COMPARISON OF ACCELERATION AND LIQUEFACTION POTENTIAL ESTIMATED BY RSPM WITH OBESERVED ONES DURING THE 2011 GREAT EAST JAPAN EARTHQUAKE Shouta IKARASHI 1, Susumu YASUDA 2, Keisuke ISHIKAWA 3 ABSTRACT The authors implemented Japan s Nation-wide Electronic Geotechnical Database System (NEGDS), developed by the Japanese Geotechnical Society, to several cities in Japan to construct representative soil profile models (RSPMs) as described by Yasuda et al. (2011). During the study, the M w =9.0 Great East Japan Earthquake hit Japan. The authors applied the RSPM to one- dimensional seismic response analyses to estimate the distribution of acceleration in Urayasu City, where most of the severe damage occurred due to liquefaction. Liquefaction potential was also estimated with the RSPM. A total of 396 digital boring datum was provided by the Urayasu City government and others. One-dimensional seismic response analysis revealed that measured accelerations were slightly higher than estimated ones at several sites in Urayasu. However, generally, the estimated accelerations coincided with the measured accelerations, al-though the data are scattered. -Therefore, it can be said that, generally, an RSPM is useful in estimating the distribution of maximum surface acceleration. The distribution of liquefaction potential estimated by RSPM also coincided with the actual distribution of liquefied zones during the Great East Japan Earthquake of 2011. INTRODUCTION The Mw=9.0 Great East Japan Earthquake occurred in the Pacific Ocean approximately 130 km off the northeast coast of Japan s main island on March 11, 2011. The hypocentral region of this quake was approximately 500 km in length and 200 km in width. Many houses and lifelines were damaged by soil liquefaction, landslides occurred, dams failed, and river dikes settled. The damage was not limited to the Tohoku district of northeastern Japan; the Kanto district surrounding Tokyo was also damaged. A significant amount of land has been reclaimed in the Tokyo Bay area since the seventeenth century. Liquefaction occurred in a wide area of the reclaimed land along Tokyo Bay, though the epicentral distance was very large, approximately 380 to 00 km, as shown in Fig. 1. Tokyo Bay area was only approximately 110 km from the boundary of the rupture plane, because the plane was very wide. Although liquefaction has been induced during past earthquakes, such as the 1923 Kanto Earthquake and the 1987 Chibake-toho-oki Earthquake, the Great East Japan Earthquake is the first on record to cause liquefaction in such a wide area and to severely damage houses, lifelines, and roads. Many seismometers have been installed in and around the Tokyo Bay area. Figure 2 shows ground surface accelerations measured by the seismograph in Urayasu (2011). From the figure, it can be seen that a 150~270 gal about. 1 Graduate student, Tokyo Denki University, Saitama, Japan, 13rmg01@ms.dendai.ac.jp 2 Professor, Tokyo Denki University, Saitama, Japan, yasuda@g.dendai.ac.jp 3 Assistant, Tokyo Denki University, Saitama, Japan, ishikawa@g.dendai.ac.jp 1
The authors implemented the Nation-wide Electronic Geotechnical Database System (NEGDS), developed by the Japanese Geotechnical Society, to several cities in Japan to construct representative soil profile models (RSPMs). Then, the authors applied the RSPM to conduct a onedimensional seismic response analysis to estimate the distribution of acceleration in Urayasu City, where most of the severe damage occurred due to liquefaction. Liquefaction potential was also estimated with the RSPM. Figure 1. Epicenter and rupture plane of the Great East Japan Earthquake of 2011 (Adapted from Geospatial Information Authority of Japan). 177.8gal 176.gal Old coast line 232.0gal 178.1gal 255.0gal 185.3gal 271.1gal 176.0gal 169.3gal 169.3gal 225.1gal 215.5gal 167.6gal 156.5gal 163.6gal 151.6gal Figure 2. Maximum ground surface acceleration in Urayasu (measured) 2
Authors should be written like A.Mehmet and M.Ahmet 3 METHOD In 2006, a five-year nation-wide inter-agency project began in Japan, which was called the Integrated Geophysical and Geological Information Database and was led by the National Research Institute for Earth Science and Disaster Prevention (NIED), with the participation of the National Institute of Advanced Industrial Science and Technology (AIST), the Public Works Research Institute (PWRI), the University of Tokyo, Tokyo Institute of Technology, and the Japanese Geotechnical Society (JGS). Recently, digital geotechnical databases with digital maps have been developed at various organizations in Japan, as shown in Fig. 3. The scheme undertaken by the JGS is to link the existing geotechnical databases. The JGS established a technical committee for this purpose in 2006, and the project was named the Nation-wide Electronic Geotechnical Database System (NEGDS). The system is essentially a collection of ground models, sized at of 250 m 250 m in plan. The data is free from ownership and copyright, and permission is granted for disclosure of information on private properties. At a minimum, the following information is required to be provided for each 250-m mesh: latitude and longitude of the mesh, -ground elevation, soil layers, SPT N-values, groundwater levels, and depth to bed rock. For the moment, the soil types choices are: gravel soil, sandy soil, clayey soil, organic soil, volcanic clay, peat, artificial material, and rock. Figure shows a computer screen during the construction of a representative soil profile model (RSPM). The procedure to construct an RSPM is as follows: 初期作成済み Constructed in the 1 st stage 別途作成 Constructed in the 2 nd stage 作成中 To be constructed Sapporo Niigata Fukuoka Tokyo Osaka Matsuyama Figure 3. Areas where digital databases have been developed and NEGDS is implemented (1) Select a mesh to construct the model. A standard 250-m mesh map published by the Geographic Survey Institute of Japan is used. (2) Place digital boring data in the mesh. If necessary, place boring data in the neighbor meshes as well. Inappropriate data are deleted by considering topography and geology. -Depth of bedrock in each boring is estimated for engineering purposes. (3) The system automatically constructs the RSPM to the depth of bedrock by averaging soil type, SPT N-value, and groundwater level. The RSPM is not constructed by selecting one borehole inside the area; rather, the system constructed surrounding ground conditions. Thus, the RSPM represents average ground conditions in the area and not the ground condition at the center of the area. It should also be noted that, although a model may be constructed automatically, judgments based on knowledge of the local geology and soil conditions should be exercised.
2. Acquire soil data inside the mesh and adjacent meshes Representative soil characteristics Detailed surface data (Including >3points) 1. Setting soil layers for modeling 3. Modeling (Division and averaging of layers) EXPLANATORY NOTES Spatial image of the model y Sand Clay Gravel Organic Soil Volcanic Soil Peat Artificial Material Rock z x Soil 土質 Type values 試験値 12 13 Sand 砂 11 12 13 3 Clay 粘土 5 5 Sand 12 砂 25 0 13 Figure 2. Operating Figure. Operating window of the NEGDS supporting In order to check usability and applicability of the software for this study, a series of trials were conducted at various ground conditions. This was necessary because the prototype software was developed for a typical ground of the Osaka Plain, where a relatively large plain is developed with almost uniformly distributed thick Holocene deposits. The software has been modified and upgraded based on the trials. Selected areas for the trials are Osaka, Fukuoka, Sapporo, Matsuyama, Niigata, and Tokyo, as shown in Fig. 3. DESCRIPTION OF THE GROUND IN URAYASU CITY The history of reclamation work in Urayasu City, where many house were seriously damaged in the Great East Japan Earthquake, is summarized in Fig. 5. Urayasu City is divided into three towns: Motomachi, Nakamachi, and Shinmachi. These names translate to, middle, and new towns, respectively. The ground of Old Town has been formed naturally at an estuary of the Edo River. In contrast, zones A, B, and C in Nakamachi were reclaimed from 1965, and zones D, E, and F in Shinmachi were reclaimed from 1972. The area of each zone is 1.0 to 3.5 km 2, and the total area of Nakamachi and Shinmachi is 1.36 km 2, which is approximately 75% of the area of Urayasu City. According to an engineer who supervised the reclamation work in Urayasu, soils dredged from the bottom of the sea were filled to a height of approximately sea level in the reclamation work. Then the filled surface was covered with hill sands transported by boat from the Boso Peninsula. Figure 6 is a schematic drawing of the dredging work. Soils from the sea bottom just outside zones C, F, and D were excavated by a cutter, drained of water with a pump, transported by a conveyor pipe, and then discharged from the pipe. As the dredged soils contained much water, coarse grains soils were probably deposited near the mouth of the conveyor pipe, and fine grains soil were probably deposited far from the pipe mouth. Moreover, the pipe mouth was frequently re-positioned, resulting in very non-homogeneous strata.
Authors should be written like A.Mehmet and M.Ahmet 5 B area A area Area A 2.18km 2 B 3.05km 2 C 3.50km 2 Total 8.73km 2 C area E area D area Area D 2.2km 2 E 2.21km 2 F 1.00km 2 Total 5.63km 2 F area Figure 5. History of reclamation in Urayasu (Quoted from the report by the Technical Committee organized by Urayasu City (2012)) Sea level Pump dredger Floating pipe Discharge pipe Distributing pipe Discharge opning Cutter Pump Sea bottom Sea wall Reclaimed soil Figure 6. Schematic drawing of dredging work at Urayasu City.
Analyzed maximum surface acceleration(gal) Elevation T.P. (m) Number of meshes DESCRIPTION OF MODELING OF URAYASU In case of Urayasu City, the RSPM was constructed within an area of approximately 5 km 5 km, as shown in Fig. 7. A total of 396 digital boring datum were provided by Urayasu City government and others. Figure 8 is histogram of the amount of boring data in each mesh. In the construction of the RSPM, depth of bedrock was judged from SPT N-value. If the SPT N-value of a layer was greater than 50, the layer was judged as the bedrock for engineering purposes. Figure 9 shows a series of RSPMs along the A-A line. The zones where sand boils were observed during the Great East Japan Earthquake coincided exactly with the area of reclaimed land. In the reclaimed zone, a 6-9 m thick filled layer is the uppermost deposit. It consists mainly of hill sand and dredged sandy soil with low SPT N-values of 2_8. A -8 m thick alluvial sand layer (A S ) is deposited under the filled layer. It consists of sands with SPT N-value of 10 to 20. A 10-0 m thick very soft alluvial clay layer (Ac) is deposited under the A S layer. Its thickness increases toward sea. Old coast line 5 0 35 30 25 20 15 10 5 0 1 2 3 5 6 7 8 9 10 11 12 13 1 15 16 17 18 19 20 Boring number Figure 7. Boring sites and modeled meshes in Urayasu City Figure 8. Histogram of boring data in each mesh A Natural deposition layer Reclaimed zone A A B old coast line A B 00 300 Motomachi Nakamachi Shinmachi Legends Sandy soil Clayer soil Gravel Organic soil Volcanic soil Peat Artificial Material Rock Figure 9. Representative soil profile models along A-A line Figure 200 10 illustrates the estimated soil cross section in Urayasu. One soil cross section shown in Fig. 10 is the section along the B-B line where many houses and lifelines suffered severe damage due to liquefaction. The zones where sand boils were observed coincide exactly with the area of reclaimed land, which is the sea side from old sea wall. In the natural land inland from the old sea wall, the same 100 A S layer as in the reclaimed land is deposited from the ground surface. If the A S layer had liquefied, sand boils should have appeared in the natural land. However, no sand boils, nor damage to houses and lifelines, occurred in the natural land. Therefore, it can be inferred that the A S layer did not liquefy 0 0 100 200 300 00 Measured maximum surface acceleration(gal) 6
Authors should be written like A.Mehmet and M.Ahmet 7 during the Great East Japan Earthquake (al-though some loose parts in the reclaimed land might have done so) and that some part of the dredged sandy soil under the water table might have liquefied. The estimated soil cross sections in Figs. 9 and 10 are similar. From this, RSM, which is a simple ground model in reproducing the earmark of the layer of Urayasu. B B B: fill F: dredged and filled sand As: Alluvial sand Ac: Alluvial clay Ds: Diluvial sand Dc: diluvial clay Lm: Loam Figure 10. Estimated brief soil cross sections along B-B line (Yasuda et al., 2012) SEISMIC RESPONSE ANALYSES USING RSPM IN URAYASU One- dimensional seismic response analyses were carried out using the computer program SHAKE. Shear wave velocity, VS, was estimated from SPT N-values by the relationship introduced in the Specification of Highway Bridge in Japan. The G/G0 - γ and the h - γ relationships were estimated by the relationship proposed by Yasuda and Yamaguchi (1985). Input soil data for the analyses are shown in Table 1. The VS of bedrock was assumed to be 350 m/s. Seismic records during the Great East Japan Earthquake were obtained not only at the ground surface but also under the ground. The seismic wave recorded at Yumeno-shima, approximately 5 km west of Urayasu City, was used for the analyses. Because the seismic wave was recorded at GL-89.8 m, the wave at GL-52.2 m which is the bedrock of the RPSM was estimated by SHAKE. Figure. 11 shows the input wave thus estimated. The peak acceleration is 138.52 cm/s 2. Table 1. Input soil data for the analysis Lowland γ t (kn/m 3 ) D 50 (mm) Sandy soil 18 0.15 Clayey soil 16 0.02 Organic soil 12.3 0.02 Artificial soil 18 0.15
Analyzed maximum surface acceleration(gal) Analyzed maximum surface acceleration(gal) Figure 11. Input seismic wave Figure 12 shows the distribution of the analyzed maximum surface acceleration A max. Figure 13 shows the relationships between measured A max and analyzed A max. Measured accelerations were slightly higher than analyzed ones in several sites. However, the analyzed accelerations generally coincided with the measured accelerations, although the data are scattered. Therefore, it can be said generally that the RSPM is useful to estimate the distribution of the maximum surface acceleration. 00 300 Motomachi Nakamachi Shinmachi old coast line 200 00 300 Motomachi Nakamachi Shinmachi 100 0 0 100 200 300 00 Measured maximum surface acceleration(gal) Figure 12. Analyzed maximum surface acceleration 200 100 Figure 13. Relationship between measured A max and analyzed A max PLvalues m 2 PLvalues m 2 20 or more 20 or more 15 or 15 more or less more than 20 less than 20 5 or more less than15 5 or more less than15 0 or more less than5 0 or more less than5 Old coast line 0 0 100 200 300 00 Measured maximum surface acceleration(gal) Figure 1. Method to judge liquefied and nonliquefied zones in Urayasu Figure 15. Estimated liquefaction potential, P L, in Urayasu 8
Authors should be written like A.Mehmet and M.Ahmet 9 The authors started to investigate liquefied zones in the Tokyo Bay area on the day after the earthquake (Yasuda et al., 2013). In this investigation, the roads where boiled sands were observed, and those where they were not, were marked on maps as shown in Fig. 1. The zones surrounded by red lines were judged to be liquefied. As mentioned above, liquefaction occurred in Nakamachi and Shinmachi, and did not occur in Motomachi. In the estimation of liquefaction potential, P L, liquefaction strength was evaluated based on SPT N-value and fines contents. As shown in Fig. 15, estimated P L is greater than 5 in Nakamachi and Motomachi, and actual P L is less than 5 in Motomachi. CONCLUSIONS In this study, we created a model of the ground from detailed drilling data and conducted onedimensional seismic response analyses using this ground model. With one-dimensional seismic response analysis, it was possible to obtain a maximum ground acceleration that coincided with that observed at the site. Further, in a simple liquid assessment, it was found that P L value is larger in reclaimed land, indicating the range is a good representation of the liquid in practice. In order to generate estimates for earthquake disaster prevention that correspond with actual measurements, it is important to create RSM using past ground survey data. REFERENCES Urayasu City. (2012) Data Compiled by the Technical Committee on Measures against Liquefaction, http://www.city.urayasu.chiba.jp/menu1132.html, 11p. (in Japanese) Yasuda S, Watanabe H, Yoshida N. (2011) Seismic zoning for ground motion in Tokyo based on Nation-wide Electric Geotechnical Database System, 5 th International Conference on Earthquake Geotechnical Engineering Yasuda S and Yamaguchi I. (1985) Dynamic soil properties of undisturbed samples, Proceeding of the 20 th Japan National Conference. on SMFE, 539-52. (in Japanese) Yasuda S, Harada K, Ishikawa K, Kanemaru Y,(2012) Characteristics of the Liquefaction in Tokyo Bay Area by the 2011 Great East Japan Earthquake, Soils and Foundations, Vol.52, Issue 5, 793-810. Yasuda S, Todo, H, Mimura, M, Yamamoto, K, (2010.) Nation-wide Electronic Geotechnical Database System in Japan. Proceedings of the 17 th Southeast Asian Geotechnical Conference, 181-18. Yasuda S, Hagiya S, (2011.) Estimation of soil cross section in the Liquefied Area along Tokyo Bay during the 2011 Tohoku-Pacific Ocean Earthquake. Proceedings of the 8 th Geo Kanto 2011, Kanto Branch of the Japanese Geotechnical Society, 87-90 (in Japanese).