Predictions from a collection of the latest climate models on average say that ice will be nearly gone from the Arctic by the 2030s. But when you don’t include man-made – or ‘anthropogenic’ – CO2 emissions’ ‘forcing’ effect, those models show a much icier picture. “This clearly shows that if you don’t consider anthropogenic forcing, the ice won’t decline that fast,” said Muyin Wang from the University of Washington. “It should be oscillating around a much higher level.”
These findings echo some that Muyin and her Seattle colleague James Overland, from the US National Oceanic and Atmospheric Administration (NOAA) Pacific Marine Environmental Laboratory, made in 2009. Then, James and Muyin used climate models that formed the basis for the Intergovernmental Panel on Climate Change’s fourth assessment report, which was published in 2007. “Because of this report’s success a lot more modelling groups around the world started doing simulations,” Muyin told me. Scientists are now bringing their improved old models together with new ones in a project to compare them. Having found the old models bad at reproducing measured shrinkage of Arctic ice at the end of the 20th century, James and Muyin wanted to see if the new and improved ones could do any better.
It’s important to be able to reproduce real data to be confident in models’ predictions, Muyin said. “If you are interviewing someone for a job, you look at their resumé, to see if they did a good job in the past,” she explained. “Then you know that they can do the job going forward. It’s a similar idea here, if models can simulate the past climate, then they’re the models we want to use in the projection.”
In a paper published in the scientific journal Geophysical Research Letters on Tuesday, they started from 32 different models, and compared them with satellite data on sea ice coverage. Overall, their resumés were slightly better than the older models: For the period from 1981-2005, the average of all these models was near the ice coverage actually seen, whereas the older models had overestimated the values. But the highest and lowest estimates in both groups were still very similar.
When James and Muyin looked at the entire group of new models, some showed that there would still be Arctic ice in 2100. So the scientists did what they had done in 2009 – only looked at the models that successfully simulated recent ice loss. “We analysed only 7 models that performed better at simulating the sea ice extent,” Muyin said. “If you want to look at the Arctic you should concentrate on these models, rather than looking at the whole group of 32 models.” Those seven, in a scenario where climate-change causing greenhouse gas emissions stayed very high, showed a rapid decline. The average time until the minimum sea ice area in the Arctic fell below a million square kilometres was 28 years from 2007, with predictions ranging from 14-36 years.
Muyin and James write that they are ‘discouraged’ that the models aren’t better at reproducing the Arctic melt we’ve seen. But Muyin said that there are reasons why it’s difficult. “There are still a lot of things we don’t fully understand about the Arctic climate system,” she explained. “We still need to find out which are the most important processes that control the sea ice growth and disappearance. That’s very challenging for us to do. Also, for the Arctic, observational data are short and sparse. That makes the modelling harder than if we had more than several decades of continued, systematic observations.”
However, by pointing out which ones are the best at simulating the Arctic, Muyin is helping make sure predictions for the area are as reliable as they can be. With a record minimum coverage reached this year, that’s more important than ever for areas where Arctic ice affects society, the environment and the economy. “When I go to meetings, other researchers do recognise that these models are better at simulating the ice, so it does have an impact,” she said. “They are still useful, and they are the only tools we have.”
Muyin Wang and James E. Overland (2012). A sea ice free summer Arctic within 30 years: An update from CMIP5 models Geophys. Res. Lett. DOI: 10.1029/2012GL052868