I overlaid today’s DMI data on the NSIDC graph from yesterday morning. NSIDC should be almost brushing up against the mean line, for the first time in more than six years. However, NSIDC made a change yesterday during the day which pushed the ice extent to the left, and the mean off to the right – and thus created a large gap.
Instead of brushing up against the mean line, today’s data will look like this.
Walt Meier at NSIDC said this
“A key point is that this change doesn’t actually change the data at all”
It most certainly does change the data. Crossing above the mean line would be symbolically a very important indicator, given that five years ago people at NASA predicted a possible ice free Arctic in 2012. I am having a very tough time believing that the timing of this is an accident. I have never seen NSIDC change the graph mid-day before. It normally updates at 8am MDT and stays that way the rest of the day.
This whole thing makes them look incompetent.
So the test system got accidentally put into production? But he’s not sure and has to confirm? Is anybody in charge?
…and now they’re going from a method that uses extrapolation for the endpoint, to one that might use interpolation when data is missing?
Why not just plot the data they have, when they have it? And don’t plot the data they don’t have, when they don’t have it (rather than make up data that will have to be revised later).
I don’t know what others think but I’d rather a centered moving average, but not extended to the last data point so there’s no endpoint issues.
This is a trivial exercise in presenting data straightforwardly and they’ve stuffed it up, it should have been sorted out 30 years ago.
James Hansen said how things would evolve from now on : thinking that the sceptics were winning the case by means of excellent public relations, the warmists are going to change all the numbers to “prove” their superstition, and claim in the MSM that AGW is REAL, and DANGEROUS. Even if nobody sees it. That ensures the continuous flow of tax money in the politicians’ pockets, without forgetting to fill heir own.
Wouldn’t it be fun if the halt of the sea ice melt continues for a few weeks? I wonder what they’ll do then?
I love the idea of not changing the data but changing the average. How about we challenge them to show that the older data produces the new average data, as they have shifted it.
In other words, when they shift the average, does this mean that they have gone through and altered the older data to produce this new, shifted average?
I suspect what they’ve done is just shift their entire spreadsheet by two days and re-generated the graphs. However, that means the mean data gets shifted twice – once because the underlying data gets shifted forward, and then again because they applied the correction to the whole spreadsheet. A simple email to them pointing out politely that the mean’s been shifted by four days instead of two should get it fixed.
The reason the most recent data point vanished is presumably because that data wasn’t actually in and validated yet (i.e. the graph was previously being calculated from a projection rather than from real data). It’ll get filled in in due course – we’re just going to have to get used to there being a bit more of a delay on the NSIDC graph from now on.
Wow thanks for clearing it up. What a relief to know it is all our misunderstanding of their data managment procedures.
Here is their article: http://nsidc.org/arcticseaicenews/2012/04/daily-graph-changes/
I think it is interesting as well that with the change the 2007 line now no longer touches the standard deviation. The old version had 2007 reaching standard deviation several times in May (fluctuated back and forth, but clearly hitting it as the animated gif shows).
So not only did they make this year’s line farther from the average, but now 2007 never even reached within the standard deviation. If the data didn’t change, how is it that 2007 did reach standard deviation, but now does not?