DR wants to know what this season looks like....after the end.
There is still a month to go until last year's minimum, not to mention warm water induced bottom melt that could continue well into October in some regions. We may yet see a surface manifestation of the anomalously mild fall and early winter, unless it has been neatly off-set (in the detrended sense) by cloudy summer weather.
If not, it's already worth contemplating what a hypothetical 'The Year in Review 2016/17' forum might contain. Since open water on 15 Sep 16 was nearly half the area of the Inner Basin (1st graphic), it's worth tracking what became of it over the yearly cycle.
This region should be called the Arctic Ocean but isn't. Some committee decided in the 1950's to throw in the Barents and Greenland seas and maybe the Bering and inland Canadian waters as well. Consequently to determine an accurate areal figure for the Inner Basin (to enable percentile scoring of sea ice concentration maps), it becomes necessary to mask out extraneous cells on an equal area grid.
By 29 Dec 16, open waters of the Inner Basin had frozen up and so defined the first year ice FYI. Under the
cafeteria tray model of the Arctic Ocean (first in, last out FILO or here first freeze, last melt FFLM), the open water latest to freeze (ie the Chukchi) is the first to melt, with the overall annual pattern
eFYI .. mFYI .. lFY ..
lFYI .. mFYI .. eFYI for early, middle and late ice.
Actually that monthly classification could be refined considerably using the 107 daily maps of the refreeze period, though weekly gives a more manageable partition of FYI ice classes. Bulk and shear movement of the ice pack, along with compaction and dispersion, can confuse the issue in some years though not so much in this one. Note too late-to-form ice can catch and even pass up early ice if transported to colder waters and weather.
The animation below marks up newly formed FYI in 5-day increments for both the freeze and melt seasons (using boolean differencing on a tiled-up overlay that pairs each interval with the previous). Twenty frames in mid-winter were uninformative (stationary) and were dropped.
It would be good to assign distinct colors to new ice to each of the 21 freeze frames to see what becomes of the FYI partition during the 22 melt frames. However that gets complex in a hurry as the melt season only time-reverses the freeze season to a 0th approximation, if that. (However it's not to easy to tell them apart running the animation backwards.)
To a certain extent, this process is captured by the diffuse colors of the annual Hycom thickness maps as there's little confusion between sub-2m FYI and older ice. That's included below for similar dates as the UH AMSR2 animation of FYI ice classes. The latter does not provide support for the Hycom ice finger discussed up-forum.