Hitoshi Oda

Subducting oceanic crust on the Philippine Sea plate in Southwest Japan

Abstract It is well-known that subcrustal earthquakes occur in Southwest Japan owing to subduction of the Philippine Sea (PHS) plate. Observing the earthquakes at a single station in the Chugoku district, we frequently find a distinct pair of later P and S phases after weak initial P and S phases, respectively. These later phases were observed only for earthquakes with particular source-receiver geometries; the events in the west Seto Inland sea, epicentral distances from 120 km to 240 km, and the source-receiver direction almost parallel to the strike of the Nankai trough. The feature of initial and later phases was summarized as follows: (1) The later phases are observed for subcrustal earthquakes occurring at depths between 40 km and 60 km, but not for those below the depth, (2) “Single-station” apparent velocities of initial P and S phases are 8.0 km/s and 4.9 km/s, and they are somewhat higher than ordinary velocities of the uppermost mantle. (3) Single-station apparent velocities of later P and S phases are, on the other hand, 6.6 km/s and 3.8 km/s, and they are comparable to the velocities of the continental lower crust and the main layer of oceanic crust. These features could not be explained unless the subcrustal earthquakes take place within the low-velocity oceanic crust overlying the high-velocity oceanic plate. This suggests that the oceanic crust subducts beneath the west Seto Inland sea as a part of the PHS plate.

Regional variation in surface wave group velocities in the Philippine Sea

Abstract Group velocities of Rayleigh and Love waves have been measured within the period range 30–80 s to study the regional variation of seismic wave velocity structures of the crust and upper mantle beneath the Philippine Sea. These group velocity measurements were made by applying the moving window method to long-period seismograms recorded at six WWSSN stations, which are located around the Philippine Sea plate. We divided the Philippine Sea region into seven small areas on the basis of tectonic features, age and topography of ocean floor, and determined the group velocities of the Rayleigh and Love waves in each region as a function of period. The surface wave group velocities in the Philippine Sea plate are higher than those in tectonic regions such as the Mariana arc, the Ryukyu-Japan arc and the Philippine trench. The low group velocities in tectonically active regions are attributable to the presence of an extremely low velocity layer in the upper mantle. In the Philippine Sea plate, the group velocities in the western region (the west Philippine basin) are higher than those in the eastern region (the Shikoku-Parece Vela basin). This result is interpreted as being due to the difference in the lithospheric thicknesses of the two regions. The lithospheric thickness estimated from the group velocity data of the Rayleigh wave is about 55 km in the western region and 40 km in the eastern region. The northern part of the west Philippine basin, covering the Amami plateau, the Daito ridge and the Oki-Daito ridge, has a thick crust, which is needed to explain the fact that the short-period group velocities are lower than those in the western and eastern regions of the Philippine Sea plate. These features are commonly found in the seismic velocity structures inferred from the group velocity data of Rayleigh and Love waves.

S wave splitting observed in southwest Japan

Abstract Shear wave polarization anisotropy was investigated for S and ScS waves recorded at seismological station SBK in the Chugoku district, southwest Japan. Polarization direction of the fast component and time lag of shear wave which was split by elastic anisotropy were measured by a waveform cross-correlation method. Comparison between splittings of the S and ScS waves reveals that the shear wave anisotropy is higher in the upper mantle above the depth of 400 km than in the lower mantle below it. The fast components of the split ScS waves were polarized nearly in the NW–SE direction, irrespective of azimuth angles of earthquake epicenters. This polarization direction agrees with that of the fast components of the direct S waves, but the S waves radiated from deep earthquakes in the Izu–Ogasawara arc region show the polarization anisotropy of the NE–SW direction which is different from that of ScS waves. The NW–SE polarization direction is attributed to the velocity anisotropy in the upper mantle caused by subduction of the Philippine Sea plate beneath southwest Japan and the NE–SW polarization direction is interpreted as being due to the anisotropy in the back arc mantle under the Izu–Ogasawara region which is a subduction zone of the Pacific plate. The measured polarization anisotropy is due to deformation-induced lattice preferred orientation which is caused to olivine crystals in the upper mantle by subduction of oceanic plate.

Application of the Dugdale crack model to the stress-strain relations and bulk moduli of a granite

Variations of dynamic and static hulk moduli and compressional strain of a rock sample of granite in response to hydrostatic pressure at room temperature are summarized as follows: 1. (1) “ nonlinear” increase of compressional strain becomes “linear” at σ0 (border pressure, for this rock, σ0 = 1.36 kbar and is the pressure which divides the compression curve in two parts: the crack-active part and the crack-closed part). 2. (2) static bulk modulus obtained from compressional strain data ϵ drastically increases and levels off at pressures higher than the border pressure σ0, 3. (3) pressure dependence of dynamic bulk modulus obtained from acoustic measurements has a trend similar to that of the static one, and 4. (4) border pressure σ0 from (3) is just equal to that from (1) and (2). To interpret these phenomena and understand the underlying reason for these variations, two crack models were examined, i.e., an elliptical crack model with constant aspect ratio and the Dugdale crack model. The difference in closing processes for the two models results in different changes in effective bulk modulus and compressional strain due to the applied pressure. If the total number of cracks cannot change with increasing pressure, the drastic increase in bulk modulus and nonlinear part of the stress-strain curve at low pressure can be attributed to reduction of crack width or crack length. It is shown that the Dugdale crack model explains better the observed results. Theoretical calculations show that the crack density of the rock sample is 0.42 (dimensionless) and that the yield stress of 2.09 kbar acts in the proximity of tips of Dugdale cracks. These values seem reasonable when we consider the values of the porosity of 0.16% and the border pressure of 1.36 kbar.

Stripping Analysis of Ps-Converted Wave Polarization Anisotropy

Abstract The polarization anisotropy of the Ps -converted wave tells us about anisotropic properties of the region between a seismic station at which the Ps phase is observed and a velocity discontinuity at which the Ps conversion takes place. However, it is very difficult to know the anisotropic structure as a function of depth from the polarization anisotropy data because the data are obtained as a quantity integrated along the ray path from the station to the velocity discontinuity. In this study, I present a method to estimate seismic anisotropy of the layers using the shear-wave splitting of Ps phases that originate at velocity discontinuity interfaces in a flat-layered structure. After a description of the method, analyses of synthetic waveforms show that it is of benefit in the study of anisotropic structure of shear waves.

Effect of occurrence rate of acoustic emissions on their statistical behavior

The parameterm in Ishimoto-Iidas relation was investigated for acoustic emissions (AEs) occurring in rock samples under uniaxial compression. In the experiment, we found: 1) The large AEs are counted without serious error but the number of small AEs is systematically underestimated at high AE rates, 2) the frequency distribution of maximum AE amplitudes becomes nonlinear in logarithmic scale with increasing AE rate, and 3) there exists a strong negative correlation betweenm-value and AE rate. The miscount of small AEs was interpreted as due to overlap of the large and small AEs. We call the miscount “masking effect”. A statistical analysis based on the masking effect showed that them-value decreases more effectively as the AE rate increases, and thus the masking effect is a possible origin both for the nonlinear frequency distribution of maximum AE amplitudes and for the negative correlation ofm-value with AE rate. We emphasize that one should be careful of the masking effect to examine correctly the change, ofm-value. In order to eliminate the masking effect, AEs should be measured by a measurement system with low sensitivity. Even if the masking effect is eliminated, them-value decreases before the main fracture of a rock sample. Them-value is a key parameter to predict the main fracture.

Elasticity measurements on the material analogous to iron in the inner core of the Earth by FT‐ultrasonic spectroscopy

Iron constituting the inner core of the Earth is probably of e‐phase with hexagonal closest packing structure stable only above 10 GPa. The elastic properties of e‐iron can be studied by using analogue materials, solid solutions of iron with ruthenium, which take hcp at relevant condition. Single crystalline ruthenium sample was grown by means of floating zone method and its elastic parameters and internal frictions were determined with an FT (Fourier Transform) ultrasonic spectroscopy introduced successfully to RST (Resonant Sphere Technique). The potentiality of the new method is discussed in relation to the study of physical properties of iron in high pressure phase.

Frequency-dependence of velocity and attenuation of elastic waves in granite under uniaxial compression

Velocity as well as attenuation factorQ−1 ofP-wave in a dry granitic rock sample under uniaxial compressions were measured in the range of frequency between 100 kHz and 710 kHz by using the pulse transmission technique. Above the stress of 0.5 σf, where σf is the fracture stress, theP-wave velocity decreases with increasing axial stress, whereasQ−1 increases. Particularly, the change ofQ−1 is greater for high frequency than for low frequency. At a given stress level, the higher the frequency, the higher theP-wave velocity and the largerQ−1. This result means that the velocity decrease with increasing stress is smaller for higher frequency. Because of this frequency-dependence of velocity decrease, theP-wave in the rock under dilatant state shows dispersion. The body wave dispersion is more remarkable at higher stress, and is not found in a homogeneous material with no cracks. Thus the disperison is attributed to the generation of cracks. When the frequency-dependence ofQ−1 is approximated asfn in the present frequency range, the exponentn takes a value from 0.63 to 0.77.

Three-dimensional P- and S-wave attenuation structures around the source region of the 2016 Kumamoto earthquakes

We investigate the three-dimensional P- and S-wave attenuation (