This season will be interesting for calibrating how important early melt season albedo is compared to inherited sea ice thickness. Snowcover is much more than 2016, but sea ice thickness is much less.
In particular, here's how I would compare 2016 and 2017 for each region using worldview imagery as of 5/18, starting with the land areas where 2016 has a signficant lead:
1. Alaska / Yukon / Mackenzie River Delta: 2016 has a slight ~5 day lead in snowmelt, particularly noticeable on Alaska's north slope, where 2016 had an extraordinarily early melt.
2. Nunavut / Hudson Bay shores / Quebec: 2016 has about a week lead in snowmelt.
3. Karelia / White Sea coast / Barentsz Sea coast: 2016 has a whopping 3-week lead in snowmelt.
4. Arctic coast from Gulf of Ob to East Siberian Sea: 2016 might have a lead of a couple of days. Snow extent is roughly comparable with 2017, although the snow appears a bit grayer in 2016 at this point.
5. Chukotka: 2016 has about a 4-day lead.
But what about sea ice? Here, the tables are turned, with a few exceptions:
1. Chukchi: 2017 has about a 1-week lead in terms of extent, although in terms of quality the difference looks even more stark. 2017 is full of rubble already, whereas 2016 was more compacted. And we know that thickness in 2017 is less too.
2. Beaufort: 2016 has about a 3-week lead in terms of extent, although we know that 2017 is missing any core of really thick ice in this region that 2016 had, that allowed areas like the "Big Block" in 2016 to persist almost all summer despite being surrounded by blue ocean for months.
3. CAA: About the same, both in terms of sea ice color and extent of breakup at each end of the Northwest Passage.
4. CAB: 2017 looks like it is about 2 weeks ahead. 2016 didn't show any substantial crack from the Beaufort all the way to north of Ellesmere Island until the first week of June.
5. Nares Strait: 2017 is about 2 months ahead. You have to skip ahead all the way to about July 18th in 2016 to see the sort of network of cracks and pulverized floes that 2017 already has. 2017 looks shockingly bad in this sector.
6. Hudson / Baffin Bay: (Not important for final extent, but still important for intervening albedo until then)...2016 and 2017 are about the same.
7. Greenland Sea / Barentsz Sea: Hard to do a comparison here because it is so at the mercy of short-term fluctuations based on wind drift and Fram export. 2016 clearly in the lead, though.
8. Kara Sea: Similar extent in each year, although 2016 looks slightly more rubble-ey and grayer, so I give a few days' lead to 2016.
9. Laptev Sea: Polynya extent is about the same. 2017's floes look more dispersed, although they also seem to have more of a blue-ish tint, whereas 2016's look slightly grayer/yellower, suggesting to me that 2017 is a few days' behind on surface snowmelt, albeit thinner overall.
10. East Siberian Sea: 2017 looks shockingly bad. In terms of extent, it is about a month ahead of 2016. The fast-ice line is hugely cut back towards the coast compared to 2016, the polynya is much larger, floes are more dispersed. After taking into account comparable sea ice color tint (indicating comparable surface snowmelt with 2016 maybe slightly grayer/yellower and farther ahead), I'd estimate 2017 is about three weeks ahead of 2016 here overall.
Overall, what I see is that weather has favored 2016's melting over 2017's so far, as indicated by more advanced surface snowmelt on both land and water in most places. However, that might be more than outweighed by:
A. Less inherited sea ice thickness for 2017.
B. A change in the weather pattern to catch 2017's snowmelt up to 2016 and give 2017 a sunny June/July (whereas 2016 generally had a cloudy summer after the early melt).
C. Some combination of the two above.
What really intrigues me is the Nares breakup already in 2017. It's not like that area is getting much absorbed sunlight. Albedo is still high. It has to be a combination of bottom melt and low thickness to begin with. If the rest of the CAB is that bad, we could see it deteriorate in a flash in July/August.