The LMD/IPSL has recently started a collaboration with economists within the French CIRED (Centre International de Recherche sur l'Environnement et le Developpement - a CNRS laboratory), and in particular with its director, J.C. Hourcade, who has been a lead author in several chapters of the IPCC Report (Group III). This collaboration aims at evaluating the impact of climate change, and the strategies to mitigate it. Several impact models will be considered within this collaboration, ranging from simple models (e.g., DIAM) to relatively detailed ones (e.g., IMAGE). All of these "integrated models" attempt to link physical and economical aspects of global climate changes. A common challenge for them is to incorporate the uncertainties of climate model scenarios into their climate change damage estimates.
Uncertainties in climate model simulations are related to both model deficiencies and fundamental problems (partial unpredictability of the system). The spread between the different climate model simulations is arguably one of the best measures currently available to estimate uncertainties in climate projections. The CMIP data base offers a unique opportunity to document at least part of those uncertainties, because similar experiments have been made with a wide range of models.
Further study and development of integrated models is clearly needed. In this study we seek to evaluate a variety of integrated model damage estimates over selected regions, making use of the CMIP data base to estimate uncertainties in climate simulation scenarios.
The CMIP data base will be used as follows: Monthly mean precipitation and surface temperature will be analyzed from both 1xCO2 and 2xCO2 scenarios. The mean response (average over the different models), the range of the response (the standard deviation across models), and the extremes will be evaluated. These measures (not the individual simulations) will be used for our study of a suite of impact models.
CMIP data requirements: Monthly mean precipitation and surface air temperature.