In support of the U.S. National Assessment of the Potential Consequences of Climate Variability and Change, climate scenarios were prepared to serve as the basis for evaluating the vulnerability of environmental and societal systems to changes projected for the twenty-first century. Since publication of the results of the assessment at the end of 2000, the National Research Council's report Climate Change Science: An Analysis of Some Key Questions, and the U.S. government's U.S. Climate Action Report—2002 have both relied on the assessment's findings. Because of the importance of these findings, it is important to directly address questions regarding the representativeness and usefulness of the model-based projections on which the findings were based. In particular, criticisms have focused on whether the climate models that were relied upon adequately represented twentieth-century conditions and whether their projections of conditions for the twenty-first century were outliers. Reexamination of the approach used in developing and evaluating the climate scenarios indicates that the results from the two primary climate modeling groups that were relied upon allowed the generation of climate scenarios that span much of the range of possible future climatic conditions projected by the larger set of model simulations, which was compiled for the IPCCs Third Assessment Report. With the set of models showing increasing agreement in their simulations of twentieth-century trends in climate and of projected changes in climate on subcontinental to continental scales, the climate scenarios that were generated seem likely to provide a plausible representation of the types of climatic conditions that could be experienced during the twenty-first century. Warming, reduced snow cover, and more intense heavy precipitation events were projected by all models, suggesting such changes are quite likely. However, significant differences remain in the projection of changes in precipitation and of the regional departures in climate from the larger-scale patterns. For this reason, evaluating potential impacts using climate scenarios based on models exhibiting different regional responses is a necessary step to ensuring a representative analysis. Utilizing an even more encompassing set of scenarios in the future could help move from mainly qualitative toward more certain and quantitative conclusions.
Lawrence Livermore National Laboratory, Livermore, California
The Pennsylvania State University, University Park, Pennsylvania
NOAA National Climate Data Center, Asheville, North Carolina
Marine Biological Laboratory, Woods Hole, Massachusetts