CONTINUOUS OBSERVATION OF MICROTREMOR AT A GEO- DISASTER SITE BEFORE AND AFTER THE 20 OFF THE PACIFIC COAST OF TOHOKU EARTHQUAKE Yoshiya HATA, Koji ICHII 2, Yoshihisa MARUYAMA 3 and Hisakazu SAKAI 4 INTRODUCTION The geo-disaster caused by the March, the 20 off the Pacific coast of Tohoku Earthquake is too extensive and widely distributed, its outline and details are recently revealed. In view of this, by focusing on the geo-disaster occurrence near the downstream basin of the mainstream Tone River, Japan, a case history is reported. In particular, we focused on the geo-disaster site including soil liquefaction in Fusa District, Abiko City, Chiba Prefecture (e.g., Koseki et al., 202, Koarai et al., 20, Koarai, 203; see Site of Interest in Figure ). This is because we carried out microtremor measurement 4 days ago (before the 20 main shock) in Fusa District. Therefore, if we carried out microtremor measurement at same site in Fusa District after the 20 main shock, by comparing the obtained microtremor data before and after the 20 main shock, we can perform the direct valuation of the difference in the ground shaking characteristics. It is thought that the difference expresses the recovery process of the ground shaking characteristics at the measurement site in Fusa District after the 20 main shock. By receiving the large distortion history due to a large-scale earthquake, the phenomenon in which the shear modulus (the velocity of elastic wave) of the sand ground falls is confirmed in the past laboratory test (e.g., Tokimatsu and Hosaka, 986). On the other hand, although the clay ground does not generate the extreme soil liquefaction by cyclic load, causing the shear modulus fall due to decrease of the effective stress because excess pore water pressure rises is known (e.g., Houlsby and Wroth, 99; Nakayama et al., 998). However, there are few examples where the phenomenon was checked on the actual ground condition. There is almost no example which detected quantitatively the process which the shear wave velocity of the affected and lowered ground recovers with time progress after a large-scale earthquake (Arai et al., 2006). That is, at present, there are many unknown portions about the shear modulus (the shear wave velocity) fall and its recovery process in the actual ground which experienced the effect of the strong motion. It is very interesting on geotechnical engineering to clarify the fall and the recovery process at an actual field site. Moreover, this is important on earthquake engineering as a study which evaluates the change of ground shaking characteristics in the damage area due to a large-scale earthquake. Based on the above background, in this study, first, microtremor measurements were carried out periodically for,000 days after the 20 main shock at the site of interest in Fusa District. Based on the obtained microtremor data, then, we performed spectral analyses in order to compare the characteristics of the microtremor H/V spectra before and after the 20 main shock. Furthermore, we Graduate School of Engineering, Osaka University, Suita, Japan, hata@civil.eng.osaka-u.ac.jp 2 Graduate School of Engineering, Hiroshima University, Higashihiroshima, Japan, ichiikoji@hiroshima-u.ac.jp 3 Graduate School of Engineering, Chiba University, Chiba, Japan, ymaruyam@tu.chiba-u.ac.jp 4 Faculty of Engineering and Design, Hosei University, Tokyo, Japan, hisakai@hosei.ac.jp
Ibaraki Pref. Tokyo Site of Interest Chiba Pref. The 20 Tohoku EQ. 20/03/ 4:46 M W 9.0 Chiba Pref. Off Ibaraki Pref. EQ. 20/03/ 5:5 M J 7.6 Hamadori Fukushima Pref. EQ. 20/04/ 7:6 M J 7.0 Figure Location information of Fusa District. evaluated the fall and its recovery process of the shear wave velocity in Fusa District where experienced the effect of the strong motion by focusing on change of the characteristics of the microtremor H/V spectra such as predominant frequency. Finally, we carried out effective stress analysis during the 20 main shock and its consolidation analysis after the 20 main shock in order to investigate the change of the natural frequency of the ground model at the site of interest in Fusa District before and after the 20 main shock. SEISMIC DAMAGE AND STRONG MOTION RECORDS In Fusa District, Abiko City, Chiba Prefecture, both of liquefied sites and non-liquefied sites were confirmed by MLIT and JGS (20), and they were located in an alternative manner as shown in Figure 2. In Figure 2, their location is also plotted on an old map that was compiled based on survey results conducted in 907. Here, based on the result of author s reconnaissance survey on 3 March 20, the trace of the soil liquefaction such as boil sand is not confirmed at the site of interest. Figure 2 reveals that the liquefied sites are located at old swamps or their vicinities. According to relevant literature (Harada, 2004), these swamps appeared and were later reclaimed as follows. First, a flood in 870 caused breakage of the Tone River dike for longer than 80 m and submerged the whole area, leaving a swamp called Kirisho-numa with an area larger than 5 ha. The area suffered also from subsequent flood disasters. Next, a project was initiated in 952 to improve the Tone River channel and dike. The dredged soil in this project was transported through pipes to reclaim the above swamp. Finally, it was developed as residential area. The geology characteristics of the deep subsurface ground structure around Fusa District is shown in Figure 3. Since Fusa District is located in the junction of 3 rivers (Tone River, Tega River and Kokai River), and geological time scale is also young, the aging effect in Fusa District is not expectable. Then, Fusa District differs in the geology characteristics as compared with the surrounding 3 sites for strong motion observation station (see Figure 3). Therefore, in Figure 4, the ground motion characteristics due to the previous 3 earthquakes (the 20 off the Pacific coast of Tohoku Earthquake (M W 9.0), the 20 off Ibaraki Prefecture Earthquake (M J 7.6) and the 20 Hamadori Fukushima Prefecture Earthquake (M J 7.0)) at the sites of the strong motion observation station is also various. 2
Ground Liquefaction Site by MLIT and JGS (20) N JMA Tone Town Site of Interest SK-net Inzai City Site of Interest K-NET Shiroi 0 Figure 2 Liquefied sites in Fusa District. 0 00 m Figure 3 Geological features around Fusa District. km JMA Tone Town 269 30.7 5.0 93 5.3 4.5 Not Recorded N Site of Interest SK-net Inzai City 376 4.3 5.5 59 8.3 4.8 Not Recorded K-NET Shiroi 486 30.7 5.5 20 4.8 4.7 97 5.7 4. PGA[Gal] PGV[cm/s] JMA Seismic Intensity Top 20/03/ 4:46; The 20 Tohoku Earthquake (MW 9.0) Mid 20/03/5:5; Off Ibaraki Prefecture Earthquake (MJ 7.6) Btm 20/04/ 7:6; Hamadori Fukushima Pref. EQ. (MJ 7.0) Figure 4 Disribution of the observed ground motion indices around Fusa District. MICROTREMOR MEASUREMENT To investigate the difference of ground shaking characteristics such as natural frequency of the ground before and after the 20 main shock, microtremor measurements were carried out at the site of interest in Fusa District (see Figure 2). At the site of interest, the boring survey is performed in November, 978 and it is the soft ground which mainly consisted of silt as shown in Figure 5 (Hata et al., 202; Okamoto and Ikeda, 2000). The later section (see SIMULATION OF RECOVERY PROCESS section) describes the result of test for the physical properties in all soil layers (.Upper fine sand layer, 2.Silt layer, 3.Silt mixed with sand layer and 4.Middle fine sand layer) shallower than engineering bedrock of Bottom fine sand layer. We used for high-sensitivity VSE-5D seismometers manufactured by Tokyo Sokushin Co., Ltd. for microtremor measurement. As the characteristics of the instrument for microtremor measurement, dimension, feature, principle and specification are shown in Figure 6, Figure 7, Figure 8 and Table. The date which measured is total 2 days before and after the 20 main shock as follows.
Depth (m) 0 SPT-N 60.6 Density (g/cm3) 0 Vs(m/s) 400 2. Fine Sand 30.78 Silt 00.69 Silt mixed with Sand 40.82 Fine Sand 240.97 Fine Sand 370 2.0 Figure 5 Ground condition at the site of interest in Fusa District. - Figure 6 Dimension of the adopted microtremor seismometers. 00 Velocity VSE-5D-(0V/m/s) 0 Sensibility Accelerometer(5V/G) Sensibility(/30 of Velocity) 0. 0.0 0. 0 00 Frequency (Hz) Figure 7 Features of the adopted microtremor seismometers. 4
Table List of specification of the adopted microtremor seismometers. Model VSE-5D- Frequency range 0. to 70Hz Mode of operation Horizontal and Vertical Max. measuring range ± 0.m/s Output Velocity (Low): 00V/m/s, (High): 000V/m/s Acceleration: 5mV/Gal Max. output voltage ± 0V Linearity 0.03% of Full scale Resolution 0-7 m/s 2 (0-5 Gal) Dynamic range Approximately 40dB Damping ration h = about 0000% Calibration coil Sensitivity: 600μ A m/s 2 (6μ A/Gal) Coil resistance: 550Ω (± 20%) Max. output voltage ± 0V Power requirements ± 5VDC Current consumption Approximately 5mA Cross axis sensitivity 0.03 G/G Sensitivity of temperature coefficient 0.0 %/ C Temperature coefficient of zero-shift 0.05 %/ C Temperature range -0 C to +50 C Allowable shock 30G (less than 0.sec) Dimension 55 69.5 72 mm Figure 8 Principle of the adopted microtremor seismometers. 20/0/29 (4 days ago) 20/03/3 (2 days later) 20/03/5 (4 days later) 20/03/7 (6 days later) 20/03/9 (8 days later) 20/03/22 ( days later) 20/03/26 (5 days later) 20/03/30 (9 days later) 20/04/03 (23 days later) 20/04/0 (30 days later) 20/04/7 (37 days later) 20/04/30 (50 days later) 20/05/28 (78 days later) 20/06/9 (00 days later) 20/08/3 (55 days later) 20/09/25 (98 days later) 202/0/07 (302 days later) 202/07/06 (483 days later) 202//8 (68 days later) 203/06/9 (83 days later) 203/2/05 (,000 days later) As shown in Photograph, we performed all measurement at night. The measurements were conducted for NS, EW and UD components. The mean of the horizontal 2 components were adopted in the following calculation of the H/V spectral ratio. The measurement was done for 00 minutes ( 63.84 seconds 36 sections), and the sampling frequency was 00Hz.
Photograph Measurement condition of micrtremor at the site of interest. EVALUATION OF MICROTREMOR H/V SPECTRA The process to calculate a microtremor H/V spectral ratio is as follows. First, a high-pass filter of 0.Hz is adopted, and 20 time sections of 63.84 seconds each were extracted from the original data considering recorded noise. Next, Fourier amplitude spectra for these 20 time sections were calculated with a Parzen window (band width of 0.05Hz). Finally, a microtremor H/V spectral ratio was calculated as the average of 20 spectra. Here, the frequency range to evaluate microtremor H/V spectral ratio is from 0.2Hz to 0Hz considering the performance of the instrument for microtremor measurements (see Figure 6, Figure 7, Figure 8 and Table ). The comparison before and after the 20 main shock of the evaluated microtremor H/V spectra is shown in Figure 9. In Figure 9, in the time zone comparatively near the 20 main shock occurring, the difference in the predominant frequency of both H/V spectra is large. On the other hand, if the time passes from the 20 main shock occurring, both H/V spectra are generally in agreement. It is thought that the change of predominant frequency expresses the recovery process of the shear modulus (shear wave velocity) of the subsurface layers shallower than the engineering bedrock. Figure 0 plots time change of the predominant frequency of the evaluated H/V spectra after the 20 main shock based on the predominant frequency of the evaluated H/V spectra before the 20 main shock. In Figure 0, we can understand that ratio of the predominant frequency before and after the 20 main shock is approaching.0 over about 200 days. That is, it suggests that the shear modulus of the subsurface layers is recovered after the 20 main shock for about 200 days. We considered that the recovery process of the shear modulus was dependent on dispersion of excess pore water pressure in the subsurface ground based on mainly silt layer. In the following section, in order to confirm our consideration, we approached simulation of change of the predominant frequency by performing effective stress analysis during the 20 main shock and its consolidation analysis after the 20 main shock. SIMULATION OF RECOVERY PROCESS Effective stress analysis during the 20 main shock and its consolidation analysis after the 20 main shock were performed to understand how dispersion of excess pore water pressure after the 20 main shock affected the change of the predominant frequency (natural frequency) before and after the 20 main shock. To investigate the influence of the dispersion of excess-pore-water-pressure, after the end of the dynamic analysis due to the 20 main shock, we performed the consolidation analysis based on static condition with the dispersion of excess-pore-water-pressure. A -D finite element model for the horizontally layered ground at the site of interest in Figure was constructed based on the in-situ and laboratory test results. Table 2 shows the parameters of 6
0 0 0 0 Microtremor H/V Spectra 20/0/29 (4 days ago) 20/0/29 (4 days ago) 20/0/29 (4 days ago) 20/0/29 (4 days ago) 20/03/3 20/03/5 20/03/7 20/03/9 (2 days later) (4 days later) (6 days later) (8 days later) 0. 0. 0. 0 0. 0. 0 0. 0. 0 0. 0 0 0 0 0 Microtremor H/V Spectra 20/0/29 (4 days ago) 20/0/29 (4 days ago) 20/0/29 (4 days ago) 20/0/29 (4 days ago) 20/03/22 20/03/26 20/03/30 20/04/03 ( days later) (5 days later) (9 days later) (23 days later) 0. 0. 0. 0 0. 0. 0 0. 0. 0 0. 0 0 0 0 0 Microtremor H/V Spectra 20/0/29 (4 days ago) 20/0/29 (4 days ago) 20/0/29 (4 days ago) 20/0/29 (4 days ago) 20/04/0 20/04/7 20/04/30 20/05/28 (30 days later) (37 days later) (50 days later) (78 days later) 0. 0. 0. 0 0. 0. 0 0. 0. 0 0. 0 0 0 0 0 Microtremor H/V Spectra 20/0/29 (4 days ago) 20/0/29 (4 days ago) 20/0/29 (4 days ago) 20/0/29 (4 days ago) 20/06/9 20/08/3 20/09/25 202/0/07 (00 days later) (55 days later) (98 days later) (302 days later) 0. 0. 0. 0 0. 0. 0 0. 0. 0 0. 0 0 0 0 0 Microtremor H/V Spectra 20/0/29 (4 days ago) 20/0/29 (4 days ago) 20/0/29 (4 days ago) 20/0/29 (4 days ago) 202/07/06 202//8 203/06/9 203/2/05 (483 days later) (68 days later) (83 days later) (,000 days later) 0. 0. 0. 0 0. 0. 0 0. 0. 0 0. 0 Frequency (Hz) Frequency (Hz) Frequency (Hz) Frequency (Hz) Figure 9 Comparison of the evaluated microtremor H/V spectra before and after the 20 main shock.
Ratio of st Peak Frequency..0 0.9 0.8 0.7 0.6 20/0/29 0.5 20/03/3 0.4-4 -27-3 0 00 000 Before the main shock The 20 main shock Consolidation analysis after the 20 main shock The Hamadori EQ. (20/04/ Mj7.0) 20/04/0 0 3 20/04/7 203/2/05 After the main shock Time after the 20 Tohoku Earthquake (day) Figure 0 Distribution of the predominant frequency of the microtremor H/V apectra with time after the 20 main shock, and the simulation result by use of the FEM analysis. L:FineSand 3. m Ground Water Table (Surface) Drainage Boundary Condition Multiple Point Constraints (MPC) were given for 2 nodes in a same depth to have the same movement. L2: Silt 26.2 m Base (Input Motion Site) Viscous Boundary Condition L3: Sand mixed L3: with Sand L4: Fine Sand 3.2 m 4.0 m Figure The -dimensional finite element model of the ground of interest. 200 PGA= 92 Gal Acc. (Gal) 0 [N45 E] JMA Tone Town -200 0 0 20 30 40 50 60 70 80 90 00 0 20 Time (s) Figure 2 The time history of input earthquake motion. the ground models for the analysis. The strain space multiple mechanism model (Iai et al., 20) was used for all the soil layers. A detailed description of the parameters of the multiple shear mechanism model can be found in Iai et al. (20). In Table 2, parameters such as t, V s, c,, m and H max were determined directly from the above-mentioned in-situ and laboratory tests (Abiko City, 203a; 203b; 203c; 203d, Hata et al., 202). On the other hand, model parameters related to excess pore water 8
Table 2 Ground characteristics and model parameters based on the soil tests at the target site. L L2 L3 L4 Geological classification Fine Sand Silt Silt mixed with Sand Fine Sand Wet unit weight t (kn/m 3 ) 7.4 6.6 7.8 9.3 Poisson's ratio 0.33 0.33 0.33 0.33 Shear wave velocity V S (m/s) 30 00 40 240 Confining pressure m G 0.5 0.5 0.5 0.5 dependence coefficient m K 0.5 0.5 0.5 0.5 Porosity n 0.54 0.7 0.6 0.45 Coefficient of permeability k (m/s) 8.8 0-5 3.2 0-9 4.7 0-9 7.3 0-7 Internal friction angle (deg.) 35.9 34.3 39.5 42.3 Phase angle m (deg.) 28.0 28.0 30.0 33.0 Cohesion c (kpa).2 6.2 3. 0.0 Upper limit for hysteretic damping H max 0.246 0.44 0.82 0.260 cm d 0.55 0.35 0.40 0.55 r dc.40.38.39.40 r d 0.45 0.35 0.40 0.45 q.00.00.00.00 Parameters for dilatancy q 2 0.50 0.50 0.50 0.50 lk 2.00 2.00 2.00 2.00 r k 0.50 0.50 0.50 0.50 S 0.005 0.005 0.005 0.005 c.73.7.72.73 20/ 03/2 20/ 03/3 20/ 03/5 20/ 03/7 20/ 03/9 20/ 03/22 20/ 03/26 20/ 03/30 20/ 04/03 20/ 04/0 EPWPR (-).0 0.8 0.6 0.4 0.2 Figure 3 Distribution of the excess-pore-water-pressure ratio after the 20 main shock. 0.0
pressure ( d cm, r dc, r d, q, q2, lk, r k, S and c ) were determined so that the liquefaction resistance based on the laboratory tests can be reproduced in the simulation of the tests. Regarding the boundary condition, as shown in Figure, the bottom of the model was set to be viscous, and Multiple Point Constraints (MPC) were given for 2 nodes at a same depth to have the same displacement. We performed -D effective stress analyses for preliminary case. Observed acceleration time histories for horizontal components (N45 E direction: Abiko City, 203b) at JMA Tone Town site (see Figure 4) shown in Figure 2 was input at the viscous bottom boundary. Other numerical conditions are summarized as follows. The time step interval of 0.00 seconds was adopted to ensure numerical stability. The time integration was done using the Wilson- method ( =.4), and a small number of Rayleigh damping ( = 0, = 0.0038) was used to ensure the stability. The initial conditions were obtained by performing a static analysis with gravity using the same constitutive model used for the seismic response calculation. As a result of the analysis, distributions of excess-pore-water-pressure ratio (EPWPR) at 0 calculation step of 20/03/2, 03/3, 03/5, 03/7, 03/9, 03/22, 03/26, 03/30, 04/03 and 04/0 (see MICROTREMOR MEASUREMENT section) were obtained as shown in Figure 3. In Figure 3, we can understand that the excess-pore-water-pressure ratio is decreasing with the time progress after the 20 main shock. In the each time section after the 20 main shock, we calculated the natural frequency of the -D ground model with dispersion of the excess-pore-water-pressure, and added to Figure 0. In Figure 0, from 4 days ago to 0 days ago, we can recognize the similarity of the calculation curve to the measured values. However, in other time sections, difference between the calculation curve and the measured values is comparatively large. As one of the reasons for this difference, in this preliminary analysis, taking only the 20 main shock into consideration as an input earthquake motion is mentioned. In particular, we have to consider the 20 off Ibaraki Prefecture Earthquake and the 20 Fukushima Hamadori Earthquake (see Figure ) as input earthquake motions in addition to the 20 off the Pacific coast of Tohoku Earthquake. SUMMARY AND CONCLUSIONS During the 20 off the Pacific coast of Tohoku Earthquake (M W 9.0), damage due to soil liquefaction was reported in Fusa District, Abiko City, Chiba Prefecture, where located in the junction of 3 rivers (Tone River, Tega River and Kokai River). In this study, to understand the recovery process of the shear modulus with dispersion of the excess-pore-water-pressure, microtremor measurements at the site of interest were carried out periodically for 000 days after the 20 main shock. Then, the -D effective stress analysis during the 20 main shock and its consolidation analysis after the 20 main shock based on the results of in-situ and laboratory tests were conducted to understand how dispersion of excess pore water pressure after the 20 main shock affected the change of the predominant frequency (natural frequency) before and after the 20 main shock. The results of the study at current stage can be summarized as follows. ) According to the comparison results of microtremor H/V spectra, the distribution of ratio of the predominant frequency before and after the 20 main shock is approaching.0 for about 200 days. It suggests that the shear modulus of the subsurface layers of the ground is recovered after the 20 main shock for about 200 days. 2) Using a ground model constrained by the in-situ and laboratory test results, the D-FEM analysis of the ground of interest was carried out. According to the results, the recovery process of the shear modulus was dependent on dispersion of excess-pore-water-pressure since in the subsurface ground was based on mainly silt layer. However, when only the 20 main shock is adopted as an input earthquake motion, we cannot fully simulate the recovery process of the shear modulus of the ground. As a future study, we will consider the 20 off Ibaraki Prefecture Earthquake and the 20 Fukushima Hamadori Earthquake as input earthquake motions in addition to the 20 off the Pacific coast of Tohoku Earthquake. 0
REFERENCES Abiko City (203a) Disclosure data from examination committee for liquefaction countermeasure in Abiko City, The 2nd committee meeting on January 28, 203, 66p. Abiko City (203b) Disclosure data from examination committee for liquefaction countermeasure in Abiko City, The 3rd committee meeting on March 26, 203, 75p. Abiko City (203c) Disclosure data from examination committee for liquefaction countermeasure in Abiko City, The 4th committee meeting on May 28, 203, 68p. Abiko City (203d) Disclosure data from examination committee for liquefaction countermeasure in Abiko City, The 5th committee meeting on August 22, 203, 74p. Arai, H., Sekiguchi, T. and Tokimatsu, K. (2006) Recovery process of subsurface Vs values at strong gtound motion stations in Ojiya after the 2004 Niigata-ken Chuetsu Earthquake, Proc. of the 2th Japan Earthquake Engineering Symposium, 44-47, Tokyo, Japan. Harada, K. (2004) Kirisho swamp and flood control in Abiko, Abiko from your photo albums (in Japanese). Hata, Y., Nozu, A., Ichii, K., Maruyama, Y. and Sakai, H. (202) Evaluation of site effects in Fusa District, Abiko City, Chiba Prefecture, Proc. of the GeoKanto202, 9(A0023), Tokyo, Japan. Houlsby, G. T. and Wroth, C. P. (99) The variation of shear modulus of a clay with pressure and overconsolidation ratio, Soils and Foundations, 3(3): 38-43. Koarai, M., Nakano, T., Otoi, K., Une, H., Kawamoto, R. and Daigo, K. (20) Liquefaction damage of the Great East Japan Earthquake and utilization of time serial geographic information, Journal of the Geospatial Information Authority of Japan, (22), 27-4. Koarai, M. (203) Liquefaction damages of the East Japan Great Earthquake and evaluation of liquefaction risk using geospatial information, GSJ CHISHITSU NEWS, 2(2), 36-366. Koseki, J., Miyashita, Y., Deng, J. and Araki, H. (202) Liquefaction along Tone river basin in Japan, caused by the 20 off the Pacific coast of Tohoku Earthquake, Proc. of 2nd International Conference on Performance- Based Design in Earthquake Geotechnical Engineering, 2-30, Taormina, Italy. Ministry of Land, Infrastructure, Transport and Tourism, Kanto Regional Development Bureau and The Japanese Geotechnical Society (20) Actual condition elucidation of the liquefaction phenomenon of the Kanto district by the 20 off the Pacific coast of Tohoku Earthquake, Survey Slip No.083 in appendixes: Abiko City-, Tone Town- and Inzai City-4. Nakayama, Y., Nishida, K., Nishigata, T. and Inoue, K. (998) Decrease and recovery in deformation modulus of clay subjected to repeated loading, Jour. of JSCE, (589/III-42), 3-40. Okamoto, K. and Ikeda, H. (2000) On the origin of silt-bed channel segment in the lower Tone River, Bulletin of the terrestrial environment research center the University of Tsukuba, (), 35-4. Tokimatsu, K. and Hosaka, Y. (986) Effects of sample disturbance on dynamic properties of sand, Soils and Foundations, 26(): 53-64.