The DMSP satellite is still operating, but the SSM/I sensor is not
Regular readers will recall that on Feb 16th I blogged about this graph of arctic sea ice posted on the National Snow and Ice Data Center sea ice news page. The downward jump in the blue line was abrupt and puzzling.
Click for larger image
Today NSIDC announced they had discovered the reason why. The sensor on the Defense Meteorological Satellite Program (DMSP) satellite they use had degraded and now apparently failed to the point of being unusable. Compounding the bad news they discovered it had been in slow decline for almost two months, which caused a bias in the arctic sea ice data that underestimated the total sea ice by 500,000 square kilometers. This will likely affect the January NSIDC sea ice totals.
Figure 1. High-resolution image Daily Arctic sea ice extent map for February 15, 2009, showed areas of open water which should have appeared as sea ice. Sea Ice Index data. About the data. Please note that our daily sea ice images, derived from microwave measurements, may show spurious pixels in areas where sea ice may not be present. These artifacts are generally caused by coastline effects, or less commonly by severe weather. Scientists use masks to minimize the number of “noise” pixels, based on long-term extent patterns. Noise is largely eliminated in the process of generating monthly averages, our standard measurement for analyzing interannual trends. Data derived from Sea Ice Index data set.
Figure 2.
High-resolution imageDaily total Arctic sea ice extent between 1 December 2008 and 12 February 2009 for Special Sensor Microwave/Imager SSM/I compared to the similar NASA Earth Observing System Advanced Microwave Scanning Radiometer (EOS AMSR-E) sensor. —Credit: National Snow and Ice Data Center
Dr. Walt Meier of NSIDC had planned to do a guest post here on WUWT, but this evening, with the magnitude of the problem looming, he’s asked to defer that post until later. I certainly can’t fault him for that. He’s got his hands full. Hopefully they have a contingency plan in place for loss of the sensor/space platform. I applaud NSIDC for recognizing the problem and posting a complete and detailed summary today. I’ve resposted it below in its entirety. Note that this won’t affect other ice monitoring programs that use the Advanced Microwave Scanning Radiometer (EOS AMSR-E) sensor, which is on an entirely different platform, the AQUA satellite.
UPDATE: 2/19 Walt Meier writes with a clarification: “One detail, though perhaps an important [one]. I realize that it is bit confusing, but it is just one channel of the sensor that has issues. And it isn’t so much that it “failed”, but that quality degraded to the point the sea ice algorithm – the process to convert the raw data into sea ice concentration/extent – failed on Monday.” – Anthony
From NSIDC Sea Ice News:
As some of our readers have already noticed, there was a significant problem with the daily sea ice data images on February 16. The problem arose from a malfunction of the satellite sensor we use for our daily sea ice products. Upon further investigation, we discovered that starting around early January, an error known as sensor drift caused a slowly growing underestimation of Arctic sea ice extent. The underestimation reached approximately 500,000 square kilometers (193,000 square miles) by mid-February. Sensor drift, although infrequent, does occasionally occur and it is one of the things that we account for during quality control measures prior to archiving the data. See below for more details.
We have removed the most recent data and are investigating alternative data sources that will provide correct results. It is not clear when we will have data back online, but we are working to resolve the issue as quickly as possible.
Where does NSIDC get its data?
NSIDC gets sea ice information by applying algorithms to data from a series of Special Sensor Microwave/Imager (SSM/I) sensors on Defense Meteorological Satellite Program (DMSP) satellites. These satellites are operated by the U.S. Department of Defense. Their primary mission is support of U.S. military operations; the data weren’t originally intended for general science use.
The daily updates in Arctic Sea Ice News & Analysis rely on rapid acquisition and processing of the SSM/I data. Because the acquisition and processing are done in near-real time, we publish the daily data essentially as is. The data are then archived and later subjected to very strict quality control. We perform quality control measures in coordination with scientists at the NASA Goddard Space Flight Center, which can take up to a year. High-quality archives from SSM/I, combined with data from the earlier Scanning Multi-channel Microwave Radiometer (SMMR) data stream (1979–1987) provide a consistent record of sea ice conditions now spanning 30 years.
Data error sources
As discussed above, near-real-time products do not undergo the same level of quality control as the final archived products, which are used in scientific research published in peer-reviewed journals. However, the SSM/I sensors have proven themselves to be generally quite stable. Thus, it is reasonable to use the near-real-time products for displaying evolving ice conditions, with the caveat that errors may nevertheless occur. Sometimes errors are dramatic and obvious. Other errors, such as the recent sensor drift, may be subtler and not immediately apparent. We caution users of the near-real-time products that any conclusions from such data must be preliminary. We believe that the potential problems are outweighed by the scientific value of providing timely assessments of current Arctic sea ice conditions, as long as they are presented with appropriate caveats, which we try to do.
For several years, we used the SSM/I sensor on the DMSP F13 satellite. Last year, F13 started showing large amounts of missing data. The sensor was almost 13 years old, and no longer provided complete daily data to allow us to track total daily sea ice extent. As a result, we switched to the DMSP F15 sensor for our near-real-time analysis. For more information on the switch, see “Note on satellite update and intercalibration,” in our June 3, 2008 post.
On February 16, 2009, as emails came in from puzzled readers, it became clear that there was a significant problem—sea-ice-covered regions were showing up as open ocean. The problem stemmed from a failure of the sea ice algorithm caused by degradation of one of the DMSP F15 sensor channels. Upon further investigation, we found that data quality had begun to degrade over the month preceding the catastrophic failure. As a result, our processes underestimated total sea ice extent for the affected period. Based on comparisons with sea ice extent derived from the NASA Earth Observing System Advanced Microwave Scanning Radiometer (EOS AMSR-E) sensor, this underestimation grew from a negligible amount in early January to about 500,000 square kilometers (193,000 square miles) by mid-February (Figure 2). While dramatic, the underestimated values were not outside of expected variability until Monday, February 16. Although we believe that data prior to early January are reliable, we will conduct a full quality check in the coming days.
Sensor drift is a perfect but unfortunate example of the problems encountered in near-real-time analysis. We stress, however, that this error in no way changes the scientific conclusions about the long-term decline of Arctic sea ice, which is based on the the consistent, quality-controlled data archive discussed above.
We are actively investigating how to address the problem. Since we are not receiving good DMSP SSM/I data at the present time, we have temporarily discontinued daily updates. We will restart the data stream as soon as possible.
Some people might ask why we don’t simply switch to the EOS AMSR-E sensor. AMSR-E is a newer and more accurate passive microwave sensor. However, we do not use AMSR-E data in our analysis because it is not consistent with our historical data. Thus, while AMSR-E gives us greater accuracy and more confidence on current sea ice conditions, it actually provides less accuracy on the long-term changes over the past thirty years. There is a balance between being as accurate as possible at any given moment and being as consistent as possible through long time periods. Our main scientific focus is on the long-term changes in Arctic sea ice. With that in mind, we have chosen to continue using the SSM/I sensor, which provides the longest record of Arctic sea ice extent.
For more information on the NSIDC sea ice data, see the following resources on the NSIDC Web site:
- Sea Ice Index
- Why is the Sea Ice Index product used to study sea ice?
- Do your data undergo quality control?
“…fake but accurate”, comes to mind.
Steve in SC (05:06:45) :
This is perfectly logical to maintain the same data record as long as possible. If you can imagine the difficulty of digitizing the data, programming the image recognition to identify ice & edges based on different levels of contrast, brightness, saturation, refraction, clouds, etc., you can’t switch sensors and hope to duplicate results of another method. Accuracy in this sense is a misnomer. You develop all of these routines to get a program that matches observations as best you can, and it’s probably reasonable to draw conclusions from the data, but proving “accuracy” would be enormously difficult since you really can’t observe the whole arctic to verify what is ice and what is not.
So you pick a method and stick with it so the results are repeatable year to year and day to day and hopefully are adjusted for sensor anomalies. Changing sensors would be like using a camera with a green filter for 30 years then switching to an IR filter. It isn’t going to operate the same, nor will your analysis programs. If there is a switch, you need both data sets to overlap for a long time to get the ability to correlate the new data to the old data. The newer stuff is probably more “accurate” due to higher resolution (pixel wise) and higher resolution (more bits per pixel), and probably a different sensor (wavelength). No two technologies are going to match data pixel for pixel, not even close.
These are all things that are commonly done with industrial vision systems used in automation. They work great if everything stays the same. Change the lighting, the part surface finish, color, get the camera dirty, change the angle, zoom, move it slightly, see a feature that wasn’t there before, whatever, and you have to re-train everything. Small changes like this can throw off a recognized part area or distance by a huge percentage if it is a small area or only a few pixels. It is anything but easy to set up.
Now try that on a planet area that is dark and cold half the year, bright and dark once a day each in the summer, has clouds, different areas per pixel, different angle of incidence of the sun every day…
The data can be saved if they can deconstruct a new sensor sensitivity curve every day since degradation started. That might allow them to take the data, apply the curve to boost the signal where it was before degradation, and end up with reasonably decent data to feed the old recognition routines… The folks at NSIDC are going to be busy for a while.
Again I’m reminded how misinformation takes on a life of its own. The original, incorrect story gets the above-the-fold headlines and the corrections get buried in page 17 next to the community announcements.
Anybody remember when the New York Times ‘discovered’ the north pole was ice free and declared this a sign of impending doom from global warming? How much fanfare did that get, and how many people missed the story about how the pole itself is often ice free during the summer, even during the coldest years?
Points of interest.
1) This episode demonstrates that the “sceintific community” now includes the scrutiny conducted on the internet. Experts and novice contributors, in many fields from around the globe, are providing a level of scrutiny never seen.
This is an enormous benefit to science.
2) The lesser scrutiny provided by the establishment sceintific community and their institutions and publications have not served as an adequate check and balance.
3) The soon to be corrected sea ice trend by the NSIDC will reflect other sources which show sea ice to be nearing the average of the last 30 years contradicting the projected calamity anticipated by Hansen et al.
4) Many goverments are in the process of adopting sweeeping policies derived from the earlier projections of massive sea ice loss from human caused global warming .
5) The media must convey the updated science and trends which contradict the projections, urgency and need for many of those costly policies.
6) Any and all additional unfounded alarms of imminent disaster by AGWers must be responded to with the critisism they deserve.
With perpetrators being held accountable in every forum possible.
DR (08:04:40) :
Accuracy vs precision…..hmm.
If the measuring device has a linearity scale degradation, I fail to understand how it can be used to determine long term trends unless the error and uncertainty is a known value.
Perhaps you should first find out how the device is used?
RSS still has a step change in their data in 1992. It sticks out like a sore thumb, yet has not been corrected; its trend is still notably higher than UAH.
And UAH has a annual deviation from RSS during the austral summer, probably due to their treatment of the antarctic sea ice, this leads to a lower trend than RSS, it also sticks out like a sore thumb.
Is there a comparison of data between the two satellites for say the last 5 years?
Yes many. Happy hunting.
I read this as simply meaning that it’s better to accumulate as long a record as possible using the same instruments. Once the data stops being collected from this source then everyone has to process multiple data sets and use various methods to adjust between them. If NSIDC changes their data source they’ll have to formally adopt one adjustment method and hope it correctly minimizes differences between the data. As their graph with data from both sources shows, they do have very similar data from AMSR-E, but someone has to rigorously prove equivalence before they should consider cutting their data set short.
You’ll also note that they mention how old their satellite and its sensors is. They know their bird is aging, so they know they won’t have to collect this data for much longer. Without knowing the expected lifetime, I expect it is no more than a few years and certainly not as many decades. They may as well collect their data while it is flowing. They and others will in the meantime figure out how to best process all the data.
Anna v,
“I also want to add my voice in thanking Dr. Meier for the oneness of this discourse.
“sorry that “oneness” should be “openness”. Something wrong with my dictionary click.”
Actually, I liked the original. It is Zen-like. Oneness. As in, we are all in this together. We need accurate data to make good decisions . . .
It’s always been my contention and personal and professional experience that the real “peer review” begins after publication, when everyone and their mother can take a shot at what’s been published, using whatever faculties and experience they have to do their own “due diligence” – at least for the purpose of trying to enhance their own understanding of reality, as well as for trying to see what is best to do overall.
Not that Anthony is anyone’s mother, but I think his work also tends to strongly verify the usefulness of this approach, which I would also classify as a Classical vs Faux Liberal m.o., one which therefore rejects taking the words of “experts” as the “given truth” then simply repeating the words as comforting or propagandistic memes. Imo, why even presume that you have something to add to the discussion if that’s all you are going to do?
I hope this will answer once and for all whether or not bloggers- who now serve as the only independent voices on AGW issues- should or should not point out when things look questionable: The answer is, “Yes.”
bluegrue,
What was unusual in the 1920’s regarding the Arctic is unusual in the early years of this century, as well.
The only difference is that AGW fear mongering has confabulated to make normal weather fluctuations be part of some great cliamte apocalypse.
It still seems to me that the real point at issue is being missed here.
The real issue is Quality Control in published data.
Surely anyone with a brain cell looking at the data being published would have been put on notice that 1 million Sq Km of ice disappering in such a short timescale was worthy of investigation.
Had they done an investigation (and sensibly held off publishing the data pending the said investigation) then all this could have been avoided.
As others have said this may well have been going on for some time and questions have reasonably been asked. That appropriate analysis has apparently not been conducted should cause the NSIDC to re-evaluate its policies.
All government agencies should look at the quality control issue much more seriously. This is not an isolated case – in fact it is wholly reflective of the very lax approach that is taken in this area and it would not be tolerated in private industry.
If we are to try and understand the impossibly complex climate system we should at the very least ensure we have good accurate data being provided by government funded agencies.
It is unacceptable IMHO that these bodies take Quality Control so lightly.
See especially the post by
John S. (07:52:23) :
which says it all for me.
Re arctic ice flatline: I recall anecdotal evidence at “SC24” of high latitude NH clouds in December. At the time, I noted that these could impede cooling and explain the pause in growth.
Re Dr. Steig: More transparency on his part would be a welcome improvement, but I would not automatically lump the esteemed Dr. Steig with others in this field.
Re sensor failure: I worked in aerospace in the early days. Sensor failure is rather more subtle than some posters here realize, evincing itself only gradually, usually in the worst possible direction and, of course, at the worst possible time.
Re Dr. Meier: An insider’s look at this subject would be most beneficial. Dr. Meier has probably had a look at the range of misunderstandings here and can give us useful insights.
This raises an important question. When Nansen made their big downwards adjustment in December as reported on WUWT, was that an attempt to line up with NSIDC?
http://wattsupwiththat.com/2008/12/13/something-is-rotten-in-norway-500000-sq-km-of-sea-ice-disappears-overnight/
[snip]
REPLY: Phil – no more posting privileges for you until you address your comment about the headline. – Anthony
Phil.
See my comment attached to yours above.
Phillip Bratby (22:52:12) :
So although they have consistent data far 30 years, they are now recognising that the data provide “less accuracy”. The question is how accurate are that 30 years of data. Can the statement “the scientific conclusions about the long-term decline of Arctic sea ice” be relied on, since they are based on less accurate data?
A good analogy would be with taking the temperature for the past 40 years with one thermometer and then using a more accurate one for the past 10 years. You cans still see the trend by looking at the 40 years of data even though the raw numbers aren’t as accurate.
However, is there no correlation between the two measurements? If there is, why not correct the old data to match the more correct data source and then use that going forward? We have seen a lot of other measurements manipulated, supposedly to correct them, so why not in this case?
“Dave (07:15:02) :
I don’t know if someone already asked but what about the Antarctic sea ice?
Do they have 2 different satellites: 1 for the arctic and 1 for antarctica, or is it the same satellite and was Antarctic sea ice data also wrong the last 45 days?”
Good question, Dave. I suspect the NSIDC satellite is a polar orbiter so it could conceivably pass over the south pole.
WUWT?
http://nsidc.org/data/seaice_index/images/daily_images/N_timeseries.png
It’s baaaaack. NSIDC has no notice I can find for this recent update.
REPLY: No, it is not. On the main page of http://nsidc.org/arcticseaicenews/ “no data available” is still showing. That is simply the last large format image they produced. – Anthony
In a lab situation, if one gets a instrument that is newer, more accurate, and more precise, one would do a calibration.
That is, run both instruments for a period. Figure out a transform that converts the data from the old instrument to the same baseline as the new instrument. Once a decent calibration has been performed, one has essentially dramatically improved the accuracy of the older instrument – and the older data. Not the precision – the old instrument will still produce data with a larger scatter. But the average of the calibrated data from the old instrument should be hitting the baseline of the new instrument.
This allows apples-to-apples comparisons and would seem to allow combination of the two instruments for climatological purposes. This is precisely the same (Missing!) step in changes at all of the surface stations.
Is this something where there is sufficient data to make a calibration attempt publicly available?
J. Peden wrote:
“it’s possible to reasonably predict that the AMSR-E data says something different from what the NSIDC prefers ”
Except that the graph in Dr. Meier’s post shows that the AMSR-E sea ice extent
was virtually the same as the SSM/I sea ice extent until sometime in January.
Scrutiny.
To me thats the key. The guys putting this data out are allowing it to be subjected to scrutiny.
Those of us who look at it are allowed to ask questions about it and challenge the data. Those challenges and questions should not be seen as a challenge to the personalities of those that put this data out there, but a robust questioning of the systems used. Those questions should be used to make the system of producing that data more reliable, enhancing our scientific knowledge.
This seems to have happened here. The guys putting the stuff out found it difficult to question their own results. So others raised the questions for them. Interesting how we can sometimes not question ourselves, and then when someone does it for us, we realise that we were only fooling ourselves. This isnt a criticism of anyone, I have been there myself, many times. I try not to let it happen now, but we build up our view of the truth of the world, and occasionally that gets challenged. First we defend our view of the truth, then defend it irrationally, but after some thought we have to cave in and accept the new reality. Our new world view that we make up becomes better than the previous one.
But this can only happen if we allow what we believe to be true to be subject to scrutiny, and are willing to be criticised for it.
This example is a case of the system working.
Now get that sensor fixed! (Or something).
I’m glad to see that Dr. Meier has had a change of heart and decided that this issue actually is worth blogging about. I’ll look forward to reading his thread. I think it will be very useful (in a positive way) to see his world through his eyes. This crew may be a bit rowdy at times, but, for me at least, it goes a long way when such individuals in positions of authority take a little time to share insights, perspectives, etc. That’s certainly in the spirit of this blog.
Seems as though the blind faith placed on satellite data may be a little misplaced, as I have suspected all along (see John S. ‘s post above). What’s needed is a weather satellite survey – http://www.satellitestations.org ? Maybe these things are crashing into each other more often than we have heard about.
Seriously, in such a harsh environment, it’s surprising that there aren’t more frequent sensor and other equipment failures. When it’s impossible to manually examine and test equipment, maintenance must be a nightmare.
Despite the billions spent on them, the satellite data must still be independently verified by other methods, such as ground weather stations and (didn’t know about these) ice patrol boats. The importance of the surfacestations.org project cannot be overestimated. The condition and accuracy of the ground weather station network is going to have a profound effect on the future for all of us, everywhere.
Why not use both sensors? Then you have a comparison between the two, and could possibly build a more accurate historical chart.
OT: Volcanism Blog and Fresh Bilge report renewed eruption of Chaiten.