Why do climate models differ from the real world?

Earlier this year, weather and climate agencies around the world declared 2014 the warmest year on record, even though the increase in global mean temperature has slowed. This warming “hiatus” has puzzled climate scientists, as it deviates from climate models which project a continuing temperature increase. Climate expert Jochem Marotzke visited MIT last week to deliver the 15th annual Henry W. Kendall Memorial Lecture, “Recent Global Temperature Trends: What do they tell us about anthropogenic climate change?” in which he discussed the hiatus as well as the abilities and limitations of climate models.

Marotzke is a director at the Max Planck Institute for Meteorology in Hamburg, Germany, and was an MIT Department of Earth, Atmospheric and Planetary Sciences (EAPS) faculty member in the 1990s. He has spent his career researching the role of the ocean in climate and climate change, and recently expanded his interests to include multi-year to decadal climate prediction. “If you look at other central indicators of global climate, such as sea ice melt, ocean heat uptake, and sea-level rise, they show that global warming is continuing,” Marotzke said. “But this particular indicator, global surface temperature, is rising at a much lower rate now. This is something that as a climate research community we need to take seriously; we need to understand it and communicate the issues about it.”

For the past 15 years, increases in global mean surface temperature have slowed contrary to what climate model simulations predicted. Known as the warming “hiatus”, this phenomenon is largely due to natural variability: Cyclical climate processes such as La Niña and fluctuations in the amount of solar radiation reaching Earth’s surface can disrupt the warming trend. Additionally, the oceans absorb an enormous amount of excess heat energy trapped by the atmosphere — as much as 93 percent, Marotzke said. Light-reflecting aerosols from volcanoes also contribute to the slowdown.

The failure of climate models to predict this hiatus has long perplexed scientists and bred some public mistrust in climate models. Climate-change skeptics claim the hiatus is proof that global warming doesn’t exist, and that climate models overestimate greenhouse gases’ warming effects. Marotzke ardently disagrees. He shared with the audience a study published in Nature earlier this year in which he and co-author Piers Forster of the University of Leeds analyzed 114 model simulations of 15-year global mean temperature trends since the beginning of the 20th century. If their analysis showed that models consistently overestimated or underestimated the amount of warming compared to real-world observations, then the models must have a systematic bias.

Fortunately the simulations performed fairly well, producing a range of predictions for each 15-year period in which actual observed temperature trends for those periods fell. Even if the observed trends at times fell close to range edges, they were not biased to one side or the other. Although the models didn’t accurately predict the current warming hiatus, which is not unusual, they also failed to predict other accelerated warming or hiatus events. In fact, the models underestimated warming in some periods compared to the observations. “The claim that models systematically overestimate warming from increasing greenhouse gas concentrations is unfounded,” said Marotzke.

To find out what these simulated short-term temperature trends actually tell us about the climate, Marotzke and Forster performed a multiple regression analysis, which aimed to identify the most significant factors contributing to the trend. For shorter 15-year periods, the analysis found random natural variability in the climate system had the largest influence — approximately three times the impact of all other physical factors combined. Only when Marotzke and Forster analyzed model simulations of global mean temperature trends spanning 62 years did differences in factors including ocean heat absorption, greenhouse gas concentration, and aerosol pollution begin to make a noticeable difference.

In other words, modeling 15-year-long periods only shows the impact of natural variations in the climate system. To see anthropogenic influences on climate change, we have to look at the bigger picture. “The hiatus masks anthropogenic warming,” said Marotzke. “It is a huge distraction, but an incredibly fascinating one.”

The 15th annual Henry W. Kendall Memorial Lecture was sponsored by the MIT Department of Earth, Atmospheric and Planetary Sciences and the MIT Center for Global Change Science. The lecture series honors the memory of Professor Henry Kendall (1926-1999), a 1990 Nobel Laureate, a longtime member of MIT’s physics faculty, and an ardent environmentalist. A founding member and chair of the Union of Concerned Scientists, he played a leading role in organizing scientific community statements on global problems, including the World Scientists’ Warning to Humanity in 1992 and the Call for Action at the Kyoto Climate Summit in 1997.

This article is published in collaboration with MIT News. Publication does not imply endorsement of views by the World Economic Forum.

To keep up with the Agenda subscribe to our weekly newsletter.

Author: Cassie Martin is a Science Writer at MIT News.

Image: A view is seen from the Amazon Tall Tower Observatory (ATTO) in Sao Sebastiao do Uatuma in the middle of the Amazon forest in Amazonas state. REUTERS/Bruno Kelly.