NIIGATA-KEN CHUETSU EARTHQUAKE OF OCTOBER 23, 2004

Transcription

1 Dedicated to the people of Chuetsu Region of Niigata -ken, who lost their lives, injured or suffered from this earthquake A RECONNAISANCE REPORT ON NIIGATA-KEN CHUETSU EARTHQUAKE OF OCTOBER 23, 2004 Ömer AYDAN Tokai University Department of Marine Civil Engineering Shizuoka, Japan November 2004

2 CONTENTS ABSTRACT 1 INTRODUCTION 2 GEOGRAPHY 3 GEOLOGY 4 TECTONICS 5 SEISMICITY 6 FAULTING MECHANISM AND OBSERVATIONS 7 STRONG MOTION RECORDS AND THEIR CHARACTERISTICS 8 GEOTECHNICAL DAMAGE 9 DAMAGE TO BUILDINGS 10 DAMAGE TO TRANSPORTATION FACILITIES 11 DAMAGE TO LIFELINES 12 CONCLUSIONS ACKNOWLEDGEMENTS REFERENCES 2

3 ABSTRACT An earthquake with a magnitude of 6.8 occurred in Chuetsu region of Niigata Prefecture of Japan on October 23, 2004 at 17:56 on JST. The earthquake caused the loss of more than 37 lives and injured more than 2500 people. The earthquake inflicted heavy damage to Kanetsu Expressway and Hokuriku Shinkansen Line and Joetsu railway line, and the Shinkansen train traveling at a speed of 200km/h was derailed. The earthquake was caused by a blind-thrust fault, which was not indicated on the active fault map of Japan. It had an unusual after-shock activity and at least 4 large after-shocks having a magnitude greater than 6 took place. The most heavily damaged towns were Kawaguchi with an intensity of 7 on the intensity scale of Japan Meteorological Agency (JMA), Tokamachi and Ojiya City. Yamakoshi village was heavily damaged due to slope failures induced by the earthquake. High ground accelerations with pronounced directivity effects did occur although the magnitude of earthquake is relatively small. Due to close proximity of the epicenter, the shaking effects become more pronounced. Furthermore, the U-shape of valley may further amplify and elongate the ground shaking Permanent deformations of ground associated with faulting did occur although no distinct ground ruptures observed. The permanent ground deformations as well as ground shaking caused some structural damage to residential houses, buildings, bridges, roads, highways, railways, expressways, lifelines. Among them, the derailment of Shinkansen train is of great importance, which will definitely have a high impact on the rehabilitation and retrofitting of existing Shinkansen lines as well high-speed railway transportation in other countries. Widespread ground liquefaction was observed throughout the epicentral region. The liquefaction caused lateral spreading and structural damage on various structures due to relative movements, uplifting or settlement. Many rock and soil slope failures took place, particularly in the mountainous area and along Shinano River and its streams. These slope failures destroyed roadways, expressways and railways. The large failure scale slope failures were directly associated with structural discontinuities in rock mass such as existing faults, bedding planes, and folds. 3

4 1 INTRODUCTION The Niigata-ken Chuetsu earthquake occurred on October 23, 2004 at 17:56 on JST and it had a magnitude (Mj) of 6.8 on the magnitude scale of Japan Meteorological Agency. The earthquake caused the loss of more than 37 lives and injured more than 2500 people. The Kanetsu Expressway and Hokuriku Shinkansen Line, Joetsu railway line were heavily damaged and the Shinkansen train traveling at a speed of 200km/h was derailed, which was the first railway accident in the 40-years old history of the bullet train, Shinkansen. The earthquake had an unusual after-shock activity and at least 4 large after-shocks having a magnitude greater than 6 took place for this M6.8 earthquake. The most heavily damaged towns were Kawaguchi with an intensity of 7 on the intensity scale of Japan Meteorological Agency (JMA), Tokamachi and Ojiya City. Yamakoshi village was heavily damaged due to slope failures induced by the earthquake. The area with a total population of 300,000 people is less populated as compared other parts of Japan. The earthquake caused the collapse of old wooden houses with heavy roofs as also observed in the 1995 Kobe earthquake. There were no known active faults in the area on the active fault maps of Japan. However, the folding process was suggested to be actively proceeding. The author had a chance to visit the area from October 29 till November 1, Except Tokamachi and Yamakoshi village***, the author visited most part of the earthquake-inflicted area. This report covers both scientific and engineering aspects of the earthquake. The material presented in this report is the interpretation and compilation of the available materials and information released by mass media, various earthquake-related institutes in Japan and other countries as well as his own observations, measurements and computations. *** The author was extremely annoyed by the restrictions imposed on academic earthquake investigators by the authorities during the site investigations of damage to various roadways and railways. 4

5 2 GEOGRAPHY The earthquake was located in the Chuetsu region of Niigata Prefecture of Japan as shown in Figure 2.1. The area is less populated compared with other regions of Japan. The overall population of the earthquake-affected region is about 300,000. The main cities in the region are Nagaoka and Ojiya (Figure 2.2). The other large towns affected by the earthquake are Kawaguchi, Tokamachi, Koide, Mitsuke, and Muikamachi. The heavy damage was observed in Yamakoshi village with a population of 2000 due to the many slope failures induced by the earthquake. Ojiya city, which is just located nearby the epicenter, has a population of The most severely damaged town is Kawaguchi with a population of The main river is Shinano River, which has several large streams such as Uono-kawa, Wanazu-kawa. The Shinano River flows from south to north and joins to Japan Sea in Niigata city. Figure 2.1 Location of the earthquake (re-arranged from ReliefWeb map of UN) 5

6 Figure 2.2 Location of cities and villages in the epicentral area (maps from JNTO) 6

7 3 GEOLOGY The epicentral area consists of Neogenic and Quaternary deposits (Figure 3.1 and 3.2). Neogenic formations are Araya, Kawaguchi, Ushigakubi, and Shiroiwa, Uonuma formations from bottom to top. In the Shinano River Lowland, the thickness of Neogenic sediments is more than 5000m. These formations are overlain by Pleistocene terrace deposits, which are covered by Holocene alluvium. These Neogenic formations and Quaternary deposits overlay the pyro-clastic volcanic basement rocks. The Quaternary deposits generally consist of clay, silt, sand and gravel. The Neogenic formations are heavily folded and their anticline and syncline axes are aligned NE-SW. The Shinano River flows through the syncline axis and Tokamachi, Ojiya and Nagaoka are situated in the Shinano valley. Another syncline, which has similar trends, is to the east, and Sumon, Koide, Yamato and Muikamachi towns are located on this syncline, along which Uono Stream of Shinano River flows. The anticline and synclines axes are tilted to NE. Uono stream changes its flow direction from NE to NW at Koide town and joins Shinano River nearby Kawaguchi town. This segment of the stream seems to follow a sinistral fault segment, which starts from Yuno valley and extends to Kashiwazaki. The strikes of the known faults, which have a main trend of NE-SW, are similar to the folding axes. Figure 3.1: A geological cross section nearby Ojiya (from Sato et al. 2001) 7

8 Figure 3.2: Geology nearby Ojiya (from Sato et al. 2001) 8

9 4 TECTONICS The plate-tectonic interpretation of Japan and its close vicinity is illustrated in Figure 4.1. The earthquake seems to take place along a collision boundary between Euro-Asian Plate and North American Plate (?). The earthquake area is on the eastern part of Fossa-Magna and underwent extensive folding, which has been still proceeding. The folding axes are ruptured by seemingly sinistral lateral strike-slip faults along the valleys of Nagano-Joetsu, Yuzawa-Kashiwazaki, Yunozawa-Ojiya and Nagaoka-Teradomari. These valleys may be interpreted as kink bands. Figure 4.1: Plate tectonics interpretation of Japan and its close vicinity (modified from NIED) The topography of Sadogashima Island to the north of earthquake area implies the existence of a dextral-type fault with an offset of 18-20km. There is no recognized active fault in the earthquake-inflicted area on the active fault map of Japan. However, folds in this area are regarded as "active folds" (Figure 4.2). Most of faults are in the vicinity of the Uonuma Hills, e.g. the Yukyuzan, Katagai, Yamamotoyama, Tokamachi and Muikamachi Faults and they are reverse faults dipping to NW. 9

10 Figure 4.2: Folding and faults in the epicentral area (modified from JGS & NIED) The author noticed some faults during the investigation (Figure 4.3). One of these faults was observed on the eastern side of Shinano-river nearby the landslide area and the second one nearby Horinouchi town along the Kanetsu expressway. One of the surfaces is associated with a thrust fault and the other surface is a sinistral lateral strike-slip fault. Faulting mechanisms inferred from the faults in Horinouchi are shown in Figure

11 (a) Thrust fault nearby the landslide area in Ojiya (b) Thrust and sinistral faults nearby Horinouchi along Kanetsu expressway Figure 4.3: Faults observed during site investigations Faulting mechanism of earthquakes between 1997 and Oct 23, 2004 obtained by F-Net of NIED are shown in Figure 4.5. A close-up view of the faulting mechanisms of the region inset in Figure 4.5 is shown in Figure 4.6. As noted from the figure earthquakes are dominated with thrust faulting. Nevertheless, some small earthquakes having sinistral lateral strike slip-faulting mechanism for NW-SE striking faults are also observed in the close vicinity of the epicentral area. It is also of great interest that 11

12 earthquakes occurred before October 23, 2004 has similar faulting mechanism to the main event. Furthermore, faulting mechanisms shown in Figure 4.4 are quite similar to those shown in Figure 4.6. Figure 4.4: Inferred faulting mechanisms for faults nearby Horinouchi 12

13 Figure 4.5: Faulting mechanism of earthquakes obtained by F-Net of NIED around the epicentral area 13

14 Figure 4.6: Faulting mechanism of earthquakes obtained by F-Net of NIED in the epicentral area 14

15 5 SEISMICITY Earthquakes with a magnitude greater than 6 in Japan Sea Region are given in Table 5.1. Most earthquakes occurred are of offshore type except Sanjo earthquake (Figure 5.1). However, from time to time shallow earthquakes do occur in the epicentral area. The 1961 Nagaoka earthquake (M5.2) caused 5 deaths (Figure 5.2). Table 5.1: Past large earthquakes in the vicinity of the epicentral area Year Location Magnitude Death 1940 Off Cape Kamui SW of Hokkaido Japan Sea Niigata Sanjo Zenkoji 7.4 Several thousands Figure 5.1: Areas and location of historical earthquakes 15

16 Figure 5.2:Iso-seismal contours of JMA seismic intensities It is pointed out that there are several seismic gapes along the collision zone between Euro-Asian plate and North American Plate. The 2004 Chuetsu earthquake filled one of these seismic gaps as shown in Figure 5.3. The author plotted the cumulative magnitude of earthquakes around the epicenter with a radius of 100km between 1973 and Oct. 23, 2004 cataloged by NEIC in Figure 5.4. The following function was fitted to the seism city. 2 M = 0.572( t 1973) It is of great interest that the above functional form is very similar to that of the accelerating creep stage of rocks in laboratory tests. Figure 5.5 shows the epicentral 16

17 distributions of the earthquakes. Although two events are seen in the epicentral area, most of earthquakes are far away from the epicenter of the main shock. Following the main-shock many aftershocks took place. Figure 5.3: Seismic gaps in the vicinity of epicentral area (modified from Yomiuri Newspaper) 17

18 MAGNITUDE YEAR Figure 5.4: Variation of magnitude and cumulative magnitude with time CUMULATIVE MAGNITUDE Figure 5.5: Pre-seismicity in the epicentral area and aftershock distributions (base map NASA) 18

19 6 FAULTING MECHANISM AND AFTER-SHOCK ACTIVITY The focal plane solutions computed by different institutes in Japan and other countries are listed in Table 6.1 and they are shown in Figure 6.1. The fundamental mode of faulting was thrust-type. Many institutes could not distinguish the causative fault from the computed mechanisms. However, the after-shock activity indicated that the fault having a NE-SW strike and inclined towards NW was contemplated as the causative fault. The subsequent large aftershocks had similar characteristics to the main fault and they migrated along a NE direction following the main shock. The focal plane solutions for some after-shocks indicated of sinistral lateral strike slip near the northern tip of the fault (Figure 6.2). However, the inclination of this fault is about 47-53degrees and it is steep for thrust faults (Figure 6.3). The fault plane is about 23km long and 14km wide. The fault terminated at a depth of 5km from the ground surface and it may be designated as a blind-thrust fault. Since the top Neogenic layers are folded and very soft, these soft layers may accommodate the relative ground deformations so that fractures on the ground surface could not become visible on the ground surface. The investigation by the active fault research group of Japan Geological Survey and other institutes could not find any noticeable ground fractures associated with thrust faulting in the earthquake-affected area. They concentrated their investigations on the Obirao fault area where the fault ruptures were expected if the causative fault was extrapolated to the ground surface. However they noticed some peculiar ground deformations in the vicinity of Hirokami village. Table 6.1: Parameters of main shock computed by various institutes Institute Latitude Longitude Depth Magnitude NP1 NP2 (km) (Mw) Strike/dip/rake Strike/dip/rake JMA /58/92 28/32/86 NIED /47/93 27/43/87 ERI /57/89 45/33/92 USGS /47/ /45/75 HARVARD /51/93 23/39/86 19

20 Figure 6.1: Faulting mechanism computed by various institutes (arranged from EMSC) Figure 6.2: Faulting mechanism of main and large aftershocks by F-Net of NIED 20

21 (a) Areal after-shock activity (b) A cross-section of locations of aftershocks along A-B line Figure 6.3: Aftershock activity and inferred causative fault 21

22 7 STRONG MOTION RECORDS AND THEIR CHARACTERISTICS The strong motion networks of Japan Meteorological Agency, K-Net, Kik-Net of NIED Japan Railways and Japan Highways recorded very high accelerations in the epicentral area. The maximum ground acceleration was observed in Tokamachi and it was more than 1700 gal while the ground acceleration was more than 1500 gal in Ojiya. Figure 7.1 shows the borehole logs and geotechnical characteristics of ground conditions at strong motion stations of Ojiya and Toka-machi. At both stations soft soil is about 2-3m deep. Therefore, the amplification effect of soft soil beneath the stations should be negligible except the well-known effect of free surface, that is, the amplification of ground surface deformation is twice that at far field for elastic response. Figures 7.2 and 7.3 show the acceleration records and integrated velocity records released by K-Net of NIED. As noted from these figures they are strikingly different from each other. Ojiya is on the hanging-wall of the causative fault and it has numerous large cycles of acceleration. On the other hand, Tokamachi is at the south-end of the causative fault and it had an impulsive character. Figure 7.4 shows the contours of the absolute and EW, NS and UD maximum ground accelerations recorded by K-Net and Kik-Net. These records clearly indicate that the characteristics of ground acceleration differ from each other and there is a strong directivity effect associated with rupturing process. Figure 7.5 shows acceleration response spectra for Ojiya, Toka-machi and Nagaoka records with a damping ratio of 5%. Toka-machi and Nagaoka response spectra have some similarities while their absolute values are different. While the peak acceleration spectra values are almost same for Ojiya and Toka-machi, the periods at the peak accelerations are different from each other. In other words, the structures having shorter period should be more affected in Toka-machi while structures with longer periods would be affected by the shaking in Ojiya. Nevertheless, the response acceleration values are 4 times that of the base acceleration. 22

23 (a) Ojiya (b) Tokamachi Figure 7.1: Geological borehole logs at Ojiya and Tokamachi strong motion stations 23

24 Figure 7.2: Acceleration and velocity records for Ojiya station (K-Net) 24

25 Figure 7.3: Acceleration and velocity records for Tokamachi station (K-Net) 25

26 (a) Absolute maximum acceleration (b) EW maximum acceleration (c) NS maximum acceleration (c) UD maximum acceleration Figure 7.4: Contours of absolute and EW, NS and UD components of maximum ground accelerations 26

27 RELATIVE ACCELERATION RELATIVE ACCELERATION RELATIVE ACCELERATION EW DIRECTION PERIOD (seconds) OJIYA TOKAMACHI NAGAOKA NS DIRECTION PERIOD (seconds) UD DIRECTION OJIYA TOKAMACHI NAGAOKA PERIOD (seconds) OJIYA TOKAMACHI NAGAOKA Figure 7.5: Response acceleration spectra for Ojiya, Tokamachi and Nagaoka stations 27

28 8 GEOTECHNICAL DAMAGE 8.1 Soil Liquefaction and Lateral spreading Soil Liquefaction was widespread all over the epicentral area. The soil liquefaction was observed from on both sides of Shinano River from Tokamachi to Nagaoka plain. Liquefaction was also observed at Joetsu city, Teradomari, Kashiwazaki along Japan Sea shores in north and Koide and Muikamachi in south. Since the soil layers are inclined towards Shinano-River, the lateral spreading and permanent displacement of ground induced by soil liquefaction took place. The soil liquefaction and associated lateral spreading caused extensive damage to linear structures such as railways, highways, embankments and viaducts of expressway and Shinkansen railways. The topsoil is mostly silty or clayey layer, which may act as a lid to prolong shaking and excess pore-pressure dissipation. Therefore, the sand volcanoes sometimes were missing although ground rupturing and ejection of mud were observed. Liquefied soil ranged from clayey soil to gravelly sand, including sand. Some sampling was done at several locations nearby Shinkansen line (Katada) where Shinkansen TOKI No: 325 derailed and Takiya. The physical and geotechnical properties of these soil samples are given in Table 8.1 and grain-size distributions are shown in Figure 8.1. Figure 8.2 shows the relation between the hypo-central distance and earthquake magnitude for liquefied sites together with liquefaction state limits developed for Turkish earthquakes (Aydan et al. 1999). Figure 8.3 shows several views of soil liquefaction in the earthquake-affected area. As it is noted from pictures, the ejection of gravels and clayey material was observed in the epicentral area and beneath Shinkansen line where the train was derailed. Furthermore, the splashing of muddy material on the columns of viaducts of Shinkansen line was 4m high in some places. Table 8.1: Properties of liquefied soil samples collected from sand volcanoes Soil Sample Dry Unit Weight Porosity Mean Grain Size Friction Angle Location (kn/m 3 ) (%) D 50 (mm) ( o ) Katada-manhole Katada-Shinkansen Takiya-field

29 CUMULATIVE PERCENTAGE (%) Clay Probable Liquefaction Limits Most Liquefiable Limits Silt Sand Gravel GRAIN SIZE (mm) Figure 8.1: Grain size distribution of samples collected from sand volcanoes CUMULATIVE PERCENTAGE (%) HYPOCENTRAL DISTANCE (km) Niigata-ken Chuetsu Earthquake Observation NO LIQUEFACTION LIMITED LIQUEFACTION MAGNITUDE M s Figure 8.2: The relation between the hypo-central distance and earthquake magnitude together with liquefaction state limits for Turkish earthquakes MODERATE LIQUEFACTION SEVERE LIQUEFACTION 29

30 Figure 8.3: Examples of liquefaction (pictures by Aydan and Niigata University) 30

31 Figure 8.4 shows two aerial photographs taken nearby Ojiya and Nagaoka along Shinano River. In these pictures taken by several Aerial Surveying Companies, the sand volcanoes and ground ruptures associated with lateral spreading are clearly visible. (a) Nagaoka (b) Ojiya Figure 8.4: Aerial photographs of liquefaction and lateral spreading (from Aerial Surveying Companies) 31

32 8.2 Slope Failures One of the most striking characteristics of this earthquake is extensive slope failure. The slope failures caused extensive damage on roadways, railways, and expressways as well destroying homes. The most extensive slope failures were observed in Yamakoshi village. The area mainly consists of Neogenic mudstone. This mudstone near ground surface is highly weathered and it had become clayey soil. The slope failures in this area are associated with bedding planes which dips NE with an inclination ranging between degrees. Before the earthquake, Typhoon No.23 passed over the earthquake epicentral area and resulted in very heavy rainfall. Therefore, rock mass and soil layers are expected to be fully saturated at the time of earthquake. The slope failures may be categorized as illustrated in Figure 8.5 a) Curved deep-seated slope failures (shear or sliding Figure 8.6) b) Shallow seated slope failures (flexural toppling, combined sliding and shearing Figure 8.7) c) Sliding failure (plane sliding, wedge sliding Figure 8.8) d) Rock-falls (toppling, bending failure Figure 8.9) Figure 8.5: Classification of slope failures (arranged from Aydan 1989) 32

33 Figure 8.6: Aerial photographs of slope failure at Yamakoshi and along Shinano River 33

34 Figure 8.7 Slope failures 34

35 Figure 8.8: Planar Rock Slope failures 35

36 Figure 8.9: Wedge, toppling and combined rock slope failures 36

37 8.3 Embankment Failures and Settlements Embankment failures and settlements were widespread. Several typical examples are shown in Figure The highways, Kanetsu expressway and railways are mainly built on the embankments. The embankments mostly failed in the form of curved sliding. Some failures were associated with the liquefaction of soil beneath the embankments. Figure 8.10: Embankment failures 37

38 There are also some earth dams built for irrigation and drinking water purposes. The failure and settlement characteristics of embankments of these structures are basically the same. Some examples are shown in Figure Figure 8.11: Embankment failures associated with earth dams 38

39 9 DAMAGE TO BUILDINGS 9.1 Wooden houses Mainly old wooden houses were either collapsed or heavily damaged. The failure was quite similar to those observed in 1995 Kobe earthquake. The failure mechanism mainly involved hinging of wooden columns at the base and also at the interface between 1st floor and 2nd floor as a result of large horizontal earthquake forces. The traces of such mechanism can be seen widespread in the epicentral area. In addition some relatively new wooden houses either collapsed or were heavily damaged. The main cause was the weak floor situation observed at the first floor used as garage, workshop or storage (Figure 9.1). Some houses either collapsed or were heavily damaged by the slope or embankment failure, on which they were built. This type of failure observed in mainly hilly regions and also Yamakoshi village, which suffered from extensive slope and embankment failures. Figure 9.1: Collapse of wooden houses 39

40 9.2 RC Buildings RC buildings with or without steel frames in the epicentral region are few. They are used as schools, public offices and a few residential buildings. The number of stories is mostly 3 to 5. Some residential buildings had 10 stories. Furthermore, the infill walls of the RC buildings were constructed with shear-walls. Therefore, their behavior should had been close to rigid boxes on softer ground. The observations and investigations in many cities and towns indicated that the RC buildings were generally none or slightly damaged in spite of high ground accelerations. Damage to RC buildings was caused by column failure at the ground floor. In some of buildings, the quality of concrete and density of stir-ups were not in accordance with regulations (Figure 9.2). Figure 9.2: Damaged RC buildings 40

41 10 DAMAGE TO TRANSPORTATION FACILITIES 10.1 Railways Joetsu local and Shinkansen lines pass through the epicentral area and it runs along the eastern bank of Shinano River next to hills. Joetsu local line is directly built on existing ground and embankments. The damage to Joetsu local line was quite extensive in the epicentral region. The damage to the railway line was mostly associated with the lateral spreading of ground, the failure or settlement of embankments and slope failure (Figure 10.1). Some structural damage was associated with the poles of pantographs, which were settled, buckled or broken. The rails were bended or buckled due to lateral spreading of ground. Figure 10.1: Damage to railways Most part of Joetsu Shinkansen line passes through long tunnels and viaducts over rivers in the epicentral region. The line extends from the portal of Takiya tunnel to 41

42 Niigata mainly on viaducts. The viaducts are framed RC structures. The viaducts between the northern portal of Takiya tunnel and Nagaoka station was damaged at several locations. The section between Niigata and Tsubame-Sanjome stations is non-damaged and presently Shinkansen operates along this section. The viaducts segments of the Shinkansen line were deformed due to lateral spreading of ground due to ground liquefaction as presented in Chapter 8 and these viaduct segments were offset at several locations which should had caused some offsets of the rails. However, they could not be confirmed on-site due to restrictions imposed by JR authorities. The RC columns of viaducts were mainly ruptured at the locations, where columns are connected to upper railway platform (Figure 10.2). Such locations constitute constructions joints, which have almost none or small adhesion and all earthquakeinduced loads should had been transferred to reinforcement bars. This type failure was also observed in 1995 Kobe earthquake and they were also repeated in this earthquake. Following the 1995 Kobe earthquake, columns were jacketed by steel plates. However, it might be very difficult to recognize or identify any damage to jacketed columns if they are damaged. One of the important events in this earthquake is the derailment of Shinkansen train. TOKI-325 of Joetsu Shinkansen train was traveling at a speed of 200 kph when the earthquake hit. The event took place when the train appeared from the northern portal of Takiya Tunnel and it stopped at a distance of about 1500m from the portal. The wagons at the ends of the train derailed with widespread damage to the railways (Figures 10.3 and 10.4). Following these events, it was explained that train would derail if the accelerations waves greater than 100mm displacement amplitude with 1Hz frequency acted on the train laterally. Some emphasis was put upon the height of piers of the viaducts as they become higher from the tunnel portal towards Nagaoka station. There is no doubt that the shaking period of the viaducts would become larger as the height of structure becomes higher. Furthermore, the liquefaction of foundation soil should also cause the prolongation of acceleration waves. In addition permanent ground deformations can cause offsets of the rails, which should result in further undesirable shaking and tilting effects on the train as observed on derailed AMTRACK train during Hector Mine earthquake in These should further make the train vulnerable to derailment. Besides these effects, the train may start rocking and also tilting when high lateral accelerations act on the train. A simple analysis indicates that rocking of a stationary wagon of Shinkansen may start rocking at a lateral acceleration of 380 gal. By considering the measured acceleration in the region, the ground acceleration would be sufficient to tilt the Shinkansen train. 42

43 Figure 10.2: Damaged columns of viaducts of Shinkansen line Figure 10.3: Illustrations of Shinkansen derailment (modified from Yomiuri Newspaper) 43

44 Figure 10.4: Maximum ground accelerations along Joetsu Shinkansen and views of derailed Shinkansen train (arranged from Yomiuri Newspaper and Internet web-sites) 44

45 10.2 Roads, Highways, Expressways Roads in cities, towns and villages were heavily damaged due to ground settlement or manhole uplift due to soil liquefaction (Figure 10.5). Some of damages were also caused by slope failures in the hilly parts of the cities and towns. The uplift of manholes of sewage system was as much as 1000mm in Ojiya, Katada and Tokamachi. Such uplifts were also observed in Nagaoka city and even in Teradomari town along Japan Sea. Figure 10.5: Damage to roads due to liquefaction and embankment settlement (Pictures from various web sites on INTERNET) 45

46 Highways R17, R117, R351, R291 and R352 pass through the epicentral region. These highways built on existing ground or embankments were damaged in various parts due to either the failure or settlement of embankments and slope failure (Figure 10.6). The highways leading to Yamakoshi village were extensively damaged by the slope failure. In some parts, the highways were included within the sliding masses. The settlements were quite amplified at the locations where embankments are in contact with rigid bridge platforms. The relative settlement at many such locations was more than 30cm. Figure 10.6: Damage to roadways due to slope and embankment failure (Pictures from various web sites on INTERNET) 46

47 Kanetsu expressway passes through the epicentral region while Joshinetsu expressway follows the route from Joetsu along Japan Sea and joins with Kanetsu expressway at Nagaoka. Both expressways were built on either directly existing ground or embankments. The damage of these expressways was basically similar to those observed on highways and ordinary roads (Figure 10.7). The surfacing of the expressways was wavy due to permanent deformation of embankments. Furthermore they were ruptured at the joints and spreaded laterally. Figure 10.7: Damage to Kanetsu expressway due to slope and embankment failure (Pictures from various web sites on INTERNET) 47

48 10.3 Tunnels Since the area is quite hilly and it is well-known sites of landslides, many tunnels were built in association with highways, expressways and railway lines. Although underground structures are known to be resistant to earthquakes except at locations where relative deformation of ground (i.e. faulting) takes place, many tunnels suffered some damage to a certain extent. It was almost impossible to visit these sites due to restrictions imposed by JR and JH authorities. The pictures shown in Figure 10.8 are those collected from some news media and some web sites. The tunnels of Railways and Roadways run almost parallel to the NW-SE direction, which is almost parallel to the strike of the causative faults. Damage on tunnels was associated with non-reinforced concrete linings. The concrete linings spalled at the sidewalls and roof. Although the effect of high ground accelerations may play some certain roles on the spalling of concrete lining, the permanent deformation of rock mass may also be one of the causes of damage. Figure 10.8: Damage to roadway and railway tunnels (Pictures from various web sites on INTERNET) 48

49 10.4 Bridges There are several railway and roadway bridges spanning over the Shinano River and its streams. The long span bridges are located nearby Ojiya and Kawaguchi area. There was no bridge collapse in-spite of high ground accelerations. The bridges are generally designed as redundant structures. Therefore none of girders of the bridges are fallen during this earthquake. The concrete piers of the bridges, spanning over Shinano River or its streams with large span and their longitudinal axis is perpendicular or sub-perpendicular to the strike of the causative fault, were ruptured at the construction joints near at their mid-height (Figure 10.9). The same phenomena were observed in the piers of bridges spanning over Muko River in the 1995 Kobe earthquake. It is generally expected that piers are weaker in bending rather shearing and the bending failure should occur at their base. The rupturing of piers at their mid-level may imply high shear forces along the longitudinal axes of the bridges associated with permanent ground deformations associated with in-land earthquakes. Figure 10.9: Damage to bridges (Pictures from Aydan and various web-sites) 49

50 11 DAMAGE TO LIFELINES 11.1 Electricity The electric supply to the earthquake stricken area was shutdown following the earthquake. Except heavily damaged areas, the electricity was restored after three days following the earthquake. The electricity distribution system in the region is done through poles and high-voltage transmission lines. Although concrete poles were either inclined as a result of partial bearing failure of the surrounding ground or collapsed (Figure 11.1). The poles having transformers mounted on were particularly suffered more and also the poles were inclined in heavily liquefied areas. Nevertheless, the repair works on electricity distribution system could be accomplished in a short period of time. The main transmission lines are elevated and supported through 40-50m high steel pylons. In-spite of wide spread ground liquefaction or slope failures, there was no damage to pylons seen during the author s investigations or reports by various investigation groups. Figure 11.1: Damage to electric transmission lines (Pictures from various web-sites) 50

51 11.2 Water Network The water networks were particularly damaged in Ojiya city, Tokamachi and Kawaguchi towns. The damage to water network is generally associated with pipe breakage at joints (Figure 11.2). Figure 11.2: Damage to water network 51

52 11.3 Natural Gas Network The natural gas systems consist of spherical storage tanks and distribution network. There was almost no damage to spherical gas storage tanks all over the epicentral region although slight damage could be seen at their foundations and connection pipes (Figure 11.3). Although the region is less populated and the networks are not so extensive, the gas supply could not be in operation in Ojiya and Kawaguchi. Since permanent ground deformations due to either soil liquefaction or associated with and faulting did take place, the regional gas company together with help from other gas companies of Japan have been now checking the gas pipe network block by block. The natural gas system is expected to be fully in operation after 1 month according to local gas company officials. Figure 11.3: Non-damaged spherical gas tanks 52

53 11.4 Sewage The sewage systems were heavily damaged by the liquefaction of back-fill soil (Figure 11.4). The uplift of manholes could be observed even in Teradomari town along Japan Sea. The uplift of manholes was as much as 1000mm. This implies both long-duration large ground shaking. The local authorities now have been checking the sewage system and replacing the sections underwent uplift or ruptured. Since the permanent ground deformations occurred due to lateral spreading of ground as a result of liquefaction, the joints of the sewage system suffer some leakage and need repairs Figure 11.4: Damage to sewage systems 53

54 11.5 Telecommunications Telecommunication system suffered the same problem due to heavy telecommunication traffic. Besides the conventional telephone system, the mobile telephone systems of Docomo, J-Phone and AU are used widespread. Particularly Docomo system suffered from heavy telecommunication traffic following the earthquake. However, the systems become normal next day. 54

55 12 CONCLUSIONS 2004 Niigata-ken Chubu earthquake is an in-land earthquake and caused widespread damage particularly in the epicentral area, which extends from Nagaoka in north to Tokamachi in south and Teradomari in west Koide in East. The main characteristics of this earthquake can be summarized as follows: 1) The faulting was of blind-thrust type. However, no ground ruptures were observed on the grounds surface associated with faulting. This may be due to soft 5km thick Neogenic folded sedimentary rocks overlain basement rocks. 2) Very large aftershocks have been observed and they are still continuing. It seems that the faulting plane propagating in NE direction as it terminated at a depth of 5km. 3) High ground accelerations with pronounced directivity effects did occur although the magnitude of earthquake is relatively small compared with, for example, 1995 Kobe earthquake. This feature of the in-land earthquakes is quite distinct with that of offshore earthquakes. Due to close proximity of the epicenter, the shaking effects become more pronounced. Furthermore, the U-shape of valley may further amplify and prolong the ground shaking as measured in Ojiya strong motion stations. 4) Permanent deformation of ground associated faulting did occur although no distinct ground ruptures observed. In addition to that of the faulting, the ground liquefaction resulted in widespread permanent ground deformations towards Shinano River and its streams. 5) Widespread ground liquefaction was observed throughout the epicentral region. The liquefaction caused lateral spreading and damage on various linear structures through relative movements, uplifting or settlement. 6) Many rock and soil slope failures took place, particularly in the mountainous area and along rivers. These slope failures destroyed roadways, expressways and railways. The large failure scale slope failures were directly associated with structural discontinuities in rock mass such as existing faults, bedding planes, folds. In some places, debris of slope failures resulted in natural dams. 7) The permanent ground deformations as well as ground shaking caused some structural damage to residential houses, buildings, bridges, roads, highways, railways, expressways, lifelines. Among them, the derailment of Shinkansen train is of great importance, which will definitely have a high impact on the rehabilitation and retrofitting of existing Shinkansen lines as well high-speed railway transport in other countries. 55

56 ACKNOWLEDGEMENTS This report has been written with a sole purpose of putting scientific and engineering information from different sources and author s his own observations and computations in a unified manner. The author is particularly great full to various institutes and people involved with various aspects of this earthquake and observations to make their information, data and computations accessible through INTERNET listed in References. If any material presented in this article goes un-referred, it is not done intentionally. This site investigation has been carried out with the collaboration of Prof. Dr. M. Hamada and his research staffs. The author also would like acknowledge the computation center of Marine Science and Technology School of Tokai University for their understanding and help putting this report on the INTERNET. 56

57 REFERENCES Aero Asahi Corporation (2004): Asahi Newspaper (2004): Asia Air Survey Co. (2004): Aydan, Ö. (1989): The stabilisation of rock engineering structures by rockbolts. Doctorate Thesis, Nagoya University, Engineering Faculty, p.204. Aydan, Ö., Y. Shimizu, Y. Ichikawa (1989). The Effective Failure Modes and Stability of Slopes in Rock Mass with Two Discontinuity Sets. Rock Mechanics and Rock Engineering, 22(3), Aydan, Ö., Ulusay, R., Hasgür, Z., and Taþkýn, B., A site investigation of Kocaeli Earthquake of August 17, Turkish Earthquake Foundation, TDV/DR 08-49, 180p. ERI (2004): IISEE (2004): GSJ(2004): JGS(2004): JSCE(2004): Japan National Travel Organization (2004): Maps of Chubu Region. K-net (2004): KiK-Net (2004): Mainichi Newspaper (2004): Kokusai Kogyo Co. (2004): Nakanihon Air Service Co. (2004): Nakano AI System Co. (2004): NASA(2004): Earthquakes near Niigata, Japan, (img_id=16709). NIED(2004): Niigata University (2004): Original Research Ideal Survey (2004): ReliefWeb Map (2004): Glide Number-EQ Sato, H.P., Abe, K., Otaki, O. (20043): GPS-measured land subsidence in Ojiya City, Niigata Prefecture, Japan. Eng. Geolo., 60, USGS (2004): Yahoo Japan (2004): Yomiuri Newspaper (2004): 57