From the beginning of human attempts to shape the water environment to human benefit, water management has dealt with the variability of the native supply of water and the variability of demands for the use of water (Stakhiv, 1998). Great strides have been made in dealing with even extreme water regimesparticulary droughtsthrough interventions on the supply and demand sides (e.g., Stern and Easterling, 1999). Drought management planning is playing an increasing role in many water management agencies, lowering their susceptibility to drought impacts. Thus, in some ways the prospects of a change in the resource baseperhaps characterized by lower mean supplies and higher variabilityrepresent only a sharpening of traditional challenges to water management. There are three important differences, however. First, future climate change is highly uncertain at spatial and temporal scales that are relevant to water management: All we know is that the future may not necessarily be like the recent past. Second, as noted above, the potential pervasiveness of these changes across large regions presents challenges that preclude some traditional steps of adaptation and requires innovative approaches that go beyond experience to date. Third, climate-induced effects may be nonlinear, carrying potential for surprises beyond those incorporated in traditional water management.
The ability to adapt to climate variability and climate change is affected by a range of institutional, technological, and cultural features at the international, national, regional, and local levels, in addition to specific dimensions of the change being experienced. Among the most important features are the following:
Whether adaptation takes place or not may be heavily influenced by the occurrence
of extreme events. Such events often are catalysts for change in management
and may serve two roles. First, they may expose failings in the current water
management system. Second, they may raise the perception among decision makers
of the possibility of climate changeeven if they cannot be attributed
directly to climate change.
Recent experience with extreme events (e.g., the Chinese floods of 1998, the Rhine floods of 1996 and 1997, the eastern European floods of 1997 and 1998, and the Mozambique floods of 2000) shows that many societies are extremely exposed to loss and damage during extreme events, especially floods. At first, it may appear that this implies that existing adaptive techniques, as widely used by water managers, are not working as expected to minimize risk and loss (some loss will always be inevitable because no flood protection scheme can provide complete protection): Adaptation is not working. However, there is extensive evidence that social vulnerability to extreme events is serious and increasing (Munasinghe and Clark, 1995; Hewitt, 1997; Tobin and Montz, 1997; Haughton, 1998; La Red, 1999; Mileti, 1999) and that this exposure to hazards has been significantly increased by public and private development with insufficient regard for known hazards (Hewitt, 1997; Marsden, 1997; Pulwarty and Riebsame, 1997). In the United States there was more damage from hurricanes between 1990 and 1995 than there was between 1970 and 1990, after adjustments for inflation (Pielke, 1997), even though both periods had low hurricane frequency (Landsea et al., 1996). Changnon et al. (1997) analyzed the dramatic increase in dollar losses of insured property in the United States, which reached US$840 billion in the 19901994 period, and conclude that changes in weather and climate were not primary causes. Detailed meteorological analyses came to the same conclusion for flooding losses (Changnon, 1998; Karl and Knight, 1998).
Thus, societies failure to adapt to extreme events in the broadest sense (i.e., by allowing risk-prone development) appears to have been largely responsible for increased damages, and that failure has not improved with time (Changnon and Changnon, 1998; Pielke and Landsea, 1998; Kunkel et al. 1999). It also appears that political decisions may have produced maladaptive results (Wiener, 1996; Hewitt, 1997; Mileti, 1999). In the United States, insurance has been a leading instrument for hazard awareness and post-event recovery. After 30 years of promotion, education, and subsidized premiums, only 20% of residents in floodplains were insured by the late 1990s (LeCompte and Gahagan, 1998; Pasterick, 1998). These failures to take advantage of insurance suggest that even wealthy societies adapt poorly to foreseeable hazards.
The residual damages of hazard events also are inequitably distributed across populations. This was shown clearly by studies of Hurricane Andrew in Florida (Peacock et al., 1997), leading the director of the Pan American Health Organization to state that those who lost the most had the least to lose (PAHO, 1999). Hurricane Mitch devastated Central America in 1998, exhibiting the extreme vulnerability of that region (La Red, 1999; UNICEF, 1999). Among the responsible factors were lack of land-use planning, deforestation, and inappropriate consumption and production systems (Hewitt, 1997; Mileti, 1999; PAHO, 1999).
Thus, available evidence concerning the effectiveness of adaptation to meteorological and geologic hazards indicates poor levels of individual and social adaptation to hazards. This failing extends well beyond the water management sector as conventionally defined and can be argued to reflect weaknesses in development control, planning guidance, public education, and fiscal incentives. The foregoing examples indicate that having the ability to adapt to change is not the same as actually adapting to change: The tools often are not used, for a variety of reasons.
Other reports in this collection