I'm not posting to launch into a long boring discussion of the pack floe by floe. I think there's a more objective way to cut to the heart of the issue. The data I use are IJIS Sea Ice Extent index, data, and CT Area, data.
Over at Neven's Sea Ice blog there is a measure of compactness of the pack that's been used for some time, termed CAPIE, links. CAPIE stands for Cryosphere Today Area Per IJIS Extent. I've followed that for in reading that blog, but have long thought the index is upside down, because we're moving towards a more dispersed pack.
CAPIE is CT Area divided by IJIS Extent. What I've been following for a while, but somehow forgot to post in a blog post last year, is NSIDC August Sea Ice Extent divided by the August average for CT Area. This has been used to try to overcome weather, while keeping valid information about the sea ice state at the end of the melt season. The September averages are too late because they would include information about the first week or two of re-freeze.
Here is the August dispersion index from 1979 to 2012.
It can be seen that there is a strongly increasing trend of dispersion, heightened by a step jump following the 2007 crash.
This is to be expected with the continued thinning of the ice and increase in Open Water Formation Efficiency. Why it should appear in the ratio of Extent/Area may need to be explained. Wikipedia's page on Measurement of Sea Ice, link, states the following:
To estimate ice area, scientists calculate the percentage of sea ice in each pixel, multiply by the pixel area, and total the amounts. To estimate ice extent, scientists set a threshold percentage, and count every pixel meeting or exceeding that threshold as "ice-covered." The National Snow and Ice Data Center, one of NASA’s Distributed Active Archive Centers, monitors sea ice extent using a threshold of 15 percent.So for extent any region of over 15% ice concentration is tallied as full ice, whereas for area the same region will have a lower area value if it contains pixels with a concentration of ice less than 15%. This makes extent a somewhat blunt tool, but because of this more statistical confidence can be had from the measure of extent, which is why the professional scientists use it. However it can be appreciated that as the pack develops more open water in the form of loose floes in water, or cracks between floes, area will decrease more than extent will.
There is not such a clear trend of the Dispersion Index earlier in the melt season, but because the principle behind it (more fragmentation means a higher ratio) is so simple it seems reasonable to apply this to the current state of the pack.
The Dispersion Index for the first two weeks of July does not support the idea that 2013 is exceptionally dispersed. However as the plot of daily dispersion index, or inverse of CAPIE, shows, it is only from the end of this month that the spread should increase leading to the rising trend in August Dispersion Index.
This is not a graph I intend to update. Click to enlarge.