Climate Models: A User's Guide

TitleClimate Models: A User's Guide
Publication TypeReport
Year of Publication2007
AuthorsFarber, DA
Series TitleUC Berkeley Public Law Research Paper
Document Number1030607
Date Published2007
InstitutionUniversity of California, Berkeley
KeywordsClimate change, environmental policy, modeling

This article has two goals: providing legal and policy analysts with a basic understanding of the types of computer models that are used in studying climate change, and thinking through the uses and limitations of these models for courts and agencies. Because this article covers a good deal of fairly diverse terrain, it may be helpful to identify four key "take away" points:

1. Climate models establish a lower end estimate for global temperature impacts, but the distribution is less clearly bounded on the high side - or in simpler terms, the high-end risk may be considerable. The models are better at predicting temperature patterns than precipitation patterns, and global predictions are considerably firmer than more localized ones.

2. Economic models are much less advanced, and their conclusions should be used with caution. Unfortunately, economists are not always carefully about incorporating uncertainty into their policy recommendations.

3. Climate scientists have created a unique institutional system for assessing and improving models, going well beyond the usual system of peer review. Consequently, their conclusions should be entitled to considerable credence by courts and agencies.

4. Model predictions cannot be taken as gospel. There is considerable residual uncertainty about climate change impacts that cannot be fully quantified. The uncertainties on the whole make climate change a more serious problem rather than providing a source of comfort. The policy process should be designed with this uncertainty in mind. For instance, rather than focusing on a single cost-benefit analysis for proposed regulatory actions, the Office of Management and Budget (OMB), which oversees federal regulatory policy, might do better to require the development of standardized scenarios for agencies to use.