GEOLOGICAL CHARACTERISTICS AND PROBLEMS IN AND AROUND OSAKA BASIN AS A BASIS FOR ASSESSMENT OF SEISMIC HAZARDS

SPECIAL ISSUE OF SOILS AND FOUNDATIONS 15-28, Jan. 1996 Japanese Geotechnical Society GEOLOGICAL CHARACTERISTICS AND PROBLEMS IN AND AROUND OSAKA BASIN AS A BASIS FOR ASSESSMENT OF SEISMIC HAZARDS KOICHI NAKAGAWAD, KIYOJI SHION0ii), NAOTO INOUEiii) and MASATO SANOiv) ABSTRACT A number of civil engineering structures in a large area Kobe-Hanshin and Awaji in southwest Kinki, southwest Japan, were seriously damaged by the 1995 Hyogoken-Nambu earthquake. Geological studies were made to reveal both geological and geotechnical characteristics in and around the Osaka sedimentary basin. A geological map of the Osaka Plain and the surrounding area was compiled for the geotechnical engineering use. A Bouguer gravity anomaly map is presented with additional gravity measurements for an evaluation of the subsurface geologic structure. Data on the depth of the basement rocks was collected from information on deep drilling for the geologic survey or hot springs, reflection seismic exploration and surface geological survey. A configuration map of the basement rock forming the bottom of the Osaka basin was also compiled from the depth data collected and an analysis of the correlation between the depth and the Bouguer anomaly was attempted to outline the relief image. In order to assess the contribution of underground structure with faults to the distribution of structural damage, the focusing process of seismic wave was examined by using the ray theory applied to model the Uemachi Fault zone, which is located in a dense urban region in the Osaka area. As a result, the efficiency of the focusing based on a common underground structure with fault was substantiated qualitatively. Key words: basement structure, focusing, geology of Osaka Basin, gravity anomaly, Hyogo-ken Nambu earthquake, underground structure, velocity structure (IGC: D7/E8) INTRODUCTION The 1995 Hyogoken-Nambu earthquake with the Japan Meteorological Agency, JMA, magnitude of 7.2 struck the Kobe-Hanshin area and Awaji Island, and seriously damaged a great number of civil engineering structures and other facilities. Geological factors, in general, control much structural damage, resulted from subsoil behavior during an earthquake. As a significant feature of the distribution characteristics of the damage, it can be pointed out that the most seriously damaged structures were concentrated in soft ground such as fill areas, especially reclaimed fills and/ or near an active fault zone (Nakagawa et al., 1995a). Figure 1 represents the distribution of active faults, seismic intensity of JMA scale seven, locations of soil liquefaction and seriously affected zones in and around the Osaka area. Highly urban zones in the Kobe and Hanshin areas, for which the subsurface profile is characterized by the soft ground formed in the younger structural basin commonly seen in most large cities in Japan, were hit directly and destroyed totally by the earthquake. A number of civil engineering structures were especially damaged by soil liquefaction and lateral displacement at the fill ponds, valleys and river channels as well as in man-made lands along the coastal zone (Mitamura et al., 1995). In addition, severe structural damage due to strong ground motion was observed in the narrow zone along known active faults such as the Nojima Fault, Yokooyama Fault, Ashiya Fault, Koyo Fault, Arima-Takatsuki Tectonic Line etc. and their extensions or along unknown buried faults (e.g., Seo, 1995; Hirano and Hada, 1995 and so on). There are some geological or geotechnical publications describing those areas which comprise the damaged area (e.g., Kasama and Kishimoto, 1974; Kobe Municipal Office, 1980; Huzita and Kasama, 1983; Nakaseko et al., 1984; Kansai Br. of Japanese Society of Soil Mechanics and Foundation Engineering, JSSMFE Professor, Faculty of Science, Osaka City University, Sugimoto 3-3-138, Sumiyoshi-ku, Osaka 558. Associate Professor, ditto. Graduate student, ditto. Research Engineer, Department of Geology, Osaka Office, Suncoh Consultant Co. Ltd., Nishinakajima 6-6-1, Yodogawa-ku, Osaka 532. Manuscript was received for review on October 17, 1995. Written discussions on this paper should be submitted before August 1, 1996, to the Japanese Geotechnical Society, Sugayama Bldg. 4F, Kanda Awaji-cho 2-23, Chiyoda-ku, Tokyo 101, Japan. Upon request the closing date may be extended one month. 15

16 NAKAGAWA ET AL. Fig. 1. Distribution map of active fault, seismic intensity of JMA scale 7, liquefaction and so on. Legend; A: zone of intensity VII in JMA scale, B: liquefied lateral displacement locations, C: zone of seriously damaged structures compared with surrounding area, D: active fault, E: epicenter of the 1995 Hyogoken-Nambu earthquake. 1: Baba Fault, 2: Minoo Fault, 3: Satsukiyama Fault, 4: Jyumantsuji Fault, 5: Nakayama Fault, 6: Najio Fault, 7: Rokko Fault, 8: Arino-Ougo Fault, 9: Yamada Fault, 10: Yubunedani Fault, 11: Ibayama Fault, 12: Kogoyama Fault, 13: Kashiodani Fault, 14: Egeyama Fault, 15: Sakamoto Fault, 16: Konoyama Fault, 17: Uchihata Fault, 18: Wakagashi Fault, 19: Kusumoto Fault, 20: Higashiura Fault, 21: Asano Fault, 22: Ikunami Fault, 23: Shizuki Fault and Kansai Geotechnical Consultants Association, 1987; The Research Group for Active Faults in Japan, 1991; Kansai Br. of JSSMFE et al., 1992; Itihara, 1993; Adachi et al., 1994; etc.). The amount of soil/rock boring data, however, is not adequate to able to define and identify each laver in the younger sedimentary cover. Very little in- on the deep geological structure is available in area as the present fine. In the Kansai Area, the Research Committee on the Application and Technology for Under-ground Space which selected Professor Kenzo Toki of Kyoto University as its chairman was established. Many types of geological and geotechnical information have been collected and a comprehensive data base of the subsurface geological information for the Kansai Area has been prepared (Kansai Br. of JSSMFE et al., 1992, Kansai Br. of Japanese Society of Soil Givil Engineerings JSCE, 1993). The committee completed its original purpose in 1993. The administration of the data base has been taken over by a new committee, Geo-Database Information Committee in Kansai (GeoDICK) which selected Mr. Tomomitu Fujii of Kinki regional construction Bureau, Japanese Ministry of Construction as the chairman and Professor Toshihisa Adachi of Kyoto University as the head of the geotechnical research section which has been investigating some geological/ geotechnical characteristics of the Kansai subsurface profile related to the disasters. Such subsurface information is a basic input required for careful land-use planning of an urban area. In addition, the Committee on Earthquake Observation and Research in the Kansai Area (CEORKA) which selected Prof. K. Toki as its chairman has also been organized by many members of public corporations, local government bodies, universities and so on. Many valuable data for the present earthquake were obtained by means of the wide network measuring system established with the aim of studying the so-called real time seismology (Toki et al., 1995; Irikura et al., 1995; etc.). Many information facts mentioned above can be utilized to evaluate important factors that influence earthquake hazard in highly urban areas in Japan. In the present study, the geologic structure of the Osaka sedimentary basin was investigated based or geotechnical considerations. GEOLOGICAL SETTING General Geology Under the Large Cities in Japan Most of the cities in Japan have suffered from large earthquakes along the inland active faults or the tectonic plate boundary. Even when the magnitude is not so great, an earthquake occurring immediately under a city may result in very serious damage. Many geological surveys and geophysical explorations are required to reveal the geology, distribution of buried active faults and dy-

GEOLOGICAL CHARACTERISTICS ON SEISMIC HAZARDS 17 Fig. 2. Schematic diagram showing a typical ground section for cities in the alluvial plain in Japan considering earthquake hazard namic material properties of subsurface materials under several main cities in Japan. Since many big cities in Japan have been developed on lowland which is underlain by thick alluvial deposits, geotechnical properties of the subsoils are essentially similar. In or around each of these alluvial plains many active faults exist and often large earthquake occur along them such as last January in Kobe. Not only some cities of Japan but also those of the other countries on the island arc or on the lowland set up at orogenic belt which can be represented by the Circum Pacific Rim are situated on such subsoils which have very similar geological setting. These areas have experienced large earthquakes associated with thrusting, strike-slip or normal faulting on the tectonic plate boundary, and large earthquakes along the active faults in the inland region as illustrated in Fig. 2. Widespread distribution of alluvial deposits in inland areas is thought to result from Quaternary crustal movement as well as the recent worldwide sea level change. The tectonic movements have resulted in the formation of sedimentary basins at many locations on the Japanese island arc since the Neogene period. Accordingly, the sedimentary cover including alluvial beds has been subjected to deformation by basement movements. In other words, the Pleistocene overlain by thick Holocene deposits has accumulated in the sedimentary basin which has been formed as a result of neo-tectonic movement and has been divided into smaller blocks by active faults. The Holocene deposits consist mainly of the sediments deposited during sea level rise, which corresponds to the event called the Frandrian transgression, 7000 to 4000 years ago. The large alluvial plains associated with the thick alluvium have generally been formed in the tectonic sedimentary basin which has existed in a regional neo-tectonic belt since Pliocene or Pleistocene time. Accordingly, the subsurface materials in many of the Japanese big cities have the same or similar geological and geotechnical characteristics. The Holocene alluvial deposits are partially composed of very soft material with a high water content, which are susceptible to a large decrease in bulk volume and also easily subjected to heavy damage from an earthquake. Serious seismic damage problems may be associated with these deposits because of their low shear strength and high compressibility. From the view point of earthquake engineering, geo-technically important factors include geomorphological features, subsurface geological structure, configuration of basement rock existing at depth, distribution of active faults, sensitive soil, loose sand, fill, surface groundwater and so on. As a loose saturated sandy layer existing near the ground surface is susceptible to liquefaction and a compressive clayey layer is susceptible to land subsidence by overburden pressure, appropriate measurements of ground water level play a very important role in evaluating the seismic stability of the subsoil profile. Eventually, the Holocene alluvial plains underlying the major Japanese cities have involved both seismically more active fields and seismically weaker passive fields as shown in Fig. 3. Fig. 3. Plate boundaries around the Japanese archipelago. Contour line indicate the depth (km) of the Wadati-Benioff zone (seismic plane): ST: Sagami Trough, T: Tsuruga Bay, I: Ise Bay, Epicenter of the 1995 Hyogoken-Nambu earthquake

18 NAKAGAWA ET AL. Geology of Osaka Plain and Surrounding Area Awaji and Kobe-Hanshin provinces are located nearby the west vertex of the so-called "Kinki Triangle", defined by Huzita (1962) to be a triangle composed of three vertices, the northern one corresponding to Tsuruga Bay, facing the Sea of Japan, the southwestern one to Awaji Island and the southeastern one to Ise Bay, facing the Pacific Ocean. The Kinki Triangle block is tectonically located between the Southwest Japan block, affected by the subducting slab of the Philippine Sea Plate and the Central-Northeast Japan block affected by the subducting slab of the Pacific Plate. Accordingly, this triangular region is forced to deform in a very complex manner as is shown in Fig. 4. The Kinki Triangle is also characterized as one of the most highly dense zones of active faults within the Japanese Islands and consists of some subsiding, uplifting or tilting blocks bounded by numerous active faults, which run mainly in the NE-SW, N-S, and ENE- WSW directions (Huzita et al., 1973). The present crustal stress conditions in this region are of the uniform stress field with E-W trending compression stress axis, which has been determined from the focal solution of many micro-earthquakes or the technique of stress release by means of over-coring of the basement rocks. The tectonic stress in the field like this region may therefore produce the strike-slip faults in the NE-SW direction, reverse faults or thrust in the N-S direction and normal faults in the E-W and so on. The Osaka Basin topographically is a typical sedimentary basin, in which an alluvial plain facing Osaka Bay on the west is widely developed, and the edge is bounded by uplifting mountain ranges, which are Mts. Rokko, Hokusetsu, Ikoma, Kongo, Izumi-Katsuragi and Awaji Island. Their boundaries correspond to one of the three fault types mentioned above. In the lowland area a thick soil deposit has been accumulated since late Tertiary time covers the basement rocks which are composed of some granitic rocks or pre-tertiary formations. These base- rocks are widely exposed in the surrounding mountainous regions. The geological map of this area is shown in Fig. 5. The map was drawn by compiling many investigations such as Geographical Survey Institute (1966), Huzita (1966), Sano (1980), Huzita and Kasama (1982), Geographical Survey Institute (1983), Huzita and Kasa- Fig. 4. Cross-section of the underground structure across the Kinki Area, Southwest Japan ma (1983), Calamity Science Institute (1984), Huzita and Maeda (1984), Huzita and Maeda (1985), Itihara et al., (1986), Kansai Br. of JSSMFE and Kansai geotechnical Consultants (1987), Mizuno et al., (1990), Itihara (1993), The Research Group for Active Faults of Japan (1991), Takahashi et al., (1992), Iwasaki et al., (1992), Itihara (1993), Iwabuchi (1995), Kansai Br. of JSSMFE (1995). The Osaka Group and the upper Pleistocene lie below the lowland area. Distinctive landform features of terraces and hills, where the Osaka Group and the upper Pleistocene are exposed sporadically, have been developed between the lowland and the mountainous region. These geological structures and morphological features have resulted mainly from the latest Cenozoic crustal movements, which have a strong relationship to the process of the present Hanshin Awaji Great earthquake disaster. The basement rocks consist of mainly the Ryoke granitic rocks and partially of Paleozoic, Mesozoic and Miocene sedimentary rocks and some volcanic rocks. The sedimentary cover overlying the basement rock consist of the Osaka Group, the upper Pleistocene and Holocene alluvial deposits. The Osaka Group is divided into three sub-groups, which are the lower, middle and upper sub-groups. The stratigraphic features and Ethology are shown in Table 1. The geologic column (OD-1) at the type locality of Osaka Group is shown in Fig. 6. The borehole location is south of the Yodo River and near the coast of Osaka Bay in Osaka City (Fig. 6). The bottom of the borehole with 900 m in depth did not reach the basement rock. The OD-2 drill hole 650 m deep at the Uemachi uplift, which is thought to have been lifting up since middle Pleistocene, did reach the basement. Data on the depth to the buried basement rocks have been revealed by means of geophysical exploration and drilling of a hot spring (Nakagawa, 1990; Itihara, 1991). Some buried active faults under the alluvial plain have been identified by seismic reflection technique in Osaka Bay (Iwasaki et al. 1994) and the inland area (Yoshikawa et al., 1987 a; Yoshikawa et al., 1987 b; Nakagawa et al., 1994; Kawasaki et al., 1994). Seismic Wave Velocities in the Osaka Sedimentary Basin Recently several types of geophysical exploration have been performed in the Osaka Plain by measuring gravity, reflected or refracted seismic wave, ground microtremor, magneto-telluric surreys, etc. The Bouguer gravity anomaly in the Osaka Plain will be described in the next section. Some seismic reflection explorations were performed in the northern part across the Uemachi Fault, which runs under the center of Osaka City with approximate strike of N-S direction (Yamamoto et al., 1992; Nakagawa et al., 1993). A mass dropping method was applied to generate the P-wave, and a large power S-vibrator was also used to generate the S-wave. The following results were obtained. The throw of the fault is about 700 meters. The depth of the basement on the west side of the fault is about 1500 meters, which coincides with the result of drilling for a hot spring. The velocity of P-wave

GEOLOGICAL CHARACTERISTICS ON SEISMIC HAZARDS 19 changes from 1500 m/ sec near the ground surface to 2400 m/ sec near the bottom of the sedimentary cover. The velocity of S-wave changes from 300 to 1000 m/ sec with depth. Figure 7 shows the depth profile of the S- wave reflection. The P-wave profile was interpreted as shown in Fig. 8. Letters A, B and C in the figure correspond of the upper Pleistocene, the upper and middle sub-groups, and the lower sub-group of the Osaka group respectively. Figure 9 shows the relationship between wave velocity and the depth. As compressional wave transmitted in saturated subsurface layer has a rather low damping characteristic, the A B Fig. 5. Geological Map of the Osaka Plain and geological profile across Osaka Bay. The lithology and stratigraphy are described in Table 1. The name of the Fault is shown in Fig. 1. The location of the profile is shown on the geological map. The intersection of two lines showing profile location corresponds to the epicenter of the 1995 Hyogoken-Nambu earthquake. Note that the vertical scale is expanded to three times the horizontal scale

20 NAKAGAWA ET AL. Table 1. Stratigraphy in and around Osaka area Fig. 7. S-wave reflection profile across the Uemachi Fault along the Yodo River (Nakagawa et al. 1993) Fig. 8. P-wave reflection profile and interpretation crossing the Uemachi Fault along the Yodo River (Yamamoto et al., 1992), V, and Vs represent compressional and shear velocities in km/sec, respectively Fig. 6. Schematic geological column diagram for the Osaka group based on actual drilling location, OD-1. The location is indicated in Fig. 16. No attached to letter Ma is the name of marine clay layer in ascending order Fig. 9. Relationship between seismic wave velocities and the depth from the reflection survey at the Yodo River

GEOLOGICAL CHARACTERISTICS ON SEISMIC HAZARDS 21 Fig. 10. Relationship between shear wave and compression wave velocities in the Osaka sedimentary basin obtained from different methods of measurement. Circle and triangle designate data from seismic reflection exploration at Yodo River, bar is from Vp in reflection and Vs in inversion of Rayleigh wave due to tremor and dot is averaged from P-S logging. See the text for equations corresponding to the solid and broken lines in this figure Fig. 11. The Bouguer Anomaly in Hanshin and Osaka area. Numerical values in figure show the anomaly with contour interval of 1 mgal attenuation of amplitude along the path length is smaller, which results in the acquisition of a high resolution image for the profile. Shear wave velocity structure below the ground surface is very important to analyze the dynamic response during an earthquake. Unfortunately, it is very difficult to perform the S-wave exploration to a large depth because a high powered S-wave generator is very expensive at the present time. Accordingly, some estimate regarding the distribution of S-wave velocity insitu is required. To that end, we can utilize the P-wave velocities obtained from the velocity analysis of P-wave reflective exploration, which can be performed relatively easily. Figure 10 represents the relationship between the P-wave velocity and S-wave velocity which were collected from the results of different measurements such as the seismic reflection methods mentioned above, inversion methods from dispersion of Rayleigh wave induced by microtremor(horike, 1985; Aoki et al., 1990) and P-S logging (Iwasaki, 1988) using the suspension technique (Kitsunezaki, 1980). The solid straight line and broken curved line in Fig. 10 are regression lines calculated from the plot by means of least square method as follows; where 17, and V, are shear wave velocity and compressional wave velocity represented in km/ sec, respectively. BASEMENT STRUCTURE OF THE OSAKA SEDIMENTARY BASIN One significant feature of the Hanshin-Awaji great earthquake disaster as mentioned above is that most of the damage to structures was concentrated in soft subsoil such as fill deposits, especially reclaimed fills and/ or near (a) (b) an active fault zone. Considering the zonal distribution characteristics of structural damage and ground failure due to this earthquake, the extensive collection of additional information about the underground structure is necessary to alleviate the problem. The Bouguer anomaly in and around the Osaka area has been compiled by Nakagawa et al. (1991). In addition, gravity around the Kobe and Hanshin areas was measured after the earthquake occurred. The gravity data at 4800 points was used to analyze the underground structure in the Osaka sedimentary basin, which is mainly comprised of Mesozoic granitic basement and post Miocene sedimentary layers covering the basement. Figure 11 shows the Bouguer anomaly in and around the Osaka Plain. In this gravity analysis the data collected and published by Professor Ryuuiti Shichi of Nagoya Univ. (Shichi and Yamamoto, 1994) is partially included. Basement configuration plays an important role in the concentration, dispersion and transmission of seismic rays. Figure 12 shows the relationship between the depth of basement and observed gravity anomaly. Accordingly, the basement structure of the Osaka sedimentary basin was estimated from the Bouguer anomaly. In general, regional gravity effects should be removed from the measured gravity values to know the relationship between the Bouguer anomaly and the depth of basements, it is possible to directly interpolate the depth of basement from the Bouguer anomaly by deducting the regional effect. In order to find the regional gravity effect in Osaka and the Hanshin area, the following relationship is assumed: where GB is the Bouguer anomaly. X and Y are distances from the origin (135 30'E, 34 40'N) measured eastwards (1)

22 NAKAGAWA ET AL. where R 2 is the correlation coefficient. The configuration of basement rock in the Osaka sedimentary basin is shown in Fig. 14 by means of contour lines. The relief image is shown in Fig. 15. There is a high possibility of the existence of faults in the zones where anomaly contour lines are concentrated. The present result based on Bouguel anomaly observation is consisted with information on the depth of base- Fig. 12. Relationship between depth to top of basement rock and the Bouguer Anomaly before removing the regional gravity effect and southwards, respectively, while Z is the depth of the 5ement rocks. The constants from A to L were determined by the least square method using the data (X, Y, Z, GB) at 22 points where the depth of the basement rocks were known or the basement rocks were outcropped. Values of constants are given in Table 2. The residual Bouguer anomaly (after the regional gravity effect has been removed from the observed anomaly using eq. (1)) shows a linear relationship with the depth of basement rocks as shown in Fig. 13. Eventually, the following relationship between the depth of basement Z and the corrected Bouguer anomaly GB was obtained: Fig. 14. Configuration of top surface of basement rock in the Osaka sedimentary basin. Numerical values in figure show the depth of basement rock with contour interval of 0.1 km Fig. 13. Relationship between the depth to the top of basement rock and the Bouguer Anomaly after removing the regional gravity effect Fig. 15. Relief picture of the imaged configuration of top surface of basement rock in the Osaka sedimentary basin Table 2. Value of constants A `L for the equation giving the regional effects of the Bouguer anomaly in the Osaka Basin and the surrounding area

GEOLOGICAL CHARACTERISTICS ON SEISMIC HAZARDS 23 ment rocks in both Hanshin and Osaka areas approximated from some kind of geological surveys. Furthermore, the subsurface structure must be examined in detail by applying Talwani's method to the residual Bouguer anomaly (Talwani et al., 1959). CONCENTRATIONS OF SEISMIC RAY DUE TO HORIZONTALLY HETEROGENEOUS SUBSURFACE STRUCTURE Structure Model Composed of Homogeneous Layers with Curved Interfaces As previously mentioned, many man made structures were damaged especially by liquefaction and associated lateral displacement of filled ponds, valleys and river channels as well as in man-made lands along the coastal zone. Severe structural damage due to strong ground motions was also observed in the narrow zone along known active faults and their extensions or along unknown buried faults. Even in the Osaka area 30-50 km east of the epicenter, many engineering structures were damaged by soil liquefaction and lateral displacement in reclaimed filled river channels and irrigation ponds in inland areas, especially along the old Yodo River. Further serious damage to many engineering structures due to strong ground motions was observed in the narrow zone along active faults as well as in reclaimed land along the coastal zone as shown in Fig. 16. The damage seems to be very Fig. 16. Distribution of damage to man-made structures in the Osaka area. A: Residential blocks where damage to man made structures was observed together with non-extensive subsurface failures. B: Blocks where extensive damage to man made structures was observed together with subsurface failures due to liquefaction, lateral displacement or slope instability. C: Active fault. D: Old coast line in 1885. E: Old coast line in the 1680s. F: Reclaimed river channel. G: Topographic contour line. H: Zone where damage was relatively larger than those in the adjacent area. Based on surveys by the Osaka City University Investigation Team on the Hanshin Great earthquake disaster.

24 NAKAGAWA ET AL. much linked to the focusing of seismic waves. The paths and amplitudes of seismic waves were examined on a velocity structure model composed of four homogeneous layers with curved interfaces defined by referring to the result of the seismic reflection survey. Fig. 8 in the previous section shows an interpretation of the reflection profile across the Uemachi Fault along the Yodo River. As the Uemachi Fault is about 45 km east from the epicenter and trends nearly in the north-south direction, it is reasonable to consider that seismic waves come from the west in a direction perpendicular to the trend of the fault. In order to examine the focusing of seismic rays, a two-dimensional structure model composed of four homogeneous layers is assumed with curved interfaces referring to the seismic reflection profile mentioned above. In order to examine the propagation of seismic waves, both shapes of sedimentary layers and propagation velocities in the layers must be determined with the high accuracy. The shape of the interface particularly an important role in the propagation of seismic It is difficult however to detect the detailed structure in the vicinity of the fault zone with the present technical level of seismic reflection experiments. Therefore referring to the deformation of the upper part of sedimentary layers, interfaces are simply approximated using an error function erf (ii) as follows: where Table 3. Model of underground structure across the Uemachi Fault It is noted that Ai is always positive because Ni is directed toward the (j + 1)-th layer. Figures 17(1) and (2) give examples of ray fields for P and S waves, respectively. They show that seismic rays tend to focus at the vicinity of the fault zone in the subsided block regardless of the directions of incident waves although the focusing site depends on the incident angle. Considering the distance from the epicenter, it may be assumed that seismic waves are incident on the basement rocks at about 40 from the west. Ray fields around the Uemachi Fault are approximated by Fig. 17(1) (c) for P waves and by Fig. 17(2) (c) for S waves. These figures represent that seismic rays focus strongly in a narrow zone just west of the fault zone to cause intense ground motions. The result shows good agreement with the distribution of damages. In addition, it can be seen from the figures that arrivals of seismic rays are very limited in a wide zone above the fault zone, indicating weak ground (a) (a) Table 3 gives values of P and S wave velocity in layers and parameters defining functions for three interface. Ray Tracing in a Model Composed of Homogeneous Layers with Curved Interfaces It is sufficient to consider the effect of refraction at the interface in the present model since a seismic ray goes straight within a layer of a constant velocity and changes its direction at the interface. Let ni be a unit vector parallel to the ray within the j-th layer. When a ray is transmitted from the j-th layer of velocity v./ into the (j + 1)-th layer of velocity vj+1 through an interface, according to Cerveny et al. (1977)'s formulation, the relation between a vector nj for the incident wave and ni+i for the transmitted wave is expressed by (b) (c) (b) (c) where Mj= vi+i/ vj, /Vj is a unit vector normal to the j-th interface and directed toward the (j + 1)-th layer, and (1) Fig. 17. Refraction pattern of seismic rays for P (1) and S (2) waves incident at equal intervals on the bottom of the model of homogeneous layers separated by curved interfaces. Parameters of velocity structure are given in (a). (a): Incident angle (5=0'; (b): (2)

GEOLOGICAL CHARACTERISTICS ON SEISMIC HAZARDS 25 motions. This explains partly observations that damage is slight in the Uemachi Uplift east of the fault, although this slight damage is mainly due to the effect of the good subsoil conditions; the sedimentary layers in the Uemachi Uplift are thin and stiff relative to those in the western subsided block. It is probable that the focusing of seismic waves also explains relatively high damage in the area south of the Itami Fault. The effect of focusing however fails to explain damage in the zone south of the Nobata Fault because the zone lies in the uplifted block. (b) Quantitative Estimation of Ray Density and Amplitude In order to evaluate the effect of focusing quantitatively, the amplitude of SH wave incident on the basement rock was calculated as plane waves considering only two effects of the transmission coefficient at the interface and the geometrical displacement, based on Cerveny et al. (a) (1977)'s ray method. When SH waves are incident at an angle ľj on an interface from a medium of density p; and velocity Ĉj and are transmitted at an angle Of into a medium of density pi+i and velocity M+ 1, the transmission coefficient R; is calculated by The total effect of transmission coefficient becomes Fig. 18. Quantitative estimate- of focus of seismic rays for SH waves incident at an angle of 40 on the bottom of the model: (a) where m is the number of interfaces. The effect of geometrical spreading is evaluated by Refraction pattern of seismic rays. (b) Density of rays approximated by numbers of rays in 75 small intervals among 4000 rays which arrive at the surface within a horizontal distance from 0 m to 4000 m; (c) Normalized amplitude of 4000 rays at the surface on a log scale where O is the angle of incidence on the j-th interface, 0; is the angle of transmission. The factor F representing the change in a cross-sectional area of a ray tube is approximated by where A x represents the difference of x-coordinates of the points where two neighboring rays impinge at the bottom of the model and A x' represent the difference of x- coordinates of the points where two rays arrive at the ground surface, and 6 gives an angle of incidence on the surface. Figure 18 represents one example of these calculations. Figure 18(a) shows the refraction pattern of S waves incident on the basement from the west at angle of 40. Figure 18(b) shows the density of rays arriving at the surface. The density is approximated by the number of rays in 75 small intervals among 4000 rays of SH waves which arrive at the surface within a horizontal distance from 0 m to 4000 m. It is emphasized in Fig. 8(b) again that an extremely high concentration of rays occurs at the narrow zone just west of the fault zone but the arrival is limited in a wide zone above the fault zone. The amplitude of SH wave at the ground surface is estimated by computing Q and L for 4000 rays of SH waves arriving within a horizontal range from 0 m to 4000 m. As Q and L give the same values in the area where interfaces are parallel to each other on both sides of the fault zone, the amplitude of each ray is normalized by that of the ray through horizontal layers in both sides of the fault zone. Figure 18(c) gives the normalized amplitude for each individual ray at the surface on a log scale. It should be noted that we discuss here only the effect of focusing of seismic rays but we do not consider the amplification due to the multiple-reflection in the calculation of the amplitude. It is expected based on Fig. 18(c) that the amplitude increases by over one order of magnitude at the very narrow focusing location just west of the fault zone and decreases by about one tenth in a wide zone above the fault zone. The above calculation however seems to overestimate ground motions when we consider an assumption of the ray theory that the wave length is considerably smaller than any characteristic quantity of length dimension such as radius of curvature of the interface and thickness of the layer. Because the wave length, say ten meters long even for 10 Hz wave, is not negligible in the present model, we should consider that contrast of the amplitude is not so extremely large as Fig. 18(c) shows. Further as shown in Fig. 18(a), it is very

26 NAKAGAWA ET AL. (a) (b) (c) (d) graben structure, the fault scarps of both sides are especially nearly vertical, very intensive concentration of the upward seismic rays inner zone of the trough. A similar graben structure is seen in the zone along Arima-Takatsuki Tectonic Line (Sangawa, 1978; Kawasaki et al., 1994). The trend of the fault system is ENE-WSW and seems to be a normal fault essentially from the regional tectonic stress field. This suggests the possibility that focusing of seismic ray due to depression of the basement formed by fault movements explains extensive damage observed in the narrow zone along the Itami and Nobata Faults shown in Fig. 16 and also severe damage zone represented by intensity VII (JMA scale) in Takarazuka City and damage zone in Kawanishi City along the Arima-Takatsuki Tectonic Line. In addition, regarding the severest structural damage zone with intensity of VII in the Kobe-Hanshin area as shown in Fig. 1, as this strip zone corresponds to the most concentrated zone of the isobath lines of the basement rocks (see Fig. 14), the focusing effect mentioned above the damages at this area can be made. Fig. 19. Examples of seismic ray trace depending on basement structure with incident angle 40 degree into basement: (a) and (c): horst type, (b) and (d): trough type; Velocity ratio = Vcover/ Vbasement = 0.5 likely that the cross-sectional areas of ray tubes become zero for some rays arriving at the focus spot, that is, the caustics are generated. An extremely high peak of amplitude shown in Fig. 19 reflects the caustic. As assumptions of the ray theory break down around the caustic, the extremely high amplitude due to caustics should not be in- to represent the real ground motion. order to simulate ground motions exactly, the frequency-dependent wave fields within an inhomogeneous media must be investigated by the perturbation method and/ or the finite difference /finite element methods to solve equations of dynamic elasticity (see Aki and Richards, 1980). Our approximated procedure based on a simple ray theory however suggests the possibility that focusing of a seismic ray generates the narrow zone of intense ground motions in the vicinity of the abrupt change in depth of substrata around the fault zone and ground motions become relatively weak in the uplifted block. This explains observations that damage of buildings and houses are extensive in several narrow zones along the Uemachi Fault and Butsunenjiyama Fault but weak in the areas east of the faults. As is shown in the figure, seismic rays incident from the east also tend to focus in the subsided block. It is expected therefore that the strong focusing occurs within a narrow trough filled with soft sediments. Some examples of the displaced basement block with faulting are demonstrated in Fig. 19. In case of CONCLUSIONS As a result of the geological investigation and geophysical consideration of the Osaka sedimentary basin and surrounding area regarding the Hanshin-Awaji great earthquake disaster the following conclusions can be drawn. The geological map and sections for geotechnical engineering use of the Osaka sedimentary basin and surrounding area were drawn by compiling the previous studies and new information after the earthquake occurred. The seismological section across the Uemachi Fault was described based on the results of reflective seismic exploration and drilling for either a geologic survey or hot spring. A relationship between shear and compressional wave velocities in the Osaka basin is generalized from the data of the velocity analysis of seismic reflection wave, the inversion of Rayleigh wave dispersion indicated by microtremor. The equation is expressed as where V, and V, are shear and compressional wave velocities in km/ sec, respectively. The Bouguer anomaly map for this area was drawn by compiling previous studies and results of additional measurement in the Kobe-Hanshin area. The regional effects were evaluated by comparison with actual depth of the basement at 22 and removed from the Bouguer anomaly to interpolate the depth over the area. Consequently, the isobath map and relief image diagram of the basement were pictured. The focusing effects of seismic rays due to horizontally heterogeneous structure were examined by using the ray theory and a model as the underground structure of the Uemachi Fault. The results indicate the concentration of seismic rays at western part near the fault and explains the actual higher damaged zone along the Uemachi-But-

GEOLOGICAL CHARACTERISTICS ON SEISMIC HAZARDS 27 sunenjiyama Fault. In addition, the underground structure showing depression of seaward basement block at the heavy damaged zone with intensity VII in Kobe-Hanshin area is suggested to be one of the important causes of serious structural damage. ACKNOWLEDGMENTS We wish to thank members of the Osaka City University investigation team on the Hanshin Great earthquake disaster, particularly, Drs. S. Masumoto, M. Mitamura, Miss. N. Kitada, Mrs. K. Ryoki and S. Senda for many helpful discussions and considerable assistance and we also wish to thank Dr. R. Shichi of Nagoya University for his help to use in part the database on the gravity. The authors are also grateful to the reviewer of this paper for helpful suggestions. REFERENCES 1) Adachi T., Nakagawa K., Suwa S. and Yamamoto K. (1994): "Underground structure and geological characteristics on deep grounds in Kansai Area," in "Application and technology on the underground space in Kansai Area", Kansai Br., Japan Soc. Civil Eng., pp. 35-66 (in Japanese). 2) Aki, K. and Richards, P. G. (1980): Quantitative Seismology II, W. H. Freeman and Co., San Francisco. 3) Aoki, Y., Ozaki, M., Tsujimoto, A., Horike, M. and Takeuchi, Y. (1990): "Estimation of S-wave velocity structure in the Osaka Plain," Proc. Kinki Br., Arch. Inst. Japan, pp. 341-344 (in Japanese). 4) Calamity Science Institute (1984): Geological Survey of the Submarine Strata at the Kansai International Airport in Osaka Bay, Central Japan, Calamity Science Institute (in Japanese). 5) Cerveny, V., Molotkov, I. A. and Psencik, I. (1977): Ray Method in Seismology, Univerzita Karlova,- Praha. 6) Geographical Survey Institute (1966): "Land-use map (Tochijokenzu)," Part of Kobe Area (in Japanese). 7) Geographical Survey Institute (1983): "Land-use map (Tochijokenzu)," Part of Northwestern Osaka (in Japanese). 8) Hirano, M. and Hada, S. (1995): "Fault activity and displacement in western Kobe City caused by Hyogoken-Nambu earthquake," The Proceedings of Symposium on the Great Hanshin Awaji Earthquake and its Geo-environments, The Committee of Environmental Geology, Geological Society of Japan, pp. 74-84 (in Japanese with English abstract). 9) Horike, M. (1985): "Inversion of phase velocity of long-period microtremors to the s-wave velocity structure down to the basement in urbanized area," Jour. Phys. Earth., Vol. 33, pp. 59-96. 10) Huzita, K. (1962): "Tectonic development of the median zone (Setouti) of southwest Japan, since the miocene," Jour. Geosci., Osaka City Univ., Vol. 6, pp. 103-144. 11) Huzita, K. (1966): Tectonic Significance of Osaka Bay, on Papers in honor of Professor Susumu Matsushita upon his retirement, pp. 133-141 (in Japanese). 12) Huzita, K. and Maeda, Y. (1971): "Itami terrace with special reference to the Late Pleistocene transgression in Japan," Jour. Geosci., Osaka City Univ., Vol. 14, pp. 53-70. 13) Huzita, K., Kishimoto, Y. and Shiono, K. (1973): "Neotectonics and seismicity in the Kinki area, Southwest Japan," Jour. Geosci., Osaka City Univ., Vol. 16, pp. 93-124. 14) Huzita, K. and Kasama, T. (1982): Geology of the Osaka-Seihokubu District, Quadrangle Series, Scale 1:50,000, Geol. Survey, Japan (in Japanese with English abstract). 15) Huzita, K. and Kasama, T. (1983): Geology of the Kobe District, Quadrangle Series, Scale 1:50,000, Geol. Survey, Japan (in Japanese with English abstract). 16) Huzita, K. and Maeda, Y. (1984): Geology of the Suma district, Quadrangle Series, scale 1:50,000, Geol. Surv. Japan, (in Japanese with English abstract). 17) Huzita, K. and Maeda, Y. (1985): Geology of the Osaka-Seinambu district, Quadrangle Series,scale 1:50,000, Geol. Surv. Japan (in Japanese with English abstract). 18) Irikura, K. and Y. Fukushima (1995): "Attenuation characteristics of peak amplitude in the Hyogoken-Nambu earthquake determined by near-field strong motion modeling, Preliminary results," Jour. Natural Disas. Sci., Vol. 16(3), pp. 47-58. 19) Itihara, M., Ichikawa, K. and Yamada, N. (1986): "Geology of the Kishiwada district," With Geological Sheet Map at 1:50,000, Geol. Surv. Japan (in Japanese with English abstract). 20) Itihara, M. (1991): "Quaternary geological map of Osaka and surrounding area," Urban Kubota, No. 30 (in Japanese). 21) Itihara, M. (1993): "The Osaka Group," edited by M. Itihara, Sogensha Co. Ltd. (in Japanese with English abstract). 22) Iwabuchi, Y., Kasuga, S., Uchida, A., Kokuta, S., Okino, K., Shimura, E. and Nagata, S. (1995): "Faults and seafloor deformation near the source region of 1995 Hyogoken-Nambu earthquake," The Proceedings of Symposium on The Great Hanshin- Awaji Earthquake and its Geo-environments, The Committee of Environmental Geology, Geological Society of Japan, pp. 85-90 (in Japanese with English abstract). 23) Iwasaki, Y. (1988): "In-situ seismic velocity and its relationship with soil characteristics in Osaka Bay," Earthquake Engineering and Soil Dynamics II, Geotech. Special Pub. No. 20, ASCE, pp. 290-304. 24) Iwasaki, Y., Kagawa, T., Sawada, S., Matsuyama, N., Ohshima, K., Ikawa, T. and Onishi, M. (1994): "Basement structure by airgun reflection survey in Osaka Bay, Southwest Japan," ZISIN (Jour. Seism. Soc. Japan), Vol. 46, pp. 395-403 (in Japanese with English abstract). 25) Iwasaki, Y., Matsuyama, N., Kagawa, T. and Sawada, S. (1992): "Geological structure and developing process of Osaka Basin," Symposium on the Geologic structure and the state of the arts on soil characteristics of the Kansai Ground, pp. 119-126 (in Japanese). 26) Jackson, P. S. (1970): "The focusing of earthquakes," Bull. Seism. Soc. Am., Vol. 61, pp. 685-695. 27) Kansai Br., JSCE (1994): "Application and technology on the underground space in the Kansai Area," Text of the technical course, Kansai Br., JSCE (in Japanese). 28) Kansai Br., JSSMFE and Kansai Geotech. Consultants Assoc. (1987): The Ground of Osaka (Revised edition), Kansai Br., JSSMFE (in Japanese). 29) Kansai Br., JSSMFE and Committee on application and technology on the underground space in the Kansai Area (1992): The Ground of Kansai (Kansai Jiban), Kansai Br., JSSMFE (in Japanese). 30) Kansai Br., JSSMFE (ed) (1995): The submarine strata (Kaitei Jiban) Kansai Br., JSSMFE (in Japanese). 31) Kasama T., Kishimoto (1974): Kobe and the Earthquake, Kobe Municipal Office, (in Japanese). 32) Kawasaki, S., Toda, S., Nakagawa K. and Kobayashi, Y. (1994): "Seismic reflection exploration in the eastern part of the Arima- Takatsuki Tectonic Line, Kinki District, Central Japan," ZISIN (Jour. Seism. Soc. Japan), Vol. 47, pp. 173-182 (in Japanese with English abstract). 33) Kitsunezaki, C. (1980): "A new method for shear wave logging," Geophysics, Vol. 45, pp. 1489-1506. 34) Kobe Municipal Office (1980): The Ground of Kobe, Kobe Municipal Office (in Japanese). 35) Mitamura, M. and Osaka City University Investigation Team of the Hanshin Great Earthquake Disaster (1995): "Ground failures due to the 1995 Hyogo-ken Nambu earthquake and artificial alteration of topography," The Proc. Symp. on the Great Hanshin-Awaji Earthquake and its Geo-environments, The Committee of Environmental Geology, Geological Society of Japan, pp. 209-214 (in

28 NAKAGAWA ET AL. Japanese with English abstract). 36) Mizuno, K., Hattori, H., Sangawa, A. and Takahashi, Y. (1990): "Geology of the Akashi district." With Geological Sheet Map at 1:50,000, Geol. Surv. Japan (in Japanese with English abstract). 37) Nakagawa, K. (1990): "Ground survey at large depth in Kansai Area, Southwest Japan. In special reference to Osaka Ground," Proc., Symp., Ground characteristics at the depths in Kansai Area, Kansai Br., JSSMFE, pp. 63-88 (in Japanese). 38) Nakagawa, K., Mitamura, M., Harada, T., Shirakata, K. and Yokota, H. (1993): "Shear-wave vibroseis survey of deep structures in Osaka basin," Proc. 88-th SEGJ Conf., Soc. Expl. Geophys. Japan, pp. 418-423 (in Japanese). 39) Nakagawa, K., Mitamura, M., Uda, H., Uno, K., Toda S., Kawasaki S. and Yokota H. (1994): "Seismic reflection survey on north part of Ikoma Fault, east end of Osaka Plain," Proc. 90th SEGJ Conf., pp. 201-205 (in Japanese with English abstract). 40) Nakagawa, K. and Osaka City University Investigation Team on the Hanshin Great Earthquake Disaster (1995a): "Earthquake ground failure in and arround Osaka area and the subsurface structure," Proc. of Urgent Joint Meeting on the Great Hanshin Eearthquake, Japan Soc. Geol., Japan Soc. Eng. Geol., Fault Res. Data Center and OCU. Invest. Team on Great Hanshin Earthq. pp. 92-108 (in Japanese). Nakagawa, K. and Osaka City University Investigation Team on the Hanshin Great Earthquake Disaster (1995b): "A relation between earthquake ground failure and structure," Proc. Symp., on the Great Hanshin-Awaji Earthquake and its Geo-environments, the Committee of Environmental Geology, Geological Society of Japan, pp. 233-238 (in Japanese with English abstract). 42) Nakagawa, K., Ryouki, K., Muto, N., Nishimura, S. and Ito, K. (1991): "Gravity anomaly map and inferred basement structure in Osaka Plain, Central Kinki, Southwest Japan," Jour. Geosci. Osaka City Univ., Vol. 34, pp. 103-117. 43) Nakaseko, K., Takemura, K., Nishiwaki, N., Nakagawa, Y. and Yamauchi, M. (1984): "Stratigraphy of the submarine strata at the Kansai International Airport in Osaka Bay off Senshu, Central Japan," Geological survey of the submarine strata at the Kansai International Airport in Osaka Bay, Central Japan, Rept., Calamity Sci. Inst., pp. 191-198 (in Japanese). 44) Sangawa, A. (1978): "Fault topography and Quaternary faulting along the middle and eastern parts of the Arima-Takatsuki Tectonic Line, Kinki district, Central Japan," Geographical Review of Japan, 51-10, pp. 760-775 (in Japanese with English abstract). 45) Sano, M. (1980): "Geological structure of the Northern Nara Basin," Guidebook on Land-developments and geotechnical problems of the region from Nara to Kyoto Region, Kansai Br., Japan Soc. Engin. Geol., pp. 13-25 (in Japanese). 46) Seo, M. (1995): "Damage to ground along the Arima-Takatsuki Tectonic Line," Proc., Urgent Joint Meeting on Hanshin Great Earthquake Disaster. Japan Soc., Geol., Japan Soc. Eng. Geol., Kansai Geotech. Cons. Assoc., Fault Res. Data center and Osaka City Univ. on Invest. Team, Hanshin Great Earthq., pp. 56-61. 47) Shichi, R. and Yamamoto, A. (1994): "A gravity database of Southwestern Japan compiled at Nagoya University," Rept. Geol. Survey, Japan, No. 280, pp. 1-28 (in Japanese with English abstract). 48) Takahashi, Y., Sangawa, A., Mizuno, K. and Hattori, H. (1992): "Geology of the Sumoto district," With Geological Sheet Map at 1:50,000, Geol. Surv. Japan (in Japanese with English abstract). 49) Talwani, M., Worzel, J. L. and Landisman, M. (1959): "Rapid gravity computation for two dimensional bodies with application to the Mendosino submarine fracture zone," Jour. Geophys. Res., Vol. 64, pp. 49-59. 50) The Research Group for Active Faults of Japan (1991): Active Faults in Japan. Sheet Maps and Inventories (Revised Edition), Univ. of Tokyo Press, Tokyo (in Japanese with English Summary). 51) Toki, K., Irikura, K. and Kagawa, T. (1995): "Strong motion records in the source area of the Hyogoken-nambu earthquake, January 17,1995, Japan," Jour. Natural Disas. Sci., 16(2), pp. 23-30. 52) Yamamoto, E., Nakagawa, K., Mitamura, M., Toda, S., Nishida, T., Terada, Y., Uda, H., and Yokota, H. (1992): "Seismic reflection method at the Central Osaka Plain Yodo River (Juso `Kunijima) Line-," Proc. 1992's Annual Meeting, Kansai Branch, Japan Soc. Eng. Geol., pp. 185-188 (in Japanese). 53) Yoshikawa, S., Iwasaki, Y., Ikawa, T., Kuroda, T. and Yokota, H. (1987): "Reflection seismic survey of the Median Tectonic Line in Wakayama Prefecture, Japan," Proc., 77th SEGJ Conf., pp. 46-49 (in Japanese with English abstract). 54) Yoshikawa, S., Machida, Y., Teramoto, M., Yokota, H., Nagao, H. and Kagiwara, M., (1987): "Reflection seismic survey in Osaka City," Proc. 77th SEGJ Conf., pp. 114-119 (in Japanese with English abstract).