Much previous work on adapting to climate change ignores the context in which people must evaluate the risks of changing their livelihood strategies or the risks of avoiding change. New adaptive strategies to climate change do not appear de novo and in a vacuum: actors decide whether the new experience is comparable with anything they have experienced or culturally known and classified in the past. If yes, the adaptive behaviour will probably repeat the previously used, proved scripts. If no, differences in the new experience will trigger more complex assessment to see whether the new experience can be fitted to the existing cultural categories and behaviour or requires more contextual assessment. If the latter is the case, the person/family must evaluate whether the benefits of producing new scripts and routines outweigh the costs, whether risk and uncertainty are increased and how large the pay-off will be if the new routine is useful versus potential failure of a well-tried routine. These risks are assessed depending on the constraints of the individual as a member of a household and a community, and the external constraints he/she faces.
To understand human dimensions of climate change, we need to begin by examining the adaptive mechanisms of human populations to environmental change, the differential responses to the magnitude and the frequency of perceived and actual changes, and the differences between adaptive responses at the individual level and those visible at the population level. Adaptive responses to environmental change are mediated by multiple factors, for example, perception of change in cultural and linguistic dimensions, such as whether people have experienced that type of change, whether it is easily understood and interpreted by existing cues and an appropriate lexicon and whether there is a repertoire of responses to that specific change or to a family of similar changes. If flooding is a regular occurrence in a region, established residents are more likely to have a specific terminology and the ways to cope than new residents who have never experienced a flood, or if flooding occurs infrequently, the memory of past events may not be part of inter-generational cultural knowledge.
We know that individual-level responses to environmental changes are time dependent and heterogeneous, with some individuals coping quickly and others taking one or more generations to adapt. What is true for the adoption of technological innovations (Rogers 2003
) is also true for behavioural adaptation to environmental change. While a few individuals may adapt quickly to an environmental shift, a population will take at least one generation to become fully adjusted (Moran 2000
). Individuals and households also vary in their capacities to respond: younger adults cope with the change more quickly than elderly adults because it is part of the new experiences and cultural scripts they incorporate as they mature, whereas elders have longer contradictory experiences, set behavioural routines and have more to lose if the change is short term rather than permanent.
In addition, households have different economic levels, thus different capacities to take risks (Cancian 1989
). Those at the top of the social ladder have a lot to risk in terms of cost if they change their practices for the short term, and the consequences of a mistake can result in a downward slide in status and wealth. Other community members who aspire to be at the top of the ladder are more eager to take risks with high potential for positive pay-off. Thus, they are more likely to innovate in their strategies, with hopes of racing ahead. Should their risk pay off, the wealthier families may follow to protect their positions, but some will resist change for some time.
One of the challenges to human adaptation to climate change is to understand the scale of the problem: most environmental perception is local rather than global, and is manifested in experience with changes in precipitation and temperature and observation of crop responses to current conditions (see diverse examples in Magistro et al. (2001)
). Moreover, the changes are likely to occur within a very short time frame, normally 2 or 3 years. At this temporal scale, there is much fluctuation in temperature and precipitation; thus the data are ‘noisy’ and hard to interpret. Local perspective confounds this because within any given watershed or region, people know that properties and households can have different experiences with crop responses, precipitation and temperature.
We need to recognize these relationships, in the Amazon and elsewhere, as complex cross-scale problems (Brondizio et al. in review
). The literature on human–environment interaction has had a tendency to treat human responses to change at a local, self-contained level, understandably an attractive approach from an analytical perspective, but limited in capturing the linkages and vertical interplay created by a growing functional interdependency of resource use systems and ecosystems (Berkes 2006
; Young 2006
). Achieving a link between local and global understanding of climate change requires connections within a broad information network, and a trust that the information from both levels is credible (Cash et al. 2006
). When the national and global forecasts contradict local experience, people resist believing them because they do not match personal understanding and experiences of climate patterns.
Research on human adaptation to climate events has been encouraged by the Climate and Global Change Program of the US National Oceanic and Atmospheric Administration to understand how people respond to weather forecasts. This has resulted in a growing body of evidence for the roles of institutions, culture, social class and perception in how people respond to forecasts (e.g. Nelson & Finan 2000
; Orlove et al. 2000
; Lemos et al. 2002
; Moran et al. 2006
). Where robust local institutions exist and are linked to larger institutions, the flow of information is quicker and more effective (Ostrom 2005
; Toniolo 2005
). Where a connection exists between climate factors and cultural practices, adaptation is likely to be quicker (Magistro et al. 2001
). Most of the research has identified obstacles to people believing and acting upon the forecasts, which come largely from the mismatch of local experience and national and global forecasts. In the Amazon, research and forecasting models for phenomena like ENSO and climate change, in general, are produced at coarse resolutions and scale: global; continental; and macroregional. Information at these levels is often all that is communicated to rural communities and farmers, given the limitations of, on the one hand, information dissemination systems and, on the other hand, data and models in generating locally precise data to most places on the planet (Pfaff et al. 1999
; Stern & Easterling 1999
However, the growing interdependence of resource use systems, commodity markets and climate requires attention to mechanisms operating at larger scales but influencing local-level adaptation. For instance, it is important to understand how farmers engage in collective action and interact with other sectors of the population within the landscape and affected by each other's land-use behaviours, such as the use of fire or management of a watershed. We also need to consider the way farmers receive and interpret information derived from global and national forecasts projecting interannual and seasonal climate patterns or weekly and daily weather information. The media play a central but not always informative role in disseminating this information, particularly in areas such as the Amazon, where the scale tends to be broad, data resolution is coarse and other forms of dissemination are infrequent and have limited reach. A farmer's interpretation and trust of such information depend on the spatial and temporal scales with which it is presented, its relevance to local-level decision making, clarity of language and terminology and even how politicized it appears (e.g. serving particular economic interests).