Lessons learnt from Kobe

Brian Langham from NZ Wood visited the Japanese city of Kobe six months after its devastating 6.8 magnitude earthquake in 1995. His report on the lessons learned from the earthquake and its implications for New Zealand (a version of which was published at the time in the Dominion newspaper) is reproduced here in the light of its parallels with the recent Christchurch earthquake.

The destruction caused by the unexpected earthquake in Kobe this year (1995) serves warning that Wellingtonians and other New Zealanders should not be complacent about such a disaster.

While people living close to major fault lines around the world wait for the “big one”, are the dangers of smaller ruptures of secondary faults being overlooked and leaving people unprepared? The devastation caused in Kobe six months ago could have important lessons for New Zealand cities.

Kobe is a city whose younger population claim hardly to have experienced even a small tremor. It came without warning, and nobody was expecting it to happen there. All eyes were on Tokyo waiting for the “big one”.

The parallels with New Zealand are most disconcerting. While most people focus on Wellington, how many people in Auckland know they have active faults just 45 km away capable of producing bigger quakes than that which collapsed motorways in Los Angeles in 1994? How many residents of Christchurch or Dunedin have arranged meeting plans with their families in the event of a similar disaster?

As both the Kobe and Los Angeles quakes showed, earthquakes often occur where people do not expect them. Seismologists had been concentrating on the main Pacific fault in Japan – thought to have caused 143,000 deaths in Tokyo and Yokohama in 1923 – or the San Andreas fault in California.

But both the recent quakes occurred not along these major inter-plate faults but on the secondary intra-plate faults that spiderweb a single plate. These secondary faults are usually much less active. The problem, however, is that there are so many of them.

While intervals between quakes along the main faults are measured in centuries, the secondary faults may only cause a major earthquake every 1000 to 5000 years. In comparison, the likelyhood of an earthquake of seven or over on the Richter scale occuring in Wellington in the next 30 years is greater than one in ten. Although the faults nearest to Auckland may only shift significantly every 10,000 years, they are capable of producing a magnitude 6.5 to 7.0 quake when they do.

In earthquake prone countries throughout the world there are so many of these secondary faults seismologists now warn that they present a combined risk that could be at least as great as the “Big One”. Secondary faults such as these run right through Dunedin. Plenty of them circle the Canterbury Plains

The consequences of not being prepared are no more apparent than in Kobe this year. While in Tokyo 27 percent of homes kept emergency supplies, in Osaka the number had shrunk to 2.6 percent. Tokyo’s army and civilian officials conducted annual exercise drills to test co-ordination but Kobe officials conceded they did not.

Professor Hiroyuki Kameda from the Urban Earthquake Hazard Research Centre at Kyoto University says earthquake engineers had been warning of the seismic hazard of the region for many years but the risk had not been seriously considered.

“Many people said that this area was free from earthquake threat, but it is not true,” he says. “We had been experiencing only some sort of quiescent period after a major earthquake took place in 1946.”

Six months on, while the city has got itself back to business amongst the scenes of demolition, guidelines for city reconstruction are only now being released for public discussion. Realists expect it will take between 10 and 20 years for the city to recover fully.

Largely due to the pre-dawn timing, nearly 90 percent of those who died were killed at home in traditional wooden houses. More than 80,000 wooden houses – almost 10 percent of all homes in the area – either collapsed totally or were substantially destroyed.

Construction of traditional Japanese houses is quite different to typical New Zealand houses, which could be expected to survive better.

Rather than four-by-two framing and studs, Japanese use light framing and lattice walls covered by plaster. The roofs, in contrast, are often heavy ceramic tiles which are usually held firmly in place by a heavy layer of clay mortar, to prevent them being blown away by typhoons.

This combination of heavy roofs and light walls proved fatal .

Newer houses that were built with four-by-twos as in New Zealand suffered almost no damage. Likewise, prefabricated houses proved very safe.

Makoto Watabe, chairman of the Earthquake Disaster Committee of the Architecture Institute of Japan says while the number of wooden house collapses was very high, they killed relatively fewer people than the reinforced concrete buildings which collapsed.

Approximately one person was killed for every 16 collapsed wooden houses. Assuming an average of three people lived in each house, that is about a two percent mortality rate, he says.

“In the case of reinforced concrete construction, while there weren’t as many collapses, that percentage went up to 15 percent – a big difference.”

In short, while there may have been less chance of a reinforced concrete building collapsing, if you were in one that did, you had a greater chance of being killed than if you were in a collapsed wooden house.

The performance of larger buildings can be seen to validate the changes made to Japan’s seismic engineering code in 1981. Those built after that time suffered little damage while older buildings all too often sustained fatal damage.

Many collapsed completely, many others had an entire middle floor simply sandwiched flat – so called “soft storey collapses”. These occurred at points of weakness. Often this was the ground floor in designs where the ground floor was used for car parking and would be virtually without walls, merely pillars. In other cases it would be at a point in the building where the design strength changed.

Unlike New Zealand, Japanese construction often “jackets” reinforced concrete pillars with an outer casing of steel. In buildings such as the old Kobe City Hall built in 1957, such steel jacketing was used for the first five floors, but above that the beams were unjacketed. The sixth floor simply gave way at the point where the reinforcement changed with the top three floors crushing the sixth floor leaving it less than a metre in height.

Unfortunately many of the crucial planning records for water, gas and electrical supply were kept on this floor, hampering emergency response and restorative efforts, until a brave soul crawled through the crevices to recover as much as possible.

The performance of the old City Hall with the 50 storey plus new City Hall only metres away built only three years ago gives some reason to hope that modern building codes have got it right.

While the older building suffered major damage, the new tower was left completely undamaged. Not one pane of glass was cracked, which may give some comfort to those who might be walking along the glass canyons of Lambton Quay in the event of a major quake. The widespread fear that while new seismic codes may leave buildings structurally intact, a deadly shower of glass and cladding would rain down onto the street, may thankfully not be realised.

The relatively superior performance of high-rise buildings built to the 1981 codes compared with older low and mid-rise buildings is not accepted unreservedly by all experts, however, as complete endorsement of the current codes.

Professor Kawamura of Kobe University says maybe it was simply that high-rise buildings benefited by the short, sharp duration of shaking that resulted from a close, shallow quake. There was not time for the building to oscillate as it would if the wave was slower and more prolonged.

If the main ocean fault over 200 km away had moved, tall buildings could have been damaged much more than lower buildings.

He feels the codes still neglect the need for strong sheer walls – the walls that provide strength between the main columns. This was seen in the poor performance of buildings with comparatively open floors such as used for garaging and retail spaces. He believes the latest codes may actually provide for less shear walls than previously making the buildings no safer.

So how do the Japanese seismic codes for buildings compare with those in place in New Zealand?

A direct comparison is complicated by the differing design philosophies used in construction in the two countries. The Japanese codes specify a higher threshold before which the buildings will deform than the New Zealand codes. The New Zealand codes, however, take more account of the “ductile” capacity of the structure – the amount of deformation the building can safely sustain without it collapsing.

The New Zealand earthquake engineers’ report following a reconnaissance mission to Kobe says despite these differences in approach they have confidence that the current design standards in New Zealand are adequate.

But the performance of buildings in Kobe emphasises the need to strictly enforce those codes. The poor performance of older buildings in Kobe that were designed to inferior codes reinforces the need to assess New Zealand’s older buildings for seismic resistance in accordance with today’s standards.

Following their trip to Kobe, the New Zealand earthquake engineers’ reconnaissance team warned that New Zealanders must not be complacent in their considerations of earthquakes. “The attitude or belief “that it will not happen to us” or “it will most likely only occur in Wellington” needs to be eradicated,” their report says.