It takes a long time for a mass of deep frozen ice to warm back up to the melt temperature. While this is happening, melting is modest. But once all the ice is potential melt (i.e. 32 degrees) then it gallops. On top of that the ice is steadily thinning so that the available mass is declining in a non linear manner.
For example, if the removal rate is 5% of the original mass M then we have the following effect:
year 1 .95M
year 2 .90M equals 5.26%
year 3 .85M equals 5.55%
year 4 .80M equals 5.88%
year 5 .75M equals 6.25%
year 6 .70M equals 6.66%
year 7 .65M equals 7.14%
You see were this is going. The point is that the surface area has dropped from 5m to 3m in the past thirty years and the actual mass has dropped by at least 60%. The truth is that this has mostly happened over the last half of that time period. This suggests that the removal rate has been around 3% of the original maximum ice cover and it is now in the last stages of removal and collapse.
I wrote three years ago that I would be brave and say that the sea ice would be flushed out by 2015 when everyone was saying 2100. I am beginning to think that I was too conservative. We need a cold spell up there.