5:15 PM - 6:30 PM
[SSS02-P03] Transversely isotropic velocity structure of marine sedimentary layers estimated by Rayleigh wave ellipticity of ambient noise
Keywords:Rayleigh wave ellipticity, seismic interferometry
Recent development of OBS networks such as S-net and DONET now enables us to measure ambient noise cross-correlations of high-resolution on the sea bottom. Spatiotemporal variations of seismic velocity and vertical transversely isotropy in marine sediment help us to understand dynamic processes associated with plate interaction in a subduction zone such as crack alignment caused by the bending of a subducting plate, fluid migration and dehydration of constituent minerals. The ellipticity of Rayleigh waves (i.e., the ratio of vertical and horizontal amplitudes, V/H) is useful to study seismic velocity of sedimentary layers because V/H is more sensitive to shallow structures than conventional phase or group dispersion data (e.g., Lin et al., 2014).
In this study, we first formulated partial derivatives of V/H for vertical transversely isotropic media including sea water. It is well known that P-wave velocity has little effect on the ellipticity in isotropic media. In the VTI media, however, the effect or distinguishment of PH and PV wave velocities cannot be neglected in comparison to conventional SV wave velocity in media of a certain degree of VTI. This is why, the partial derivative or sensitivity of P wave in isotropic media is the sum of those of PV and PH waves with opposite signs.
Next, since short-period Rayleigh waves have sensitivity for shallow parts, we applied the seismic interferometry to the noise data recorded at S-net stations. The cross-correlations among S3 stations perpendicular to the trench axis clearly show surface wave signals with relatively good symmetry of causal and anti-causal waveforms in the period range from 5 to 12 sec (Fig.1). Their group velocity is very small, about 0.2-0.4 km/s. This signals probably correspond to Scholte waves. Scholte waves propagate parallel to the water-sediment interface, when the shear velocity of the solid is lower than the compressional velocity in the water (1.5 km/s), that is, very loose sediments.
Finally, we estimated V/H spectrum ratios of the extracted Scholte waves at S-net stations and inverted a new velocity model of sedimentary layers. Prior to the V/H inversion, we estimated a velocity model of the basement beneath sediment by the measure phase velocity dispersion. The area in the subarray S3 of deep ocean near the Japan trench axis has small V/H values. This indicates slow velocity regions there, probably very and/or thick soft sediments. The estimated PV and SV wave velocities can explain these low V/H values in the subarray S3. PV and SV wave velocities in shallow sediments are 1.5 ~ 2.0 km/s and 0.50 ~ 1.0 km/s respectively (Fig.2). The estimated SV wave velocity, much is lower than the water P wave velocity can explain the observed low group velocity of Scholte waves.
We expect that the SV and PV wave velocity of the shallow ocean-bottom sediment to vary significantly in space and time. This is why our partial derivatives and the measurement of V/H with ambient noise data of S-net should be useful to retrieve and spatial-temporal variations in the region around the Japan trench.
In this study, we first formulated partial derivatives of V/H for vertical transversely isotropic media including sea water. It is well known that P-wave velocity has little effect on the ellipticity in isotropic media. In the VTI media, however, the effect or distinguishment of PH and PV wave velocities cannot be neglected in comparison to conventional SV wave velocity in media of a certain degree of VTI. This is why, the partial derivative or sensitivity of P wave in isotropic media is the sum of those of PV and PH waves with opposite signs.
Next, since short-period Rayleigh waves have sensitivity for shallow parts, we applied the seismic interferometry to the noise data recorded at S-net stations. The cross-correlations among S3 stations perpendicular to the trench axis clearly show surface wave signals with relatively good symmetry of causal and anti-causal waveforms in the period range from 5 to 12 sec (Fig.1). Their group velocity is very small, about 0.2-0.4 km/s. This signals probably correspond to Scholte waves. Scholte waves propagate parallel to the water-sediment interface, when the shear velocity of the solid is lower than the compressional velocity in the water (1.5 km/s), that is, very loose sediments.
Finally, we estimated V/H spectrum ratios of the extracted Scholte waves at S-net stations and inverted a new velocity model of sedimentary layers. Prior to the V/H inversion, we estimated a velocity model of the basement beneath sediment by the measure phase velocity dispersion. The area in the subarray S3 of deep ocean near the Japan trench axis has small V/H values. This indicates slow velocity regions there, probably very and/or thick soft sediments. The estimated PV and SV wave velocities can explain these low V/H values in the subarray S3. PV and SV wave velocities in shallow sediments are 1.5 ~ 2.0 km/s and 0.50 ~ 1.0 km/s respectively (Fig.2). The estimated SV wave velocity, much is lower than the water P wave velocity can explain the observed low group velocity of Scholte waves.
We expect that the SV and PV wave velocity of the shallow ocean-bottom sediment to vary significantly in space and time. This is why our partial derivatives and the measurement of V/H with ambient noise data of S-net should be useful to retrieve and spatial-temporal variations in the region around the Japan trench.