To more accurately simulate global temperature rise, scientists at the French National Center for Scientific Research, University of Southampton and Royal Netherlands Meteorological Institute developed a new prediction model.
The model operates differently than most climate models, using advanced statistical analysis to survey 20th and 21st century climate simulations. The model identifies analogs of current climate conditions and uses them to predict future temperatures.
Many models require the power of supercomputers and take weeks to run simulations. The new model runs on a laptop and spits out results in a fraction of a second. And the results are comparable to the predictions of more robust models.
In tests, the new model accurately simulated the global warming hiatus.
Unlike other global warming models, which focus on the effects of manmade carbon emissions, the new simulations account for the probability of weather phenomena absent human impacts — what scientists call “the internal variability of the climate.”
This internal variability can “temporarily reinforce the long-term global warming trend,” researchers wrote.
Due the low probability of extreme cold events, the new simulations predict higher than normal temperatures of the next five years.
The simulations also showed the next half-decade is more likely than not to feature extreme sea surface temperatures, increasing the likelihood of tropical storm activity.
Researchers detailed the new predictions and their modeling techniques in the journal Nature Communications.
Traditionally, climate science research is the domain of governments and large academic centers, but the latest findings could democratize climate modeling.
“The important numerical efficiency of the method — a few hundredths of a second on a laptop — opens the possibility for real-time probabilistic predictions carried out on personal mobile devices,” researchers wrote in their new paper.
