How have human societies reacted to climate change in the past and what lessons may we learn for the future from the interactions between climate variability
North-eastern North America
North-eastern North America
Central America (1 °S-9°N)i n I I I I I I I I I I I
Figure 3.9 Biomass burning reconstructions for different parts of the world expressed as a charcoal index, derived from stratigraphic data, compared with the atmospheric CO2 concentration curve from the Taylor Dome, Antarctica (Carcaillet etal. 2002).
290 "I Taylor Dome 280- (Antarctica)
290 "I Taylor Dome 280- (Antarctica)
10 000 5000 0
Calibrated years before present
10 000 5000 0
(notably the incidence of damaging extremes such as severe drought) and the changing patterns of social organization developed as a response to the resulting environmental stresses? This theme too is controversial. At one extreme, those who favor a degree of environmental determinism can point to many instances
Figure 3.10 Estimates ofthe increase in biospheric carbon storage between the Last Glacial Maximum and the present day. Note the contrast between the range of sedimentologically derived estimates and those based on modeling and isotope studies. PMIP, Paleo-model Intercomparison Project. (Reproduced from Oldfield F. (2005) Environmental Change: Key Issues and Alternative Approaches. Cambridge University Press, Cambridge.)
where the transformation, decline or collapse of societies, even whole civilizations, coincides with major climatic changes or periods of intense and persistent drought (e.g. Hodell et al. 1995; Weiss 1997; Cullen et al. 2000; deMenocal 2001; Weiss and Bradley 2001; Haug et al. 2003; Huang et al. 2003; Bradley, this volume; Beer and van Geel, this volume). By contrast, most authorities with a background in social sciences stress instead the degree to which poorly adapted patterns of social organization or inappropriate response strategies appear to be more direct and significant factors in social decline or collapse (e.g. Redman 1999). To a large degree, these differences reflect different academic backgrounds and training: there is a tendency to favor interpretations that lie within the scope of the disciplines with which we are familiar, although the historical rejection of determinism in the wake of its eventual discrediting during the middle decades of the 20th century may have reinforced the second type of attitude. If we take a wider and more balanced view and detach the issue from the emotive concept of "collapse", it is clearly unrealistic to suppose that climate variability has had no effect on human welfare and social viability - there are too many examples from the more recent past, for which well-documented historical evidence is available. Equally, explanations based entirely on climate determinism have been shown to be inadequate wherever interdisciplinary studies have been carried out that span biophysical and cultural/socio-economic research. The key lies in trying to unravel and learn from the interactions between the biophysical and the social systems. This poses formidable challenges.
One of the most recent attempts to present an overview is that of Diamond (2005). He examines diverse case studies of societal collapse in the light of a wide range of possible causative factors, of which climatically induced stress is only one. The others include eight broad categories of unsustainable damage to the environment, the impact of hostile neighbors, interactions with trade partners, and the nature of the societal response to environmental problems. In almost all the cases o 5
Palynology and sedimentology 2000 n data
Isotopic budget (no CO= effects)
Isotopic budget (including CO= effects)
Figure 3.11 Bahn and Flenley's (1992) conceptual model of the interactions between human activities and their environment that could have led to the dramatic
population decline on Easter Island. The shaded boxes identify processes and interactions that are less specific to Easter Island. (Based on Oldfield 2005.)
Figure 3.11 Bahn and Flenley's (1992) conceptual model of the interactions between human activities and their environment that could have led to the dramatic population decline on Easter Island. The shaded boxes identify processes and interactions that are less specific to Easter Island. (Based on Oldfield 2005.)
he considers, environmental damage and societal reactions form a constant thread. The role of climate varies from case to case, with sustained drought a key factor in many, including those considered by the more "determinist" authors referred to in the previous paragraph. In each case, however, the notion of a single, simple response leading to total demise never provides a fully satisfactory explanation. Moreover, in the most extreme case, that of Easter Island (Bahn and Flenley 1992, 2003; Figure 3.11), it seems unlikely that climate played any significant role. Although much of Bahn and Flenley's scheme is particular to Easter Island, the processes and interactions portrayed in, and linking, the shaded boxes are not without parallels in the present day, for example, in southern Sudan where climate variability imposes additional stresses.
Although it is tempting to explain the demise of Norse settlements in Greenland by climatic deterioration during the early stages of the Little Ice Age, it is clear that many other factors played an important role, including land degradation, adoption of a life-style too dependent on European contact and supplies (both of which dwindled as sea-ice in the North Atlantic made navigation more difficult), the replacement of walrus ivory by Africa elephant ivory, and failure to broaden and adapt diets in parallel with the Inuit who survived through the period of Norse abandonment (Barlow et al. 1997; Pringle 1997; Diamond 2005; Orlove 2005). Similar evidence for a complex of interactive processes, both biophysical and cultural, lies at the heart of many other analyses of cultural transformation or demise. Rosen's (1995) analysis of the collapse of early Bronze Age societies in southern Levant stresses an inability to adapt appropriate technological responses to a diminished water supply. Hassan's (1986, 1997) recent studies show that the failure of the Nile floods around 2150 bc appears to have been the main reason for the demise of the Old Kingdom, but out of the crisis that drought created came a period of radical social change of long-term importance for emerging concepts of social justice (http://www.bbc.co.uk/history/).
Severe drought still seems to be the dominant reason for the collapse of prehistoric communities in the Saharan region (Hassan 1986, 1997; Cremaschi and Di Lernia 1999), although there is no simple link between climate and cultural dynamics and one of the most significant responses to desertification appears to have been the movement of people into the Nile Valley, leading to the appearance of Neolithic sites in the Delta and central Sudan. Response via migration has also characterized some sub-Saharan societies. Tyson et al. (2002) point to southward migration by Iron Age agriculturalists following changed rainfall patterns in southern Africa, notably the southern migration of the Sotho-Tswana speaking people during the first few centuries of the past millennium and earlier. These migrations appear to be linked to the anti-phase incidence of rainfall north and south of the Equator documented by records from Lake Naivasha in Kenya and Cold Air cave in the Makapansgat Valley, South Africa. Verschuren et al. (2000) link their reconstruction of hydrologic change in Lake Naivasha to Webster's (1979) reconstruction of periods of prosperity, famine, and migration, but Robertshaw et al.'s (2004) evaluation of the records from the region once more stress multi-causation and the importance of the administrative structures in place during different periods of famine and disease.
One example of inferred collapse is that of the Mapungubwe agro-pastoralist society in the Limpopo Valley, which lasted for some 300 years before its demise around ad 1280-90. Huffman (1996) ascribes this mainly to drought and O'Connor and Kiker (2004), using a modeling approach, suggest that both crop failure and destabilization of pastoralism may have been involved. Scott and Lee-Thorp (2004), however, point out that the evidence for societal collapse is not unambiguous and the suggestion that drought may have been responsible is not borne out by the latest paleoclimatic evidence.
Two well-documented examples of societal collapse come from the New World, that of the Anasazi in the south-west of the USA and the Maya in Yucatan during the 12th and 10th centuries ad, respectively. Drought has been invoked in both cases (see e.g. Hoddell et al. 1995), but a fuller review of the evidence (Diamond 2005) reveals, in both cases, discrepancies in a simple climate-collapse hypothesis. In the case of the Maya, using numerical values expressing estimates of different types and degrees of vulnerability, both climatically and anthropogenically related,
Me-Bar and Valdez (2005) infer an 80 percent increase in vulnerability during the 9th century ad (the period of cultural demise) compared with the level of vulnerability during the late pre-Classical period some 600 years earlier.
For the Anasazi culture, severe drought may have been the "last straw" but earlier there were severe droughts that the Anasazi survived. There is also evidence of evacuation of sites before drought set in. The demise of the Anasazi is now believed to have also involved other factors such as overexploitation of local resources, overextended trade routes, internecine strife, religious/ideological conflict, and the possible pull of Kachina-based religion to the south in the areas now occupied by the related Hopi and Zuni peoples (see e.g. Kohler 1988, 1992; van West 1991). There have been several attempts to deepen understanding of the processes involved through model-based simulations that are of wider interest for their methodological implications. In Dean et al.'s (1999) study, households act as the agent units. The authors conclude that this approach allows some testing of hypotheses (or at least their reinforcement or negation in light of the degree of compatibility with the available empirical evidence), sheds light on the importance of interaction between social and environmental factors, identifies examples of equifinality, permits experimental manipulation of behavioral modes and their effects on responses to environmental variability, and makes it possible to explore previously ignored, unspecified or discounted factors. They claim only partial success in relation to their main goal, to generate target outcomes consistent with the archaeological evidence and identify rules of agent behavior that generate these. Axtell et al. (2002) claim to have come closer to achieving this goal.
There is drama in the collapse of past civilizations, for the remains - the cliff dwellings of Mesa Verde, or the great temples of the Mayan or Khmer cultures -are evocative and their desertion is inexplicable without much research. In so far as generalizations are possible, the key to their understanding lies in unraveling the complex relationships between environmental change and human cultures. Some poorly adaptive behaviors recur in many of the case studies now available - the evolution of exploitative hierarchies out of touch with the cultural and environmental consequences of their roles in society and demands of the rest of society, overexploitation of a limited resource base in response to population growth, environmental damage resulting from this, short-term responses to shortage without regard to long-term consequences, and failure to incorporate memory of previous periods of environmental stress into the repertoire of survival strategies employed. Couple a combination of these behavior types to the stresses imposed by severe drought especially, and collapse has often been the result. Climatic factors have not been implicated in every case, but there are relatively few in which they can be ruled out.
Fascinating though the consideration of past collapse is, there is surely an urgent need to balance these studies with ones devoting at least as much attention to documenting successful survival strategies during periods of climatically imposed stress, as well as to learning more from situations where irreversible damage to the resource base of a society has been avoided and long-term productivity maintained (see e.g. Crumley 2000).
The role of people in the Holocene | 81 A role for Holocene paleoresearch
We turn now to an examination of the role paleoresearch may play in addressing the issues raised in the foregoing sections. The first four themes above hinge, in large part, on the reconstruction of changes in the structure, function, and extent of past ecosystems, of the disturbance regimes to which they have been subjected, and, in the case of the third theme, of the hydrologic and erosional responses to such changes. It is my belief that, amongst other things, addressing these issues calls for a shift in emphasis from using paleobiological techniques, such as pollen or macro-fossil analysis, primarily for climate reconstruction towards using them primarily as the basis for ecological reconstruction. This would allow other evidence for past climate change (chironomids, tree rings, and stable isotopes, for example) to generate trajectories of variability independent of the evidence for ecosystem changes. This is a prerequisite of any attempt to disentangle evidence for climate change from that portraying ecosystem responses (see also Birks, this volume). It is also essential if we are to deepen our understanding of past human impacts on ecosystems. This requires further efforts to characterize and detect the imprint of human activities in the results of paleobiological research, and to link such evidence to the archaeological record.
These are also the important components of any contribution paleoscientists may make to the unraveling of problems posed by the fifth theme, in view of the frequency with which ecological degradation has been invoked as a contributory factor in societal collapse. In addition though, there is a need for a much higher degree of interdisciplinarity, spanning the biophysical and social sciences.
Achieving better empirically based quantification of feedbacks from land-cover change both to climate directly and via the impact on atmospheric greenhouse gas concentrations depends, in the first instance, on developing a more robust framework for inferring past land-cover and, in so far as possible, ecosystem structure and function from palynological data. One step in the right direction is the attempt to improve the basis for estimating landscape openness from pollen data (Gaillard et al. 1998; Sugita et al. 1999; Brostrom et al. 2004). Nielsen and Odgaard (2005) summarize a parallel approach comparing the degree of success achieved by different approaches to quantitative, pollen-based vegetation reconstruction by comparing the results for ad 1800 with cartographic evidence from the period. They claim that their Extended R-value (ERV) model captures the main trends in vegetation and land-cover, and also allows extrapolation, incorporates background pollen explicitly, and may be more applicable to large lakes than an alternative partial least-squares regression approach. There is an urgent need to extend the application of pollen-based land-cover reconstructions on a much wider spatial scale than achieved hitherto and this is currently part of the agenda of the POLLANDCAL initiative using the Europe Pollen Data Base (Gaillard, personal commmunication). A further step must be the application of similar approaches to other regions, especially in areas of long-standing human impact in Eurasia. Given the pivotal role of fire in the regeneration and disturbance histories of many ecosystems, there is also a need to integrate more reconstructions of land-cover changes with reconstructions of fire incidence. There is also a continued need for improved chronologies to establish the phasing and precise links between changes in biophysical and social systems. More studies need to go beyond the demonstration of compelling coincidences between extreme climatic events and cultural changes.
One of the crucial issues concerns the role of modeling in these areas of paleo-research. In some cases, the possible roles for modeling are relatively easy to define. Given sufficiently robust reconstructions of changes in past ecosystem structure and function, the estimation of consequences for climate and for atmospheric greenhouse gas concentrations using models derived from present-day observations of the relevant processes becomes a logical next step. Indeed, there are already examples of this type of linkage, but the empirical basis for inferring changes through time in past vegetation and land-cover is still relatively weak. The potential value of studies at the interface between pollen-based reconstructions of land-cover and models of vegetation-atmosphere interactions works in two directions. Good quality empirical reconstructions of both vegetation and regional climate could be used to test the likely reliability of models designed to explore the future impact of land-cover change on projected future climate. Equally, well-constrained models of the role of past land-cover change in driving regional climate change could be of great value in establishing the role of anthropogenic land-cover transformation in past desertification.
Given the susceptibility of human societies to changes governed by indeterminacy, what, realistically, is the role of modeling in the area of human-environment interactions described above? Several authors use agent-based modeling techniques, which are, to some degree, common to the methodologies of modelers studying past societies and those using integrated impact assessment to develop scenarios of future human-environment relationships (e.g Holman et al. 2005; Matthews 2005). Holman et al.'s approach (Figure 3.12) rests in part on the
Figure 3.12 The "Drivers-Pressure-State-Impact-Response" (DPSIR) conceptual framework used by Holman et al. (2005) in developing future regional integrated impact assessments for the UK, but equally applicable as a framework within which to study past interactions between societies and changing environmental pressures. (From Holman et al. (2005) with kind permission from Springer Science and Business Media.)
identification of "sensitivity of impacts and interactions using historical analogs, e.g. previous droughts, changes in crop subsidies, increases in housing stock". Matthews' article is a future-oriented study using multi-agent modeling applied to simulated subsistence systems. One important criterion for successful models that seek to link biogeophysical and cultural processes is the capacity to incorporate adaptive learning in the model as simulations evolve through time. One possible approach uses cellular automaton-type models (Dearing et al. 2006). Wirtz and Lemmen (2003) apply, with impressive success, a rule-based multi-agent model with adaptive capacity to simulate the spread of Neolithic agriculture. If we take the view that human-environment systems are special cases of self-organizing and emergent systems, there may be potentially applicable models developed within ecology (e.g. Breckling et al. 2005).
Was this article helpful?