- Email: [email protected]
P-wave tomography, anisotropy and seismotectonics in the eastern margin of Japan Sea
Tectonophysics 489 (2010) 177–188
Contents lists available at ScienceDirect
Tectonophysics j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / t e c t o
P-wave tomography, anisotropy and seismotectonics in the eastern margin of Japan Sea Zhouchuan Huang a,b,⁎, Dapeng Zhao a,⁎, Norihito Umino a, Liangshu Wang b, Toru Matsuzawa a, Akira Hasegawa a, Takeyoshi Yoshida c a b c
Research Center for Prediction of Earthquakes and Volcanic Eruptions, Tohoku University, Sendai 980-8578, Japan School of Earth Sciences and Engineering, Nanjing University, Nanjing 210093, China Institute of Mineralogy, Petrology and Economic Geology, Tohoku University, Sendai 980-8578, Japan
a r t i c l e
i n f o
Article history: Received 6 July 2009 Received in revised form 5 April 2010 Accepted 17 April 2010 Available online 24 April 2010 Keywords: Eastern margin of the Japan Sea Tomography Anisotropy Crustal earthquakes Fluids
a b s t r a c t To understand the seismotectonics in the eastern margin of the Japan Sea, we determined the first highresolution 3-D P-wave velocity structure and azimuthal anisotropy under the Japan Sea off Northeast Japan using 175,425 high-precision P-wave arrival times from 2833 local earthquakes recorded by 330 seismograph stations. P-wave arrival times from 145 suboceanic earthquakes relocated with sP depth phase are crucial to determine the structure of the crust and uppermost mantle under the Japan Sea. Our results show that strong velocity variations exist in the crust and uppermost mantle under the eastern margin of the Japan Sea. Many large crustal earthquakes occurred in or around low-velocity zones which may represent weak sections of the seismogenic crust. The P-wave azimuthal anisotropy is complex under the Japan Sea, which may also indicate the complex crustal structures there. In the eastern margin of the Japan Sea, the strong heterogeneities in the crust and upper mantle revealed by seismic tomography may reflect the complicated geologic structures such as the alternate rift zones, ridges, basins, horsts, grabens, volcanics, and continental fragments which were produced during the back-arc spreading, opening of the Japan Sea and the present compressional stage of the Honshu arc. Crown Copyright © 2010 Published by Elsevier B.V. All rights reserved.
1. Introduction Japan is an earthquake country and has suffered heavily from seismic hazards during the long history (e.g., Utsu, 1982; Usami, 2003). Earthquakes that occurred in or around Northeast (NE) Japan are caused by the active subduction and collisions among three lithospheric plates (Fig. 1), where the Pacific plate is subducting from the east beneath the North American and Amurian plates (Hasegawa et al., 2009). Large interplate earthquakes occur actively in the forearc region under the Pacific Ocean, while large intraplate earthquakes occur in the upper crust under the Japan Islands and in the eastern margin of the Japan Sea (EMJS) (Fig. 1). Seismic tomography is proved to be an effective tool to study the structural heterogeneity in the crust and upper mantle and its relationship to seismic and volcanic activities (see Zhao, 2009; Hasegawa et al., 2009 for comprehensive reviews). There is a correlation between the spatial distribution of large crustal earthquakes (M5.7–8.0) under
⁎ Corresponding authors. Research Center for Prediction of Earthquakes and Volcanic Eruptions, Tohoku University, Sendai 980-8578, Japan. Fax: +81 22 264 3292. E-mail addresses: [email protected] (Z. Huang), [email protected] (D. Zhao).
the Japan Islands during 1885–2000 and structural heterogeneities in the crust and uppermost mantle revealed by seismic tomography (Zhao et al., 2002, 2010). So far the 3-D seismic velocity structure under the land area of the Japan Islands has been well determined, but the structures under the Pacific Ocean and the Japan Sea are not well known because very few seismic stations exist in the surrounding oceanic regions and thus the suboceanic earthquakes are poorly located by the routine procedure of the seismic networks that are installed only on the Japan land area. Umino and Hasegawa (1994) and Umino et al. (1995) made an important breakthrough in the seismological study of the oceanic regions around the Japan Islands. They detected sP depth phase from three-component seismograms of shallow earthquakes that occurred under the Japan Sea and the Pacific Ocean recorded by the seismograph stations in NE Japan (Fig. 2). The sP depth phase is a seismic wave that is radiated upward from an earthquake hypocenter as an S-wave and is reflected from a bounce point on the surface or seafloor and at the same time converted from the S-wave to a P-wave which finally arrives at a seismic station. The bounce point of the sP depth phase is close to the earthquake epicenter, and so its travel time is very sensitive to the focal depth. Thus the arrivaltime data of sP depth phase are very useful to locate the suboceanic earthquakes which occur outside a seismic network, because the
0040-1951/$ – see front matter. Crown Copyright © 2010 Published by Elsevier B.V. All rights reserved. doi:10.1016/j.tecto.2010.04.014
178
Z. Huang et al. / Tectonophysics 489 (2010) 177–188
Fig. 1. Distributions of large crustal earthquakes (black stars) during 830–2010 and active faults in Northeast Japan and the Japan Sea. The occurrence years and magnitudes of the large earthquakes are also shown. Gray triangles show the active volcanoes in Northeast Japan. Inset map: tectonic background of Northwest Pacific and Northeast Asia. The dashed line indicates the estimated border of the Amurian plate (Wei and Seno, 1998). Arrows show the directions of the plate motions relative to Northeast Japan.
focal depth of such an earthquake is the most difficult parameter to determine among the four hypocentral parameters (i.e., origin time, latitude, longitude, and focal depth). Therefore the detection of an sP depth phase from a seismogram is just like installing a new seismometer (at the sP bounce point) close to the epicenter. Umino and Hasegawa (1994) and Umino et al. (1995) showed that the suboceanic earthquakes under the Japan Sea and the Pacific Ocean could be located very accurately with the sP depth phase data, and the location accuracy is comparable to that of earthquakes under the seismic network in NE Japan (hypocenter location error b3 km). Umino et al. (1998) further relocated many more crustal earthquakes that occurred in a broad region in EMJS off NE Japan by using the sP depth phase data collected from three-component seismograms. Zhao et al. (2002) used suboceanic earthquakes precisely relocated with sP depth phase data to determine the first 3-D P-wave velocity (Vp) structure under the NE Japan forearc region from the Japan Trench to the Pacific coast, and they showed that the method is very powerful for tomographic imaging outside a seismic network. Later, more detailed 3-D P- and S-wave velocity and Poisson's ratio images in Tohoku and Hokkaido forearc areas under the Pacific Ocean were determined by using more sP depth phase data in addition to a large
number of P and S-wave arrival times (Mishra et al., 2003; Wang and Zhao, 2005; Zhao et al., 2007). Umino and Hasegawa (1994) and Umino et al. (1998) relocated the aftershocks of the 1993 Hokkaido–Nansei–Oki Earthquake (M 7.8) and other crustal earthquakes in EMJS using sP depth phase data. However, so far no one has determined the 3-D velocity structure under the Japan Sea. In this work, we tried to determine the first 3-D Vp structure under EMJS by using a large number of P-wave arrival times from the suboceanic earthquakes investigated by Umino et al. (1998) and the earthquakes that occurred under the NE Japan land area. Our first tomographic images in EMJS shed light on the structural heterogeneities and seismotectonics of this region. 2. Data and method Fig. 3 shows the hypocentral distribution of the 2833 earthquakes used in this study. These events consist of two groups. The first group includes 145 crustal events which occurred in EMJS (Fig. 3a) during April 1992 to December 1997 and have 442 sP depth phase arrival times collected by Umino et al. (1998) and 4740 P-wave arrival times recorded by the Tohoku University
Z. Huang et al. / Tectonophysics 489 (2010) 177–188
179
Fig. 2. An example of three-component seismograms (b) of a crustal earthquake under the Japan Sea recorded by a seismic station in NE Japan (a). The epicentral distance is 140 km. The three dots show the first P and S arrivals and the sP depth phase.
seismic network. The second group consists of 2688 events which occurred under the NE Japan land area during January 1992 to April 2003 and have 170,685 P-wave arrival times. Each of the events was recorded by more than 15 seismic stations and their hypocentral locations are accurate to 1–2 km. The total number of 2833 events in our data set generated 175,425 P-wave arrival times recorded by 330 seismic stations in NE Japan (Fig. 3d). These seismic stations are permanent and portable stations operated by Tohoku University and other Japanese universities, Japan Meteorological Agency, and the High-Sensitivity Seismic Network (Hi-net). The picking accuracy of the P-wave arrival times is estimated to be 0.05–0.15 s. We used the tomographic method of Zhao et al. (1992, 2007) to determine the 3-D Vp structure under the study region (Figs. 1 and 3). The starting one-dimensional (1-D) velocity model for the tomographic inversion was derived from Zhao et al. (1992, 2007) (Fig. 4). Lateral depth variations of the Conrad and Moho discontinuities and the upper boundary of the subducting Pacific slab were taken into account in the 3-D ray tracing and the tomographic inversion, because the existence of the three discontinuities under NE Japan has been established well and their geometries have been determined well (see Zhao et al., 2007
and Hasegawa et al., 2009 for details). Under the Japan Sea, the depths of the Conrad and Moho discontinuities are modified based on the results of seismic explosion experiments (Nishisaka et al., 2001; Iwasaki and Sato, 2009). To express the 3-D velocity structure, a 3-D grid was set up in the study area with a lateral grid interval of 30–50 km. Meshes of grid nodes were set at 10, 25, 40, 65, 90, 120, 160 and 200 km depths in the crust and mantle wedge, and at 5 and 25 km under the upper boundary of the subducting Pacific slab, following the approach of Zhao et al. (1992, 2007). Hypocenter parameters and velocity perturbations at the grid nodes were taken as unknown parameters. The velocity perturbation at any point in the model was calculated by linearly interpolating the velocity perturbations at the eight grid nodes surrounding that point. An efficient 3-D ray tracing technique (Zhao et al., 1992) was used to compute travel times and ray paths accurately. The damped least-squares method was used to invert the large and sparse system of observation equations that relate the observed arrival times to the unknown parameters. The 3-D Vp structure and hypocenter parameters were determined simultaneously in an iterative inversion process. The surface topography and station elevations were taken into account in the 3-D ray tracing and the tomographic inversion.
180
Z. Huang et al. / Tectonophysics 489 (2010) 177–188
Fig. 3. (a) Epicentral distribution of 2688 events (crosses) that are located accurately by the seismic network and 145 suboceanic events (dots) that are relocated by using the sP depth phase data. (b) East–west and (c) north–south vertical cross-sections of the hypocenter locations for the earthquakes shown in (a). (d) Distribution of 330 seismic stations (gray circles) used in this study. The solid triangles show the active volcanoes.
3. Velocity structure and resolution tests Fig. 5 shows the hypocenter distribution of the 145 crustal events in EMJS we relocated as compared with those determined by the seismic network of Tohoku University. The 145 events were relocated by applying the earthquake location scheme of Zhao et al. (1992, 2007) with 3-D ray tracing to both P and sP arrival-time data for the 1-D velocity model (Fig. 4) and the final 3-D velocity model (Figs. 7 and 8). The seafloor topography and sediment layer under the Japan Sea were also taken into account in the earthquake relocation by using the model CRUST2.0 (Laske et al., 2003) which is the updated version of the model CRUST5.1 (Mooney et al., 1998). The formal uncertainties of the hypocenter locations are smaller than 3 km for all the 145 events, thanks to the use of sP depth phase data. The hypocenters of the 145 events located by the seismic network show
a diffusive distribution in a depth range of 0–58 km. After relocation with the sP depth phase, however, most of the events are located shallower than 15 km, only a few events close to the coastline under southern NE Japan have focal depths of 15–21 km (Fig. 5e). The differences in the hypocenter locations resulted from the 1-D and 3-D velocity models are relatively small, which do not exceed 3 km for most of the events (Fig. 5). A few linear dipping features are visible in the hypocenter distribution of the relocated events under EMJS (Fig. 5c–e), which may represent active faults there (Umino and Hasegawa, 1994; Umino et al., 1998). We conducted many checkerboard resolution tests to examine the resolution scale of the tomographic images with our data set following the approach of Zhao et al. (1992, 2007). We first assigned positive and negative velocity anomalies of 3% to all the 3-D grid nodes, then calculated synthetic arrival times for the checkerboard
Z. Huang et al. / Tectonophysics 489 (2010) 177–188
181
synthetic test is the same as that of the checkerboard resolution test (Fig. 6) except for the input model which was constructed from the obtained tomographic results (Figs. 7 and 8) as follows. For the grid nodes with velocity anomalies ≥ +3%, they are changed to 6%, while for those with velocity anomalies ≤ −3%, they are changed to −6%. For the grid nodes with velocity anomalies between −3% and +3%, they are changed to 0%. The velocity perturbation at any point in the model is calculated by linearly interpolating the velocity perturbations at the eight grid nodes surrounding that point. Thus the input model is constructed (Fig. 9a–g). The results of the synthetic test (Fig. 9a′–g′) show that the main features of velocity anomalies in the input model (and also in the real tomographic images) are well recovered, suggesting that those anomalies are reliable features. 4. Anisotropic tomography
Fig. 4. The starting P-wave velocity model adopted for the tomographic inversion. Hc, Hm and Hslab denote depths of the Conrad and Moho discontinuities and the upper boundary of the subducting Pacific slab. The depth variations of the three discontinuities determined by the previous studies (Zhao et al., 1992, 2007; Nishisaka et al., 2001; Iwasaki and Sato, 2009) are adopted in this study. A high-velocity anomaly of 4% is added to the subducting Pacific slab in the starting velocity model (dashed line).
model, and then inverted the synthetic data to see if the assigned velocity anomalies could be recovered or not. The numbers of seismic stations, events and ray paths in the synthetic data are the same as those in the real data set. Random errors with a normal distribution having a standard deviation of 0.1 s were added to the synthetic arrival times before the tomographic inversion. Fig. 6 shows the test results at six depths in the crust and mantle wedge and two layers in the subducting Pacific slab, which indicate that our tomographic model has a resolution of 30–50 km in the horizontal direction and 10–30 km in depth. Fig. 7 shows the tomographic images at 10, 25 and 40 km depths together with the distribution of large crustal earthquakes (M ≥ 6.0) that occurred during 830 to 2010 in the present study area (Utsu, 1982; Usami, 2003). Distributions of active faults and active volcanoes are also shown. The results show that strong lateral heterogeneities exist in the crust and the uppermost mantle under EMJS. A belt of high-velocity (high-V) anomalies is visible along the Japan Sea coast, extending from the surface to the upper mantle. Several low-velocity (low-V) anomalies also exist in the crust and uppermost mantle under the Japan Sea. Although the correlation between the tomography and the distribution of large crustal earthquakes is not very obvious, many large earthquakes seem to be located in or around the low-V zones, as pointed out earlier by Zhao et al. (2002, 2010). However, some large earthquakes are located in or around the high-V zones (Fig. 7). Fig. 8 shows seven E–W vertical cross-sections of P-wave tomography down to 130 km depth together with seismicity, large crustal earthquakes, low-frequency microearthquakes (Hasegawa and Yamamoto, 1994), and active volcanoes in the vicinity of each profile. To confirm the main features of the obtained 3-D Vp structure, we made one more synthetic test (Fig. 9). The procedure of the
We further determined the 3-D P-wave anisotropic tomography beneath the EMJS region using the method of Wang and Zhao (2008). Compared with the isotropic tomography, two additional unknown parameters are added at each grid node to represent the P-wave azimuthal anisotropy. The LSQR algorithm (Paige and Saunders, 1982) was applied to solve the large sparse system of observation equations that relates the arrival-time data with the unknown parameters for the hypocenter locations and the isotropic and anisotropic velocity variations at every grid nodes. We first made the checkerboard resolution test as mentioned above for the anisotropic tomography (Wang and Zhao, 2008). At every grid nodes of the input velocity model (Fig. 10), we assigned isotropic velocity anomalies of ± 3% alternatively and two anisotropic parameters which represent the fast Vp directions of 22.5° or 112.5° with anisotropic velocity variations of ∼ 4.24% (see Wang and Zhao, 2008 for details). The test results show that both the pattern and amplitudes of the isotropic and anisotropic velocity anomalies are well recovered at the depths of 10 and 25 km (Fig. 10). At 40 km depth, the resolution becomes lower under the Japan Sea. Fig. 11 shows the distribution of P-wave fast-velocity directions (FVDs) together with the isotropic velocity anomalies at different depths under the EMJS. The general pattern of the isotropic velocity images is similar to that obtained from the isotropic tomography (Fig. 7), e.g., the high-V belt along the Japan Sea coast and the low-V zones under the EMJS, though the amplitude of the velocity anomalies becomes smaller. The anisotropic tomography results (Fig. 11) show that the FVDs are complex in the upper crust, while in the lower crust and uppermost mantle under EMJS, the FVDs with large amplitudes are generally oriented in NW–SE to E–W directions. This issue is discussed in detail in the next section. 5. Interpretation and discussion Until the early Miocene, Japan was part of the Asian margin (Sato, 1994). The northeast Honshu arc was constructed by backarc spreading during 21–18 Ma when Yamato basin formed, and by the subsequent rifting during 19–13.5 Ma (when Northern Honshu rift system formed) and island-arc uplifting up to the present (Ohtake et al., 2002; Yamada and Yoshida, 2004; Yoshida et al., 2005). The EMJS region has been under the convergent tectonics since the Pliocene under the E–W compressional stress regime (Tamaki and Honza, 1985). The relative motion of 9– 17 mm/year between the Amurian and North American plates is accommodated at EMJS as shown by the GPS observations (Wei and Seno, 1998; Jin et al., 2007). Sagiya et al. (2000) studied the crustal deformation of the Japan Islands using GPS measurements and suggested that EMJS has a much larger strain rate than the surrounding areas. Hence the EMJS region has a high seismic
182
Z. Huang et al. / Tectonophysics 489 (2010) 177–188
Fig. 5. Hypocenter distributions of the 145 shallow earthquakes under the eastern margin of the Japan Sea in (a) map view and in (b) north–south and (c–e) east–west vertical crosssections. The open circles show the hypocenters located by the seismic network of Tohoku University, while blue and red crosses denote the hypocenters determined in this study for the 1-D and 3-D velocity models using P and sP depth phase data (see text for details).
potential and there may be a common mechanism for the large earthquakes there (Ohtake et al., 2002). Previous studies suggest that the generation of a large earthquake is not a pure mechanical process, but is closely related to the physical and chemical properties of materials in the crust and upper mantle, such as magma, fluids, etc. Zhao et al. (2002, 2010) studied large crustal earthquakes under the Japan Island and concluded that most of the earthquakes occurred in or around the low-V zones which may represent the weak zones in the crust. Under the volcanic front, the crustal weakening may be caused by the upward intrusion of the magma and fluids from the mantle wedge which is caused by the subduction of the Pacific Plate and dehydration of the slab (Hasegawa et al., 2005, 2009; Zhao et al., 2002, 2010). However, the correlation between the velocity anomalies and the distribution of large crustal earthquakes and active faults becomes lower under the Japan Sea. Although many earthquakes are located in or around the low-V zones, some are in the high-V areas (Fig. 7). Several factors may cause the lower correlation. Firstly, the tomographic resolution is lower under the Japan Sea compared with that under the land area, and so some smaller velocity anomalies may not be imaged well. Secondly, the hypocenter locations for some of the large historic earthquakes are less accurate. Note that the hypocenter locations of the large earthquakes that occurred after about 1900
represent the initial points of earthquake ruptures because the earthquakes were located by the modern seismic network in Japan, and they may be accurate to about 10 km or less. In contrast, the hypocenters of large historic earthquakes before 1900 represent the centroid locations of the rupture zones because they were estimated from the macroscopic damages of the earthquakes, and so they may have uncertainties of tens of kilometers (Usami, 2003). Finally, the discrepancy itself may be the clue that the structure under the Japan Sea is very complex and the ruptures of the large crustal earthquakes are affected by various factors. The P-wave anisotropic images show another piece of evidence for the structural heterogeneity under the EMJS. The anisotropic image in the upper crust is complex, while the FVDs with large amplitudes are E–W or NW–SE in the lower crust and uppermost mantle, which are similar to those revealed by Wang and Zhao (2008) and to the S-wave splitting results (Nakajima et al., 2006). The upper-crust anisotropy is mainly caused by the stress-induced microcracks, the cracks and fractures accompanied with the active faults, and the intrinsic rock anisotropy resulting from preferred orientation of the minerals (Kaneshima, 1990). Under the EMJS, the complex anisotropic structures may be caused by the frequent occurrence of large crustal earthquakes and the formation and closing of active faults that change the stress status and the orientations of the cracks and fractures, as well as by the intrusion
Z. Huang et al. / Tectonophysics 489 (2010) 177–188
183
Fig. 6. Results of a checkerboard resolution test for P-wave velocity structure at six depths in the crust and mantle wedge (a–f) and at two layers in the subducting Pacific slab (g, h). The dashed lines in (d–f) indicate the upper boundary of the Pacific slab. The depth to each layer is shown at the upper-left corner of each map. The velocity perturbation scale is shown below (f).
and transportation of the magma and fluids that change the preferred orientations of the minerals. Detailed tomographic images were determined for the source areas of the 2004 and 2007 Niigata earthquakes (both with M 6.8) in southwestern part of NE Japan (Fig. 1) (e.g., Nakajima and Hasegawa, 2008; Xia et al., 2008). An anomaly with pronounced low-velocity and high Poisson's ratio in the lower crust and uppermost mantle was revealed under the Niigata source areas, which also exhibits high conductivity and high He3/He4 ratio (Ogawa and Honkura, 2004; Uyeshima et al., 2005; Horiguchi and Matsuda, 2008), indicating the existence of fluids under EMJS rising from the mantle wedge as a result of slab dehydration and corner
flow in the mantle wedge. The results suggest that high-temperature arc and back-arc magmas as well as fluids released from the ascending flow in the mantle wedge can affect the rupture nucleation of large crustal earthquakes. This may be one reason for the frequent occurrence of large crustal earthquakes in the EMJS (Fig. 7). Many normal faults and fault-related rifts and grabens developed in Early to Middle Miocene, simultaneous with backarc spreading of the Japan Sea (Okamura et al., 1995). The normal faults reactivated as reverse faults during the inversion stage due to an increase in compressional stress since 1–2 Ma ago (Tamaki and Honza, 1985). Some large thrust faults during
184
Z. Huang et al. / Tectonophysics 489 (2010) 177–188
Fig. 7. P-wave velocity images at depths of (a) 10, (b) 25, and (c) 40 km under the study area. Red and blue colors denote low and high velocities, respectively. Red stars denote large crustal earthquakes (M ≥ 6.0) that occurred from 830 to 2008 (Utsu, 1982; Usami, 2003). Black triangles denote active arc volcanoes. Active faults are shown by thin lines. Low-frequency microearthquakes within 5 km of each depth are shown by red dots. The velocity perturbation scale and the earthquake magnitude scale are shown at the bottom.
large earthquakes may cut through the whole crust (Fukao and Furumoto, 1975), and seawater may permeate down to the deep crust through these active faults during many earthquake cycles in the long geological history, which may be another source of crustal fluids in addition to the fluids rising from the mantle wedge (Zhao et al., 2002). In our tomographic images, low-V zones are visible widely in the crust under the Japan Sea (Figs. 7 and 8), suggesting that the velocities are reduced by the fluids. The seismic velocity anomalies under EMJS may also reflect the complicated geologic structures such as the alternate rift zones, ridges, basins, horsts, grabens, volcanics, and continental fragments which were produced during the back-arc spreading, opening of the Japan Sea and the current compressional stage of the Honshu arc (Ohtake et al., 2002; Yoshida et al., 2005). In the uppermost mantle, significant high-V anomalies (∼4– 6%) are imaged under the Japan Sea side, while low-V zones are revealed under the NE Japan inland area (Figs. 7 and 8). The result is well consistent with that of seismic explosion experiments carried out across the EMJS, which shows higher subMoho P-wave velocity under the Japan Sea (∼8.0 km/s) than that under the NE Japan land area (∼ 7.5–7.7 km/s) (Nishisaka et al., 2001; Iwasaki and Sato, 2009). Mineral physics studies of lowercrust mafic xenoliths sampled in EMJS suggested that the high-V anomalies in the lower crust under the Japan Sea coast (Figs. 7 and 8) represent cooled igneous rocks formed in the back-arc igneous period (21–13.5 Ma) (Nishimoto et al., 2008). 6. Conclusions We determined the first P-wave tomography and azimuthal anisotropy under the eastern margin of the Japan Sea using a large number of arrival-time data from local earthquakes in the region. Data from 145 suboceanic earthquakes under the Japan Sea which are precisely relocated by using sP depth phase data made it possible to determine the 3-D structure under the
Japan Sea. Our results show that strong lateral heterogeneities exist in the crust and uppermost mantle under the EMJS region, which may reflect the complicated geologic structures such as the alternate rift zones, ridges, basins, horsts, grabens, volcanics, and continental fragments which were produced during the opening of the Japan Sea and the present compressional stage of the Honshu arc. The structural heterogeneities may have affected the seismogenesis in the region. Many large crustal earthquakes occurred in or around lowvelocity zones which may represent weak sections of the seismogenic crust. However, the correlation is reduced under EMJS due to the lower tomographic resolution, less accurate hypocenter locations and the very complex crustal structures under the Japan Sea. The complex P-wave anisotropy also reflects the strong heterogeneities under the EMJS. In the arc and back-arc areas of NE Japan, the crustal weakening may be caused by both high-temperature arc magma and fluids released from the mantle diapirs rising from the mantle wedge. Sea water that permeates down to the deep crust through active faults during many earthquake cycles in the long geological history may be another source of crustal fluids. These results indicate that the generation of large crustal earthquakes is closely related to the tectonic environment and the physical and chemical properties of materials in the crust and upper mantle, such as magma, fluids, etc. Future studies using data from a portable seismic network of oceanbottom seismometers in the Japan Sea are expected to clarify the seismotectonics, back-arc magmatism and mantle dynamics in the EMJS region. Acknowledgements We thank the Hi-net, Tohoku University and Japanese National University seismic networks for providing the waveform and arrival-time data used in this study. This work was supported
Z. Huang et al. / Tectonophysics 489 (2010) 177–188
185
Fig. 8. Seven vertical cross-sections of P-wave velocity image from the surface down to 130 km depth along the profiles shown on the insert map. Red and blue colors denote slow and fast velocities, respectively. Red stars and black triangles denote large earthquakes and active volcanoes within a 30-km width along each profile. Note that the focal depths of the large crustal earthquakes are set at 10 km because the accurate focal depths are unclear for most of these historic earthquakes. White and red circles denote the normal earthquakes and the low-frequency microearthquakes that occurred within a 10-km width of each profile. The horizontal bar atop each cross-section shows the land area where seismic stations are located. The three curved lines denote the Conrad and Moho discontinuities and the upper boundary of the subducting Pacific slab. The velocity perturbation scale and the earthquake magnitude scale are shown at the bottom.
partially by a grant (Kiban-A 17204037) to D. Zhao from Japan Society for the Promotion of Science (JSPS), a grant (40634021) from the National Natural Science Foundation of China, and by the Scientific Research Foundation of Graduate School of Nanjing University. We are grateful to Drs. J. Wang, G. Jiang, S. Miura, T.
Okada, J. Nakajima and N. Uchida for helpful discussions. Prof. M. Liu (Editor) and two anonymous reviewers provided thoughtful comments and suggestions which improved the manuscript. Most of the figures were made by using the Generic Mapping Tools (Wessel and Smith, 1991).
186
Z. Huang et al. / Tectonophysics 489 (2010) 177–188
Fig. 9. Results of a synthetic resolution test for the cross-sections shown in Fig. 8 (see the text for details). The input model is shown in (a–g), while the inversion results are shown in (a′–g′). Crosses and circles denote the high- and low-velocity anomalies, respectively. The scale of the anomalies is showed at the bottom. The other labeling is the same as that in Fig. 8.
References Fukao, Y., Furumoto, M., 1975. Mechanism of large earthquakes along the eastern margin of the Japan Sea. Tectonophysics 25, 247–266. Hasegawa, A., Yamamoto, A., 1994. Deep, low-frequency microearthquakes in or around seismic low-velocity zones beneath active volcanoes in northeastern Japan. Tectonophysics 233, 233–252.
Hasegawa, A., Nakajima, J., Umino, N., Miura, S., 2005. Deep structure of the northeastern Japan arc and its implications for crustal deformation and shallow seismic activity. Tectonophysics 403, 59–75. Hasegawa, A., Nakajima, J., Uchida, N., Okada, T., Zhao, D., Matsuzawa, T., Umino, N., 2009. Plate subduction, and generation of earthquakes and magmas in Japan as inferred from seismic observations: an overview. Gondwana Res. 16, 370–400. Horiguchi, K., Matsuda, J., 2008. On the change of 3He/4He ratios in hot spring gases after the Iwate-Miyagi Nairiku earthquake in 2008. Geochem. J. 42, e1–e4.
Z. Huang et al. / Tectonophysics 489 (2010) 177–188
187
Fig. 10. The input model (a–c) and output results (d–f) of a checkerboard resolution test for P-wave anisotropic tomography. The azimuth and length of the bars represent the fastvelocity direction and the anisotropic amplitude, respectively. The scale for the anisotropic amplitude is shown at the bottom. The other labeling is the same as that in Fig. 6.
Iwasaki, T., Sato, H., 2009. Crust and upper mantle structure of island arc being elucidated from seismic profiling with controlled sources in Japan. Zisin 61, S165–S176. Jin, S., Park, P., Zhu, W., 2007. Micro-plate tectonics and kinematics in Northeast Asia inferred from a dense set of GPS observations. Earth Planet. Sci. Lett. 257, 486–496. Kaneshima, S., 1990. Origin of crustal anisotropy: shear wave splitting studies in Japan. J. Geophys. Res. 95, 11121–11133. Laske, G., Masters, G., Reif, C., 2003. CRUST2.0: A new global crustal model at 2 x 2 degrees. http://mahi.ucsd.edu/Gabi/rem.dir/crust/crust2.html. Mishra, O.P., Zhao, D., Umino, N., Hasegawa, A., 2003. Tomography of northeast Japan forearc and its implications for interplate seismic coupling. Geophys. Res. Lett. 30 GL017736. Mooney, W., Laske, G., Masters, G., 1998. CRUST5.1: a global crustal model at 5 x 5 degrees. J. Geophys. Res. 103, 727–747. Nakajima, J., Hasegawa, A., 2008. Existence of low-velocity zones under the source areas of the 2004 Niigata–Chuetsu and 2007 Niigata–Chuetsu–Oki earthquakes inferred from travel-time tomography. Earth Planets Space 60, 1127–1130. Nakajima, J., Shimizu, J., Hori, S., Hasegawa, A., 2006. Shear-wave splitting beneath the southwestern Kurile arc and northeastern Japan arc. Geophys. Res. Lett. 33, L05305. Nishisaka, H., Shinohara, M., Sato, T., Hino, R., Mochizuki, K., Kasahara, J., 2001. Crustal structure of the Yamato basin and the margin of the northeastern Japan Sea using ocean bottom seismographs and controlled sources. Zisin 54, 365–379. Nishimoto, S., Ishikawa, M., Arima, M., Yoshida, T., Nakajima, J., 2008. Simultaneous high P–T measurements of ultrasonic compressional and shear wave velocities in
Ichino-megata mafic xenoliths: their bearings on seismic velocity perturbations in lower crust of northeast Japan arc. J. Geophys. Res. 113, B12212. Ogawa, Y., Honkura, Y., 2004. Mid-crustal electrical conductors and their correlations to seismicity and deformation at Itoigawa–Shizuoka Tectonic Line, central Japan. Earth Planets Space 56, 1285–1291. Ohtake, M., Taira, A., Ohta, Y., 2002. Active Faults and Seismo-Tectonics of the Eastern Margin of the Japan Sea. Univ. Tokyo Press, Tokyo, p. 201. Okamura, Y., Watanabe, M., Morijiri, R., Satoh, M., 1995. Rifting and basin inversion in the eastern margin of the Japan Sea. Island Arc 4, 166–181. Paige, C., Saunders, M., 1982. LSQR: an algorithm for sparse linear equations and sparse least squares. ACM Trans. Math. Software 8, 43–71. Sato, H., 1994. The relationship between late Cenozoic tectonic events and stress field and basin development in northeast Japan. J. Geophys. Res. 99, 22261–22274. Sagiya, R., Miyazaki, S., Tada, T., 2000. Continuous GPS array and present-day crustal deformation of Japan. Pure Appl. Geophys. 157, 2303–2322. Tamaki, K., Honza, E., 1985. Incipient subduction and deduction along the eastern of the Japan Sea. Tectonophysics 119, 381–406. Umino, N., Hasegawa, A., 1994. Aftershock focal depths of the 1993 Hokkaido–Nansei– Oki earthquake estimated from sP depth phase at small epicentral distances. J. Phys. Earth 42, 321–329. Umino, N., Hasegawa, A., Matsuzawa, T., 1995. sP depth phase at small epicentral distances and estimated subducting plate boundary. Geophys. J. Int. 120, 356–366. Umino, N., Hori, S., Hasegawa, A., 1998. Depth Distribution of Crustal Earthquakes in the Eastern Margin of Japan Sea Determined by Using sP Depth-Phase Data. Abstract of
188
Z. Huang et al. / Tectonophysics 489 (2010) 177–188
Fig. 11. P-wave anisotropic tomographic images at depths of (a) 10, (b) 25, and (c) 40 km under the study area. The azimuth and length of the bars represent the fast-velocity direction and the anisotropic amplitude, respectively. The scale for the anisotropic amplitude is shown at the right bottom. The other labeling is the same as that in Fig. 7.
a Presentation at the Joint Meeting of Japan Earth and Planetary Sciences held in May 1998 in Chiba, Japan. Usami, T., 2003. Materials for Comprehensive List of Destructive Earthquakes in Japan. University of Tokyo Press, Tokyo. Utsu, T., 1982. Catalog of large earthquakes in the region of Japan from 1885 through 1980. Bull. Earthquake Res. Inst. Univ. Tokyo 57, 401–463. Uyeshima, M., Ogawa, Y., Honkura, et al., 2005. Resistivity imaging across the source region of the 2004 mid-Niigata prefecture earthquake (M6.8), central Japan. Earth Planets Space 57, 441–446. Wang, J., Zhao, D., 2008. P-wave anisotropic tomography beneath Northeast Japan. Phys. Earth Planet. Inter. 170, 115–133. Wang, Z., Zhao, D., 2005. Seismic imaging of the entire arc of Tohoku and Hokkaido in Japan using P-wave, S-wave and sP depth-phase data. Phys. Earth Planet. Inter. 152, 144–162. Wei, D., Seno, T., 1998. Determination of the Amurian plate motion. In: Flower, M.F.J., et al. (Ed.), Mantle Dynamic and Plate Interactions in East Asia. Geodynamics Series: Am. Geophys. Union, 27, p. 419. Wessel, P., Smith, W., 1991. Free software helps map and display data, EOS Trans. AGU 72, 441.
Xia, S., Zhao, D., Qiu, X., 2008. The 2007 Niigata earthquake: effect of arc magma and fluids. Phys. Earth Planet. Inter. 166, 153–166. Yamada, R., Yoshida, T., 2004. Volcanic sequences related to Kuroko mineralization in the Hokuroku district, Northeast Japan. Resour. Geol. 54, 399–412. Yoshida, T., Nakajima, J., Hasegawa, A., Sato, H., Nagahashi, Y., Kimura, J., Tanaka, A., Prima, O., Ohguchi, K., 2005. Distribution of Quaternary volcanoes and mantle structures in the NE Honshu arc. Quatern. Res. 44, 195–216. Zhao, D., 2009. Multiscale seismic tomography and mantle dynamics. Gondwana Res. 15, 297–323. Zhao, D., Hasegawa, A., Horiuchi, S., 1992. Tomographic imaging of P and S wave velocity structure beneath northeastern Japan. J. Geophys. Res. 97, 19909–19928. Zhao, D., Mishra, O.P., Sanda, R., 2002. Influence of fluids and magma on earthquakes: seismological evidence. Phys. Earth Planet. Inter. 132, 249–267. Zhao, D., Wang, Z., Umino, N., Hasegawa, A., 2007. Tomographic imaging outside a seismic network: Application to the northeast Japan arc. Bull. Seismol. Soc. Am. 97, 1121–1132. Zhao, D., Santosh, M., Yamada, A., 2010. Dissecting large earthquakes in Japan: role of arc magma and fluids. Island Arc 19, 4–16.
Contents lists available at ScienceDirect
Tectonophysics j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / t e c t o
P-wave tomography, anisotropy and seismotectonics in the eastern margin of Japan Sea Zhouchuan Huang a,b,⁎, Dapeng Zhao a,⁎, Norihito Umino a, Liangshu Wang b, Toru Matsuzawa a, Akira Hasegawa a, Takeyoshi Yoshida c a b c
Research Center for Prediction of Earthquakes and Volcanic Eruptions, Tohoku University, Sendai 980-8578, Japan School of Earth Sciences and Engineering, Nanjing University, Nanjing 210093, China Institute of Mineralogy, Petrology and Economic Geology, Tohoku University, Sendai 980-8578, Japan
a r t i c l e
i n f o
Article history: Received 6 July 2009 Received in revised form 5 April 2010 Accepted 17 April 2010 Available online 24 April 2010 Keywords: Eastern margin of the Japan Sea Tomography Anisotropy Crustal earthquakes Fluids
a b s t r a c t To understand the seismotectonics in the eastern margin of the Japan Sea, we determined the first highresolution 3-D P-wave velocity structure and azimuthal anisotropy under the Japan Sea off Northeast Japan using 175,425 high-precision P-wave arrival times from 2833 local earthquakes recorded by 330 seismograph stations. P-wave arrival times from 145 suboceanic earthquakes relocated with sP depth phase are crucial to determine the structure of the crust and uppermost mantle under the Japan Sea. Our results show that strong velocity variations exist in the crust and uppermost mantle under the eastern margin of the Japan Sea. Many large crustal earthquakes occurred in or around low-velocity zones which may represent weak sections of the seismogenic crust. The P-wave azimuthal anisotropy is complex under the Japan Sea, which may also indicate the complex crustal structures there. In the eastern margin of the Japan Sea, the strong heterogeneities in the crust and upper mantle revealed by seismic tomography may reflect the complicated geologic structures such as the alternate rift zones, ridges, basins, horsts, grabens, volcanics, and continental fragments which were produced during the back-arc spreading, opening of the Japan Sea and the present compressional stage of the Honshu arc. Crown Copyright © 2010 Published by Elsevier B.V. All rights reserved.
1. Introduction Japan is an earthquake country and has suffered heavily from seismic hazards during the long history (e.g., Utsu, 1982; Usami, 2003). Earthquakes that occurred in or around Northeast (NE) Japan are caused by the active subduction and collisions among three lithospheric plates (Fig. 1), where the Pacific plate is subducting from the east beneath the North American and Amurian plates (Hasegawa et al., 2009). Large interplate earthquakes occur actively in the forearc region under the Pacific Ocean, while large intraplate earthquakes occur in the upper crust under the Japan Islands and in the eastern margin of the Japan Sea (EMJS) (Fig. 1). Seismic tomography is proved to be an effective tool to study the structural heterogeneity in the crust and upper mantle and its relationship to seismic and volcanic activities (see Zhao, 2009; Hasegawa et al., 2009 for comprehensive reviews). There is a correlation between the spatial distribution of large crustal earthquakes (M5.7–8.0) under
⁎ Corresponding authors. Research Center for Prediction of Earthquakes and Volcanic Eruptions, Tohoku University, Sendai 980-8578, Japan. Fax: +81 22 264 3292. E-mail addresses: [email protected] (Z. Huang), [email protected] (D. Zhao).
the Japan Islands during 1885–2000 and structural heterogeneities in the crust and uppermost mantle revealed by seismic tomography (Zhao et al., 2002, 2010). So far the 3-D seismic velocity structure under the land area of the Japan Islands has been well determined, but the structures under the Pacific Ocean and the Japan Sea are not well known because very few seismic stations exist in the surrounding oceanic regions and thus the suboceanic earthquakes are poorly located by the routine procedure of the seismic networks that are installed only on the Japan land area. Umino and Hasegawa (1994) and Umino et al. (1995) made an important breakthrough in the seismological study of the oceanic regions around the Japan Islands. They detected sP depth phase from three-component seismograms of shallow earthquakes that occurred under the Japan Sea and the Pacific Ocean recorded by the seismograph stations in NE Japan (Fig. 2). The sP depth phase is a seismic wave that is radiated upward from an earthquake hypocenter as an S-wave and is reflected from a bounce point on the surface or seafloor and at the same time converted from the S-wave to a P-wave which finally arrives at a seismic station. The bounce point of the sP depth phase is close to the earthquake epicenter, and so its travel time is very sensitive to the focal depth. Thus the arrivaltime data of sP depth phase are very useful to locate the suboceanic earthquakes which occur outside a seismic network, because the
0040-1951/$ – see front matter. Crown Copyright © 2010 Published by Elsevier B.V. All rights reserved. doi:10.1016/j.tecto.2010.04.014
178
Z. Huang et al. / Tectonophysics 489 (2010) 177–188
Fig. 1. Distributions of large crustal earthquakes (black stars) during 830–2010 and active faults in Northeast Japan and the Japan Sea. The occurrence years and magnitudes of the large earthquakes are also shown. Gray triangles show the active volcanoes in Northeast Japan. Inset map: tectonic background of Northwest Pacific and Northeast Asia. The dashed line indicates the estimated border of the Amurian plate (Wei and Seno, 1998). Arrows show the directions of the plate motions relative to Northeast Japan.
focal depth of such an earthquake is the most difficult parameter to determine among the four hypocentral parameters (i.e., origin time, latitude, longitude, and focal depth). Therefore the detection of an sP depth phase from a seismogram is just like installing a new seismometer (at the sP bounce point) close to the epicenter. Umino and Hasegawa (1994) and Umino et al. (1995) showed that the suboceanic earthquakes under the Japan Sea and the Pacific Ocean could be located very accurately with the sP depth phase data, and the location accuracy is comparable to that of earthquakes under the seismic network in NE Japan (hypocenter location error b3 km). Umino et al. (1998) further relocated many more crustal earthquakes that occurred in a broad region in EMJS off NE Japan by using the sP depth phase data collected from three-component seismograms. Zhao et al. (2002) used suboceanic earthquakes precisely relocated with sP depth phase data to determine the first 3-D P-wave velocity (Vp) structure under the NE Japan forearc region from the Japan Trench to the Pacific coast, and they showed that the method is very powerful for tomographic imaging outside a seismic network. Later, more detailed 3-D P- and S-wave velocity and Poisson's ratio images in Tohoku and Hokkaido forearc areas under the Pacific Ocean were determined by using more sP depth phase data in addition to a large
number of P and S-wave arrival times (Mishra et al., 2003; Wang and Zhao, 2005; Zhao et al., 2007). Umino and Hasegawa (1994) and Umino et al. (1998) relocated the aftershocks of the 1993 Hokkaido–Nansei–Oki Earthquake (M 7.8) and other crustal earthquakes in EMJS using sP depth phase data. However, so far no one has determined the 3-D velocity structure under the Japan Sea. In this work, we tried to determine the first 3-D Vp structure under EMJS by using a large number of P-wave arrival times from the suboceanic earthquakes investigated by Umino et al. (1998) and the earthquakes that occurred under the NE Japan land area. Our first tomographic images in EMJS shed light on the structural heterogeneities and seismotectonics of this region. 2. Data and method Fig. 3 shows the hypocentral distribution of the 2833 earthquakes used in this study. These events consist of two groups. The first group includes 145 crustal events which occurred in EMJS (Fig. 3a) during April 1992 to December 1997 and have 442 sP depth phase arrival times collected by Umino et al. (1998) and 4740 P-wave arrival times recorded by the Tohoku University
Z. Huang et al. / Tectonophysics 489 (2010) 177–188
179
Fig. 2. An example of three-component seismograms (b) of a crustal earthquake under the Japan Sea recorded by a seismic station in NE Japan (a). The epicentral distance is 140 km. The three dots show the first P and S arrivals and the sP depth phase.
seismic network. The second group consists of 2688 events which occurred under the NE Japan land area during January 1992 to April 2003 and have 170,685 P-wave arrival times. Each of the events was recorded by more than 15 seismic stations and their hypocentral locations are accurate to 1–2 km. The total number of 2833 events in our data set generated 175,425 P-wave arrival times recorded by 330 seismic stations in NE Japan (Fig. 3d). These seismic stations are permanent and portable stations operated by Tohoku University and other Japanese universities, Japan Meteorological Agency, and the High-Sensitivity Seismic Network (Hi-net). The picking accuracy of the P-wave arrival times is estimated to be 0.05–0.15 s. We used the tomographic method of Zhao et al. (1992, 2007) to determine the 3-D Vp structure under the study region (Figs. 1 and 3). The starting one-dimensional (1-D) velocity model for the tomographic inversion was derived from Zhao et al. (1992, 2007) (Fig. 4). Lateral depth variations of the Conrad and Moho discontinuities and the upper boundary of the subducting Pacific slab were taken into account in the 3-D ray tracing and the tomographic inversion, because the existence of the three discontinuities under NE Japan has been established well and their geometries have been determined well (see Zhao et al., 2007
and Hasegawa et al., 2009 for details). Under the Japan Sea, the depths of the Conrad and Moho discontinuities are modified based on the results of seismic explosion experiments (Nishisaka et al., 2001; Iwasaki and Sato, 2009). To express the 3-D velocity structure, a 3-D grid was set up in the study area with a lateral grid interval of 30–50 km. Meshes of grid nodes were set at 10, 25, 40, 65, 90, 120, 160 and 200 km depths in the crust and mantle wedge, and at 5 and 25 km under the upper boundary of the subducting Pacific slab, following the approach of Zhao et al. (1992, 2007). Hypocenter parameters and velocity perturbations at the grid nodes were taken as unknown parameters. The velocity perturbation at any point in the model was calculated by linearly interpolating the velocity perturbations at the eight grid nodes surrounding that point. An efficient 3-D ray tracing technique (Zhao et al., 1992) was used to compute travel times and ray paths accurately. The damped least-squares method was used to invert the large and sparse system of observation equations that relate the observed arrival times to the unknown parameters. The 3-D Vp structure and hypocenter parameters were determined simultaneously in an iterative inversion process. The surface topography and station elevations were taken into account in the 3-D ray tracing and the tomographic inversion.
180
Z. Huang et al. / Tectonophysics 489 (2010) 177–188
Fig. 3. (a) Epicentral distribution of 2688 events (crosses) that are located accurately by the seismic network and 145 suboceanic events (dots) that are relocated by using the sP depth phase data. (b) East–west and (c) north–south vertical cross-sections of the hypocenter locations for the earthquakes shown in (a). (d) Distribution of 330 seismic stations (gray circles) used in this study. The solid triangles show the active volcanoes.
3. Velocity structure and resolution tests Fig. 5 shows the hypocenter distribution of the 145 crustal events in EMJS we relocated as compared with those determined by the seismic network of Tohoku University. The 145 events were relocated by applying the earthquake location scheme of Zhao et al. (1992, 2007) with 3-D ray tracing to both P and sP arrival-time data for the 1-D velocity model (Fig. 4) and the final 3-D velocity model (Figs. 7 and 8). The seafloor topography and sediment layer under the Japan Sea were also taken into account in the earthquake relocation by using the model CRUST2.0 (Laske et al., 2003) which is the updated version of the model CRUST5.1 (Mooney et al., 1998). The formal uncertainties of the hypocenter locations are smaller than 3 km for all the 145 events, thanks to the use of sP depth phase data. The hypocenters of the 145 events located by the seismic network show
a diffusive distribution in a depth range of 0–58 km. After relocation with the sP depth phase, however, most of the events are located shallower than 15 km, only a few events close to the coastline under southern NE Japan have focal depths of 15–21 km (Fig. 5e). The differences in the hypocenter locations resulted from the 1-D and 3-D velocity models are relatively small, which do not exceed 3 km for most of the events (Fig. 5). A few linear dipping features are visible in the hypocenter distribution of the relocated events under EMJS (Fig. 5c–e), which may represent active faults there (Umino and Hasegawa, 1994; Umino et al., 1998). We conducted many checkerboard resolution tests to examine the resolution scale of the tomographic images with our data set following the approach of Zhao et al. (1992, 2007). We first assigned positive and negative velocity anomalies of 3% to all the 3-D grid nodes, then calculated synthetic arrival times for the checkerboard
Z. Huang et al. / Tectonophysics 489 (2010) 177–188
181
synthetic test is the same as that of the checkerboard resolution test (Fig. 6) except for the input model which was constructed from the obtained tomographic results (Figs. 7 and 8) as follows. For the grid nodes with velocity anomalies ≥ +3%, they are changed to 6%, while for those with velocity anomalies ≤ −3%, they are changed to −6%. For the grid nodes with velocity anomalies between −3% and +3%, they are changed to 0%. The velocity perturbation at any point in the model is calculated by linearly interpolating the velocity perturbations at the eight grid nodes surrounding that point. Thus the input model is constructed (Fig. 9a–g). The results of the synthetic test (Fig. 9a′–g′) show that the main features of velocity anomalies in the input model (and also in the real tomographic images) are well recovered, suggesting that those anomalies are reliable features. 4. Anisotropic tomography
Fig. 4. The starting P-wave velocity model adopted for the tomographic inversion. Hc, Hm and Hslab denote depths of the Conrad and Moho discontinuities and the upper boundary of the subducting Pacific slab. The depth variations of the three discontinuities determined by the previous studies (Zhao et al., 1992, 2007; Nishisaka et al., 2001; Iwasaki and Sato, 2009) are adopted in this study. A high-velocity anomaly of 4% is added to the subducting Pacific slab in the starting velocity model (dashed line).
model, and then inverted the synthetic data to see if the assigned velocity anomalies could be recovered or not. The numbers of seismic stations, events and ray paths in the synthetic data are the same as those in the real data set. Random errors with a normal distribution having a standard deviation of 0.1 s were added to the synthetic arrival times before the tomographic inversion. Fig. 6 shows the test results at six depths in the crust and mantle wedge and two layers in the subducting Pacific slab, which indicate that our tomographic model has a resolution of 30–50 km in the horizontal direction and 10–30 km in depth. Fig. 7 shows the tomographic images at 10, 25 and 40 km depths together with the distribution of large crustal earthquakes (M ≥ 6.0) that occurred during 830 to 2010 in the present study area (Utsu, 1982; Usami, 2003). Distributions of active faults and active volcanoes are also shown. The results show that strong lateral heterogeneities exist in the crust and the uppermost mantle under EMJS. A belt of high-velocity (high-V) anomalies is visible along the Japan Sea coast, extending from the surface to the upper mantle. Several low-velocity (low-V) anomalies also exist in the crust and uppermost mantle under the Japan Sea. Although the correlation between the tomography and the distribution of large crustal earthquakes is not very obvious, many large earthquakes seem to be located in or around the low-V zones, as pointed out earlier by Zhao et al. (2002, 2010). However, some large earthquakes are located in or around the high-V zones (Fig. 7). Fig. 8 shows seven E–W vertical cross-sections of P-wave tomography down to 130 km depth together with seismicity, large crustal earthquakes, low-frequency microearthquakes (Hasegawa and Yamamoto, 1994), and active volcanoes in the vicinity of each profile. To confirm the main features of the obtained 3-D Vp structure, we made one more synthetic test (Fig. 9). The procedure of the
We further determined the 3-D P-wave anisotropic tomography beneath the EMJS region using the method of Wang and Zhao (2008). Compared with the isotropic tomography, two additional unknown parameters are added at each grid node to represent the P-wave azimuthal anisotropy. The LSQR algorithm (Paige and Saunders, 1982) was applied to solve the large sparse system of observation equations that relates the arrival-time data with the unknown parameters for the hypocenter locations and the isotropic and anisotropic velocity variations at every grid nodes. We first made the checkerboard resolution test as mentioned above for the anisotropic tomography (Wang and Zhao, 2008). At every grid nodes of the input velocity model (Fig. 10), we assigned isotropic velocity anomalies of ± 3% alternatively and two anisotropic parameters which represent the fast Vp directions of 22.5° or 112.5° with anisotropic velocity variations of ∼ 4.24% (see Wang and Zhao, 2008 for details). The test results show that both the pattern and amplitudes of the isotropic and anisotropic velocity anomalies are well recovered at the depths of 10 and 25 km (Fig. 10). At 40 km depth, the resolution becomes lower under the Japan Sea. Fig. 11 shows the distribution of P-wave fast-velocity directions (FVDs) together with the isotropic velocity anomalies at different depths under the EMJS. The general pattern of the isotropic velocity images is similar to that obtained from the isotropic tomography (Fig. 7), e.g., the high-V belt along the Japan Sea coast and the low-V zones under the EMJS, though the amplitude of the velocity anomalies becomes smaller. The anisotropic tomography results (Fig. 11) show that the FVDs are complex in the upper crust, while in the lower crust and uppermost mantle under EMJS, the FVDs with large amplitudes are generally oriented in NW–SE to E–W directions. This issue is discussed in detail in the next section. 5. Interpretation and discussion Until the early Miocene, Japan was part of the Asian margin (Sato, 1994). The northeast Honshu arc was constructed by backarc spreading during 21–18 Ma when Yamato basin formed, and by the subsequent rifting during 19–13.5 Ma (when Northern Honshu rift system formed) and island-arc uplifting up to the present (Ohtake et al., 2002; Yamada and Yoshida, 2004; Yoshida et al., 2005). The EMJS region has been under the convergent tectonics since the Pliocene under the E–W compressional stress regime (Tamaki and Honza, 1985). The relative motion of 9– 17 mm/year between the Amurian and North American plates is accommodated at EMJS as shown by the GPS observations (Wei and Seno, 1998; Jin et al., 2007). Sagiya et al. (2000) studied the crustal deformation of the Japan Islands using GPS measurements and suggested that EMJS has a much larger strain rate than the surrounding areas. Hence the EMJS region has a high seismic
182
Z. Huang et al. / Tectonophysics 489 (2010) 177–188
Fig. 5. Hypocenter distributions of the 145 shallow earthquakes under the eastern margin of the Japan Sea in (a) map view and in (b) north–south and (c–e) east–west vertical crosssections. The open circles show the hypocenters located by the seismic network of Tohoku University, while blue and red crosses denote the hypocenters determined in this study for the 1-D and 3-D velocity models using P and sP depth phase data (see text for details).
potential and there may be a common mechanism for the large earthquakes there (Ohtake et al., 2002). Previous studies suggest that the generation of a large earthquake is not a pure mechanical process, but is closely related to the physical and chemical properties of materials in the crust and upper mantle, such as magma, fluids, etc. Zhao et al. (2002, 2010) studied large crustal earthquakes under the Japan Island and concluded that most of the earthquakes occurred in or around the low-V zones which may represent the weak zones in the crust. Under the volcanic front, the crustal weakening may be caused by the upward intrusion of the magma and fluids from the mantle wedge which is caused by the subduction of the Pacific Plate and dehydration of the slab (Hasegawa et al., 2005, 2009; Zhao et al., 2002, 2010). However, the correlation between the velocity anomalies and the distribution of large crustal earthquakes and active faults becomes lower under the Japan Sea. Although many earthquakes are located in or around the low-V zones, some are in the high-V areas (Fig. 7). Several factors may cause the lower correlation. Firstly, the tomographic resolution is lower under the Japan Sea compared with that under the land area, and so some smaller velocity anomalies may not be imaged well. Secondly, the hypocenter locations for some of the large historic earthquakes are less accurate. Note that the hypocenter locations of the large earthquakes that occurred after about 1900
represent the initial points of earthquake ruptures because the earthquakes were located by the modern seismic network in Japan, and they may be accurate to about 10 km or less. In contrast, the hypocenters of large historic earthquakes before 1900 represent the centroid locations of the rupture zones because they were estimated from the macroscopic damages of the earthquakes, and so they may have uncertainties of tens of kilometers (Usami, 2003). Finally, the discrepancy itself may be the clue that the structure under the Japan Sea is very complex and the ruptures of the large crustal earthquakes are affected by various factors. The P-wave anisotropic images show another piece of evidence for the structural heterogeneity under the EMJS. The anisotropic image in the upper crust is complex, while the FVDs with large amplitudes are E–W or NW–SE in the lower crust and uppermost mantle, which are similar to those revealed by Wang and Zhao (2008) and to the S-wave splitting results (Nakajima et al., 2006). The upper-crust anisotropy is mainly caused by the stress-induced microcracks, the cracks and fractures accompanied with the active faults, and the intrinsic rock anisotropy resulting from preferred orientation of the minerals (Kaneshima, 1990). Under the EMJS, the complex anisotropic structures may be caused by the frequent occurrence of large crustal earthquakes and the formation and closing of active faults that change the stress status and the orientations of the cracks and fractures, as well as by the intrusion
Z. Huang et al. / Tectonophysics 489 (2010) 177–188
183
Fig. 6. Results of a checkerboard resolution test for P-wave velocity structure at six depths in the crust and mantle wedge (a–f) and at two layers in the subducting Pacific slab (g, h). The dashed lines in (d–f) indicate the upper boundary of the Pacific slab. The depth to each layer is shown at the upper-left corner of each map. The velocity perturbation scale is shown below (f).
and transportation of the magma and fluids that change the preferred orientations of the minerals. Detailed tomographic images were determined for the source areas of the 2004 and 2007 Niigata earthquakes (both with M 6.8) in southwestern part of NE Japan (Fig. 1) (e.g., Nakajima and Hasegawa, 2008; Xia et al., 2008). An anomaly with pronounced low-velocity and high Poisson's ratio in the lower crust and uppermost mantle was revealed under the Niigata source areas, which also exhibits high conductivity and high He3/He4 ratio (Ogawa and Honkura, 2004; Uyeshima et al., 2005; Horiguchi and Matsuda, 2008), indicating the existence of fluids under EMJS rising from the mantle wedge as a result of slab dehydration and corner
flow in the mantle wedge. The results suggest that high-temperature arc and back-arc magmas as well as fluids released from the ascending flow in the mantle wedge can affect the rupture nucleation of large crustal earthquakes. This may be one reason for the frequent occurrence of large crustal earthquakes in the EMJS (Fig. 7). Many normal faults and fault-related rifts and grabens developed in Early to Middle Miocene, simultaneous with backarc spreading of the Japan Sea (Okamura et al., 1995). The normal faults reactivated as reverse faults during the inversion stage due to an increase in compressional stress since 1–2 Ma ago (Tamaki and Honza, 1985). Some large thrust faults during
184
Z. Huang et al. / Tectonophysics 489 (2010) 177–188
Fig. 7. P-wave velocity images at depths of (a) 10, (b) 25, and (c) 40 km under the study area. Red and blue colors denote low and high velocities, respectively. Red stars denote large crustal earthquakes (M ≥ 6.0) that occurred from 830 to 2008 (Utsu, 1982; Usami, 2003). Black triangles denote active arc volcanoes. Active faults are shown by thin lines. Low-frequency microearthquakes within 5 km of each depth are shown by red dots. The velocity perturbation scale and the earthquake magnitude scale are shown at the bottom.
large earthquakes may cut through the whole crust (Fukao and Furumoto, 1975), and seawater may permeate down to the deep crust through these active faults during many earthquake cycles in the long geological history, which may be another source of crustal fluids in addition to the fluids rising from the mantle wedge (Zhao et al., 2002). In our tomographic images, low-V zones are visible widely in the crust under the Japan Sea (Figs. 7 and 8), suggesting that the velocities are reduced by the fluids. The seismic velocity anomalies under EMJS may also reflect the complicated geologic structures such as the alternate rift zones, ridges, basins, horsts, grabens, volcanics, and continental fragments which were produced during the back-arc spreading, opening of the Japan Sea and the current compressional stage of the Honshu arc (Ohtake et al., 2002; Yoshida et al., 2005). In the uppermost mantle, significant high-V anomalies (∼4– 6%) are imaged under the Japan Sea side, while low-V zones are revealed under the NE Japan inland area (Figs. 7 and 8). The result is well consistent with that of seismic explosion experiments carried out across the EMJS, which shows higher subMoho P-wave velocity under the Japan Sea (∼8.0 km/s) than that under the NE Japan land area (∼ 7.5–7.7 km/s) (Nishisaka et al., 2001; Iwasaki and Sato, 2009). Mineral physics studies of lowercrust mafic xenoliths sampled in EMJS suggested that the high-V anomalies in the lower crust under the Japan Sea coast (Figs. 7 and 8) represent cooled igneous rocks formed in the back-arc igneous period (21–13.5 Ma) (Nishimoto et al., 2008). 6. Conclusions We determined the first P-wave tomography and azimuthal anisotropy under the eastern margin of the Japan Sea using a large number of arrival-time data from local earthquakes in the region. Data from 145 suboceanic earthquakes under the Japan Sea which are precisely relocated by using sP depth phase data made it possible to determine the 3-D structure under the
Japan Sea. Our results show that strong lateral heterogeneities exist in the crust and uppermost mantle under the EMJS region, which may reflect the complicated geologic structures such as the alternate rift zones, ridges, basins, horsts, grabens, volcanics, and continental fragments which were produced during the opening of the Japan Sea and the present compressional stage of the Honshu arc. The structural heterogeneities may have affected the seismogenesis in the region. Many large crustal earthquakes occurred in or around lowvelocity zones which may represent weak sections of the seismogenic crust. However, the correlation is reduced under EMJS due to the lower tomographic resolution, less accurate hypocenter locations and the very complex crustal structures under the Japan Sea. The complex P-wave anisotropy also reflects the strong heterogeneities under the EMJS. In the arc and back-arc areas of NE Japan, the crustal weakening may be caused by both high-temperature arc magma and fluids released from the mantle diapirs rising from the mantle wedge. Sea water that permeates down to the deep crust through active faults during many earthquake cycles in the long geological history may be another source of crustal fluids. These results indicate that the generation of large crustal earthquakes is closely related to the tectonic environment and the physical and chemical properties of materials in the crust and upper mantle, such as magma, fluids, etc. Future studies using data from a portable seismic network of oceanbottom seismometers in the Japan Sea are expected to clarify the seismotectonics, back-arc magmatism and mantle dynamics in the EMJS region. Acknowledgements We thank the Hi-net, Tohoku University and Japanese National University seismic networks for providing the waveform and arrival-time data used in this study. This work was supported
Z. Huang et al. / Tectonophysics 489 (2010) 177–188
185
Fig. 8. Seven vertical cross-sections of P-wave velocity image from the surface down to 130 km depth along the profiles shown on the insert map. Red and blue colors denote slow and fast velocities, respectively. Red stars and black triangles denote large earthquakes and active volcanoes within a 30-km width along each profile. Note that the focal depths of the large crustal earthquakes are set at 10 km because the accurate focal depths are unclear for most of these historic earthquakes. White and red circles denote the normal earthquakes and the low-frequency microearthquakes that occurred within a 10-km width of each profile. The horizontal bar atop each cross-section shows the land area where seismic stations are located. The three curved lines denote the Conrad and Moho discontinuities and the upper boundary of the subducting Pacific slab. The velocity perturbation scale and the earthquake magnitude scale are shown at the bottom.
partially by a grant (Kiban-A 17204037) to D. Zhao from Japan Society for the Promotion of Science (JSPS), a grant (40634021) from the National Natural Science Foundation of China, and by the Scientific Research Foundation of Graduate School of Nanjing University. We are grateful to Drs. J. Wang, G. Jiang, S. Miura, T.
Okada, J. Nakajima and N. Uchida for helpful discussions. Prof. M. Liu (Editor) and two anonymous reviewers provided thoughtful comments and suggestions which improved the manuscript. Most of the figures were made by using the Generic Mapping Tools (Wessel and Smith, 1991).
186
Z. Huang et al. / Tectonophysics 489 (2010) 177–188
Fig. 9. Results of a synthetic resolution test for the cross-sections shown in Fig. 8 (see the text for details). The input model is shown in (a–g), while the inversion results are shown in (a′–g′). Crosses and circles denote the high- and low-velocity anomalies, respectively. The scale of the anomalies is showed at the bottom. The other labeling is the same as that in Fig. 8.
References Fukao, Y., Furumoto, M., 1975. Mechanism of large earthquakes along the eastern margin of the Japan Sea. Tectonophysics 25, 247–266. Hasegawa, A., Yamamoto, A., 1994. Deep, low-frequency microearthquakes in or around seismic low-velocity zones beneath active volcanoes in northeastern Japan. Tectonophysics 233, 233–252.
Hasegawa, A., Nakajima, J., Umino, N., Miura, S., 2005. Deep structure of the northeastern Japan arc and its implications for crustal deformation and shallow seismic activity. Tectonophysics 403, 59–75. Hasegawa, A., Nakajima, J., Uchida, N., Okada, T., Zhao, D., Matsuzawa, T., Umino, N., 2009. Plate subduction, and generation of earthquakes and magmas in Japan as inferred from seismic observations: an overview. Gondwana Res. 16, 370–400. Horiguchi, K., Matsuda, J., 2008. On the change of 3He/4He ratios in hot spring gases after the Iwate-Miyagi Nairiku earthquake in 2008. Geochem. J. 42, e1–e4.
Z. Huang et al. / Tectonophysics 489 (2010) 177–188
187
Fig. 10. The input model (a–c) and output results (d–f) of a checkerboard resolution test for P-wave anisotropic tomography. The azimuth and length of the bars represent the fastvelocity direction and the anisotropic amplitude, respectively. The scale for the anisotropic amplitude is shown at the bottom. The other labeling is the same as that in Fig. 6.
Iwasaki, T., Sato, H., 2009. Crust and upper mantle structure of island arc being elucidated from seismic profiling with controlled sources in Japan. Zisin 61, S165–S176. Jin, S., Park, P., Zhu, W., 2007. Micro-plate tectonics and kinematics in Northeast Asia inferred from a dense set of GPS observations. Earth Planet. Sci. Lett. 257, 486–496. Kaneshima, S., 1990. Origin of crustal anisotropy: shear wave splitting studies in Japan. J. Geophys. Res. 95, 11121–11133. Laske, G., Masters, G., Reif, C., 2003. CRUST2.0: A new global crustal model at 2 x 2 degrees. http://mahi.ucsd.edu/Gabi/rem.dir/crust/crust2.html. Mishra, O.P., Zhao, D., Umino, N., Hasegawa, A., 2003. Tomography of northeast Japan forearc and its implications for interplate seismic coupling. Geophys. Res. Lett. 30 GL017736. Mooney, W., Laske, G., Masters, G., 1998. CRUST5.1: a global crustal model at 5 x 5 degrees. J. Geophys. Res. 103, 727–747. Nakajima, J., Hasegawa, A., 2008. Existence of low-velocity zones under the source areas of the 2004 Niigata–Chuetsu and 2007 Niigata–Chuetsu–Oki earthquakes inferred from travel-time tomography. Earth Planets Space 60, 1127–1130. Nakajima, J., Shimizu, J., Hori, S., Hasegawa, A., 2006. Shear-wave splitting beneath the southwestern Kurile arc and northeastern Japan arc. Geophys. Res. Lett. 33, L05305. Nishisaka, H., Shinohara, M., Sato, T., Hino, R., Mochizuki, K., Kasahara, J., 2001. Crustal structure of the Yamato basin and the margin of the northeastern Japan Sea using ocean bottom seismographs and controlled sources. Zisin 54, 365–379. Nishimoto, S., Ishikawa, M., Arima, M., Yoshida, T., Nakajima, J., 2008. Simultaneous high P–T measurements of ultrasonic compressional and shear wave velocities in
Ichino-megata mafic xenoliths: their bearings on seismic velocity perturbations in lower crust of northeast Japan arc. J. Geophys. Res. 113, B12212. Ogawa, Y., Honkura, Y., 2004. Mid-crustal electrical conductors and their correlations to seismicity and deformation at Itoigawa–Shizuoka Tectonic Line, central Japan. Earth Planets Space 56, 1285–1291. Ohtake, M., Taira, A., Ohta, Y., 2002. Active Faults and Seismo-Tectonics of the Eastern Margin of the Japan Sea. Univ. Tokyo Press, Tokyo, p. 201. Okamura, Y., Watanabe, M., Morijiri, R., Satoh, M., 1995. Rifting and basin inversion in the eastern margin of the Japan Sea. Island Arc 4, 166–181. Paige, C., Saunders, M., 1982. LSQR: an algorithm for sparse linear equations and sparse least squares. ACM Trans. Math. Software 8, 43–71. Sato, H., 1994. The relationship between late Cenozoic tectonic events and stress field and basin development in northeast Japan. J. Geophys. Res. 99, 22261–22274. Sagiya, R., Miyazaki, S., Tada, T., 2000. Continuous GPS array and present-day crustal deformation of Japan. Pure Appl. Geophys. 157, 2303–2322. Tamaki, K., Honza, E., 1985. Incipient subduction and deduction along the eastern of the Japan Sea. Tectonophysics 119, 381–406. Umino, N., Hasegawa, A., 1994. Aftershock focal depths of the 1993 Hokkaido–Nansei– Oki earthquake estimated from sP depth phase at small epicentral distances. J. Phys. Earth 42, 321–329. Umino, N., Hasegawa, A., Matsuzawa, T., 1995. sP depth phase at small epicentral distances and estimated subducting plate boundary. Geophys. J. Int. 120, 356–366. Umino, N., Hori, S., Hasegawa, A., 1998. Depth Distribution of Crustal Earthquakes in the Eastern Margin of Japan Sea Determined by Using sP Depth-Phase Data. Abstract of
188
Z. Huang et al. / Tectonophysics 489 (2010) 177–188
Fig. 11. P-wave anisotropic tomographic images at depths of (a) 10, (b) 25, and (c) 40 km under the study area. The azimuth and length of the bars represent the fast-velocity direction and the anisotropic amplitude, respectively. The scale for the anisotropic amplitude is shown at the right bottom. The other labeling is the same as that in Fig. 7.
a Presentation at the Joint Meeting of Japan Earth and Planetary Sciences held in May 1998 in Chiba, Japan. Usami, T., 2003. Materials for Comprehensive List of Destructive Earthquakes in Japan. University of Tokyo Press, Tokyo. Utsu, T., 1982. Catalog of large earthquakes in the region of Japan from 1885 through 1980. Bull. Earthquake Res. Inst. Univ. Tokyo 57, 401–463. Uyeshima, M., Ogawa, Y., Honkura, et al., 2005. Resistivity imaging across the source region of the 2004 mid-Niigata prefecture earthquake (M6.8), central Japan. Earth Planets Space 57, 441–446. Wang, J., Zhao, D., 2008. P-wave anisotropic tomography beneath Northeast Japan. Phys. Earth Planet. Inter. 170, 115–133. Wang, Z., Zhao, D., 2005. Seismic imaging of the entire arc of Tohoku and Hokkaido in Japan using P-wave, S-wave and sP depth-phase data. Phys. Earth Planet. Inter. 152, 144–162. Wei, D., Seno, T., 1998. Determination of the Amurian plate motion. In: Flower, M.F.J., et al. (Ed.), Mantle Dynamic and Plate Interactions in East Asia. Geodynamics Series: Am. Geophys. Union, 27, p. 419. Wessel, P., Smith, W., 1991. Free software helps map and display data, EOS Trans. AGU 72, 441.
Xia, S., Zhao, D., Qiu, X., 2008. The 2007 Niigata earthquake: effect of arc magma and fluids. Phys. Earth Planet. Inter. 166, 153–166. Yamada, R., Yoshida, T., 2004. Volcanic sequences related to Kuroko mineralization in the Hokuroku district, Northeast Japan. Resour. Geol. 54, 399–412. Yoshida, T., Nakajima, J., Hasegawa, A., Sato, H., Nagahashi, Y., Kimura, J., Tanaka, A., Prima, O., Ohguchi, K., 2005. Distribution of Quaternary volcanoes and mantle structures in the NE Honshu arc. Quatern. Res. 44, 195–216. Zhao, D., 2009. Multiscale seismic tomography and mantle dynamics. Gondwana Res. 15, 297–323. Zhao, D., Hasegawa, A., Horiuchi, S., 1992. Tomographic imaging of P and S wave velocity structure beneath northeastern Japan. J. Geophys. Res. 97, 19909–19928. Zhao, D., Mishra, O.P., Sanda, R., 2002. Influence of fluids and magma on earthquakes: seismological evidence. Phys. Earth Planet. Inter. 132, 249–267. Zhao, D., Wang, Z., Umino, N., Hasegawa, A., 2007. Tomographic imaging outside a seismic network: Application to the northeast Japan arc. Bull. Seismol. Soc. Am. 97, 1121–1132. Zhao, D., Santosh, M., Yamada, A., 2010. Dissecting large earthquakes in Japan: role of arc magma and fluids. Island Arc 19, 4–16.