Source process and near-source ground motions of the 2005 West Off Fukuoka Prefecture earthquake

Transcription

1 LETTER Earth Planets Space, 8, 93 98, 26 Source process and near-source ground otions of the 2 West Off Fukuoka Prefecture earthquake Kiiyuki Asano and Tootaka Iwata Disaster Prevention Research Institute, Kyoto University, Gokasho, Uji, Kyoto , Japan (Received August 8, 2; Revised October 4, 2; Accepted October 12, 2; Online published January 27, 26) A large shallow crustal earthquake occurred at the western off-shore of Fukuoka Prefecture, northern Kyushu, Japan, at 1:3 on March 2, 2 (JST). Source rupture processes of the ainshock and the largest aftershock on April 2, 2, are estiated by the kineatic wavefor inversion of strong otion seisogras. The rupture of the ainshock started with relatively sall slip, and the largest slip was observed at the southeast of the hypocenter. The inverted source odels showed that both of the ruptures of the ainshock and the largest aftershock ainly propagated to southeast fro the hypocenters, and the rupture area of those events did not overlap each other. Three-diensional ground otion siulation by the finite difference ethod considering three-diensional bedrock structure was also conducted to see the spatial variation of the near-source ground otion of the ainshock. The result of the siulation shows that expected groud otions are relatively large in and arround Genkai Island, Shikanoshia Island, and the center of Fukuoka City copared to the other area because of the rupture heterogeneity and the deep basin structure in Fukuoka City. Key words: the 2 West Off Fukuoka Prefecture earthquake, source process, kineatic wavefor inversion, strong ground otion siulation. 1. Introduction At 1:3 on March 2, 2 (JST), a large shallow crustal earthquake (M JMA 7.) occurred beneath the Sea of Genkai, western off-shore of Fukuoka Prefecture, northern Kyushu, Japan. Though this earthquake occurred beneath the sea, it brought severe strong ground otions to the nearsource region, such as Genkai Island, Shikanoshia Island, and the central district of Fukuoka City. At 6:11 on April 2, 2, the largest aftershock (M JMA.8) occurred just below Shikanoshia Island near the southeastern edge of the fault plane of the ainshock. The ain purpose of this paper is to understand strong otion generation process of these earthquakes. At first, the source process of the 2 West Off Fukuoka Prefecture earthquake is studied using the strong otion seisogras obtained by Japanese nation-wide strong otion seisograph networks, K-NET (Kinoshita, 1998) and KiKnet (Aoi et al., 21). These networks are installed and operated by the National Research Institute for Earth Science and Disaster Prevention (NIED). The source process of the largest aftershock on April 2, 2, is also estiated in the sae anner. Relationship between the fault rupture of the ainshock and that of the largest aftershock is discussed. Finally, a three-diensional ground otion siulation using a finite difference ethod reveals the spatial variation of ground otions in the near-source area. Copyright c The Society of Geoagnetis and Earth, Planetary and Space Sciences (SGEPSS); The Seisological Society of Japan; The Volcanological Society of Japan; The Geodetic Society of Japan; The Japanese Society for Planetary Sciences; TERRAPUB. 2. Source Process of the Mainshock 2.1 Methodology The source rupture process of the ainshock is estiated by the kineatic linear wavefor inversion with ultiple tie windows (Sekiguchi et al., 2). This ethodology is originally based on the technique developed by Hartzell and Heaton (1983). The ethodology eployed here was described in detail by Sekiguchi et al. (2). A planar fault plane odel is assued referring to the aftershock distributions. The length and width of the fault plane are 26 k and 18 k, respectively. According to the oent tensor solution by the F-net (NIED), the strike and dip of the fault plane are assued to be 122 and 87, respectively. The rupture starting point is fixed at the hypocenter location (33.7 N, E, 14 k) deterined by the Institute of Seisology and Volcanology, Faculty of Sciences, Kyushu University (ISV). The fault plane is divided into 117 subfaults of2k 2 k. The teporal slip history at each subfault is expressed by a series of 6 soothed rap functions, which has the rise tie of 1. s, separated by. s. The rupture front propagation velocity triggering the rupture of the first tie window is selected so as to iniize the residual of data fitting. The spatio teporal soothing to reduce instability or excess of coplexity, which reduces the differences aong slip aounts close in space and in tie, is also included in the analysis following Sekiguchi et al. (2). ABIC approach (Akaike, 198) is used to select an appropriate strength of the soothing constraint. The rake angles variation are also liited within ±4 using the nonnegative least-squares ethod (Lawson and Hanson, 1974). The Green s functions between each subfault and each station are calculated by the discrete wavenuber ethod 93

2 94 K. ASANO AND T. IWATA: SOURCE PROCESS OF THE 2 WEST OFF FUKUOKA PREFECTURE EARTHQUAKE NGS22 NGS2 k FKO1 NGS23 FKO FKOH3 FKO6 FKO7 FKOH4 SAGH1 FKOH8 SAG2 SAGH3 SAGH4 SAG SAGH2 Along Dip (k) -1 - N122 E Along Strike (k) Fig. 2. Final slip distribution of the ainshock estiated fro the inversion. The open star indicates the rupture starting point. The allows show the slip vectors of the hanging wall relative to the foot wall. The interval of contours is Fig. 1. Map showing studied area. The black and gray stars indicate the epicenters of the ainshock and the largest aftershock, respectively. The solid line shows the projection of the fault plane of the ainshock. Solid triangles indicate the locations of strong otion stations used for the wavefor inversion. FKOH3 FKOH8 FKO FKO FKOH4 FKO1 FKO6 SAGH1 Obs. Syn (Bouchon, 1981) with the reflection and transission atrix ethod (Kennett and Kerry, 1979). Data at nine stations of the K-NET and seven stations of the KiK-net are used for the wavefor inversion (see Fig. 1). For the KiK-net stations, uphole seisograph data is used. Fro aong any available stations, stations which have enough quality to retrieve the source process were selected by the visual inspection. Observed digital acceleration data are integrated into the ground velocities in the tie doain and bandpass filtered with a Chebyshev filter between. and 1. Hz. We inverted 16 s of the S-wave portion fro 1 s before the direct S-wave arrival. Theoretical Green s functions are also bandpass filtered in the sae anner. 2.2 Underground structure odel To calculate the Green s functions, a one-diensional underground structure odel is assued for each station referring to the underground structure odel proposed by Nakaichi and Kawase (22). They used the underground structure odel to evaluate the strong ground otions for a scenario earthquake on the Kego fault in Fukuoka City, so we think this odel would be appropriate for the studied area in this paper. Several low-velocity layers for superficial layers at each station are also introduced based on the borehole logging inforation released by the K-NET and KiK-net. So, the velocity odels of the sedientary part differ slightly fro one station to another. 2.3 Result Figure 2 shows the final slip distribution on the fault surface estiated by the inversion. The rupture front propagation velocity, which triggers the rupture of the first tie SAGH2 SAGH4 SAG NGS SAGH3 SAG2 NGS2 NGS Fig. 3. Coparison between observed (black traces) and synthesized (gray traces) velocity wavefors of the ainshock. The axiu aplitude of each coponent of observed wavefors are shown above each trace in c/s. The horizontal axis is tie (s). window, was selected to be 2.1 k/s. This is approxiately equal to 6% of the shear-wave velocity at the depth of the rupture, that is a little slower than the average rupture velocity (72%) of shallow crustal earthquakes epirically obtained by Geller (1976). The rupture ainly propagated to the southeastward. The asperity or large slip area with the axiu slip of 3.2 was observed at southeast of the hypocenter, and relatively saller slip was observed in the vicinity of the hypocenter. The slip direction is alost pure left-lateral strike slip, which is consistent with the regional stress field in northern Kyushu (Seno, 1999). Figure 3 shows the coparison between the observed and synthesized ground velocities in. 1. Hz. The synthesized wavefors atch well the observed ones at ost stations. We think that the source odel obtained here is reliable to represent the source rupture process of this earthquake. However, the large aplitude of velocity pulse in the NS coponent at Fukuoka (FKO6), which is located inside the Fukuoka basin, could not be reproduced well.

3 K. ASANO AND T. IWATA: SOURCE PROCESS OF THE 2 WEST OFF FUKUOKA PREFECTURE EARTHQUAKE 9.~1. s 7.~9. s 1.~3. s 3.~4. s 4.~6. s 6.~7. s 9.~1. s Total /s Fig. 4. Snapshots of the teporal rupture progression on the fault at tie step of 1. s. The contour interval of the slip velocity is.3 /s. The open star indicates the rupture starting point. The figure at botto right shows the final slip distribution. Figures 4 and show the teporal rupture progression on the fault. The rupture started fro the hypocenter with the sall slip velocity, and the entire rupture continued for approxiately 1 s. The asperity ruptured at approxiately 3. s after the initiation of the rupture. The total seisic oent was estiated to be N (M W 6.6). The characteristics of the rupture process of this event entioned in this section is consistent with the analysis of the initial and ain rupture phases by Takenaka et al. (26), in which they deterined the rupture tie and the location of the ain rupture using the onset data of strong otion records. Other studies on the rupture process of this event (Horikawa, 26; Kobayashi et al., 26; Sekiguchi et al., 26) also show siilar results to our source odel. 3. Source Process of the Largest Aftershock on April 2, Fault odel and data The source rupture process of the largest aftershock is estiated by the sae approach presented for the ainshock. The length and width of the fault plane are assued to be 8 k and 8 k, respectively. According to the oent tensor solution by the F-net, the strike and dip of the fault plane are assued to be 132 and 9, respectively. The rupture starting point is fixed at the hypocenter location (33.67 N, E, 13 k) deterined by the ISV, Kyushu University. The fault plane is divided into 64 subfaults of 1 k 1 k. The teporal slip history at each subfault is expressed by a series of 6 soothed rap functions, which has the rise tie of.6 s, separated by.3 s. Integrated ground velocities are bandpass filtered with a Along Dip Direction 3.s Along Strike Direction Fig.. Moent rate functions of each subfault of the ainshock. The open star indicates the hypocenter. Along Dip (k) N132 E Along Strike (k) 1 Fig. 6. Final slip distribution of the largest aftershock on April 2, 2, estiated fro the inversion. The open star indicates the rupture starting point. The allows show the slip vectors of the hanging wall relative to the foot wall. The interval of contours is.2. Chebyshev filter between.1 and 1. Hz. We inverted 6 s of the S-wave portion fro 1 s before the S-wave arrival. The underground structure odel to calculate Green s functions is the sae for the ainshock odeling. 3.2 Result and discussions Figure 6 shows the final slip distribution on the fault surface estiated by the wavefor inversion. The rupture front propagation velocity, which triggers the rupture of the first tie window, was selected to be 2.6 k/s. The rupture velocity was not constrained well because of its sall spatial extent. As in the case of the ainshock, the rupture propagated to the southeastward. The northwest of the hypocenter had alost no slip on the fault. Figure 7 shows the coparison between observed and synthesized ground velocities in.1 1. Hz. The synthesized wavefors fit the observed ones fairly well. Figure 8 shows the oent rate functions at each subfault. The asperity was located in the vicinity of the hypocenter. The teporal rupture history was also rather siple, and the duration of source tie function at each subfault is approxiately 1 s on the asperity. The total seisic oent was obtained to be N (M W.6). Figure 9 shows the spatial slip distributions of the ainshock and the largest aftershock. The rupture area of the largest aftershock and that of the ainshock do not overlap x1 17 N/s

4 96 K. ASANO AND T. IWATA: SOURCE PROCESS OF THE 2 WEST OFF FUKUOKA PREFECTURE EARTHQUAKE FKOH3 FKOH8 FKO FKO7 SAGH2 SAGH4 SAG NGS Obs. Syn FKOH FKO FKO SAGH SAGH SAG NGS NGS Depth (k) 13 E 13.2 E 13.4 E 33.8 N 33.8 N Distance (k) Depth (k) M>2 (N=124) 2/3/2 1:3 ~2/3/21 1:2 2/3/2 1:3 ~2/4/19 6:1 2/4/2 6:11 ~2/4/21 6:1 Distance (k) Fig. 7. Coparison between observed (black traces) and synthesized (gray traces) velocity wavefors of the largest aftershock. The axiu aplitude of each coponent of observed wavefors are shown above each trace in c/s. The horizontal axis is tie (s). Along Dip Direction 2.1s Along Strike Direction Fig. 8. Moent rate functions of each subfault of the largest aftershock. The open star indicates the hypocenter. each other. Before the largest aftershock on April 2, few aftershocks occurred arround the hypocenter of this event. The strike of the spatial series of aftershocks which occured after the largest aftershock is slightly differs fro that of aftershocks before the largest aftershock. The spatial and teporal pattern of aftershock distributions suggests that the largest aftershock ight rupture the area in which the ainshock did not rupture. 4. Three-Diensional Ground Motion Siulation 4.1 Ground otion siulation using the finite difference ethod To see the spatial variation of strong ground otions in the near-source region, a three-diensional ground otion siulation is carried out with the use of the finite difference ethod using staggered grids with nonunifor spacing developed by Pitarka (1999). The source odel obtained above is adopted to siulate the ground otion generation on the fault. A three-diensional underground structure odel is constructed by anually digitizing the gravity baseent depth contour ap, which was produced by Koazawa (2) fro gravity anoalies. Figure 1 shows the depth of the bedrock surface with 4.1x1 16 N/s Fig. 9. Map showing hypocenters of aftershocks whose agnitude are larger than 2.. The open circles are the hypocenters of aftershocks deterined by the JMA. The colors of circles depend on the periods. Vertical section along the strike direction of the ainshock (N122 E) with the large slip area obtained by the kineatic wavefor inversion of the ainshock and the largest aftershock is also shown in the botto figure. The area in which the slip is larger than.8 is colored orange for the ainshock, and the area in which the slip is larger than.2 is colored light blue for the largest aftershock. The open star indicates the epicenter of the ainshock used in this study. the shear-wave velocity (V S ) of 28 /s. Three sedientary layers (V S = 6, 11, 17 /s) are assued above the bedrock. Material paraeters assued for this layered underground structure odel basically follow that of Nakaichi and Kawase (22), and those paraeters of each layer are listed in Table 1. h is the depth of the bedrock surface shown in Fig. 1. The diensions of the odel space are 71, 71, and 28 in the x, y, and z directions. The grid spacing is set to.1 k in the x and y directions and.1 k and.2 k in the z direction in the depth intervals of to 2 k and 2 to 4 k, respectively. The tie step is.682 s, the total nuber of tie steps is 2933, and the duration of synthetic wavefors is 2 s. The target frequency range of the calculation is up to 1. Hz. The siulated distribution of axiu horizontal velocity at surface is shown in Fig. 11. The area where large ground otion was expected fro the siulation extended to the southeastward fro the fault because of the relatively deep basin structure in the center of Fukuoka City as well as the forward directivity effect of the fault rupture. Genkai Island and Shikanoshia Island are located in the area with large ground otion above the fault. The localized area with relatively large ground otions were also expected at the northwest of the hypocenter and around the hypocenter. Those are ainly controlled by the teporal fault rupture process. 4.2 Discussion Figure 12 shows the distribution of strong otion stations in Fukuoka City by the K-NET, the Japan Meteorological Agency (JMA), and Fukuoka City Governent. Recently, the JMA and local governents of all over Japan have deployed seisic intensity and strong ground otion observation syste (e.g., Nishiae, 24). In Fukuoka City, strong ground otions during the ainshock were densely

5 K. ASANO AND T. IWATA: SOURCE PROCESS OF THE 2 WEST OFF FUKUOKA PREFECTURE EARTHQUAKE E 33.7 N Genkai Is E Sea of Genkai 13.4 E 13. E 33.7 N Shikanoshia Is N 9 Hakata Bay 97 FKO EDF Kego fault 33.6 N N 13.2 E 1 k 13.3 E 13.4 E 33. N 13. E Fig. 1. Depth contour ap showing the surface of the bedrock with the shear-wave velocity of 28 /s. The interval of contours is 2. Fig. 12. Map showing the locations of strong otion stations in Fukuoka City. The station FKO6 belongs to the K-NET. The station EDF belongs to the JMA. The stations were deployed by Fukuoka City Governent. The broken line indicates the surface trace of the Kego fault (Nakata and Iaizui, 22). Table 1. Model paraeters of each layer of the underground structure odel used in the finite difference calculation. Depth V P V S ρ Q () (/s) (/s) (kg/ 3 ) h h h h is the depth of the bedrock surface shown in Fig c/s FKO6 (K-NET) EDF (JMA) 92 (Hakata) 93 (Chuo) 94 (Minai) 9 (Nishi) 96 (Jonan) EW-cop. NS-cop. UD-cop. Obs Syn Obs Syn Obs Syn Obs Syn Obs Syn Obs Syn Obs Syn Obs (Sawara) Syn Fig. 13. Coparison of observed ground velocities with synthetic ones at stations in Fukuoka City. All traces are bandpass filtered between. 1 Hz. Top and botto traces for each station show the observed and synthetic wavefors, respectively. The peak ground velocity of each trace is shown at above each trace in c/s. 2s Fig. 11. Spatial distribution of sliulated axiu horizontal velocity at surface. The interval of contours is 1 c/s. recorded by these organizations. Figure 13 shows the coparison between observed and synthetic ground velocities (. 1 Hz) at the stations of the K-NET, the JMA, and Fukuoka City Governent. Synthetic wavefors fit observed ones fairly well except several coponents. The observed aplitude of NS coponents at FKO6 and 93 are exceptionally large copared with those at other stations. The siulation result could not succeed to reproduce the large aplitude of these coponents. As shown in Fig. 12, the Kego fault runs through the central part of Fukuoka City (e.g., Nakata and Iaizui, 22; Oniki, 1996; The Research Group for Active Faults of Japan, 1991). Several studies reported that the depth of Quaternary deposits accuulated on the northeast side of the Kego fault are deeper than that on the southwest side (e.g., Karakida et al., 1994; Nakaichi and Kawase, 22; Oniki, 1996). The axiu thickness of Quaternary deposits along the Kego fault is approxiately 6. Since the assued shear-wave velocity at the surface in Table 1 is 6 /s that corresponds to Paleogene sandstone, the Quaternary deposit was not included in this siulation. It

6 98 K. ASANO AND T. IWATA: SOURCE PROCESS OF THE 2 WEST OFF FUKUOKA PREFECTURE EARTHQUAKE should cause the underestiation of synthetic ground velocities. Moreover, the horizontal irregularity of the bedrock structure caused by the Kego fault is not explicitly considered in our three-diensional structure odel, because the odel is based on the gravity data and rather sooth. Such a lateral irregurarity causes the localized aplification along the fault trace in soe case (e.g., Kawase, 1996; Pitarka et al., 1998). The isatches of synthetic and observed ground velocities entined above are found only at stations located at northeast side of the Kego fault. These isatched ight be attributed to the insufficient odeling with lack of the Quaternary deposits and the coplex subsurface structure. More detailed odeling of the underground structure in Fukuoka City would be necessary to iprove the propagation and aplification characteristics inside the basin.. Conclusions The source rupture odels of the 2 West Off Fukuoka Prefecture earthquake and its largest aftershock were estiated by the kineatic wavefor inversion of strong otion seisogras. The rupture area of the ainshock and that of the largest aftershock do not overlap each other. The ruptures of both events ainly propagated to the southeast direction fro the hypocenters, and the asperity of the ainshock was observed at southeast of the hypocenter, and relatively saller slip was observed in the vicinity of the hypocenter. The forward siulation of the ground otions in the near-source region by the finite difference ethod showed that larger ground velocities were expected in Genkai Island, Shikanoshia Island, and the central district of Fukuoka City copared to other region. The feature and duration of the synthetic ground velocities in Fukuoka City atched the observed ones fairly well. However, absolute aplitude of the NS coponents of the synthetic ground velocities at several stations located at the footwall side of the Kego fault were underestiated copared with observed aplitudes. This underestiation would be caused by the insufficient odeling of coplex subsurface structure in this area. Acknowledgents. We are deeply indebted to the K-NET and the KiK-net operated by the NIED for releasing the strong otion data. We are also grateful to the JMA, Fukuoka City Governent, and Prof. Hiroshi Kawase at Kyushu University for strong otion data recorded in Fukuoka City. We used the hypocenter inforation of the ISV, Kyushu University and the JMA, and the oent tensor catalog of the F-net by the NIED. We would like to thank Dr. Haruko Sekiguchi and Dr. Arben Pitarka for kindly allowing us to use their original codes. This work has used Active Fault Shape File fro Nakata and Iaizui (22) (product serial nuber: DAFM1483). The Generic Mapping Tools (Wessel and Sith, 1998) was used to draw all the figures in this paper. The coents fro Prof. Hiroshi Takenaka and Dr. Alessio Piatanesi were helpful in iproving the anuscript. This study is partially supported by Grant-in-Aid for Special Purposes (1781, PI. Prof. Kawase, Kyushu Univ.) and the Special Project for Earthquake Disaster Mitigation in Urban Areas fro the Ministry of Education, Culture, Sports, Science, and Technology, Japan. References Akaike, H., Likelihood and the Bayes procedure, in Bayesian Statistics, edited by J. M. Bernardo, M. H. DeGroot, D. V. Lindley, and A. F. M. Sith, pp , University Press, Valencia, Spain, 198. Aoi, S., K. Obara, S. Hori, K. Kasahara, and Y. Okada, New Japanese uphole/downhole strong-otion observation network: KiK-net, Seis. Res. Lett., 72, 239, 21. Bouchon, M., A siple ethod to calculate Green s functions for elastic layered edia, Bull. Seis. Soc. A., 71, , Geller, R. J., Scaling relations for earthquake source paraeters and agnitudes, Bull. Seis. Soc. A., 71, , Hartzell, S. H. and T. H. Heaton, Inversion of strong ground otion and teleseisic wavefor data for the fault rupture history of the 1979 Iperial Valley, Carifornia, earthquake, Bull. Seis. Soc. A., 73, , Horikawa, H., Rupture process of the 2 West Off Fukuoka Prefecture, Japan, earthquake, Earth Planets Space, 8, this issue, 87 92, 26. Karakida, Y., S. Toita, S. Shioyaa, and K. Chijiwa, Geology of the Fukuoka District, with Geological Sheet Map at 1:,, 192 pp., Geological Survey of Japan, Tsukuba, 1994 (in Japanese with English suary). Kawase, H., The cause of the daage belt in Kobe: The basin-edge effect, constructive interference of the direct S-wave with the basininduced diffracted/rayleigh waves, Seis. Res. Lett., 67(), 2 3, Kennett, B. L. N. and N. J. Kerry, Seisic waves in a stratified half space, Geophys. J. Roy. Astr. Soc., 7, 7 83, Kinoshita, S., Kyoshin-net (K-NET), Seis. Res. Lett., 69, , Kobayashi, R., S. Miyazaki, and K. Koketsu, Source processes of the 2 West Off Fukuoka Prefecture earthquake and its largest aftershock inferred fro strong otion and 1-Hz GPS data, Earth Planets Space, 8, this issue, 7 62, 26. Koazawa, M., Gravity field around the source region of the 2 west off Fukuoka earthquake, Rep. Coord. Co. Earthq. Pred., 74, 7 9, 2 (in Japanese). Lawson, C. L. and R. J. Hanson, Solving Least Squares Probles, 34 pp., Prentice-Hall, New Jersey, Nakaichi, S. and H. Kawase, Broadband strong otion siulation in Fukuoka City based on a three-diensional basin structure and a hybrid ethod, J. Struct. Constr. Eng., AIJ, 6, 83 91, 22 (in Japanese with English abstract). Nakata, T. and T. Iaizui (eds.), Digital Active Fault Map of Japan, 6 pp. with 2 DVD-ROM, University of Tokyo Press, Tokyo, 22 (in Japanese). Nishiae, Y., Observation of seisic intensity and strong ground otion by Japan Meteorological Agency and local governents in Japan, J. Jpn. Assoc. Earthq. Eng., 4, 3 (Special Issue), 7 78, 24. Oniki, F., Precise location and subsurface features of the Kego fault in Fukuoka-city, north Kyushu, Japan, Active Fault Research, 1, 37 4, 1996 (in Japanese with English abstract). Pitarka, A., 3D elastic finite-difference odeling of seisic otion using staggered grids with nonunifor spacing, Bull. Seis. Soc. A., 89, 4 68, Pitarka, A., K. Irikura, T. Iwata, and H. Sekiguchi, Three-diensional siulation of the near-fault ground otion for the 199 Hyogo-ken Nanbu (Kobe), Japan, earthquake, Bull. Seis. Soc. A., 88, , Sekiguchi, H., S. Aoi, R. Honda, N. Morikawa, T. Kunugi, and H. Fujiwara, Rupture process of the 2 West Off Fukuoka Prefecture earthquake obtained fro strong otion data of K-NET and KiK-net, Earth Planets Space, 8, this issue, 37 43, 26. Sekiguchi, H., K. Irikura, and T. Iwata, Fault geoetry in the rupture terination of the 199 Hyogo-ken Nanbu earthquake, Bull. Seis. Soc. A., 9, , 2. Seno, T., Syntheses of the regional stress fields of the Japanese islands, Island Arc, 8, 66 79, Takenaka, H., T. Nakaura, Y. Yaaoto, G. Toyokuni, and H. Kawase, Precise location of the fault plane and the onset of the ain rupture of the 2 West Off Fukuoka Prefecture earthquake, Earth Planets Space, 8, this issue, 7 8, 26. The Research Group for Active Faults of Japan, Active Faults in Japan: Sheet Maps and Inventories (Revised Edition), 437 pp., University of Tokyo Press, Tokyo, 1991 (in Japanese with English suary). Wessel, P. and W. H. F. Sith, New, iproved version of Generic Mapping Tools released, Eos Trans. A. Geophys. Union, 79, 79, K. Asano (e-ail: k-asano@egdpri1.dpri.kyoto-u.ac.jp) and T. Iwata