Climate Change 2001:
Working Group II: Impacts, Adaptation and Vulnerability
Other reports in this collection

18.3.3. Models, Analogs, and Empirical Analysis of Autonomous Adaptation

Adaptation to rapid anthropogenic climate change may be a new challenge, but individuals, societies, and economies have adapted—in various ways and with various degrees of success—to changed and variable environmental conditions throughout history. These experiences in adaptive behavior provide information on the processes, constraints, and consequences of adaptations.

Knowledge of the processes by which individuals or communities actually adapt to changes in conditions over time comes largely from analog and other empirical analyses (Wigley et al., 1981; Glantz, 1996; Meyer et al., 1998; Tol et al., 1998; Smit et al., 1999; Yohe and Dowlatabadi, 1999; Bryant et al., 2000). These studies indicate that autonomous adaptations tend to be incremental and ad hoc, to take multiple forms, to be in response to multiple stimuli (usually involving a particular catalyst and rarely climate alone), and to be constrained by economic, social, technological, institutional, and political conditions.

Conceptual models of adaptation processes describe sequential relationships and feedback involving climatic and nonclimatic stimuli, system sensitivities and impacts, tactical and strategic adaptations, and net or residual impacts. They also indicate conditions that constrain or facilitate various kinds of adaptation (e.g., Carter, 1996; Smit et al., 1996; UNEP, 1998; Schneider et al., 2000). Spatial analogs have been used to gain insight into adaptation, by transferring experience from existing climatic regions to places where such climate may be found in the future. The contributions and limits of spatial analogs are known (Schneider, 1997; Rayner and Malone, 1998). Some ecological and paleoecological studies reconstruct species or community dynamics over hundreds and thousands of years (e.g., MacDonald et al., 1993).

Temporal analog or case studies document adaptive responses to climatic stimuli in resource-based economic sectors and communities over periods of several decades (e.g., Glantz, 1988; Olsthoorn et al., 1996; Changnon et al., 1997). Other empirical analyses have examined adaptive behavior in key sectors such as agriculture in light of climatic variability and extremes over even shorter time periods (e.g., Appendi and Liverman, 1996; Smit et al., 1997; Bryant et al., 2000).

These direct empirical analyses of adaptation processes tend to start with the system of interest, then assess its sensitivity and adaptability to climate and other stimuli. This analytical strategy is consistent with vulnerability assessment (Downing et al., 1996; Adger, 1999; Handmer et al., 1999; Kelly and Adger, 1999), the "adjoint approach" (Parry, 1986), and "shift-in-risk" perspectives (Warrick et al., 1986). These studies have yielded some important insights about adaptation.

For systems such as agriculture, forestry, water resources, and coastal zone settlements, the key climatic stimuli are not average conditions but variability and extremes. A direct climatic condition prompts adaptation less often than the economic and social effects or implications of the climatic stimuli that are fundamental in triggering adaptive responses. Nonclimatic conditions are important in moderating and sometimes overwhelming the influence of climate stimuli in the decisionmaking of resource users. Decisions on adaptation are rarely made in response to climate stimuli alone. These findings are important for predicting autonomous adaptations and for improving adaptation assumptions in impact models.

In estimating future adaptations and developing adaptation policies (see Section 18.4), it is helpful to understand factors and circumstances that hinder or promote adaptation. As Rayner and Malone (1998) conclude, the consequences of a climate event are not direct functions of its physical characteristics; they also are functions of "the ways in which a society has organized its relation to its resource base, its relations with other societies, and the relations among its members." To understand vulnerability in archeological, historical, and contemporary contexts, Rayner and Malone (1998) identify the most promising research strategy:

" …explicitly to focus attention on the process of adaptation—or, on the other hand, of failure to adapt—that partly condition the impact of the climatic stress in particular societies…cases in which societies appear to have been seriously damaged by, or even totally succumbed to, climatic stress should not be taken to demonstrate the determining influence of climate. It is essential to consider ways in which these societies might have coped better, and to focus on the political, cultural, and socioeconomic factors which inhibited them from doing so" (Ingram et al., 1981).

Following this approach, McGovern's (1991) reexamination of the extinction of Greenland settlements found that the stress imposed by climate shifts was indeed severe but was within the theoretical ability of the colonies to have coped, by means that were available to them. Why they failed to employ those adaptive means emerges as the key question, still incompletely answered, in explaining the collapse: "It did get cold and they did die out, but why?" (McGovern, 1991). Intervening between the physical events and the social consequences is the adaptive capacity and hence vulnerability of the society and its different groups and individual members.

18.3.4. Costs of Autonomous Adaptation

As assessments of climate impacts (commonly measured as "costs" that include damages and benefits) increasingly have incorporated expected adaptations, and particularly as impact models and "integrated assessment" models have shown the potential of adaptation to offset initial impact costs, interest has grown in calculating the costs of autonomous adaptations. Whether climate change or another climate stimulus is expected to have problematic or "dangerous" impacts depends on the adaptations and their costs (Leary, 1999). Climate change impact cost studies that assume adaptation also should include the "adjustment of costs" of these adaptations (Reilly, 1998).

Tol and Fankhauser (1997) provide a comprehensive summary of analyses of the costs of autonomous, mainly (but not exclusively) reactive adaptations, undertaken privately (i.e., not adaptation policies of government). A common basis for evaluating impact costs is to sum adaptation costs and residual damage costs (Fankhauser, 1996; Rothman et al., 1998). Procedures for defining and calculating such adaptation costs are subject to ongoing debate. Tol and Fankhauser (1997) note that most approaches consider equilibrium adaptation costs but ignore transition costs. Hurd et al. (1997) include market and nonmarket adaptation in their assessment of impact costs. Most research to date on adaptation "costs" is limited to particular economic measures of well-being (Brown, 1998). Any comprehensive assessment of adaptation costs (including benefits) would consider not only economic criteria but also social welfare and equity.

Cost estimation for autonomous adaptations is not only important for impact assessment; it also is a necessary ingredient in the "base case," "reference scenario," or "do-nothing option" for evaluations of policy initiatives, with respect to both adaptation and mitigation (Rayner and Malone, 1998; Leary, 1999; Smit et al., 2000).

18.3.5. Lessons from Adaptation Experiences

Research in many sectors and regions indicates an impressive human capacity to adapt to long-term mean climate conditions but less success in adapting to extremes and to year-to-year variations in climatic conditions. Climate change will be experienced via conditions that vary from year to year, as well as for ecosystems (Sprengers et al., 1994) and human systems (Downing et al., 1996); these variations are important for adaptation. Thus, although human settlements and agricultural systems, for example, have adapted to be viable in a huge variety of climatic zones around the world, those settlements and systems often are vulnerable (with limited adaptive capacity) to temporal deviations from normal conditions (particularly extremes). As a result, adaptations designed to address changed mean conditions may or may not be helpful in coping with the variability that is inherent in climate change.

All socioeconomic systems (especially climate-dependent systems such as agriculture, pastoralism, forestry, water resources, and human health) are continually in a state of flux in response to changing circumstances, including climatic conditions. The evidence shows that there is considerable potential for adaptation to reduce the impacts of climate change and to realize new opportunities. In China's Yantze Valley, 18th-century regional expansions and contractions on the double-cropping system for rice represented adaptive responses to the frequency of production successes and failures associated with climatic variations (Smit and Cai, 1996). Adaptation options occur generally in socioeconomic sectors and systems in which the turnover of capital investment and operating costs is shorter and less often where long-term investment is required (Yohe et al., 1996; Sohngen and Mendelsohn, 1998).

Although an impressive variety of adaptation initiatives have been undertaken across sectors and regions, the responses are not universally or equally available (Rayner and Malone, 1998). For example, the viability of crop insurance depends heavily on the degree of information, organization, and subsidy available to support it. Similarly, the option of changing location in the face of hazard depends on the resources and mobility of the affected part and on the availability and conditions in potential destination areas (McGregor, 1993). Many response strategies have become less available; many others have become more available. Individual cultivator response to climate risk in India has long relied on a diverse mix of strategies, from land use to outside employment (sometimes requiring temporary migration) to reciprocal obligations for support; many of these strategies have been undermined by changes such as population pressure and government policy, without being fully replaced by others—illustrating the oft-remarked vulnerability of regions and populations in transition (Gadgil et al., 1988; Johda, 1989). In areas of China, many historical adaptations in agriculture (e.g., relocating production or employing irrigation) are no longer available as population pressures increase on limited land and water resources (Fang and Liu, 1992; Cai and Smit, 1996). In Kenya, effective smallholder response to drought has shifted from traditional planting strategies to employment diversification (Downing et al., 1989).

Not only is there rarely only one adaptation option available to decisionmakers (Burton and Cohen, 1993) but also "rarely do people choose the best responses—the ones among those available that would most effectively reduce losses—often because of an established preference for, or aversion to, certain options" (Rayner and Malone, 1998). In some cases there is limited knowledge of risks or alternative adaptation strategies. In other cases, adoption of adaptive measures is constrained by other priorities, limited resources, or economic or institutional barriers (Eele, 1996; Bryant et al., 2000; de Loë and Kreutzwiser, 2000). Recurrent vulnerabilities, in many cases with increasing damages, illustrate less-than-perfect adaptation of systems to climatic variations and risks. There is some evidence that the costs of adaptations to climate conditions are growing (Burton, 1997; Etkin, 1998). There is strong evidence of a sharp increase in damage costs of extreme climatic or weather events (Berz, 1999; Bruce, 1999). Growing adaptation costs reflect, at least in part, increases in populations and/or improvements in standards of living, with more disposable income being used to improve levels of comfort, health, and safety in the short run. It is not clear whether the expansion in adaptations is likely to be effective and sustainable in the long run. In any event, although adaptations to changed and variable climatic conditions are undertaken, they are not necessarily effective or without costs.

Many adaptations to reduce vulnerability to climate change risks also reduce vulnerability to current climate variability, extremes, and hazards (El Shaer et al., 1996; Rayner and Malone, 1998). Measures that are likely to reduce current sensitivity of climate variations in Africa also are likely to reduce the threat of adverse impacts of climate change (Ominde and Juna, 1991):

" Most analysts in the less-developed countries believe that the urgent need, in the face of both climate variation and prospective climate change, is to identify policies which reduce recurrent vulnerability and increase resilience. Prescriptions for reducing vulnerability span drought-proofing the economy, stimulating economic diversification, adjusting land and water uses, providing social support for dependent populations, and providing financial instruments that spread the risk of adverse consequences form individual to society and over longer periods. For the near term, development strategies should ensure that livelihoods are resilient to a wide range of perturbations." (Rayner and Malone, 1998)

Examples of current adaptation strategies in agriculture with clear applications to climate change are given by Easterling (1996) and Smit et al. (1997), including moisture-conserving practices, hybrid selection, and crop substitution. In the water resources sector, Stakhiv (1996) shows how current management practices represent useful adaptive strategies for climate change. Some analysts go further to point out that certain adaptations to climate change not only address current hazards but may be additionally beneficial for other reasons (e.g., "no regrets" or "win-win" strategies) (Carter, 1996).

Societal responses to large environmental challenges tend to be incremental and ad hoc rather than fundamental (Rayner and Malone, 1998). In all of the climate analog cases examined by Glantz (1988), "Ad hoc responses were favored over long term planned responses. As a result, there has been a tendency to ‘muddle through.' This has not necessarily been an inappropriate response, but it is probably more costly in the long term than putting a long-term strategy together in order to cope with climate-related environmental change." In each case, moreover, action was not taken without a catalyst or trigger that dramatically indicated the seriousness of a threat (Glantz, 1988). Other studies also indicate the ad hoc nature of adaptations and the importance of a catalyst (Wilhite et al., 1986; Glantz, 1992; Kasperson et al., 1995). These findings suggest that problems that demand early or long-term attention often fail to receive it, and the most efficient responses are not taken. That the earlier action would have been more efficacious, however, presupposes that the best strategy was evident to the decisionmakers and that premature responses closing off useful options would not have been taken instead (Rayner and Malone, 1998). There is little evidence that efficient and effective adaptations to climate change risks will be undertaken autonomously.

A consistent lesson from adaptation research is that climate is not the singular driving force of human affairs that is sometimes assumed—but neither is it a trivial factor. Climate is an important resource for human activities and an important hazard. Climate change is a source of significant stresses (and perhaps significant opportunities) for societies, yet it has always been only one factor among many. The consequences of a shift in climate are not calculable from the physical dimensions of the shift alone; they require attention to human dimensions through which they are experienced (Rayner and Malone, 1998; Bryant et al., 2000). The significance of climate change for regions depends fundamentally on the ability and likelihood of those regions to adapt.

To what degree are societies likely to adapt autonomously to avoid climate change damages? Some studies show faith in market mechanisms and suggest considerable capacity of human systems to adapt autonomously (Ausabel, 1991b; Mendelsohn et al., 1996; Yohe et al., 1996; Mendelsohn, 1998; Mendelsohn and Neumann, 1999). Other studies highlight the constraints on "optimal" autonomous adaptation, such as limited information and access to resources, adaptation costs, and residual damages; these studies emphasize the need for planned, especially anticipatory, adaptations undertaken or facilitated by public agencies (Smith et al., 1996; Reilly, 1998; Tol, 1998a; Fankhauser et al., 1999; Bryant et al., 2000; Schneider et al., 2000)

Other reports in this collection