JpGU-AGU Joint Meeting 2020

Presentation information

[J] Poster

S (Solid Earth Sciences ) » S-SS Seismology

[S-SS12] Seismic wave propagation: Theory and Application

convener:Kiwamu Nishida(Earthquake Research Institute, University of Tokyo), Kazuya Shiraishi(Japan Agency for Marine-Earth Science and Technology), Takao Nibe(JAPEX), Kaoru Sawazaki(National Research Institute for Earth Science and Disaster Resilience)

[SSS12-P18] Propagation characteristics of two-dimensional scalar waves in alternating layered structures

*Kazuki Kidoguchi1, Jun Kawahara1 (1.Graduate School of Science and Engineering, Ibaraki University)

Keywords:alternating layers, scalar waves, wave propagation characteristics

Deep events in the subducting Pacific plate often produce anomalously large intensity on the Pacific seaboard of eastern Japan. The cause was first explained by the Utsu model, assuming a high-Q subducting plate compared with the mantle wedge above it [Utsu, 1967; Utsu and Okada, 1968]. It was shown later that the waveform records in the regions of large intensity show a low-frequency (f < 0.25 Hz) onset followed by large-amplitude high-frequency (f > 2 Hz) later waves [e.g., Iidaka and Mizoue, 1991; Abers, 2000]. The Utsu model cannot explain this feature. Furumura and Kennett [2005] conducted two-dimensional numerical wave simulations of deep events in the Pacific plate with von Karman-type random heterogeneities. They showed that the heterogeneities with a longer correlation length (10 km) in the downdip direction along the subducting plate and a much shorter correlation length (0.5 km) in thickness reproduced very well the characteristics of the observation. This implies that such a lamina structure acts as an efficient waveguide for high-frequency waves by multiple scattering.

We examined this high-frequency waveguide effect using numerical simulations based on alternating layered structures, which are simplified models of lamina structures. We simulated two-dimensional scalar waves radiated from a point source with a Ricker-wavelet time function, adopting the velocity-stress finite-difference scheme with a staggered grid [Virieux,1984]. An alternating layered structure was made of layers with the same thickness, whose wave velocities was 10% either higher or lower than the mean velocity 4.5 km/s. In every simulation, the total thickness of the layered structure was 59.95 km and the dominant wavelength of the radiated waves (defined as the mean velocity divided by the central frequency of the Ricker wavelet) was 2 km. We changed the layer thickness from 4.95 km down to 0.1 km in respective simulations, thereby changing the wavelength relative to the layer thickness. We deployed 20 km-long receiver arrays from the source in the directions both parallel and normal to the layers (hereafter, horizontal and vertical, respectively).

We obtained the following results. As the relative wavelength became shorter, horizontally propagating waves showed more clearly the dispersion and broadening of waveforms, together with the delay of maximal peaks. These features would denote the high-frequency waveguide effect stated above. For wavelength longer than the layer thickness, vertically propagating waves also showed waveform broadening due to the generation of coda waves, but the forms of the direct waves remained nearly unchanged except for the amplitude attenuation. The generation of coda waves was suppressed as the wavelength became shorter than the layer thickness. We also observed that the propagation velocity for the onset of the horizontally propagating waves became higher than the mean velocity, if the wavelength was shorter than the layer thickness. This would be because the direct waves selected the fastest paths (higher-velocity layers) in the layered structures.