Problems with satellite sea ice monitoring confirmed by NSIDC – no timeline for fix

From NSIDC and the “worth blogging about” department

Sensor on F-17 experiencing difficulties, sea ice time series temporarily suspended

NSIDC has suspended daily sea ice extent updates until further notice, due to issues with the satellite data used to produce these images. The vertically polarized 37 GHz channel (37V) of the Special Sensor Microwave Imager and Sounder (SSMIS) on the Defense Meteorological Satellite Program (DMSP) F-17 satellite that provides passive microwave brightness temperatures is providing spurious data. The 37V channel is one of the inputs to the sea ice retrieval algorithms, so this is resulting in erroneous estimates of sea ice concentration and extent. The problem was initially seen in data for April 5 and all data since then are unreliable, so we have chosen to remove all of April from NSIDC’s archive.

It is unknown at this time if or when the problem with F-17 can be fixed. In the event that the sensor has permanently failed, NSIDC is working to transition to either the DMSP F-18 or possibly the JAXA Advanced Microwave Scanning Radiometer 2 (AMSR2) on the Global Change Observation Mission – Water (GCOM-W) satellite. Transitioning to a different satellite will require a careful calibration against the F-17 data to ensure consistency over the long-term time series. While this transition is of high priority, NSIDC has no firm timeline on when it will be able to resume providing the sea ice time series. For background information on the challenges of using data in near-real-time, see the ASINA FAQ, “Do your data undergo quality control?


Here’s what the failure looks like:

arctic-nsidc-sat-uptick antarctic-nsidc-glitch-DMSP17

Interestingly, the failure in 2009 and this one both came to the attention of NSIDC on a Monday morning.

69 thoughts on “Problems with satellite sea ice monitoring confirmed by NSIDC – no timeline for fix

  1. Some liberal was trying to make “adjustments” to the data and they got it backwards ? OOPS !
    [no, your’e dead wrong – it’s a simple instrument failure -Anthony]

  2. Anthony, quick question: Every time there’s more than 100 comments below a blog post, I can no longer see the comments (although I do receive them in my inbox). Is that a browser issue on my side, or a WordPress thing?

    On-topic: It really is a shame the SSM/I broke down just now, as “an early melt event over the Greenland ice sheet occurred this week, smashing by a month the previous records of more than 10% of the ice sheet melting” (source: Polar Portal) because of an early ‘heat wave’. It would’ve been interesting to see how this melting event stacks up in the NSIDC data set, which differs from the DMI’s.
    Let’s hope that NSIDC can quickly resume daily reporting of sea ice concentration and extent numbers. And let’s hope that this failure doesn’t threaten the long-term time series, or else we’ll only have the ASMR-2 sensor to provide passive microwave data (and I remember how it sucked when its predecessor AMSR-E failed).

    • I think it might be a browser issue, or perhaps not enough memory to render everything.
      I’ve found older browser versions don’t do so well with lots of comments that cause a very long scrolled page.

    • Nevena, if you are using Windows10 ……. then me thinks that is your problem.
      The next time your CP stalls when viewing WUWT, …… check your Task Manager activity of Internet Explorer.

  3. Is there no ongoing effort to reconcile the various satellite services?
    There are ice monitoring satellites from Russia, Japan and the EU, plus maybe from China.
    On the one hand I’m happy that there are independent data sources being compiled, yet it is frustrating to have an interruption to something this basic. Seems like a project for some small government such as Iceland.

  4. Boy oh boy, we need some competition to all the keepers of the data. The ice extent glitch provides another opportunity to fiddle the data. We will end up frozen down to Martha’s Vineyard with the graphics still showing the continuing death spiral in arctic ice. I note, too, that the BOM of Australia didn’t update their ENSO graph (red line graph) this Monday. It has been showing a free-fall toward La Nina for the last several months. I’m nervous about the suddent ‘stop’. The Karlization of temperatures are likely to embolden the data fudgers in other data sets. It won’t do to have the “Pause” disappeared if the ice continues to expand. It won’t do to have the equatorial Pacific cool off either.

    • This year winter seems to have started up again from the beginning of April in a LOT of places including about a third of the US, Parts of Canada, Russia, etc. If the Arctic resumed growing ice for a week in April, would we know it? The bits that are missing– that make the ice 7% below the mean are outside the main bowl. I guess there’s no followup by anybody w a camera to see what those areas by the Barentz Sea or the Bering look like, huh?

      • Penelope
        Yes the NH has certainly gone into climate cooling since the start of the month. With areas of high pressure pushing northwards up towards the Arctic and then sending down cold air on their eastern side.
        Looks like NE America will suffer a other blast of cold air on the 19/20th.

  5. “Here’s what the failure looks like:”
    There was no thought of failure when the extent declined near the peak, only when it bends upward.

  6. Simple issue, the satellite has flipped! Its upside down, arctic freezing, antarctic thawing!
    Or maybe the world has flipped and we just haven’t noticed?
    Good luck to whoever has to sort it out!
    Have fun!

  7. During the next 7 days looks like there is going to be a large movement air both in and out of the Arctic.
    Which will cause bursts of warming in the Arctic a little like the recent one in Greenland but not as strong. Which may well quicken the pace of the sea ice decline.

  8. The problem was initially seen in data for April 5 and all data since then are unreliable….
    Really?….so it couldn’t have been a slow degrading over weeks? months?…………..years??

    • Niet comrade!
      Please, this simple modelling aberration is under correction.
      Rest assured normal service will resume, once the clouds have been dispersed.
      Interesting to have two other similar aberrations during the south ocean max.
      Funny old stuff this floating water, seems to be a law onto its self?

      • Nick, great comparison. Can you expand that trace with unsmoothed data? Sometimes satellite glitches can be orbital. IE every so many hours between complete cycles. What would be of interest is anything that demonstrates a regular beat of some kind, but only raw data would show it.

      • Pamela,
        “Can you expand that trace with unsmoothed data? “
        I have graphed the numbers as published. I understand JAXA is smoothed prior to publication, over three days I think. And I think NSIDC is not.

      • Unfortunately, the NSIDC isn’t allowing access to daily data, even if limited to no more than this April 1.

      • Pamela,
        Sorry, I rather missed the point of your query. The link I gave has a radial plot going back several years. Choice of JAXA or NSIDC. I showed just the latest bit. Above it, there is a table of the numbers as published. If you scroll down, there is a rather extensive table of actual numbers published. They are in difference form – ie they show today’s extent in absolute and before that, daily differences. You can reconstitute the numbers if you like, but the differences generally are a more sensitive indicator.
        I’ve made it show NSIDC up to the fault period. It may revert tonight when it tries to read data again.

    • 37 years of observing sea ice from satellite and nobody ever thought of that? Well, no. There are other satellites, in the visible, which can be used to check. Plus fishermen who operate near the ice edge, plus polar natives who hunt at the ice edge, plus oil companies considering drilling, scientific research cruises, tourist cruises, the Navies of many countries, ….

  9. Forgive me if I’m wrong, but isn’t there a polarizing filter on this satellite. I do seem to remember adjusting this mechanism being responsible for upticks in one hemisphere and corresponding downticks in the other hemisphere in the ice extent graphs. Looks like they’ve tried to adjust it and the “stop” command isn’t working

  10. I’ve seen these “glitches” in the past. They were maybe 2 years ago, or longer ago. Who is to say these errors were not present 2-3 years ago?
    Well, as we used to say when I was a Navy Officer (Lt Jg – damage control assistant on an ammunition ship Wrangell AE-12, off the coast of Vietnam) in 1969, “that’s good enough for government work”.

    • From Wiki:
      Wrangell finished out the year with three operating-periods off the coast of Vietnam and in the Tonkin Gulf. The largest underway replenishments were conducted with such men-of-war as the battleship New Jersey (BB-62) and the carriers Constellation (CV-64) and Ranger (CV-61). In addition, she armed many smaller units — destroyers, guided missile destroyers, and Coast Guard cutters. One of the latter, USCGC Winnebago, sent her whaleboat across to Wrangell on 21 November while off the Mekong Delta in what Wrangell’s command history termed “what may well be history’s smallest rearming.” Winnebago’s request: five rounds of 5-inch ammunition!

  11. Given their dismal track record, it’s perfectly logical and rational to assume CAGW alarmists are conducting another round of data tampering, however, another possible explanation is that a strong solar storm that started in early April could have damaged some weather satellites:
    It just seems strange that both NINO 3.4 SST and sea-ice satellite data went out at exactly the same time during a strong solar storm.

  12. Sorry, hold on.
    It’s my understanding that the ERROR in a broken device is supposed to produce the SAME error at the same time. Why is it erroring out in the opposite direction of the active trend on both charts? That the magnitude of the error is the same but in opposite directions and compliant to the same curves implies a mathematical error in the processing of the data, not in the sensor itself.

    • The satellite does not observe sea ice concentration itself. Like any satellite, it observed the energy in some wavelength/frequency band(s). One then has to apply an algorithm to translate what the satellite does observe in to something you’re interested in. The NSIDC makes use of 19V, 19H, 22V, 37V, 37H channels, by way of the NASA Team 1 algorithm. (All frequencies in GHz, H and V mean horizontal and vertical polarization.)
      The algorithms for the Arctic and Antarctic are different. And at this time of year, the ice packs are quite different. The Arctic is warm and retreating, the Antarctic is cold and expanding. It’s hardly surprising that erroneous data are having different effects.

  13. Marcus
    April 13, 2016 at 2:10 pm
    “..Weird how they seem to be exactly opposite of each other ?”
    Yes this struck me as odd too. I find it a little difficult to see how an instrument failure could have almost equal and opposite effects on each hemisphere. Grab the data for later tampering verification.
    Whatever gets done I’ll bet that global sea ice will take a hit when it’s “fixed.”

  14. Special Sensor Microwave Imager and Sounder (SSMIS) on the Defense Meteorological Satellite Program (DMSP) F-17 ,
    The F-35 doesn’t work all that well either.

  15. “Transitioning to a different satellite will require a careful calibration against the F-17 data ”
    and will give us leeway to make it appear to be ‘worse than we thought’.

  16. So what’s the problem with the BOM (Oz) ENSO graph? It was plummeting downward to a La Nina when reporting their got “interrupted. I think we are being prepared for further karlization of all the cooling data. You can’t kill the pause and then leave growth of ice extent and cooling of the eastern Pacific, can you?

  17. Thanks, Samuel. I’m on Win7, with the latest version of Firefox and Chrome. If I’m the only person with this problem, it’s not that important.

  18. Conspiracy theorizing dutifully showing up in the comments, it seems apt to mention an old article of mine, How large would the sea ice conspiracy have to be?
    F-17 was launched November 2006, so these problems are showing as it approaches its 10th birthday. Given that the design life was 5 years, it’s doing awfully well for a geezer. The surprise is that it was still giving good data through to this month, and that it is still doing so on almost all its channels. Not that an elderly satellite is having some data quality issues.

  19. Glitches, errors, sensor fall-off are OK as long as the down trend is maintained. Data for the last three months is suspect; how long does it take to realize the data is atypical? Didn’t the last failure occur during winter? Much too convenient for my taste.

  20. The apparent source of the problem has been found, the attached is the announcement by NSIDC:
    “Near-Real-Time DMSP SSMIS Daily Polar Gridded Sea Ice Concentrations, Version 1
    Notice (04/05/2016): Notice (04/19/2016): Daily sea ice concentration updates have improved. On 04/05/2016 a change in the solar panel position to shade the nitrogen tank on board the Defense Meteorological Satellite Program (DMSP) F-17 satellite was made. In doing so, the integrity of the vertically polarized 37 GHz channel (37V) of the Special Sensor Microwave Imager and Sounder (SSMIS) was compromised. This is a primary channel used in the sea ice processing. On 04/13/16 an additional change in the solar panel position was made.This change has improved the problems we were seeing in the 37V GHz channel. The affected daily files from 04/05 – 04/13 have been removed from distribution.”

Comments are closed.