Investigations into the recent observed slower rate of global warming have largely been focussed on variability in the Pacific basin. Climate models also show similar slowdowns focussed in the Pacific (e.g. Meehl et al. 2011).
But, is this the only type of simulated slowdown? How different can regional patterns of temperature change be for the same global change?
We have used the FAMOUS GCM to perform a large ensemble of simulations with increasing CO2 at 1%/year. In these ensembles it is possible to find many periods of 15-years with an identical global mean temperature trend.
Fig. 1 shows the regional temperature trends for four different periods, each with a global mean temperature trend of 0.000K/decade. Figs. 2 & 3 are the same but for a global trend of 0.100K/decade & 0.200K/decade. The different periods have vastly different patterns, almost anti-correlated in some cases, even though the global trend is identical.
Of course, FAMOUS is only one GCM, which is rather simple by modern standards and has larger-than-average variability, but this diversity of trends is also seen in other GCMs (e.g. Deser et al. 2012).
This type of analysis also highlights the value of large ensembles of climate projections for quantifying near-term changes in climate. If you only had a handful of such simulated periods, you would not effectively sample the full range of plausible near-term trends, even if the precise global trend was known.
Welcome to the zoo of short-term trends!