The trend for the gain in each of the three months is similar generally upward. Looking at the last few years and saying the trend is downwards is highly likely to be noise giving a spurious result. (You can see the large amounts of noise in this graph.) Last couple of years having higher minimum against the trend also makes looking at just a few comparisons to deduce the trend dangerous.
Not sure I agree with you there, if we are to let the numbers themselves decide.
There seems to be a sea change at or just before 2007, from which year the first half of the X to 2014 years will have more rapid ice gain than the second half. That looks like a trend to me.
That means November 2007–2010 will have more ice growth than November 2011–2014, 2008–2010 more than 2011–2014, 2009–2011 more growth than 2012–2014, 2010–2012 more than 2013–14 and finally 2011–2012 more rapid ice gain than 2013–2014.
If you include more than 8 years, the trend is the opposite (last half of the selection will have quicker growth).
The graph I posted earlier with discussion of main movement being downwards but also a wider bottom and steeper sides gives much better understanding of the situation.
Just because the trends are similar so far does not stop us understanding that it is October that is going to turn downwards first and it might not take much lower minimums to cause that to occur.
This doesn't seem to be based in the data, Crandles. The trend is for faster melt in May and June, slower melt in July, August and September, continuing in faster refreeze in October, turning to slower refreeze in November, then back again to faster refreeze in December and April (with January to March in the 'Undecided' category).
My guess is that January to March will be the first to tip over into predominantly slower growth than the earlier years in the data, while October will keep growing ever faster until the summer crash expands into October 1st. Data suggest August will crash before October, though.