Meanwhile
http://www.nature.com/news/climate-change-the-case-of-the-missing-heat-1.14525
"Stark contrast
On a chart of global atmospheric temperatures, the hiatus stands in stark contrast to the rapid warming of the two decades that preceded it. Simulations conducted in advance of the 201314 assessment from the Intergovernmental Panel on Climate Change (IPCC) suggest that the warming should have continued at an average rate of 0.21 °C per decade from 1998 to 2012. Instead, the observed warming during that period was just 0.04 °C per decade, as measured by the UK Met Office in Exeter and the Climatic Research Unit at the University of East Anglia in Norwich, UK.
The simplest explanation for both the hiatus and the discrepancy in the models is natural variability. Much like the swings between warm and cold in day-to-day weather, chaotic climate fluctuations can knock global temperatures up or down from year to year and decade to decade. Records of past climate show some long-lasting global heatwaves and cold snaps, and climate models suggest that either of these can occur as the world warms under the influence of greenhouse gases."
So chaotic climate fluctuations can knock global temperatures up or down from year to year and decade to decade.
"But none of the climate simulations carried out for the IPCC produced this particular hiatus at this particular time. That has led sceptics and some scientists to the controversial conclusion that the models might be overestimating the effect of greenhouse gases, and that future warming might not be as strong as is feared. Others say that this conclusion goes against the long-term temperature trends, as well as palaeoclimate data that are used to extend the temperature record far into the past. And many researchers caution against evaluating models on the basis of a relatively short-term blip in the climate. If you are interested in global climate change, your main focus ought to be on timescales of 50 to 100 years, says Susan Solomon, a climate scientist at the Massachusetts Institute of Technology in Cambridge."
10-20 years is a short trend when there is no warming or there is cooling, but when there is warming is a clear signal.