Coping with Seasonal-to-Interannual Climatic Variation
The effects of climatic variations on any social system result from the combination of experienced weather-related conditions and the ways that the social system anticipates and responds to these conditions. Throughout human history, societies have expected seasonal changes similar to the local historical averages and a certain amount of variation around these averages, but, despite their efforts to forecast these variations, they have not typically counted on much skill in predicting them. Thus, they have organized themselves to expect climatic surprises and to deal with their impacts after the fact. The newly developing scientific skill in climate forecasting may fundamentally change the ways social systems cope with climatic variation by reducing the magnitude or frequency of surprise and by providing more time to prepare for climatic events. The results are likely to be beneficial overall; however, there may be different effects on different social systems and on different individuals and organizations operating within those systems.
To understand the effects of climate forecasts on human well-being and their potential to benefit people, it is therefore important to begin by examining how social systems currently cope with climate variability. Such coping involves both activities undertaken in anticipation of climatic uncertainty, sometimes called ex ante or risk management strategies, and responses to experienced climatic events on the part of individuals and organizations, sometimes called ex post or crisis response strategies. The net result of these coping strategies may or may not be an improvement in outcomes for the society or for specific segments of it. A
variety of insurance mechanisms create net social benefits by spreading risk over a risk-averse population, and many public investments in infrastructure, public health, and hazard management programs effectively reduce climate-related damages. However, some individual or community-level risk-management or crisis response activities can have adverse impacts on other parties, so that the actions do not necessarily improve overall societal well-being.
There have been many studies of the ways particular social systems cope with particular kinds of climatic variations, but there is as yet no general theory of such coping. This chapter begins to develop a framework for analyzing coping systems by distinguishing between ex ante and ex post strategies, identifying some subtypes within these, and distinguishing among the actions of individuals and of public and private organizations, the behavior of markets or informal exchange relationships, and the roles of legal and other institutions. The chapter examines available knowledge about coping systems for climate variability in order to characterize the state of knowledge; identify ways in which coping strategies may shape the impacts climatic variations have on the people and groups that use them; and define gaps in knowledge that, if filled, could help increase the usefulness of climate forecasting for humanity.
We first examine human coping mechanisms in several weather-dependent. sectors of human activity, including agriculture and water management. We then briefly discuss some systems of human activity that have a primary function of coping with climate variability, such as insurance and emergency preparedness. The chapter shows the wide variety of coping strategies and identifies some of the factors that determine the coping strategies available to particular actors and that shape the outcomes they experience from climate variations. These factors include the availability of insurance and insurance-like systems for making up for losses, integration into global markets, the cognitive and economic resources available to actors engaged in an affected activity, and the ways in which these resources are distributed.
Coping in Weather-Sensitive Sectors
Human activities are sometimes affected directly by climatic events, such as when great floods destroy lives and property. Many of the important effects of climatic events are indirect, however, operating through biophysical processes on which human welfare depends. Examples include the effects of climate on crop production, fisheries, forests, water resources, and the ecology of pests and diseases. This section illustrates the variety of systems that humanity has developed to cope with the
effects of climate variability on weather-dependent sectors of human life and indicates the general state of knowledge about them.
Agriculture, including both plant cultivation and livestock production, is a sector that is heavily dependent on the amount and timing of rainfall, which in many areas of the world are highly variable. For example, the dry rangelands of Africa, which receive less than 600mm of rain per year, experience some of the greatest climatic variability on the continent. El Niño/Southern Oscillation (ENSO) events have caused droughts in southern Africa with a frequency of three to six years since the 1950s (Trenberth and Shea, 1987; Scoones, 1992). In the semiarid tropical zone of India, cultivation must wait for the onset of monsoon rains because of the hardness of the soil, and the timing of the monsoon onset is highly variable. Some agricultural systems are also highly sensitive to climate parameters other than rainfall, such as the occurrence of killing frosts, the length of the growing season, and the number of growing degree-days.
In all areas of the world and at all levels of economic development, human cultures inhabiting variable environments have developed strategies and behaviors designed specifically to ameliorate the effects of climatic variability on their subsistence (Galvin, 1992; Halstead and O'Shea, 1989). Indeed, in a variety of cultures and environments that exist under the stress of high climatic variability, primary cultural characteristics such as social relations, land tenure systems, institutions, laws, and land use practices are organized as coping mechanisms for dealing with climatic variability (Minc and Smith, 1989; Legge, 1989; Blaikie and Brookfield, 1987; Halstead and O'Shea, 1989; Fratkin et al., 1994).
The methods by which individuals directly engaged in agricultural production cope with climatic variability can be classified according to whether these strategies and behaviors affect production (the sensitivity of agricultural output and incomes to climatic events) or consumption (the ability of agriculturists to acquire food and other goods and services in spite of climate-related fluctuations in their agricultural production). Coping mechanisms can also be classified by the timing of the actions relative to the occurrence of the climate event. Actions taken prior to the realization of a particular climate event, such as the onset of the monsoon or unusually heavy rainfall (ex ante or risk management actions), are based on expectations of the likelihood of bad or good events, which are in turn based on primarily historical experience. Activities that take place after the event has occurred (ex post) attempt to ameliorate or exploit what has already occurred.
Table 3-1 provides a diagram of this four-way classification of the various coping strategies employed by individuals and production units engaged in agricultural production and lists a number of coping strategies of each type that are employed across different societies of the world. The specific array of strategies observed in particular parts of the world will differ, but in all societies some strategies appearing in each quadrant of the table are used.
An important feature of coping systems is that the strategies in the four quadrants are interdependent. For example, if farmers could alter their crop mix or inputs without cost to take advantage of climatic events after they occur, they would have less need to engage in production practices that reduce the sensitivity of their incomes to climatic variability. Similarly, if farmers' incomes were perfectly insured against reductions due to adverse climate outcomes, they would need to engage less in other ex ante coping strategies that reduce the risk of income loss, and they would have less need to accumulate assets as a buffer against income loss.
Another important feature of agricultural coping evident in the table is that many of the ex ante coping strategies that reduce sensitivity to climatic variations are undertaken mainly to reduce the risk of extreme negative events. For example, buying insurance involves continually paying a small cost to reduce this risk; crop diversification and other hedging
strategies involve forgoing potential advantage from positive climatic events to reduce the risk of disaster.
For most areas of the world, ex post strategies have limited value or are very costly. For example, U.S. citrus growers occasionally use grove heaters or, more frequently, spray trees with water to avoid the consequences of frost (Miller, 1988). African herders who experience adverse climate outcomes respond by migration, even making extraordinary movements under severe drought stress, including leaving the pastoral system until the perturbation passes (Coughenour et al., 1985; Ellis et al., 1987; Galvin, 1992).
Many of the ex ante production techniques listed in the table are common across many societies around the world. An example is hedging strategies to spread the risk of extremely negative climatic events. African pastoralists spatially separate their herds, and Indian farmers use diversified seed types and farm on multiple plots. Similarly, in the Great Plains of the United States, many farmers incorporate drought-resistant but low-profit grain sorghum with their drought-susceptible but highprofit corn-soybean rotations in anticipation of the adverse consequences of drought for their incomes. And both U.S. and African farm households are characterized by diversified occupational portfolios, with family members engaged in both agricultural and nonagricultural activities. The worldwide pervasiveness of such ex ante hedging strategies for both production and consumption suggests that the cost-effectiveness of ex post strategies is limited in most societies and that insurancean alternative ex ante strategyis either incomplete or more costly than the other ex ante strategies.
The size and distribution of the impacts of climatic variability depend strongly on the array of coping strategies available to and employed by agricultural producers. These in turn vary according to agroclimatic conditions and the structure of markets and other institutions. Groups facing the same climatic variability are more or less vulnerable to extreme negative climatic events depending on their ability to make use of particular coping strategies and methods. For example, low-income farmers in developing countries, who comprise a large proportion of the world population, are less able than their wealthier neighbors to accumulate assets while meeting minimum subsistence requirements; such poor farmers are thus less able to maintain their consumption by drawing from their savings levels when they experience particularly low levels of rainfall (Rosenzweig and Wolpin, 1993). Since many of these countries lack developed insurance markets, an inability to accumulate assets in anticipation of bad years makes poor farmers especially vulnerable. Because of their great vulnerability, the poor in less developed countries may benefit
most from improved climate forecasts, provided that they can gain access to resources needed to respond appropriately to forecast information.
By contrast, producers who have accumulated wealth or are well-insured may benefit little from skillful climate forecasts in an extremely bad crop year because their climatic risks are already covered. In the United States, for example, federally subsidized crop insurance to cover climatic risk has been available to U.S. farmers in its present from roughly since 1948 (Easterling, 1996). Moreover, U.S. agricultural history is marked by instances in which the federal government has provided insurance-like income support to farmers suffering income losses from extreme climatic events. In this case, the cost of unfavorable climatic conditions may be shared widely among taxpayers and the benefits of improved forecasts in bad years may flow mainly to the national treasury as avoided costs. In good years, however, farmers may be able to use skillful forecasts to increase their output. (Subsidized insurance and income-support programs may alter farmers' coping strategies by encouraging them to gamble with high-vulnerability crops, because they gain the benefits while the treasury takes the risks.)
The agricultural sector in low-income countries does not often benefit from government assistance in the form of insurance or insurance-like coping strategies, although governmentally organized drought relief is not uncommonfor example, during the 1990-1991 drought in Zimbabwe (Magadza, 1994). Poor countries usually cannot afford to invest much in the institutions for societal buffering against climatic variation. Little institutional buffering occurs in the form of ex ante preparations, such as appropriate subsidies, insurance, and infrastructure for delivering relief, or even ex post relief such as loans or food shipments. Notwithstanding or perhaps because of government neglect, people in developing countries have perfected very sophisticated nongovernmental insurance-like coping strategies that accompany traditional ex ante production diversification. These strategies must be taken into account in assessing the value of improved climate forecasts.
Formal and informal nongovernmental social institutions such as obligatory sharing within groups and community self-help organizations are important local buffers. In the African livestock sector, herding families in the areas with more favorable local climate conditions adhere to social obligations to provide assistance to those in less favored areas. Many pastoral societies generate a strong sense of social interdependence, establish obligations to help and support less fortunate friends and relatives during times of need, and develop strong norms of reciprocity. These insurance-like institutions have been extremely effective over time (Coughenour et al., 1985; Ellis et al., 1987; Galvin, 1992). In the semiarid tropics of India, where sedentary agriculture is practiced and weather
shocks can affect many villages over a wide region, cultural traditions provide a similar type of informal insurance that results in the transfer of resources from households in villages with more favorable climate outcomes to those in villages, sometimes a great distance away, with less favorable outcomes. The tradition of exogamous marriage is helpful here, as financial aid can come to households who have had adverse weather outcomes from the households in which married daughters reside, which may be located in distant villages (Rosenzweig and Stark, 1989).
The effects of climatic variation also depend on agricultural producers' access to and use of hedging strategies. For example, although farmers worldwide diversify their crops, some countries have more sophisticated systems than others for fine-tuning that diversification (e.g., agricultural universities and hybrid seed industries that produce and advise on the use of diverse seeds). Farmers in some countries have ready access to commodities futures markets that allow them to lock in prices for some of their crop in advance of climatic variations. However, not all farmers with access to this strategy use itsome prefer to hedge by varying production practices or developing sources of nonfarm income (Weber, 1997). Irrigation, a hedging strategy in some regions, is available mainly to producers in areas in which public or collective investments have been made in the necessary infrastructure and effective institutions exist to maintain and manage the system.
The interdependence among the different methods for coping with climatic variability and the scope for engaging in them must be taken into account in evaluating the effects of climatic variation and the potential gains from improved climate forecasting. In addition, the combinations of individual and cultural coping strategies, developed over centuries and often serving populations well, can be fragile with respect to changes in environment and society. For example, the exploitation of resources over wide geographical areas that is a central coping strategy of pastoral societies in Africa has been constrained by population growth, which has encroached on the land used by pastoralists. This has increased their vulnerability to climatic fluctuations.
Improved climate forecasts may have complex effects on agricultural societies, extending beyond agricultural production. For example, an increased scope for taking ex ante production actions (e.g., diversification of income sources) may reduce the need for other ex ante measures on the production and consumption sides (e.g., crop insurance, norms of reciprocity). To the extent that the provision of informal insurance and consumption maintenance is a strong component of the organization of social relations in many societies, there may be important ramifications for social relations in these societies from introducing better forecasting skill. Some of the social consequences of improvements in forecast skill can be
anticipated, but others are not evident. For example, on one hand, increased forecast skill may increase demand for seeds that are more sensitive to rainfall or temperature, thus raising average incomes and increasing savings and consumption. On the other hand, the reduction in the costs of adverse climate events because of improved forecasts reduces the need for savings. It will be a challenge to estimate the aggregate and distributed effects of improved climate forecasts and their effects on traditional coping strategies, and then to design forecast information so that people benefit from the forecasts.
A critically important difference between fisheries and agriculture or herding is the fact that the fish stocks themselves are usually not privately owned. Rather, commercial and sport fisheries are almost always publicly managed common-pool resources. In a few cases, commercial harvesters have devised private methods of policing their own harvest rates (e.g., Acheson, 1988) and access to some sport fisheries is effectively limited by private property owners. In the more general case, public regulation of the fishery arises to control the tendency for competing harvesters to overfish. Overfishing in an economic sense involves devoting too much effort to fishing, so that the value of the harvest, net of harvesting cost, is not maximized (Gordon, 1954; Cheung, 1970). Economic overfishing often results in biological overfishing as well, sometimes leading to catastrophic collapses of commercial fish stocks. This inherent tension between the private incentives of the harvesters and efficient management of the fishery means that harvesters' coping strategies and their desired responses to climatic opportunities may not result in a socially beneficial outcome.
The traditional goals of fishery management have been to constrain both biological and economic overfishing. Most fishery management schemes have emphasized biological conservation, although economic goals have received considerable attention in recent decades. Achieving these goals often has proved to be quite difficult. In contrast to simple theoretical models, real fish populations fluctuate, sometimes radically, for reasons unrelated to harvesting. Climatic variations often play a role in these natural fluctuations, although the role is more immediate and apparent for some fish populations (e.g., Peruvian anchovies) than for others. In addition, the effects of climatic variations on fisheries are usually difficult to observe. Except for anadromous fish stocks, marine fish populations remain hidden from view, so that the size of breeding stocks must be inferred largely from harvest information. When fishery managers have only a very uncertain picture of abundance, their estimates of
optimal harvest rates are subject to considerable error. In such circumstances, a conservative approach to setting allowable harvests would reduce the risk of biological overharvesting, and thus jeopardy to future harvests. Conservative fishery managers, however, frequently encounter intense pressure from elements of the harvesting community who may expect to gain more from an immediate increase in allowable harvest than from an uncertain investment in the size of the breeding stock.
The economic objective of fishery managementto increase economic rent by reducing harvesting costs relative to the value of the harvestperhaps has been more difficult to achieve than the biological objectives. Fishery managers have found that, when regulations limit effort along one dimension (e.g., days open to fishing), competition reappears along other dimensions (e.g., more boats or larger, faster boats).
Although it is a challenging task to achieve efficient management of a fishery that is confined to a single jurisdiction, further complications emerge when the targeted fish population migrates across international boundaries or straddles the boundary between a national jurisdiction and the international commons of the open ocean. In the case of a coastal fish population that migrates across international boundaries, harvesting in each jurisdiction affects the availability of fish in the other jurisdiction. If these nations harvest the shared stock competitively, they will tend to squander its potential value. Recognizing that possibility, they may attempt to work out a cooperative division of the harvest, but maintaining cooperation is particularly difficult when there are large natural variations in the size, location, or migratory patterns of the fish population.
Uncertainty regarding the magnitude and sources of variations in fish stocks is often a stumbling block to cooperative harvest management. For example, when the availability of fish declines, it may not be immediately apparent if the cause was excessive harvesting by the neighboring nation or a natural fluctuation in abundance. In addition, the parties may have different information or beliefs about how the stock is changing and they may have a strategic interest in concealing that information from one another, or in promoting a particular interest-laden interpretation of the biological facts.
In such circumstances, it is possible that improved information on the links between climatic variations and fish populations could reduce uncertainty and allow the parties to forge a common view as to their best joint harvesting policy. If so, the likelihood of breakdowns in cooperation and associated economic losses might diminish. The extent to which improved seasonal-to-interannual climate forecasts can contribute to improved fishery management is likely to depend on the nature of the management institutions and on the clarity of the links between climate and changes in the fish population.
Forests and Other Ecosystems
El Niño can have major effects on forests and other ecosystems, as seen from recent experience and from paleoenvironmental data, including analyses of pollen, coral, and tree ring records around the world. For example, tree ring records in the U.S. Southwest show the correlation of the width of tree rings with precipitation and with the dry and wet years associated with El Niño. The dates of fires can also be reconstructed through tree ring analyses. In the U.S. Southwest, forest fires often occur when wet winters associated with El Niño and the buildup of vegetation are followed by dry periods associated with La Niña (Swetnam and Betancourt, 1990, 1992).
The 1982-1983 and 1997-1998 El Niño events clearly showed the effects of climatic variations on forest conditions in Austral-Asia and Latin America. In 1982-1983, more than 400,000 hectares of forest burned in East Kalimantan, Indonesia, and wildfires also devastated parts of Australia and southern Brazil. In 1997-1998, fires destroyed forests in Indonesia, the Philippines, Mexico, and Brazil. The World Wildlife Fund estimated the area burned in Indonesia at 6 million hectares, and in Brazil, about 5 million hectares of forest burned in the state of Roraima.
In addition to the obvious damage to the forestry industries of these regions, the impacts on biodiversity are serious. In Indonesia, the fires threatened several species, including endangered orangutans. In Mexico, the Chimalapas nature reserve, one of the regions with the highest biodiversity in North America, was severely damaged by fires in 1998. Costa Rica is concerned about the long-term effects of drought on biodiversity and ecotourism. Although natural vegetation is often adapted to climatic variability (Nicholls et al., 1991), human activity has sometimes increased the vulnerability of biodiversity to drought-induced fires. Policies of fire suppression to protect timber resources, homes, and tourist sites have led to the buildup of fuel and to more serious fires in the long run.
Agricultural encroachment on forests, especially through clearing by burning, has significantly increased the risk of forest fires. Forest managers have attempted to respond to climate variability by trying to obtain a better understanding of natural fire history and using historical knowledge and climate predictions to decide when to reduce fuel buildup through controlled burns. Governments have attempted to impose fire bans, including laws against the traditional slash-and-burn clearing of agricultural lands, and have invested extra resources in their firefighting services in dry years.
Marine ecosystems are heavily influenced by climatic variability, as noted in the discussion of fisheries above. Many of the species that feed on fish fluctuate with fish and marine phytoplankton populations in an El
Niño year. For example, in Chile and Peru, thousands of seabirds died during the 1982-1983 El Niño, and the valued ecosystem of Ecuador's Galapagos Islands was disrupted (Ribic et al., 1992; Trillmich and Limberger, 1985). In California, the warmer coastal waters of El Niño years reduce the fish populations that support seals and other marine mammals, resulting in die-off and reproductive failures (Shane, 1994). Coral reefs are also vulnerable. They experience bleaching under warm water stress and can have high mortality rates in El Niño years (Glynn, 1984). Some species, however, such as shrimp and scallops, flourish in the warmer waters of these years. Managing fluctuations in marine mammal and bird populations is difficult, especially when conservation might involve cutting back on commercial fisheries. Groups have attempted to rescue a few mammals and provide emergency food supplies to birds.
Many riparian and grassland ecosystems are also highly sensitive to climatic variability. Coping systems affecting livestock production on grassland ecosystems are discussed in an earlier section. However, there is significant climate-related variability in the populations of less-managed species in riparian and grassland systems, including breeding birds and amphibians in marshes and wetlands and grassland wildlife populations ranging from rodents to grazers to large carnivores. Severe droughts in southern Africa, for example, are often associated with large-scale mortality of wildlife.
Water Supply and Flood Management
Climate-driven variability in supply is a fundamental characteristic of surface water resources. Various water management entities around the world have planned their infrastructure and operating procedures in response to expected variations in hydrologic conditions. In the United States, these entities range from individual irrigators and domestic water users who control their own water supply systems to federal agencies that oversee the operation of complex multiunit, multiple-purpose water storage, control, and delivery systems. Institutional contexts, which differ markedly between the arid western and humid eastern states, shape the efforts of water users and the large variety of public water managers to cope with variable streamflows. Similarly, other countries have developed institutions and infrastructure for water control and allocation that are the product of particular physical, climatic, and social circumstances. Such arrangements include small-scale traditional irrigation systems that are often managed according to complex allocation rules designed specifically to cope with the effects of variable water supplies, as well as large-scale modern irrigation projects, typically managed by agents of the central government.
Because each party's use and manipulation of a stream or other water source may affect the interests of all other users of that resource, water uses are typically governed by a body of law, custom, and related institutions that define the rights and obligations of each entity. Such institutions have evolved in response to the characteristics of local water supplies, the demands placed on those supplies, and the types of conflicts that have arisen between competing water users.
In the western United States, the scarcity and variability of water supplies, coupled with the predominance of out-of-stream consumptive uses (e.g., crop irrigation), led to adoption of the prior appropriation system of water law (Chandler, 1913; Tarlock, 1989). As streamflows fluctuate, whether due to drought, regular seasonal variation, or sporadic storms, the availability of water to any particular user is determined by the position of that user's right in the priority hierarchy. The oldest rights have highest priority. The familiar statement of the principle is: ''first in time, first in right.'' In the eastern states, the relative abundance of surface water, together with the historical importance of instream water uses (e.g., to power mills) favored the riparian system of water law. Under that system, owners of stream-side properties share coequal rights to reasonable use of the water resource (Clark, 1970; Tarlock, 1990; Rose, 1990). In the modern era, the riparian doctrine states of the United States have largely shifted to a system of state-issued permits (Abrams, 1990; Sherk, 1990). These different legal traditions continue to shape water use and management and to affect the impacts of a drought and the options for responding to it (Miller et al., 1997). The drought management tools available in the western United States include short-term water transfers from willing sellers to willing buyers. Such voluntary marketing generally is not possible in the riparian tradition states, where state agencies may play a central role in allocating water supplies during drought emergencies.
Responses to seasonal-to-interannual variations in water supply historically have taken the form of long-term investments in surface water storage, groundwater pumping capacity, and transbasin diversions. In the western United States, where rapid population growth has surged on watersheds in which farmers and ranchers long ago appropriated the reliable streamflows, municipal water suppliers and other new users often permanently purchase water rights from irrigators and other senior users to obtain reliable supplies. Infrastructure investments or permanent water transfers can insulate out-of-stream water uses and some navigation and hydropower uses from many of the effects of seasonal-to-interannual variations in runoff, although often at the expense of environmental, aesthetic, and cultural values. For example, reservoir operations may remove water from streams and damage aquatic habitats,
whereas permanent sales of water rights from agriculture areas may erode the economic base for rural communities.
In recent years, growing recognition of the environmental and social costs of dams, transbasin diversions, and permanent water transfers has generated a search for more flexible alternatives. These include short-term water transferseither privately arranged or through an organized water bankand conjunctive use of groundwater and surface water sources. These conjunctive use programs frequently involve active recharge of aquifers during periods of surface water abundance.
California has recently created a series of emergency drought water banks (Miller, 1996) and purchased, in one instance, option contracts on water in anticipation of a water bank that later proved to be unneeded (Jercich, 1997). In the first water bank in 1991, the state's Department of Water Resources purchased more than 800,000 acre-feet of water, but heavy rains in March reduced demand for that water and it was forced to carry over approximately 265,000 acre-feet at a cost to the state of about $45 million (Lund et al., 1992). This experience has led the department to adopt a more cautious approach to purchasing water early in the season.
Seasonal-to-interannual climate variations also affect public stewardship of water-dependent natural resources. For example, water quality tends to decline under low-flow conditions, affecting many aquatic species that are sensitive to the temperature increases, loss of habitat, and changes in water chemistry that intensify when streamflows decline. In the eastern United States, specific statutes, court decisions, and conditions on state-issued water diversion or discharge permits may determine the range of options available to public agencies to manage the effects of climatic variations on water quality and riparian ecosystems. In the western states, management is constrained by a system of water quantity allocation that makes no allowance for water quality (Tarlock, 1989). Although the majority of western states allow public agencies to acquire water rights and use them to maintain streamflow levels, most of those rights have such junior status that they may be of little practical significance (Wilkinson, 1989). Agencies can cope more effectively where laws allow private groups to purchase senior agricultural water rights and donate them to instream use in perpetuity (Colby, 1993), or where public agencies have used water banks to acquire water for environmental purposes, as in California and Idaho (Miller, in press).
Flood events tend to develop over shorter time scales than droughts, although unusually high seasonal precipitation resulting in heavy winter snowpack is likely to increase the probability of flooding. The applicability of seasonal-to-interannual forecasts to flood management may depend on the extent to which decision makers can alter their activities in response to information on changes in seasonal flood risks. Flood manage-
ment has typically coupled long-term structural measures (e.g., dams and levees) and flood plain management with short-term warnings of impending flood events and emergency responses to actual floods.
Recent large flood events have spurred considerable rethinking of flood management policies and infrastructure design in the United States (Mount, 1995; Changnon, 1996; Pielke, 1996). Systems of dams and levees, which limit flood damage in most years, have been blamed for exacerbating the devastation caused by great floods, such as the Mississippi flood of 1993. Flood damages have increased in inflation-adjusted terms over the course of this century; however, it is not clear whether per capita or wealth-adjusted vulnerability to flooding has increased or decreased (Pielke, 1996).
The United States and other developed countries typically manage flooding as a hazard to be avoided and controlled; some societies in the developing world have designed their agricultural activities to make use of annual cycles of flooding. Villagers along the Senegal River, for example, plant their crops on bottomlands as the annual flood waters recede (Magistro, 1998). In such cases, more accurate forecasts of the timing and extent of annual flooding might help such societies to anticipate good and bad agricultural years, giving them additional lead time to implement such coping strategies as migration to cities to seek additional income.
More accurate long-term forecasts of regional flood probabilities might allow more effective planning and deployment of emergency flood management and relief operations and perhaps improved prioritization of federal levee repair and maintenance investments. However, currently available long-term flood outlooks are neither well understood nor effectively used by many public and private decision makers (Changnon, 1996; Pielke, 1997).
Human health is sensitive to several types of climatic variation. Some sensitivities are to extreme events. Extreme temperatures cause hypothermia or heat stress in unprotected individuals, and precipitation shortfalls can bring droughts that reduce crop yields, resulting in famine and malnutrition. Climate can also affect human health more indirectly through its effect on ecosystems. An important instance is changes in the ecology of infectious disease organisms or their vectors that can precipitate disease outbreaks. An illustration is the story of the hantavirus outbreak of 1993. A prolonged drought in the U.S. Southwest in the early 1990s reduced the populations of animals such as owls, coyotes, and snakes that prey on rodents. When the drought yielded to intense rains associated with the 1992-1993 El Niño, the grasshoppers and piñon nuts
on which local rodents feed became more abundant. The result, when combined with the drop in predators, was a tenfold increase in rats and mice (Levins et al., 1993; Epstein, 1994) and the emergence of a "new" diseasecalled hantavirus pulmonary syndromestemming from a virus and transmitted through rodent droppings.
The effects of climatic variations on ecosystems have been shown to be related to outbreaks of malaria (Bouma et al., 1994a, 1994b; Hales et al., 1996), dengue fever, and other mosquito-borne diseases (Loevinsohn; 1994), which spread when appropriate rainfall conditions or higher daytime minimum temperatures favor mosquito breeding and survival. Climate variations, by altering functional relationships within the marine food web (Roemmich and McGowan; 1995), may increase the risks to humans from paralytic, diarrheal, neurologic, and amnesic shellfish poisoning (Epstein et al., 1993b) and cholera (Colwell, 1996). It is at least suggestive that domoic acid poisonings, resulting from diatom blooms that produce toxins in seafood, appeared in Canada in the El Niño year of 1987 (Todd, 1989; Todd and Holmes, 1993), and related phenomena occurred in California, Argentina, and Scandinavia in the El Niño year of 1992 (Ludlohm and Skov, 1993; Carreto and Benevides, 1993). The cold phase of ENSO can also create conditions, such as intense rains and flooding following prolonged drought, that are optimal for breeding insect vectors of dengue fever and Venezuelan equine encephalitis and for rodent transmission of leptospirosis (Epstein et al., 1995). Many such associations have been documented, and where the ENSO signal is closely correlated with weather patterns, predictive models of conditions conducive to disease outbreaks may be useful. The "ENSO Experiment" begun in spring 1997 by the National Oceanic and Atmospheric Administration's Office of Global Programs coordinates scientific work by health researchers, ecologists, and meteorologists examining the relationships between ENSO and a variety of infectious diseases and marine ecological disturbances.
Human coping with disease has primarily involved year-round precautions such as individual maintenance of good nutrition, food refrigeration, and collective programs of sewage treatment, water chlorination, testing for red tide and fecal coliform bacteria, air quality testing and alerts, mass vaccination, and the like. Some coping activities also involve seasonal routines, such as the use of mosquito netting and insect repellents and alerts for heat waves and extreme cold. Public health systems do, however, also respond to forecasts of disease outbreaks, for example, with annual programs to develop and disseminate vaccinations against the influenza strains considered most likely to infect a population in a given winter. Thus, public health is a potential beneficiary of improved climate forecasting.
Other Weather-Sensitive Sectors
Many other classes of human activity are affected by climatic events and have developed coping strategies. An important example is the sensitivity of households and firms to weather- and climate-induced emergency conditions such as floods, hurricanes, and drought-induced wildfires. A large body of research on disaster preparedness and response has produced a number of syntheses, new theoretical approaches, and major advances in applying what is known. This research gives a good picture of the levels of preparedness (ex ante coping) typically found among households (Drabek, 1986; Mileti and O'Brien, 1992; Mulilis and Duval, 1995) and organizations (Wenger, 1986; Gillespie and Streeter, 1987; Lindell and Meier, 1994) in the United States and the nature of post-disaster (ex post) response efforts, both for households (Perry and Mushkatel, 1986; Phillips, 1992; Tierney, 1993; Morrow and Enarson, 1995) and organizations (Drabek, 1986; Drabek and Dynes, 1994; Wenger et al., 1989). We discuss community-level preparedness and response, in the context of institutions for coping, in the next section.
Households, even those that are trying to prepare for disasters, in fact do very little. Household preparedness (Turner et al., 1986; Palm et al., 1990; Mileti and Darlington, 1995; Russell et al., 1995), and that of organizations as well (Mileti et al., 1993; Drabek, 1995; Perry and Lindell, 1996), is constrained by the low salience of hazards, the competition of preparedness with more pressing concerns, and inadequate resources. Households whose members belong to nonminority groups do more to prepare than other households (e.g., Perry and Greene, 1982; Perry, 1987), but the reasons for this remain unknown.
Post-disaster response among households and organizations is shaped by a variety of social, social-psychological, and cognitive processes, including prior disaster experience and the existence of government mandates. Research in the United States consistently shows that social solidarity remains strong even in the most trying of circumstances, and few situations occur that completely break down social bonds and eliminate the feeling of responsibility people have for one another. This is true at both individual (Dynes et al., 1990; O'Brien and Mileti, 1992) and organizational levels (Wenger at al., 1989; Tierney, 1993) and helps account for the prevalence of volunteerism, self-reliance, and the emergence of social groups after disasters. Thus, strong efforts at response are often mounted even where there is a low level of preparedness.
Various industries are sensitive to climatic variations. For example, in the energy industry, suppliers of natural gas and electricity are affected by changes in their seasonal demand profilesa cold winter or hot summer will increase demand for energy, which companies may be able to
supply with greater reliability and at lower cost if they can accurately forecast demand. Suppliers of hydroelectric power are also sensitive to streamflow at their dam sites, and distributors of electric power are sensitive to severe storms that may bring down power lines.
Local gas distribution companies that are sensitive to demand in extreme weather conditions cope, regulations permitting, by charging weather-normalized rates to dampen fluctuations in revenue across warm and cold winters. They also hedge by using multiple suppliers, storing gas in summer for use in winter, and keeping enough gas on hand for a winter that is 10 percent colder than average. Gas distributors commonly also use 7-day forecasts of weather and heating degree-days for planning, but they have not typically used the weather service's 3-month forecasts (Changnon et al., 1995, Golnaraghi, 1997). Electricity distribution companies use 10-day weather forecasts, mainly to anticipate major storms, and hydroelectric power producers forecast water inflow to reservoirs (Golnaraghi, 1997).
There has been little systematic study of coping in other climate-sensitive industries, such as construction. However, it is reasonable to presume each such sector uses a variety of coping strategies, both ex ante and ex post, developed out of past experience with unpredicted climatic variations.
Institutions For Coping With Climate Variability
Societies cope with climatic variability on a level beyond the affected individuals, organizations, and sectors by developing institutions that help those actors cope better. This section discusses a few of the important institutions that perform this function.
Disaster Insurance and Reinsurance
An important part of the coping system in many countries is the system of property and casualty insurance. Insurers offer financial compensation to subscribers who suffer from extreme climatic events such as floods, droughts, and hurricanes, thus reducing the risks that face actors who are insured.
Insurers can buffer the effects of climate variation in several ways. One is through their primary function of spreading risks over a large pool of subscribers. They can also influence subscribers to take other actions to reduce their own sensitivity, for example, by offering lower premiums for hurricane coverage of homes that meet standards of stormworthiness or that are located in municipalities that adopt and enforce building codes that reduce hurricane risk.
Insurance firms are themselves vulnerable to climate variations through extreme events that cause simultaneous covered losses for a large proportion of their subscribers. To cope with this risk, some insurers purchase reinsurance from other companies or government agencies. Reinsurers are also vulnerable to climatic events that affect a large portion of their subscribers. A special case in the United States is the federal government, which issues millions of policies under the National Flood Insurance Program, advertises to issue more, and acts as its own reinsurer, thus spreading the risk of major floods among taxpayers in general.
In the United States, there has been sharply increased attention to the vulnerability of the insurance and reinsurance industries since Hurricane Andrew in 1992, which caused $16 billion in insured property lossesmore than twice the losses of the worst case the industry expected (Changnon et al., 1997). An immediate result was restrictions on coverage that made insurance a less reliable coping strategy for potential subscribers. Since 1992, the property insurance industry has responded by creating the Institute for Business and Home Safety, which is concerned with improving building codes and conducts research on improved building materials and techniques, and by beginning to support basic science, establishing a Risk Prediction Initiative at the Bermuda Biological Station (Golnaraghi, 1997; also on the internet at <http://www.bbsr.edu/agcihome/rpi/rpihome.html>). In addition, several firms in the financial industry are offering products for managing and transfering the financial risks of catastrophic exposure. Insurance firms have also asked regulatory authorities for permission to base rates on expectations of future loss rather than only on historical experience. Some insurers have hired climate scientists on their staffs, and many employ risk-modeling companies that, among other things, interpret climate forecast information in terms of its implications for the risk profile of particular insurance companies (Golnaraghi, 1997). These innovations may increase the reliability of commercial insurance as a coping strategy for other sectors.
Emergency Preparedness and Response
Many societies help affected sectors cope with climatic variations by creating general systems of emergency preparedness and response. These include national weather services, which forecast storms and other significant weather-related events (including, recently, climatic variations), thus enabling better ex ante responses. Other national or regional organizations in some countries perform a similar function by providing fire danger and flood warnings. Local emergency response organizations that provide fire, rescue, and emergency medical services are also part of the emergency preparedness and response system. These organizations
prepare and then respond ex post to climatic and other emergencies regardless of the cause or the sectors of a community that are affected. Regional and national governments may also provide emergency response services, such as fighting forest fires and maintaining order in communities devastated by floods or hurricanes. Governments sometimes offer disaster relief payments or subsidized loans for reconstruction after disasters. And in addition, nongovernmental organizations such as the Red Cross stand ready to aid in ex post response.
A large body of research has examined systems of emergency preparedness and response and developed general knowledge about how they function and the conditions under which they function most effectively. Although the situation is improving, studies in the United States show that, with notable exceptions, disaster preparedness at the local level is usually not well maintained, that emergency planners tend to have low prestige, and that relatively few resources are allocated to disaster preparedness and response (Rossi et al., 1982; Labadie, 1984; Gillespie, 1991). The research record provides little information on the status of state- and national-level preparedness. The adequacy of local response varies based on the degree of pre-event preparedness in place (Mileti and Sorenson, 1990).
A great deal is known about the factors that influence organized and effective community response to disaster. Effective response results from preparedness within a variety of organizations and by networks of organizations. Preparedness within an organization is enhanced if the organization is not surprised by events. Thus, organizations are more likely to perform well the work required in an emergency if the individual or role responsible for each task is well specified; if the definition of the emergency work domain is clearly set forth for all divisions and actors; if authority to perform the required work is clearly marked; if priorities among tasks and work have been clearly established; and if roles, tasks, authority, domains, and priorities are well understood by organizational actors and legitimated in advance of the emergency rather than negotiated during a disaster. These conditions may be achieved through training or because emergency work matches nonemergency work. Effectiveness is also enhanced if communication channels between divisions are open, clear, and frequently used, allowing efficient sharing of critical information that appears during an emergency (Mileti and Sorenson, 1990).
Disaster preparedness and response also require the effective operation of networks of organizations. Such networks respond best if they are well integrated before a disaster occurs and if they maintain sufficient flexibility to respond to surprise. Network integration means that the roles and tasks of each member organization, authority for relations between organizations, and priorities for tasks and work between organiza-
tions are defined in advance; that linkages between member organizations are well understood; and that adequate resources are available to support interorganizational linkages. Networks function better if there is consensus on the tasks expected of each member; if each member has adequate resources to do the expected work; if the cost to each member for membership is low; and if the leaders of member organizations need not fear loss of organizational autonomy as a result of participating in the network. Effective networks tend to include boundary personnel (people who have the job of interacting with other organizations), individuals who belong to several organizations in the network, interorganizational boards and committees, and a superorganizational board. Ideally, interorganizational interactions are frequent and reciprocal rather than one-way, and communication patterns are clear, open, and broad as to content. In addition, networks are more likely to be effective if they are composed of smaller numbers of organizations that are compatible in terms of goals, function, and scope and if they have been initiated by their member organizations rather than created by outside request or legislative mandate (Mileti and Sorenson, 1987).
Market institutions have not been much studied as coping mechanisms for climatic variability, but this is one of the functions they serve. Two examples illustrate. One is the emergence of global markets for grains and other foods. These markets reduce the dependence of human populations on food grown nearby and therefore their dependence on local climatic conditions. They also allow producers to benefit from climatically induced food shortages elsewhere by supplying food to those areas. These effects, however, are contingent on the ability of producers and consumers to participate in the global markets. For consumers, this means having money to purchase food at market prices and access to distribution networks; for producers, it means the ability to ship their products. Thus, markets alone do not insulate the poor from the effects of climatic variation nor secure benefits for producers in remote areas. Nevertheless, to the extent that global food markets function well, they spread the risks and benefits of climatic variability worldwide.
A second example of how markets help cope with climatic variation is the functioning of commodities futures markets. These markets allow producers and distributors of food and other weather-sensitive commodities to hedge against climatically induced variations in production by guaranteeing themselves the price or availability of a known quantity of the commodity at a later date. As with food markets, futures markets do not benefit everyone equally. To benefit from the potential to hedge, an
actor must have a sophisticated knowledge of how the markets work and must have a large enough interestor an association with others who together have a large enough interestto make the minimum transactions the market allows.
This chapter shows that both the effects of climatic events on human populations and activities and the potential usefulness of climate forecast information are shaped by sets of coping strategies that have been developed over long periods of time and that are in constant development and change. Specifically: