For more than a decade, climate change doubters have pointed to the last 15 years of slowed global warming as proof it’s a myth and that climate modelling “overestimated” warming and thus was wrong.
But it turns out the modelling is spot on, a major review, published today in Nature Climate Change, has found.
The problem was that unlike the modelling, critics did not factor in periods when the climate system was El Nino-dominant (when the Pacific Ocean surface temperature rises), as opposed to La Nina-dominant (when it cools).
Lead author Dr James Risbey, a synoptic climatologist at the CSIRO, led a team of six who reviewed climate data in 15-year blocks between 1880 and 2012 and then compared it with the modelling. During El Nino-dominant periods, which can last up to 30 years, the modelling “underestimated” warming, just as it’s “overestimating” during the current “La Nina” period.
But when Dr Risbey and the team overlaid and averaged out a number of climate prediction models – there are around 40 used to contribute to main international archive – he found the fluctuations caused by the Nino/Nina effect averaged out. While 15-year temperature trends speed up and slow down in response to natural climate variations, based on the climate’s preference for El Nino- or La Nina-dominance, a true picture of the warming trend emerges in line with predictions.
There’s nothing unique about the last 15 years he says, it’s just part of a longer-term pattern.
“In short, climate models have provided good estimates of recent and past trends,” he says.
Dr Risbey said the misinterpretation used by the sceptics hinged on confusion between climate projections and climate forecasts.
“Climate forecasts attempt to track the sequence of changes in internal climate variations like El Nino and La Nina. In contrast with a climate forecast, climate projections have no information about the phase (sequence and timing) of internal climate variability,” he said.
“These natural cycles are important if you’re going to get the right trend.
“One will find periods when the observed rate of warming speeds up and exceeds the model average rate, and periods when the observed rate of warming slows down (the recent 15 year period) and is less than the model average rate.
“That’s exactly what you would expect, because the models all have their own cycles of natural variability.”
The point, he says, is that you can’t just grab one 15-year period to prove a point, the analysis need to look at a number of them. It’s the difference between shorter term oscillations and longer term trends.
“You could just as easily take the 15 years before that to argue they’re underestimating the warming,” he said.
What’s not in doubt, his study shows, is that global warming is happening and over the long-term, the modelling got it right.
“There’s an unmistakable warming trend over the last 100 years and that warming trend is well simulated by the models for the past, so there’s no reason to distrust the magnitude of future warming trends based on the past 15 years,” Dr Risbey said.