Second lowest Arctic sea ice minimum on the 10th or was that the 7th?

[xmetman]

The Arctic sea ice looks to have reached its minimum on the 7th September, which is four days earlier than average. The sea ice extent bottomed out at 4.083 million square kilometres making it the second lowest since records started in 1978 - well that's according to the data file that I've just downloaded!

Strangely, according to the data that I download from the National Snow and Ice Data Center [NSIDC] the minimum occurred three days later on September 10th. As I said in my introduction on the 7th the value was 4.083, but according to the news item that I've included below, the value on the 10th was 4.14 million square kilometres and tied it with the year 2007, which according to the data file is third.

Here's how I see the latest annual Arctic sea ice extent rankings using the latest data that I've just downloaded ending on the 15th of September.

And here's a snapshot of what the data file looks like that I download just to prove to myself I'm not going mad.

All these daily values translate into the following chart with the minimum occurring on the seventh and not the tenth of September.

I have monitored polar sea ice using data from the NSIDC for almost five years now and never found much problem until June of this year when the NSIDC transitioned to using data from the DMSP F-18 satellite due to issues with the F-17 satellite. I'm now totally confused, I realise that all the differences are very trivial when you are dealing with millions of square kilometers of ice, but I still wonder where I'm going wrong, If you know then please drop me a line!

[Alastair B. McDonald]

There was a problems with one of the satellites in the spring.
http://moregrumbinescience.blogspot.co.uk/2016/04/2016-tough-on-sea-ice-satellites.html
I wonder if that has caused a problem.

[xmetman]

Alistair

It did and the data disappeared for a while, and then they came up with version v2 of the format for the new satellite and that's the data I download.

ftp://sidads.colorado.edu/DATASETS/NOAA/G02135/north/daily/data/

Bruce.

[John Hall]

Interesting data. I was confused by the bottom row in the table for a moment, until I realised that the data for 1978 probably only began on 26th October (1978 being the first year of the record).

[Alastair B. McDonald]

I've been checking for a minimum on this map:
http://nsidc.org/arcticseaicenews/charctic-interactive-sea-ice-graph/

I get 1.370 on 10th.

I suspect that the 7th was just finger trouble. It does happen!

[Alastair B. McDonald]

 I should have written

I get 4.137 on the 10th

Is it that your data is from NOAA rather than NSIDC?

[xmetman]

Alistair

I've triple checked that!

The link from the NSIDC from the "Get Data" button on the web page below takes you to that same FTP folder and the same file with the minimum occurring on the 7th (4.083)  and not the 10th (4.16). Your value of 4.137 is from the 9th in this data!

The reference to NOAA is just a folder name in the FTP structure.

Bruce.

[xmetman]

Someone (John Dann on UKSW) has kindly pointed out to me that the NSIDC now use a trailing five-day mean to construct all their graphs and ranked tables of the extent and time that minima and maxima sea ice extent occurred. So it is the same data that I download and use it’s just smoothed. Here’ a further snippet from their website that explains matters.

It looks like they adopted this new way of working in June of this year,  because a footnote further down that same news item goes on to say:-

In my opinion using a running mean has only muddied the waters regarding sea ice extent, should they have used a centred or a leading moving mean rather than a trailing mean when referring to maximum or minimum extents? Why not just accept the daily values warts and all, if they're good enough to construct a five-day running mean why not just continue using the daily values? The daily minimum and maximum will most likely never coincide with the minimum or maximum gleaned from the five-day trailing mean, which to me is just plain confusing.