Cover Image
close this bookCrucibles of Hazard: Mega-Cities and Disasters in Transition (UNU, 1999, 544 pages)
View the document(introduction...)
View the documentAcknowledgements
View the document1. Introduction - James K. Mitchell
View the document2. Natural disasters in the context of mega-cities - James K. Mitchell
View the document3. Urbanization and disaster mitigation in Tokyo - Yoshio Kumagai and Yoshiteru Nojima
View the document4. Flood hazard in Seoul: A preliminary assessment - Kwi-Gon Kim
View the document5. Environmental hazards in Dhaka - Saleemul Huq
View the document6. Natural and anthropogenic hazards in the Sydney sprawl: Is the city sustainable? - John Handmer
View the document7. Disaster response in London: A case of learning constrained by history and experience - Dennis J. Parker
View the document8. Lima, Peru: Underdevelopment and vulnerability to hazards in the city of the kings - Anthony Oliver-Smith
View the document9. Social vulnerability to disasters in Mexico City: An assessment method - Sergio Puente
View the document10. Natural hazards of the San Francisco Bay mega-city: Trial by earthquake, wind, and fire - Rutherford H. Platt
View the document11. There are worse things than earthquakes: Hazard vulnerability and mitigation capacity in Greater Los Angeles - Ben Wisner
View the document12. Environmental hazards and interest group coalitions: Metropolitan Miami after hurricane Andrew - William D. Solecki
View the document13. Findings and conclusions - James K. Mitchell
View the documentPostscript: The role of hazards in urban policy at the millennium - James K. Mitchell
View the documentAppendices
View the documentContributors
View the documentOther titles of interest

13. Findings and conclusions - James K. Mitchell

"Urbanization holds out the bright promise of an unequaled future and the grave threat of unparalleled disaster."

(Wally N'Dow, Secretary-General, Habitat II; quoted in United Nations
Commission on Human Settlements, 1996, p. xxi)

The preceding case-studies provide convincing testimony about the importance of environmental hazards as mega-city problems. It is now time to bring together the main findings and to discuss their implications for researchers, managers, and others who will share responsibility for guiding the development of mega-cities in the twenty-first century. In the pages that follow, emphasis will be placed on the role of natural hazards because they raise some of the most difficult intellectual and policy issues and because they have been largely neglected by the global community of urban experts. This does not imply that other kinds of hazard are lacking in importance, for the evidence shows that many types of urban environmental threat are worsening and all are becoming increasingly intertwined. However, a natural hazards perspective brings the issue of limits to human agency clearly into view at the same time as it illuminates the action-forcing nature of disjunctive events. In discussions of technological and social hazards, these matters are often obscured by expansive interpretations of human agency and by the privileging of structures (especially social structures) as causal factors.

Herein it will be argued that mega-city hazards are profuse, burdensome, symbolically potent, incompletely understood, and addressed by public policies that typically make use of just a few types of possible adjustment. They are also closely bound up with many other urban issues and have acted as catalysts for change in a wide range of policy contexts. Despite noteworthy successes in some places, the bulk of existing mega city hazard-management systems are either barely holding their own or seriously deficient. In the future, all such systems will be grievously challenged by the changing nature of hazards themselves and by the profound transformations of large cities that are now occurring throughout the world. The consequences of misjudging natural hazards in mega-cities of the twenty-first century are likely to be worse than anything yet experienced.

Hazards in 10 mega-cities: A review

The 10 mega-cities examined in this volume are a diverse group. Not only do they differ in terms of culture, population, area, growth rates, economic status, and political functions (see chap. 1); they are also subject to varied mixes of hazards. Consider, for example, Tokyo and Sydney. Though both are affluent industrialized cities of roughly similar age, their hazard profiles are quite different. Tokyo has a history of catastrophic destruction and its residents are much concerned with staving off future disasters. Here earthquakes, snowstorms, and typhoon-driven floods threaten one of the world's chief business centres - a densely built-up city that is distinguished by remarkably little social pathology. Sydney, in contrast, has historically been free from truly catastrophic natural disasters but is increasingly beset by technological risks and social tensions. Bush-fires and flash-floods now affect what is arguably the world's most dispersed mega-city - a place whose residents pay only selective and intermittent attention to environmental hazards. Equally compelling contrasts mark the hazard profiles of Dhaka and Lima, Seoul and Los Angeles, London and San Francisco, Mexico City and Miami. Such a wide spectrum of settings and experiences provides a rich array of knowledge about the ways in which metropolitan societies have sought to come to terms with hazardous environments.

Not only are the mega-cities diverse; they are beset by complex hazard problems:

· the agents of hazard are many and the mixes varied;

· relationships between hazards and competing urban problems vary widely among different cities;

· urban hazards issues and interest groups are volatile, especially in the largest cities;

· human-modified natural processes of urban areas are imperfectly understood;

· although they are typically viewed as "public policy problems," mega-city hazards are also underlain by contentious scientific and philosophical issues about the nature of change in urban environments and about the role of nature in human-constructed settings.

In the crucible of the mega-city, those ingredients interact and are transformed, giving rise to distinctive patterns of hazard.

Mega-city hazards are profuse

Environmental hazards are present in profusion among all of the case-study cities. Every one is affected by at least two or three different kinds of natural hazard and some (e.g. Tokyo, Mexico City) by as many as seven or eight. Floods, earthquakes, and windstorms are the most common damaging phenomena; different pairs of these three constitute major risks in all of the places studied (e.g. floods and windstorms in Dhaka; earthquakes and floods in Los Angeles; windstorms and earthquakes in Tokyo). Other risks that have triggered significant disasters include (in descending order): (1) slope failures; (2) droughts/water shortages; (3) wildfires, subsidence, and fog; (4) tsunamis; (5) volcanoes; and (6) snow. Fire has been the most common and destructive human-made urban hazard, especially fire as a consequence of external or internal warfare, though natural factors also play a role. Among the technological hazards, air pollution is perhaps the most widespread, the most frequently recognized, and the most generally destructive (UNEP/WHO, 1994), while terrorism and other violent crimes are important social hazards in most mega-cities.

Mega-city hazards are important

For many urban dwellers, either natural disasters are historical curiosities that, it is hoped, will not be repeated or they are improbable fantasies that it is believed will never leave the movie screen. But the record of actual events is both recent and very real. Hazards are not only present, they are undeniably important phenomena both inside the case-study cities and beyond their boundaries.

Losses of life and property are one index of importance. During the twentieth century, human fatalities attributable to single extreme events have been as high as tens to hundreds of thousands in large cities (e.g. Tokyo in 1923; Mexico City in 1985). Moreover, the potential for catastrophic urban death tolls is growing rapidly. But the material and economic impacts of environmental hazards in big cities loom even larger. For example, more than 40 per cent of Tokyo was completely destroyed by earthquakes and fires on two occasions in the past 75 years. Recent record-setting bills for hurricane damage in Miami (US$ 30 billion) and wildfire in Oakland (US$ 2 billion) and Kobe (over US$ 100 billion) testify to the potential for even larger economic costs. Of course, economic loss totals tend to be biased in favour of richer cities - which have more expensive properties at stake. But even modest-seeming losses can be devastating to mega-cities of the third world, as evidenced by the experience of Mexico City. Though by no means the most expensive earthquake to affect a large city, the one that struck Mexico City in 1985 came at a time when the country was facing severe economic problems. Resources were diverted to recovery in the city, thereby hampering national development for a number of years.

Between 1985 and 1995 the myth that modern cities have successfully conquered the hazards of Nature was refuted by a series of massive losses (see table 1.1 above). During this period, earthquakes, hurricanes, wildfires, and other natural extremes inflicted human casualties and record property damage on half a dozen major metropolitan areas in the United States, Mexico, and Japan. In several of these the economic losses reached proportions that had serious implications for national budgets and the stability of the global reinsurance system.

The recent record of technological hazards and disasters in mega-cities is more ambiguous. Chronic technological hazards, especially those associated with locally generated environmental pollutants, are, of course, almost synonymous with large cities and have attracted a great deal of attention from analysts in recent years. Some acute technological disasters also stand out in mega-cities, including explosions of gas fuel tanks in Mexico City (1984) and of sewer lines in Guadalajara (1992), as well as Bhopal's catastrophic chemical leak (1984). Large cities have been hit by massive and prolonged electricity power failures (e.g. New York in 1965 and 1977; London in 1987), which sometimes spawned secondary disasters. However, as far as the mega-cities in this sample are concerned, acute technological disasters seem to have been less deadly and less damaging than natural disasters. Much the same might be said of mega-city epidemics and other biological hazards, although major scares such as outbreaks of hepatitis in Lima and plague in Surat (India) suggest that worse is possible. Of course, like the rest of the world, mega-cities are at risk from technological and biological threats that occur on continental and global scales, such as the Chernobyl nuclear accident, the AIDS pandemic, and the spectre of global climate change.

If the mega-city record of technological and biological hazards is ambiguous, the same cannot be said about urban social hazards. In particular, there is growing concern about the vulnerability of large cities to acts of terrorism, internal warfare, and rancorous public protest. Table 13.1 identifies some of the larger cities that have recently experienced these kinds of event. Big-city facilities have long been prime targets of inter-state warfare, internal coups, and popular anti-government uprisings, but they are now also becoming battlegrounds in a whole new generation of problems. These include: conflicts among ethnic, religious, and other groups recently freed from the constraints of Cold War politics; as well as deprived or alienated members of a growing urban underclass. Access to powerful weapons of mass destruction and the availability of soft targets such as easily disrupted electronic information technologies and urban infrastructures are contributory factors. Thus far, public buildings, business offices, shopping centres, and transportation facilities have been the most common targets, but there are fears that basic life-support systems (e.g. food, water, and energy supplies) could easily be disrupted - or hazardous facilities deliberately destroyed - with catastrophic results.

Table 13.1 Cities with major incidents of social violence since 1989

Civil/internal war or urban terrorism

Riots or street protests by civilian population

External warfare

Baku

Madrid

Belgrade

Baghdad

Beijing

Manchester

Bombay


Bogota

Mogadishu

Calcutta


Buenos Aires

Monrovia

Dhaka


Cairo

Moscow

Jakarta


Colombo

New York

Los Angeles


Kabul

Paris

Yangon


Karachi

Phnom Penh



Kinshasha

Port au Prince



Lahore

Tibilisi



Lima

Tokyo



London




Note: Case-study cities in italic.

Losses are not the only measure of importance. The malign ubiquity of extreme events is manifest in other ways. Among the case-study cities are several in which nearly every occupant was affected negatively by extreme events for extended periods. The experience of Dhaka in 1987 and 1988, when three-quarters of the city was covered by floodwaters for many weeks, is a case in point. It is not so much that some Dhaka residents died, many lost their possessions, and others contracted waterborne diseases, or that space satellite images of the submerged city quickly conveyed its plight to an international audience of concerned scientists and curious laypeople. Rather, it is that the floods also became yardsticks for measuring other experiences in the lives of victims and managers alike. In other words, they become marker events for resident populations.

Very often such natural disasters have ushered in major changes of built environments and far-reaching institutional reforms. Lima was almost entirely reconstructed after the catastrophic earthquake of 1746. The pivotal significance of the 1906 earthquake for the city of San Francisco is too well known to warrant further comment, but more recent disasters have also reconfigured the landscape of other Bay Area communities. For example, the Loma Prieta earthquake (1989) radically changed both the Marina District of San Francisco and the lowlands of Oakland, while the firestorm of 1992 extensively remade upscale neighbourhoods in the hills above the city of Oakland. The urban fabric of modern Tokyo is largely a product of rebuilding after the 1923 Kanto earthquake and the fire bombing raids of the Second World War. Miami hurricanes of the 1940s gave a strong push to massive water-management schemes that encouraged metropolitan expansion, and the hurricane of 1992 significantly accelerated the Hispanicization of Dade County's population. In London, the inner city's physical layout owes much to the Great Fire (1666)1 and the economic decline of the port was sharply assisted by the 1940s' blitz. More recently, a catastrophic 1953 east coast storm surge paved the way for construction of the Thames flood barrier, which, in turn, helped to secure investments in the rejuvenated Docklands development area of east London. Mexico City's 1985 earthquake provided opportunities for grass-roots citizen organizations to take on responsibilities for relief and reconstruction, which had formerly been the exclusive preserve of government agencies, and to use the experience as a base for challenging government authority in other realms of public policy.

1 New street patterns and new construction materials both appeared in the wake of the fire. The changes would have been more profound if Charles II had been both rich enough and powerful enough to impose a formal reconstruction programme on the city. As it was, massive Baroque-style reconstruction schemes proposed by Sir Christopher Wren and others were never carried out (Knox, 1993, p. 92).

Some of the changes that took place in the wake of natural disasters had effects well beyond the affected cities. For example, the 1933 Long Beach (Los Angeles) earthquake triggered state legislation (The Field Act) that subsequently revolutionized school building standards and siting criteria throughout California, as well as inspiring similar changes elsewhere. The Kanto earthquake promoted the development of country-wide civil defence programmes and the creation of a national disaster day (1 September), which is marked throughout Japan by emergency-management exercises and simulations that involve tens of thousands of citizens.

Unfortunately, changes that occurred in the wake of major disasters were not always beneficial. Rubble from the Tokyo fire of 1657 was used to create filled land in the Koto Delta, which is now one of the most troublesome earthquake- and flood-prone locations in Tokyo. After the Second World War, many of Tokyo's bombed-out neighbourhoods were filled with cheaply constructed wooden housing that remains a perennial problem for today's hazard managers. Fiscal policies adopted by the Dhaka Municipal Committee of 1840, in an effort to address a crisis of sanitation and economic decline, helped push residents out of the old city into adjacent flood-prone areas; the new polders that have recently been constructed near Dhaka may also increase the city's catastrophe potential while providing needed space for expansion. The much-afflicted Marina District of San Francisco is founded on rubble from the 1906 earthquake and from a subsequent Exposition that was held to mark the city's resurrection. Aqueducts that were intended to reduce the earthquake vulnerability of San Francisco's water-supply system cross the very faults that will likely rupture during future quakes. Drainage works undertaken during the nineteenth century in response to historic flooding have accelerated soil compaction and subsidence in much of Mexico City. The lesson is unmistakable: when a big disaster affects a major metropolitan area its consequences can be permanent and profound - both positively and negatively - for the city itself and for the larger society.

On occasion, natural disasters may turn into something more than just marker events - they may become metaphors that place environmental hazard near the centre of a city's identity. The conjunction of natural and human disaster impacts in Los Angeles in 1992 (i.e. earthquakes in the flatlands, wildfires in the hills, and riots in the ghettos) not only inflicted record-setting economic losses but reinforced the city's civic reputation as a place of environmental and social insensitivity that continually flirts with apocalyptic demise. Although Los Angeles may be the most flamboyant example of a mega-city that has embraced disaster as a central component of urban identity, it is not the only one. Foundational disaster imagery and myth-making are also present to varying degrees in Tokyo, Seoul, San Francisco, Dhaka, Mexico City, Lima, and perhaps elsewhere. It is not possible here fully to assess the implications of such characterizations for urban populations, but they should not be neglected in the formulation of future urban strategies and policies - especially those that have to do with hazard management (see the final sections of this chapter).

The management of mega-city hazards is truncated

Certain mega-cities (e.g. Greater San Francisco and Tokyo) possess civic cultures that are both sensitive to hazards issues and active in mounting appropriate responses, but most city governments and institutions are far less committed to the containment and reduction of hazards as primary goals of public policy. Typically, the gap that separates awareness of risks from responses to mega-city hazards is wide and is filled with mediating factors that tend to retard the adoption of improved public policies. Existing public policies strongly favour professionalized warning, evacuation, and emergency-management programmes for a wide range of acute threats backed up by separate sophisticated engineering technologies for different chronic risks. London's movable flood barrier, Dhaka's embankments and polders, Seoul's flood pumping stations, and Los Angeles' sediment catchment basins are typical examples of mega-city flood adjustments. Hazard insurance and hazard-resistant building practices also are sometimes available in the more affluent mega-cities. In most places the emphasis is on expert-driven systems and formal procedures carried out by public institutions, although a high degree of co operation by well-informed lay citizens is often expected. Lay residents of large cities also participate in informal, traditional, or folk measures, such as networks for providing mutual aid.

Many improvements to the formal public adjustments are possible, including the upgrading of emergency services and the installation of hazard-warning and evacuation technologies in cities that do not yet possess them, as well as the development of appropriate methodologies for assessing hazards and incorporating risk-management strategies into public budgets, plans, statutes, and other regulatory devices. However, even in relatively well-provisioned mega-cities of Japan, North America, and Europe, the areal and demographic coverage of formal public sector hazard-management programmes is incomplete, and the extent to which they address the premier hazard concerns of resident populations is often uncertain.

What else that might be done remains missing from the preferred range of management alternatives? Broadly speaking, the neglected approaches involve non-expert systems, informal procedures, non-structural technologies, private sector institutions, and actions taken by individuals, families, neighbourhood groups, firms, and similar entities. Among others, these include measures that: (1) encourage hazard-sensitive decisions about site selection, land management, and facility operations; (2) control the installation and replacement of infrastructure; (3) relieve institutional and social inequities that shift hazard burdens onto certain (already disadvantaged) groups; (4) buttress local grass-roots capacities for hazard management; and (5) promote less environmentally stressful non-structural hazard-mitigation technologies. In addition, there is a lack of initiatives that jointly address different kinds of hazard, a slowness to integrate hazards management with other problem-solving urban programmes, and a failure to investigate other roles that hazard plays in the lives of urban residents.

It is not that the latter approaches are entirely absent in mega-cities. For example, local hazard-management programmes staffed by laypersons have made useful contributions to hazard reduction in Mexico City and Los Angeles. Neighbourhood hazard volunteer organizations have been pioneered in Tokyo, along with local stores that sell emergency supplies and community centres that stock low-techology emergency-management equipment. Internet pages now carry lists of profit-making companies that make a living from serving the hazard-protection needs of residents in some US cities, and some commercial organizations have established trade shows of hazard-management technology that is suitable for private as well as public sector adoption. However, these kinds of facilities and activities by no means exhaust the list of alternatives that are possible and they today account for only a small minority of existing responses.

Much is still unknown about mega-city hazards

Of the four basic components of hazard (i.e. risk, exposure, vulnerability, response), information about risks and responses in mega-cities is generally more plentiful and accurate than is information about exposure and vulnerability. Likewise, the information base for mega-cities in more developed countries (MDCs) is typically better than for mega-cities in less developed countries (LDCs). But the relationship between information availability and public policy is not a simple matter and is continually changing under the impress of scientific investigations, the experience of disasters, and fluctuations in public anxiety about different types of hazard.

Information about risks and responses is relatively good in the mega-cities of MDCs, at least with respect to long-established natural hazards. For example, the general probability of earthquakes and of areas susceptible to destructive ground motion is well known in places such as Tokyo, San Francisco, and Los Angeles, though the details are far from complete. Large faults are well documented but many important small faults go undetected until earthquakes occur, and geologists are only now beginning to explore the complex interference ground motion patterns that appear in lowland basins when seismic waves are reflected off surrounding mountains. Likewise, the information necessary to construct reliable earthquake-resistant high-rise buildings is readily available, but much less is known about fail-safe construction practices for highway overpasses, underground infrastructure systems, and other lifelines (Platt, 1991). Compared with risks and responses, exposure and vulnerability are highly volatile components of urban hazard that sometimes respond quickly to demographic and land-use changes (Montz and Gruntfest, 1986). Prospects for identifying changes in exposure have improved with the advent of remote sensing and GIS (Geographic Information System) technologies, although most MDC mega-cities are only now beginning to apply these tools and their potential is far from being realized (GeoHazards, 1994; Gruber and Haefner, 1995; Oppenheimer, 1994). Socio-spatial patterns of earthquake vulnerability are poorly known in virtually all MDC mega-cities.

Compared with the MDCs, many mega-cities in LDCs often lack very basic information about hazards. For example, data about flooding in Seoul and in Dhaka are sparse. Despite the fact that floods have claimed hundreds of lives in Seoul in recent years, the flood risk has not been assessed by public agencies or research organizations. The case-study of Seoul's flood risk reported in this volume is believed to be the first effort of its kind in the city's history and certainly the first to provide even rudimentary quantitative documentation of the relationship between land development practices and the severity of flooding. It is also clear that the growing flood hazards of Dhaka have not been analysed for purposes of disaster mitigation. International agencies and the national government have devoted much more attention to rural flooding in Bangladesh, although the city of Dhaka now contains around 8 million people and rapid infilling of open space in low-lying neighbourhoods is feeding a major potential for urban flood disaster. Likewise, the case-study of social and economic vulnerability to earthquakes in Mexico City represents the first known attempt to address this kind of knowledge gap in a mega-city of the developing world.

Information about informal groups and non-governmental organizations that are engaged in urban hazard management is also woefully deficient. Many such groups are national or international in scope (e.g. religious, fraternal, charitable, and public service organizations); some are indigenously urban (e.g. neighbourhood associations, urban cooperatives, local union welfare groups), others have been transferred from rural areas by migrants to the cities (e.g. family and locality kinship networks); and some are emergent groups that are still developing in situ, either in response to specific hazards or for other purposes. Given the growing importance of the informal economic sector in many large cities, especially throughout the LDCs, and corresponding shrinkages in the role of the state, especially in MDCs, such groups may become more important as hazard managers in the future. Of course, other outcomes are also possible. For example, cities have often acted as agents of "modernization," encouraging the replacement of informal organizations by formal ones and lay competence by professionalization. The struggle to create more effective hazard management in mega-cities may be as much a contest between outdated professional institutions and emergent ones as between experts and laypeople or among citizen groups that are differently empowered. These are matters that are not well understood at present.

The contrasts between scientific information about hazards in mega-cities of poor countries and rich countries are real enough but it would be a mistake to exaggerate them. The comparative information advantage of MDC mega-cities is eroding both because of improved scientific infrastructures in some LDCs and because of institutional barriers that impede further advances in the application of scientific information in many MDC mega-cities. For example, in Latin America scientific information about earthquake risks, micro-zonation, and structural vulnerability is improving rapidly, partly as a result of research initiatives stimulated by the International Decade for Natural Disaster Reduction (GeoHazards, 1994). Conversely, MDC mega-cities are often hampered by public institutions that lack the authority to apply available hazards information at the metropolitan scale. Some US cities, including San Francisco and Los Angeles, have developed metropolitan hazards-research and -management organizations but many others (e.g. Miami, New Orleans, Houston) do not possess them.

In certain important respects there are no significant information gaps between MDC and LDC mega-cities. Indeed, all mega-cities start from similar bases of ignorance when it comes to hazard "surprises." These unprecedented events are, by definition, poorly known everywhere. A great deal of additional research will be necessary before it is possible to assess the effect of surprises associated with climate change, new industrial chemicals, or hybrid hazards that combine natural, technological, and social risks. Likewise, there is a general lack of information about the relationship between hazards and other urban issues in all kinds of large cities. The complex nature of contemporary mega-cities virtually ensures that hazard policies and programmes rarely stand alone as claimants of public support. Competition among different urban constituencies with quite different agendas is the rule, and public policies inevitably involve compromises and trade-offs among responses to different issues. By the same token, the complexity of mega-cities affords scope for the creation of coalitions and collaborative ventures among unusual partners. Some contributors to this volume (e.g. Solecki in chap. 12) hint at the possibilities that changing coalitions might portend, but the general research literature is thus far largely silent about such matters. In summary, lack of information about mega-city hazards is clearly a serious problem, but it is often overshadowed by the volatility of the urban crucible, which continually recatalyses even the known components of hazard.

Mega-city hazards are changing

Although information gaps prevent a definitive assessment, it is very clear that mega-city hazards are in transition at the close of the twentieth century.2 Everywhere, the composition of hazards is changing; the management of hazards is changing; and the ways people think about hazards are changing (Mitchell, 1995). Nowhere is this more true than in the world's largest cities. At the same time, the global population is undergoing a vast demographic shift towards a predominantly urban world and the process of urbanization is itself in flux. Some of the more important changes are summarized in table 13.2.

2 Parts of this section have been summarized in Mitchell (1995).

Table 13.2 Trends of mega-city hazards

* Mixes of natural, technological, and social hazards are increasingly common

* Risks are changing slowly

* Loci of hazard are shifting markedly

* Differentially vulnerable groups are becoming polarized and segregated

* Public support for hazard-management initiatives may be faltering

* Overlaps among hazard and urbanization issues offer opportunities for managerial intervention

Changes of interactivity

Urban hazards and disasters are becoming an interactive mix of natural, technological, and social events. In the past, these events were analysed as separate types of phenomena, but that approach may no longer be desirable or possible. Most potential victims - and many hazard-management agencies - do not make such distinctions; for laypersons, urban hazard is usually a composite and unstable class of events. Experience is carried over from one type of event to another and coping measures frequently straddle several types of hazard.

Interactivity occurs in different ways. Most commonly there is a burgeoning of hybrid hazards composed of different mixes of natural, technological, and social risks.

Recent representative examples from the United States include: fiery floods in the Houston metropolitan area of Texas; the incorporation of toxic chemicals into Mississippi flood detritus in St. Louis; and problems in Chicago's central business district involving the collapse of outworn canal walls, inundation of derelict underground rail tunnels, and subsequent electrical power failures. More of these kinds of event are likely to occur in the future, thereby raising serious difficulties for reductionist analyses that treat the components of complex hazards as if they were separate phenomena. In any event, the sheer numbers and diversity of mega-city hazards mean that populations at risk customarily must simultaneously juggle with the prospects of many different kinds of threat (see chap. 1). Some of what urban residents know about hazards comes from direct experience and visual inspection, but many of the cues are indirect and are derived from images of risk and hazard projected by mass media sources for purposes that are not connected with hazard management (e.g. partisan politics, advertising, entertainment). The hybridization of mega-city hazards is not solely an objective process that involves overlaps between risks that are both active and real; it is also the subjective conflation of experienced and imagined events and the reinvention of past hazards to fit the explanations of the present.

Interactivity need not require direct physical involvement among different kinds of event; it can occur indirectly, as when experience gained with one type of hazard informs adjustments to another. The widening cascade of effects that flow from urban terrorist campaigns provides a good example of this form of interactivity. Terrorist incidents have raised the hazard awareness of residents in places such as London, New York, and Paris in complex ways. For example, bombings of New York's World Trade Center (26 February 1993), the Federal Building in Oklahoma City (19 April 1995), and the Olympic Park in Atlanta (27 July 1996) have not only led to tighter security precautions at major public facilities throughout the United States but raised the general visibility of emergency-management agencies. Thus, in successive State of the Union speeches (1995, 1996) made by the US President, the Federal Emergency Management Agency has become both a lauded exemplar of compassion and efficiency in government and a pivot of political campaigns to encourage citizens to take increased responsibility for their own safety. Anxiety about the general vulnerability of densely populated urban areas was also fuelled by poison gas attacks carried out by the Aum Shinrikyo cult in Japan (20 March 1995). In London, Irish Republican Army bombings and fires in the Underground railway system have had repercussions for public safety measures that resulted in increased security against a wide range of natural and human extremes (e.g. heightened readiness among public safety services, upgraded facilities and techniques for treating traumatic injuries, more stringent building standards). They have also affected the business property insurance market and have encouraged owners and tenants of risk-susceptible buildings to give public safety a prominent place among their priorities. With urban terrorism a continuing problem in mega-cities from Colombo to Lima, the prospects for increased adoption of overlapping or mutually supportive adjustments to natural, technological, and social threats have probably increased.

Changes of risk

Risk (as measured by the probability of experiencing an extreme event) is among the more stable components of urban hazard. That is not to say that the urban physical environment remains unchanged. In many cities, geological substrata are sinking, river beds are rising in response to accelerated sedimentation, natural flood-detention systems are being converted to other uses, and vegetated slopes are being cleared for housing developments. For example, subsidence in response to underground water withdrawals is a continuing problem in Mexico City. In Lima, the bed of the Rimac River is slowly rising as sediment accumulates in the channel; seepage from the elevated river is weakening foundations and walls of nearby buildings, which are predominantly built of dried mud. The disappearance of open spaces to provide building lots for a voracious housing market is held to be directly connected with increased runoff and flood losses in Seoul. Infilling of floodwater-detention ponds in Dhaka is also believed to be cutting the time between low water and peak discharge during major floods and is thought to be increasing the damage potential of lesser floods. Moreover, the spectre of global climate change and sealevel rise threatens a significant number of large cities. However, despite their continuing importance, changes in physical risks probably pose fewer problems and fewer uncertainties for urban populations than do changes in the other three components of hazard.

Changes of exposure

Large changes in exposure and vulnerability are perhaps the most striking of the case-study findings. In virtually all of the mega-cities that were analysed, the locus of hazard is shifting - often rapidly. For example, despite popular beliefs that associate the city of San Francisco with catastrophic earthquakes, other communities on the eastern fringes of the mega-city now have much larger potentials for disaster. In Tokyo, the inner and outer suburbs are increasingly at risk because of both rapid expansion of the city's periphery and lack of attention to hazard-sensitive design in the newer developments. By contrast, the central city and adjacent neighbourhoods have been the focus of major investments in emergency preparedness, hazard mitigation, and other disaster-reduction alternatives. Unlike North American cities, which appear to be dying from the centre towards the periphery - with parallel increases in exposure and vulnerability to hazard - Tokyo appears to be renewing itself and upgrading its disaster resilience from the centre outwards. In Greater London, a shortage of buildable land since the Second World War has encouraged occupance of previously open floodplains everywhere except in a protected green belt and despite land-use planning controls. At the same time there has been a movement of the locus of economic investment and reinvestment downstream (eastward) along the Thames Valley. This is making London increasingly more susceptible to tidal flooding, especially if predictions about sealevel rise (connected with global climate change) are confirmed by subsequent events.

Changes of vulnerability

Increasingly, polarization and spatial segregation of groups that have different degrees of vulnerability to disaster are becoming the norms for most cities. In almost every case it is clear that poor and rich urban neighbourhoods are diverging with respect to the types of hazard they face and the degrees of risk they are exposed to. Contrasting types of hazard in rich and poor areas are most obvious in the mega-cities of developing countries (e.g. Mexico City, Lima, Dhaka), where the sheer volume of recently arrived poor migrants from rural areas and their association with hazardous marginal lands are also prominent features. Available information does not permit a definitive assessment of polarization and hazard segregation in Tokyo and Seoul, but there too evidence suggests that "safety gaps" are opening up between different social groups. Polarization and segregation with respect to income and risk are also increasingly visible in European and North American mega-cities. The contrast between poor earthquake-prone inner-city populations in Los Angeles and affluent suburbs at risk from wildfires and slope failures has received much attention from analysts both in this book (chap. 11) and elsewhere (Davis, 1995; Kirp, 1997). Similar but less well-known contrasts characterize Greater Miami, where poor south-side neighbourhoods bore the brunt of hurricane Andrew whereas the richer north side escaped relatively unscathed. Indeed, as far as the hazard consequences of polarization and segregation are concerned, it can be argued that the cities of the developed world are beginning to resemble those of the developing world.

Changes in efficacy of hazard management

Responses to hazard include measures adopted by formal public and private institutions and adjustments adopted by private individuals. Most of the case-study authors report increasing difficulty in developing and sustaining public support for hazard-management initiatives. In mega-cities, patterns of response to hazard are complex and varied. Some cities (e.g. Tokyo) have adopted multiple reinforcing public adjustments, involving sophisticated technologies, backed up by high levels of training and self-reliance on the part of civilian populations that together are designed to address all parts of the disaster cycle (i.e. preparedness, emergency management, recovery, mitigation). Others possess no effective formal programmes for natural-hazard management (e.g. Dhaka). Most occupy intermediate positions on the spectrum of response. However, hazard reduction appears to be losing ground almost everywhere in recent years, even in the best-protected cities of Japan, Europe, and North America. For example, the pace of hazard-related legislation and associated administrative actions has slowed noticeably in Tokyo since the 1970s; so too has the construction of large "hazard-proof " building complexes. Whether the disastrous effects of the 1995 Hanshin earthquake will undermine public confidence in existing Japanese approaches to hazard management or bring about renewed investments in alternative measures remains to be seen. In London, there has been a decided weakening of government support for planning and development controls, including those that affect the use of hazard-prone land and funding for emergency services. Despite Sydney's long-established formal commitment to principles of comprehensive planning and significant investment in institutions that respond to floods, bush-fires, and other sudden emergencies, the city seems unable to grapple with the chronic hazards associated with long-term environmental deterioration.

Counterposed against the very real problems of mounting actions that are exclusively focused on hazards are some promising opportunities for jointly attacking hazards and other overlapping urban issues. In such cases, shared constituencies help to keep hazard reduction high on the public agenda. Moreover, it is when hazard issues overlap with other issues that analysts have opportunities to examine the link between hazard systems and the larger problematiques of which they are a part.

The momentum to change public policy or to find the resources necessary to implement existing policy is usually short lived in most large cities. It is often present for a limited period during a major disaster but wanes quickly thereafter. Although disaster recovery and improved preparedness can stand alone as public issues during these periods because they command the support of many victims or other affected groups, it is often difficult to sustain public interest and involvement in disaster policy-making for very long at other times. However, when hazard issues overlap with other issues there is considerable potential for mobilizing a large and continuing joint constituency. The post-hurricane rebuilding of Greater Miami was effectively assisted by the joining together of separate groups concerned about poverty, immigration, migrant labour, crime, national defence, the peacetime role of military forces, mass transit, building code enforcement, intergovernmental relations, and tourism, among other matters (Mitchell, 1995). Though the broad interest group coalitions that were born during the emergency have not persisted, the hurricane's differential impact on these groups has created the potential for new cooperative alignments in local politics that might be exploited by the managers of hazards and urban communities (see chap. 12 in this volume). If disaster-reduction efforts are to succeed, urban leaders and managers need to have a better understanding of the potential for these kinds of intersecting issues to emerge and the probability that some combinations may be especially felicitous for the adoption of innovative disaster-reduction measures. The sustainability of cities in the face of disaster is as much a function of enhancing institutional and behavioural capacities to deal with uncertainty as it is an outcome of material investment in hazard-management technologies and physical infrastructure.

Spatial parameters of mega-city hazards

The spatial parameters of mega-cities have been singled out for attention for two main reasons. First, the internal and external spatial organization of cities - including their populations, physical environments, and activities - is one of the fundamental themes of urban research. Secondly, the "goodness of fit" between people and environments, which is a central problem of hazards research, has a strong spatial component. In other words, issues of space and place are particularly germane to the analysis of hazards and of urbanization. Some of these matters have already been touched on above (e.g. the shifting locus of mega-city hazard), but they will be revisited here in order to highlight a whole family of changes that are explicitly spatial in nature. The discussion that follows examines mega-city hazards at a number of spatial scales from global to neighbourhood.

Global and national linkages of mega-city hazards

On 12 May 1997, a well-defined tornado was captured on television as it passed the downtown district of Miami without causing significant damage. The image of a powerful vortex spinning behind glass and steel high-rises that are the product of Miami's recent emergence as an international centre of finance for Latin America vividly underlines one aspect of the connections between the global economic system and the hazardousness of mega-cities. In this case, the outcome may have been positive from a hazard perspective. The influx of new investment helped to rebuild a business district that is now more resistant to extreme winds. What of other effects elsewhere?

First, the global economic system has a definite but uneven influence on mega-city hazards. Investments and economic functions are being redistributed among and within countries, thereby stimulating changes in national and local levels of exposure and vulnerability to hazards. For example, loss of investment in Los Angeles leads to the closure of firms that contribute to - and are victimized by - environmental hazards. But it also means reductions in local public and private resources for managing hazards and speeding recovery from disasters. Conversely, London in recent years has attracted a disproportionate share of new investments in urban hazard management precisely because it contains so many globally important financial services that are vulnerable to disruption. Such divergent experiences illustrate the difficulties of assessing net effects of global economic changes on mega-city hazards. Mexico City's recent economic history reinforces the point. In the period between 1940 and 1980, during the era when import-substitution economic policies were in vogue in Mexico, large numbers of new businesses established themselves in the city's inner districts. But recently, with the advent of export-oriented policies, many of Mexico's larger, more productive, and more profitable firms have migrated elsewhere, especially to the strip of land adjacent to the US border. The city still confronts a legacy of serious natural hazards that were much aggravated by recent meteoric urban growth, but it does so without the involvement of some of the large-scale economic enterprises that it had previously monopolized. These examples suggest that global economic forces help to change the calculus of hazard in mega-cities but do not produce any one predominant outcome.

A second way in which external influences can affect the management of mega-city hazards is via the extension of national government authority over local public policy. Here again the evidence is difficult to interpret both because local politics often have a disproportionately large effect on national policy-making in primate cities that are also national capitals and also because the constitutional division of responsibilities for hazard management between different levels of government varies among countries. For example, in the past, national government policies largely shaped the development of places such as Seoul, Lima, and Dhaka, but elsewhere metropolitan government institutions such as the Greater London Council pursued independent orientations. With the re-emergence of laissez-faire philosophies of national government during the 1980s, the latter role became less evident in London, although the departure in May 1997 of a British political regime that was sympathetic to free market ideology may signal a possible reversal of this trend. Even in Japan, which remains very much a centrally directed bureaucratic state, the city (prefecture) of Tokyo has a high degree of autonomy, as evidenced by its proliferation of specially tailored metropolitan hazard-management laws and planning initiatives. In short, although the evidence is mixed, mega-city hazard-management policies and programmes are just as likely to be different from those pursued by national governments as they are to be similar.

A third way in which external influence comes to bear on mega-city hazards is through the operation of broad cultural notions about the societal role of environmental hazards. From a public policy standpoint, the main issue here is whether the myths are appropriate for the management of mega-city hazards. Many seem to be out of phase with urban realities. For example, the Australian national myth of a colonizing frontier people who have contested and subjugated a harsh land may not be well suited to the relatively benign precincts of Sydney. There the most pressing environmental problems are now more likely to arise through human mismanagement than because of natural risks. Hazards are largely missing from the dominant British environmental myth, which runs to bucolic images of well-cultivated fields, neat gardens, and an undemanding climate. Yet London remains susceptible to very serious flood and storm hazards. Finding and reinforcing appropriate hazard myths for mega-cities deserve a significant place on the agenda of urban management in the next decade.

Urban hinterlands supply yet another set of external influences on mega-city hazards. For Puente (chap. 9 in this volume), a mega-city is more or less vulnerable according to the degree of its dependence on its hinterland. For example, rural disasters are a potent generator of migrants to urban areas. Dhaka receives a seasonal influx of rural flood victims, many of whom remain to swell the city's permanent population and increase its long-term vulnerability. The migration of rural victims from rural Peru to Lima under pressure from Shining Path guerrillas, or from Haiti to Miami in response to civil unrest, are other examples. Elsewhere the stimulus might be drought or famine, perhaps combined with civil war (e.g. Ouagadougou or Khartoum). The preceding examples are of cities in LDCs, but the principle extends to all mega-cities. In the case of London or New York, hazard hinterlands are much larger. Both of these cities are bases for governmental and non-governmental organizations that have global responsibilities for disaster relief and both have been subject to terrorist incidents that reflect geographically distant conflicts. The size of the hazard hinterland increases as does the city's centrality to the global urban network.

Another set of spatial properties derives from the position of mega-cities within systems of cities. Mega-cities tend to be at or near the top of national urban hierarchies, where they generally have superior access to national resources. This enhances their resilience against disasters. For example, national policy for disaster relief was liberalized when Britain's largest city - London - was badly affected by a severe storm in 1987 (Mitchell et al., 1989). Miami, the eleventh-largest US metropolitan area, did not receive nearly such favourable treatment from national policy makers in the wake of a much more destructive storm - hurricane Andrew (1992). Mega-cities also lie at different distances from the core of the global economic system. This affects their capacities to mobilize international aid in the wake of disasters and to secure other forms of external support for hazard-mitigation programmes. The general spillover effects of responses to security threats that were taken by private financial institutions in London have already been noted. Similar changes occurred in the San Francisco Bay Area's Silicon Valley after the Loma Prieta earthquake. On the other hand, the extent to which international businesses have upgraded adjustments to hazard in cities such as Dhaka and Lima is not known but is suspected to have been much less.

Urban form as a factor of mega-city hazard

Now let us turn to the mega-city as an urban spatial form. In this respect there are great variations among the case-study mega-cities. Some are relatively compact, densely populated, and intensely developed places (e.g. Lima), whereas others sprawl in all directions at low density (e.g. Sydney). Some pivot around one or two central nodes (e.g. Mexico City), whereas others have multiple urban nuclei (e.g. Los Angeles). Some are set on more or less flat land with few major barriers to surface movement (e.g. Miami); others are punctured by rivers, coastal embayments, steep slopes, and similar impediments (e.g. San Francisco Bay Area). Some are served by well-integrated transportation grids with many alternative routes, modes, and interchange points (e.g. London); others squeeze traffic into a handful of paralysing bottlenecks (e.g. Seoul). The hazard consequences of these different layouts are considerable.

Each urban form offers a different mix of advantages and disadvantages to hazard managers and potential victims. For example, compared with compact cities such as Dhaka, low-density ones such as Los Angeles are serviced by complex infrastructure networks (e.g. electricity, water, gas) that may be disrupted more often and more easily by a wider range of locally prevalent hazards. But dispersed urban forms lend themselves to emergency-management systems that optimize vehicular mobility and the transfer of relief aid into affected neighbourhoods from external sources. Problems can arise when movement is impeded; winding tree-lined roads in the low-density hilly suburbs of Oakland proved to be deathtraps for escaping motorists and barriers to incoming emergency vehicles during a recent wildfire! Conversely, Dhaka's heavily built-up inner city, served by narrow, congested streets and often improvisational infrastructures, raises nearly insurmountable access problems for "rapid response" teams that must be dispatched from elsewhere and offers advantages to neighbourhood-based emergency systems.3 Shortages of open space in densely built-up cities put severe strains on emergency evacuation systems that use such places as accessible assembly points where temporary shelters can be erected and neighbourhood feeding stations established. Diseconomies of scale certainly come into play sooner for large hazard engineering projects in low-density cities than in compact mega-cities, but this does not mean that such projects are precluded. The cost of building large protective structures (e.g. dykes) around lightly occupied neighbourhoods is daunting, but this has not prevented even poor cities such as Dhaka from investing in such measures when all existing urban lands are already developed at high densities. The point here is not that the advantages and disadvantages of either type of city form are permanent; rather it is necessary to design hazard-management systems that are best adjusted to the changing needs of specific mega-cities. It is also necessary to be careful about applying policy guidelines for urban development that do not take much account of hazard criteria. For example, as Parker notes in chapter 7, London does not conform to the model of urban development preferred by the European Commission (i.e. high-density, mixed-use, compact). The reasoning may be sound from the perspective of amenity preservation and the efficient delivery of services, but the hazard implications of such urban forms are far from clear.

3 This does not mean that both practices are in fact avoided. The "poldering" of new suburban neighbourhoods behind large embankments in Dhaka is one obvious example of the former. Because of urban congestion, Calcutta's official ambulance services are mainly dispatched from their bases to carry non-critical patients to hospitals, whereas victims whose lives are at stake are generally brought to medical facilities by passing motorists.

Internal urban dynamics

Core - periphery relations in the city, the differentiation of neighbourhoods, and the sequential occupation of urban districts by different land uses and populations have frequently been of interest to urban analysts. Information about environmental hazards throws new light on all three topics. Perhaps the most interesting finding relates to patterns of upgrading and renewal of urban areas in response to concerns about natural hazard. Particularly distinctive is the migration of vulnerability in Tokyo. There the zone of structural vulnerability to hazards is moving outwards toward the suburbs and the zone of social vulnerability is differentiating inwards toward particular marginal populations that frequent places with special hazard characteristics. Though less clear cut, a similar process seems to have been occurring in Miami before hurricane Andrew devastated the south side of the metropolitan area and accelerated the transfer of more resilient middle-class populations to north-side locations. Variants may be in progress in the San Francisco Bay Area and in Los Angeles, where the exposure and vulnerability of buildings to a suite of natural hazards are also very high along the outer urban margins and where there has been less investment in upgrading the hazard resistance of neighbourhoods near the urban cores.

Mention of urban margins and marginal populations recalls the marginalization theory of hazard ecology (Wisner, chap. 11 in this volume). This suggests that marginal places (i.e. peripheral, hazardous, or otherwise undesirable) and marginal people (i.e. disadvantaged groups) often coincide, usually because social forces push those without resources to the literal and figurative "edges" of community life. It is this that leads Wisner to assert that, in so far as Los Angeles is concerned, "there are worse things than earthquakes." However, evidence from the full range of mega-cities seems less clear. Sometimes areas that lie at the geographic heart of a great city, and are much in demand by many potential users, are also highly hazardous (e.g. the filled lands of Tokyo and the former lake beds of downtown Mexico City). Sometimes, too, the outer periphery of a mega-city is exposed to great physical hazard but is also much sought after for homesites because it possesses offsetting amenities or other values (e.g. the San Francisco Bay Area's Oakland Hills; the new polder developments of Dhaka). Yet there is no denying that the dry lakebed neighbourhoods of Mexico City's north-east side or the pueblos jovenes in peripheral districts of Lima are occupied by poor people and often subject to serious environmental hazards. Given the implicit logic of marginalization as a theory of hazard, these cases deserve close analysis. The various kinds of margins need to be more clearly distinguished4 and the relationships between them and conditions of social advantage or disadvantage spelled out.

4 Among others, margins might include: (1) the physical edges of the city, namely, the outer edge, which is in contact with rural districts; inner edges adjacent to physical boundaries such as coasts, rivers, and steep slopes; (2) time - distance edges that connote remoteness from opportunities for work, necessary facilities, and the like; (3) perceptual edges such as boundaries between social areas occupied by different - perhaps hostile - groups and places located beside zones of exclusion such as high-crime districts or red-light districts; and (4) zones of physical difficulty such as areas exposed to natural risks and hazards.

The space - time continuum

"Compression of the space - time continuum" has been identified as a distinguishing feature of worldwide societal change in the late twentieth century. This term connotes both: (1) a speeding up of the circulation of information, capital, and people; and (2) a reduction of distinctions between different places. These processes are believed to be occurring more or less everywhere but are said to be most evident in urban areas. The record of hazards and hazard responses in this sample of mega-cities provides considerable support for the first of these notions but much less for the second.

Urban patterns of exposure and vulnerability are sensitive barometers of speeded-up circulation and they are the dimensions of mega-city hazard that are changing fastest and most dramatically. New spatial patterns and widening gaps between population subgroups are detectable nearly everywhere. Hazard effects attributable to the accelerated circulation of capital and people are more apparent than those attributable to speeded-up information. Sometimes the results are positive - that is, they lead to improved protection against environmental hazards for workers and residents (e.g. in the financial districts of London and Miami); sometimes they are negative, as in the experience of cities that lost taxable investments that paid for safety services and provided residents with the financial resources to buffer hazard (e.g. parts of Los Angeles). But there are signs that exchanges of information between MDC and LDC mega-cities may be working to reduce hazards in some places (e.g. Latin American cities). At present such exchanges mainly involve the physical movement of experts and students between information-sending and information-receiving countries by means of international conferences, courses for foreign students, and consultant visits abroad. The first Internet Conference on Urban Hazards was held in 1996 under the sponsorship of the International Decade for Natural Disaster Reduction (Geneva) and signalled the beginning of faster exchange of ideas and knowledge.

Conversely, the hazard landscapes of mega-cities do not appear to be becoming more alike. Residents confront a kaleidoscope of choices and constraints produced by layered combinations of: physical risks; human-modified environmental processes; the forms and dynamics of the built urban environment and its social ecology; and existing and emerging patterns of human exposure, vulnerability, and response to hazards, among others. For managers of urban hazard, the central problem remains one of adjustment between environments and peoples that started out both diverse and changeable and have only become more so in recent decades.

Summary of case-study findings

The picture of mega-city hazards that emerges from the case-studies is complex and susceptible to different interpretations. Its main features include: (1) the diversity of risks that confront urban populations and growing interactivity among those risks; (2) the extent to which previous urban disasters (especially natural ones) have had deep and long-lasting repercussions on built environments and societal institutions as well as more obvious immediate human effects; (3) the build-up of catastrophe potential in mega-cities; (4) the narrowness of existing urban hazard-management policies and programmes; (5) important gaps in scientific information; and (6) reorganization of the urban ecology of environmental hazard, most notably reflected in shifting patterns of exposure and vulnerability. In short, according to the case-study evidence, the environmental hazards of large urban areas are already highly important and they are changing in ways that will increase their significance during the twenty-first century. Urban managers would do well to pay attention to these trends and to include hazard management among their priorities.

In the meantime, there are a few simple steps that can be taken to pave the way toward improved management of environmental hazards in mega-cities. There is a pressing need for "horizontal" exchanges of information about hazards and hazard-management policies among individual mega-cities. Such exchanges offer valuable opportunities for learning about the effectiveness of different policy tools in different hazard contexts. For example, London and Seoul might compare the use of green belts as hazard-management devices; Tokyo and Miami might exchange data about the taxation of peri-urban agricultural land; Mexico City and Seoul might compare the impacts of programmes for decentralizing the economic and political functions of capital cities; Sydney and Los Angeles might focus on the nexus of transportation - air pollution - emergency-management problems that affects both cities; Dhaka and Lima could share experiences about links between squatters, housing policies, and hazardous sites.

Historical data on previous hazard adjustments and data about present hazard-management strategies adopted by individuals and small firms are scarce in most mega-cities. We need carefully to sort out adjustments that failed because they were fundamentally overwhelmed or mismatched with the new conditions created by recent urbanization from those that atrophied from lack of awareness by public officials and new urban residents, or from lack of investment, or from other causes that might still be rectified or redressed. In an era when private sector initiatives are being encouraged by public policies (e.g. the World Bank's new urban-management programme), promising innovations at the grass roots also need to be identified and nurtured. Researchers can begin both activities now.

The potential of theoretical constructs such as "urban metabolism" and "urban carrying capacity" is largely unexplored from a hazards-management perspective. To date, these notions have been chiefly applied to the exchange of resources and wastes between urban areas and surrounding territories and to the environmental impact of urban resource development projects. But they might profitably be extended to include the impact of hazards on the carrying capacity of mega-city infrastructures. Hazards that impair the conduits on which metabolism depends or that reduce the carrying capacity of urban infrastructures, even on a temporary basis, might be taken into account by urban sustainable-development programmes (Berke, 1995; MacLaren, 1996).

New departures

Beyond the immediate findings lie other observations that open up new questions and suggest new directions for future research. Broadly speaking these fall into two categories, with pairs of contrasting themes: (1) the effects of hazards on mega-cities and of mega-cities on hazards; and (2) the ways in which hazard mitigation might be added to the agenda of mega-city managers while hazards managers might take on board some of the broader concerns of mega-city management.

Quite apart from their obvious impacts - destructive and otherwise - natural hazards, natural disasters, and other natural extremes affect mega-cities and the people who live there in three distinctive ways that have received inadequate scrutiny by researchers. First, they are important agents of diversification in urban settings. Secondly, they provide vivid examples of contingencies that affect human life. Thirdly, they are fertile sources of metaphors and myths about urban existence.

It is often alleged - and not without cause - that large urban areas are increasingly structured by global economic and political forces that encourage uniformity of landscapes and lived experiences. Whether the process is unilinear or bifurcated - towards one type of urban entity or towards "more developed" and "less developed" variants - similarity appears to be growing. Natural hazards are a potent antidote to this process, for they segment urban space into areas of differing risk that are continually being modified and shifted as people search out and apply a growing roster of adjustments that change net risks and alter levels of exposure and vulnerability.

Likewise, an entirely praiseworthy concern for alleviating the everyday problems of big cities blinds many urban analysts to the vast importance of unexpected, infrequent, and otherwise exceptional events that compel leaders and residents to change focus and redirect their energies toward different ends. Natural hazards and disasters are prime examples of a class of change factors that frustrate long-term schemes of planning and management that are based on assumptions of continuity. Of course, local government has often been considered an institutionalized form of crisis management, mostly characterized by muddling through. But the numbers and significance of events that continually bubble to the top of the urban crucible are such that more than crisis-response is now necessary. The public and private sectors of metropolitan areas should learn to take disjunctive events into account systematically and deliberately, not just as inconvenient disruptions of "normalcy." In other words, mega-cities need to equip themselves with more than preparedness systems, emergency operations centres, and disaster-response capabilities. The more contingencies occur in cities, the more cities will have to develop flexible responses across the full spectrum of governmental responsibilities. Broadly construed, hazard mitigation - in all its forms and for a broad range of events - should become a continuing basic part of urban governance.

Finally, the ultimate importance of environmental hazards for big-city populations may be their symbolic value as fertile sources of metaphors and myths about appropriate human behaviour in an uncertain universe. Too often, myths about great cities and disasters have tended toward the heroic and the apocalyptic. They have crystallized into unchanging behavioural guides and public attitudes that may have been historically useful but are now ill suited to the fast-breaking multiple crises of the postmodern world. The phoenix myths of Chicago or Seoul or San Francisco have outlived their former purposes and now stand as half-truths at a time when more subtle lessons about the ongoing nature of human adjustments to environmental constraints might offer appropriate pointers toward the long-term sustainability of urban areas.

The effects of large-scale urbanization on natural hazards are manifold, but three new avenues are particularly worthy of further investigation:

1. the changing (spatial) ecology of urban hazard;

2. the role of indirect information about hazard that is presented in the urban landscape and the urban experience; and

3. opportunities for raising the low visibility of hazard as a topic of urban governance.

The bulk of previous research on human adjustments to natural hazard has tended to adopt an all or nothing approach toward the role of group processes. Hazard adjustments are regarded either as preponderantly the product of individual decision-making or as primarily the creations of powerful social forces that are brought to bear by the dominant on the dominated. Neither of these views accords with the reality of life in a complex urban environment where both factors operate and the entire setting is quite evidently undergoing far-reaching change. So it is necessary to employ investigative tools that are more sensitive to the joint effects of individual and group processes. Given the ample evidence that the socio-spatial organization of mega-cities is in flux, it is worthwhile to invest more effort in mapping the various elements of hazards. Patterns of risk, exposure, vulnerability, and response are known to be changing, but the changes are poorly documented. If long-term strategies for hazard mitigation are to be effective, it will be necessary for them to be periodically readjusted to take account of changing land use and occupancy, new technologies, emerging social issues that bear upon hazard management, and other factors. Better understanding of urban hazard ecology is a prerequisite for this process.

Mega-cities can be thought of as complex information-transfer nodes that display and transmit a vast array of signs, signals, and messages. In such settings, direct messages about hazard (e.g. hazard maps, warnings, informational pamphlets, public service announcements by aid agencies) are apt to be diluted or lost before they reach the intended recipients. At the same time, the urban environment is rich in indirect information about hazard that may be more effective in influencing decisions and behaviour than the formally targeted messages. Sources of indirect information include, among others: events observed in the street or evidence of previous occurrences; knowledge acquired in schools, libraries, museums, or other educational establishments; news reports picked up from the mass media; warning notices and symbols; advertising signs and broadcasts; consumer product information; entertainment programmes and performances; gossip among family members, friends, and neighbours; rumours passed along by others; acquired folk beliefs, etc. Although there has been a large amount of research on the perception of specific risks and on judgements about certain sets of risks, little attention has been focused on links between the broad array of indirect influences on risk judgement in cities and assessments of particular natural hazards or on the interplay of indirect information and direct information pertaining to the full range of mega-city risks. The contextual embeddedness of mega-city hazards warrants greater attention by the research community.

A third aspect of urbanization that deserves further exploration as an influence on hazard management is the emerging structure of interest group competition in city governance. As recognized by the United Nations Commission on Human Settlements (1996) and other organizations, public policy in contemporary cities arises through struggle among interest groups that focus on one or more of the following subjects: economic growth, environmental protection, and social justice (Campbell, 1996). To a large extent, economic interest groups exist not just because of a desire for profit but also because of previous experience with economic stagnation and fears that such conditions could occur again. Likewise, social justice interest groups exist not just because of a commitment to the ideal of social harmony but also because of experience with social injustice and fears of worse to come. However, urban environmental protection interests are much less united about the goals of environmental protection and they are rarely concerned about hazards that arise in Nature; instead their focus is on stemming the deleterious effects of human mismanagement of the physical world. In other words, natural hazards are largely unrepresented in the urban governance models of contemporary analysts. The tendency to marginalize the physical environment as a variable in urban affairs is already marked in many cities, and the low salience of natural hazards in city politics is disturbing and requires further analysis.

This brings us to the second pair of issues, namely, the insertion of hazard considerations into urban management and the broadening of urban hazards management to take account of other goals. If the central problem of contemporary urban management is to resolve conflicts among economic, environmental, and justice interests, the central problem of contemporary hazards management is to shift the emphasis of public policy from reaction to anticipation - from dealing with emergencies after they occur to addressing the problems that "cause" them. What happens when these agendas are combined?

A very likely possibility is that tensions will be generated among existing urban policy interest groups. For example, judged by previous experience, economic interests may favour policies that improve preparedness and emergency-management alternatives that reduce foreseeable disruptions to business operations and markets but may be less willing to invest in mitigation measures whose pay-off lies far in the future. Environmental interest groups may prefer the inverse priorities. So little is known about the hazard perspectives of interest groups concerned with urban social justice that a wide range of orientations is possible; they might favour either, neither, or both alternatives or they might promote quite different ones. It is premature to fix on any single distribution of preferences, but the likelihood of incompatibilities among the principal interest groups is significant. Again there is a need for additional research, especially with respect to the stance of justice interest groups toward issues of urban hazard.

Environmental hazards have been, are, and will continue to be important problems for the world's mega-cities. Indeed, their visibility is likely to increase over the next several decades. It will be important to monitor how well urban societies respond to the challenges that they pose, because hazards and disasters possess special significance as public issues. To a significant degree the success of urban sustainable-development initiatives will be determined by the ability of mega-city leaders and their diverse constituencies to join in metropolitan-wide systems of governance and management that are responsive to disjunctive changes as well as to continuing problems. Nor will it be possible to exempt mega-cities from the broader global agenda of sustainable development, which requires humans to balance the demands that we make upon the physical environment with the limits that the environment sets on the possible. As agents of disjunctive change and dramatic constraints on human activities, natural hazards and disasters set demanding tests for the emerging institutions of urban living.

REFERENCES

Berke, Philip R. 1995. "Natural-hazard reduction and sustainable development: A global assessment." Journal of Planning Literature 9(4): 370 - 382.

Campbell, Scott. 1996. "Green cities, growing cities, just cities? Urban planning and the contradictions of sustainable development." Journal of the American Planning Association 62(3): 296 - 312.

Davis, Mike. 1995. "The case for letting Malibu burn." Environmental History Review Summer: 1 - 36.

GeoHazards. 1994. The Quito, Ecuador earthquake risk management project: An overview. San Francisco: GeoHazards International.

Gruber, Urs and Harold Haefner. 1995. "Avalanche hazard mapping with satellite data and a digital elevation model." Applied Geography 15(2): 99 - 114.

Kirp, David L. 1997. "There goes the neighborhood: After the Berkeley fire, an architectural disaster." Harper's Magazine 294 (March): 45 - 53.

Knox, Paul. 1993. "Cities of Europe." In Stanley D. Brunn and Jack F. Williams, Cities of the world: World urban development. New York: HarperCollins, pp. 88 - 150.

MacLaren, Virginia W. 1996. "Urban sustainability reporting." Journal of the American Planning Association 62(2): 184 - 202.

Mitchell, James K. 1995. "Coping with natural hazards and disasters in U.S. mega-cities: Perspectives on the twenty-first century." GeoJournal 37(3): 303 - 312.

Mitchell, James K., Neal Devine, and Kathleen Jagger. 1989. "A contextual model of natural hazards." Geographical Review 79(4): 391 - 409

Montz, Burrell and Eve Gruntfest. 1986. "Changes in American urban floodplain occupancy since 1958: The experiences of nine cities." Applied Geography 6(4): 325 - 338.

Oppenheimer, Clive. 1994. "Discussion meeting on natural hazard assessment and mitigation: The unique role of remote sensing, The Royal Society, London, 8 - 9 March, 1994." Disasters 18(3): 294 - 297.

Platt, Rutherford H. 1991. "Lifelines: An emergency management priority for the United States in the 1990s." Disasters 15(2): 172 - 176.

United Nations Commission on Human Settlements. 1996. An urbanizing world: Global report on human settlements 1996. New York: Oxford University Press.

UNEP/WHO (United Nations Environment Programme/World Health Organization). 1994. "Air pollution in the world's mega-cities." Environment 36(2): 4 - 13, 25 - 37.