1229/11/A

EARTHQUAKE RESPONSE OF ANCIENT FIVE-STORY PAGODA

STRUCTURE OF HORYU-JI TEMPLE IN JAPAN

Koji NAKAHARA2 , Toshiharu HISATOKU1, Tadashi NAGASE2

and Yoshinori TAKAHASHI2

ABSTRACT

The architectural style of the five-story pagoda was brought into Japan from India via China in the
6th century. In Japan no pagodas have ever suffered serious damage from earthquakes. Even in
the Great Hanshin-Awaji Earthquake, there were no reports of serious damage to wooden pagodas
in Hyogo, Kyoto and Nara. This fact must be scientifically explained. Since the end of the Meiji
era, many researches have studied the earthquake resistance of five-story pagodas. And several
factors of earthquake resistance of them has been pointed out , such as friction damping and
sliding effect of the wooden joints , base isolation effects , balancing toy effects of deep eaves ,
bolt fastening effect of the center column and so on. This paper attempts to evaluate and examine
the earthquake behaviors of the ancient Japanese 5-story pagoda of Horyu-ji Temple through the
response analyses.

1. INTRODUCTION

The architectural style of the five-story pagoda was

introduced with Buddhism from India via China
around the mid 6th century. During the years since
then, about 1,300 years, many five-story pagodas
encountered several huge scale earthquakes. There
exist, however, no historical documents that report
any toppling incidents of five-story pagodas except
some damages in the ornamental element called
kurin in Japanese in the top structure.
Even in the Great Hanshin-Awaji Earthquake
Disaster of 1995 inflicted by an earthquake that
registered a seismic intensity of M 7.2, there had
been no reports on major damages to wooden
pagodas close to the affected areas of Hyogo
Prefecture or old pagodas in Kyoto and Nara.
These five-story pagodas are a mysterious existence
for modern people in many respects. To understand
one aspect of it, we conducted a simulation analysis
of why the five-story pagoda of Horyuji Temple ,
shown in Figure. 1 , still standing in Ikaruga-no-sato Figure. 1: Photograph of the five story
in Nara has been so resistant to earthquakes. d

1
Head Office, Takenaka Corporation, 2Building Design Department, Takenaka Corporation,
3-10, 2-Chome, Nishi-Hom-Machi, Nishi-Ku, Osaka, 550-0005, Japan Email: nakahara.kouji @takenaka.co.jp

1
 2. THE FORM AND STRUCTURE OF THE FIVE-STORY PAGODA OF HORYUJI TEMPLE

Five-story pagodas were built to enshrine Buddha’s ashes (the skeletal remains of Buddha) and said to inherit
the form of stupa, a style of tombs in ancient India. The five-story pagoda of Horyuji , regarded as the oldest
existing wooden pagoda in Japan , was rebuilt around A.D. 711 after the original one was lost in a fire. Figure. 2
shows plans and sectional views of the pagoda. It boasts a total height of 32.55 m from its top to the top of its
podium or 107.44 shaku, an older unit of length for Japanese.The plan of its structure is square with the length
of its side, 5.45 m in the first through forth stories and 3.64 m in the fifth story. The first story is surrounded by
a structure called mokoshi ( an extra eave ), an addition to the main structure that is covered with lean-to roofs.
A center column supports its top structure, sorin.
The features of the five-story pagoda of Horyuji are described in the four points listed below. Further two more
impressive features are the pliant impression suggestive of a flexible nature of its structure and deep eaves.
1. The ratio of the total height to the width of the main structure in the first story is 5.1.
2. The ratio of the width of the main structure in the top story and that of the first story is 0.51.
3. The ratios between the lengths of eaves and the widths of the main structure are 2.2 in the first story
to 3.0 in the fifth story.
4. The ratio of sorin to the total height is 1:3.4.
The following six points can be listed as its structural features:
1. The main structural elements consist of wood.
2. There are many joints or connections such as the “kumimono” or complex joints connecting many
wood members.
3. A framework in which each story is independent and no column ties them together.
4. The center column supports the ornamental structure on the top independently of the main structure.
5. The columns in the first story are not tied down to the foundation.
6. Its natural periods are around 1 second, and these are rather long considering the height of its
structure.
In the original structure, the center column was buried in a deep hole in the ground, but it now stands on a base
stone in the podium.

5.66m
3.14m

Sorin

Kumimono 9.69m
5th 14.27m
(joints)
32.55m 2nd story 5th story
Shin -bashira
4th 6.42m
(center-colum)

3rd

2nd main
structure

extra eave
1st
Base stone 10.85m

Section 1st story

Figure. 2: Plan and sectional views of the pagoda

2
 3. ANALITYCAL FOCUS OF THE STRUCTURAL ELEMENTS CONTRIBUTING TO THE
EARTHQUAKE RESISTANCE

Many scientific researches have been conducted on the earthquake resistance of five-story pagodas since the end
of the Meiji era (around A.D. 1900). Dr. Muto[3] thought the friction damping effect of the wooden joints was
an important factor in making them earthquake resistant. After Dr. Ishida[2], the center column acts as a bolt
fastening the whole structure and adding a restrainsing effect of shearing deformations among individual stories.
According to the analyses conducted by Tanabashi[5], the factors increasing the resistance of the structure were
the scale effect of the five-story structure , a characteristic of flexible structure and the wood joints’ capacity for
allowing plastic deformations through slipping or gaps in them. Dr. Ueda[6] considered that each structurally
independent stories are mounted on top the other was able to allow each one to act like a balancing toy,
cancelling the inertia force of each story out among them. And Dr. Omori proposed that the compound
penduilam system , the center column and the main structure , gives TMD effect after researches of pagodas in
Nikko-ji Temple and Senso-ji Temple.
Based on this background, seven factors , listed below and illustrated in Figure. 3, have been considered in the
analysis.

1. Sliding between the base stones and columns contributing the earthquake resistance (base isolations)
2. Slipping and gaps in the wooden joints
3. Friction damping effect of wooden joints
4. Balancing toy effect ( due to deep eaves )
5. Oscillation of the whole structure like a snake dance
6. Collision between the center column and the main structure , making a bolt effect
7. Center column TMD effect

Slipping
K ƒ
ƒ ^ _ ƒ
ƒ “ pƒ [
Friction

Sliding
Š ‚è x ³

Base Isolation Slip joint Friction damper

Balancing toy s TMD

‚ l‚
‚ c

Snake dance Shin-bashira Tuned Mass Damper

[Balancing toy]

Figure . 3: Vibration control devices of the pagoda

3
 4. ANALISYS MODEL

4.1 MODEL OF FIVE-STORY PAGODA

The structural model used in this analysis is the two
dimensional frame shown in Figure. 4, and the
column and beam members are assumed to be Shin-basira
elastic. The mass of each story is concenntrated at
several node points. The stiffness of the model frame
are given as a result of referring to the Micro Tremor Surrounding
Measurement data by Dr. Uchida et al.[7], and the frame
vibration period and modes of the first, second and
third are shown in Table 1 and Figure. 5. The
damping ratio of the column and beam members are Friction
assumed to be at 4% to critical based on the Damper
experimental observation by Dr. Uchida et al.[7] Kannuki
Also, the direction of analysis is set to be 45degree effect
to the main structure , according to the report[8] by
Dr. Yamabe and Dr. Kanai , the direction of the
Base
principal vibration of pagodas is considered
Isolation
45degree to the main structure . The earthquake
input motion is the N-S component of the 1995
Hyogo-ken Nanbu Earthquake, observed at Kobe
station of Japan Meteorological Agency. The peak Figure . 4: Analysis model of the pagoda
ground accelaration of the input excitaion is
818cm/sec2. The sway angle of each story at which
the frame collapses is assumed to be 1/50 rad. 5

Table 1: Natural periods of the pagoda 4

Natural period(sec.) 3rd 2nd
Micro-tremor[7] Analysis
1st 1.11 1.13 3 1st
story

2nd 0.42 0.49

3rd 0.24 0.33 2

4.2 VIBRATION CONTROLE DEVICES

1
4.2.1 Slip joint and friction damper
The whole amount of sliding and friction effects are
attributed to the top parts of the columns of the
analysis model. As shown in Figure. 6, they are -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5
assumed to act as a friction damper with the participation factor
hysteresis characteristic of bi-linear frame with a 0.4 Figure . 5: Participation factor of the model
coefficient of friction within ±1.5 cm
displacements, and they are represented by a non- Q
linear spring sketched in the solid line in Figure. 6
and an equivalent damping factor. ＋Qs

4.2.2 Base isolation -1.5cm δ

Base isolation effect is also assumed by sliding and
+1.5cm
friction between bottom parts of colums at first story
and the foundation is modeled using a non-linear -Qs Qs=CfΣW
spring shown in the solid line in Figure. 6. Qs :Sliding Load
Cf : Friction coef.=0.4
4.2.3 Shin bashira W : Upper weight
The gap between the center column and the main
structure is set to be ±1 cm based on the restoration
Figure . 6: Force-displacement characteristics
report of Horyu-ji.[1] Gap elements are placed
of friction damper
between the center column and beams in the
analysys model.

4
4.2.4 Balancing toy effect
A balancing toy is equilibrated stably by gravity. When the balancing toy is excited and begins to rotate , the
restoring momemt about the point O is applied as sketched in Figure. 7 (a). When a rotation angle θ is minute
, the balancing toy effects are given by simple linear springs in the vertical direction as illustrated in Figure. 7
(b).
M = mg ⋅ b − mg ⋅ a

θ
O A k
O k
B B
A A θ0
mg
mg cos θ0
k≈
mg l sin θ0
a b l l

(a) (b)

Figure . 7: Model of balancing toy effect

5. ANALISYS RESULTS

In Figure. 8 (a), (b), the envelope of maximum displacement responses against the earthquake of each model are
shown in comparison with that of a conventional rigid frame structure model with its fundamental period of
1.1sec. Figure. 9 illustrates the envelope of maximum displacement against the earthquake of the composite
model , including all vibration control effects , in comparison with that of the rigid frame model. And Time
history of relative displacements of 3rd story and 5th story of the composite model against the earthquake at
around the principal shock are shown in Figure. 10.
As can be seen in Figure. 8, the maximum displacement response of each model that includes only one vibration
control effect is smaller than that of the rigid frame structure. The maximum response of each model , however ,
exceed the limit sway angle of 1/50 at which the frame may collapse.
Fig. 9 indicates that the maximum responses of the composite model is reduced to 56% of the maximum
response of the rigid frame model, and the response sway angles are below 1/50, proving the integrated effect of
many resistance factors that has been pointed out by researchers. Fig. 10 shows the time history of the relative
displacement responses of 3rd and 5th stories of the composite model. The amplitude of the 3rd story is larger
than that of 5th story at around the principal shock of the ground earthquake motion. This indicates that the
intermediate stories behaves like as soft stories and act as isolators.

5th 5th
TMD slip and friction
4th 4th
1/50rad.
1/50rad.
Story

Story

3rd 3rd
snake dance base isolation
2nd 2nd
rigid frame rigid frame
1st 1st

0 10 20 30 40 50 60 0 10 20 30 40 50 60
M ax. Displacement(cm) M ax. Displacement(cm)
(a) TMD effect and Snake dance effect (b) Slip and friction effect and Base isolation effect

Figure . 8: Comparison of maximum displacements of each model

5
 1/50rad.
5th 3rd
5th 15.0

relative displacement (cm)

4th
10.0
pagoda
5.0
Story

3rd
0.0
2nd rigid frame -5.0

-10.0
1st
-15.0
0 10 20 30 40 50 60 30sec 35sec 40sec
M ax. Displacement(cm) time (30sec - 40 sec)

Figure . 9: Maximum displacement Figure . 10: Time history of relative

of the composite model displacements of the composite model

6. CONCLUDING REMARKS

The simulation study of the earthquake resistance of Horyuji’s five-story pagoda proved that the pagoda has
escaped the fate of collapsing in seismic excitation through an integrated effect of many resistance factors
against earthquakes as has been pointed out by many researchers.
It should be emphasized that th mos effective factor for earthquake resistance of the pagoda may be the efforts
of the faithful who have preserved the structure for such a long period of 1,300 years.

7. REFERENCES

1. Committee of the restoration work for the national treasures of Horyuji. (1955), The report on the
restoration work for the national treasures of Horyuji , vol.13 , Japan.
2. Dr. Ishida, S. (1993), “The critical behavior of the wooden pagoda under earthquakes”, Technical
report No. 15, AIJ Kinki branch , July, pp71-85.
3. Dr. Muto, K. (1949), “Five-Story Pagodas and Earthquake Resistance”, Journal of the Disaster
Prevention 11 , Japan.
4. Dr. Omori, H. (1921), “About the seismic vibration of five-story pagodas”, Journal of Architeecture
and Building Science 415 , AIJ, Japan , pp.219-226.
5. Dr. Tanabashi, R. (1960), “Earthquake Resistance of Traditional Japanese Wooden Structure”, Special
lecture of 2WCEE , July , Japan.
6. Dr. Ueda, A. et al (1996), Why five story pagodas hardly collapsed, Shincho-sha., Japan.
7. Dr. Utida, A. , Kawai, N. , Maekawa, H. (1996), “Dynamic Characteristics of Traditional Wooden
Building (Part2 : Micro Tremor Measurement on Horyu-ji Pagoda) ”, Summaries of Tech. Papers, 1996
Annual Meeting, AIJ , Japan.
8. Dr. Yamabe, K. , Dr. Kanai, K. (1988), “Study on the Aseismic Properties of the Gozyunotos
(pagodas)”, Journal of the college of the Industrial Technology , Nihon Univ , Japan , pp.91-110.