Just a short post. In previous post, I lamented whether I was the only one who wondered about the trend change of Arctic sea ice and trying to visualize it. Doing an ultimate search using several search engines, I finally found a graph that visualized that same trend change, although in a bit different way. Surprisingly enough, I found it on the NSIDC website (scroll down until “September 2019 compared to previous years”):
Although the graph is similar to the ones that I made, there are some differences. The NSIDC uses the September monthly mean extent values, while I use the values of the day of the year with the smallest extent. September is the month with the lowest extent, so that is close enough. Another difference is that their graph only uses straight lines, while I tried to fit the trend with a curved line. This because when I look at the Arctic sea ice extent and volume data, then I have the impression that this process is definitely not linear, probably polynomial or it might even be sinusoidal.
It is of course possible to fit the curve with a straight line as the NSIDC did (the thin dark blue dashed line in the NSIDC graph), but this doesn’t fit as snugly as the polynomial trend line that I used.
What the NSIDC does different in regard to their other graphs is that they compare the linear trend of the period 2000-2012 with the linear trend of the period 2007-2019. This shows that the former trend is steeply down and the latter parallel to the x-axis (there wasn’t a trend up or down in the 2007-2019 period).
Unfortunately, this NSIDC graph seems to be a one-off. It was made as part of the analysis of the month September 2019 and as far as I can find, it is not repeated, now buried under several years of more recent analyses.
Trust, yet verify 