Found the Billelo equation which summarized looks like:
Ice thickness = SQRT(InitialThickness^2+(FDD/804.2082))
(Thank you, DOSBAT)
Assuming an initial thickness of zero, comparing the product of 3650 vs 4529 FDDs, that suggests a reduction of 1.1M of ice.
That seems a bit high to me, though it may actually apply to areas starting with no ice.
(corrected numbers - arithmetic fail
)
However, as this is an average across the entire Arctic, assuming a starting thickness of ~1.5 meters, to approximate average thickness and then applying 3650 and 4249 respectively, I get 2.67M - 2.45M = 0.22M less ice formed.
Where you have starting ice, this method may be more skillful, as it appears to take in account of the reduction in thermal transfer caused by existing ice.
My instinct is, to average the two values (Billelo & Lebedev) to provide an approximation - which suggests there's ~0.43M/1.4' less ice now as compared to last year, and quite a bit less compared to the average.
[Post Script - some additional thoughts] Thinking at a qualitative rather than quantitative level, the effect of increased heat earlier in the year, when there is less or no ice, will actually be more pronounced. This in particular is what's suggested by the Bellelo equation, which gives increasingly greater effect for the FDD's with thinner ice.
By extrapolation, once we have ice in place, even if it is fairly thin, the effect of the FDD's starts lower and decays faster with increasing thickness; once you get past about a meter of thickness, hundreds of additional FDD's gives only a couple of 10's of centimeters of additional ice.
QED, by crushing the fall freeze, the worst of the damage which could be done to the pack, has already been done. This was just the opposite of what I had been thinking as I'd hoped that a cool down in the last half of the refreeze might help us recover. I am now much more pessimistic.