CONSTRUCTION METHODS AND EARTHQUAKE ENCOUNTER RECORDS OF JAPANESE TRADITIONAL TIMBER THREE STOREY PAGODAS

CONSTRUCTION METHODS AND EARTHQUAKE ENCOUNTER RECORDS OF JAPANESE TRADITIONAL TIMBER THREE STOREY PAGODAS Chee Siang Tan 1, Kaori Fujita 2 ABSTRACT: A study to assess the seismic resistance of traditional timber three storey pagodas in Japan, which is commonly said to be inferior to that of the five storey pagodas. First, the earthquake encounter of each pagoda and possible damage caused are determined. Then, the construction methods of the pagodas and the way they correspond to the aforementioned records are analysed. Lastly, the results are compared to that of the five storey pagodas. It is true to say that three storey pagodas too exhibit exceptional earthquake resistance, despite there being several collapse records. KEYWORDS: Three Storey Pagodas, Construction Methods, Earthquake Encounter Records, Seismic Resistance 1 INTRODUCTION 12 Five storey pagodas in Japan are said to possess exceptional earthquake resistant properties [1]. Despite having experienced multiple earthquakes of Japan Meteorological Agency (JMA) intensity scale 6 and above, no collapse were recorded apart from some minor damages. There is still no satisfactory scientific explanation for the above phenomenon, even though studies and experiments have been conducted since Meiji era [1]. On the other hand, few studies were related to the three storey pagodas, possibly owing to the existence of collapse records due to earthquake amongst the relatively abundant numbers of three storey pagodas in Japan. The authors believe that it is important to also extend the research to three storey pagodas which share the same fundamental structural properties, starting by assessing their earthquake resistance and studying their construction methods. It is hoped that the comparison studies between the five and three storey pagodas would provide some hints on the supposedly superior earthquake resistance of the former to that of the latter. 1.1 AIM The aim is to collect and study the data of the construction method and earthquake encounter record of the traditional timber three storey pagodas in Japan, in 1 Chee Siang Tan, Graduate School of Engineering, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo, 113-8656 Japan. Email: tcsiang@hotmail.com 2 Kaori Fujita, Assoc. Prof., Graduate School of Engineering, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo, 113-8656 Japan. Email: fujita@buildcon.arch.t.u-tokyo.ac.jp order to assess their earthquake resistance compared to their five storey counterparts. 1.2 RESEARCH METHOD Three storey pagodas designated as national treasures and important cultural properties by the government, 57 of them in total, all dated back from or before Edo Period are selected. Earthquakes with JMA intensity scale 5 and above that occurred around each pagoda, and had caused severe damage and numerous casualties were extracted from reference 3. These were summarised into a chronological table together with seismic damage recorded and repairs executed, based on official reports and documents (references 6 and 8). Construction method and details of each pagoda were then analysed together with the aforementioned records and compared to that of the five storey pagodas. Similar study [1] on five storey pagodas had been conducted by M. Oyama and K. Fujita in the year 2002, and will serve as the main reference and comparison material for this study. 2 EARTHQUAKE RECORDS 2.1 EARTHQUAKE ENCOUNTER RECORD The chronological table (figure 8, 9) shows the earthquakes encountered by each pagoda since their respective year of completion, together with damage and repair records. At least 80% of the three storey pagodas had encountered earthquake with JMA intensity 6 and above, with 5 encounters as the maximum. In average, the earthquake encounter is 2.54 times per pagoda, with 145 earthquake encounters in total. All three storey pagodas have encountered earthquake with JMA intensity 5 and above, with Hoki-ji, the oldest amongst

all three storey pagodas, encountering as many as 32 times. In average, the earthquake encounter is 11.68 times per pagoda, with 666 earthquake encounters in total. The encounter frequency is once every 226 years for intensity scale 6 and above, and once every 60 years for intensity scale 5 and above. It should be noted that the list is not exhaustive and precise for earthquakes occurred prior to modern recording system, thus it is possible that the above figures should register higher frequency. Earthquakes without ascertained intensity, especially those dated back before 1500s, but had caused vast damage were considered as intensity 5 and above. 2.2 SEISMIC DAMAGE Table 1: Damage recorded in three storey pagodas Name Damage Count Damage Type Construction Method Oppo-ji 1 Finial Stack-up Hiyoshi 1 Finial Stack-up Shrine Sanmyo-ji 1 Finial Stack-up Jimoku-ji 1 Collapse Stack-up Hoki-ji 1 Finial Stack-up Yakushi-ji 4 Finial, Stack-up East Pagoda inclination Henjo-in 1 Inclination Long Column Table 1 shows the seismic damage cited from reference 6 and 8. Out of the 57 pagodas studied, only 7 pagodas were recorded to have been damaged by earthquake, with 10 cases in total. The damage pattern mostly involves the breakage of metallic finials (sōrin in Japanese), and in specific cases, the collapse (Jimoku-ji) and inclination (Yakushi-ji East Pagoda) of the pagoda itself. The metallic finial connected to the central pillar is often damaged at the joint as a result of the collision between the roof and the central pillar during earthquakes. The east pagoda of Yakushi-ji has a unique feature i.e. the decorative pent roofs (mokoshi in Japanese) attached on each storey. As a result, the height of each storey increases and the columns are thus taller and slenderer, possibly attributing to its vulnerability to earthquake and even strong wind. Amongst the 7 pagodas, all are of stack-up construction method except Henjō-in. Henjō-in s damage is mainly contributed by foundation problem (subsidence) rather than structural problem and will thus be omitted. The damage caused by strong wind such as typhoon is also shown in the chronological table as comparison. Wind damage is comparatively more common, with 20 cases recorded in 14 pagodas. The damage mainly involves the roof, and occasionally the metallic finial. 2.3 REPAIRS Amongst the 36 pagodas with available official repair documents and reports, there were 253 repairs recorded in total, with average 7 repairs per pagoda and frequency of once every 96 years. However, as can be seen from the chronological table, the repairs rarely correspond to the earthquake encounters. Decay and re-roofing are commonly cited as reason for the repairs. Most pagodas were dismantled and repaired after the Act on Protection of Cultural Properties took effect in 1950. 3 CONSTRUCTION METHODS 3.1 CENTRAL PILLAR (SHIN-BASHIRA) All pagodas in Japan are built with a unique long column that passes through the centre of the pagoda, supporting the metallic finial that extends high above the roof. The central pillar, also known as shin-bashira in Japanese, is structurally separated from the rest of the pagoda, thus it does not carry any load of the pagoda save for the metallic finial. The 4 earliest three storey pagodas, all dated back before the 9 th century have their central pillars built on the ground level, on top of a base stone. Pagodas after that however have their central pillars built above the ceiling of the first storey. Central pillars have long been hypothesised as the main quake resistant component of pagodas, serving as some sort of latch bolt and damper [4]. Through experiments, it has been proven that central pillar oscillates with a phase different to that of the rest of the pagoda, and as a result providing some sort of damping effect during earthquakes [5]. Judging by the fact that pagodas of both types suffered similar seismic damage, there is no clear difference between the two types of central pillar in terms of seismic resistance. Central Pillar Metallic Finial Figure 1: Central pillar of Taimadera (W) and Myotsu-ji Table 2: Seismic damage and central pillars Name Damage Type Central Pillar Oppo-ji Finial First Storey Hiyoshi Shrine Finial First Storey Sanmyo-ji Finial First Storey Jimoku-ji Collapse First Storey Hoki-ji Finial Ground Yakushi-ji East Pagoda Finial, inclination Ground 3.2 CONSTRUCTION METHODS Length of Finial Total Height The main components of a typical pagoda framework include the support columns (12 outer columns and 4 inner columns surrounding the central pillar), bracket complex and rafters. Each storey has its own individual columns that are not connected to their corresponding pillars below and above. Construction method of pagodas could be categorised into 3: stack-up method, long column method, and yagura method. Each method

Total Height (m) could be easily distinguished by observing the position of the support columns on the upper storeys. Stack-up method is the oldest and most common construction method. Pagodas built using this method have their columns placed on top of the rafters, in other words, on top of the roof of the storey below. Long column method pagodas have their columns built directly on top of the beam-like bracket complexes, called tsunagi-hijiki in Japanese. As the result, the columns are longer compared to that of the stack-up method, thus the name long column method. In yagura method pagodas, bracket complexes are replaced by beams supporting the columns. There are also combinations of 2 or more methods, usually with the outer columns built using the stack-up method, while the inner columns built using the long column or yagura method. 35 out of the 57 pagodas are built using stack-up method, 9 are of the long column method, 9 being the combination of stack-up and either long column or yagura method, and 1 of yagura method. The rest remains unknown. Long column method was invented later than stack-up method, with the first known case being from Muromachi Period. Long column and yagura method have the advantage of shortening the construction period by discarding the need to consider the balance between the weight of the roof and the upper storeys, as well as the need to complete the roof before continuing to the construction of the upper storeys [2]. Interestingly, all three storey pagodas that suffered seismic damage are of stack-up method. Central Pillar Inner Column Outer Column have receded raigō-bashira, and 5 have none. Inner columns play an important role in carrying the load of upper storeys, thus structurally speaking their omission is unfavourable towards the pagoda s stability. However, All 6 pagodas that have suffered seismic damage except Sanmyō-ji carry shiten-bashira on the lowest storey. 3.4 INNER COLUMNS ON UPPER STOREYS There are pagodas that have all their inner columns omitted on all storeys, while some have their shitenbashira omitted only on the lowest storey but retained on the upper storeys and vice versa. 19 out of the 57 pagodas have shiten-bashira on their 2 upper storeys omitted. Although the omission of inner columns gives an impression of a less stable structure, all 6 pagodas with seismic damage record carry shiten-bashira on their upper storeys. Further study is needed to determine the role of inner columns in preventing seismic damage. 3.5 SIZE AND ROOFING MATERIAL The total height, length of metallic finial, total area, length of eaves, and slenderness ratio of support columns of each pagoda were analysed. The graphs below show that the 4 oldest pagodas are comparatively larger in scale. The tallest of all, the east pagoda of Yakushi-ji, has slender support columns notwithstanding its gigantic scale of 34m. All pagodas with seismic damage record except Sanmyō-ji generally have taller than average height, longer finial and eaves. On the other hand, pagodas that adopted new construction methods such as the long column method and the omission of inner columns in general are of smaller scale and have shorter eaves. 29 out of the 57 pagodas have tiled roof. Including five storey pagodas, tiled roof pagodas are more susceptible to seismic damage. Those with long column and yagura construction method, as well as those without inner columns are more inclined to having lighter roof such as copper plate and thatched roof. Figure 2: Typical plan of a pagoda Central Pillar Bracket Inner Column Outer Column Corner Rafter Figure 3: From left: stack-up, long column and yagura 3.3 INNER COLUMNS ON LOWEST STOREY Originally there are 4 inner support columns on each storey. These 4 inner columns are called shiten-bashira. However, later down the history there came pagodas that have their front 2 inner columns on the lowest storey omitted, remaining only the back 2 columns called raigōbashira. There are also pagodas that have no inner columns at all. 36 out of the 57 pagodas have shitenbashira on the lowest storey, 11 have raigō-bashira, 5 Long column without damage record Long column with damage record Others without damage record Others with damage record 35 30 25 20 15 10 700 1200 1700 Figure 4: Total height of three storey pagodas

Slenderness ratio of 3rd Storey Column Length of 3rd storey Eaves (m) Lenght of Finial (m) 11 10 9 8 7 6 5 4 3 700 900 1100 1300 1500 1700 Figure 5: Length of finial of three storey pagodas 4 3.5 3 2.5 2 1.5 1 700 900 1100 1300 1500 1700 Figure 6: Length of 3 rd storey eaves 0.35 0.3 0.25 0.2 0.15 0.1 0.05 700 1200 1700 Figure 7: Slenderness ratio of 3 rd storey outer column 4 MAIN DIFFERENCE BETWEEN THREE AND FIVE STOREY PAGODA Central pillar that reaches ground level is much more common in five storey pagodas regardless of construction year. Also, inner support column is never omitted for five storey pagodas. Therefore it is said that the construction and design of five storey pagodas are more prudent and conservative, while three storey pagodas are inclined to being experimented with new construction methods [2]. Similar to the three storey pagodas, there are no seismic damage records for long column and yagura method five storey pagodas. 5 CONCLUSIONS The 57 pagodas studied showed very few cases of seismic damage despite experiencing multiple times of huge earthquake, in fact, fewer in terms of proportion compared to that of the five storey pagodas (5 pagodas, 7 cases out of 22 pagodas [1] ). This proves their exceptional earthquake resistance. This is especially true in cases of long column and yagura construction method without any seismic damage recorded. On the other hand, large scale pagodas with heavy roof i.e. tiled roof are apparently more susceptible to seismic damage. Three storey pagodas have been subjected to various new construction methods and designs. In such cases, they are usually reduced in scale and weight, possibly to ensure stability. It is not completely true to say that five storey pagodas possess superior earthquake resistance just because there exist several cases of three storey pagodas collapsing due to earthquake. It can only be deduced that there exists a certain level of variants when it comes to the earthquake resistance of three storey pagodas. Further studies should be done to study the pagodas that collapsed during earthquake. ACKNOWLEDGEMENT The assistance provided by the members of Fujita Lab. and the university library is greatly appreciated. REFERENCES [1] M. Ooyama, K. Fujita: Damage Characteristics from Historic Earthquakes, AIJ (C-1), 249-250, 2002. [2] M. Hamashima: Japanese Pagodas, Chuo Koron Bijutsu Press, 2001. [3] T. Usami: Historical Earthquake in Japan, University of Tokyo Press, 1996. [4] NPO Forum for Wood Architecture: Dynamic Behaviour of the Five Storey Pagoda, NPO Forum for Wood Architecture, 2004. [5] K. Fujita & others: Micro Tremor Measurement of Ikegami Honmonji Temple, AIJ(C-1), 251-252, 2002. [6] (Respective) Board of Education (or The Japanese Association for Conservation of Architectural Monuments or Repair Committee): Repair Reports of each Pagoda, (Respective) Board of Education (or The Japanese Association for Conservation of Architectural Monuments or Repair Committee), various publication years (from 1954 to 2000). [7] H. Ota: Volume 11 & 12 of Japanese Architecture History, Chuo Koron Bijutsu Press, 1984 & 1999. [8] K. Iida: Tensho Earthquake, Nagoya University Press, 1987. [9] Agency for Cultural Affairs, National Treasures & Important Cultural Properties Index, Agency for Cultural Affairs, 2009. [10] Homepage of Each Temple (Shrine). [11] Koyasu-to s photos, Minka Saisei http://www.reform-apple.com/html/kominka/report55/kiyomizureport55-2010.4.htm [12] Photo album of pagodas, http://www.d1.dion.ne.jp/~s_minaga/touba3.htm

Figure 8: Earthquake encounter record chronological table part 1

Figure 9: Earthquake encounter record chronological table part 2