LAUSR

The global warming potential (GWP) is widely used in policy analysis, national greenhouse gas (GHG) accounting, and technology life cycle assessment (LCA) to compare the impact of non-CO 2 GHG emissions to the impact of CO 2 emissions. While the GWP is simple and versatile, different views about the appropriate choice of time horizon—and the factors that affect that choice—can impede decision-making. If the GWP is viewed as an approximation to a climate metric that more directly measures economic impact—the global damage potential (GDP)—then the time horizon may be viewed as a proxy for the discount rate. However, the validity of this equivalence rests on the theoretical basis used to equate the two metrics. In this paper, we develop a new theoretical basis for relating the GWP time horizon and the economic discount rate that avoids the most restrictive assumptions of prior studies, such as an assumed linear relationship between economic damages and temperature. We validate this approach with an extensive set of numerical experiments using an up-to-date climate emulator that represents state-dependent climate-carbon cycle feedbacks. The numerical results largely confirm the theoretical finding that, under certain reasonable assumptions, time horizons in the GWP of 100 years and 20 years are most consistent with discount rates of approximately 3% and 7% (or greater), respectively.