Below are the rankings for 925hPa air temperatures north of 72.5N, based on the ncep reanalysis for since 1948.
[...]
Extremely interesting statistics!!
Compare 2011 and 2015
Year May..June...July.Aug.MJJA
2011 7.......5.....1.....2.....1
2015 27....11.....2....26....13
These years shared similarities since the beginning of the season (remember Vid?). Even the shape of the final pack has a lot in common. And July month.
However see August 26th position! And yet the huge wreck that three weak cold storms caused over the already broken MYI Beaufort ice, and the compaction from the relatively cold Laptev area. Then the final storm.
And May in 27th position, which saw a warmth wave in Alaska and Canada, with South winds that opened up the Beaufort sea and Chukchi kicking in albedo effect, a preconditioning element that, in absence of it, melting would have been delayed.
Then snow cover preconditioning due to Winter weather. Not accounted for in these statistics.
Overall averaged temperatures are extremely misleading.
Local temperatures 850hpa-925hpa and land surface temps can be significant though, especially in Peripheral areas during May-June.
This year we also had the out-of-mind hot Pacific and the few systems it brought to the Arctic, sometimes via North America, sometimes warm advection, other times not-so-warm systems like the big Chukchi August storm. You can have more and more heat being transported to the Arctic Ocean, under what seems overall cool temperatures. Then the ocean stores heat and releases it later in the season.