講演情報
[S06-06]Full Waveform Inversion of Extra-long Offset Seismic Data for Imaging 2024 Noto Earthquake (M7.6) Rupture Zone
*Ehsan Jamali Hondori1, Hamzeh Mohammadigheymasi2, Tatsuya Ishiyama3, Jin-Oh Park2 (1. Geoscience Enterprise Inc (GSE), 2. Atmosphere and Ocean Research Institute, The University of Tokyo, 3. Earthquake Research Institute, The University of Tokyo)
The 2024 Noto earthquake (M7.6) occurred in the Noto Peninsula and triggered a tsunami along the eastern coast of the Sea of Japan on January 1st, 2024. The distribution of the aftershocks denoted a possible reverse fault being the main source of the earthquake and extending over 150 km from western coast of the Noto peninsula to the northeast offshore Sea of Japan. A previous research in this area had been conducted a decade ago to study the structures and active tectonics of compressional reactivated faults in the back-arc in the Sea of Japan by using first arrival traveltime tomography of multichannel seismic data (Ishiyama et al., 2017). Despite their unique seismic velocity model and the corresponding depth profiles, the low resolution nature of the traveltime tomography method limits the quantitative interpretation of the subsurface. In contrast to the tomographic approach which only relies on the first arrival time, seismic full waveform inversion (FWI) uses all the information coming with seismic data, including both phase and amplitude. Moreover, it uses the whole seismic waveform from various wave paths with different features such as refractions, reflections, diffractions, and all the scattered energy in the subsurface. In this paper we apply 2D acoustic full waveform inversion on the same dataset which was acquired back in 2013 to build a higher resolution seismic velocity model, which makes the structural interpretation easier and gives a better seismic depth section compared to the traveltime velocity model.
The frequency domain implementation of the FWI allows to update the velocity model by only using a few discrete frequency components of seismic data. The multichannel seismic data which we used in this work had been acquired using a two-vessel strategy, one for towing the airgun seismic source another for towing a receiver cable with 156 channels spaced at every 12.5 m intervals (Ishiyama et al., 2017). The maximum source to receiver offset in this data is around 12 km, which makes it capable of imaging deep targets using FWI, the method which traditionally requires long offsets to allow diving waves to build the large-scale model. The preliminary results that we obtained show clear velocity contrast at the fault boundaries and enhance the resolution of the previous traveltime tomography velocity model. The enhanced resolution in our velocity model allowed us to not only cross-check the model with previously obtained seismic depth sections, but also deliver more accurate depth migrated images which enhance structural interpretation. In the future work we will apply various FWI strategies using this data, including a hierarchical cascading of source to receive offsets in the input data to see the effect on the FWI velocity model.
Reference
Ishiyama et al., 2017, Structures and active tectonics of compressionally reactivated back-arc failed rift across the Toyama trough in the Sea of Japan, revealed by multiscale seismic profiling, Tectonophysics, 710–711, 21–36.
The frequency domain implementation of the FWI allows to update the velocity model by only using a few discrete frequency components of seismic data. The multichannel seismic data which we used in this work had been acquired using a two-vessel strategy, one for towing the airgun seismic source another for towing a receiver cable with 156 channels spaced at every 12.5 m intervals (Ishiyama et al., 2017). The maximum source to receiver offset in this data is around 12 km, which makes it capable of imaging deep targets using FWI, the method which traditionally requires long offsets to allow diving waves to build the large-scale model. The preliminary results that we obtained show clear velocity contrast at the fault boundaries and enhance the resolution of the previous traveltime tomography velocity model. The enhanced resolution in our velocity model allowed us to not only cross-check the model with previously obtained seismic depth sections, but also deliver more accurate depth migrated images which enhance structural interpretation. In the future work we will apply various FWI strategies using this data, including a hierarchical cascading of source to receive offsets in the input data to see the effect on the FWI velocity model.
Reference
Ishiyama et al., 2017, Structures and active tectonics of compressionally reactivated back-arc failed rift across the Toyama trough in the Sea of Japan, revealed by multiscale seismic profiling, Tectonophysics, 710–711, 21–36.