I let excel do the 5 year moving average. If I did them I would have placed each data points two years earlier at the middle of the range rather than the end.
The April average to May average shows a remarkable upturn with the volume melting tripling.
May to June show a similar upturn close to doubling.
June to July is a bit unclear whether it is beginning to show an upturn.
July to Sept surprises me. It looks like there is a discontinuity at about 1997/98, but that timing might be suggestive of a large spike due to the super El Nino. That with random noise might give the impression of a discontinuity.
Anyway the strength of the decline from 1979 to 1996 still seems surprising given the lack of any trend in the total season melt unless there is some explanation. Therefore I am wondering if less MYI (perhaps particularly in ESS and Laptev?) was encouraging early season melt but once season was in advance of where it usually was after the early season melt, the melt slowed down to compensate for some reason(s). I would like to better understand those reasons for the slower melt in the later part of the melt season. The only thing that has occurred to me and/or been suggested by Dr Zhang is that with less ice, there is less area and less edge where the melt occurs.
Any other thoughts on whether there might be a discontinuity or reasons for a late melt season slowing of melt?
I was using the 5 year moving average as well, but wasn't breaking things down as finely month to month. What I noticed primarily with the 5 year moving average, was when computing anomalies from that, the dramatic break that takes place around/about 2000. The net loss, April-Sept starts making a consistent upturn, and the total melt increases. The Sept-April refreeze does show a significant, almost complete recovery, but it seems like the volatility of the refreeze/melt is oscillating around some *almost* predictable level. I'll continue tinkering with the data.
Once again, Albedo, energy loss, etc. is no doubt playing a part during any specific year, but it still seems the net energy available in the arctic establishes some sort of focal point for the variations in volume, +/- variability tied to characteristics of the specific season.
Even with an uptick (2013) I'm not sure the refreeze will take that value high enough, that the coming melt season won't see a drop below 2013 levels, possibly lower still. Even with radiative loss, the net inputs into the system seem quite high. Once we start getting insolation back, the equation becomes quite a bit more dire, I think. Those 20+ degree anomalies have to be having and effect. They may not be melting, but I suspect the imported heat is reducing the refreeze, and pumping energy into the region which replaces that lost through radiation. I'd love to have some sort of reliable heat data.