Increased greenhouse gas emissions may drive up atmospheric temperatures by as much as 35% more than climate scientists previously believed, according to more than a half-dozen of the new climate models that will inform the next major assessment from the Intergovernmental Panel on Climate Change (IPCC).

“The findings—which run counter to a 40-year consensus—are a troubling sign that future warming and related impacts could be even worse than expected,” Weather Underground reports. The calculations show the planet’s equilibrium climate sensitivity (ECS), defined as the air temperature increase that would result from a doubling of greenhouse gas concentrations, at 5.3°C, far above the previous threshold of 4.0°.

The 35% increase in sensitivity comes from the second version of the Community Earth System Model (CESM2) from the U.S. National Center for Atmospheric Research (NCAR).

“It is imperative that the community work in a multi-model context to understand how plausible such a high ECS is,” said the authors of a paper published last month in the journal Geophysical Research Letters. “What scares us is not that the [new model] is wrong…but that it might be right.”

“At least eight of the global-scale models used by IPCC are showing upward trends in climate sensitivity,” Weather Underground writes, citing an online column by IPCC lead author Joëlle Gergis, a scientific advisor to Australia’s Climate Council. “Researchers are now evaluating the models to see whether the higher ECS values are model artifacts or correctly depict a more dire prognosis.”

“The model runs aren’t all available yet, but when many of the most advanced models in the world are independently reproducing the same disturbing results, it’s hard not to worry,” Gergis wrote.

The Weather Underground report traces the significance of the new models back to 1979, when a landmark U.S. climate study placed the planet’s equilibrium climate sensitivity at between 1.5 and 4.5°C. The IPCC’s five full assessment reports since then have “largely agreed with this conclusion, although a few individual models have gone outside the range.”

The next assessment report is due in 2021 or 2022, based on research published through the end of this year. “It does indeed look like many of the latest models will have ECS values higher than the IPCC ‘likely range’ of 1.5 to 4.5°C,” said University of Exeter mathematician Peter Cox, lead author of a 2018 study on temperature variability around long-term warming. “It seems that the new models with high ECS have more low-level cloud that tends to burn off under climate change, producing an amplifying feedback on warming.”

In his own study last year, Cox concluded the ECS was unlikely to move outside its long-accepted range of 1.5 to 4.5°C, Weather Underground notes. Even now, “it is worth noting that observational constraints from both the temperature trend and temperature variability still suggest ECS of around 3°C,” said Cox. “So climate science has a conundrum to solve here.”

The impact of cloud cover has been a long-standing uncertainty in climate modelling, and the Weather Underground post digs into some of the recent research, along with the details of the latest NCAR model. 

That work “is based on tests of nearly 300 model configurations, with a focus on how well the models simulated pre-industrial climate and how well they reproduced the main global temperature trends of the last century,” explains WU meteorologist Bob Henson. “These trends include warming from 1920 to 1940, a period of roughly steady global temperature with regional cooling in the mid-20th century, and a more sustained global warming since the late 20th century.”

NCAR “also took into account new estimates of aerosol emissions (soot and other particles and droplets),” he adds. “Aerosol pollution tends to cool the climate overall, both by blocking sunlight directly and by serving as nuclei for clouds that block sunlight more effectively.”

The model did indeed show a stronger cooling effect from aerosols. “However, the stronger aerosol-related cooling also led to an unrealistic portrayal of 20th century climate. When the model was reconfigured in response, it produced a more accurate reproduction of 20th- and 21st-century climate, including cloud behaviour—but with a higher ECS, which pointed to a more ominous portrayal of future change.”

That’s what has the NCAR scientists concerned. If the higher ECS in the new models turns out to be on the right track, “it’s really bad news,” said Andrew Gettelman, a co-author of the Geophysical Research Letters paper. “It means we are going to be on the warm end of projections, with larger impacts for any given emissions trajectory.”