Sunday, September 23, 2012

Shaken, not Stirred

The Great Hanshin earthquake, or Kobe earthquake, occurred on Tuesday, January 17, 1995. It made rather a mess of things - more than 6400 people lost their lives, and about ten trillion yen ($100 billion) in damage was caused, 2.5% of Japan's GDP at the time.

The strength of this earthquake moved the ground 7 inches horizontally and 4 inches vertically. Buildings and structures of all types suffered tremendous damage, as the following collection of photos illustrate:  


Kobe, an industrial manufacturing center, was a part of Japan firebombed into a wasteland by the US in WW II, and had been rebuilt in the post war period fairly hurriedly - you might say that corners were cut in certain instances. Despite the normal Japanese fastidiousness, and their vast and sophisticated use of concrete generally speaking, the expressway support columns had lacked certain key pieces of rebar reinforcement:


Seems like it has plenty of rebar, judging from the photos, but I think not enough horizontal banding was the issue.

Older buildings, constructed before the 1980 promulgation of a national building code for Japan, suffered from insect damage in the timbers, lack of shear bracing, and in some cases had been retrofitted with heavier tile roofs that exceeded the design for the structure as originally configured. In a report entitled Lessons in the Strengthening and Reinforcement of Historical Buildings from Rescue Projects Following the Great Hanshin Earthquake, Professor Nobuo Ito said:

"In the case of shrines, the collapse of the one-bay Nagare style Yakujin Honden of the Rokko Hachiman Shrine is worthy of note. The way in which it had collapsed, just as if its legs had been kicked from under it, makes one think that it was thrown off balance in an instant and the body of the building was crushed by the weight of the roof. There can be little doubt that, as Dr. Kuroda has reported, the main cause was that the roof covering had been changed from cypress bark roofing to heavy classic roof-tile, but it is thought that the rather unbalanced Nagare roof form was also a contributory factor."

Few owners chose to retrofit older structures with uprated bracing systems, regardless of code changes, as there is a cultural preference in Japan to demolish and build anew when it comes to most houses, rather than do retrofits. And these things do cost money after all, and severe earthquakes in the region were relatively uncommon.

While houses of all sorts of construction types, traditional and modern, and varying materials - concrete, brick, stone, 2x and ply, and timber-frame - were trashed along the primary lines of seismic activity, the Pre-fab and 2x construction industries were quick to take advantage of the sudden need for new buildings in the greater Kobe area - and elsewhere - and advertised their systems in part by showing pictures of failed traditional structures. As traditional structures are considerably more costly to build anyhow, a usual disincentive in any market, this marketing campaign turned out to be one of the proverbial nails in the coffin of traditional building in Japan. The industrialized building approaches thereby gained a greater foothold in the Japanese market, a trend that has continued to this day. 


This earthquake demolished homes of 200,000 people. Professor Ito, in the aforementioned report, noted that damage was most devastating in wooden buildings with:

  1.  poor foundations and sills
  2.  insufficient diagonal bracing
  3.  inadequate connecting elements
  4.  insufficient areas of solid wall
  5.  heavy roofs.
Looking specifically at two kinds of traditional Japanese architecture, Ito first considered thatched roof farmhouses:
"Among other things, in the earth-floored area (doma) of farmhouses, posts are close together and the surrounding walls are more or less continuous, making this part of the house comparatively strong, whereas the reception rooms (zashiki) at the other end of the house are more open, with fewer walls and posts. For this reason, it is reported, there was a characteristic trend for damage to be concentrated in the reception rooms.
Another point that has been noted with respect to farmhouses is the strength of thatched vernacular houses. The Kosaka house, in Ashiya, was one of the few thatched farmhouses still surviving in the area of severe shocks, and it suffered only slight damage. The center of gravity of a thatched vernacular house as a whole is low. Moreover the whole roof frame is a flexible network held together with rope, and as a result the roof frame can evade earthquake forces, making it highly earthquake resistant."
And then he looked at Japanese traditional detached houses characteristic of Kobe:
"In the Kansai region, where there had been no earthquakes for a long period, countermeasures against typhoons had been given priority. The use of comparatively heavy tiled roofs, despite the light eaves detailing and the wide south facing openings, seems to have been an adaptation to cope with such local climatic conditions. Sukiya houses of the kind referred to as modern Japanese-style houses were built in large numbers in Hanshin area until the last war. These houses developed as if each were competing to be the most delicate and the least enclosed, but it cannot be denied that, with their lack of solid walls and ill-balanced layouts, they were ill-conceived for an encounter with an earthquake on this scale."
A final point Ito makes in respect to traditional architecture, a plus and a minus at the same time, concerns relative repairability:
"With the kind of vernacular houses mentioned above, there is a need to inform widely that one of the advantages of the traditional method of building is that, even in the case of severe damage such as cracking of the walls and inclination of the posts, it is possible to repair the structure. Craftsmen known as house pulling carpenters (hikiya daiku) make a profession of moving buildings, but they can also undertake this kind of straightening of the frame of a building. Since they can easily correct a degree of inclination of the structural frame that ordinary carpenters would regard as hopeless, we have come to appreciate their true worth in the aftermath of this earthquake. In the Hanshin area there are no longer any such craftsmen, and it is deeply regret that there are not more hikiya daiku."
The Japanese have taken steps to thoroughly analyse buildings and their performance in seismic events. In 2003, a massive structure, the Hyogo Earthquake Engineering Research Center, aka 'E-defense', was completed:


This structure is purpose-built as a giant test bed for entire structures. And when I say 'bed', I mean a huge mobile bed is inside this building, atop of which structures of various types are built, and then evaluated by subjecting them to movements approximating seismic events. 

Here's a video clip, or three, to show what goes on when the shake test occurs:

Another traditional frame, completely bare:

Here's a pair of completely traditional, fully-detailed houses, and here the test is at 100% of the seismic force of the Hanshin Earthquake:

There are several approaches being taken to making structures more resistant to earthquakes. One approach is to make the structure as rigid as possible. Another involves sophisticated rollers or springs in the foundation to dampen uplift and sideways shifting loads. The traditional Japanese approach has evolved over a long time to produce structures which behave flexibly in earthquakes, like the proverbial willow shedding the load of snow by bending while the oak has its branches snap off under a similar loading. The difficulty in making design decisions in this area is balancing the cost of making a structure more resistant to earthquakes against the likely frequency of the event. Few would want a structure to withstand the worst earthquake likely in 1000 years, if the building is unlikely to last more than 150 years otherwise. And given that many parts of Japan deal with typhoons on a more frequent basis, trying to design structures to perform well in those conditions only complicates the picture. Then there are tsunami - how would one make a structure, other than some sort of bunker, that could resist the sort of wave seen in the recent disaster in Japan?

Culturally, the Japanese have developed a sort of fatalism about such things. Expressions such as 'shigata ga nai' (nothing can be done") and 'shoganai' ("it can't be helped") are common in everyday conversation, and, some would say, more-or-less representative of philosophical outlook. Even 'sayōnara' literally means, "well, if it must be so." There are many other examples in the Japanese language where fatalism is expressed, far too numerous to mention here.

I think it is intriguing though to see the 'E-Defense' structure and the willingness to put building systems under reality testing. While it can't be cheap to do this sort of thing, hopefully some highly useful data will come out of those tests perhaps showing new ways forward in building design and construction, and potentially saving lives in the process.

Thanks for coming by the Carpentry Way. Comments always welcome.


  1. Chris

    How odd to see those apartment buildings still perfectly square, but so desperately out of plumb! There may be such a thing as too much diagonal reinforcement!


  2. Tom,

    pleased to receive your comment. Besides the issue of whether the building holds together or not, there is the matter of people being killed and crushed not by the structure, but by things inside the structure that come loose during the seismic event. Chandeliers, large bookcases, etc., can become deadly when thrown about rapidly. And if the building is ultra rigid, then the seismic shocks will be transmitted even more directly to occupants. It's interesting to observe in videos of other tests at E-Defense (not shown above) how the behavior of furnishings inside the houses can cause danger.

    And a lot of people think it is safest to stand under a doorway or similar, but if the building collapses, these locations are not actually so safe. Apparently it is better to position yourself next to an object which will provide some adjacent space if the ceiling comes crashing down. Better to be next to a car than under a car, that sort of thing.


  3. I have a friend that is Hikiya Daiku, and he also supplemented his skills to also be what is referred to as a "Tobi", or a person that specializes in doing work above ground level, sometimes well above ground level! I met this interesting man when removing tree tops. I believe that he must be the only Hikiya Daiku in the area, and there doesn't seem to be much demand for their work any longer, probably both in part due to the expense, and also the difficulty in finding such specialists, to the point that there work seems to have been largely forgotten.

    Living in earthquake central here, especially after the latest big one in Northern Japan, there seems to be more and more concern about living in houses with the traditional heavy tiled roofs. As mentioned in your post, cited as one cause for very negative earthquake consequences. Though they still can be seen everywhere, possibly now some movement away from them. Of course there is already the strong movement away from timber frame itself, though more often than not it is for dubious reasons. Chris, apart from ''officialdom'' what are your own thoughts on the heavy roofs? In a high frequency earthquake region, can timber frame houses be built to counter the possible negative consequences of the heavy roofs in the event of a big shaker coming along, as they certainly will do? I'm wondering what construction plan you would advise? During the Northern quake, i was far enough away to not experience damage, but standing outside of my shop at the time, I suddenly felt sick to my stomach. You do get a very odd combination of movements with the quakes.

    1. Dennis,

      yes, the hikiya daiku has exactly the same problem as many doing traditional work - demand drying up.

      I think heavy roofs are fine so long as the structure is designed for them. Heavy roofs and 'sukiya', with its tendency towards overly-slender structural elements is, to my way of thinking, an odd and ill-suited combination. If you want the sukiya look, then keep the structure to one floor, and use a metal roof.

      Heavy tile roofs in Japan have posed problems for centuries due to their weight. On massive structures, the weight of the roof will deform structural members over time, especially hip rafters. One solution has been to redesign tiles to be lighter. Another has been to beef up the structure using a lot of metal, even metal columns with concrete filling, etc.

      In the end though, while you can design and brace a structure with a heavy tile roof to perform adequately in a high-frequency earthquake region, it makes far more sense to me to use a lighter roof and try and keep things to one floor. I really like copper-shingled roofs myself. Much lighter than the tile, and with comparable lifespan. There are other metals used for roofing as well of course, most cheaper than copper (terne, steel) and some more costly (titanium).


  4. Hi Chris,

    Thanks for a wonderful post. It is rather serendipitous as well, as I have been having several conversations of late with engineers, architects, and G.C. about the true realities of Asian architecture. There is so much miss information and misinterpretation, that I am routinely dispelling myths of building methodologies East to West.

    Your fourth photo down is one I will be keeping for educational purposes. It clearly shows what I have been stating for decades about the simple reality of “horizontal” versus “oblique” bracing systems. Often I am confronted with the statement, “this design you are submitting has no proper bracing,” or “the building is not rigid enough to be safe.” This often comes from architects or engineers that are only indoctrinated into Western thinking modalities. I try to use the metaphor of Martial Arts, which seems to help them understand. Eastern combat styles are much more deflecting and flexible; while the Western fighting styles tend to be more ridged and braced, often with the back leg in and oblique position to the body.

    I have also noted through observation that oblique bracing, (which only works in compress unless it is designed as a “strut”) is truly too ridged, often acting as a fulcrum weakening a frame, and when it fails, it does so catastrophically. Unlike horizontal bracing, (see fourth photo,) which is flexible and forgiving, often allowing occupants time to respond to a seismic or metrological occurrences and further allocate the buildings repair.

    From a construction approach, the placement of Nuki 貫 beams (a pass through horizontal member for those not familiar with Japanese design,) facilitates design flexibility and ease of assembly. When assembling a timber frame, I would much rather deal with 貫 than all the different bracing assemblies within bent systems of Western frames.

    I could go one but would rather wait to see what others have to say. Thanks once again for your wonderful efforts on this blog.



    1. Jay,

      great to hear from you. I think if you are going to propose non-Western structural systems to Western engineers, you are going to have to convince them with empirical data for starters, and there's virtually none of that in the English language. So, uphill is your path. You face the same problem as those who wish to be proponents for cob, light clay straw, hemp-crete, etc. - engineers are not going to sign off on it unless they can be confident about performance. With next-to-no data to go on, the only recourse is modelling and finite element analysis, building model walls and testing them with a simulator, etc. That costs money and time of course. If, as an engineer, your professional license was on the line, and a lawsuit could sink your business permanently, you would be prudent about where you placed your stamp. Trying to convince them with metaphors about Eastern martial arts are generally going to be a waste of time I'm afraid.

      The irony here is that engineers will not sign off on western timber-framed structures either, regardless of bracing, as they cannot be made adequately rigid enough. The pegs will always allow too much movement at the joints. For a barn or utility structure, a little building sway in the wind is not much of an issue, however for modern houses with their expensive triple-glazed sealed windows and carefully controlled air exchanges such wall movements cannot be allowed. The primary way in which the timber frame structure can exist in such an situation is as a temporary scaffold and then a largely decorative element, as seen in the usual timber-framed, SIP-wrapped structural system we have grown to know and, uh, love (ahem!).

      I don't quite agree with your conclusion about oblique bracing. Oblique bracing IS tunable for flex. Yes, short and thick braces - the typical knee brace - are rigid, and can act as fulcrums to lever nearby joints apart if subject to shearing loads. They are also notoriously weak at their connection points. A 'strut', by definition, is subject to compression alone (a 'tie' being designed to resist tension alone), and a strut is how a wooden brace functions in most timber frame designs. Wood is most suited to functioning in that manner in a structure, unless the connections are to rely upon metal and/or adhesives.

      If a more flexible bracing arrangement was desired, then the braces can be lengthened, slimmed, curved, etc. Again though, the idea of a flexible brace is not going to fly with structural engineers if a primary goal is to keep the building from moving under shear loads so as to preserve the integrity of doors and windows. I've always liked longer braces myself, such as you see in German traditional wall framing.

      I also like nuki, however it is most critical that the joins between nuki and posts be tight for good structural performance; that tightness in fit however makes assembly difficult.

      Japanese framing has employed diagonal bracing and trusses for the past 150 years to one extent or another, and the vast majority of Japanese timber frames done in the post-war period, residences and temples alike, make use of diagonal reinforcement. Most of that is concealed in the structure, in the walls and up in the roof, as are the metal reinforcements. There's no need to be completely avoidant of diagonal bracing in an Japanese/Chinese timber frame - a point of harmony can be achieved.

      So, if you want to bring these sorts of Asian framing systems, which have great virtues, into the 21st century Western context, then you are going to have to adapt them in ways which work within this context, including oversight by the structural engineering community. One way is enclosing the timber structure with SIPs, however I believe there are better solutions than that which give adequate rigidity and good thermal performance.



All comments are moderated, so if you're planning to spam this, know now that your clicking and pasting is in vain. I do read the comments before posting, so your mission is doomed from the outset. All this time and effort trying to put your inane spam onto blogs -- is this how you want to spend your time on earth?

Please do me the courtesy of appending your name to your comment, even if posting under the 'anonymous' option. No name = deleted.

Comments NOT accepted include:

-those containing links unrelated to blog content
-spam of any kind, or ham for that matter
-did I mention that attempted spam postings will be non-starters?