387227a0Nature3876630199705152272290028-0836199710.1038/387227a01476-4679199715 May 1997ukNatureNatureNATUREnatureNature is a weekly international journal publishing the finest peer-reviewed research in all fields of science and technology on the basis of its originality, importance, interdisciplinary interest, timeliness, accessibility, elegance and surprising conclusions. Nature also provides rapid, authoritative, insightful and arresting news and interpretation of topical and coming trends affecting science, scientists and the wider public./nature/journal/v387/n6630issueJournal homeArchiveCurrent issueAdvance online publicationPrivacy policySubscribeNature Publishing GroupSupplementsCurrent issue387227a0Engineering design lessons from Kobe
AU  - Chandler, Adrian M.Department of Civil & Environmental Engineering, University College London, Gower Street, London WC1E 6BT, UK.Bringing old building stock into line with modern standards of earthquake-resistant design is a daunting and expensive task. But the damage caused by the Kobe earthquake shows that doing nothing will be even more costly.Kobe's city administrators, construction engineers and 1.5 million inhabitants are all determined to turn the aftermath of the disaster that struck the city on 17 January 1995 to the city's gain. Already more than two years into an ambitious ten-year redevelopment programme, Kobe is set to become Japan's first redesigned twenty-first century city, and perhaps therefore its safest in terms of the risk of further earthquake damage. The planning brief highlights "disaster prevention living zones": wide streets creating fire breaks and a duplication of the main routes through the city by both road and rail. It outlines a new Kobe airport, accelerated extensions to an embryonic subway system and a port containing steel-plated earthquake-resistant berths reserved for emergency services.


It is to be hoped that the insights gained by engineers from the Kobe earthquake will be instrumental and effective in reducing future earthquake damage and loss of life in Japan and elsewhere. There is little doubt that an earthquake with similar destructive force could, and probably will, hit Tokyo within the next 20 years. A large earthquake near the capital, last experienced in 1923, is now long overdue. Some earthquake engineers argue that, at present, Tokyo is in practice little better prepared than was Kobe, and that a similar close hit would cause enormous devastation and loss of life. Kobe has served as a timely warning, and its lessons must be quickly and effectively learned.


The urgent need to upgrade the older (more vulnerable), badly designed building stock is evident in all the main earthquake zones in the world. This may take the form of strengthening existing structures, or of introducing an additional form of response control that will reduce the vibrations and damage caused by earthquake ground motions. In the past ten years, research into earthquake response control methods has become very diverse, and it might be more effective if funding were to be more narrowly channelled into the exploration and implementation of tried and tested, cost-effective, useful techniques for retrofitting older buildings, such as base isolation and some varieties of energy-absorbing frame devices, as I will discuss here.


Lessons from Kobe must be learned, and quickly, as there is a real danger of research into new technology for earthquake resistance in countries such as Japan and the United States becoming unduly sophisticated and diverse, which may draw attention away from the reality that simple observation of a few basic guidelines can effectively reduce the impact of such events on the built environment.


Figure 1 Overall damage statistics in Kobe earthquake, by percentage, for reinforced concrete (RC), steel (S) and composite (SRC) buildings (from ref. 1).


Kobe earthquake and its aftermath


Early in the morning of 17 January 1995, an earthquake of moment magnitude 6.9 hit south-central Japan, causing considerable casualties and damage. Japan had not experienced an earthquake disaster on such a scale for 46 years, during which time the loss of life owing to building failures in earthquakes was just six people a year. As a result, the extent of damage and loss came as a shock to most people, including structural design engineers and risk-assessment specialists. The total death toU in Kobe exceeded 6,000, with 35,000 people injured. The total direct economic loss resulting from the disaster has been put conservatively at $130 billion, making Kobe the most costly naturally occurring disaster in human history, and the most devastating ever to hit a developed country. All this happened in about 20 seconds, which was the duration of strong ground shaking for the earthquake. Adding the direct and indirect losses from business disruption and loss of productivity in the city and port is likely to take the final cost to more than $200 billion.


Virtually  all  Japan  is  classified  as  a medium- or high-risk earthquake zone. In theory, Kobe is in the same high-risk zone as Tokyo, 420 km to the east. But because no recorded damaging event had hit Kobe for the preceding 1,000 years, earthquake-prediction specialists in Japan rated the chances of such an event as remote, and certainly less than the probability of a repeat of the 1923 great Kanto earthquake, which killed 142,000 people in the capital city.


The 1995 Kobe earthquake was particularly severe because of a combination of three factors. First, it was shallow: the main shock was just 14 km below the surface. Second, the horizontal and vertical ground accelerations were exceptionally strong: these are measured in terms of percentage of g, the acceleration due to gravity, and averaged 50-90 per cent of g horizontally (50-150 per cent greater than conventional design values) and 30 per cent of g vertically over a widespread area, including virtually the whole of Kobe city. Third, the strong motion propagated northeastwards for 50 km from the epicentre (close to Awaji Island, 40 km west-southwest of the city) along a known fault line running through soft alluvial deposits, leading directly to the heart of downtown Kobe. Recorded ground movements highlight the linear band of intense shaking.


More than 56,000 buildings (mainly houses) were totally destroyed by the earthquake, with a further 110,000 severely damaged. The Kobe shock was by far the most devastating earthquake of modern times, measured in terms of damage inflicted on building stock that was, by global standards, constructed to a high standard, designed for appreciable earthquake loadings, and predominantly post-1950 construction. The damage would probably have been significantly heavier had the duration of the earthquake been longer than 20 seconds.


Engineering lessons


What are the implications of the Kobe earthquake as an indicator of the potential for earthquake disaster in other parts of Japan, in particular Tokyo? Despite the particular seismological conditions of the Kobe event (outlined above), it is by no means inconceivable that ground motions of equal severity may strike Tokyo from a larger-magnitude, deeper subduction-zone earthquake, or even from a shallow, moderate-magnitude event similar to Kobe. As in Tokyo, most 'engineered' buildings in Kobe are six to twelve storeys high, with relatively few above fifteen storeys high. These engineered buildings suffered a surprisingly high level of damage1, with failure or heavy damage caused to 40-50 per cent of structures made of reinforced concrete, steel or steel/reinforced concrete (composite), as shown in Fig. 1. The use of steel/reinforced concrete as a construction technique for earthquake-resistant buildings, in which steel column sections are encased in a concrete column, is limited mainly to Japan - the standard elsewhere is reinforced concrete, which uses a grid of steel bars to strengthen the concrete columns.


The engineering performance of modern (post-1981) buildings designed to resist a strong earthquake was generally as expected in Kobe, with most surviving the event with minor or moderate levels of damage. The post-1981 Japanese seismic code closely matches the procedures adopted in the United States, Europe and New Zealand, where dense reinforcement in concrete columns is required, including closely spaced steel hoops to encourage ductility (the ability to deform after yield without severe damage or collapse). Hence the most modern Japanese buildings suffered only minor, non-structural damage in the Kobe earthquake, and, with the exception of a single building, did not collapse.


Buildings and transportation structures that collapsed were mostly, but not exclusively, designed and built to less stringent standards than required by modern Japanese codes, reflecting the progressive move towards better earthquake-resistant design technology over the past 30 years or so. For example, recorded damage to reinforced concrete buildings2 shows that 57 per cent of those constructed before 1971 suffered collapse or very severe damage; this reduces to 21 per cent for those built between 1971 and 1980, and to near zero for those built to the most modern post-1981 earthquake code standards. The damage to the latter group of structures was estimated overall to have been only one-fifth of that which occurred in the older, pre-1971 buildings.


The damage levels in pre-1971 buildings in the worst-hit parts of Kobe are representative of those to be expected in any future earthquake in a large Japanese or US city, and in other earthquake-prone regions of the world. The proportion of damaged buildings is similar to, or even higher than, that from earthquakes during the past 15 years in Mexico, Armenia, Turkey and other parts of the world that are considered much less well developed than Japan or the United States in terms of earthquake-resistant building technology2. This disturbing result explains why there are important questions being asked throughout the world about the adequacy of strengthening and replacement policies for older buildings in earthquake regions3.


The Kobe earthquake, together with that exactly one year earlier in Northridge, California (moment magnitude 6.7), will therefore be remembered as being of special sig nificance largely for three reasons. First, it caused an unexpected level of damage in a part of the world generally thought to be prepared for earthquakes; second, the direct economic damage costs from the Kobe event far exceeded those of any previous earthquakes; and third (but perhaps most important), Kobe has posed questions about the rate at and urgency with which older buildings, clearly much more vulnerable to earthquake damage and collapse, are being replaced or adequately strengthened, in Japan and elsewhere.


Figure 2 Severely damaged steel-framed building in Kobe.


Unexpected engineering failures


The age of buildings that failed cannot alone explain the high damage levels to engineered buildings in the Kobe earthquake. There were also some notable collapses and failures in modern steel and concrete high-rise buildings, and these are prompting a radical and far-reaching rethink of design methodologies for earthquake-resistant structures3. There are four important areas being subjected to this critical scrutiny.


First, there was a disturbingly high proportion of failures in steel buildings in Kobe (Figs 1,2). Damage to welded plate connections in the bracing members of several high-rise buildings was unexpectedly severe, and in several cases buildings could have collapsed if the duration of strong shaking had been a few seconds longer. This type of damage, which could have had catastrophic consequences, echoed the concern over similar types of damage to steel buildings recorded in the 1994 Northridge earthquake4. Some mid- and high-rise steel condominium structures incorporated innovative structural systems consisting of steel frames in which the column and girder members were large steel trusses. Damage observed included the brittle fracture of tubular columns and fracturing of wide-flange diagonal bracing elements; such damage also occurred in taller, commercial steel structures.


Second, there is cause for concern that this was an earthquake of only moderate magnitude, yet owing to the close proximity of Kobe to the event's epicentre, buildings would have been subjected to forces between 150 and 250 per cent of their designed ultimate resistance. The effect of a much stronger earthquake on one of the large cities of Japan or California could be expected to be yet more devastating5. Building codes do not consider explicitly the effect of very close proximity earthquakes (so-called 'direct hits') and may tend, as a consequence, to underestimate the exceedingly high localized ground accelerations close to the fault slip, even in moderate-magnitude earthquakes.


Engineers, particularly from the United States and Europe, are therefore increasingly concerned that existing codes do not fully account for the distinct features of near-field large-magnitude earthquakes. In particular, these are the exceptionally high horizontal ground accelerations, some approaching 100 per cent of g; coupling of the horizontal ground accelerations with vertical accelerations that in Kobe were, in some instances, one-and-a-half times the horizontal movement values (these vertical motions are ignored by codes and can lead to failure of columns in buildings and bridges by inducing unaccounted-for tensile forces); and low-frequency 'fling' or 'killer-pulse' ground motions near the earthquake fault, which although of relatively low acceleration can produce excessively large deflections in stiff structures such as elevated bridges and buildings with reinforced concrete structural walls, by acting like a large horizontal static force applied to the structure over a duration of perhaps one or two seconds, sufficient in many cases to cause the structure to suffer severe damage and in the worst cases to collapse.


Third, the use of steel/reinforced concrete columns in combination with conventional reinforced concrete, as advocated in Japanese building codes, has been brought into question. The former are modern buildings formed of I-section or multiple T-section columns surrounded by mass concrete or reinforced concrete. The main problem appears to lie in the difficulty of connection between the two types of construction, with steel/reinforced concrete generally being used in Japan only up to the sixth-floor level and reinforced concrete construction from the seventh floor upwards. This was probably the main cause of the unusually high number of mid-height failures of modern concrete buildings in Kobe (Fig. 3).


Finally, much of the newer construction in Kobe, particularly of larger buildings, is on very soft, recent alluvial soil and on recently constructed near-shore islands. Most of the serious earthquake damage to larger commercial and industrial buildings and the transportation infrastructure occurred in areas of soft soils and reclaimed land. The worst industrial damage occurred at or near the waterfront, owing to ground failures arising from liquefaction, lateral spreading and settlement. Although the engineering profession has tried hard to develop methods for strengthening filled areas to resist failures during earthquakes, most have been put into practice without the benefit of being adequately tested in strong earthquakes. The performance of these techniques was mixed, but the failures costly-most retaining walls along the port failed, and the consequent ground settlement led to damage of many buildings and port structures.


Innovative structural solutions


Because of the perceived lower risk of earthquakes in Kobe than in Tokyo, only two buildings in the affected region were constructed with base isolators, a technique that uses high-damping rubber bearings to support the columns and thereby reduces the level offerees transmitted to the structure by the ground motions. Both these buildings survived the earthquake undamaged, but given their distance from the epicentre (more than 35 km), neither experienced the full force of the seismic loadings. Nevertheless, designers were encouraged by the observation that the isolators performed exactly as intended, and the buildings were completely in service after the earthquake.


There is great unexploited potential to incorporate this and other innovative construction methods into new buildings for earthquake regions. The additional cost considerably exceeds that of providing conventional earthquake resistance, but may easily be justified on the basis of improved performance and lower repair costs, particularly for vital structures and facilities such as hospitals and fire stations.


Base isolation is one of an increasing number of methods of modifying structures (including the upgrading or retrofitting of existing buildings, discussed below) to mitigate the disastrous effects of strong earthquakes on buildings and occupants. Many passive-damping (energy-dissipating) devices and active control methods using computer-controlled hydraulic actuators strategically placed in or around the building have been researched in Japan, the United States, New Zealand and elsewhere over the past 20 years. All have their problems, some related to additional costs which may amount to 10-20 per cent of the total construction bill (although compared with potential post-earthquake repair or reinstatement costs for conventional buildings, these are clearly acceptable), whereas others relate to the maintenance and operation of sophisticated methods (particularly active control, which relies on computer-controlled devices), during the lead up to and occurrence of a major earthquake.


Even in Japan, where the industrial research corporations have developed and advanced the technology of these modern methods of earthquake control more, perhaps, than anywhere else, the number of buildings using a form of response control (active or passive) remains small, being considerably less than 1 per cent of the total building stock at risk. At government-funded earthquake research centres in the United States, New Zealand and Europe, there has been a massive expansion of interest in the past 10 years in the development and testing of both passive and active control methods. What is lacking is a sufficiently widespread will among building owners and developers to implement these techniques and to fund the additional up-front costs or the supplementary costs to upgrade existing vulnerable structures. This may partly be the result of the sheer diversity of methods now available, with no single technique being sufficiently effective or versatile to be considered the clear market leader.


Figure 3 Concrete-frame structure with a mid-height storey collapse.


What for the future?


There is a widespread view among engineers that owners and developers must demonstrate greater willingness in the future to pay for proper engineering design and construction to ensure a safe structure. The cost of adequate earthquake resistance is very low - only a few per cent of the building cost - if it is designed into the structure initially, but very high if the building must be upgraded at a later date. In fact, retrofitting an existing building can cost almost as much as constructing a new one. For a large city in Japan or California, the total costs of upgrading all building stock with an effective useful life of upwards of 50 years could be enormous, perhaps $50-100 billion. Clearly, this cost and the associated construction programme would have to be spread over many years, and could not be funded by conventional government or private sources. Despite this apparent financial deterrent, the lesson from Kobe is that government economists and urban planners must urgently upgrade older, more vulnerable building stock and infrastructure, which has been shown to suffer up to 10 times more earthquake damage than structures built to modern design-code standards in Japan and the United States.


The catastrophic damage and disruption caused in Kobe can be taken as an illustration of the eventual cost of ignoring the problem of older, vulnerable buildings in large cities in earthquake regions. In this context, the expense of upgrading could be regarded as an acceptable price to pay. Engineers have the technology and capability to implement such a programme, but so far the wholehearted political will to go ahead has been lacking. Perhaps Kobe will provide the stimulus for a swing of opinion, but memories are short, and action is required sooner rather than later to prevent a repeat in other earthquake-threatened cities.


Elnashai, , A. S.HEI News No. 15, 6-7 (1996).Earthquake Engineering Field Investigation Team (EEFIT) The Kobe, Japan Earthquake of 17 January 1995 (Inst. Struct. Eng., London, May 1997).Hayward, , D.New Civil Eng.16-20 (13 June 1996).Earthquake Engineering Field Investigation Team (EEFIT) The Northridge, California Earthquake of 17 January 1994 (Inst. Struct. Eng., London, May 1997).Scawthorn, , C.et al.EQE Summary Report (EQE International, San Francisco, April 1995).