The decline in Arctic sea ice has become an iconic symbol of global warming. You don’t have to look far on the internet to find predictions by scientists, campaigners and commentators about how soon the region will become ice free in the summer. Unfortunately for those predictions, the Arctic ice has not been listening.
The Arctic sea ice has probably reached its greatest extent for this year. It usually occurs at the end of March – last year it was March 21st. There have been some reports that this year’s maximum extent was the lowest since satellite monitoring began in 1979, and it certainly looks low hovering around 13 million km2 for over a month, see Fig 1 (click on image to enlarge). But looking back over past behavior its maximum extent was similar last year and in 2011 and 2005-7 (Fig 2). Hence this year’s extent is not that unusual being similar to that observed ten years ago!
The extent of minimum sea ice is also doing something very interesting – there are hints of a “pause.”
When satellite observations of Arctic ice extent began in 1979 it was obvious that a long-term decline was already underway. That decline appeared to be monotonic until the mid-2000s when, for a while at least, it seemed to have accelerated. The ice extent in the summer of 2007 was a record low, and was accompanied by cries from some quarters of imminent collapse.
The same was said in 2012 when another low was observed. However 2012 was an unusual year as an intense storm occurred in August and its effects on concentrating the ice cover can be clearly seen in the data. Likewise 2007 was an exceptional year.
We now know that year had what was later called an “unusual atmospheric pattern,” that is clear skies under high pressure that promoted a strong melt and at the same time winds brought warm air into the region.
These exceptional years became statistically important as using them to guide a straight line through the Arctic ice decline made its gradient even steeper.
A New “Pause?”
Examining the sea ice extent data for the past eight years it is obvious that there has not been any statistically significant downward trend, even though there is more noise (interannual variability) in the data. There are interannual variations but they do not form a trend. For the 2002 – 2006 period the annual differences are mostly in the extent of maximum and not minimum ice cover. The period 1990 – 1996 displays much more interannual variability. The main difference between the ice-curves is that in recent years there has been an increase in the gradient around the beginning of June.
Of the general decline and the interannual variability how much is due to external forcing and how much to internal variability? Estimate from climate models give about equal measure to forcing and internal variability, Kay et al (2011), Stroeve et al (2012). That 50% internal variability is almost never illustrated graphically when presenting Arctic ice data.
That the minimal extent of Arctic ice has “paused” is admitted by Swart et al (2015)
“…from 2007–2013 there was a near-zero trend in observed Arctic September sea-ice extent, in large part due to a strong uptick of the ice-pack in 2013, which has continued into 2014.”
Swart et al (2015) maintain that “cherry-picking” such short periods can be “misleading about longer-term changes, when such trends show either rapid or slow ice loss.”
The situation with this “pause” in Arctic ice reminds me of the early days of the annual average surface temperature “pause.” When it was first raised, around 2007 with then an estimated 5-year duration, it was dismissed as being cherry-picking and being well within the internal variability of the models, Researchers then looked at similar periods throughout the surface temperature data and the climate models and asked what the likelihood was of a period of no change, just like what Swart et al (2015) have done for the Arctic ice. Back in 2008 the UK Met Office said that climate models regularly showed eight-year pauses but not ten years. Then it was a ten-year pause and, of course, the models were able to explain that after all, and so on even as many viewed the “pause” as increasingly problematical. The analytical approach to the “pause” in Arctic ice is repeating that of the surface temperature “pause.”
Something may indeed have changed in the pattern of Arctic ice melting. The decline that was already in progress when satellite observations were started in 1979 show that Arctic ice was shrinking even before human effects were strong (Fig 3), although the decline between 1979 and about the mid-1990s is not that significant! The extent of minimum Arctic melting may have paused over the past eight years. It will be interesting to see if it continues in the future. But whatever happens the big question will remain. Is it caused by internal variability masking continuing human-induced-sea-ice loss? Or has internal variability over decadal periods since 1979 been misinterpreted as human-induced decline?