Arctic Ocean could be ice-free much sooner than expected

As this year’s Arctic sea ice is confirmed to be fifth lowest on record, new research suggests the ice could be almost completely gone in summer in 30 years

UPDATE 4/4/09: Please see the end of this post for insightful comments from James White, a paleoclimatologist and director of the Institute of Arctic and Alpine Research here at the University of Colorado. I emailed Jim when I was working on this post, but he got back to me after I had finished. I include his comments in full below. 

Ice thickness for the Arctic in September now (upper) compared with conditions of a nearly ice-free Arctic Ocean in 30 years (lower).

With Arctic sea ice already dwindling much more quickly than the IPCC has predicted, researchers  in Seattle  say it could be gone in September by the late 2020’s — three times more quickly than previous estimates. 

The researchers, Muyin Wang of the University of Washington, and James Overland of NOAA, used six models that have proved most reliable in matching what has actually happened to sea ice in recent years. Starting with the what they called the “extreme” minimums of summer Arctic sea ice extent reached in September 2007 and 2008, they modeled what might happen over time, taking into account different scenarios of greenhouse gas emissions and natural variability.

“The two sequential years of extreme low sea ice extent at the end of summer in 2007 and 2008 are indeed evidence that the Arctic may be on a fast track for increased September sea ice reduction over the next 30 years,” Wang and Overland write in their paper, which was published today in Geophysical Research Letters. 

A positive feedback mechanism contributes to more rapid melting, the researchers say. With less ice coverage in summer, more heat is absorbed from the atmosphere into the ocean. Then, as the Arctic slides toward autumn and sea ice shrinks to its seasonal minimum, more of that heat is released back into the atmosphere, reinforcing the cycle. 

Wang and Overland note that natural variability is a factor in their simulations. But without forcing from emissions of greenhouse gases, major sea ice loss does not occur. 

“The Arctic is changing faster than anticipated,” Overland says, quoted in a press release from the American Geophysical Union.  ”It’s a combination of natural variability, along with warmer air and sea conditions caused by increased greenhouse gases.”

The new research comes in the wake of this announcement from the National Snow and Ice Data Center : 

“Arctic sea ice extent reached its maximum extent for the year, marking the beginning of the melt season. This year’s maximum was the fifth lowest in the satellite record.”

The ice reached its maximum extent of 5.85 million square miles on February 28. That was 278,000 square miles below the 1979 to 2000 average, or an area roughly equivalent to the size of Texas. 

NSDIC says a more detailed analysis of winter sea ice conditions will be coming next week.

UPDATE: After writing this post, James White, director of the Institute of Arctic and Alpine Research provided some commentary about this new study. Here is his response, in full: 

“This is an attempt to update the model predictions for an ice free (in summer) Arctic. We were caught short in 2007 when the sea ice shrunk by a third, and this was not predicted.  While the models used in the IPCC AR4 have non-linear behavior — that is, sea ice reaches a threshold after which it goes away at a faster pace — none of them got the timing right; all predicted that we had much more time before the fast decline would begin.

Part of the reason for this mismatch is probably getting the forcings right — for example, a warmer ocean is important and just how fast it warms is hard to predict and model. Also strong, persistent winds at the right place and right time can break up thin sea ice in a hurry, and when that occurs in the model world it is apparently not the same as when that occurs in the real world.

There are other problems as well, including the albedo of sea ice which affects surface melting, etc. The other side to this is predicting when the sea ice will reach a critical thinness after which it is vulnerable to events like what happened in 2007 when atmospheric winds (and warmer ocean temperatures?)  tore into the thin sea ice and eroded it. Getting that timing right is a tough job.

Modeling climate and the environment is an evolutionary process. We do our best, which we know is not as good as we would like to do, and as we learn more, we evolve, fix the models, and try again. This paper is telling us that our new, best predictions for an ice free Arctic are in a couple of decades. Do they have it right? Nobody knows. Certainly you shouldn’t set your watch to their timetable. Is this a better prediction than we had before? Absolutely yes. Which means all of our policy decisions about an ice free Arctic need to be revisited because their timetable is much faster than the one we were using.