ground motion characteristics duringthe 2011 off the pacific coast of tohokuearthquake

7/27/2019 GROUND MOTION CHARACTERISTICS DURINGTHE 2011 OFF THE PACIFIC COAST OF TOHOKUEARTHQUAKE

1/44

GROUND MOTION CHARACTERISTICS DURING

THE 2011 OFF THE PACIFIC COAST OF TOHOKUEARTHQUAKE

Hiroyuki Gotoi) and Hitoshi Morikawaii)

i) Assistant Professor, Disaster Prevention Research Institute,Kyoto University, Japan.

ii) Associate Professor, Department of Built Environment,Tokyo Institute of Technology, Japan.

ABSTRACT

The 2011 off the Pacific coast of Tohoku Earthquake (MW 9.0) caused

severe damage not only because of the great tsunami, but also because of

the ground motions. In the present paper, we report some details of the

source mechanism and the ground motions during the earthquake, and dis-

cuss the ground motion characteristics related to the damage in the Tsukidate

and Furukawa areas. Damage grades in Tsukidate, where the largest peak

ground accelerations were observed, are not severe in comparison to those in

Furukawa. The pseudo-velocity response spectrum for the ground motions

observed in the Furukawa area is similar to that in the JR Takatori records

for the 1995 Kobe earthquake.

1


2/44

Key words: The 2011 off the Pacific coast of Tohoku Earthquake, earth-

quake, ground motion, peak ground acceleration, pseudo velocity response

spectrum, frequency contents (IGC: A-01, C-00)

INTRODUCTION

On 11 March 2011, a large earthquake hit the eastern part of mainland

Japan. The earthquake, and particularly the great tsunami caused by it,

resulted in the death of more than ten thousand persons. Structures were

also damaged not only due to the tsunami, but also due to the ground mo-

tions, e.g., elevated bridges of the Tohoku Shinkansen (Takahashi, 2011) and

geotechnical structures (Tohata et al., 2011). Strong ground motions during

the earthquake were observed over almost the entire county of Japan. At

least 18 stations observed peak horizontal accelerations of over 980 cm/s2,

and two stations observed seismic intensities of over 6.5 on the Japan Mete-

orological Agency (JMA) scale. It should be noted that the damaged areas

do not correspond to the locations observing the large accelerations and in-

tensities.

2


3/44

Residential damage, due to the ground motions, was concentrated in the

Furukawa area of Osaki City and the Sanuma area of Tome City, which are

both located in inland areas of Miyagi Prefecture (Goto, 2011a). The seismic

intensities observed at the seismic stations located in these areas were in the

range of 5.5-6.5. On the other hand, no significant damage was reported in

the Tsukidate area of Kurihara City, where a seismic intensity of over 6.5

was observed (Goto, 2011a). This means that the damage grades are not

consistent with the JMA intensities.

In this article, we review the source process and the ground motions

during the earthquake, and then discuss the ground motion characteristics

related to the damage in the Tsukidate and Furukawa areas.

CHARACTERISTICS OF SOURCE PROCESS

The earthquake, officially named The 2011 off the Pacific coast of To-

hoku Earthquake by JMA, occurred at 14:46 (JST, GMT+9) on 11 March

2011. The hypocenter was located in the offshore region of the eastern part

of mainland Japan, and the depth was 24 km. The Moment magnitude (MW)

was 9.0 (Hirose et al., 2011), making it the largest earthquake observed in

3


4/44

Japan since 1900, and the 4th largest earthquake in the world since 1900. Fig-

ure 1 shows the distribution of aftershocks occurring within 24 hours after

the main shock, which almost covers the fault zone of the main shock. This

implies that the dimensions of the fault zone were about 500 km in length

along the coast line by about 200 km in width. Figure 1 also shows the

centroid moment tensor (CMT) solutions estimated by JMA (Hirose et al.,

2011) and F-net organized by the National Research Institute for Earth Sci-

ence and Disaster Prevention (NIED). They indicate that the mechanism was

a reverse fault with a compressional axis in an almost east-to-west direction,

which is estimated as N103E by JMA and N110E by NIED, respectively.

The location of the seismic fault and the mechanism imply an inter-plate

earthquake on the plate boundary between the North American plate and

the Pacific plates.

The Headquarters for Earthquake Research Promotion had warned of

the occurrence of an earthquake in offshore Miyagi Prefecture with a prob-

ability of 99% and a scale of MJMA 7.5 within 30 years. The event was

expected to be similar to Miyagi-oki earthquake (MW 7.6) of 1978. However,

the events corresponding to the Tohoku earthquake were not evaluated. The

earthquake ruptured over several segments, which had been evaluated as

4


5/44

independent earthquakes.

Figure 2 shows the distribution of seismic intensities on the JMA scale.

Intensities over 6.5 were recorded at K-NET MYG004, located in Kurihara

City of Miyagi Prefecture, and at KiK-net TCGH16, located in Haga Town

of Tochigi Prefecture. Intensities over 6.0 were widely recorded in Miyagi,

Fukushima, Ibaraki, and Tochigi Prefectures (see Figure 1) (Hoshiba et al.,

2011). The observation of an intensity over 6.5 occurred for only the second

time since the 1995 Kobe earthquake in Japan; the other was the record

in Kawaguchi during the Mid-Niigata Prefecture earthquake of 2004. The

intensity distribution is not simply attenuated from the east coast, but also

varies strongly along the coast line. This implies that the source rupture

process was not spatially uniform.

Figure 3 shows the envelopes of the borehole accelerograms observed

by KiK-net (NIED). In order to highlight the high frequency components,

the accelerograms are filtered in the range of 0.5-10 Hz, and the envelopes

are averaged with moving windows of 2.0s. Stations were selected to be al-

most parallel to the strike direction of the seismic fault estimated by JMA

(N193E). First and second wave groups, with an interval of about 50 sec-

onds, were mainly observed at the northern stations, e.g., wave groups being

5


6/44

observed at the earliest time at IWTH05. The wave groups are not clear at

the southern stations, but a third wave group was mainly observed at the

southern stations, e.g., wave groups being observed at the earliest time at

TCGH13. This implies that the rupture process of the earthquake involved

at least three significant events radiating high-frequency energy. The first

two pulses mainly affected the ground motions observed in Iwate and Miyagi

Prefectures, while the third pulse affected the ground motions observed in

Ibaraki and Tochigi Prefectures.

Several research groups have reported on the source process during the

earthquake. GPS data and low-frequency ground motions imply that a large

slip region is located on the shallower side of the seismic fault (e.g. Miyazaki

et al., 2011; Suzuki et al., 2011; Yoshida et al., 2011). They estimated a max-

imum slip of about 35-50 m, which is almost in the same order as the geodetic

data observed just above the hypocenter (Sato et al., 2011). On the other

hand, ground motions corresponding to the frequency components higher

than 0.1 Hz imply that several strong motion generation areas (SMGAs) are

located on the deeper side of the seismic fault (e.g. Asano and Iwata, 2011;

Kurahashi et al., 2011). A high-frequency radiation area is also estimated by

the back-projection method to be located on the deeper portion of the fault

6


7/44

(e.g. Ishii, 2011; Wang and Mori, 2011; Zhang et al., 2011). This implies

that the locations of seismic wave radiation areas depend on the frequency

band (Koper et al., 2011); the high-frequency radiation is generated along

the deeper side, while low-frequency radiation and a large slip are generated

on the shallower side. The full rupture process, considering both the high-

and low-frequency radiation is estimated by Ide et al. (2011).

The ground motions observed in Miyagi and Iwate Prefectures consist of

two major wave groups with an interval of about 50 seconds (see Figure 3).

The origins are estimated to be on the deeper side of the hypocenter, and

to be located close to each other (Asano and Iwata, 2011; Kurahashi et al.,

2011). This means that a time lag of about 50 seconds was required between

the events. Goto et al. (2011c) suggest that reflection waves from the free

surface affected the generation of the latter event.

CHARACTERISTICS OF GROUND MOTIONS

Ground motions during the earthquake were observed by several seismic

networks organized by NIED, JMA, Port and Airport Research Institute

(PARI), Building Research Institute (BRI), National Institute for Land and

Infrastructure Management (NILIM), and other organizations. Figures 4 and

7


8/44

5 show the distributions of peak ground accelerations (PGA) and peak ground

velocities (PGV), respectively, in the horizontal components calculated from

the available records of the ground motion, respectively. PGA values do

not simply attenuate from the east coast; they vary strongly along the coast

line, similar to the seismic intensity distribution. Large PGAs were mainly

observed in the regions of 1) Miyagi Prefecture, and 2) Tochigi and Ibaraki

Prefectures, which are close to the locations of the SMGAs. Si et al. (2011)

compared the observed PGA and PGV to the ground motion prediction

equation by Si and Midorikawa (1999), which was developed using data from

earthquakes with MW 5.8-8.3. They summarized that PGAs observed within

100 km of the seismic fault are consistent with the equation for MW 9.0,

whereas PGVs within 300 km are weaker than that.

Table 1 lists the top 8 records of PGA in horizontal components, and Fig-

ure 6 shows the accelerograms of the top 4 PGA records: K-NET MYG004,

K-NET MYG012, PARI Onahama-G, and K-NET IBR003. The largest

PGA, 2765 cm/s2, was observed at K-NET MYG004. It is almost three times

larger than the gravitational acceleration. Even the second largest PGA ob-

served at K-NET MYG012 is about twice larger than the gravitational ac-

celeration. K-NET MYG004 and MYG012 are located in Miyagi Prefecture,

8


9/44

while PARI Onahama-G and K-NET IBR003 are located in Fukushima and

Ibaraki Prefectures, respectively (see Figure 4). As discussed in the previ-

ous section, waveforms of K-NET MYG004 and MYG012 contain two wave

groups because of the two SMGAs located downdip of the hypocenter. On

the other hand, waveforms of PARI Onahama-G and K-NET IBR003 contain

one wave group.

The frequency contents of the ground motions are discussed via the

pseudo-velocity response spectrum (pSv), which is defined by dividing the

acceleration response spectrum (Sa) by angular natural frequency :

pSv(T) =1

Sa(T) =

T

2Sa(T) (1)

The pseudo-velocity response spectrum gives the same information as plot-

ting the acceleration response spectrum when the oblique line has a gradient

of 1/2, a so-called tri-partite plot (e.g. Chen and Scawthorn, 2002). It is

not theoretically equal to the velocity response spectrum (Sv), whereas sta-

tistically it is very close, except for the long natural periods (Hudson, 1962).

The acceleration and displacement response spectra are flat in the shorter

and longer natural periods, respectively, while the pseudo velocity response

spectrum gives a clear peak around the transition periods. We focus on the

peak value and the corresponding period of pSv to discuss the frequency

9


10/44

contents.

Figure 7 shows the pseudo velocity response spectra of the K-NET

MYG004 and the MYG012 records. The lines corresponding to each record

are EW and NS components. The damping coefficient is set to be 0.05. The

response spectra are compared to the JR Takatori records taken during the

1995 Kobe earthquake, which were observed in the damage belt, the area

with a seismic intensity of 7 on the old JMA scale. In the area, more than

30% of the wooden houses were reported as having collapsed on the basis

of field reconnaissance (e.g. Kawase, 1996; JMA, 1997). pSv values of K-

NET MYG004 and MYG012 exceeded the JR Takatori records at periods

shorter than 0.5s, whereas those in the range of 1.0-2.0s were almost 1/4 of

those of the JR Takatori records. This means that the frequency content of

the records with the large PGAs is not similar to that of the JR Takatori

records. Figure 8 shows the distribution of the peak periods of the pSv ex-

ceeding 50 cm/s, which are categorized into either periods of either less than

or greater than 1.0s. Most of the records show peak periods of less than 1.0s.

This means that the high-frequency radiation less than 1.0s was significant

during the earthquake and caused the large PGAs. On the other hand, sev-

eral peak periods of over 1.0s were observed, although the PGAs were not

10


11/44

very large.

Figure 9 shows the distribution of the peak values and periods of pSv in

its maximum component. The color of the symbols highlights the periods in

1.0-2.0s in order to emphasize the records similar to the JR Takatori records.

Large pSv values with periods in the range of 1.0-2.0s are concentrated in

the northern part of Miyagi Prefecture. We focus on the typical records with

the large pSv values and the period range of 1.0-2.0s, JMA Furukawa and

K-NET TCG006, as indicated in Figure 9. PGAs of JMA Furukawa and K-

NET TCG006 are 568 cm/s2 and 421 cm/s2, respectively, which are almost

half of the gravitational acceleration. Figure 10 shows the pSv of the JMA

Furukawa and the K-NET TCG006 records compared to the JR Takatori

records. Both peak periods are almost similar to the peak period of the JR

Takatori records. The peak value of JMA Furukawa is almost similar to that

of the JR Takatori records, whereas that of K-NET TCG006 is a little less.

In the following section, we discuss the differences in structural damage

around the seismic stations of K-NET MYG004 and JMA Furukawa, which

are typical of the stations observing the largest PGA and the significant

characteristics of pSv, respectively.

Long-duration ground motion is also a typical characteristic of this

11


12/44

earthquake, because the generation of the seismic waves continued about

160s (Suzuki et al., 2011) due to the rupture propagation of such a large

scale seismic fault. This caused liquefaction damage, especially in the Tokyo

Bay area (Tohata et al., 2011). Figure 11 shows accelerograms at K-NET

CHB008 and CHB024 located in the Tokyo Bay area. An acceleration of over

50 cm/s2 acceleration continued for about 50s at CHB008. CHB024 shows a

drastic change in amplitude, a reduction in high-frequency components, and

spike-like phases after about 100s, which all corresponds to the characteris-

tics of the records on the liquefied ground (e.g., Holzer et al., 1989; Iai et al.,

1995).

DAMAGES DUE TO GROUND MOTIONS IN TSUKIDATE

AND FURUKAWA AREAS

We focus on the detailed damages due to the ground motions in two

typical areas of Miyagi Prefecture, Tsukidate and Furukawa. The largest

PGA site, K-NET MYG004, is located in the Tsukidate area, and the re-

markable site in terms of pSv, JMA Furukawa, is located in Furukawa area.

We performed damage reconnaissances not only for residential structures,

but also for Japanese shrines and temples in both areas. Focus was placed

12


13/44

on the elements common to the shrines and gravestones at temples, because

these are almost uniformly distributed in all of Japan. In the case of the

shrines, the several elements are addressed; such as torii, a Shinto gate made

of wood and/or stones, suibansha, a basin containing water for ritual pu-

rification of mouth and hands before entering the shrine grounds, komainu,

statues of lions or lion dogs that guard the shrine, honden, the Main Hall

or kami hall where the kami (god) is enshrined, and sando, the approach to

the shrine or honden, lined with trees or lanterns (Young et al., 2004). The

difference in damage patterns emphasizes the relative difference in damage

grades between the areas, although the quantitative research has been lim-

ited in the damage grade of the honden at the Japanese shrine (Yamada et

al., 2008), and for the damage ratio of the gravestones (e.g. Ishiyama, 1982;

Midorikawa and Fujimoto, 1996; Kaneko and Hayashi, 2000). Therefore, we

compare the actual damages to the honden and the gravestones with the es-

timation curves developed quantitatively in previous studies quantitatively,

and introduce patterns of the other damages developed qualitatively.

Tsukidate area

13


14/44

The Tsukidate area is located in Kurihara City, Miyagi Prefecture, which

is about 60 km north of Sendai City. There is one seismic station, K-NET

MYG004, which observed the largest PGA of 2765 cm/s2 and a seismic in-

tensity of 6.6 on the JMA scale.

Damage reconnaissance around the seismic station was performed on 2

April 2011. Figure 12 shows the location of K-NET MYG004 and the damage

along the reconnaissance trails. The trails cover the downtown of Tsukidate.

We observed a few fallen pieces of block walls, peeled off exterior faces of

stores, cracks on the columns of residential houses, broken windows, and the

caving in roads, which are marked as squares in Figure 12. There were no

collapsed or severely damaged houses along the reconnaissance trails.

One shrine, Tsukidate Hachiman Shrine, and one temple, Sorinji Tem-

ple, are located in the area. At Tsukidate Hachiman Shrine , only one ko-

mainu fell down, and the damage of the honden was categorized into D1 in

the damage rank by Okada and Takai (1999) (Figure 13). At the Sorinji

Temple, the gravestones were classified into three categories: a) collapse of

the main gravestone, b) movement of the gravestone or the other damage,

and c) no damages. The number of gravestones corresponding to each cat-

egory was a) 36, b) 111, and c) 457. Ratios of the collapsed and damaged

14


15/44

gravestones are 0.06 and 0.24, respectively. Collapsed gravestones around

the observed station in the JR Takatori records from the Kobe earthquake

were reported in the ratio of 0.70 by Midorikawa and Fujimoto (1996), which

is higher than the ratio for the Tsukidate area.

Furukawa area

The Furukawa area is located in Osaki City, Miyagi Prefecture, which is

in between the Tsukidate area and Sendai City. Furukawa Station of the JR

Tohoku Shinkansen and JR Riku-u-tosen is located at the center of down-

town. There are four seismic stations, K-NET MYG006, JMA Furukawa,

Furukawa Gas, and NILIM Furukawa. However, the records for the main

shock at Furukawa Gas were unfortunately missed (Goto et al., 2011b). K-

NET MYG006, JMA Furukawa and NILIM Furukawa observed PGAs of

583 cm/s2, 568 cm/s2, and 483 cm/s2, and seismic intensities of 6.1, 6.2, and

6.1, respectively.

Damage reconnaissance in the area around the seismic stations of K-

NET MYG006 and JMA Furukawa was performed on 4 days, namely, 14, 15,

29, and 31 March 2011. Figure 14 shows the location of the seismic stations

15


16/44

and the significant damage along the reconnaissance trails. Furukawa Gas

is located about 1 km west of the area, and NILIM Furukawa is located

about 2 km east of the area. Uplifted manholes and uneven walkways due

to liquefaction were observed in the western area of JR Furukawa Station.

Severe residential damage was observed in the area nearby the liquefaction

area. We classified the residential damage into four categories, namely, 1)

totally collapsed houses, 2) severely damaged houses, 3) raw land, and 4)

other. The total number of houses in categories 1)-3) was 37.

Three shrines and three temples are located in the area. Figures 15-17

show the details of the shrine damage to the shrines. The marked squares

indicate collapsed stone monuments, while the arrows show the collapse direc-

tions. Several stone monuments, i.e., stone lanterns and komainu, collapsed

at the shrines. Wooden honden at the Konpira and Nishinomiya Shrines,

and the suibansha of the Nishinomiya Shrine were severely damaged. The

damage to the honden at Konpira, Kanaya, and Nishinomiya Shrines was

categorized into D3, D1, and D3 in the damage rank by Okada and Takai

(1999), respectively. We classified the damage levels of the gravestones at

the Zuisenji Temple into the same categories as the Sorinji Temple, and the

number of graves corresponding to each category was a) 193, b) 64, and c)

16


17/44

93 in the half-yard of the gravestone garden. The ratios of the collapsed

and damaged gravestones are 0.55 and 0.73, respectively, which are higher

than for Sorinji Temple. Although gravestones at Ichijyoji Temple were also

severely damaged, the number of damaged gravestones at Seiganji Temple

was very few.

The aspect ratios for 10 samples randomly selected from the gravestones

at Zuisenji Temple averaged 0.44, defined by the ratio of width to height of

the stone. It is similar to the form of typical Japanese gravestones, averaging

0.40 of the aspect ratio. Several researcheres proposed a relation between the

ground motion indices and the damage ratio of the typical gravestones as a

damage curve, based on the actual damages during the historical earthquakes

(e.g., Midorikawa and Fujimoto, 1996; Kaneko and Hayashi, 2000). The

damage curve proposed by Kaneko and Hayashi (2000) is as follows;

R = [(ln PGV ln76)/0.4], (2)

where R is the damage ratio, and denotes the cumulative distribution

function of the normal distribution. The unit for the PGV is cm/s. The

observed PGVs at K-NET MYG004 and JMA Furukawa are 106 cm and

84 cm/s, respectively, and they estimate the damage ratios as 0.80 and 0.60

using Equation (2), respectively. The value is consistent with the one for

17


18/44

Zuisenji Temple, whereas it is larger than the actual damage ratio for Sorinji

Temple.

Yamada et al. (2008) proposed the damage curve for the honden cali-

brated by the damages during the 2007 Niigataken Chuetsu-oki earthquake.

It is defined by a ratio of higher than D3 in the damage rank within a 2 km

radius of the observation site;

R = [(ln PGV ln 100)/0.31]. (3)

The observed PGVs at K-NET MYG004 and JMA Furukawa lead to estima-

tions of damage ratios of 0.57 and 0.28, respectively. In the Tsukidate area,

only one sample, D1, is available. In the Furukawa area, three samples, D3,

D1, and D3, indicate the ratio of 0.67, which is larger than the estimation.

The quantitative values imply that the damage is not exactly consistent

with the observed PGVs. In addition, the damage patterns for the structures

and the shrines suggest that the ground motions in the Tsukidate area. At

least, the patterns imply that the ground motion characteristics should not

be discussed only based on PGAs and PGVs.

DISCUSSION

18


19/44

Figure 18 shows the velocity waveforms observed at K-NET MYG004

in the Tsukidate area, and K-NET MYG006 and JMA Furukawa in the

Furukawa area. The waveforms contain two wave groups with an interval of

about 50s. The peak velocities are about 100 cm/s, 90 cm/s, and 80 cm/s,

respectively. PGAs and PGVs of K-NET MYG004 are larger than the records

for the Furukawa area, whereas the indicies seem to contradict the damage

patterns in both areas. This implies that PGAs and PGVs were not indicative

of the damage grade around the seismic stations.

Figure 19 shows the pseudo response velocity spectra (h=0.05) of K-

NET MYG004, MYG006, and JMA Furukawa. Periods at the peak values

for K-NET MYG006 and JMA Furukawa are in the range of 1.0-1.5s, whereas

those for K-NET MYG004 are less than 0.5s. As shown in Figures 7 and 10,

the spectrum shapes of JMA Furukawa are more similar to those of the JR

Takatori records than K-NET MYG004. The ground motions observed at the

damaged sites during the Kobe earthquake and also in Furukawa area contain

similar frequency components. Sakai (2009) suggests that the components

in the period ranging from 1.0-1.5s affect the structural damages, and this is

consistent with the observations in the Furukawa area.

In Furukawa, the damage grades around the seismic stations were slightly

19


20/44

different; K-NET MYG006 is not located in the severely damaged region,

while JMA Furukawa is (see Figure 14). The ratios of the horizontal com-

ponent of the ground motions observed at JMA Furukawa to the horizontal

component of the ground motions observed at K-NET MYG006 during the

same event, namely H/H spectrum ratios, are calculated. Figure 20 shows

the H/H spectrum ratios during the main shock and the other significant

earthquakes: 1) a MW 7.0 intra-slab earthquake that occurred in the Pacific

plate on 26, May 2003, 2) a MW 7.1 inter-plate earthquake that occurred on

the plate boundary between the Pacific and North American plates on 16,

August 2005, and 3) a MW 7.0 intra-crust earthquake that occurred on 14,

June 2008. The spectrum ratio is calculated after filtering with a moving

window of length 0.2s in width. The periods at the peaks of the H/H spec-

trum ratios are about 0.5s for the other events, whereas those for the main

shock are larger than 0.5s, especially about 1.0s in the NS component. PGAs

at K-NET MYG006 were 268 cm/s2 during the 2003 earthquake, 151 cm/s2

during the 2005 earthquake, and 315 cm/s2 during the 2008 earthquake, re-

spectively, which are about a half the PGA during the main shock, 583 cm/s2.

The longer peak periods during the main shock imply the nonlinear behaviors

of the shallow soils. The peak period of the H/H spectrum ratios is almost

20


21/44

close to the peak period of pSv at K-NET MYG006 and JMA Furukawa.

This means that ground motions in 1.0-1.5s around JMA Furukawa were

emphasized because of the soil nonlinearity compared to those at K-NET

MYG006.

CONCLUSION

The 2011 off the Pacific coast of Tohoku Earthquake caused severe dam-

age not only because of the great tsunami, but also because of the ground

motions. The observed ground motions imply at least three significant events

generating high-frequency radiation, which are consistent with the location

of the SMGAs. They gave rise to two wave groups in the waveforms at the

northern stations, and a third wave group at the southern stations.

PGAs observed during the earthquake were extraordinarily large. The

largest PGA observed at K-NET MYG004 was about three times larger than

the gravitational acceleration. The pseudo-velocity response spectra of the

records are different from those of JR Takatori because the ground motions

with large PGA give shorter peak periods than 1.0s. On the other hand,

several peak periods over 1.0s were observed, although the PGAs were not

large. The pseudo-velocity response spectrum of the JMA Furukawa records

21


22/44

was similar to that of the JR Takatori records with the peak period of 1.3s.

We have emphasized that the spectrum shapes recorded in the damaged areas

during this earthquake and the Kobe earthquake are similar, and that the

results are consistent to those in the summary by Sakai (2009). We have

also pointed out that the ground motions in 1.0-1.5s at JMA Furukawa were

emphasized because of the soil nonlinearity.

ACKNOWLEDGMENTS

We thank the anonymous reviewer as well as the editor, Professor Akira

Murakami for their thorough comments. We also thank many organizations

for their contributions to the strong ground motion observations, namely,

K-NET, KiK-net organized by NIED, JMA, PARI, BRI, NILIM, ERI of

the University of Tokyo, and AIST. Damage investigations were conducted

in collaboration with Associate Professor Yasuko Kuwata, Associate Profes-

sor Akihiro Takahashi, Associate Professor Yoshikazu Takahashi, and the

geotechnical team of the JSCE Great Tohoku Earthquake damage reconnais-

sance members. We appreciate the information from and the discussions with

many researchers, and also the kind cooperation of the people of Furukawa

and Tsukidate.

22


23/44

References

Asano, K. and Iwata, T. 2011. Analysis of strong ground motion records from

the 2011 off Pacific coast of Tohoku, Japan Earthquake, Proc. of 2011 SSA

Annual Meeting.

Chen, W.F. and Scawthorn, C. 2002 Earthquake Engineering Handbook,

4 1319, CRC Press.

Goto, H. 2011. Characteristics of Earthquake and Ground Motion, Report of

Urgent Damage Reconnaissance by JSCE Great Tohoku Earthquake Dam-

age Reconnaissance Member, 4, 1-36. (in Japanese)

Goto, H., Morikawa, H., and Kuwata, Y. 2011. Estimation of ground motion

at Furukawa Gas during 2011 Tohoku-oki Earthquake, Proc. of 31st JSCE

Earthquake Engineering Conference, 1-010, 1-7. (in Japanese with English

abstract)

Goto, H., Yamamoto, Y., and Kita, S., Dynamic rupture simulation of the

2011 off the Pacific coast of Tohoku Earthquake: multi-event generation

within dozens of seconds, Earth Planets Space, submitted.

Hirose, F., Miyaoka, K., Hayashimoto, N., Yamazaki, T. and Nakamura,

23


24/44

M. 2011. Outline of the 2011 off the Pacific coast of Tohoku Earth-

quake (Mw9.0) -seismicity: foreshocks, mainshock, aftershocks, and in-

duced activity-, Earth Planets Space, 63, 513-518.

Holzer, T.L., Youd, T.L., and Hanks, T. C. 1989. Dynamics of liquefaction

during the 1987 Superstition Hills, California, Earthquake, Science, 244,

56-59.

Hoshiba, M., Iwakiri, K., Hayashimoto, N., Shimoyama, T., Hirano, K.,

Yamada, Y., Ishigaki, Y. and Kikuta, H. 2011. Outline of the 2011 off the

Pacific coast of Tohoku Earthquake (MW 9.0) -Earthquake Early Warning

and observed seismic intensity-, Earth Planets Space, 63, 541-545.

Hudson, D.E. 1962. Some problems in the application of spectrum techniques

to strong-motion earthquake analysis, Bull. Seism. Soc. Am., 52, 417430.

Iai, S., Morita, T., Kameoka, T., Matsunaga, Y., and Abiko, K. 1995. Re-

sponse of a dense sand deposit during 1993 Kushiro-Oki earthquake, Soils

and Foundations, 35, 115-131.

Ide, S., Baltay, A. and Beroza, G.C. 2011. Shallow dynamic overshoot and en-

ergetic deep rupture in the 2011 Mw 9.0 Tohoku-Oki earthquake, Science,

332, 14261429.

24


25/44

Ishii, M. 2011. High-frequency rupture properties of the MW 9.0 off the Pacific

coast of Tohoku Earthquake, Earth Planets Space, 63, 609614.

Ishiyama, Y. 1982. Criteria for overturning of bodies by earthquake excita-

tions, Trans. of AIJ, 317, 114.

Japan Meteorological Agency. 1997. Technical report of the Japan Meteoro-

logical Agency. (in Japanese)

Kaneko, M. and Hayashi, Y. 2000. Proposal of a curve to describe overturn-

ing ratios of rigid bodies, J. Struct. Constr. Eng., AIJ, 536, 55-62. (in

Japanese with English abstract)

Kawase, H. 1996. The cause of the damage belt in Kobe: the basin-edge ef-

fect, constructive interference of the direct S-wave with the basin-induced

diffracted/Rayleigh waves, Seism. Res. Lett., 67, 2534.

Koper, K.D., Hutko, A.R., Lay, T., Ammon, C.J. and Kanamori, H. 2011.

Frequency-dependent rupture process of the 2011 Mw9.0 Tohoku Earth-

quake: Comparison of short-period P wave backprojection images and

broadband seismic rupture models, Earth Planets Space, 63, 599602.

Kurahashi, S. and Irikura, K. 2011. Source model for generating strong

25


26/44

ground motions during the 2011 off the Pacific coast of Tohoku Earth-

quake, Earth Planets Space, 63, 571576.

Midorikawa, S. and Fujimoto, K. 1996. Isoseismal map of the Hyogo-ken

Nanbu earthquake in and around Kobe city estimated from overturning of

tombstones, J. Struct. Constr. Eng., AIJ, 490, 111-118. (in Japanese with

English abstract)

Miyazaki, S., McGuire, J.J. and Segall, P. 2011. Seismic and aseismic fault

slip before and during the 2011 off the Pacific coast of Tohoku Earthquake,

Earth Planets Space, 63, 637642.

Okada, S. and Takai, N. 1999. Classifications of structural types and damage

patterns of buildings for earthquake field investigation, J. Struct. Constr.

Eng., AIJ, 524, 65-72. (in Japanese with English abstract).

Sakai, Y. 2009. Reinvestigation on period range of strong ground motions

corresponding to buildings damage -including data of the 2007 Noto-hanto

and Niigata-ken Chuetsu-oki Earthquakes-, J. Struct. Constr. Eng., AIJ,

74, 1531-1536. (in Japanese with English abstract)

Sato, M., Ishikawa, T., Ujihara, N., Yoshida, S., Fujita, M., Mochizuki,

26


27/44

M. and Asada A. 2011. Displacement above the hypocenter of the 2011

Tohoku-Oki earthquake, Science, 332, 1395.

Si, H. and Midorikawa, S. 1999. New attenuation relationships for peak

ground acceleration and velocity considering effects of fault type and site

condition, J. Struct Constr. Eng., AIJ, No.523, 63-70. (in Japanese with

English abstract)

Si, H., Koketsu, K. and Miyake, H. 2011. Attenuation characteristics of peak

ground motion during the 2011 Tohoku, Japan, Earthquake, Proc. of 2011

SSA annual meeting.

Suzuki, W., Aoi, S., Sekiguchi, H. and Kunugi, T. 2011. Rupture process of

the 2011 Tohoku-Oki mega-thrust earthquake (M9.0) inverted from strong-

motion data, Geophys. Res. Lett., 38, L00G16.

Takahashi, Y. 2011. Structural damage of Tohoku shinkansen viaducts by

the off the Pacific Tohoku earthquake, EERI Tohoku Japan Earthquake &

Tsunami clearinghouse.

Tohata, I., Goto, H., Kazama, M., Kiyota, M., Nakamura, S., Wakamatsu,

K., Wakai, A., Yasuda, S. and Yoshida, N. 2011. Earthquake news on

gigantic Tohoku Pacific Earthquake in Japan, ISSMGE Bulletin, 5, 4666.

27


28/44

Wang, D. and Mori, J. 2011. Rupture process of the 2011 off the Pacific

coast of Tohoku Earthquake (Mw 9.0) as imaged with back-projection of

teleseismic P-waves, Earth Planets Space, 63, 603607.

Yamada, M., Hayashi, Y., Morii, T., Park, S., Kambara, H., Onishi, Y. and

Shimizu, H. 2008. Ground motion estimation in the near-source region

during the 2007 Niigata-ken Chuetsu-oki earthquake, Proc. of 14th World

Conference on Earthquake Engineering, 02-0046.

Yoshida, K., Miyakoshi, K. and Irikura, K. 2011. Source process of the 2011

off the Pacific coast of Tohoku earthquake with the combination of tele-

seismic and strong motion data, Earth Planets Space, 63, 577582.

Young, M., Young, D. and Yew, T.H. 2004. Introduction to Japanese Archi-

tecture, PERIPLUS, Singapore.

Zhang, H., Ge, Z. and Ding, L. 2011. Three sub-events composing the 2011

off the Pacific coast of Tohoku Earthquake (Mw9.0) inferred from rupture

imaging by back-projecting teleseismic P-waves, Earth Planets Space, 63,

595598.

28


29/44

Table 1: Top 8 records of PGA in horizontal components

PGA [cm/s2] Station

2765 K-NET MYG004

1970 K-NET MYG012

1913 PARI Onahama-G

1844 K-NET IBR003

1807 K-NET MYG013

1614 K-NET IBR013

1425 K-NET TCG009

1425 K-NET FKS016

29


30/44

138 140 142 144

36

38

40

100 km

CMT solution (NIED)

CMT solution (JMA)

0 25 50

depth [km]

Iwate Prefecture

Miyagi Prefecture

Fukushima Prefecture

Ibaraki Prefecture

Tochigi Prefecture

Figure 1: Aftershock distribution within 24 hours after the main shock, and

CMT solutions estimated by JMA and F-net (NIED).

30


31/44

138 140 142 144

36

38

40

100 km 4.5

5.0

5.5

6.0

6.5

7.0JMA intensity

Figure 2: Distribution of seismic intensities on the JMA scale.

31


32/44

0

100

200

300

400

500

[km]

0 50 100 150 200

time [s]

0.510Hz envelope (smoothing in 2s window)

IWTH08

IWTH13

IWTH02

IWTH04

IWTH05

MYGH06

MYGH08

MYGH09

FKSH17

FKSH18

FKSH09

FKSH11

IBRH12

TCGH13

TCGH16IBRH11

IBRH19

140 142

36

38

40

100 km

IBRH19

IBRH11

TCGH16

TCGH13

IBRH12

FKSH11

FKSH09

FKSH18

FKSH17

MYGH09

MYGH08

MYGH06

IWTH05

IWTH04

IWTH02

IWTH13

IWTH08

Figure 3: Envelopes of borehole accelerograms in EW component observed

by KiK-net (NIED). Stations are selected to be almost parallel to the strike

direction of the seismic fault.

32


33/44

138 140 142 144

36

38

40

100 km

KNET MYG004

KNET MYG012

PARI OnahamaG

KNET IBR003

0

300

600

900

PGA [cm/s/s]

Figure 4: Distribution of peak ground accelerations (PGA) in horizontal

components.

33


34/44

138 140 142 144

36

38

40

100 km 0

25

50

75

100

PGV [cm/s]

Figure 5: Distribution of peak ground velocities (PGV) in horizontal com-

ponents.

-2000

0

2000

0 50 100 150 200

acceleration[cm/s2]

time [s]

K-NET MYG004 N161oE

K-NET MYG012 N92oE

PARI Onahama-G N127oE

K-NET IBR003 N323oE

(Max. 2765 cm/s2)

(Max. 1970 cm/s2)

(Max. 1913 cm/s2)

(Max. 1844 cm/s2)

Figure 6: Accelerograms of top 4 PGA records of the maximum component.

34


35/44

10

100

1000

0.1 0.5 1 2 5

pSv(h=0.0

5)[cm/s]

period [s]

K-NET MYG004K-NET MYG012

1995 JR Takatori

Figure 7: Pseudo-velocity response spectra (h=0.05) of K-NET MYG004 and

MYG012, compared to JR Takatori records during 1995 Kobe earthquake.

The two lines plotted together for each event correspond to EW and NS

components.

35


36/44

138 140 142 144

36

38

40

100 km

peak period of pSv

(> 50cm/s)

< 1.0s

> 1.0s

Figure 8: Distribution of peak periods of pSv over 50 cm/s categorized into

periods of either less than or greater than 1.0s.

36


37/44

138 140 142 144

36

38

40

100 km

KNET TCG006 (359cm/s, 1.3s)

JMA Furukawa (423cm/s, 1.2s)

0.1 1 10

sec.

pSv (h = 0.05)

300cm/s

150cm/s

Figure 9: Distribution of peak pSv and periods. Color of the marks highlights

the periods in 1.0-2.0s.

37


38/44

10

100

1000

0.1 0.5 1 2 5

pSv(h=0.0

5)[cm/s]

period [s]

JMA FurukawaK-NET TCG006

1995 JR Takatori

Figure 10: Pseudo-velocity response spectra (h=0.05) of JMA Furukawa and

K-NET TCG006, compared to JR Takatori records during 1995 Kobe earth-

quake. The two lines plotted together for each event correspond to EW and

NS components.

0

500

0 50 100 150 200

acceleration[cm/s

2]

time [s]

K-NET CHB008 EW

(Max 157 cm/s2)

K-NET CHB024 EW

(Max 203 cm/s2)

Figure 11: Accelerograms in Tokyo Bay area

38


39/44

K-NET MYG004

Sorinji temple

Tsukidate hachiman shrine

evidence of ground motion

200m

reconnaissance trail

Osaki city hall

Figure 12: Location of seismic station, K-NET MYG004, and the significant

damages observed in Tsukidate area.

N

Figure 13: Damage to Tsukidate Hachiman Shrine. Symbol of gate is the

location oftorii, squares are komainu, symbol of house is honden, and shaded

area is sando.

39


40/44

200m

Zuisenji temple

Nishinomiya shrine

Konpira shrine

Kanaya fudoson shrine

JMA Furukawa

collapsed residence

severe damaged residence

raw land

sand boiling

uplifted manhole

K-NET MYG006

reconnaissance trail

JR Furu

Ichijyoji temple

Seiganji temple

Figure 14: Location of seismic stations, K-NET MYG006 and JMA Fu-

rukawa, and the significant damages in Furukawa area.

N

Figure 15: Damage to Konpira Shrine. Symbol of gate is the location of torii,

squares are komainuand stone monuments, symbols of house are honden and

suibansha, and shaded area is sando.

40


41/44

N

Figure 16: Damage to Kanaya Shrine. Symbol of gate is the location of torii,

squares are komainu and stone monuments, symbol of house is honden, and

shaded area is sando.

N

Figure 17: Damage to Nishinomiya Shrine. Symbol of gate is the location of

torii, vertical markers are wooden lanterns, symbol of house is honden, and

shaded area is sando.

41


42/44

0

100

200

300

0 30 60 90 120 150 180

velocity[cm/s]

time [s]

K-NET MYG004 EW

K-NET MYG006 EW

JMA Furukawa EW

(Max. 49.2 cm/s)

(Max. 91.9 cm/s)

(Max. 76.7 cm/s)

0

100

200

300

0 30 60 90 120 150 180

velocity[cm/s]

time [s]

K-NET MYG004 NS

K-NET MYG006 NS

JMA Furukawa NS

(Max. 104.8 cm/s)

(Max. 58.5 cm/s)

(Max. 81.1 cm/s)

Figure 18: Velocity waveforms of K-NET MYG004, MYG006, and JMA

Furukawa.

42


43/44

10

100

1000

0.1 0.5 1 2 5

pSv(h=0.0

5)[cm/s]

period [s]

K-NET MYG004K-NET MYG006JMA Furukawa

Figure 19: Pseudo-velocity response spectra (h=0.05) of K-NET MYG004,

MYG006, and JMA Furukawa. The two lines plotted together for each event

correspond to EW and NS components.

43


44/44

0.5

1.0

2.0

0.2 0.5 1 2 5

H/H

spectrumr

atio

period [s]

JMA Furukawa / MYG006

main shock

other events

Figure 20: H/H spectrum ratios between JMA Furukawa and K-NET

MYG006 during the main shock and the other significant events on 26, May

2003, 16, Aug. 2005, and 14, Jun. 2008.

44

ground motion characteristics duringthe 2011 off the pacific coast of tohokuearthquake

Documents