- Email: [email protected]
Precise aftershock distribution of the 2004 Mid-Niigata prefecture earthquake—Implication for a very weak region in the lower crust
Physics of the Earth and Planetary Interiors 172 (2009) 345–352
Contents lists available at ScienceDirect
Physics of the Earth and Planetary Interiors journal homepage: www.elsevier.com/locate/pepi
Precise aftershock distribution of the 2004 Mid-Niigata prefecture earthquake—Implication for a very weak region in the lower crust Yoshihisa Iio a,∗ , Takuo Shibutani a , Satoshi Matsumoto b , Hiroshi Katao a , Takeshi Matsushima b , Shiro Ohmi a , Fumiaki Takeuchi a , Kenji Uehira b , Kinya Nishigami a , Masatoshi Miyazawa a , Bogdan Enescu a , Issei Hirose a , Yasuyuki Kano a , Yuhki Kohno c , Ken’ichi Tatsumi a , Tomotake Ueno d , Hiroo Wada a , Yohei Yukutake d a
Research Center for Earthquake Prediction, Disaster Prevention Research Institute, Kyoto University, Gokasho Uji 611-0011, Japan Institute of Seismology and Volcanology, Faculty of Sciences, Kyushu University, Shin’yama, Shimabara, Nagasaki 855-0843, Japan c Graduate School of Sciences, Kyushu University, Fukuoka 812-8581, Japan d Graduate School of Science, Kyoto University, Kyoto 606-8502, Japan b
a r t i c l e
i n f o
Article history: Received 12 January 2008 Received in revised form 29 June 2008 Accepted 16 October 2008 Keywords: The 2004 Mid-Niigata prefecture earthquake Aftershock distribution Intraplate earthquake Weak zone Lower crust
a b s t r a c t The 2004 Mid-Niigata prefecture Earthquake (Mjma 6.8) occurred in the region of large strain rates (>0.1 ppm/y contraction) in the intraplate region in Japan. The mainshock was followed by four major aftershocks with Mjma >= 6.0. The hypocenters of the mainshock and two large aftershocks that occurred in the central part of the aftershock region were located near the lower limit of the earthquake distribution, while hypocenters of the other two aftershocks near both ends, are located near its upper limit. Furthermore, the fault planes of the latter two aftershocks were confined within the upper half of the upper crust. Also, the lower limit of the aftershock distribution is deepest in the central part and becomes shallower toward the NNE and SSW ends. These data can be explained by the hypothesis that a localized stress concentration occurred near the bottom of the seismogenic region only in the central part. The stress concentration may be generated by the deformation in the very weak region of low strength in the lower crust beneath the central part of the aftershock region. © 2008 Elsevier B.V. All rights reserved.
1. Introduction Shallow intraplate earthquakes cause enormous damage; however, the process by which intraplate earthquakes are generated has been poorly understood and is still controversial, in contrast to interplate earthquakes. There are two end-member models of the process by which intraplate earthquakes are generated (Iio and Kobayashi, 2002). The first is called the regional stress model, in which broad-scale uniform regional stress generates intraplate earthquakes on pre-existing faults in the upper crust (e.g., Sykes, 1978; Hinze et al., 1988; Johnston and Kanter, 1990; Zoback, 1992). The second is the local stress model, in which local stress concentrations generate intraplate earthquakes (e.g., Campbell, 1978; Liu and Zoback, 1997; Stuart et al., 1997; Kenner and Segall, 2000). Iio and Kobayashi (2002) claims that the regional stress model cannot explain some of important problems concerning intraplate earthquakes, e.g., why and how is the stress accumulated on
∗ Corresponding author. Tel.: +81 774 38 4200; fax: +81 774 38 4190. E-mail address: [email protected] (Y. Iio). 0031-9201/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.pepi.2008.10.014
intraplate earthquake faults, despite the fact that stress at nearby plate boundaries is released quasi-periodically and is apparently not accumulated over the recurrence interval of interplate earthquakes? and why intraplate earthquakes do not trigger successive large earthquakes in adjacent regions? On the other hand, a local stress model, that assumes strength heterogeneities in the lower crust beneath the earthquake fault, can explain the above problems (Iio et al., 2002, 2004). The deformation in the weak zone in the lower crust can generate local stress concentrations in the upper crust and intraplate earthquakes can be generated at long recurrence intervals. Recently, the heterogeneities in the lower crust beneath intraplate earthquake faults were associated with conductivity and seismic velocity anomalies, and it was found that strain rates detected by GPS were larger in the region above the heterogeneities than those in the surrounding region (Iio et al., 2002; Hasegawa et al., 2005). Iio et al. (2002) inferred that the lower crust beneath the large strain rate region (NKTZ, the Niigata–Kobe tectonic zone; Sagiya et al., 2000) is weakened by high water content and that the large strain rate is due to the lower strength of the lower crust beneath the region compared to the surrounding region.
346
Y. Iio et al. / Physics of the Earth and Planetary Interiors 172 (2009) 345–352
Table 1 List of the major earthquakes. No. 1 2 3 4 5
2004 2004 2004 2004 2004
Date
Origin time
Longitude
Latitude
Depth
Mjma
Error (km)
October 23 October 23 October 23 October 23 October 27
17:56:00.005 18:03:12.423 18:11:55.529 18:34:05.509 10:40:49.979
138.8304 138.9621 138.7909 138.8974 139.0025
37.3052 37.3679 37.2745 37.3173 37.2989
11.120 4.550 2.667 13.472 13.873
6.8 6.3 6.0 6.5 6.1
0.254 0.950 4.556 0.424 0.238
Mjma is the magnitude determined by the Japan Meteorological Agency. Error is the standard error in hypocentral depth.
The 2004 Mid-Niigata prefecture (Chuetsu) Earthquake (Mjma 6.8) occurred on 23 October 2004 in the region of large strain rates in the intraplate region in Japan (NKTZ). The earthquake was followed by numerous aftershocks, including four aftershocks with Mjma (Japan Meteorological Agency Magnitude) greater than or equal to 6.0 (Kato et al., 2005; Aoki et al., 2005). The origin times and hypocenters of these five major events are listed in Table 1. The aftershock distribution was very complicated. Three or four major fault planes with almost parallel strikes are inferred from preliminary studies of the aftershock distribution using temporary seismic stations (Shibutani et al., 2005a; Sakai et al., 2005; Kato et al., 2005). One of the major differences among these studies is the hypocentral depth and the dip direction of the fault plane of the M6.0 aftershock (No. 3 in Table 1) that occurred at 18:11 on October 23, in the southern part of the aftershock region. It is important to determine the depth of the M6.0 fault, since this problem is closely related to the stress field in and around the aftershock region and the process by which this earthquake sequence was generated. We will precisely analyze the aftershock distribution, in particular, the depth of the
M6.0 fault, and estimate the stress field and its heterogeneities, that may be the fundamental origin of the 2004 Mid-Niigata prefecture earthquake, in order to clarify the process by which intraplate earthquakes are generated.
2. Data, method and results We located aftershock hypocenters based on 3D velocity models of P- and S-waves, which were obtained from a tomographic inversion method developed from SIMUL3M (Thurber, 1983). Initial values of hypocenters and station corrections were determined with a joint hypocenter determination method (Shibutani et al., 2005a). The distribution of 74 stations used in this study is shown in Fig. 1. DP.TDOM, DP.OJKW and DP.YMKS are temporary stations and the others are the permanent online stations operated by the National Research Institute for Earth Science and Disaster Prevention (NIED), JMA, University of Tokyo and Tohoku University.
Fig. 1. Map showing the locations of the stations (crosses with square: temporary online, crosses: permanent online) used. Gray dots denote the aftershocks.
Y. Iio et al. / Physics of the Earth and Planetary Interiors 172 (2009) 345–352
We selected 2544 events from October 1, 1997 to November 10, 2004, in and around the source region (Fig. 1), including 506 aftershocks, and determined simultaneously the hypocentral parameters, 1D velocity models and station corrections, using a JHD technique (Kissling et al., 1994). The events before the 2004 Mid-Niigata prefecture earthquake occurred mainly outside of the source region. The reason we used these events is to precisely determine velocity structures around the source region. This is justified by the fact that the difference between velocity structures before and after the mainshock was extremely small in the source region of an intraplate earthquake, where a dense seismic network had been installed before the mainshock (Yamawaki et al., 2004). Then, we carried out a tomographic inversion of the travel times using the same data set. The grid intervals are 3 × 3 × 3 km in the vicinity of the aftershock area. The initial velocity model for the tomography was obtained by resampling the 1D models obtained from the JHD method. The RMS of the travel time residuals decreased from 0.18 s for the 1D inversion to 0.10 s for 3D inversion. Using the tomographic method, we were able to determine more accurately the hypocenters considering the lateral velocity heterogeneities. Almost all the standard errors of the hypocentral locations by this tomographic method were smaller than those from the JHD method. Finally we relocated 7257 aftershocks based on the derived 3D velocity model. In Fig. 2, we plotted hypocentral distributions of the aftershocks that were selected with the following conditions: (1) the standard errors in the x, y, and z directions are smaller than 0.5 km, (2) the magnitude was greater than or equal to 0.5, (3) the resolution is greater than 0.5, and (4) the time period is from 23 October to 31
347
December 2004. It should be noted that the hypocenters of the mainshock and the major aftershocks with Mjma >= 6.0 (the major events) are plotted independent of errors in hypocentral determination. The resolution is the diagonal element of resolution matrix of the tomographic inversion method that uses a damped least square approach. The condition (3) was introduced to remove events with poor resolutions. Fortunately, the resolutions of the hypocentral parameters are almost 1 for all the located events and all the events satisfied the condition (3). This is because damping factors to the hypocentral parameters are set to be sufficiently small. It is found that the hypocenters of the three major events (Nos. 1, 4, and 5) occurring in the central part of the aftershock distribution are deep and located near the lower limit. On the other hand, the other two major events (Nos. 2 and 3) in the NNE and SSW portions of the aftershock region are very shallow, near the top of the aftershock distribution. Furthermore, it seems the lower limit of the aftershock distribution is deepest in the central part and becomes shallower toward the NNE and SSW ends, except for a small cluster located near the NW corner of the aftershock region with depths around 15 km. Fig. 3 displays hypocentral distributions projected on X–Z cross-sections. Hypocenters located in 2 km widths in the Y (NNE–SSW) direction are plotted in 16 panels that are aligned with the NNE section at the top-left and the SSW section at the bottom-right. The red circles are the hypocenters of the major events. Blue dotted lines are the fault planes of the major events that are located in each 2 km width. These fault planes were determined by Miyazawa et al. (2005) using the aftershock distribution by Shibutani et al. (2005a) and the NIED moment tensor
Fig. 2. Distribution of the aftershock hypocenters from the tomographic inversion method using the initial values of the hypocenters, 1D velocity structures, and station corrections determined with the JHD (joint hypocenter determination) method by Shibutani et al. (2005a). Red circles indicate the hypocenters of the major events of M >= 6.0, and the numbers correspond to those listed in Table 1.
348
Y. Iio et al. / Physics of the Earth and Planetary Interiors 172 (2009) 345–352
Fig. 3. Hypocentral distributions projected on X–Z cross-sections. Hypocenters located in 2 km widths in the Y (NNE–SSW) direction are plotted in 16 panels, with the NNE section at the top-left and the SSW section at the bottom-right. Blue dotted lines are the fault planes of the major events that are located in the 2 km widths. The fault planes were estimated by Miyazawa et al. (2005) using the aftershock distribution by Shibutani et al. (2005a) and the NIED moment tensor solutions (NIED, 2005).
solutions (National Research Institute for Earth Science and Disaster Prevention (NIED), 2004). In the bottom-left panel, it is seen that the hypocenter of the M6.0 aftershock (No. 3) is very shallow and it appears that its aftershocks align on the east-dipping plane in this and the neighboring panels. On the other hand, aftershocks are scarcely distributed on and near the estimated mainshock fault plane in the bottom three panels (Y = −10 to −4 km). It is inferred from strong motion seismograms that the slip of the mainshock is very small in this region (Hikima and Koketsu, 2005). In the northern part, it is seen that the hypocenter of the M6.3 aftershock (No. 2) is also very shallow and that its aftershocks align on the west-dipping plane. Although Hikima and Koketsu (2005) estimated that the fault plane is dipping to the east, the aftershock distribution around the hypocenter of the M6.3 aftershock is not consistent with their conclusion as shown in Fig. 3. The above results about the M6.0 aftershock (No. 3) are almost the same as those by Shibutani et al. (2005a), while some other studies had estimated that the hypocenter was deeper than 10 km and was located on the deeper part of the mainshock fault plane (Sakai et al., 2005; Kato et al., 2005). These studies expected that the
M6.0 fault was located on the extension of the mainshock fault. This discrepancy probably results from the differences in station corrections estimated for the permanent stations, which were located around the source region. Shibutani et al. (2005a) determined station corrections from the data including the events that occurred around the source region, while the other studies used only the data of the mainshock and aftershocks. Furthermore, Shibutani et al. (2005a) determined P- and S-wave velocities independently, while the other two studies (Sakai et al., 2005; Kato et al., 2005) determined S-wave velocities assuming a constant Vp /Vs ratio of 1.73. As shown in Shibutani et al. (2005a), Vp /Vs ratios in the shallow layers are much greater than 1.73 in and around the aftershock region. The assumption of the constant Vp /Vs ratio can make locations of shallow aftershocks deeper. Okada et al. (2005, 2006) also analyzed another dense temporary network by the Double-Difference tomography method and estimated that the depth is about 7 km, which is an intermediate to the studies mentioned above. This is probably because they used more permanent stations than the other two studies. They used 54 permanent stations with epicentral distance of less than 60 km, while the other two studies used only
Y. Iio et al. / Physics of the Earth and Planetary Interiors 172 (2009) 345–352
28 stations. Okada et al. (2005, 2006) also used only the data of the mainshock and aftershocks. Thus, it is possible that the structures beneath the permanent stations and their station corrections are not precisely determined. In Fig. 3, there is no aftershock activity around the mainshock fault immediately beneath the hypocenter of the M6.0 aftershock. If the M6.0 aftershock occurred on the mainshock fault plane, its secondary aftershocks (aftershocks of the M6.0 aftershock) should be located in this region, especially since the hypocenters of smaller aftershocks are more accurately determined than those of the major events, because S-arrivals can be used. Fig. 4 displays a 10-km wide depth distribution of the M6.0 aftershock and the earthquakes that occurred immediately after from 18:11 to 19:00 on October 23. It is found that small aftershocks occurred on (or near) the estimated east-dipping fault plane of the M6.0 aftershock, although the standard errors indicated by error bars are larger than 0.5 km (plot criteria in Fig. 3). The errors for the M6.0 (Table 1) are possibly due to the very shallow hypocenters. 3. Discussion It is difficult to explain the spatial change in the lower limit of the aftershock distribution by the thermal structure, since it is thought that the depth of an isothermal surface scarcely changes in the
349
Fig. 4. Depth distribution of the M6.0 aftershock and the earthquakes that occurred immediately after, from 18:11 to 19:00 on October 23 in a 10-km wide X–Z crosssection. Error bars indicate the standard location error of each aftershock.
NNE–SSW direction. The depth possibly changes in the ESE–WNW direction, since active volcanoes are located east of the aftershock region and align in the NNE–SSW direction along the axis of the island arc. This estimate is in good accordance with the distribution of heat flow in the Japanese compiled by Tanaka et al. (2004).
Fig. 5. Location of the very weak region inferred in the lower crust only beneath the central part of the aftershock region. The fault planes and hypocenters of the major events are displayed in different colors. The event numbers correspond to those listed in Table 1. The locations of the S-wave reflectors (Matsumoto et al., 2005) and the region of the low velocity anomaly (Okada et al., 2006) are also shown. The reflectors are schematically illustrated by ellipses in the vertical cross section in the lower panel, and the region in which the reflectors were investigated is also shown by the bracket in the right panel. The region of the low velocity anomaly is roughly enclosed by pink rectangles.
350
Y. Iio et al. / Physics of the Earth and Planetary Interiors 172 (2009) 345–352
Fig. 6. (Upper) Schematic illustration of the depth distribution of the shear strength (black line) and the shear stress before the mainshock (green line) in the upper crust. The shear stress after the mainshock is also shown in profile A (blue line). The dotted lines indicate dynamic friction on the fault. (Lower) Schematic illustration of the very weak region, the weak zone and the rupture propagation of the mainshock and aftershocks. The very weak region exists beneath the central part of the mainshock fault.
Instead these data could be explained by the hypothesis that a localized stress concentration occurred near the bottom of the seismogenic region in the central part of the aftershock region. In this hypothesis, the localized stress concentration is generated by the deformation in the very weak region in the lower crust beneath the central part, as shown in Figs. 5 and 6. Since the aftershock region is located within the concentrated deformation zone (NKTZ), it is likely that there exists the weak zone in the lower crust beneath the aftershock region (Iio et al., 2004; Sibson, 2007). We assume that the strength of the very weak region is much lower than that in the surrounding weak zone. The weak zone beneath the concentrated deformation zone (NKTZ), which extends about 500 km in length, is thought to be weakened by high water content (Iio et al., 2002). We assume that water is concentrated more in the very weak region than that in the surrounding weak zone and then the very weak region is more significantly weakened by higher water content. Laboratory experiments clearly show that larger water fugacity makes the viscosity of rocks in the lower crust smaller (e.g., Rybacki et al., 2006). Fig. 6 schematically illustrates vertical stress profiles above the very weak region (profile B) and the surrounding weak zone (profile A). The mainshock fault shown in this figure is not the fault plane shown in Fig. 5, but the actual slip area roughly deduced from
the slip distribution by Miyazawa et al. (2005). The magnitudes of the shear strength and dynamic friction are proportional to depth, assuming the coefficients of static and dynamic frictions are constant and independent of depth following the experimental results by Byerlee (1978) and pore pressure is hydrostatic in the upper crust (Grawinkel and Stöeckhert, 1997). The difference between the shear strength and dynamic friction increases with depth, since the coefficient of dynamic friction is smaller than that of static friction, as Byerlee (1978) estimated the coefficients of static and dynamic frictions to be 0.6 and 0.47, respectively. It is thought that dynamic friction determines the lowest level of shear stress on the fault plane and the deformation in the lower crust can raise shear stress above this lowest level in the upper crust (Iio et al., 2004). On profile B, a large stress concentration is generated at the bottom of the seismogenic region by the deformation in the very weak region. On the other hand, on profile A, the stress concentration at the bottom of the seismogenic region is relatively small and the difference between the strength and the stress before the mainshock increases with depth. Thus, when the mainshock occurred, the shallower part is broken more easily at the location of profile A. This results from the following two reasons. The first is that the difference between the static and dynamic frictions are thought to be smaller in a shallower part. The second is that the magnitude of
Y. Iio et al. / Physics of the Earth and Planetary Interiors 172 (2009) 345–352
the stress concentration generated at both sides of the mainshock fault is thought to be basically independent of depth, since the distance from the large slip area, which is located in the central part of the fault (e.g., Miyazawa et al., 2005), does not significantly change in depth at both sides of the mainshock fault. The stress profiles shown above could be the reasons why the central three major events (Nos. 1, 4, and 5) initiated at depth and why the two major events (Nos. 2 and 3) occurring in the NNE and SSW portions of the aftershock region initiated near the top of the aftershock distribution, and their faults were confined within the upper part of the upper crust. Evidence that suggests the existence of the very weak region is obtained by the imaging of S-wave reflectors and a seismic tomography. Matsumoto et al. (2005) found distinct S-wave reflectors in the lower crust in the central part of the aftershock region. Okada et al. (2006) estimated a low velocity anomaly in the lower crust only beneath the central part. A similar aftershock distribution showing a bowl shape is seen for the 2000 Western Tottori earthquake (Shibutani et al., 2005b) and the 2005 Western Fukuoka prefecture earthquake (Uehira et al., 2006). Furthermore, a low resistivity zone is recognized beneath the central part of the aftershock region of the 2000 Western Tottori earthquake (Shiozaki et al., 2003). The horizontal extent of the low resistivity zone is about 10 km along the fault. The resistivity value in the central part is estimated to be about 10 m, suggesting high water content, while those in the surrounding parts are 103 to 104 m (Shiozaki et al., 2003). The low resistivity zone might reflect the very weak region in the lower crust. These data suggest that the hypothesis holds for moderate size intraplate earthquakes. 4. Conclusions We precisely determined the aftershock distribution of the 2004 Mid-Niigata prefecture earthquake using a tomographic method. We obtained the following results. (1) The hypocenters of three major events occurring in the central part (Nos. 1, 4, and 5) are deep, near the lower limit of the aftershock distribution, while the other two major events (Nos. 2 and 3) in the NNE and SSW portions of the aftershock region are very shallow, near the top of the aftershock distribution. (2) The fault planes of the two major events occurring in the NNE and SSW portions are shallow, limited to the upper half of the upper crust. (3) There is a tendency that the lower limit of the aftershock distribution is deepest in the central part and becomes shallower toward the NNE and SSW ends. These results could be explained by the hypothesis that a localized stress concentration occurred near the bottom of the seismogenic region only in the central part of the aftershock region. The stress concentration may be generated by the deformation in the very weak region of low strength in the lower crust beneath the central part. The hypothesis presented in this paper clearly supports the local stress model in contrast to the regional stress model. If the hypothesis of the very weak region in the lower crust generally holds for inland earthquakes, the hypothesis is useful for forecasting the locations of large earthquakes, in particular the rupture initiation point. Furthermore, the hypothesis could roughly estimate their rupture extent from the size of the very weak region. Thus, it is important to carefully examine this hypothesis by studies of the location, the lateral and vertical extents, and the properties of the very weak region. Acknowledgements We are very grateful to people in Tochio City, Ojiya City, Nagaoka City (Yamakoshi Village) and Niigata Prefecture who kindly helped
351
us in the aftershock observations. We thank Yuta Asaka, Ling Bai, Mio Hori, Takeshi Katagi, Shigemitsu Matsuo, Koichi Miura, Masahiro Morita, Sun Cheon Park and Atsushi Watanabe for their help in the instrument deployment, and Aiko Nakao for her help in picking the arrival times. We utilized waveform data from permanent stations operated by NIED, JMA and the University of Tokyo and P- and S-arrival times from the JMA preliminary catalogue. We thank James Mori for his critical review of the manuscript. This work was conducted under the support of the Grant-in-Aid for Special Purposes (No. 16800054), MEXT, Japan. This work was also supported by the subsidies to earthquake research projects of the Tokio Marine Kagami Memorial Foundation. References Aoki, S., Nishi, M., Nakamura, K., Hashimoto, T., Yoshikawa, S., Ito, H., 2005. Multiplanar structures in the aftershock distribution of the Mid-Niigata prefecture Earthquake in 2004. Earth Planets Space 57, 5411–5416. Byerlee, J.D., 1978. Friction of rocks. Pure Appl. Geophys. 116, 615–626. Campbell, D.L., 1978. Investigation of the stress-concentration mechanism for intraplate earthquakes. Geophys. Res. Lett. 5, 477–479. Grawinkel, A., Stöeckhert, B., 1997. Hydrostatic pore fluid pressure to 9-km depth; fluid inclusion evidence from the KTB deep drill hole. Geophys. Res. Lett. 24, 2373–2376. 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. Hikima, K., Koketsu, K., 2005. Rupture processes of the 2004 Chuetsu (mid-Niigata prefecture) earthquake, Japan: a series of events in a complex fault system. Goophys. Res. Lett. 32, L18303, doi:10.1029/2005GL023588. Hinze, W.J., Braile, L.W., Keller, G.R., Lidiak, E.G., 1988. Models for mid-continent tectonism: an update. Rev. Geophys. 26, 699–717. Iio, Y., Kobayashi, Y., 2002. A physical understanding of large intraplate earthquakes. Earth Planets Space 54, 1001–1004. Iio, Y., Sagiya, T., Kobayashi, Y., 2004. Origin of the concentrated deformation zone in the Japanese Islands and stress accumulation process of intraplate earthquakes. Earth Planets Space 56, 831–842. Iio, Y., Sagiya, T., Kobayashi, Y., Shiozaki, I., 2002. Water-weakened lower crust and its role in the concentrated deformation in the Japanese Islands. Earth Planet. Sci. Lett. 203, 245–253. Johnston, A.C., Kanter, L.R., 1990. Earthquakes in stable continental crust. Sci. Am. 262 (3), 68–75. Kato, A., Sakai, S., Hirata, N., Kurashimo, E., Nagai, S., Iidaka, T., Igarashi, T., Yamanaka, Y., Murotani, S., Kawamura, T., Iwasaki, T., Kanazawa, T., 2005. Spatiotemporal variations of the aftershock distributions during one month after the occurrence of the 2004 Mid-Niigata prefecture earthquake. Earth Planets Space 57 (6), 551–556. Kenner, S., Segall, P., 2000. A mechanical model for intraplate earthquakes: application to the New Madrid. Science 289, 2329–2332. Kissling, E., Ellsworth, W.L., Eberhart-Phillips, D., Kradolfer, U., 1994. Initial reference models in local earthquake tomography. J. Geophys. Res. 99, 19635–19646. Liu, L., Zoback, M.D., 1997. Lithospheric strength and intraplate seismicity in the New Madrid seismic zone. Tectonics 16, 585–595. Matsumoto, M., Iio, Y., Matsushima, T., Uehira, K., Shibutani, T., 2005. Imaging of S-wave reflectors in and around the hypocentral area of the 2004 mid Niigata Prefecture Earthquake (M6.8). Earth Planets Space 57, 557–561. Miyazawa, M., Mori, J., Iio, Y., Shibutani, T., Matsumoto, S., Katao, H., Ohmi, S., Nishigami, K., 2005. Triggering sequence of large aftershocks of the Mid-Niigata Prefecture, Japan Earthquake in 2004 by static stress changes. Earth Planets Space 57, 1109–1113. National Research Institute for Earth Science and Disaster Prevention (NIED), 2004. Earthquake mechanism information. Available from:. Okada, T., Umino, N., Matsuzawa, T., Nakajima, J., Uchida, N., Nakayama, T., Hirahara, S., Sato, T., Hori, S., Kono, T., Yabe, Y., Ariyoshi, K., Gamage, S., Shimizu, J., Suganomata, J., Kita, S., Yui, S., Arao, M., Hondo, S., Mizukami, T., Tsushima, H., Yaginuma, T., Hasegawa, A., 2005. Aftershock distribution and seismic velocity structure in and around the focal area of the 2004 Mid-Niigata prefecture earthquake using double-difference tomography, and its relationship with the earthquake generation process. Earth Planets Space 57, 435– 440. Okada, T., Yaginuma, T., Umino, N., Matsuzawa, T., Hasegawa, A., Zhang, H., Thurber, C.H., 2006. Detailed imaging of the fault planes of the 2004 Mid-Niigata prefecture, central Japan, earthquake sequence by double-difference tomography. Earth Planet. Sci. Lett. 244, 32–43. Rybacki, E., Gottschalk, M., Wirth, R., Dresen, G., 2006. Influence of water fugacity and activation volume on the flow properties of fine-grained anorthite aggregates. J. Geophys. Res. 111, B03203. Sagiya, T., Miyazaki, S., Tada, T., 2000. Continuous GPS array and present-day crustal deformation of Japan. Pageoph. 157, 2303–2322.
352
Y. Iio et al. / Physics of the Earth and Planetary Interiors 172 (2009) 345–352
Sakai, S., Hirata, N., Kato, A., Kurashimo, E., Iwasaki, T., Kanazawa, T., 2005. Multifault system of the 2004 Mid-Niigata Prefecture Earthquake and its aftershocks. Earth Planets Space 57 (5), 417–422. Sibson, R.H., 2007. An episode of fault-valve behaviour during compressional inversion?—The 2004 MJ6.8 Mid-Niigata Prefecture, Japan, earthquake sequence. Earth Planet. Sci. Lett. 257, 188–199, doi:10.1016/j.epsl.2007.02.031. Shibutani, T., Iio, Y., Matsumoto, S., Katao, H., Matsushima, T., Ohmi, S., Takeuchi, F., Uehira, K., Nishigami, K., Enescu, B., Hirose, I., Kano, Y., Kohno, Y., Korenaga, M., Mamada, Y., Miyazawa, M., Tatsumi, K., Ueno, T., Wada, H., Yukutake, Y., 2005a. Aftershock distribution of the 2004 Mid-Niigata Prefecture Earthquake derived from a combined analysis of temporary online observations and permanent observations. Earth Planets Space 57, 545– 549. Shibutani, T., Katao, H., Group for the dense aftershock observations of the 2000 Western Tottori Earthquake, 2005b. High resolution 3-D velocity structure in the source region of the 2000 Western Tottori Earthquake in southwestern Honshu, Japan using very dense aftershock observations. Earth Planets Space 57, 825–838. Shiozaki, I., Oshiman, N., Kasaya, T., Uto, T., Kagiyama, T., 2003. Deep crustal resistivity structure derived from the wide band MT investigations in the eastern part of Sanin region, southwestern Honshu, Japan, IUGG2003 JSA10/30P/D-006, XX111 General Assembly of the International Union of Geodesy and Geophysics.
Stuart, W.D., Hildenbrand, T.G., Simpson, R.W., 1997. Stressing of the New Madrid seismic zone by a lower crust detachment fault. J. Geophys. Res. 102, 27623–27633. Sykes, L.R., 1978. Intraplate seismicity, reactivation of preexisting zones of weakness, alkaline magmatism, and other tectonism postdating continental fragmentation. Rev. Geophys. 16, 621–688. Tanaka, A., Yamano, M., Yano, Y., Sasada, M., 2004. Geothermal gradient and heat flow data in and around Japan. I. Appraisal of heat flow from geothermal gradient data. Earth Planets Space 56, 1191–1194. Thurber, C.H., 1983. Earthquake locations and three-dimensional crustal structure in the Coyote Lake area, central California. J. Geophys. Res. 88, 8226– 8236. Uehira, K., Yamada, T., Shinohara, M., Nakahigashi, K., Miyamachi, H., Iio, Y., Okada, T., Takahashi, H., Matsuwo, N., Uchida, K., Kanazawa, T., Shimizu, H., 2006. Precise aftershock distribution of the 2005 West Off Fukuoka Prefecture Earthquake (Mj = 7.0) using a dense onshore and offshore seismic network. Earth Planets Space 58, 1605–1610. Yamawaki, T., Nishimura, T., Hamaguchi, H., 2004. Temporal change of seismic structure around Iwate Volcano inferred from waveform correlation analysis of similar earthquakes. Geophys. Res. Lett. 31, 2004. Zoback, M.L., 1992. Stress field constraints on intraplate seismicity in eastern North America. J. Geophs. Res. 97, 11761–11782.
Contents lists available at ScienceDirect
Physics of the Earth and Planetary Interiors journal homepage: www.elsevier.com/locate/pepi
Precise aftershock distribution of the 2004 Mid-Niigata prefecture earthquake—Implication for a very weak region in the lower crust Yoshihisa Iio a,∗ , Takuo Shibutani a , Satoshi Matsumoto b , Hiroshi Katao a , Takeshi Matsushima b , Shiro Ohmi a , Fumiaki Takeuchi a , Kenji Uehira b , Kinya Nishigami a , Masatoshi Miyazawa a , Bogdan Enescu a , Issei Hirose a , Yasuyuki Kano a , Yuhki Kohno c , Ken’ichi Tatsumi a , Tomotake Ueno d , Hiroo Wada a , Yohei Yukutake d a
Research Center for Earthquake Prediction, Disaster Prevention Research Institute, Kyoto University, Gokasho Uji 611-0011, Japan Institute of Seismology and Volcanology, Faculty of Sciences, Kyushu University, Shin’yama, Shimabara, Nagasaki 855-0843, Japan c Graduate School of Sciences, Kyushu University, Fukuoka 812-8581, Japan d Graduate School of Science, Kyoto University, Kyoto 606-8502, Japan b
a r t i c l e
i n f o
Article history: Received 12 January 2008 Received in revised form 29 June 2008 Accepted 16 October 2008 Keywords: The 2004 Mid-Niigata prefecture earthquake Aftershock distribution Intraplate earthquake Weak zone Lower crust
a b s t r a c t The 2004 Mid-Niigata prefecture Earthquake (Mjma 6.8) occurred in the region of large strain rates (>0.1 ppm/y contraction) in the intraplate region in Japan. The mainshock was followed by four major aftershocks with Mjma >= 6.0. The hypocenters of the mainshock and two large aftershocks that occurred in the central part of the aftershock region were located near the lower limit of the earthquake distribution, while hypocenters of the other two aftershocks near both ends, are located near its upper limit. Furthermore, the fault planes of the latter two aftershocks were confined within the upper half of the upper crust. Also, the lower limit of the aftershock distribution is deepest in the central part and becomes shallower toward the NNE and SSW ends. These data can be explained by the hypothesis that a localized stress concentration occurred near the bottom of the seismogenic region only in the central part. The stress concentration may be generated by the deformation in the very weak region of low strength in the lower crust beneath the central part of the aftershock region. © 2008 Elsevier B.V. All rights reserved.
1. Introduction Shallow intraplate earthquakes cause enormous damage; however, the process by which intraplate earthquakes are generated has been poorly understood and is still controversial, in contrast to interplate earthquakes. There are two end-member models of the process by which intraplate earthquakes are generated (Iio and Kobayashi, 2002). The first is called the regional stress model, in which broad-scale uniform regional stress generates intraplate earthquakes on pre-existing faults in the upper crust (e.g., Sykes, 1978; Hinze et al., 1988; Johnston and Kanter, 1990; Zoback, 1992). The second is the local stress model, in which local stress concentrations generate intraplate earthquakes (e.g., Campbell, 1978; Liu and Zoback, 1997; Stuart et al., 1997; Kenner and Segall, 2000). Iio and Kobayashi (2002) claims that the regional stress model cannot explain some of important problems concerning intraplate earthquakes, e.g., why and how is the stress accumulated on
∗ Corresponding author. Tel.: +81 774 38 4200; fax: +81 774 38 4190. E-mail address: [email protected] (Y. Iio). 0031-9201/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.pepi.2008.10.014
intraplate earthquake faults, despite the fact that stress at nearby plate boundaries is released quasi-periodically and is apparently not accumulated over the recurrence interval of interplate earthquakes? and why intraplate earthquakes do not trigger successive large earthquakes in adjacent regions? On the other hand, a local stress model, that assumes strength heterogeneities in the lower crust beneath the earthquake fault, can explain the above problems (Iio et al., 2002, 2004). The deformation in the weak zone in the lower crust can generate local stress concentrations in the upper crust and intraplate earthquakes can be generated at long recurrence intervals. Recently, the heterogeneities in the lower crust beneath intraplate earthquake faults were associated with conductivity and seismic velocity anomalies, and it was found that strain rates detected by GPS were larger in the region above the heterogeneities than those in the surrounding region (Iio et al., 2002; Hasegawa et al., 2005). Iio et al. (2002) inferred that the lower crust beneath the large strain rate region (NKTZ, the Niigata–Kobe tectonic zone; Sagiya et al., 2000) is weakened by high water content and that the large strain rate is due to the lower strength of the lower crust beneath the region compared to the surrounding region.
346
Y. Iio et al. / Physics of the Earth and Planetary Interiors 172 (2009) 345–352
Table 1 List of the major earthquakes. No. 1 2 3 4 5
2004 2004 2004 2004 2004
Date
Origin time
Longitude
Latitude
Depth
Mjma
Error (km)
October 23 October 23 October 23 October 23 October 27
17:56:00.005 18:03:12.423 18:11:55.529 18:34:05.509 10:40:49.979
138.8304 138.9621 138.7909 138.8974 139.0025
37.3052 37.3679 37.2745 37.3173 37.2989
11.120 4.550 2.667 13.472 13.873
6.8 6.3 6.0 6.5 6.1
0.254 0.950 4.556 0.424 0.238
Mjma is the magnitude determined by the Japan Meteorological Agency. Error is the standard error in hypocentral depth.
The 2004 Mid-Niigata prefecture (Chuetsu) Earthquake (Mjma 6.8) occurred on 23 October 2004 in the region of large strain rates in the intraplate region in Japan (NKTZ). The earthquake was followed by numerous aftershocks, including four aftershocks with Mjma (Japan Meteorological Agency Magnitude) greater than or equal to 6.0 (Kato et al., 2005; Aoki et al., 2005). The origin times and hypocenters of these five major events are listed in Table 1. The aftershock distribution was very complicated. Three or four major fault planes with almost parallel strikes are inferred from preliminary studies of the aftershock distribution using temporary seismic stations (Shibutani et al., 2005a; Sakai et al., 2005; Kato et al., 2005). One of the major differences among these studies is the hypocentral depth and the dip direction of the fault plane of the M6.0 aftershock (No. 3 in Table 1) that occurred at 18:11 on October 23, in the southern part of the aftershock region. It is important to determine the depth of the M6.0 fault, since this problem is closely related to the stress field in and around the aftershock region and the process by which this earthquake sequence was generated. We will precisely analyze the aftershock distribution, in particular, the depth of the
M6.0 fault, and estimate the stress field and its heterogeneities, that may be the fundamental origin of the 2004 Mid-Niigata prefecture earthquake, in order to clarify the process by which intraplate earthquakes are generated.
2. Data, method and results We located aftershock hypocenters based on 3D velocity models of P- and S-waves, which were obtained from a tomographic inversion method developed from SIMUL3M (Thurber, 1983). Initial values of hypocenters and station corrections were determined with a joint hypocenter determination method (Shibutani et al., 2005a). The distribution of 74 stations used in this study is shown in Fig. 1. DP.TDOM, DP.OJKW and DP.YMKS are temporary stations and the others are the permanent online stations operated by the National Research Institute for Earth Science and Disaster Prevention (NIED), JMA, University of Tokyo and Tohoku University.
Fig. 1. Map showing the locations of the stations (crosses with square: temporary online, crosses: permanent online) used. Gray dots denote the aftershocks.
Y. Iio et al. / Physics of the Earth and Planetary Interiors 172 (2009) 345–352
We selected 2544 events from October 1, 1997 to November 10, 2004, in and around the source region (Fig. 1), including 506 aftershocks, and determined simultaneously the hypocentral parameters, 1D velocity models and station corrections, using a JHD technique (Kissling et al., 1994). The events before the 2004 Mid-Niigata prefecture earthquake occurred mainly outside of the source region. The reason we used these events is to precisely determine velocity structures around the source region. This is justified by the fact that the difference between velocity structures before and after the mainshock was extremely small in the source region of an intraplate earthquake, where a dense seismic network had been installed before the mainshock (Yamawaki et al., 2004). Then, we carried out a tomographic inversion of the travel times using the same data set. The grid intervals are 3 × 3 × 3 km in the vicinity of the aftershock area. The initial velocity model for the tomography was obtained by resampling the 1D models obtained from the JHD method. The RMS of the travel time residuals decreased from 0.18 s for the 1D inversion to 0.10 s for 3D inversion. Using the tomographic method, we were able to determine more accurately the hypocenters considering the lateral velocity heterogeneities. Almost all the standard errors of the hypocentral locations by this tomographic method were smaller than those from the JHD method. Finally we relocated 7257 aftershocks based on the derived 3D velocity model. In Fig. 2, we plotted hypocentral distributions of the aftershocks that were selected with the following conditions: (1) the standard errors in the x, y, and z directions are smaller than 0.5 km, (2) the magnitude was greater than or equal to 0.5, (3) the resolution is greater than 0.5, and (4) the time period is from 23 October to 31
347
December 2004. It should be noted that the hypocenters of the mainshock and the major aftershocks with Mjma >= 6.0 (the major events) are plotted independent of errors in hypocentral determination. The resolution is the diagonal element of resolution matrix of the tomographic inversion method that uses a damped least square approach. The condition (3) was introduced to remove events with poor resolutions. Fortunately, the resolutions of the hypocentral parameters are almost 1 for all the located events and all the events satisfied the condition (3). This is because damping factors to the hypocentral parameters are set to be sufficiently small. It is found that the hypocenters of the three major events (Nos. 1, 4, and 5) occurring in the central part of the aftershock distribution are deep and located near the lower limit. On the other hand, the other two major events (Nos. 2 and 3) in the NNE and SSW portions of the aftershock region are very shallow, near the top of the aftershock distribution. Furthermore, it seems the lower limit of the aftershock distribution is deepest in the central part and becomes shallower toward the NNE and SSW ends, except for a small cluster located near the NW corner of the aftershock region with depths around 15 km. Fig. 3 displays hypocentral distributions projected on X–Z cross-sections. Hypocenters located in 2 km widths in the Y (NNE–SSW) direction are plotted in 16 panels that are aligned with the NNE section at the top-left and the SSW section at the bottom-right. The red circles are the hypocenters of the major events. Blue dotted lines are the fault planes of the major events that are located in each 2 km width. These fault planes were determined by Miyazawa et al. (2005) using the aftershock distribution by Shibutani et al. (2005a) and the NIED moment tensor
Fig. 2. Distribution of the aftershock hypocenters from the tomographic inversion method using the initial values of the hypocenters, 1D velocity structures, and station corrections determined with the JHD (joint hypocenter determination) method by Shibutani et al. (2005a). Red circles indicate the hypocenters of the major events of M >= 6.0, and the numbers correspond to those listed in Table 1.
348
Y. Iio et al. / Physics of the Earth and Planetary Interiors 172 (2009) 345–352
Fig. 3. Hypocentral distributions projected on X–Z cross-sections. Hypocenters located in 2 km widths in the Y (NNE–SSW) direction are plotted in 16 panels, with the NNE section at the top-left and the SSW section at the bottom-right. Blue dotted lines are the fault planes of the major events that are located in the 2 km widths. The fault planes were estimated by Miyazawa et al. (2005) using the aftershock distribution by Shibutani et al. (2005a) and the NIED moment tensor solutions (NIED, 2005).
solutions (National Research Institute for Earth Science and Disaster Prevention (NIED), 2004). In the bottom-left panel, it is seen that the hypocenter of the M6.0 aftershock (No. 3) is very shallow and it appears that its aftershocks align on the east-dipping plane in this and the neighboring panels. On the other hand, aftershocks are scarcely distributed on and near the estimated mainshock fault plane in the bottom three panels (Y = −10 to −4 km). It is inferred from strong motion seismograms that the slip of the mainshock is very small in this region (Hikima and Koketsu, 2005). In the northern part, it is seen that the hypocenter of the M6.3 aftershock (No. 2) is also very shallow and that its aftershocks align on the west-dipping plane. Although Hikima and Koketsu (2005) estimated that the fault plane is dipping to the east, the aftershock distribution around the hypocenter of the M6.3 aftershock is not consistent with their conclusion as shown in Fig. 3. The above results about the M6.0 aftershock (No. 3) are almost the same as those by Shibutani et al. (2005a), while some other studies had estimated that the hypocenter was deeper than 10 km and was located on the deeper part of the mainshock fault plane (Sakai et al., 2005; Kato et al., 2005). These studies expected that the
M6.0 fault was located on the extension of the mainshock fault. This discrepancy probably results from the differences in station corrections estimated for the permanent stations, which were located around the source region. Shibutani et al. (2005a) determined station corrections from the data including the events that occurred around the source region, while the other studies used only the data of the mainshock and aftershocks. Furthermore, Shibutani et al. (2005a) determined P- and S-wave velocities independently, while the other two studies (Sakai et al., 2005; Kato et al., 2005) determined S-wave velocities assuming a constant Vp /Vs ratio of 1.73. As shown in Shibutani et al. (2005a), Vp /Vs ratios in the shallow layers are much greater than 1.73 in and around the aftershock region. The assumption of the constant Vp /Vs ratio can make locations of shallow aftershocks deeper. Okada et al. (2005, 2006) also analyzed another dense temporary network by the Double-Difference tomography method and estimated that the depth is about 7 km, which is an intermediate to the studies mentioned above. This is probably because they used more permanent stations than the other two studies. They used 54 permanent stations with epicentral distance of less than 60 km, while the other two studies used only
Y. Iio et al. / Physics of the Earth and Planetary Interiors 172 (2009) 345–352
28 stations. Okada et al. (2005, 2006) also used only the data of the mainshock and aftershocks. Thus, it is possible that the structures beneath the permanent stations and their station corrections are not precisely determined. In Fig. 3, there is no aftershock activity around the mainshock fault immediately beneath the hypocenter of the M6.0 aftershock. If the M6.0 aftershock occurred on the mainshock fault plane, its secondary aftershocks (aftershocks of the M6.0 aftershock) should be located in this region, especially since the hypocenters of smaller aftershocks are more accurately determined than those of the major events, because S-arrivals can be used. Fig. 4 displays a 10-km wide depth distribution of the M6.0 aftershock and the earthquakes that occurred immediately after from 18:11 to 19:00 on October 23. It is found that small aftershocks occurred on (or near) the estimated east-dipping fault plane of the M6.0 aftershock, although the standard errors indicated by error bars are larger than 0.5 km (plot criteria in Fig. 3). The errors for the M6.0 (Table 1) are possibly due to the very shallow hypocenters. 3. Discussion It is difficult to explain the spatial change in the lower limit of the aftershock distribution by the thermal structure, since it is thought that the depth of an isothermal surface scarcely changes in the
349
Fig. 4. Depth distribution of the M6.0 aftershock and the earthquakes that occurred immediately after, from 18:11 to 19:00 on October 23 in a 10-km wide X–Z crosssection. Error bars indicate the standard location error of each aftershock.
NNE–SSW direction. The depth possibly changes in the ESE–WNW direction, since active volcanoes are located east of the aftershock region and align in the NNE–SSW direction along the axis of the island arc. This estimate is in good accordance with the distribution of heat flow in the Japanese compiled by Tanaka et al. (2004).
Fig. 5. Location of the very weak region inferred in the lower crust only beneath the central part of the aftershock region. The fault planes and hypocenters of the major events are displayed in different colors. The event numbers correspond to those listed in Table 1. The locations of the S-wave reflectors (Matsumoto et al., 2005) and the region of the low velocity anomaly (Okada et al., 2006) are also shown. The reflectors are schematically illustrated by ellipses in the vertical cross section in the lower panel, and the region in which the reflectors were investigated is also shown by the bracket in the right panel. The region of the low velocity anomaly is roughly enclosed by pink rectangles.
350
Y. Iio et al. / Physics of the Earth and Planetary Interiors 172 (2009) 345–352
Fig. 6. (Upper) Schematic illustration of the depth distribution of the shear strength (black line) and the shear stress before the mainshock (green line) in the upper crust. The shear stress after the mainshock is also shown in profile A (blue line). The dotted lines indicate dynamic friction on the fault. (Lower) Schematic illustration of the very weak region, the weak zone and the rupture propagation of the mainshock and aftershocks. The very weak region exists beneath the central part of the mainshock fault.
Instead these data could be explained by the hypothesis that a localized stress concentration occurred near the bottom of the seismogenic region in the central part of the aftershock region. In this hypothesis, the localized stress concentration is generated by the deformation in the very weak region in the lower crust beneath the central part, as shown in Figs. 5 and 6. Since the aftershock region is located within the concentrated deformation zone (NKTZ), it is likely that there exists the weak zone in the lower crust beneath the aftershock region (Iio et al., 2004; Sibson, 2007). We assume that the strength of the very weak region is much lower than that in the surrounding weak zone. The weak zone beneath the concentrated deformation zone (NKTZ), which extends about 500 km in length, is thought to be weakened by high water content (Iio et al., 2002). We assume that water is concentrated more in the very weak region than that in the surrounding weak zone and then the very weak region is more significantly weakened by higher water content. Laboratory experiments clearly show that larger water fugacity makes the viscosity of rocks in the lower crust smaller (e.g., Rybacki et al., 2006). Fig. 6 schematically illustrates vertical stress profiles above the very weak region (profile B) and the surrounding weak zone (profile A). The mainshock fault shown in this figure is not the fault plane shown in Fig. 5, but the actual slip area roughly deduced from
the slip distribution by Miyazawa et al. (2005). The magnitudes of the shear strength and dynamic friction are proportional to depth, assuming the coefficients of static and dynamic frictions are constant and independent of depth following the experimental results by Byerlee (1978) and pore pressure is hydrostatic in the upper crust (Grawinkel and Stöeckhert, 1997). The difference between the shear strength and dynamic friction increases with depth, since the coefficient of dynamic friction is smaller than that of static friction, as Byerlee (1978) estimated the coefficients of static and dynamic frictions to be 0.6 and 0.47, respectively. It is thought that dynamic friction determines the lowest level of shear stress on the fault plane and the deformation in the lower crust can raise shear stress above this lowest level in the upper crust (Iio et al., 2004). On profile B, a large stress concentration is generated at the bottom of the seismogenic region by the deformation in the very weak region. On the other hand, on profile A, the stress concentration at the bottom of the seismogenic region is relatively small and the difference between the strength and the stress before the mainshock increases with depth. Thus, when the mainshock occurred, the shallower part is broken more easily at the location of profile A. This results from the following two reasons. The first is that the difference between the static and dynamic frictions are thought to be smaller in a shallower part. The second is that the magnitude of
Y. Iio et al. / Physics of the Earth and Planetary Interiors 172 (2009) 345–352
the stress concentration generated at both sides of the mainshock fault is thought to be basically independent of depth, since the distance from the large slip area, which is located in the central part of the fault (e.g., Miyazawa et al., 2005), does not significantly change in depth at both sides of the mainshock fault. The stress profiles shown above could be the reasons why the central three major events (Nos. 1, 4, and 5) initiated at depth and why the two major events (Nos. 2 and 3) occurring in the NNE and SSW portions of the aftershock region initiated near the top of the aftershock distribution, and their faults were confined within the upper part of the upper crust. Evidence that suggests the existence of the very weak region is obtained by the imaging of S-wave reflectors and a seismic tomography. Matsumoto et al. (2005) found distinct S-wave reflectors in the lower crust in the central part of the aftershock region. Okada et al. (2006) estimated a low velocity anomaly in the lower crust only beneath the central part. A similar aftershock distribution showing a bowl shape is seen for the 2000 Western Tottori earthquake (Shibutani et al., 2005b) and the 2005 Western Fukuoka prefecture earthquake (Uehira et al., 2006). Furthermore, a low resistivity zone is recognized beneath the central part of the aftershock region of the 2000 Western Tottori earthquake (Shiozaki et al., 2003). The horizontal extent of the low resistivity zone is about 10 km along the fault. The resistivity value in the central part is estimated to be about 10 m, suggesting high water content, while those in the surrounding parts are 103 to 104 m (Shiozaki et al., 2003). The low resistivity zone might reflect the very weak region in the lower crust. These data suggest that the hypothesis holds for moderate size intraplate earthquakes. 4. Conclusions We precisely determined the aftershock distribution of the 2004 Mid-Niigata prefecture earthquake using a tomographic method. We obtained the following results. (1) The hypocenters of three major events occurring in the central part (Nos. 1, 4, and 5) are deep, near the lower limit of the aftershock distribution, while the other two major events (Nos. 2 and 3) in the NNE and SSW portions of the aftershock region are very shallow, near the top of the aftershock distribution. (2) The fault planes of the two major events occurring in the NNE and SSW portions are shallow, limited to the upper half of the upper crust. (3) There is a tendency that the lower limit of the aftershock distribution is deepest in the central part and becomes shallower toward the NNE and SSW ends. These results could be explained by the hypothesis that a localized stress concentration occurred near the bottom of the seismogenic region only in the central part of the aftershock region. The stress concentration may be generated by the deformation in the very weak region of low strength in the lower crust beneath the central part. The hypothesis presented in this paper clearly supports the local stress model in contrast to the regional stress model. If the hypothesis of the very weak region in the lower crust generally holds for inland earthquakes, the hypothesis is useful for forecasting the locations of large earthquakes, in particular the rupture initiation point. Furthermore, the hypothesis could roughly estimate their rupture extent from the size of the very weak region. Thus, it is important to carefully examine this hypothesis by studies of the location, the lateral and vertical extents, and the properties of the very weak region. Acknowledgements We are very grateful to people in Tochio City, Ojiya City, Nagaoka City (Yamakoshi Village) and Niigata Prefecture who kindly helped
351
us in the aftershock observations. We thank Yuta Asaka, Ling Bai, Mio Hori, Takeshi Katagi, Shigemitsu Matsuo, Koichi Miura, Masahiro Morita, Sun Cheon Park and Atsushi Watanabe for their help in the instrument deployment, and Aiko Nakao for her help in picking the arrival times. We utilized waveform data from permanent stations operated by NIED, JMA and the University of Tokyo and P- and S-arrival times from the JMA preliminary catalogue. We thank James Mori for his critical review of the manuscript. This work was conducted under the support of the Grant-in-Aid for Special Purposes (No. 16800054), MEXT, Japan. This work was also supported by the subsidies to earthquake research projects of the Tokio Marine Kagami Memorial Foundation. References Aoki, S., Nishi, M., Nakamura, K., Hashimoto, T., Yoshikawa, S., Ito, H., 2005. Multiplanar structures in the aftershock distribution of the Mid-Niigata prefecture Earthquake in 2004. Earth Planets Space 57, 5411–5416. Byerlee, J.D., 1978. Friction of rocks. Pure Appl. Geophys. 116, 615–626. Campbell, D.L., 1978. Investigation of the stress-concentration mechanism for intraplate earthquakes. Geophys. Res. Lett. 5, 477–479. Grawinkel, A., Stöeckhert, B., 1997. Hydrostatic pore fluid pressure to 9-km depth; fluid inclusion evidence from the KTB deep drill hole. Geophys. Res. Lett. 24, 2373–2376. 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. Hikima, K., Koketsu, K., 2005. Rupture processes of the 2004 Chuetsu (mid-Niigata prefecture) earthquake, Japan: a series of events in a complex fault system. Goophys. Res. Lett. 32, L18303, doi:10.1029/2005GL023588. Hinze, W.J., Braile, L.W., Keller, G.R., Lidiak, E.G., 1988. Models for mid-continent tectonism: an update. Rev. Geophys. 26, 699–717. Iio, Y., Kobayashi, Y., 2002. A physical understanding of large intraplate earthquakes. Earth Planets Space 54, 1001–1004. Iio, Y., Sagiya, T., Kobayashi, Y., 2004. Origin of the concentrated deformation zone in the Japanese Islands and stress accumulation process of intraplate earthquakes. Earth Planets Space 56, 831–842. Iio, Y., Sagiya, T., Kobayashi, Y., Shiozaki, I., 2002. Water-weakened lower crust and its role in the concentrated deformation in the Japanese Islands. Earth Planet. Sci. Lett. 203, 245–253. Johnston, A.C., Kanter, L.R., 1990. Earthquakes in stable continental crust. Sci. Am. 262 (3), 68–75. Kato, A., Sakai, S., Hirata, N., Kurashimo, E., Nagai, S., Iidaka, T., Igarashi, T., Yamanaka, Y., Murotani, S., Kawamura, T., Iwasaki, T., Kanazawa, T., 2005. Spatiotemporal variations of the aftershock distributions during one month after the occurrence of the 2004 Mid-Niigata prefecture earthquake. Earth Planets Space 57 (6), 551–556. Kenner, S., Segall, P., 2000. A mechanical model for intraplate earthquakes: application to the New Madrid. Science 289, 2329–2332. Kissling, E., Ellsworth, W.L., Eberhart-Phillips, D., Kradolfer, U., 1994. Initial reference models in local earthquake tomography. J. Geophys. Res. 99, 19635–19646. Liu, L., Zoback, M.D., 1997. Lithospheric strength and intraplate seismicity in the New Madrid seismic zone. Tectonics 16, 585–595. Matsumoto, M., Iio, Y., Matsushima, T., Uehira, K., Shibutani, T., 2005. Imaging of S-wave reflectors in and around the hypocentral area of the 2004 mid Niigata Prefecture Earthquake (M6.8). Earth Planets Space 57, 557–561. Miyazawa, M., Mori, J., Iio, Y., Shibutani, T., Matsumoto, S., Katao, H., Ohmi, S., Nishigami, K., 2005. Triggering sequence of large aftershocks of the Mid-Niigata Prefecture, Japan Earthquake in 2004 by static stress changes. Earth Planets Space 57, 1109–1113. National Research Institute for Earth Science and Disaster Prevention (NIED), 2004. Earthquake mechanism information. Available from:
352
Y. Iio et al. / Physics of the Earth and Planetary Interiors 172 (2009) 345–352
Sakai, S., Hirata, N., Kato, A., Kurashimo, E., Iwasaki, T., Kanazawa, T., 2005. Multifault system of the 2004 Mid-Niigata Prefecture Earthquake and its aftershocks. Earth Planets Space 57 (5), 417–422. Sibson, R.H., 2007. An episode of fault-valve behaviour during compressional inversion?—The 2004 MJ6.8 Mid-Niigata Prefecture, Japan, earthquake sequence. Earth Planet. Sci. Lett. 257, 188–199, doi:10.1016/j.epsl.2007.02.031. Shibutani, T., Iio, Y., Matsumoto, S., Katao, H., Matsushima, T., Ohmi, S., Takeuchi, F., Uehira, K., Nishigami, K., Enescu, B., Hirose, I., Kano, Y., Kohno, Y., Korenaga, M., Mamada, Y., Miyazawa, M., Tatsumi, K., Ueno, T., Wada, H., Yukutake, Y., 2005a. Aftershock distribution of the 2004 Mid-Niigata Prefecture Earthquake derived from a combined analysis of temporary online observations and permanent observations. Earth Planets Space 57, 545– 549. Shibutani, T., Katao, H., Group for the dense aftershock observations of the 2000 Western Tottori Earthquake, 2005b. High resolution 3-D velocity structure in the source region of the 2000 Western Tottori Earthquake in southwestern Honshu, Japan using very dense aftershock observations. Earth Planets Space 57, 825–838. Shiozaki, I., Oshiman, N., Kasaya, T., Uto, T., Kagiyama, T., 2003. Deep crustal resistivity structure derived from the wide band MT investigations in the eastern part of Sanin region, southwestern Honshu, Japan, IUGG2003 JSA10/30P/D-006, XX111 General Assembly of the International Union of Geodesy and Geophysics.
Stuart, W.D., Hildenbrand, T.G., Simpson, R.W., 1997. Stressing of the New Madrid seismic zone by a lower crust detachment fault. J. Geophys. Res. 102, 27623–27633. Sykes, L.R., 1978. Intraplate seismicity, reactivation of preexisting zones of weakness, alkaline magmatism, and other tectonism postdating continental fragmentation. Rev. Geophys. 16, 621–688. Tanaka, A., Yamano, M., Yano, Y., Sasada, M., 2004. Geothermal gradient and heat flow data in and around Japan. I. Appraisal of heat flow from geothermal gradient data. Earth Planets Space 56, 1191–1194. Thurber, C.H., 1983. Earthquake locations and three-dimensional crustal structure in the Coyote Lake area, central California. J. Geophys. Res. 88, 8226– 8236. Uehira, K., Yamada, T., Shinohara, M., Nakahigashi, K., Miyamachi, H., Iio, Y., Okada, T., Takahashi, H., Matsuwo, N., Uchida, K., Kanazawa, T., Shimizu, H., 2006. Precise aftershock distribution of the 2005 West Off Fukuoka Prefecture Earthquake (Mj = 7.0) using a dense onshore and offshore seismic network. Earth Planets Space 58, 1605–1610. Yamawaki, T., Nishimura, T., Hamaguchi, H., 2004. Temporal change of seismic structure around Iwate Volcano inferred from waveform correlation analysis of similar earthquakes. Geophys. Res. Lett. 31, 2004. Zoback, M.L., 1992. Stress field constraints on intraplate seismicity in eastern North America. J. Geophs. Res. 97, 11761–11782.