Originally posted by Eladar
http://www.drroyspencer.com/2014/02/95-of-climate-models-agree-the-observations-must-be-wrong/
Eladar, can you explain the graph in your reference? The author is claiming this data proves climate models are wrong, but does not explain the basis for the conclusion, and the figure has no legend. In the data presented, the model average and the global temperature are exactly the same as projected 15 years into the future ('83-'98), and seem to mirror global fluctuations with remarkable accuracy.
After that 15-year period, the models veer off, but obviously there are many possible explanations for this. a) the models could be wrong. But the data does not support this, as if that were true, wouldn't they have been wrong prior to that point as well? b) the temp. measurements were wrong. But that does not indicate whether or not models are wrong. c) the input variables were too unpredictable based on a priori assumptions about the distant future. This is partially true [1], as predicted changes to carbon emissions and land use were not correct. d) the early models were missing important variables that changed after 1998. This is also true, as carbon sinks, land usage etc., accuracy of measurements, were not very well understood in the '80s. Modern analyses have shown they are only off by ~0.05 degrees [2,3] , which is well within a reasonable statistical range.
So I ask again: how does the author of your primary reference come to the conclusion that 95% of the models are wrong?
[1] http://www.popsci.com/is-climate-too-complex-to-model-or-predict
[2] http://www.meteo.psu.edu/holocene/public_html/Mann/articles/articles/grl53276.pdf
[3] http://oceanservice.noaa.gov/education/pd/climate/factsheets/howreliable.pdf