I blame Oren.
Great posts! Thank you so much gerontocrat. I blame myself as well, as I should have done the analysis myself instead of suggesting work for others... but your charts are much better looking than mine and the results more comprehensive.
In general I think the 50% charts are best at showing the trends for most seas.
Some commentary:
Greenland Sea
At 50% or less, a trend is visible, the number of days increasing from about 100 to about 150.
However, annual variations of export of ice down the Fram Strait presumably have quite an effect on sea ice area.
I didn't expect a trend for GS, but it seems the pre-melting of thin ice before it even reaches the Fram, and massive post-export melting in recent years while still in northerly waters, have caused this trend to materialize.
Barents Sea
The loss of sea ice area in the Barents sea is far greater.
At 5% or less, number of days increased from about 50 to well over 100,
At 15% or less, number of days increased from about 100 to approaching 200,
At 50% or less, number of days increased from about 150-200 to 300-350.
The Barents is one of the worst ground-zero locations for arctic change. The melting season has doubled in length, and about half of the Barents is now ice-free year-round (meaning of 365 days on the 50% chart). There was a major step change in 2006.
Baffin bay,Bering Sea, Okhotsk, St Lawrence
The Bering Sea data does show an upward spike in the last two or three years. (Perhaps 2018 is a continuation of that trend?)
Interesting. The Bering data might stand out more if the analysis was from Sep 1st to Sep 1st, as the 2017/2018 season had a very late refreeze and a very early melt.
The graphs do show the gradual advance of warming into the Kara, Chukchi, East-Siberian, and Laptev seas. Of note is that at no time is the extent in the Central Arctic or Canadian Archipelago at less than 50%. A similar set of graphs show the same result for Hudson Bay.
Chukchi - at least 50% of it has been seasonally ice-free historically, but now it's almost the whole sea (shown in the 5% chart). The 50% melting season has doubled in length from 75 to 150 days.
Kara - amazingly similar to the Chukchi in all measures.
ESS - its 5% and 15% charts separate the top years from all the rest. To be a record contender you must clear most of the ESS.
Beaufort - might show a clearer trend at 80% or 90%, if true will show that melt onset (+no more refreeze of opened cracks) is earlier, and perhaps refreeze completion is later.
CAA/CAB - should show a trend at 80% or 90%. I believe some parts of these regions have become seasonally ice-free, while other parts are still ice-covered year-round.
Interestingly, HB has a step change around 1995 in all the graphs, and has been stable since.
From 1979 to 2017 loss of sea ice area on any measure has been gradual. What is going to change that and when ?
I think the gradual assessment is true. But bear in mind, it's enough that the 5% chart consistently reaches 30 days in order to reach a seasonally ice-free state for each region. All external peripherals (plus Kara and Chukchi on most years) are already there. Laptev, Beaufort and ESS all made appearances on this chart, and when they line up we get an 2012.
The only regions that are still protected are the CAB and the CAA, and if they were split into sub-regions I think some of them would turn out to be seasonally ice-free as well. How long will these holdouts last? We'll be certain to tune in and find out.