17:15 〜 18:45
[SSS07-P03] マルチモード表面波の非線形波形インバージョン：振幅測定への応用
キーワード：表面波、振幅、非弾性減衰、マルチモード、非線形インバージョン
A new method of fully nonlinear waveform inversion of multimode surface waves is developed to measure the modal amplitude of surface waves, which can be used to constrain the anelastic attenuation structure in the mantle with enhanced vertical resolution. The phase speed and amplitude measurements of multimode surface waves are not straightforward issues since multiple modes overlap in the observed seismogram. However, through the full waveform fitting process with appropriate model parameters, we can extract the multimode dispersion information as indicated by earlier studies (e.g., Yoshizawa and Kennett, 2002; Yoshizawa and Ekstrom, 2010.)
Our method is an extended version of the method of multimode phase speed measurements originally developed by Yoshizawa and Ekstrom (2010), based on the fully nonlinear waveform fitting technique with a global optimization. The Neighborhood Algorithm (NA) is used as a global optimizer. In the case of phase speed measurements, we employed the path-specific 1-D S wave speed profile as an unknown model parameter to fit the phase of observed and synthetic waveforms. The best-fit 1-D S-wave profile can then be used to extract the multimode phase-speed dispersion curves. Surface-wave WKBJ theory is used to compute the synthetic seismograms based on the Earth's normal modes, which expedites the computation speeds for repeated forward modeling.
In our new method of multimode amplitude measurements, we employ the path-specific 1-D Q profile to represent the multimode amplitude variations. Over 5000 1-D Q profiles are searched using the NA by fitting the amplitude of synthetic and observed seismograms. The best-fit Q profile is then used to extract the modal amplitude information based on the formulation of anelastic attenuation term in the WKBJ synthetic seismogram. A variety of synthetic experiments suggest the utility of the method to enhance the vertical resolution of Q models and to extract reliable and stable amplitude information of the fundamental mode and several higher modes.
The method has been applied to some observed seismograms for oceanic and continental paths to extract the path-specific Q model and modal amplitude. The results indicate that the resultant Q profiles exhibit a high Q in the lithospheric depth, while the Q values decrease rapidly in the asthenospheric depth. The extracted modal amplitude information from this method can be used as the basis for future analysis of anelastic attenuation tomography, incorporating multimode amplitudes.
Our method is an extended version of the method of multimode phase speed measurements originally developed by Yoshizawa and Ekstrom (2010), based on the fully nonlinear waveform fitting technique with a global optimization. The Neighborhood Algorithm (NA) is used as a global optimizer. In the case of phase speed measurements, we employed the path-specific 1-D S wave speed profile as an unknown model parameter to fit the phase of observed and synthetic waveforms. The best-fit 1-D S-wave profile can then be used to extract the multimode phase-speed dispersion curves. Surface-wave WKBJ theory is used to compute the synthetic seismograms based on the Earth's normal modes, which expedites the computation speeds for repeated forward modeling.
In our new method of multimode amplitude measurements, we employ the path-specific 1-D Q profile to represent the multimode amplitude variations. Over 5000 1-D Q profiles are searched using the NA by fitting the amplitude of synthetic and observed seismograms. The best-fit Q profile is then used to extract the modal amplitude information based on the formulation of anelastic attenuation term in the WKBJ synthetic seismogram. A variety of synthetic experiments suggest the utility of the method to enhance the vertical resolution of Q models and to extract reliable and stable amplitude information of the fundamental mode and several higher modes.
The method has been applied to some observed seismograms for oceanic and continental paths to extract the path-specific Q model and modal amplitude. The results indicate that the resultant Q profiles exhibit a high Q in the lithospheric depth, while the Q values decrease rapidly in the asthenospheric depth. The extracted modal amplitude information from this method can be used as the basis for future analysis of anelastic attenuation tomography, incorporating multimode amplitudes.