Imaging the source regions of normal faulting sequences induced by the 2011 M9.0 Tohoku‐Oki earthquake

Transcription

1 GEOPHYSICAL RESEARCH LETTERS, VOL. 4, , doi:1.12/grl.14, 213 Imaging the source regions of normal faulting sequences induced by the 211 M9. Tohoku-Oki earthquake Aitaro Kato, 1 Toshihiro Igarashi, 1 Kazushige Obara, 1 Shinichi Sakai, 1 Tetsuya Takeda, 2 Atsushi Saiga, 3 Takashi Iidaka, 1 Takaya Iwasaki, 1 Naoshi Hirata, 1 Kazuhiko Goto, 4 Hiroki Miyamachi, 4 Takeshi Matsushima, Atsuki Kubo, 6 Hiroshi Katao, 7 Yoshiko Yamanaka, 8 Toshiko Terakawa, 8 Haruhisa Nakamichi, 8 Takashi Okuda, 8 Shinichiro Horikawa, 8 Noriko Tsumura, 9 Norihito Umino, 1 Tomomi Okada, 1 Masahiro Kosuga, 11 Hiroaki Takahashi, 12 and Takuji Yamada 12 Received 27 October 212; revised 14 December 212; accepted 18 December 212; published 29 January 213. [1] Intense swarm-like seismicity associated with shallow normal faulting was induced in Ibaraki and Fukushima prefectures, Japan, following the 211 Tohoku-Oki earthquake. This seismicity shows a systematic spatiotemporal evolution, but little is known of the heterogeneity in crustal structure in this region, or its influence on the evolution of the seismicity. Here, we elucidate a high-resolution model of crustal structure in this region and determine precise hypocenter locations. Hypocenters in Ibaraki Prefecture reveal a planar earthquake alignment dipping SW at ~4, whereas those in Fukushima Prefecture show a more complex distribution, consisting of conjugate sets of aligned small earthquakes. On the north of the hypocenter of the largest earthquake in the sequence (the M7. Iwaki earthquake), we imaged a high-velocity body at shallow depths that lacks aftershock seismicity. Based on fault source models, the large-slip region of the Iwaki earthquake is situated along a zone that roughly coincides with this high-velocity body. We delineated a separate low-velocity anomaly directly beneath the hypocenter of the Iwaki earthquake, indicating crustal fluids in this region. We hypothesize that strong crust underwent structural failure due to the infiltration of crustal All Supporting Information may be found in the online version of this article. 1 Earthquake Research Institute, University of Tokyo, Tokyo, Japan. 2 National Research Institute for Earth Science and Disaster Prevention, Tsukuba, Japan. 3 Tono Research Institute of Earthquake Science, Association for the Development of Earthquake Prediction, Mizunami, Japan. 4 Graduate School of Science and Engineering, Kagoshima University, Kagoshima, Japan. Institute of Seismology and Volcanology, Faculty of Sciences, Kyushu University, Shimabara, Japan. 6 Kochi Earthquake Observatory, Faculty of Science, Kochi University, Kochi, Japan. 7 Disaster Prevention Research Institute, Kyoto University, Uji, Japan. 8 Earthquake and Volcano Research Center, Graduate School of Environmental Studies, Nagoya University, Nagoya, Japan. 9 Faculty of Science, Chiba University, Inage, Japan. 1 Research Center for Prediction of Earthquakes and Volcanic Eruptions, Graduate School of Science, Tohoku University, Sendai, Japan. 11 Graduate School of Science and Technology, Hirosaki University, Hirosaki, Japan. 12 Institute of Seismology and Volcanology, Graduate School of Science, Hokkaido University, Sapporo, Japan. Corresponding author: A. Kato, Earthquake Research Institute, University of Tokyo, Tokyo, Japan. (akato@eri.u-tokyo.ac.jp) 213. American Geophysical Union. All Rights Reserved /13/1.12/grl fluids into the seismogenic zone from deeper levels, causing the Iwaki earthquake. Citation: Kato, A., et al. (213), Imaging the source regions of normal faulting sequences induced by the 211 M9. Tohoku-Oki, Geophys. Res. Lett., 4, , doi:1.12/grl Introduction [2] The 11 March 211 Tohoku-Oki earthquake (magnitude (M) 9.) ruptured a mega-thrust fault off the eastern shore of northern Honshu, Japan, where the Pacific plate is subducting beneath an overriding continental plate at a convergence rate of ~8. cm/year [e.g., Loveless and Meade, 21]. The large extensional stress perturbations associated with the Tohoku-Oki earthquake [e.g., Ozawa et al., 211] resulted in widespread seismicity across much of NE Japan [e.g., Toda et al., 211]. A remarkable aspect of this induced seismicity was a strong increase in swarm-like shallow normal faulting along the Pacific coast of NE Japan, most notably in Ibaraki and Fukushima prefectures (Figure 1). The progression of this induced seismicity shows a systematic spatiotemporal evolution. Specifically, an initial M.7 event occurred in the northern part of Ibaraki Prefecture about 8 minutes after the Tohoku-Oki mainshock rupture (Figure 1a), which subsequently led to the development of swarm-like activity with a M6.1 earthquake on 19 March, slightly to the north of the M.7 event. After these initial events, the activity then jumped northeastward, to the southern end of Fukushima Prefecture, close to its southern coastline, accompanied by a M6. earthquake on 23 March. Finally, in the area between the first two clusters (swarms) of earthquakes, the largest earthquake in the sequence (the M7. Iwaki earthquake) took place on 11 April 211 [Hikima, 212; Fukushima et al., 213]. The Iwaki earthquake was accompanied by a series of surface ruptures up to ~1 km long along two faults, characterized by steeply dipping fault planes [Mizoguchi et al., 212]. [3] Some previous studies have proposed that these sequences of shallow normal faulting were caused by an increase in extensional differential stress, possibly due in turn to the southwesterly extension of the large slip zone of the Tohoku-Oki mainshock rupture that occurs along the plate boundary (inset in Figure 1b) [e.g., Kato et al., 211a; Imanishi et al., 212; Kato and Igarashi, 212]. Similar processes have been proposed in relation to the 21 Maule earthquake in Chile [Farías et al., 211; Ryder

2 (a) 37.2 km 1 211/3/11-3/21 211/3/22-4/11 211/4/11-4/3 M6. M7. 37 M (b) M.7 initial event km 211/3/23(M6.) /4/12(M6.4) Fukushima 211/4/11(M7.) 37 Y = 12km Y = 6 km Y = km Y = -6 km Y = -12km Y = -18km /3/19(M6.1) Ibaraki X = km 211/3/11(M.7) zone Large slip Depth (km) M6 M4 M2 Stations Active faults Figure 1. (a) Spatiotemporal evolution of induced seismicity in the northern part of Ibaraki Prefecture and the southern part of Fukushima Prefecture. The earthquakes plotted are listed in the JMA catalog and occurred at depths shallower than 1 km, with M 1.. (b) Map of seismic stations and earthquakes used in the tomography analysis, with earthquakes shown as circles with radii scaled to earthquake magnitude and colored according to depth. The grid used in the tomographic analysis is plotted with gray crosses. The open squares indicate the locations of temporary offline (64 closely spaced squares) and permanent online seismic stations. The moment tensors (in red and white) of large events (M 6.) were determined by NIED. The red lines delineate the surface traces of major active faults. Inset map shows the location of the study area with respect to prefectures in Japan and the large-slip zone of the 211 Tohoku-Oki mainshock, from Kato and Igarashi [212]. et al., 212]. In addition to changes in regional stress patterns in the crust, it has been argued that local heterogeneities in crustal structure play an important role in controlling the spatiotemporal evolution of seismicity and associated faulting behavior [e.g., Michael and Eberhart-Phillips, 1991; Chiarabba and Amato, 23; Chiarabba et al., 29; Kato et al., 29, 21; Zhao et al., 211]. However, there are few constraints on heterogeneous crustal structure in the source region or the relationship between the heterogeneity and induced seismicity. [4] While online seismic networks of the National Research Institute for Earth Science and Disaster Prevention (NIED) and the Japan Meteorological Agency (JMA) were in operation throughout the Japanese islands, the average spacing of the stations (2 3 km) was too large to allow for a high-resolution analysis of crustal heterogeneities in the source region. Therefore, to assess the high-resolution velocity structure, we deployed a dense network of temporary seismic stations that became operational on 28 March 211 (Figure 1b). Seismic tomography analysis combined with such a dense network is a powerful tool for performing high-resolution imaging of crustal structures, as well as for determining precise earthquake locations. These seismic data allow us to evaluate and discuss the relationships between heterogeneities in crustal structure and the spatiotemporal evolution of seismicity, as well as assess the possible role of crustal fluids in these processes. 274

3 2. Data and Methods [] The seismic network deployed for this study consists of an array of 64 closely spaced seismometers (average station spacing of <4 km; Figure 1b) with a 1 or 2 Hz natural frequency, which continuously recorded three-component signals at a sampling rate of 1 or 2 Hz. The arrival times of both P- and S-waves for a total of 1348 earthquakes of M 2., which are listed in the JMA catalog (for the period between 28 March and 31 October 211), were manually picked from the waveforms observed by both the temporary and permanent seismic stations. To model the velocity structure and to determine precise hypocenters in the target region, the double-difference tomography method [Zhang and Thurber, 23] was applied to the P- and S-wave data for each earthquake. Within the target area, the grid spacing for the tomography is 3 km, both horizontally and vertically (Figure 1b). A series of tests was carried out to evaluate the resolution of the final velocity structure models (Supporting Information, Figs. S1 and S2), and based on those results, we deduced that the P-wave velocity (V p ) structure within the target area is quite well resolved. The root mean square (RMS) travel time residual reduced from.19 s to.9 s after 2 iterations. [6] We also relocated a number of additional small magnitude (M < 2.) earthquakes that were not used in the tomographic analysis. We first applied an automated picking algorithm to the waveforms of small earthquakes recorded by our dense seismic network from July to October 211. We considered the source locations of the 1348 seismic events with M 2. relocated by tomographic analysis to represent master events for these additional 748 smaller earthquakes, such that the detailed evaluation of hypocenter distributions in the present study uses a collective total of 8,86 earthquakes (Figure 2 and Supporting Information, Figure S3). For details, please see the Supporting Information, Text S1. 3. Results [7] Depth sections of the relocated hypocenters in Ibaraki reveal aligned earthquakes that define a plane dipping SW at ~4 (see Y < 3 km in Figures 2 and S3). The down-dip extent of the SW-dipping alignment is about 1 km. Minor W2S E2N Depth(km) Depth(km) Depth(km) Y = 1 km Y = 12 km Y = 9 km 4/12(M6.4) 3/23(M6.) Y = 3 km Y = km Y = -3 km 4/11(M7.) Y = -9 km 3/19(M6.1) Y = -12 km Y = -1 km Depth(km) Y = 6 km Y = -6 km Y = -18 km Cross-section(km) δvp/vp (%) 1 Figure 2. Vertical depth sections of V p velocity perturbations and nearby earthquakes. The cross-sections are constructed along lines drawn from W2S to E2N (see Figure 1b). Relocated earthquakes (superimposed gray circles) correspond to those distributed within 1. km (laterally) of each vertical cross-section. The masked areas marked by gray color on these vertical depth sections correspond to regions where model resolution is relatively low (as defined in the Supporting Information, Fig. S1). The red arrows at the top of each section correspond to the locations of surface ruptures. The red and white moment tensor solutions for the largest earthquakes are shown using a lower hemisphere projection rotated into the plane of each section. 27

4 (a) (b) (c) /23(M6.) (d) 37 HV /19(M6.1) 4/11(M7.) 4/11(M7.) km 1 Depth 3 km Depth 6 km δvp/vp (%) First (M.7) Depth 9 km Average Vp (depth 3 and 6 km) (km/s) Figure 3. Map views of observed V p perturbations at three representative depths: (a) 3 km, (b) 6 km, and (c) 9 km. Relocated earthquakes (superimposed dots) correspond to those distributed (vertically) within 1. km of each depth section. The white dashed lines denote the location of a high-velocity body (HV) that is discussed in the main text. (d) Average V p at depths between 3 and 6 km. The small dots indicate the locations of epicenters of the JMA catalog that occurred before the M7. Iwaki earthquake, but subsequent to the Tohoku-Oki rupture. The red star demarcates the epicenter of the largest event: the M7. Iwaki earthquake. The white stars indicate the locations of epicenters of other large earthquakes. The red lines demarcate the traces of major active faults. NE-dipping alignments are also visible on some sections (e.g., C1 and C11 in Fig. S3). In contrast, the relocated hypocenters in the southern part of Fukushima Prefecture show complex distributions, consisting of numerous small planar alignments of earthquakes that have various orientations (Y > 6 km in Figure 2). Although in the western region many of these planar alignments of earthquake hypocenters dip to the SW at ~4 in sections C, C6 and C8 in Fig. S3, some additional alignments are also visible on these plots, including WNW-, NW-, and SE-dipping alignments (C1 C4, C7, and C9 in Fig. S3). Some of these planar alignments appear to represent conjugate sets of planes that dip in opposite directions (C1, C3, and C4 in Fig. S3). Furthermore, the distributions of earthquake epicenters reveal a remarkable change between Ibaraki and Fukushima prefectures. For instance, below Ibaraki Prefecture, the epicenters are tightly clustered along a NNW SSE trending corridor about 7 km wide, whereas in Fukushima Prefecture the epicenters are more broadly (~4 km wide) distributed (Figure 1). [8] Aftershocks associated with the M7. Iwaki earthquake appear to be roughly aligned along planes dipping towards the SW at angles of approximately 4 (gray dashed lines in the depth sections in Figure 2), defining somewhat vague trends in comparison with the clearly aligned relocated hypocenters observed in Ibaraki Prefecture (i.e., on the depth sections in Figure 2 where Y < 3 km). The down-dip extent of the aftershock areas associated with the Iwaki earthquake is limited (<1 km). Notably, the surface extensions of the inferred fault planes defined by aftershock events roughly coincide with the locations of observed surface ruptures (arrows in Figure 2). Curiously, however, there are no aftershocks recorded at depths shallower than 4 km that might have been used to link these known surface ruptures to the inferred fault planes that dip at 4 to the SW. Because the dip angles of these surface ruptures are reported to be nearly vertical (as observed in outcrops) [e.g., Mizoguchi et al., 212], we suppose that the dip angle of these fault planes must change in a listric fashion from 4 at depth to near vertical angles as they approach the surface. Indeed, these depth variations of fault dips are supported by an analysis of coseismic radar interferograms [Fukushima et al., 213]. A similar pattern of seismicity has been reported for the 199 Grevena Ms 6.6 normal faulting earthquake in northern Greece [Chiarabba and Selvaggi, 1997]. [9] The depth sections and maps of P-wave velocities reveal low-velocity anomalies at depths of km within the study area, directly beneath earthquake clusters (Figures 2 and 3). In particular, a relatively large volume, low-velocity body appears to be present directly beneath the hypocenter of the Iwaki earthquake at depths of ~9 km ( 3km< Y < 3 km, in Figure 2). In contrast, on the northern side of the mainshock hypocenter of the Iwaki earthquake, a comparatively high-velocity body is imaged at shallower depths of 3 and 6 km (areas labeled HV and encircled by white dashed lines in Figures 3a and 3b). Within this high-velocity body, seismicity is lower than in the surrounding regions. Some of these same features of the V p model are also present in the V s model (Supporting Information, Fig. S4). 4. Discussion and Conclusions [1] Most of the observable planar alignments of earthquake in depth sections throughout the study area exhibit dip angles of 4 to the SW (Figures 2 and S3). These dip angles are not relatively optimal for normal faulting [Sibson, 2]. This might suggest that the normal faulting earthquakes occurred along weak pre-existing faults. The progression and distribution of seismicity shows some important differences between northern Ibaraki and southern Fukushima prefectures. For instance, in Ibaraki Prefecture, several fault planes dip 276

5 towards the SW, concentrated within a relatively narrow region, whereas the region beneath Fukushima Prefecture contains numerous, widely distributed fault planes with a range of dip directions. The occurrence of this complex array of fault ruptures in Fukushima Prefecture, including those linked with the M7. Iwaki earthquake, is consistent with observations of multi-segmented surface deformation revealed by analysis of coseismic radar interferograms [e.g., Fukushima et al., 213]. [11] On the northern side of the hypocenter of the Iwaki earthquake, a high-velocity body with a lack of aftershock seismicity was imaged at shallow depths of 3 and 6 km (Figures 2 and 3). In addition, a fault source model of the Iwaki earthquake, based on geodetic and strong motion waveform analysis [e.g., Fukushima et al., 213; Hikima, 212], shows that large-slip patches along the associated SW-dipping faults are also located on the northern side of the mainshock hypocenter at depths shallower than 6 km. Therefore, the large slips that took place in association with the Iwaki earthquake must have occurred along the steeply dipping portions of these faults, and the large-slip patches are roughly located within the high-velocity body imaged in this study. [12] Many previous tomographic studies have indicated a tendency for coseismic slip events to be concentrated within confined high-velocity bodies [e.g., Michael and Eberhart- Phillips, 1991; Chiarabba and Amato, 23; Kato et al., 21], even in the case of the 211 M9. Tohoku-Oki megathrust earthquake [Zhao et al., 211]. It is generally believed that such high-velocity bodies coincide with the more brittle and competent parts of the Earth s crust. The competent part of the crust in the studied area might withstand further increases in stress caused by the two adjacent clusters of seismic events in neighboring crustal domains where seismic velocities are lower than in the high-velocity body (Figure 3d). Then, after a while (a month in this case), the subsequent rupture of another earthquake (i.e., the M7. Iwaki earthquake) would take place within the high-velocity body, finally releasing the stress that had built-up between the two adjacent, previously formed seismic clusters in low-velocity regions. [13] Some of the low-velocity anomalies of the present study are imaged beneath earthquake alignments, especially beneath the hypocenter of the Iwaki earthquake (Figures 2 and 3). Other recent studies employing dense networks of seismic stations have revealed that such low-velocity anomalies lie directly beneath the seismogenic zone, or at least within its lower reaches [e.g., Nakajima and Hasegawa, 28; Kato et al., 29; Kato et al., 211b]. These types of anomalies may indicate the presence of crustal fluids. Indeed, we identified several migrations of hypocenters as a function of time during each swarm sequence (Supporting Information, Fig. S), reflecting diffusive processes of localized fluids and aseismic slip [e.g., Roland and McGuire, 29]. The low-velocity anomalies imaged in this study thus imply the potential involvement of crustal fluids in triggering the induced seismicity. Regional (larger-scale) tomography studies conducted in NE Japan [Tong et al., 212] indicate that fluids within the lower crust are infiltrating the overlying seismogenic zone from a deeper fluid reservoir located at a depth of ~2 km. Consequently, it seems likely that this type of fluid migration could result in a marked reduction in the shear strength of the overlying crust, ultimately triggering ruptures along pre-existing faults in the region [e.g., Miller et al., 24; Kato et al., 211b]. [14] Given these considerations, we hypothesize that farfield extensional deformation associated with the Tohoku-Oki mainshock triggered the sequences of normal faulting in relatively weak regions of crust in the Ibaraki and Fukushima prefectures, during the initial stages of deformation that took place prior to the Iwaki earthquake (Figure 3d). Eventually, the comparatively strong (i.e., high-velocity) parts of the crust were ruptured, probably in association with the infiltration of deeply sourced crustal fluids into the overlying seismogenic zone [Tong et al., 212], and ultimately resulting in the largest earthquake in the sequence (M7. Iwaki earthquake). These crustal heterogeneities likely played an important role in controlling the observed systematic spatiotemporal evolution of the induced seismicity across the region. [1] Acknowledgments. We are grateful to Claudio Chiarabba and Roland Bürgmann for constructive and useful comments. We thank H. Zhang for allowing us to use the tomodd code and acknowledge NIED and JMA for allowing us to use the waveform data. This study was partially supported by the Observation and Research Program for the Prediction of Earthquakes and Volcanic Eruptions under MEXT. References Chiarabba, C., and G. Selvaggi (1997), Structural control on fault geometry: example of the Gravena Ms 6.6, normal faulting earthquake, J. Geophys. Res., 12(B1), Chiarabba, C., P. De Gori, and E. Boschi (29), Pore-pressure migration along a normal-fault system resolved by time-repeated seismic tomography, Geology, 37(1), doi:1.113/g222a.1. Chiarabba, C., and A. Amato (23), Vp and Vp/Vs images in the Mw 6. Colfiorito fault region (central Italy): A contribution to the understanding of seismotectonic and seismogenic processes, J. Geophys. Res., 18(B), 2248, doi:1.129/21jb166. Farías, M., D. Comte, S. Roecker, D. Carrizo, and M. Pardo (211), Crustal extensional faulting triggered by the 21 Chilean earthquake: The Pichilemu Seismic Sequence, Tectonics, 3, TC61, doi:1.129/ 211TC2888. Fukushima, Y., Y. Takada, and M. Hashimoto (213), Complex ruptures of the 11 April 211 Iwaki earthquake (Mw 6.6) triggered by the 11 March 211 Tohoku earthquake (Mw 9.), Japan, Bull. Seismol. Soc. Am., doi: 1.178/ Hikima, K. (212), Rupture process of the April 11, 211 Fukushima Hamadori earthquake (Mj 7.) Two fault planes inferred from strong motion and relocated aftershocks, Zisin, 64, Imanishi, K., R. Ando, and Y. Kuwahara (212), Unusual shallow normalfaulting earthquake sequence in compressional northeast Japan activated after the 211 off the Pacific coast of Tohoku earthquake, Geophys. Res. Lett., 39, L936, doi:1.129/212gl1491. Kato, A., E. Kurashimo, T. Igarashi, S. Sakai, T. Iidaka, M. Shinohara, T. Kanazawa, T. Yamada, N. Hirata, and T. Iwasaki (29), Reactivation of ancient rift systems triggers devastating intraplate earthquakes, Geophys. Res. Lett., 36, L31, doi:1.129/28gl364. Kato, A., T. Miyatake, and N. Hirata (21), Asperity and barriers of the 24 Mid-Niigata Prefecture earthquake revealed by highly dense seismic observations, Bull. Seismol. Soc. Am., 1(1), , doi:1.178/ Kato, A., S. Sakai, and K. Obara (211a), A normal-faulting seismic sequence triggered by the 211 off the Pacific coast of Tohoku earthquake: Wholesale stress regime changes in the upper plate, Earth Planets Space, 63, , doi:1.47/eps Kato, A., et al. (211b), Anomalous depth dependency of the stress field in the 27 Noto Hanto, Japan, earthquake: Potential involvement of a deep fluid reservoir, Geophys. Res. Lett., 38, L636, doi:1.129/21gl Kato, A., and T. Igarashi (212), Regional extent of the large coseismic slip zone of the 211 Mw 9. Tohoku-Oki earthquake delineated by on-fault aftershocks, Geophys. Res. Lett., 39, L131, doi:1.129/212gl222. Loveless, J. P., and B. J. Meade (21), Geodetic imaging of plate motions, slip rates, and partitioning of deformation in Japan, J. Geophys. Res., 11, B241, doi:1.129/28jb6248. Michael, A. J., and D. Eberhart-Phillips (1991), Relations among fault behavior, subsurface geology, and three-dimensional velocity models, Science, 23, Miller, S. A., C. Collettini, L. Chiaraluce, M. Cocco, M. Barchi, and B. J. P. Kaus (24), Aftershocks driven by a high-pressure CO 2 source at depth, Nature, 427, , doi:1.138/nature

6 Mizoguchi, K., S. Uehara, and K. Ueta (212), Surface fault ruptures and slip distributions of the Mw April 211 Hamadoori, Fukushima Prefecture, Northeast Japan, Earthquake, Bull.Seismol.Soc.Am., 12(), , doi:1.178/ Nakajima, J., and A. Hasegawa (28), Existence of low-velocity zones under the source areas of the 24 Chuetsu and 27 Chuetsu-Oki earthquakes inferred from travel-time tomography, Earth Planets Space, 6, Ozawa, S., T. Nishimura, H. Suito, T. Kobayashi, M. Tobita, and T. Imakiire (211), Coseismic and postseismic slip of the 211 magnitude-9 Tohoku- Oki earthquake, Nature, 47, , doi:1.138/nature1227. Roland, E., and J. J. McGuire (29), Earthquake swarms on transform faults, Geophys. J. Int., 178, , doi: /j X x. Sibson, R. (2), Fluid involvement in normal faulting, J. Geodynamics, 29, Ryder, I., A. Rietbrock, K. Kelson, R. Bürgmann, M. Floyd, A. Socquet, C. Vigny, and D. Carrizo (212), Large extensional aftershocks in the continental forearc triggered by the 21 Maule earthquake, Chile, Geophys. J. Int., 188(3), , doi:1.1111/j x x. Toda, S., R. S. Stein, and J. Lin (211), Widespread seismicity excitation throughout central Japan following the 211 M = 9. Tohoku earthquake and its interpretation by Coulomb stress transfer, Geophys. Res. Lett., 38, LG3, doi:1.129/211gl Tong, P., D. Zhao, and D. Yang (212), Tomography of the 211 Iwaki earthquake (M 7.) and Fukushima nuclear power plant area, Solid Earth, 3, Zhang, H., and C. H. Thurber (23), Double-difference tomography: The method and its application to the Hayward fault, California, Bul. Seismol. Soc. Am., 93, Zhao, D., Z. Huang, N. Umino, A. Hasegawa, and H. Kanamori (211), Structural heterogeneity in the megathrust zone and mechanism of the 211 Tohoku-Oki earthquake (Mw 9.), Geophys. Res. Lett., 38, L1738., doi:1.129/211gl