HR: 1340h
AN: GC13B-1236    [Abstracts]
TI: A Performance Index for the Evaluation of Coupled Climate Models
AU: * Kim, J
EM: jkim@met.utah.edu
AF: University of Utah, 135 S 1460 E, Rm. 819 (WBB), Salt Lake City, ut 84112 United States
AU: Reichler, T
EM: thomas.reichler@utah.edu
AF: University of Utah, 135 S 1460 E, Rm. 819 (WBB), Salt Lake City, ut 84112 United States
AB: In a recent coordinated effort 21 climate modeling groups from around the world produced a coupled multi-model dataset in support of the 4th IPCC assessment report. The question addressed in this paper is how the latest model improvements and the full range of external forcings considered in the simulations affect the ability of such models to match the observations of present-day climate. The answer to this question is important to give credible estimates for future climate change. In order to evaluate the model performances, we compare various key-climatological quantities with observational data and determine normalized root-mean-square errors. In order to facilitate the comparison between the climate models, we aggregate the individual model errors into a single index, which is a good measure for the overall model performances, and which provides new insight into the present state of climate modeling. We analyze the geographical error distribution for the individual models and for the multi-model super-ensemble. Further, we investigate the interannual variability and the ENSO relationships between model results and observations. The errors in simulating the climate over the past two decades show large model-to-model differences. The large systematic differences raise the question whether the current practice of the IPCC, which includes the whole range of models in projections of future climate change, is still appropriate.
DE: 0545 Modeling (4255)
DE: 0550 Model verification and validation
DE: 1620 Climate dynamics (0429, 3309)
DE: 1626 Global climate models (3337, 4928)
SC: Global Climate Change [GC]
MN: Fall Meeting 2005