What was done
Noting that "confidence in the simulation and prediction skills of global climate models is a crucial precondition for formulating climate protection policies," the authors tested the scaling performance of seven state-of-the-art global climate models for two scenarios using detrended fluctuation analysis.  The first scenario considered only the effect of greenhouse gas forcing, taken from observations until 1990 and then increased at a rate of 1% per year thereafter.  The second scenario added the effect of aerosols, but with only direct sulfate forcing considered, where historical sulfate measurements were utilized until 1990 and then increased linearly after that.

What was learned
According to the authors, all the models used in their study have been shown to be capable of reproducing the current mean state of the atmosphere to various degrees of accuracy.  Furthermore, the authors note that the models have been "validated" by comparing their output to historical data, as well as via intercomparison among themselves.  This being the case, should we accept the model-based vision of CO2-induced global warming as gospel?  No way.  The results of the authors' analysis clearly revealed that in spite of how well the models can reproduce the current mean state of the atmosphere, and in spite of the belief of many that the models have been validated, the simulated records for both scenarios displayed "wide performance differences" and "fail[ed] to reproduce the universal scaling behavior of the observed records," as demonstrated by their underestimation of the long-range persistence of the atmosphere and overestimation of observed trends.

What it means
Based on their findings, the authors conclude that "anticipated global warming is also overestimated by the models."  The moral of this review, therefore, is don't be fooled into believing that just because models can simulate mean climate fairly well, or because they all produce the same results, they can correctly predict the future.  It just doesn't work that way.  In fact, as we learn from this study, all current models still have a long way to go before we can put sufficient faith in them to formulate enlightened energy policies.