Michael Oppenheimer

We disagree with Solomon eta . that our Policy Forum was misleading. We correctly noted that model-based numerical ranges for century sea level rise presented in the Working Group I (WIG) "Summary for Policymakers" (SPAM) (Table SPAM-3) did not account for the uncertainty resulting from potential increases in the rapid dynamic response of ice sheets. Solomon a AL. challenge this assertion by pointing instead to qualitative statements in the SPAM, implying that the latter provide a satisfactory accounting of uncertainty. But the distinction between numerical values highlighted in a prominent table and narrative qualifications of such numbers is critically important. Numbers are powerful, grabbing the readers' attention, whereas qualifications are often ignored. For example, the tabular values, indicating a maximum sea level rise of 59 cm during the 21st century, are cited frequently in the public discussion absent any qualification. We did not imply, as Solomon ET AL. argue, that the WEI SPAM omitted information from paleolithic studies in evaluating uncertainty in sea level rise beyond the 21st century. We suggested that it gave too much credence to ice sheet health articles models compared  For example, in reporting only a model-based estimate for the time scale of a long-term contribution (from Greenland), the WIG SPAM gives short shrift to the implications of observations of fast responses in the ice sheets in both Greenland and western Antarctica, narrative qualifications to the contrary notwithstanding. Such an approach understates the range of opinion in the relevant expert community on the potential magnitude and rate of the ice sheet contribution as indicated by studies reviewed during the Fourth Assessment Report (ARE) (i). Further perspective on this question is provided by the ARE Synthesis Report (2). Finally. contrary to Solomon ct Al.'s assertion, our suggestions for improving the treatment of uncertainty were made specifically with the shortcomings of ice sheet modeling in mind. It makes little sense to highlight model-based projections of sea level rise when models that are supposed to account for the ice sheet component have failed the test against reality. Other approaches provide important additional perspectives. For example, the fact that two independent semi-empirical analyses estimating uncertainties in future sea level rise have been published recently (3, 4) suggests that observation-based methods yield important insights where models are deficient. We do not propose that IPECAC conduct new research: Rather, we argue that it take full advantage of what has already been produced. IPECAC also has the flexibility to fill gaps in modeling and analysis where the completeness of assessment calls for it, and it has done so many times. In anticipation of a Fifth Assessment. and realizing that ice sheet models may not improve rapidly, IPECAC should encourage the development of more comprehensive approach to uncertainty. As it has done for other arenas, such as emissions scenarios or abrupt climate change, I P CC could spur research into empirical approaches, formalized expert elicitation, and comprehensive analysis of Paleocene extent and sea level, each carried out with a specific view toward informing quantitative judgments on the range of future sea level. Holding work-shops on this problem over the next few years would fit neatly into IPECAC tradition. Three of us are authors of ARE, well aware of the difficulty of assessment. A premise of our Policy Forum is that I P CC has done a superb job of establishing the scientific consensus. But in a high-stakes problem like global warming. governments need to calibrate policy to the full range of plausible outcomes, for sea level rise and for all other key uncertainties.