Re: Okono's last point -- this spring we had a substantial +++ice anomaly in Sea of Okhotsk. I think that is also important (as is snowcover in Eurasia). If spring Eurasian snowcover persists in southern regions more than normal that prevails in ++sea ice in Okhotsk and --SSTA across the NPAC (like this yr).
I share your hunch: I feel like the cryosphere is very likely to be involved here. That's not only for the plausible mechanisms, but also because it's novel. If this were a persistent, normal feature, we would have noticed it before 2014. It's huge.
To me, the cryosphere is the "working memory" for climate change. There are many other energy reservoirs and sinks, but ice is the most visible buffer for change on the timescale we're effecting.
I don't know how far deranged the Siberian winters we have seen are relative to historic variability. Since we've seen this persist for years, I want something that isn't just seasonal, and I want something truly extreme.
If this turns out to be a legit oscillator -- and I think it's far too early to know -- I want something variable, too. Doesn't need to be the same forcing(on/off switch, temperature dial).
Some candidates I want to look at are things that have changed markedly in recent years:
1) Sea ice extent, probably lagged
2) Actual SST's, not just anomalies
3) The disturbed North Atlantic Drift & thermohaline circulation
Other obvious candidates, but I don't know them to have changed markedly:
4) Any major and persistent surface current anomalies
5) Thermal depth profiles
6) Upwelling indices
7) Siberian snow extent, probably lagged