On the Number of Effective Climate Models

1 Projections of future climate change are increasingly based on the output of many 2 different models. Typically, the mean over all model simulations is considered as the 3 optimal prediction, with the underlying assumption that different models provide 4 statistically independent information evenly distributed around the true state. However, 5 there is reason to believe that this is not the best assumption. Coupled models are of 6 comparable complexity and are constructed in similar ways. Some models share parts of 7 the same code and some models are even developed at the same center. This contributes 8 to the well-known problem of common model biases and possibly to an unrealistically 9 small spread in the outcomes of model predictions. 10 This study attempts to quantify the extent of this problem by asking how many 11 models there effectively are and how to best determine this number. Quantifying the 12 number of effective models is achieved by evaluating 24 state-of-the-art models and their 13 ability to simulate broad aspects of 20 century climate. Using two different approaches, 14 we calculate the amount of unique information in the ensemble and find that the effective 15 ensemble size is much smaller than the actual number of models. As more models are 16 included in an ensemble the amount of new information diminishes in proportion. We 17 further find that this reduction goes beyond the problem of “same-center” models and 18 that systemic similarities exist across all models. Our results suggest that current 19 methodologies for the interpretation of multi-model ensembles may lead to overly 20 confident climate predictions. 21 22 23