Something odd is going on at the National Snow and Ice Data Center.
Look at this image:
The image is directly from NSIDC’s Artic Sea Ice News page today. Of course there’s the large drop of about 1 million sqkm of sea ice in the last couple of days that is puzzling.
If this were real, we’d also expect to see something also on Cryosphere today plots, and while that group does not do an extent graph, they do make an areal graph. It “should” show something that reflects the drop but instead goes up. WUWT?
While ice extent and area are not exactly the same, they are closely related. So one would expect to see at least some correlation. But we have zero. I suppose there could be a wind issue that is compacting sea ice, but surely there would be something in the area graph.
Something seems not right, and NSIDC owes the public an explanation as they did for a previous drop in extent change from January 15 to 26 which is currently in their Feb 3rd news release.
h/t to Joe D’Aleo and many WUWT commenters.
@ur momisugly Pamela Gray (12:24:05) :
Do you have some data about ice thickness from different places?
Walt Meier (12:06:13)
I’m not sure why you think things like this are worth blogging about. Data is not perfect, especially near real-time data. That’s not news.
Thanks for the update Walt, but I think you must know perfectly well why things like this are worth blogging about. The health of our northern ice pack has attracted alot of attention in recent years, so there are alot of people taking a real keen interest in both how it forms during winter and how it recedes during the summer. Furthermore, as others have pointed out, why could someone from NSIDC not have discovered the problem with the data before Joe public?
While you’re there I would like to pose a couple of questions regarding your “News and Analysis”. Why are all the bulletins negative? While I understand the need show possibly disturbing trends, surely you might have made a headline for example out of the speedy recovery observed last Autumn? Also, on the subject of manual adjustments, I have noticed that during the winter, sudden gains are often adjusted down the way where as during the summer, sudden losses are NEVER smoothed back up the way. If the sudden spikes are indeed due to “bad data” and manual adjustments are required, why do those adjustments always pull down the way and not up? Surely there have been some intances were the bad data has represented (for example during the summer melt), too much loss or (during the winter), not enough gain?
During the month of February, the Ice Extent MONTHLY trend charts for Antarctica began swinging wildly.
http://nsidc.org/data/seaice_index/index.html
I did an overlay of a previous chart from December to compare year by year values only to find that the % change from their 1979-2000 “mean” were not the same at all. I emailed to them a few times complaining that, first of all, using 1979-2000 as the average (“mean”) basically throws out about 30% of the available data (what is wrong with calculating the mean with ALL data from 1979-2008?). Somehow, they think 1979-2000 was somehow “normal”? They basically said no to using all the years for calculating the mean.
Then, when I told them the trend graph overlays from month to month showed different departures from the average (mean) on any particular year…they countered with:
—–Original Message—–
From: nsidc@nsidc.org [mailto:nsidc@nsidc.org]
Sent: Monday, February 9, 2009 03:01 PM
Subject: {nsidc-178955} Re: Antarctic Ice Extent Trend Graph Is Messed Up
Dear Mr. Strong, Thank you for contacting NSIDC. It appears that you are trying to overlay December with February. The points in the December graph are specific to data in December, and points in the January graph are specific to data in January. Overlaying these graphs is not recommended. You may see some differences in time with the graphs because the passive microwave satellite data we use became operational in October 1978. So Oct, Nov, and Dec graphs may show 1978, whereas the other months start in 1979. The scales are not equal between months because they are scaled dynamically to fit the data being plotted. January has higher values, so the scale now goes from -20 to 40. If you wish to do intercomparison of differnet months, we recommend you the extent values in the .txt files we make available at ftp://sidads.colorado.edu/DATASETS/NOAA/G02135. Let me know if you have any further questions. Best regards, Donna Scott NSIDC
But I sent them the overlay I did, they must have never looked at it. I used Photoshop to stretch and perfectly match the scales on each. i.e., I matched the values on both the X and Y to the correct values: each year lined up and well as the Y-scale for % change from the mean. The chart overlay was wildly different for each successive year for each plot date.
Now, if you look at the February Antarctic trend plots today, they have swings as much as 30% from “normal” .
I also took a look at their data sets. They are missing averages in some years: i.e. 1987 data is immediately followed by 1989 data with an entire month missing. This happens 3 times Nov, Dec 1987 and Feb 1988
Someone is being sloppy over there ad the NSIDC.
In industry we would put guard bands to signal a sudden shift like this to draw attention to a possibly false reading.
If the system we are measuring has a normal distribution we may apply SPC rules to further improve detection of a problem.
They changed it back, it looks more ‘normal’ today. I suspect human error and we will not hear about it from them.
Walt Meier (12:06:13) :
“The QC is partly manual.”
Walt, you need to get a new person to do the manual part of the QC if the current person cannot spot an error that big. Surely someone with a bit of intelligence actually looks at the graph before it gets posted.
I’ve watched time-sequence sea ice videos on igloo, and there is a lot of gap flicker, with occasional huge triangular… underlaps? Aside from that, ice extent isn’t a straightforward measurement. I’m not greatly surprised there are some over-shoot/under-shoot issues. Thanks to Walt Meier for the response. Yes, the overall reaction here was a bit over the top. It could be worse, Walt; close scrutiny is better than being ignored. People here care about your output. Thanks.
With the Poles getting colder and ocean currents speeding up, I would expect the Arctic ice to have trouble expanding into areas where heat is being released from the North Pacific and Atlantic oceans. An increase in ocean current velocity should also melt ice more quickly when Summer arrives at the Pole which seems to be occurring in the Antarctic http://nsidc.org/data/seaice_index/images/daily_images/S_timeseries.png. I would expect the Antarctic to replace ice quickly next month and exceed last years area in the coming Southern Hemisphere winter. If NSIDC and others try to manipulate the data they pass on to the public, I think that shortly the reality of a cooling earth will be impossible for anyone to hide.
Carsten Arnholm, Norway @ur momisugly 11:37:54
Interesting. Read it carefully, he’s sying that the Little Ice Age was a local European event, and not global in extent. I think that is the standard fall-back position of AGWers when preented with reality.
But, I’m out to buy a copy of S&T.
The presence or absence of a little annual ice in the Sea of Okhotsk in February doesn’t have any meaningful impact on the summer minimum of perennial Arctic Basin ice in September. The Arctic Basin has been saturated with ice for months, so I am not clear why NSIDC keeps highlighting the minor deviations in Northern extent in their monthly sea ice press releases.
Drift is much more important, because it does affect the summer minimum.
http://iabp.apl.washington.edu/maps_daily_track-map.html
Lee Kington (12:53:34)
We should have alerted the BBC of another catastrophic climate event.
On the one hand, some of us fret that certain agencies “manage” the data before releasing it to outside scrutiny. On the other hand we howl when they release bad data that would eventually require some “managing”. I don’t think we can have it both ways.
I personally prefer the “see it now and manage it later” approach NISIDC apparently took in this particular instance.
I am not a scientists, so I may be way off base, but could the decrease in NSIDC Arctic Sea Ice have anything to do with the Sudden Stratospheric Warming in January and the split in the Polar Highs.
It will be interesting to see if any true believers or mainstream media pick up on the NISIDC erroneous ice extent data and provide more AGW gloom and doom polemics. I can’t wait for tomorrow at WUTW.com
DanD (12:30:25) : Data ARE, Walt. Data ARE.
No Walt data is; data is an English word. English includes many words originally press-ganged from Latin, which have changed their grammatical type.
As has been pointed out far more eloquently than I can:
“The majority of writers who would dutifully pluralise `data’ in writing naturally and consistently use it as a mass noun in conversation: they ask how much data an instrument produces, not how many; they talk of how data is archived, not how they are archived; they talk of less data rather than fewer; and they talk of data with units, saying they have a megabyte of data, or 10 CDs, or three nights, and never saying `I have 1000 data’ and expecting to be understood.
If challenged, they will respond that `data is a Latin plural’. Agree to this, for the sake of professional harmony, and carry on the conversation, making sure to mention that `the telescope has data many odd images tonight’ (it’s a past participle after all), suggest looking at the data raw images (…or an adjective) and that you both examine the datorum variance (surely they recall the genitive plural); suggest they give you the datis (…the dative), so that you can redo the analysis with their datis (…and the ablative). If they object ask them to explain their sentimental attachment to the nominative plural, that they would use that in all cases, in brute defiance of good Latin grammar.
Isn’t it lucky English is now genderless, making `data’ neuter, else we’d have to memorise masculine dati (dati dati datos datorum datis datis) and feminine datae, too? Isn’t it simpler just to speak and write English?”
It’s as bad as affecting “an historical” and pronouncing the ‘h’.
Walt Meier
Thank for you prompt response. It is worth blogging about because many of us crunch multiple data sets. If there appears to be an error it is prudent to sound out and see if what may be a error is indeed that. It was an error and nothing more. No harm was done and your feed back has put the matter to rest.
I was always told to simply substitute the word ‘figures’ for ‘data’ if you wanted to explain the word. Not sure if that’s correct or if it helps. Let us not forget though that English is one of very few languages that actually doesn’t have rules set down. People forget that, but it’s a fact. You may see ‘how to say/pronounce/write’ etc. in a book on grammar, but it’s just the author’s guide – it isn’t law. English came about on the very fact that it adapts and changes. Only two decades ago people wrote “developement”, which has changed to “development”. And it’s perfectly okay to split an infinitive! I’ll take anyone outside who says it isn’t.
PeterW,
nominative singular: datum
genitive singular: dati
dative singular: dato
accusative singular: datum
ablative singular: dato
nominative plural: data
genitive plural: datorum
dative plural: datis
accusative plural: data
ablative plural: datis
…Catholic school made me do it!
Smokey-
Looks like you were forced to wear an ablative habit and neuter some nouns with a dull rhetorical knife.
Maybe someone brought this up in one of the nearly 100 comments but have a look at this on cryosphere.
Look at the Sea of Okhotsk. It looks like the ice disappeared there overnight in the graphics.
Jørgen F. (13:04:06) :
“Walt Meier (12:06:13) :
I’m not sure why you think things like this are worth blogging about. Data is not perfect, especially near real-time data. That’s not news.”
Oh no – it’s just a graph used for world politics, reallocation of billions of tax dollars/euros, closure of power plants etc. – who cares if you mess up data some times.
Don’t exaggerate, as they point out on the site the monthly data are checked and QC’d whereas the daily data are subject to errors. The NSIDC data are the only ones that I follow that use the SSM/I sensor rather than the ASMR-E (the CT comparison data do too, which often gets some all agitated on here). Missing swathes occur on both which is why having the two systems is a good crosscheck.
“I’m not sure why you think things like this are worth blogging about. Data is not perfect, especially near real-time data. That’s not news.”
Maybe it’s not news but could it fit in one of these other categories?
“…puzzling things in life, nature, science, weather, climate change, (or)technology…”
I think a QC problem at NSIDC fits quite well in all of those, includings news for that matter.
This is a blog. Anything whatsoever that the writer of the blog wants to post is perfectly alright.
How much more gracious would it have been to say, “Thanks, Anthony, for bringing this to our attention. It will be fixed shortly.”?
Now Dr. Meiers curt answer is likely to be the subject of a blog out there somewhere.
Why is it that the mistakes always seem to under-estimate ice extent (or over-estimate temperature)?
Would it be that results from mistakes in the opposite direction are minutely scrutinised and corrected prior to publication; whereas the error-stricken results picked up on this site, by Climate Audit and others, have received no such scrutiny in the unseemly haste to make them public?
Moderator, is this one-liner to far off-topic?
“If we can recall bad peanut butter, why can’t we also recall a Nobel Prize?”
I contacted Ron Goodson at Environment Canada, Edmonton office. He assured me that there has been no break-up what-so-ever of Hudson Bay ice. He looked at the 2-16-2009 image posted as NSIDC and suggested that there algorithm for translation of satellite image date to ice coverage may be at fault. There has been some low cloud coverage in the Hudson region.