LETTER Earth Planets Space, 60, 1131–1135, 2008 Source process of the 2007 Niigata-ken Chuetsu-oki earthquake derived from near-fault strong motion data Shin Aoi1, Haruko Sekiguchi2, Nobuyuki Morikawa1, and Takashi Kunugi1 1National Research Institute for Earth Science and Disaster Prevention, 3-1 Tennodai, Tsukuba, Ibaraki 305-0006, Japan 2National Institute of Advanced Industrial Science and Technology, Site 7, 1-1-1 Higashi, Tsukuba, Ibaraki 305-8567, Japan (Received December 8, 2007; Revised February 18, 2008; Accepted February 26, 2008; Online published November 18, 2008) The 2007 Niigata-ken Chuetsu-oki earthquake generated strong ground motions in Kashiwazaki and Kariwa, where the world largest nuclear power plant was in operation. Due to the complexity of the aftershock distribution, activation of the northwest-dipping fault and/or the southeast-dipping fault is proposed. To explore the fault geometry and source process of the earthquake, we performed multi-time window linear waveform inversions for both the fault planes from near-fault strong motion data. A fault plane model of 30 km in length by 24 km in width was set to cover the region of aftershock distribution within 24 h of the mainshock. Both inverted slip models provided moment magnitudes of 6.7 with a small asperity near the rupture starting point, and a large asperity approximately 10 km southwest of the rupture initiation, which is located in the region of relatively sparse aftershock distribution. Both the small and large asperities are located near the intersection between the two conjugate fault plane models, and the asperities of both models have similar radiation patterns. Therefore, the difference of the residuals between the observed and synthetic waveforms for both models was not significant, indicating that it is difficult to conclude which fault is the rupture. Key words: Source process, waveform inversion, 2007 Niigata-ken Chuetsu-oki earthquake, near-fault strong motion. 1. Introduction strike direction based on Hi-net (Obara, 2003) and F-net The 2007 Niigata-ken Chuetsu-oki earthquake occurred (Fukuyama et al., 1996) analyses (Fig. 1). The aftershock on July 16, 2007, 10:13 JST (37.557N, 138.608E, 17 km distribution for the earthquake is quite complex, so that rup- depth, MJMA = 6.8; Japan Meteorological Agency, here- ture on both the SE-dipping fault (e.g., JMA, 2007; Yuku- after JMA). The earthquake generated strong ground mo- take et al., 2008) and the NW-dipping fault (e.g., Hirata et tions in Kashiwazaki and Kariwa. During this event, al., 2007) was proposed. A series of F-net moment ten- K-NET and KiK-net strong-motion networks (Kinoshita, sor analyses indicated that major aftershocks with reverse 1998; Aoi et al., 2000, 2004) recorded ground acceler- faulting aligned with the SE-dipping fault (Matsumoto et ations at 390 and 307 stations, respectively. The maxi- al., 2007). The Geographical Survey Institute (GSI, 2007) mum peak ground acceleration (PGA) of 813 cm/s2 was reported that the most plausible source model satisfying observed at the K-NET Kashiwazaki station (NIG018), and geodetic data is a combination of two NW-dipping faults. distinct non-linear behavior was identified at the site (e.g., However, these analyses of the aftershock distribution, mo- Aoi and Morikawa, 2007; Yoshida et al., 2007). Very strong ment tensors, and geodetic data have yet to properly clarify ground motions were recorded at the foundation slabs (base the fault plane responsible for the mainshock. mats) of all seven reactors of the Kashiwazaki-Kariwa nu- Figure 2 showed that the observed peak ground accelera- clear plant operated by the Tokyo Electric Power Company tions (PGAs) and peak ground velocities (PGVs) are consis- (TEPCO), which is located on stiff soil with a Vs of ap- tent with the empirical attenuation relations of Si and Mi- proximately 500 m/s. The maximum PGA recorded at the dorikawa (1999) at fault distances larger than 20 km,in- reactor foundation slabs (base mats) was 680 cm/s2, which dependent of which fault plane is assumed. The observed exceeds the design level of the nuclear power plant. Two or PGAs and PGVs in the near-fault region, on the other hand, three pulses are clearly seen in the ground motions at these were obviously larger than predicted in the case of the SE- reactors. Similar pulses were widely observed in the near- dipping fault, yet remain in general agreement with the pre- fault area, suggesting a complex rupture process consisting dictions assuming the NW-dipping fault. However, in the of multiple subevents. near-fault region, the fault distance used in the attenuation The focal mechanism of the earthquake was estimated to relationship becomes more sensitive to the fault geometry, be a nearly pure reverse fault with a northeast-southwest such as the assumed location and size of the fault plane. In addition, the near-fault ground motions are strongly af- Copyright c The Society of Geomagnetism and Earth, Planetary and Space Sci- fected by complex source processes and the hanging-wall ences (SGEPSS); The Seismological Society of Japan; The Volcanological Society of Japan; The Geodetic Society of Japan; The Japanese Society for Planetary Sci- effect (Abrahamson and Somerville, 1996). Therefore, we ences; TERRAPUB. cannot uniquely determine which fault plane is appropriate 1131 1132 S. AOI et al.: SOURCE PROCESS OF THE 2007 NIIGATA-KEN CHUETSU-OKI EARTHQUAKE km 38.0 0 10 20 NIGB03 NIG011 NIG004 NIG016 Model A NW-dipping NIGH06 L65039 37.5 NIG017 TEPCO NIGH01 NIG018 P Model B NIG019 T SE-dipping NIG020 JMACB6 NIGH11 NIG026 NIG023 37.0 138.0 138.5 139.0 139.5 Fig. 1. Distribution of the observation stations used for the inversion analysis (solid circles). Solid squares show stations of the K-NET Kashiwazaki (NIG018) and the Kashiwazaki-Kariwa nuclear power plant (TEPCO). Rectangles show the surface projections of the two fault plane models. Star indicates the epicenter of the mainshock. Open circles denote aftershock locations within 24 h of the mainshock estimated by Yukutake et al. (2008), using manual picking of Hi-net data and hypoDD (Waldhauser and Ellsworth, 2000). Focal mechanisms were estimated by the polarity of P-wave first motion by Hi-net (left) and the moment tensor analysis of F-net data (right). Model A (NW-dipping) Izumozaki-Komeda Kashiwazaki-Nishiyama Kashiwazaki-Takayanagi NIG018 1000 Nagaoka-Oguni 100 KKZ1R2 KKZ1R2 Iizuna-Imogawa KSHSG1 Kashiwazaki- KKZ5R2 Chuo NIGH13 KKZ5R2 NIGH11 KSHSG1 100 NIG016 10 Izumozaki- ] ] s Kawanishi / s s Kariwa 10 1 PGV [cm/ PGA [cm/ 1 0.1 0.1 0.01 1 10 100 1000 1 10 100 1000 Distance [km] Distance [km] Model B (SE-dipping) Kashiwazaki-Nishiyama NIG018 Kashiwazaki-Chuo KKZ1R2 Kariwa KKZ1R2 1000 Kashiwazaki-Takayanagi 100 Izumozaki-Komeda KSHSG1 KKZ5R2 KKZ5R2 Iizuna-Imogawa KSHSG1 Nagaoka-Oguni NIGH13 Izumozaki-Kawanishi NIGH11 100 10 NIG016 ] ] s / s s 10 1 PGV [cm/ PGA [cm/ 1 0.1 0.1 0.01 1 10 100 1000 1 10 100 1000 Distance [km] Distance [km] Fig. 2. Comparisons of the observed peak ground accelerations (PGAs; left panels) and peak ground velocities (PGVs; right panels) with the empirical attenuation relations. Upper and lower panels show the cases respectively assuming Model A (NW-dipping fault) and Model B (SE-dipping fault) used in this study. Dashed and dotted lines represent the empirical attenuation relations by Si and Midorikawa (1999) and those standard deviations, respectively. PGVs are converted from the observed velocities on the ground surface into those on a stiff-soil site, where Vs is 600 m/s, following the method of Si and Midorikawa (1999). The number of PGVs is less than that of PGAs because the method requires Vs structure of the stiff-soil layer or 30 m in depth at the site. S. AOI et al.: SOURCE PROCESS OF THE 2007 NIIGATA-KEN CHUETSU-OKI EARTHQUAKE 1133 Model A: NW-dipping Model B: SE-dipping strike strike N215E N49E 0 0 m 2.5 dip 49 dip 42 2.0 10 10 1.5 1.0 0.5 20 20 0.0 2m 0210 030 0210 030 Fig. 3. Estimated total slip distribution for Model A (NW-dipping fault) and Model B (SE-dipping fault). Star indicates the rupture starting point. Arrows show the amplitude and direction of slip. Model A: NW-dipping Model B: SE-dipping Obs. Syn. EW NS UD EW NS UD 0.015 0.021 0.013 [m/s] 0.015 0.021 0.013 [m/s] 0.015 0.015 0.010 0.013 0.015 0.010 NIG004 0.015 0.018 0.015 0.015 0.018 0.015 0.013 0.013 0.009 0.013 0.007 0.008 NIGB03 0.027 0.022 0.012 0.027 0.022 0.012 0.019 0.012 0.004 0.009 0.013 0.005 NIG011 0.020 0.035 0.021 0.020 0.035 0.021 0.008 0.011 0.009 0.008 0.009 0.007 NIGH06 0.037 0.019 0.013 0.037 0.019 0.013 0.014 0.011 0.005 0.022 0.017 0.009 NIG016 0.272 0.251 0.060 0.272 0.251 0.060 0.134 0.093 0.055 0.097 0.047 0.036 L65039 0.119 0.048 0.046 0.119 0.048 0.046 0.062 0.049 0.034 0.054 0.034 0.030 NIG017 0.056 0.059 0.014 0.056 0.059 0.014 0.056 0.050 0.011 0.031 0.043 0.019 NIGH01 0.194 0.093 0.053 0.194 0.093 0.053 0.109 0.088 0.034 0.122 0.040 0.015 NIG019 0.033 0.046 0.019 0.033 0.046 0.019 0.019 0.032 0.017 0.026 0.030 0.009 NIG020 0.063 0.087 0.051 0.063 0.087 0.051 0.039 0.065 0.034 0.035 0.050 0.016 NIGH11 0.033 0.030 0.032 0.033 0.030 0.032 0.015 0.026 0.026 0.025 0.030 0.018 NIG023 0.031 0.036 0.025 0.031 0.036 0.025 0.025 0.023 0.019 0.024 0.025 0.020 NIG026 0.043 0.044 0.034 0.043 0.044 0.034 0.028 0.025 0.028 0.034 0.020 0.038 JMACB6 0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14 [sec] [sec] [sec] [sec] [sec] [sec] Fig.