challenging. Incorporating a more integrated approach into some of CCSP’s assessment activities could provide an important opportunity to learn more about how to conduct more effective integrated assessments, while also producing integrated, societally relevant outcomes.

Recommendation: The CCSP should invest in experimental applications of integrated assessments, with a specific focus on advising future applications of truly integrated, ongoing, interdisciplinary assessments in the United States.


One of the most difficult tasks in an assessment is the expression of uncertainty. In the case of climate change, uncertainties remain as part of the evolving understanding of various aspects of the greenhouse effect, its likely consequences, and the efficacy of various countermeasures. Some of these uncertainties will not be resolved for decades, if then. An effective characterization of uncertainty in assessments requires determining what kinds of uncertainty information would be useful for decision makers as well as developing quantitative or qualitative measures of uncertainty. While there is evidence that decision makers have an aversion to ambiguity, uncertainty is unavoidable in many decision-making contexts. Once decision makers understand that they are operating in an uncertain environment, they typically prefer that the conclusions of an assessment be accompanied by a description of the level and source of relevant uncertainties. The manner in which uncertainties are acknowledged and characterized will affect both the salience and the credibility of the assessment.

Ways of addressing uncertainty include clearly identifying the uncertainties; characterizing and identifying the source and magnitude of the uncertainties; expert judgments of the level of confidence; and testing this sensitivity through the development of plausible future scenarios. For example, the IPCC has attempted to deal with uncertainty by using words to indicate judgment estimates of confidence (e.g., “virtually certain” denotes a greater than 99 percent chance that a result is true, “very likely” denotes a 90-99 percent chance, etc.). Alternatively, there are formal methods for eliciting expert judgments (Morgan and Henrion 1990), but these can be quite time consuming and, therefore, can only be applied selectively.

Recommendation: Uncertainties should be well articulated in global change assessments to the extent they are understood, and the sources of the uncertainty should be described. There should be a deliberate effort to clarify the importance of alternative assumptions and to illustrate the impacts of uncertainties.

The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement