Now that 2012 is over, it is time to update a comparison of simulations and observations of global mean temperatures.
[UPDATE (17/03/13): David Rose has written an article in the Mail on Sunday which, by eye, seems to use the top left panel from the figure below, but without mention of its original source. In the article David Rose suggests that this figure proves that the forecasts are wrong. This is incorrect - the last decade is interesting and I have discussed these issues previously (as have many others) and I have even co-authored a published article about the most sensitive simulations being less likely. David also incorrectly suggests that the shaded ranges shown are 75% and 95% certainty. As labelled below, they are actually the 25-75% and 5-95% ranges, so 50% and 90% certainty respectively.]
[UPDATE (21/03/13): Mail on Sunday acknowledge that Climate Lab Book figure was used & redrawn and apologise for lack of credit in article update.]
[UPDATE (16/04/13): The Economist also used a version of the figure below in an article on climate sensitivity.]
[UPDATE (17/04/13): Two new published papers suggest that deep ocean warming is the reason why the surface temperature has not warmed so rapidly in the last 15 years: Balmaseda et al. & Guemas et al.]
[UPDATE (20/05/13): The most recent appearance of this now infamous graph was on CNBC, wrongly attributed to 'Mail Online'.]
[UPDATE (29/05/13): A complete update to this post is now available.]
Previously I compared the CMIP5 simulations (performed over the past couple of years) and observations attempting to consider the effects of the relative availability of observations. Now that we have some new observations (HadCRUT4 up to 2012) an update can be made. (Note that RealClimate have also done something similar).
The figure below (top row) shows the comparison of observations (HadCRUT4 & GISS) with the CMIP5 simulations, but masking the simulations with the HadCRUT4 observational mask for that comparison. The bottom row shows the effect of the masking on the simulated temperatures, which continues to grow into the future (assuming the current observational mask).
MASKING UPDATE: Paul Matthews suggested that I needed to explain masking a bit more. HadCRUT4 only uses available observations of temperature, and there are many missing regions. Therefore the comparison is only done using model data where observations exist that particular month (see here for more details). GISS observations however are interpolated and fill in the observational gaps to create a complete dataset which does not need masking.