Computational, multiscenario simulation is a new analytic approach to the assessment of climate change policy. Bankes (1993), Lempert et al. (1996), and Laitner and Hogan (2000) have employed this approach, as have Morgan and Dowlatabadi (1996), van Asselt and Rotmans (1997), and, to some extent, Yohe (1996). Also, the IPCC Special Report on Emissions Scenarios (IPCC, 2000b) presented a large set of very different baseline scenarios. The basic idea is to use computer simulation models to construct a range of a large number of fundamentally different scenarios of the future and, instead of aggregating the results using a probabilistic weighting, make policy arguments from comparisons of fundamentally different, alternative cases. These methods are most useful under conditions of deep uncertainty. For example, when we do not have reliable information or widespread agreement among the stakeholders about the system model, the prior probability distributions on the parameters of the system model, and/or the loss function to use in evaluating alternative outcomes (Lempert and Schlesinger, 2000).
These multiscenario simulation approaches offer the promise of a powerful synthesis between the narrative, process-oriented methods of scenario-based planning (Schwartz, 1996; van der Heijden, 1996) and quantitative tools such as decision analysis, game theory, and portfolio analysis. From the quantitative methods, multiscenario simulation draws systematic methods of handling large quantities of data and normative descriptions of good decisions. From scenario-planning, multiscenario simulation draws the insight that multiple views of the future are crucial to allow groups to transmit and receive information about highly uncertain futures. Also scenario planning shows that groups can often agree on actions to take in the face of deep uncertainty without agreeing on the reasons for these actions (Lempert and Schlesinger, 2000). For instance, multiscenario simulation can adopt a meaningful costbenefit framework for climate change, but at the same time acknowledge the deep uncertainty and differing values among stakeholders. These make it impossible to fully quantify the costs and benefits or to assign widely accepted probabilities to many of the key outcomes of interest. Such computational, multiscenario simulations are enabled by new computer technologyprimarily large quantities of inexpensive memory; fast, networked processors; and powerful visualization toolsand are only just becoming available.
Other reports in this collection