| dc.description.abstract | Serious seismic disasters may highly relate to basin amplification effects caused by the shallow earthquake or by the passage of seismic waves with rather large strong ground motion generated from large earthquake occurred distance away. Although the epicenters of these damage earthquakes are far away from the basin, the intensity within the basin is even larger than the area near the epicenter. Such amplification of seismic energy within the basin region is well-known because most seismic waves can be trapped by the basin geometry, topography and the extremely low velocity layer sitting on top of Taipei basin. Among all the previous studies, we provide relatively new approach by using the 3D hybrid discontinuous-grid finite difference computational method developed in-house to simulate the seismic waves propagating within the Taipei basin. In addition, we also implement the method of horizontal-to-vertical spectral ratio, which has been widely used by the engineering seismology, to analyze site amplification in and around Taipei basin. The main goal is to investigate the factors affecting seismic resonant frequencies (SRF) and spatial variation of amplification from 3D effects. The HVSR analysis is applied to both synthetic and Sishou Hills earthquake data collected from TSMIP of Central Weather Bureau, Taiwan.
From our 3D simulation studies, basin amplification and HVSR results are mainly affected by the basin shape, spatial variation of sedimentary layer thickness and the possession of low velocity feature carrying by the Songshan formation. Topography effect is not included in our 3D computation as strong motion records collected from Sishou Hills earthquake is mainly from stations located within Taipei basin where rather flat free-surface with elevation high of no more than ten to fifteen meter above sea surface. Full 3D waveform simulation approach automatically includes effects from 1D and 2D responses.
Over all speaking, the results from HVSR analysis of real data has lower SRF of 1.0Hz while the synthetic one poses a dominant feature of having higher SRF of 2.0Hz. Such difference is mainly attributed to the insufficient model specification on Q factors, more accurate and high resolution 3D reference velocity model of Taipei basin and lacking of highly correlated reflection, refraction and scattering signals from deeper layers in our 3D simulation. An added factor is that synthetic data is fairly narrow band in contrast to real one which has relatively wide frequency contents. Sum over all synthetic frequency responses improve the low frequency contents and more reasonably fits the real data records. Comparing HVSR analysis results from different frequency band, the resonance frequency around 0.5, 1.0, 2.0 and/or 2.5 Hz do exist in all calculations. The 2.0Hz peak value may well corresponds to the southeast (SE) part of Taipei basin where Songsan formation becomes shallower. Such response mainly corresponds to the 40 m layer thickness exist in the SE of Taipei basin where relatively simplified 1D layer model produces main wave modes due to the resonance of waves existing within the region. Providing that source radiation may slightly affect the HVSR analysis, the investigation on fundamental resonant frequency is still feasible. Shallow earthquake produces more complicate waveform records compare with simulated records from deeper one.
The results from 3D seismic simulation and site response analysis show that further detailed investigation on basin edge, lateral variation of 3D shallow velocity in the Songsan Fm., accurate shape of basin basement, more detailed velocity distribution for the deeper layers and nonlinear waveform simulation scheme are the imperative research targets in the future. | en_US |