In my view, it is silly to wait around for the entire Arctic Ocean to be below some arbitrary sq km cut-off and downright preposterous to wait around for five consecutive such years -- in terms of effects,
the future is already here in considerable measure for the Beaufort, Chukchi, Barents, and polar front areas.
How could it not be?
The first animation shows the open water boundary of a Beaufort/Chukchi wedge for Sept 1st through Oct 4th. These waters are
already seasonally ice-free for much of the insolation season and well into the fall and so are
already having their 'end-game' effects on albedo, radiation budget, cloud formation, NH atmosphere, etc.
The second animation overlays the open water boundaries over Worldview infrared cloud imagery of the same dates. As jdallen notes above, a heavy cloud cover is not altogether unexpected.
http://go.nasa.gov/2dTmCyZUsing this partition, the third animation (of Sep 10th) then classifies the cloud cover by thickness, separately for over open water and for over ice. This scheme uses just three bins for thick, thin and clear which is by no means an adequate characterization of clouds and their effect on the Arctic Ocean energy budget.
However these cloud masks, along with length of day, sun angle,and Project Sheba data from Perovich are a start on insolation reaching the surface and heat emissions being re-radiated back down from the local greenhouse effect.
The fourth image is just a visual definition of the Arctic Ocean used to convert AMRS2 UHH pixels into square kilometers for purposes of the spreadsheet columns overlaid on the first animation. Wipneus, somewhere, has a similar definition with more accurate areas. Note that official bodies such as the Int'l Hydrological Union throw in various extraneous waters such as Hudson Bay into their Arctic Ocean definition (see map at wikipedia).