Does anybody care to disabuse me of my conjecture that there is a nonlinear relationship between ice thickness and melt resistance - with decrease in melt resistance curving down faster than the linear % decline in thickness - due to qualitative differences in thinner vs. thicker ice?
The fact that ice accumulation is radically nonlinear with increasing thickness is accepted as established fact, e.g. the curve published by Thorndike 1975
.
Earlier this year I pitched the idea that the reverse is true for melting, with 1 meter thick ice melting at twice the rate of 2 meter ice (0.8 cm/day vs. 0.4 in the example shown):
.
Those who actually understand the physics of ice melt shot down that theory, explaining that the energy flows involved in summer melt are not simply the reverse of winter freeze. Correction which I gratefully accept, .... but
....even if a straight reversal of the thickness-freeze rate curve to estimate thickness-melt rate curve is too simplistic to be valid, that still leaves open the possibility, and (in my mind at least) the near certainty that the melt rate vs. thickness ratio is not a stricltly linear 1:1 ratio. I have no idea what it would be, but it I'm almost certain that the melt rate for 1 meter vs. 2 meter thick ice has to be greater than 1:1. And that ratio has to be even greater for 0.9, 0.8, 0.7 etc. meter thick ice vs 2 meter ice.
It is well documented and accepted that the chemical and structural characteristics of Arctic sea ice varies with thickness. Those qualitative differences have to make some difference to the melt rate.
This is not merely an academic question. An accelerating melt rate with declining thickness would have major consequence for acceleration of Extent and Volume losses as average thickness continues to decline as shown on the chart posted by gerontocrat at
https://forum.arctic-sea-ice.net/index.php/topic,119.msg275579.html#msg275579 (A chart which I nominate for the ASI Graphical Hall of Fame).
Which leads to a vision of the near future of the ASI showing accelerated melt to the same weather conditions and energy inputs of previous years, and even more so as continued cumulative global warming, exacerbated by Arctic amplification, increases energy inputs into melt seasons and reduces winter refreeze potential (and greater potential for Arctic cyclones, and jet stream weakening to allow warm air mass incursions, etc.).
If so, the drop from 4 million km2 September Extent to 3 million could occur in a shorter time frame than the observed trend for the drop from 5 million to 4 million. And with average ice thickness in late summer approaching 1 meter, a nonlinear melt response for thinner ice would accelerate even more for the drop from 3 million to 2 million km2, and even more than that for the drop from 2 million to 1 million km2.
(I suspect that dropping below 1 million km2 would complicate things because that final ice has resistance due to protection within bays etc. that would compensate for a thin ice melting effect).
By extrapolation, the linear Extent decline trend reaches zero decades later than the Volume trend. But of course that is impossible, because when there is no Volume, there is no ice left to create Exent. So the Extent trend has to eventually start accelerating to curve downward to catch up with Volume by the date when they both reach zero. I think that thin ice melt acceleration will be a major contributing factor (along with mobiillty for export, fracturing, surface area and possible others), that will cause that to happen.
Is there a fallacy in this line of thinking? What alternative mechanism accounts for the required unification of Extent and Volume as they approach zero. Binntho I'm talking to you! This is right up your alley and I haven't seen you post for a while.
One more conjecture. I think that as the average thickness in the High Arctic Seas, as shown in gerontocrat's graph, is approaching 1 meter in September, the accelerated thin ice melt effect, which might have been relatively inconsequential until now, will become an increasingly important influence. As a result, there will be "Extent goes poof" events of increasing scale and frequency over the next 10 years, resulting in a BOE by the early 2030s if not before.