Extreme weather events and the weather-related events they may induce, such as landslides, floods and storm surges, form an important part of what have been termed ‘natural hazards’. They have a major influence not only on the physical landscapes and human societies directly affected by them but also on the wider community through their impact on the insurance industry and the costs of emergency aid or relief. It is increasingly being recognised that their distribution in time and space is dynamic rather than static and significant changes in the frequency of extremes such as heavy daily rainfalls, droughts, extreme heat and cold and tropical cyclones are envisaged in ‘global warming’ predictions for the next century (IPCC 1996; United Kingdom Climate Change Impacts Review Group 1996; Hulme and Viner 1998). This chapter reviews the roles that geographers have played in examining the climatology of extreme events, their spatial and temporal distribution (including past, current and future changes), their impacts on natural systems and human activities, and the design and effectiveness of strategies aimed at reducing their adverse effects. With particular reference to geographical research on tropical cyclones, it highlights the problems that stem from their inherent rarity, which reduces the sample size upon which to base conclusions and advice, the mismatch between the spatial scale at which one can offer reliable advice and the spatial scale most useful for planning purposes, and the implications and uncertainties associated with recent and predicted future changes in extreme event frequency.
Extreme weather events can be divided into absolute and relative event types (Figure 4.1). Absolute extreme events are considered extreme simply because of their magnitude and nature. Thus tropical cyclones, tornadoes, hailstorms and climate-related events such as avalanches, landslides and floods are considered extreme events, even in areas where they are frequent. Relative extreme events, on the other hand, are considered extreme on the basis of their rarity at the location concerned; they are distinguished from normal events on the basis of their probability of occurrence (Smith 1997) and its inverse, the return period or recurrence interval, which is the average time interval between events of a specified magnitude. If one views events as being statistically distributed around a long-term mean value, then most events lie relatively close to the mean within what Smith terms the socio-economic band of tolerance and can be interpreted as resources, as they constitute the near-normal events on which human activities in that climate are based. Weather extremes such as unusual cold, warmth, drought or wetness are measured relative to (and hence are specific to) the local climate; what constitutes an unusual drought, for example, will therefore vary with location. Extreme events,