ARCTIC SEA ICE –Brief thoughts on DMI “shoreline masking” problem–

Over the past couple years the Danes have tried to make their measurements of sea-ice more precise, which is admirable. I have always tended to give them the benefit of the doubt, in issues regarding sea-ice, because sea-ice is serious stuff to them, as they have to deal with it. Their fishermen deal with it and are an important part of their economy, and they have a close association with Greenland, which is awash with sea-ice. But their simple attempt to be more precise has made an amazing mess.

What was involved was a so-called “mask” which runs along shorelines, and in effect disqualifies shorelines from sea-ice measurements. There are many reasons for doing this, two of which I saw when living on the coast of Maine during bitter cold winters.

1.) Big tides can make a mudflat “open water” one mild morning, but “ice-covered” the next.

2.) Big winds can shove the ice to the coast in the morning and then pull the ice away when winds shift in the afternoon.

It would take a whole staff of observers simply to measure the ups and downs along coasts. It is far, far simpler (and cheaper) to skip the bother by skipping the coasts. Skipping-the-bother was called a “mask”, and ran along all the convoluted ins and outs of the arctic shores.

The change was to make the “mask” thinner. This immediately added ice to the amount counted, because sea-ice that wasn’t counted before was now counted.

One wonders how much ice such a small change in the “mask” could be? Surely it would be a tiny amount, would it not? I myself don’t know, but have heard that 1.4 million sq km of ice was added to the amount counted, which is not a tiny amount.

Adding so much ice to the total made the old DMI “30%” graph, which wasn’t adjusted, read too high, even as the new “15%” graph was adjusted, and read lower. This disagreement put DMI in an awkward position, where their own graphs disagreed hugely, and their response was to simply discontinue the old “30%” graphs. They simply could not spend the time to fix the problem in the old graphs.

They did have to spend the time to fix the problem with the new gtaphs, because you cannot have the sea-ice “extent” jump by 1.4 million and not expect an uproar. However to simply subtract 1.4 million is too simple an answer. Why not? I made an attempt to explain why things are not so simple, on another site, as follows:

“1.4 Million sq km is a nice figure for winter sea-ice, but even in the winter there are problems with it, because when the weather patterns are meridienal even in the winter the ice may not be snug with the shore, as polynyas form even when temperatures are far below zero in places like the top of Baffin Bay or the coasts of Siberia, when off-shore winds roar.

If there is no ice at all along the shore, it doesn’t matter if you mask more or mask less; zero ice is zero ice. Any program that assumes you need to add ice along that shore, due to less masking, is an adjustment that will need adjusting.

Now May has past, and we are no longer talking about winter sea-ice. We are talking about June sea-ice, and there will be a lot more open water along the shores. It matters less and less whether you halve or double the masking along the shores, because there is no ice in the area you are masking.

In the winter such a potential glitch is minor, as polynyas swiftly skim over with thin layers of ice. In the spring polynyas stop freezing over, and appear downwind of every shoreline. (For example, downwind of Wrangle Island.)

I very much doubt that the fellows at DMI go on blithely adding 1.4 million sq km of ice as a “masking adjustment” right through the summer. After all, they aren’t nincompoops. What I mean to suggest is that the jobs becomes more difficult, due to the need for the adjusted adjustment.”

This is only a guess on my part, but I wonder if such an “adjusted adjustment” may explain why the amount of sea-ice abruptly grew in these DMI maps, between May 24 and May 25. (Hat tip to Svend Ferdinandsen.) (May 24 is to the left, and May 25 to the right.)

May 24 satcon.arc.d-05 May 25 satcon.arc.d-04

(The best way to compare the above maps is to open them in new tabs, and then switch back and forth between them. You will notice an increase in ice that weather conditions can’t explain.)

Often I notice maps show sudden appearances or disappearances of ice, such as the one I point out above. My assumption always is that the fellows behind the scenes are dealing with some difficult problem, and have had to “tweak” the way a program handles data.

It would be best if we could deal with such problems like super-engineers, beforehand, and not later, when they explode in our face. However even super-engineers know of something called “Murphey’s Law”,  (also expressed in a famous Scottish poem as (roughly) “The best laid plans of mice and men often go astray.”) All too often we have to deal with problems we never imagined would occur.

Who would have guessed simply reducing the “mask” along the arctic shoreline could cause such problems? (Likely there was some old Danish grouch warning the young idealists, but he was deemed a “wet blanket”, a “wrench in the works” , a “fossil”, and about fifteen other disparaging things I can think of, and then “fools rushed in, where wise men fear to go”. To which I reply, “Respect your elders” [especially me.])

In any case, it probably seemed a good thing initially to reduce the “mask” along the arctic shoreline by 50%. Is it not good to be more precise? Now the person or persons are facing “unintended consequences”, because people are noticing odd illogical happenings occurring in maps, and are being very rude and quoting Shakespeare and saying “something stinks in Denmark.”  This is unsympathetic and unfair.

Often, when you simply ask, or simply point out ice has vanished and/or appeared, and do so to the scientists involved with a polite email, you will receive a polite and sometimes quite lengthy and detailed explanation.

We should not stop asking questions. We especially shouldn’t automatically deride either the questioned (in this case DMI) or the questioner (in this case me.)

These things happen.


Hat-tip to the Blogger “Craig T” who produced this map over at . It does an excellent job of showing how reducing the “mask”(deep blue along the shores)  creates the impression there is more ice. (Green is ice that exists this year that didn’t exist on the same date last year.)

13 thoughts on “ARCTIC SEA ICE –Brief thoughts on DMI “shoreline masking” problem–

  1. Hi Caleb, I am enjoying reading your interesting posts, keep the good work on!
    I am a Hungarian fan of you and an amateur climate watcher. (sorry for my English)

    I’ve downloaded from the OSISAF page the HDF files of the Sea Ice Edge for 2016 may 30, and 2015 may 30. In the description of the Sea Ice Edge files I found this for the values of the 10×10 km cells: ” 1 – No ice (less than 30% ice concentration), 2 – Open ice (30-70% ice concentration), 3 – Closed ice (more than 70% ice concentration)”
    So, I counted the cells containing 2 or 3 and I have these results:
    2015 may 30: 86448 cells = 8,644,800 km2
    2016 may 30.: 88253 cells = 8,825,300 km2
    Here is my post, describing in details the method, but sorry, it is only in Hungarian:

    • Thanks for doing all that work. Judging from your results, there are 180,500 km2 extra this year. If one automatically subtracted 1,400,000 km2 as a “masking adjustment ” we would have 1,219,500 km2 less ice this year. My guess is that is where the problem arises, but that is only a guess.

      When I was only three years old the Hungarian revolution occurred, and my parents opened our house to refugees, and my mother taught them English. I actually spoke a little Hungarian as a small boy. All I can recall now is that “egan” meant “yes” and “nem sabad” meant something like “don’t do that.” I also remember they were full of joy, because they had escaped StalinKhrushchev. Some of the stories of how they escaped were amazing.

      In one story a woman wanted out because the Russians had killed her husband, and she was heading for the Yugoslav border through deep snow with her 15-year-old son. Just as they got to the border the Russian soldiers appeared in the distance behind them. The Russian soldiers raised their guns to shoot, but the Yugoslav border guards raised their guns and aimed at the Russians. The Russians lowered their guns, and the mother and child escaped.

      Whew! I tell you, those folk really appreciated freedom. They also spoiled me. That woman made a chocolate covered pastry called, “devil’s pills.”

      • This is an interesting and heart-stirring story, our history is full of such dramas, my grandfather was held 4 years in captivity by the Russians after the W.W.II., but he managed to return on foot to Transylvania (part of Hungary then, now part of Romania) and had a hard but beautiful live.

        Returning to the DMI chart, I’ve downloaded the data for 31th December 2015 to verify my calculating method, and I’ve obtained 10,325,800 km2, which is confirmed by the discontinued chart:

      • I looked at your article, using the “translate key” of my computer. (It would not translate the word “jégkiterjedés”.)

        I am not smart enough to download data. If I could download, I would try to find the day they first started using the “new masking” to see if the graph jumped up.

      • I remembered a story the refugees told about Transylvania.

        One of the refugees was royalty from the old Austro Hungarian Empire, (Count Teleki) and they had a castle in Romania. At the end of World War Two they went back to the castle, and the local people would not go near it, as they said it was haunted. The Russians had used it when they invaded and it was full of moaning and groaning. It turned out the Russians had left wounded soldiers in the castle, and they had died there.

        Many of the poorer Russian soldiers had never seen running water before. They would rip the faucets off the wall, so they could have running water when they got back to Russia.

        It must have been a very crazy time to live through.

      • Hi Caleb,
        I apologize in advance for being picky, especially with such a heartwarming post, but Stalin died in 1953 and was not around for the Hungarian Revolution, which was in 1956. Nevertheless, the person who was running the show at the time for the USSR (Nikita Khrushchev) was not that much better than Stalin.

      • By the Hammer of Thor man, you said “I was 3 when the Hungarian revolution occurred” and my brain immediately jumped to the 1848 one (hadn’t seen as much about the 1956 one) and thought “damn, you’re looking good for someone over a hundred and fifty years old!” before I googled and found the later one that makes more sense.

        Also made this from your comparo pics for ya:

  2. Caleb. I don’t agree with your view of climate change but I still enjoy reading your observations and thoughts, and not only on the Arctic ice. It’s interesting to see another perspective on life, and observe, whether or not in the end, with the ice, something hugely important is going to happen.
    There isn’t much you can be totaly sure about in this world.


  3. “jégkiterjedés” means Sea Ice Extent.

    Here is how to download and analyze the data:
    – you can download the HDF files from here:
    – to open an HDF file you have to use the HDFView software, you can download it from
    – open the HDF file with the HDFViewer
    – select all the cells (selecting one and hit CTRL-A)
    – copy and paste into an excel sheet.
    – in excel do a count for the cells with value 2 and another count for cells having the value 3.

    • Thanks. Even if I can’t find the time to do this other people who visit my site will.

      I mentioned your site over at so you might get some visitors. Hopefully none will be rude. (Probably no one will know how to be rude in Hungarian.)

      Thanks for visiting. I’ve enjoyed the cross-Atlantic talk. Now I need to rush off to work.

      • You’re wellcome!
        I’ve enjoyed too this talk, and I need to rush off to bed now. I’ts allmost midnight here.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.