Image from www.despair.com (click to buy)The failure of the GCMs to reproduce real world temperature trends is becoming quite awe inspiring. In a post at WUWT, Michaels and Knappenburger have compared actual temperature trends over periods of different lengths (but all ending in the present - either 2012 or 2013) with the output of the IPCC corpus of climate models. They find that trends as long as 27 years are inconsistent for 2012 period ends. If you bring the data bang up to date it's even worse
For data ending in the year 2013, the category of marginal inconsistency extends out to 37 years and is now flirting with lengths exceeding 50 years, and trends of lengths 11-28, 31, 33, and 34 (!) are clearly inconsistent with the climate model simulations.
Some people say you need a fifteen year trend to falsify the models, some say seventeen, others say you need thirty. It's beginning to look as if the day when all excuses fall by the wayside is fast approaching.