Policymakers require a coherent synthesis of all aspects of climate change. Researchers have spent the past decade developing integrated assessment methods to meet these needs of policymakers. An overview of the framework, including examination of impact and vulnerability, is in the SAR (Weyant et al., 1996). In addition, Rotmans and Dowlatabadi (1998) have concentrated on the broader social science components of integrated assessment; as a result, they came closer to presenting a view within which impacts and adaptation might be most fully investigated with and without relying on models. They assert, "Integrated assessment is an interdisciplinary process of combining, interpreting, and communicating knowledge from diverse scientific disciplines in such a way that the whole set of cause-effect interactions of a problem can be evaluated." Current integrated assessment efforts generally adopt one or more of four distinct methodological approaches:

• Computer-aided modeling in which interrelationships and feedbacks are mathematically represented, sometimes with uncertainties incorporated explicitly (see Chapter 19)
• Scenario analyses that work within representations of how the future might unfold [the MINK study, based on a climate analog of the dust bowl climate of the 1930s, is a classic example (see Rosenberg, 1993, for details)]
• Simulation gaming and participatory integrated assessment, including policy (see Parson and Ward, 1998, for a careful review)
• Qualitative assessments that are based on limited and heterogeneous data and built from existing experience and expertise. Cebon et al. (1998) contains a collection of papers that offer similar qualitative coverage; their insights can serve as the basis for a long-run research agenda that looks for regions and sectors in which uncertain futures most significantly cloud our view of where and when impacts might be most severe.

Schneider (1997) has developed a taxonomy of integrated assessments that creates an historically rooted taxonomy of modeling approaches. It begins with "premethodogical assessments" that worked with deterministic climate change, with direct causal links and without feedbacks. It ends with "fifth-generation" assessments that try to include changing values explicitly. In between are three other stages of development, differentiated in large measure by the degree to which they integrate disaggregated climate impacts, subjective human responses, and endogenous policy and institutional evolution.

Methodological bias is an issue in interpreting the results of integrated assessments, as it is in every research endeavor. Schneider (1997) also warns that models composed of many submodules adopted from a wide range of disciplines are particularly vulnerable to misinterpretation and misrepresentation. He underscores the need for validation protocols and explorations of predictability limits. At the very least, integrated assessments must record their underlying value-laden assumptions as transparently as possible. Including decisionmakers and other citizens early in the development of an assessment project can play an essential role in analytical processes designed to produce quality science and facilitate appropriate incorporation of their results into downstream decisions.

In the past decade, several research teams have been working on the development of such frameworks (see Tol and Fankhauser, 1998, for a compendium of current approaches). Known as integrated assessment models (IAMs), these frameworks have been used to evaluate a variety of issues related to climate policy. Although the current generation of IAMs vary greatly, in scope and in level of detail, they all attempt to incorporate key human and natural processes required for climate change policy analysis. More specifically, a full-scale IAM includes submodels for simulating:

• Activities that give rise to GHG emissions
• The carbon cycle and other processes that determine atmospheric GHG concentrations
• Climate system responses to changes in atmospheric GHG concentrations
• Environmental and economic system responses to changes in key climate-related variables.

Although IAMs provide an alternative approach to impact assessment, it is important to note that there is no competition between such integrated approaches and the more detailed sectoral and country case studies discussed in preceding sections. Each approach has its strengths and weaknesses and its comparative advantage in answering certain types of questions. In addition, there are considerable synergies between the two types of studies. Integrated approaches depend on more disaggregated efforts for specification and estimation of aggregate functions and, as such, can be only as good as the disaggregated efforts. Reduced-form integrated approaches make it relatively easy to change assumptions on the "causal chain." That is, one can identify critical assumptions upon which a policy analysis might turn.

In conducting such sensitivity analyses, one can identify where the value of information is highest and where additional research may have the highest payoff from a policy perspective. This can provide some useful guidance to the impacts community about where to direct their efforts to resolve uncertainty. At the same time, integrated models become more useful as uncertainty is narrowed (through the contributions of partial impact assessments); hence, the reduced-form representations become more realistic

IPCC Homepage