A probability and decision-model analysis of a multimodel ensemble of climate change simulations

Because of the inherent uncertainties in the computational representation o f climate and because of unforced chaotic climate variability, it is argued that climate change projections should be expressed in probabilistic form. In this paper, 17 Coupled Model Intercomparison Project second-phase exper iments sharing the same gradual increase in atmospheric CO2 are treated as a probabilistic multimodel ensemble projection of future climate. Tools com monly used for evaluation of probabilistic weather and seasonal forecasts a re applied to this climate change ensemble. The probabilities of some tempe rature- and precipitation-related events defined for 20-yr seasonal means o f climate are first studied. A cross-verification exercise is then used to obtain an upper estimate of the quality of these probability forecasts in t erms of Brier skill scores, reliability diagrams, and potential economic va lue. Skill and value estimates are consistently higher for temperature- rel ated events (e.g., will the 20-yr period around the doubling of CO2 be at l east 1 degreesC warmer than the present?) than for precipitation-related ev ents (e.g., will the mean precipitation decrease by 10% or more?). For larg e enough CO2 forcing, however, probabilistic projections of precipitation-r elated events also exhibit substantial potential economic value for a range of cost-loss ratios. The treatment of climate change information in a prob abilistic rather than deterministic manner (e.g., using the ensemble consen sus forecast) can greatly enhance its potential value.