The provision of useful guidance to inform policy requires observation and description of human contributions to global change, as well as theoretical studies of the underlying social processes that shape them. We also need to understand how global change affects human welfare. This requires not merely studies of direct exposure but also of the capacity to respond.

Causal models of social processes have large uncertainties, and pose problems that are of a qualitatively different character than those encountered in modelling non-human components of the Earth system. This is due, first and foremost, to the inherent reflexivity of human behaviour; i.e., the fact that human beings have intellectual capabilities and emotional endowments enabling them to invent new solutions and transcend established �laws� in ways that no other species can do. As a consequence, predictive models may well alter the behaviour that they seek to predict and explain � indeed, such models are sometimes deliberately used exactly for that purpose. Moreover, the diversity of societies, cultures, and political and economic systems often frustrates attempts to generalise findings and propositions from one setting to another. Representation of human behaviour at the micro (individual) and macro (collective) scale may require fundamentally different approaches (see Gibson et al., 1998).

These kinds of difficulties intrinsically limit the predictive power that can be ascribed to models of social processes. As a consequence, research on human drivers and responses to climate change cannot be expected to produce conventional predictions beyond a very short time horizon. This does not imply, however, that research on human behaviour and social processes cannot provide knowledge and insight that can inform policy deliberations. A considerable amount of basic knowledge and insight exist, and this knowledge can be used, inter alia, for constructing scenarios showing plausible trajectories and identifying the critical factors that will have to be targeted in order to switch from one trajectory to another. From the perspective of policy-makers, this can indeed be an important contribution.

To make the most of this potential, further progress is required along two main frontiers. One challenge is to develop a more integrated understanding of social systems and human behaviour. With some exceptions, the first generation of models in this area represented �the human system� by a few key variables. For example, resource use was often conceived of as a function of population size and income level. The performance of such simplistic models was by-and-large poor. It is abundantly clear that the impact of human activities as drivers of climate change depends upon a complex set of interrelated factors, including also technologies in use, social institutions, and individual beliefs, attitudes, and values. At present, it seems fair to say that we have a reasonably good theoretical grasp of important types of institutions, such as markets and hierarchies, in ideal-type form. What we need to understand better is how their impure real-world counterparts work, and to improve our understanding of the intricate interplay of institutional complexes, i.e., how markets, governments and other social institutions interact to shape human behaviour. Research in political economy clearly indicates that phenomena such as economic growth are to a significant extent affected by the functioning of interlocking networks of institutional arrangements.

Similarly, we have a fairly good grasp on particular kinds of intellectual processes - in particular, the logic of rational choice - but we are doing less well when it comes to understanding how beliefs, attitudes and values change and how change in these factors in turn affects manifest human behaviour, such as consumption patterns. To address these challenges we need more interdisciplinary research designed to integrate knowledge from different fields and sub-fields into a more holistic understanding of �the human system�. The intellectual and organisational problems involved should not be underestimated, but we are confident that the prospects for making progress along this frontier are better now than ever before.

The other main challenge is to find better ways of integrating models of the biogeophysical Earth system with models of social systems and human behaviour. Some encouraging progress has been made at this interface, particularly over the last decade. For example, there has been a rapid increase in attempts to integrate representations of human activities in models with explicit formal linkages to other components of the Earth system. Such integrated assessment models have offered preliminary characterisations of human-climate linkages, particularly through models of multiple linked human and climate stresses on land cover. Moreover, they have provided preliminary characterisation of broad classes of policy responses, and have been employed to characterise and prioritise policy-relevant uncertainties.

Yet, effective integration is frustrated by at least two main obstacles. One is incongruity of temporal and spatial scales. Social science research cannot match the long time horizons of much natural science research. On the other hand, in studying consequences for human welfare and responses to these consequences, social scientists need estimates of biophysical impacts of climate change differentiated by political units or even smaller social systems. Aggregate global-scale estimates are of limited use in this context; human sensitivity to climate change varies significantly across regions and social groups, and so does response capacity. We can expect to see some progress in alleviating the spatial resolution problem, as regional-scale models of climate change are further developed, but we have to recognise that the scale problems are fundamental and that no quick fixes are in sight. The other problem pertains to the interface between different methodological approaches. In particular, concerted efforts are required to develop better tools for coupling approaches relying on numerical modelling with �softer� approaches using interpretative frameworks and qualitative methods. Some of these differences are too profound to be eliminated, but that does not imply that bridges cannot be built. Learning how to work more effectively across these methodological divides is essential to the further development of integrated global change research. Again, some encouraging progress is being made.

IPCC Homepage