Extent is preferred by professional scientists because it is less sensitive to melt ponds impacting concentration in satellite measurements of sea ice. However I stopped using the various extent indices some time ago because I consider that in seeking to avoid such issues extent is losing valuable information.
There is a new feature of the Cryosphere Today Area series (CT Area), a June crash in anomalies, this is an important feature, which I suspect is a useful indicator of conditions that prime the ice pack for record melts.
The anomalies for CT Area are shown below. Anomalies are the difference from the long term average seasonal cycle of the ice pack.
It can be seen that in 2007, 2011 and 2012, there is a rapid drop in anomalies from after the first week in June. These years show a drop, with no recovery, in 2010 there was also a drop but a recovery in anomalies followed. This was because whatever caused the drop, afterwards there was a lesser reduction of area when compared to the long term average. The significance of these years in the longer term context can be appreciated from the earlier post in which I gave CT Area anomalies in graphics from 1980, link.
What is notable about 2007, 2012 and 2007, is that 2007 was a record breaking crash, 2011 was a 'tie' (a new record in some indices but not in others), and in 2012 there was a record breaking crash.
I've previously blogged on research into melt ponds by Rösel & Kaleschke, link. Using MODIS, they find that in 2007, 2010 and 2011, there were high levels of melt pond formation in comparison to other years (2000 to 2011).
The graphic above is from Rösel & Kaleschke, figure 3b. Along the horizontal axis are dates (18 May, 1 June, 15 June etc), the vertical axis shows area of melt ponding. 2007 (red line) and 2011 (magenta) against the average for 2000 to 2011 (black with one sigma bounds in dashed lines).
Below is a detail from the CT Area anomalies showing the June anomaly crash.
It can be seen that the anomaly crashes are happening around the same time as melt pond fraction is at its greatest. This is exactly the sort of 'contamination' of the measure of sea ice extent/area that the extent indices seek to avoid. In area the area of ice within a grid box is the fraction of ice shown by the satellite sensor, the sensor can pick up melt ponds as open water. In extent any grid box with more than 15% ice is counted as fully ice, so 'contamination' of the data by melt ponds is reduced to near zero giving a conservative measure of the amount of sea ice. However in being conservative important indicators, such as the June crash are not reflected.
Anyway, in 2007 and 2011 Rösel and Kaleschke find record June melt ponding, in the same years there is a crash in June CT Area anomalies. I have been unable to find independent information about melt pond area in 2012 in the context of previous years, but my working assumption is that the size of June CT Area anomaly crash is probably indicative of the area of melt ponding.
However as I have asserted before, the occurrence of June CT Area anomaly crashes post 2010 is supportive of the volume loss of 2010 being a real event, link.
I had thought that the June crash was due to rapid reduction of ice from the periphery of the pack. This is not the case. Using Cryosphere Today's regional breakdowns, link, it is possible to examine the change in anomalies during June 2012. Having the full series broken down into regions would be preferable, but as this is not available I will just have to make do with what I have.
The full decline in anomalies worked out by eye from the graphs is around 1.16M km^2, this is of roughly the same order as the June crash in the orange (2012) plot line in the above graphic. Of this the central Arctic shows a decline of around 0.4M km^2, a substantial fraction of the overall decline and the largest single contributor by far.
So whilst recession of the ice edge in the peripheral seas has a role, in 2012 much of the decline was due to an apparent lowering of concentration in the central Arctic, into which the ice edge had not penetrated by the end of June, i.e. University of Bremen, 1/6/12, 30/6/12.
While the exact mechanism is unclear, what is clear is that whether mainly due to low concentration ice, or melt ponds masquerading as open water, the June crash is associated with September record minima. This makes sense as melt ponds lower albedo, trapping more sunlight and warming the ice, promoting melt, and the same applies to leads between the broken floes of ice typical of summer conditions.
This year we see a record expanse of first year ice, this matters because first year ice is typically flatter than multi-year ice and more prone to developing melt ponds. Also if the ice profile is thinner opening of cracks between floes is more likely to develop open water within the pack, warming the waters within.
I have previously blogged on conditions shown by PIOMAS in March, link, and reflected in both ASCAT and the Drift Age Model (DAM) in December, link. Now DAM data is out up to the end of April.
There are ridged areas and areas of thicker ice within the expanse of first year ice, but it seems to me that the prospects for melt ponding and low concentration across that expanse are substantial this year. I expect a very large June anomaly crash starting in about three weeks. I hesitate to say I expect it to beat the 2012 record June anomaly crash, but I can't think of a logical reason for hesitation.