[Update : Actually, I just forgot to use my brain. This post is completely bogus – as explained here]
All of the turquoise on the right edge of the September graph should have changed to green over the winter (minus any 1-2 ice which was disproportionately transported out of the Arctic.) NSIDC changed their percentage scheme from 30% concentration to 15% concentration over the winter, but that won’t account for this.
Polar drift doesn’t selectively target 1-2 year old ice, so I think something is wrong. The amount on the right edge which changed from turquoise to green looks too small to me.