News Archive

Comparing climate model projections directly with observed global warming trends is comparing 'apples with oranges', according to scientists including The University of Western Australia's Stephan Lewandowsky.

Their comments, part of a paper published today in Nature Climate Change , come after some studies questioned whether climate model projections were overestimating recent temperature trends, given the ‘slowdown' in the rate of surface warming over the past 15 years.

However, the scientists say closer examination of the projection models - taking into account other key influences - reveals the models have in fact performed well in predicting global warming trends.

Winthrop Professor Stephan Lewandowsky, a cognitive scientist from UWA's School of Psychology and the University of Bristol who has been modeling the effects of scientific uncertainty in the climate system, said climate model projections represent long-term expectations that reflect temperature trends across centuries.

"Because projections have no information about the sequence and timing of internal climate variability, they average across the ups and downs of short-term trends in response to natural climate variations," Professor Lewandowsky said. "One of the most important drivers of internal variability is the El Niño-La Niña oscillation, which determines how much heat is taken up by the oceans rather than the atmosphere.

"In order to meaningfully compare model projections against recent observed trends they must therefore be synchronised with natural variability."

When the authors focused on those models that were synchronised with the El Niño-La Niña phase of the oceans, they found the models were in fact accurately estimating recent temperature trends and were even mirroring the spatial distribution of heat in the oceans.

"The results show that models selected in this way provide good estimates of 15-year trends over the past half century, including for the most recent 15 year period," lead author James Risbey said.

Dr Risbey, a synoptic climatologist from CSIRO Oceans and Atmosphere in Tasmania, said 15-year temperature trends sped up and slowed down in response to natural climate variations represented by El Niño or La Niña dominated periods.

"The preference for El Niño or La Niña periods in the climate model projections is not synchronised between models and the real world," he said.  "This means that any particular model is not expected to give the same 15-year trend as observations in any given 15-year period.

"Thus, comparing model projections directly with observed trends is comparing 'apples with oranges'."

The paper, Well-estimated global surface warming in climate projections selected for ENSO phase , was written by Dr Risbey; Professor Lewandowsky; Naomi Oreskes of Harvard University; and Clothilde Langlaise, Didier Monselesan and Terence O'Kane of CSIRO Marine and Atmospheric Research.

Media references

Winthrop Professor Stephan Lewandowsky (UWA School of Psychology - based in Bristol)
David Stacey (UWA Media Manager)  (+61 8) 6488 3229 / (+61 4) 32 637 716