Great Hanshin Earthquake

On the first anniversary of the moment magnitude (MW) 6.7 1994 Northridge Earthquake, Kobe, Japan was struck by an MW6.9 earthquake. Both earthquakes struck in the pre-dawn hours, both ruptured beneath densely populated areas, and both caused horrible damage. Yet in Kobe there were many more deaths, financial losses dwarfed those in Northridge, and the amount of destroyed building stock and infrastructure was far worse in Kobe than in Northridge.

The reasons for these differences are many, but it would be incorrect to issue a blanket condemnation of current Japanese seismic engineering practice. While engineered structures did fail due to design flaws, they were predominantly older structures built before the current Japanese building code became effective; or they frequently failed due to problems revealed to be deficiencies in California design practices by the Northridge Earthquake. Japanese seismic engineering expertise has justifiably been considered among the best in the world, and a careful examination of the damage in Kobe does not change that conclusion.

Despite differences in design and construction practices, the same general principles frequently came into play: highway collapses were often primarily due to insufficient lateral ties in the concrete columns, nonductile concrete frame buildings did much worse than ductile design, shear walls typically helped to lessen catastrophic damage, and soft soils resulted in greater damage to the structures constructed on them.

The most important lesson in both earthquakes is that the knowledge to significantly improve structures to resist earthquake damage and thereby avoid most of the deaths and financial losses exists; what is lacking is a consistent willingness to marshall the resources necessary to put that knowledge to work on the scale necessary to prevent disasters. It is an odd paradox, for time and time again it is demonstrated that it usually costs less to prepare for earthquakes in advance than to repair the damage afterwards.

Differences in Kobe and Northridge

While there are more similarities than differences in structural performance in the Kobe and Northridge earthquakes, there are important differences that explain why the Kobe Earthquake was so much more damaging. Some of the lessons from these differences apply only to Japan, others apply to all areas of the world at risk from earthquakes.

The vast majority of deaths in Kobe occurred in the collapse of housing built using traditional Japanese methods. Traditional Japanese housing construction is based on a post-and-beam method with little lateral resistance. Exacerbating the problem is the practice of using thick mud and heavy tile for roofing, resulting in a structure with a very heavy roof and little resistance to the horizontal forces of earthquakes. U.S.-style frame housing with light-weight roofs is now coming into use in Japan and newer housing constructed using these methods had little or no damage from the earthquake.

Another significant difference between the Kobe area and the Northridge area is the quality of the soils. Because of a severe shortage of available land, much of modern urban Japan, including Tokyo, is built on the worst soil possible for earthquakes. Much of the newer construction in Kobe, particularly larger buildings, is built on very soft, recent alluvial soil and on recently constructed near-shore islands. Most of the serious damage to larger commercial and industrial buildings and infrastructure occurred in areas of soft soils and reclaimed land. The worst industrial damage occurred at or near the waterfront due to ground failures-liquefaction, lateral spreading, and settlement.

The Port of Kobe was an extreme example of the problems associated with poor soils in areas prone to earthquakes. The port is built almost entirely on fill. The engineering profession has tried hard to develop methods for strengthening filled areas to resist failures during earthquakes, but most of these methods have been put into practice without the benefit of being adequately tested in strong earthquakes. The results were decidedly mixed, but the failures costly_most retaining walls along the port failed, and the related ground settlement pulled buildings and other structures apart.

Buildings

The large commercial and industrial buildings in the Kobe area, particularly those built with steel or concrete framing, are similar to buildings of the same vintage in California. The Japanese building code had a major revision for concrete-frame buildings and a more limited revision for steel-frame buildings in 1981. The Uniform Building Code, as used in California, had major changes in 1973, 1975, and several times since then. The current Japanese code requires that buildings in Japan be designed for somewhat higher force levels than does the Uniform Building Code. Both areas require design for much higher forces than most other earthquake regions of the world.

Typically, pre-1981 concrete-frame buildings performed very poorly in Kobe, with many collapses. Post-1981 buildings performed much better_some were extensively damaged, but most had light damage. The buildings that fared best, and those without significant damage, had extensive concrete shear walls.

As in other earthquakes, large commercial and industrial steel-frame buildings performed better than any other type. However, major damage and a few collapses were observed. Pre-1981 steel buildings had most of the serious known damage. Certain innovative types of steel buildings, including high-rises, had very serious damage, and collapses might have occurred if the duration of the earthquake had been a few seconds longer.

Building owners usually do not understand that the earthquake provisions of even the strictest building codes do not necessarily have reasonable performance criteria for larger and stronger earthquakes. The current regulations, including those for all of California, are typically written with the expectation that in a strong earthquake a building will be severely damaged_in fact, it is assumed the building may need to be torn down, but it should not collapse. In California, higher performance criteria are mandated for certain types of structures_schools, hospitals, police and emergency response buildings, and certain power facilities. An informed building owner can choose to use these higher criteria, and thus avoid having their high-value, heavily occupied commercial building designed, in effect, to the same earthquake performance level as a low-value farm building.

Transportation

A number of major expressways, rail lines, and bridges, some of very modern design, were severely damaged. There are no significant new lessons from the collapse and damage of the older unretrofitted bridges and elevated structures. The structural and foundation details that typically caused damage to the expressways and rail lines have been observed in numerous earthquakes, and the damage was predictable. Some of the upgrade details observed in older retrofitted structures, such as steel column jacketing, are now widely used in California for strengthening. The apparent good performance of these details in Kobe is important to ongoing U.S. programs and needs to be studied in detail.

Many bridges and bridge approaches were severely damaged. The performance of large new bridges, including cable-tied arch, braced arch, and cable-stayed bridges, should be studied extensively because this is the strongest earthquake to affect such bridges.

The Port of Kobe, much of which was new, was devastated by widespread and severe liquefaction and/or permanent ground deformation, which destroyed more than 90% of the port's 187 berths and damaged or destroyed most large cranes. Shipping will be disrupted for many months, and some shipping business will probably never return to Kobe, resulting in significant losses to the local economy.

Other Infrastructure

The electrical and telecommunications systems in Kobe and surrounding areas performed as expected based on experience from previous earthquakes. Long term power outages were isolated to the most heavily damaged areas. Facilities near the epicenter sustained damage while resiliency of the systems prevented widespread service interruption. Most of the major transmission lines skirt the heavily damaged region of Kobe_the results may have been substantially different if the epicenter was located closer to the 500 kV transmission system. There were substantial financial losses to the electrical utilities, however, because expensive specialized equipment must be replaced and the distribution network must essentially be rebuilt within heavily damaged areas of Kobe.

During the earthquake, Kobe's water system sustained approximately 2,000 breaks. Generally, ground or building failure was the cause of the severe damage to Kobe's water systems. The resulting lack of water contributed significantly to the fire problem and will be a major hardship on the population for several months. The gas system had major damage, generally caused by ground or building failure, which also contributed significantly to the fire problem.

Fire

More than 150 fires occurred in Kobe and surrounding areas in the hours after the earthquake. These resulted in several large fires, and fire fighters were for the most part unable to combat them because of streets being blocked by collapsed buildings and building debris, traffic congestion, and severe water system damage. Calm wind conditions prevented conflagrations. The United States and Japan have both sustained the largest peacetime urban conflagrations in this century's history_because of earthquakes. Fire following earthquake is a potential major agent of damage, and needs to be recognized as such by planners.

© 1997 Ricardo Cheung

since Nov. 17, 1997

Return to home