In my last post I updated a post from a year ago entitled Summer Daze in which I showed that a new pattern of sea level pressure anomalies has asserted itself in the wake of the sea ice crash of 2007. New research is out that examines this pattern, sheds some light on possible mechanisms behind it, and considers the implications.
Following my earlier post on the PIOMAS volume loss of 2010 I was digging further into whether there might be an atmospheric driver for the ensuing change in the PIOMAS seasonal cycle. I've discovered that there may be, although it's hard to separate cause and effect. In my searches I discovered that there was a new paper on Dr Jennifer Francis's publications page, the paper was entitled "The Recent Shift in Early Summer Arctic Atmospheric Circulation." Due to the change in PIOMAS seasonal cycle, with increasing calculated thickness loss over the summer the title of the paper made me ask myself if this had any bearing on the volume loss in 2010. So I emailed Dr Francis to ask if by 'recent' they meant after 2010. It turned out it was after 2007, and it had more to do with the pattern I found last year in the post Summer Daze.
I have discussed this matter with Dr Francis, and am grateful for her comments and the provision of several papers, including pre-publication sight of the new paper. However, this is my blog post, and my contact with Dr Francis does not imply that she or her colleagues endorse my understanding.
Almost inevitably, this pattern I'd seen had also caught the attention of other scientists who were not aware of my blog post. Dr Francis sent me a copy of a paper by Dr Edward Hanna and colleagues published early this year. Dr Hanna was studying the role different atmospheric and oceanic factors played in run off from the Greenland ice sheet. In the course of their study they found that run off from the ice sheet is most strongly correlated with the Greenland Blocking Index (GBI). The GBI being an index of atmospheric blocking over Greenland due to stationary and persistent high pressure. They found that since 2007 there has been a notable increase in GBI, which is effectively what I have found using another method in Summer Daze, the key issue from that being that the new summer pattern in sea level pressure is centred on Greenland.
Hanna has joined forces with Overland, Francis and Wang to examine this behaviour of the GBI. Overland and Wang have produced several important papers on changes in the Arctic atmosphere, and with Francis co-authoring a paper "Evidence linking Arctic amplification to extreme weather in mid-latitudes" earlier this year, her work seems to have been highly relevant to this issue.
The Arctic dipole is calculated using empirical orthogonal functions (EOFs), which is basically a mathematical technique for extracting patterns in order of their contribution to the overall data. The most significant EOF is that describing the Arctic Oscillation, which produces a flow around the pole due to the polar vortex, which in turn is a consequence of the spin of the Earth. Because the polar vortex cycles air around the pole it has a tendency to keep cold air collected at the pole, and reduce mixing with warmer lower latitude air. The second most significant EOF has in recent years become more prevalent, it is called the Arctic Dipole (AD). In contrast to the Arctic Oscillation the AD draws air across the pole, so rather than keeping the pole cool in its pool of cold air, the AD draws warmer air across the Arctic basin from lower latitudes.
The Arctic Dipole was first identified by Wu, more here, as the second EOF during the latter part of the 20th Century. Zhang later found that in the 21st Century the AD had moved from a secondary role to a more dominant one, more here. Jia Wang (not Muyin Wang, co author on the Overland 2012 paper) showed that the AD has a strong role in sea ice loss events from the 1990s into the 2000s, more here.
In the past I've been using Wu's definition of AD index, that the index is positive with when pressure over the Canadian Arctic Archipelago (CAA) and Greenland is high, and pressure over the Siberian sector is low. Now however Overland, who seems to be running the lead in this issue, has for several papers used the opposite definition. So from now on I'll be using Overland's definition.
Here is the timeseries of the AD index for June. The vertical axis is in standard deviations.
Timeseries of June AD index from reference 2, figure 2. Grey bars is AO index, blue line AD index.
From this the significance of the post 2007 shift can be seen. Not having the mathematical ability to calculate EOFs, I've not been able to produce such a series. However I've noted that the NSIDC have said that since 2007 the AD has been a regular feature of summers post 2007, and have said that I suspect the pattern I found in Summer Daze was indeed the AD, ref.
But why see the AD in the new summer pressure pattern? This is simply a matter of the interplay between high pressure over Greenland and the CAA, and low pressure over the Siberian sector. Because high pressures rotate air masses in a clockwise manner, and low pressures in an anti-clockwise manner the action between the two pressure systems is like two wheels on either side of a plank of wood - the upshot being that they produce air motion across the Arctic basin, rather than around it as with the Arctic Oscillation. The reason this has a detrimental impact on ice is two-fold: First the motion drags warmer sub Arctic air over the basin, second the motion of the air drags ice with it, from the Pacific Sector to the Atlantic Sector.
In the above adapted graphic of this year's June to August I've annotated the general motion from low and high pressures (purples/blues and greens/yellows respectively). The most intense purple around which I've annotated low pressure flow for convenience is due to the early August storm.
The Atlantic sector receives a large amount of warmer water from the Atlantic, this melts ice. Indeed the reason the ice edge doesn't proceed much further into the Arctic than the chain of islands from Svalbard to Zemlya Franzsa Iosifa (map) is that poleward of these islands is an abyssal deep. The Atlantic water is saltier and denser than Arctic water, so when it hits the edge of the abyss it drops out of contact with the ice edge. This means that ice pushed beyond this demarcation is destined to melt out in the summer. Furthermore ice transferred to the Atlantic sector is likely to be entrained in the transpolar drift, which pushes ice out into the Fram Strait between Iceland and Greenland, where it melts at lower latitudes.
So the pattern I described in Summer Daze is really the Arctic Dipole. I do wonder whether it's correct to talk about the AD, because the one process may not be forming all incidents of it, e.g. going back into the 20th Century. But due to the consistent pattern identified in Summer Daze and discussed in Overland et al, I think it's safe to say that since 2007 there has been one over-riding process driving the formation of the pattern.
So what is causing this pattern? As this post is long enough already I'll go over that in the next post.
 Hanna et al, 2012, "The influence of North Atlantic atmospheric and oceanic forcing effects on 1900–2010 Greenland summer climate and ice melt/runoff."
 Overland et al, 2012, "The Recent Shift in Early Summer Arctic Atmospheric Circulation."
GRL, in press.