Variations in the possible geographical impact of global warming have been highlighted by a leading climate scientist.
Computer models suggest there may be many different combinations of warming regions and cooling regions for any given assumed rate of increase in the global average surface temperature, according to Ed Hawkins of the University of Reading.
Furthermore, these computer models show that different combinations of warming and cooling regions can give rise to very similar periods of slower warming or even pauses, Hawkins told reportingclimatescience.com.
Hawkins refers to the collection of possible scenarios that can give rise to an apparent slowdown or pause in the rate of increase in global temperatures as a “zoo” of short-term trends.
“We’ve done a large number of simulations with the same model and the same assumed emissions and lots of them show pauses and slowdowns at different times and that is to be expected but they also show spatial patterns that are quite different,” Hawkins explained. “It is the same for warming periods. Simulations may be identical in terms of warming at the global level but they are very different spatially.”
Hawkins and colleagues have used the FAMOUS GCM computer model to perform a large ensemble of simulations with a scenario that assumes an increase in atmospheric carbon dioxide levels (CO2) at 1 per cent per year. In these ensembles of simulations it is possible to find many periods of 15-years with an identical global mean temperature trend, Hawkins says.
This has implications for those attempting to forecast how global warming might play out at a regional level and affect particular parts of the world. “People don’t experience the global mean temperature, they experience their local climate,” said Hawkins. “If we knew global temperature was going to rise by 0.2 degrees per decade over the next 15 years we would still not be sure where warming was going to happen. Not everywhere would warm at the same time and at the same rate,” he explained.
Warming has been slower over the past 15 years or so and that has been highlighted in certain regions and the Pacific, parts of Russia and Siberia whereas other parts of the world have experienced warming, Hawkins said.
Investigations into the recent observed slower rate of global warming have largely been focussed on variability in the Pacific basin and climate models have also been able to show similar slowdowns focussed in the Pacific. But this need not be the case, according to this new research, which will form the basis of a presentation to the Royal Society next month.
The geographical variations that can emerge with the same overall trend in global average temperatures is illustrated by the graphics on the right. The top graphic shows four different regional temperature patterns that can arise from a scenario with a global mean temperature trend of 0.000K/decade – a true pause. The lower graphic shows four different combinations of regional effects in a scenario with a global trend of 0.200K/decade. The different outcomes have vastly different patterns even though the global trend is identical.
Hawkins points out that this experiment has been done with one climate model but that the diversity of trends it illustrates is also seen in other computer models.
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.
This work also has implications for the pause. Hawkins said that periods of a slowdown in the rate of increase in global temperatures, similar to the current so called pause or hiatus in global warming, often emerge within climate model simulations. This work shows that changes in different parts of the world – such as the Atlantic or the Pacific – can also give rise to slowdowns.
When asked at what point the current slowdown would start to vary from the many possible scenarios that computer models can generate, Hawkins told reportingclimatescience.com that “if slower rates went on for another 10 years or so then that would be interesting”.