You may recall that I posted about how the National Snow and Ice Data Center has an issue with the DMSP satellite sensor channel used to detect sea ice. Cryosphere Today is a few days behind in update compared to NSIDC, and here is what their imagery now looks like before and after:
Above: Arctic “Insta-melt” Click for a larger image
Here is the link to reproduce the image above.
Larger “holes” are likely to open up in the arctic sea in the next couple of days as the sensor further degrades.
Here is what CT has to say as a caveat for the side by side images:
February 17, 2009 – The SSMI sensor seems to be acting up and dropping data swaths from time to time in recent days. Missing swaths will appear on these images as a missing data in the southern latitudes. If this persists for more than a few weeks, we will start to fill in these missing data swaths with the ice concentration from the previous day. Note – these missing swaths do not affect the timeseries or any other plots on the Cryosphere Today as they are comprised of moving averages of at least three days.
No mention of the issue on CT’s main page though. They are still commenting on George Will. They seem a bit out of touch on the sensor issue.
h/t to Garrett
UPDATE: 11:30PM 2/20 CT has removed the comments about George Will from the main page, but still no mention there of the satellite outage nor are they displaying imagery on the main page from 2/20/09 The most recent is 02/19/09. It will be interesting to see what tomorrow brings.
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
They do not have a model to detect quality control issues. Its pretty simple stuff for those who deal with large amounts of data flowing in.
Like most cases, the private sector is way ahead of the public and scientific one when it comes to finding and fixing data issues.
Somebody needs to contact Drudge and Bloomberg, THIS IS NOT NEWS! It’s not even worth blogging about.
Austin (17:59:42) :
“They do not have a model to detect quality control issues.”
I’ve seen “test” routines, which are part of processing algorithms, as well as code to control “quality” which are part of the running algorithms. That to me is a bigger problem than not having any models to draw from in the algorithms, since they apparently don’t work when a known sensor error arose. This particular sensor problem has been known and has influenced the sensor output to varying degrees over several years. It just became obvious to Anthony, not to the scientists, who thought the drop was because of weather. With that in mind, what have the algorithms been doing with data that isn’t obviously bad? If their code didn’t catch a biggie, would anyone expect it to catch minor errors? I doubt it.
how long will it take them to update? i am looking forward to the graphs reflecting this…
D. King (17:30:12) :
The sensors can simply not be trusted. The only way to tell
when they started to degrade, is to look for a jump in the
calibration bias tables.
I posted this on the other thread.
Please have a look at this comparison between AMSRE (JAXA site) and SSM/I (NSIDC) This compares monthly averages back to 2002. In general the AMSRE data is smaller than the NSIDC and the only significant time it is larger is in Sept 2008 when it is 170,000 sq km larger. A difference of 3.5% larger compared to AMSRE area at that time.
http://img16.imageshack.us/img16/9782/ssmiamsrerz8.jpg
The lowest trace (values on right) show the difference between the nsidc extent and amsre data
There is no significant DRIFT visible in data. There is an obvious difference in the algorithms used to genereate the 15% sea ice areas leading to different total areas. The nsidc data therefore does show valid CHANGES in extent up to end of December 2008 as does the AMSRE data.
Mike
More Political Science. One question? What is their funding source?
The Sensor problems at NSIDC are mainstream now.
Read this Bloomberg article:
Arctic Sea Ice Underestimated for Weeks Due to Faulty Sensor
“The recent error doesn’t change findings that Arctic ice is retreating, the NSIDC said”.
By Alex Morales
Feb. 20 (Bloomberg) — A glitch in satellite sensors caused scientists to underestimate the extent of Arctic sea ice by 500,000 square kilometers (193,000 square miles), a California- size area, the U.S. National Snow and Ice Data Center said.
The error, due to a problem called “sensor drift,” began in early January and caused a slowly growing underestimation of sea ice extent until mid-February. That’s when “puzzled readers” alerted the NSIDC about data showing ice-covered areas as stretches of open ocean, the Boulder, Colorado-based group said on its Web site.
“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,” the center said. “Although we believe that data prior to early January are reliable, we will conduct a full quality check.’’
The extent of Arctic sea ice is seen as a key measure of how rising temperatures are affecting the Earth. The cap retreated in 2007 to its lowest extent ever and last year posted its second- lowest annual minimum at the end of the yearly melt season. The recent error doesn’t change findings that Arctic ice is retreating, the NSIDC said.
The center said real-time data on sea ice is always less reliable than archived numbers because full checks haven’t yet been carried out. Historical data is checked across other sources, it said.
The NSIDC uses Department of Defense satellites to obtain its Arctic sea ice data rather than more accurate National Aeronautics and Space Administration equipment. That’s because the defense satellites have a longer period of historical data, enabling scientists to draw conclusions about long-term ice melt, the center said.
“There is a balance between being as accurate as possible at any given moment and being as consistent as possible through long time-periods,” NSIDC said. “Our main scientific focus is on the long-term changes in Arctic sea ice.”
To contact the reporter on this story: Alex Morales in London at amorales2@bloomberg.net.
http://www.bloomberg.com/apps/news?pid=20601110&sid=aIe9swvOqwIY
Link via http://www.seablogger.com
And they are still publishing admittedly corrupted data because…?
Somebody asked why we don’t bombard them with email.
Will that guy please go the the website and find the “contact us” link? ‘Cause, I can’t seem to find it.
Isn’t that odd…?
I suspect that there were sensor issues far earlier than six weeks ago.
cryosphere-science@atmos.uiuc.edu
or cryosphere-data@atmos.uiuc.edu should work
WestHoustonGeo:
The comments link on the CT site is below the “Global sea ice area” graphic on the right side of the page about two clicks down.
Fellow Geo from the PNW
thefordprefect (18:30:01) :
Mike,
Thanks for the link. I guess the problem I’m having
with this, is that to truly calibrate a passive sensor End-
to End, you must have a survived ground site with a
known emitting source. Satellite to satellite comparisons
will give you a general idea of health, but seem a little
lacking. I wish I knew more about the sensor construction.
Best,
Dave
Smokey (17:01:42) :
Neil Crafter is right. Providing clearly inaccurate information to the public makes no sense.
Unless you’re politically motivated and reality needs to be distorted.
“these missing swaths do not affect the timeseries or any other plots on the Cryosphere Today as they are comprised of moving averages of at least three days”
A moving average will cause a drop spread over 3 days if one of the days is bad.
A median of 5 days would cancel out most of the missing data.
Hopefully, someone there understands how to use the data to remove the missing parts. The technique of Median Combine is well known in the AstroImaging or Image Processing routines of departments of CalTech, which has a pipeline called CalEx.
Even Hubble’s raw data must be put through pipelines of some sort of Median or other scheme to fill in the ever-present holes left by millions of CCD pixels that are non-uniform in response.
Simple averaging…no way will it get the job done properly.
Cryosphere e-mail address
cryosphere-science@atmos.uiuc.edu
In 2001 data from Surveyor showed global warming on Mars.
The media wasn’t interested but bloggers heard about it, and mocked the global warmers with it.
Finally, the Media’s fair haired boy couldn’t take anymore mocking, so they posted a “debunking” of global warming on Mars, October 5, 2005. http://www.realclimate.org/index.php?p=192
This put the media/government nexus of professional AGW alarmism in a tight situation. Their policy had been to ignore global warming of the other planets, but their other policy had been to direct any who questioned the scientific basis of global warming toward Real Climate to search haphazard for the answers from their paid shill. The two policies were in conflict.
Realclimate had debunked something that the nexus was loathed to report. Global Surveyor was rocking the boat.
What to do? What to do?
How about kill the witness?
http://blog.wired.com/gadgets/2006/11/mars_global_sur.html
Then besmirch it’s character.
SPACE.com — Dust Storms Fuel Global Warming on MarsApr 4, 2007 … Shifting sandstorms on Mars might be contributing to a recent bout of global warming on the planet that is shrinking the southern polar …
Old news…
but could this be history repeating itself?
Though I’m sorry to see the failure of the SSMI sensor, I’m glad it will no longer be used by NSIDC to further the AGW hoax. I take a personal interest, since the sensor is based on my dissertation (Wisconsin-Madison, 1976) and research with the late Dr. Jim Hollinger of Naval Research Labs. I wrote the sensor specifications, and was manager of the Hughes Aircraft software team that developed the processing code. We turned that software over to NSIDC, free gratis (with government permission). Had I known then what they would do with the data, I’d have traveled to Illinois or to Podunk instead.
There are two reasons for my disgust with NSIDC. One is their endorsement of Al Gore’s Inconvenient Truth, posted in a blog on their website, and signed by Walt Meier (and Ted Scambos). Scientists who endorse Mr. Gore are not scientists; they’re propagandists.
The second reason is the annual Summer analysis of ever decreasing Arctic ice extent, made in August, when “melt ponds” abound on the ice surface. A couple of mm of melt water make the submerged ice look like open water. Everyone familiar with microwave remote sensing is aware of this, and the disagreement by the National Ice Center with the NSIDC analysis has been noted in the past on this blog. Sensor degradation makes such analysis even more questionable.
No sympathy for NSIDC. Bad science/propaganda is worse than none at all.
Richard C. Savage
Colorado
REPLY: Richard, thank you for your candor. I found the NSIDC link to AIT you spoke of, readers can see it here:
http://nsidc.org/news/press/20060706_goremoviefaq.html
– Anthony Watts
This and other advocacy issues is why CT is no longer a site I visit.
I am a regular visitor at IARC-JAXA, which uses the more accurate AMSR system that NSIDC seems to want to avoid.
Mr Will was comparing the END OF 2008 with the END OF 1979. Not Feb 15th of each year (which was still a month and a half away from happening when he wrote it).
Why did Cryophere Today blatantly misrepresent what the author was writing about?
And blatantly cherry pick a single days data in attempt to ‘disprove’ what Cryosphere’s own graphs clearly show to be true.
http://arctic.atmos.uiuc.edu/cryosphere/IMAGES/global.daily.ice.area.withtrend.jpg
These people (AGW proponents) have no shame.
Ah, it’ll be alright, the ice is all going to melt anyway, so why do we need those sensors? You see, the ice is melting, methane is being released which in turn causes more warming and more release and so the cycle goes until it all burns up. Remember, uncle Al told us that it was going to melt in 5 years. He warned us!
L.A. Times Says (embedded video) .. “All those people living in those big cities producing CO2” .. give me a break!
BTW, I was playing around on the Cryosphere site, and it looks to me like we currently have a whole lot of ice. I’m having a hard time believing old uncle Al !!!
Are there any Greenpeacers headed to the “butterfly” with their kayaks yet?
I know that somebody brought the December necklace to our attention just recently. Anybody got a picture of the necklace image at Cryosphere back in, what, December? Is it possible that we have to back to November to get a clean data set?
It’s the batman signal! dadadadadadadadadadadadadada…BATMAN!!!!
Mike D. (21:14:51) :
Let us hope so!