abstract

The modeling of a finite source and medium characterization plays a significant role in investigating the impact of earthquake on the observed ground motion at the surface. The strong ground motion at the recording site are heavily dependent on the size of earthquake source and the characteristics of medium through which seismic energy propagates. The main objective of this thesis is to investigate the effect of finite source of the strong ground motion (SGM) observed at the surface and the effect of finite medium on propagating seismic energy released during an earthquake. A finite source earthquake denotes a rupture that occurs across a defined area along a fault plane rather than occurring instantaneously at one point. As a result, knowing the mechanisms causing finite source earthquakes is essential for both minimizing human casualties and damage to property and enhancing our knowledge of the features of Earth's interior. Finite source modeling is needed to examine the spatial extent and the distribution of the slip along and across the fault plane accountable for an earthquake. Among various available simulation techniques, Modified Semi Empirical Technique (MSET) has evolved as an effective tool to simulate SGM records in recent years. In the present thesis, MSET has been tested to model the strong motion generation areas (SMGAs) and frequency dependent radiation pattern (FDRP) effect in modeling the finite rupture plane. In this context, the modeling of the finite source of the 2019 Hualien earthquake (Mw 6.1) has been successfully validated the proposed technique. The epicenter of the earthquake is 24.037°N, 121.65°E, and focal depth is 20 km. The rupture length and its downward extension for the Hualien earthquake are 32 km and 18 km, respectively. The highest Peak Ground Acceleration (PGA) of 379 cm/s2 for the E-W component and 515 cm/s2 for the N-S component for a station at 12 km distance from the epicenter. Both the spatiotemporal distribution of the earthquake's aftershocks and a visual examination of the recorded acceleration data indicate the presence of two SMGA within the finite fault plane responsible for this event. The finite SMGAs have been divided up into several sub-faults using the scaling laws. The simulation technique relies on the source parameters of the finite fault plane of the earthquake, specifically, two identified SMGAs, as the SGM is primarily generated by these SMGAs, necessitating their computation for modeling purposes. The source displacement spectra (SDS) is used to compute seismic moments of both SMGAs which are (4.80 ± 0.98) x 1017 Nm and (5.25 ± 0.93) x 1017 Nm, respectively, and the sum of these seismic moments is equivalent to the seismic moment of the whole rupture plane i.e. 1.05×1018 Nm. The observed acceleration ecordings at the surface need to be projected at the rock site using SHAKE91 as MSET simulates the records at the rock site only. The final modeling parameters like nucleation point, dip, and strike have been decided based on iterative forward modeling at a nearest station. Based on the iterative forward modeling and the minimum RMSEPGA i.e. root mean square error of PGA, the nucleation point is selected at the extreme left corner of the rupture planes of both SMGAs. The strike and dip of the rupture planes of both SMGAs are selected as N215.5° and 45°, respectively. A comparison has been made between the PGA calculated using simulated accelerograms and that acquired from accelerograms observed at stations located within a 100 km radius around the epicenter. Additionally, the PGA distribution trend is presented with respect to the epicentral distance and compared with the PGA derived from the attenuation relation provided by Lin and Lee (2008). The contour plot trend of the N-S and E-W components, both simulated and observed suggests a northward directivity effect caused by the rupture, consistent with the findings by Lee et al. (2020).