impacts of emissions reductions. Therefore, he believes it is essential to make value judgments—how to value nonmarket goods ranging from environmental concerns to war and conflict. There is no correct value judgment, and there will be different views on this.
Nordhaus outlined four major catastrophic risks that have emerged over the years in the literature: the reverse of the North Atlantic deep water circulation, melting of large ice sheets, abrupt climate change, and ocean carbonization. He warned that unmanaged ecosystems, such as the polar ice caps, are likely to be drastically impacted. He described a scenario in which the Greenland ice sheet for example, could reach a tipping point, meaning that consequences would potentially be irreversible (Figure 4). These types of issues will require better integration of geophysical and economic modeling. Joel Smith agreed that damage to unmanaged ecosystems could potentially be the most significant consequence of climate change. He advised caution when attempting to assign value to occurrences like species or ecosystem loss—applying existing tools and frameworks may not be appropriate for challenges of this magnitude. He also mentioned the ongoing inquiry into the potential effects of sea surface warming on hurricane formation—one study pointed to $10 billion to $20 billion a year in additional damage by 2080 (ABI, 2005). In an analysis of costs for storm water systems in a city in Honduras, Joel and colleagues found that, under an assumption of a reasonably high level of change in intensity and amount of rainfall by 2025, infrastructure costs could increase by as much as 30 percent.
John Weyant explained that policy formulation indeed has a large effect on modeling results, and cost projections vary depending on the policy regime employed. Experiments in this area initially led Howard Gruenspecht to coin the term “where and when flexibility,” which refers to allowing for interregional and intertemporal trading. Adding to this uncertainty is the fact that cost projections are very dependent on policy assumptions: even when policy formulation is controlled for and the range of modeling results is thus narrowed, cost projections still assume that policies will be adopted and implemented in a rational and efficient manner. While analyzing options for California’s renewable portfolio standard (RPS) of 33 percent, Weyant and the rest of the Technical Advisory Committee stated