I’ve been aware of this for a couple of days on our Sea Ice page, but hadn’t done anything about it since I wanted to see if it might change. When blogger Kate of Small Dead Animals noticed it and published on it, I figured it was time to start asking NSIDC some questions.
Compare this NSIDC Arctic Sea Ice extent chart…
…with this from Cryosphere Today:
The NSIDC plot has since intersected the 2007 line, but CT has no new images up since 10-27-10:
It certainly appears that there is more ice in 2010 than 2007 on the Cryosphere Today page. CT hardly ever responds to email, so I didn’t even bother asking them why the discrepancy. NSIDC’s Walt Meier though, takes our concerns seriously and responded rather quickly to my questions:
~~~~~~~~~~~~~~~~~~~~~~~~~~~
From: Walt Meier
To: Anthony
Subject: Re: you might have a problem
Sent: Oct 29, 2010 8:42 AM
Hi Anthony,
Thanks for the heads up. I looked at it and it doesn’t look like there
is any problem.
As we went through before with Steve [Goddard], looking at the images can be
misleading because they’re not on an equal area projection. There is
more ice in the central Arctic this year, but less in the Beaufort Sea,
Canadian Archipelago, and Baffin Bay. These areas roughly balance each
other out.
I also recall Cryosphere Today having an issue of changing their images,
so I don’t know if you can consistently compare them anyway – it looks
like their 2007 image is missing some ice. Attached is our concentration
images from 2007 and yesterday and there doesn’t look like much
discrepancy (apologies for the different image sizes).
walt
~~~~~~~~~~~~~~~~~~~~~~~
I fixed the size differences, and here they are:
Of course we don’t have the daily extent data from NSIDC, since they so far have refused to publish it (they do give monthly though) so, we have to be content with image comparison rather than data comparison with NSIDC.
=======================================
Walt, as I said before, you really should publish the daily data. Consider how this looks: NSIDC director Serreze screams “death spiral” to the media while at the same time holds back publicly funded data. It is the same sort of bull-headedness that got CRU in deep trouble. – Anthony
=======================================
UPDATE: Reader Lee Kington provides this blink comparator version:
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
I think the difference is in the shift in position of the ice formation so far for the season.
In 2007 the lunar declination was starting to drop from a maximum extent, (in 2005) creating a lot of the wind patterns that pushed out a lot of the ice, now in 2010 the declinational extent is down to the average ~23 degrees and we are having a lot of the loopy turbulence patterns in global circulation.
The pattern of the ice formation is more centered in the arctic ocean, from the global wind patterns of this last summer/fall pushing up from the South the remnants of the tropical storms, both warmth and moisture in across the N Atlantic into the Beaufort Sea, Canadian Archipelago, and Baffin Bay. With the flow out through Siberia extending the ice extent in that direction, with the advent of NH winter coming on the flows of warm moisture will slow to a crawl, and the ice will continue to grow in this more centered position through spring 2011.
The main difference between the 2007 and 2010 ice formation pattern is the result of where we are in the 18.6 Mn period of lunar declination, and as the declinational angle at culmination drops further to ward the minimum of ~18.5 degrees over the next several years, the patterns of blocking highs will drop off in about two years yielding more robust zonal flow patterns, isolating the polar cold from the tropics more, that will engender more years of increases in total ice cover.
Couldn’t one just do a Hysteresis plot of each image to see if there are measurably more white pixels than blue/grey pixels?
Just a thought..
Regards..
I’ve just made an image analysis comparision, using the free (and great) imageJ package, and Dr. Meier is right. In terms of ice pixel area (considering the full ice pixel spectrum) both images have the same amount. Although, has some people pointed in previous comments, there are more white pixels in 2010 (100% ice) than in 2007.
http://igloo.atmos.uiuc.edu/cgi-bin/test/print.sh?fm=10&fd=27&fy=2007&sm=10&sd=27&sy=2010
I thought everyone knew that UIUC concentration maps were;
“Sea ice concentrations less than 30% are not displayed in these images.”
2007’s PDF/CDF != 2010’s PDF/CDF
PDF = Probability Density Function
CDF = Cumulative Density Function
If you don’t understand that different PDF’s/CDF’s can cross over each other, well then, you’ve never looked at soil/sediment/crushed aggregate sieve data sets.
NSIDC uses 15%, again I thought everyone knew that;
15% != 30%
Also, I’ve come to fine UIUC’s daily sea ice areas (100%) at;
http://arctic.atmos.uiuc.edu/cryosphere/timeseries.anom.1979-2008
to be the most useful data set of the lot of available Arctic sea ice time series (especially during the growth cycle, when there are no melt ponds on top of the sea ice);
For example, a 7-day least squares fit (last 7 days) of JAXA (15%) is 53.3K/day vs UIUC which is 123.3K/day, or ~2.3X as fast in 100% area vs 15% extent growth.
I’ve also been pulling the graphical dailies from Bremen and NSIDC to plot against JAXA’s digital dailies, all three show a slow down in extent growth.
I too, would very much like for NSIDC to publish their 15% extents and 100% areas on a daily basis instead of just the monthly means of both, there’s some good data mining that could be done as per the UIUC 100% area dailies.
Paul says:
October 30, 2010 at 2:20 pm
JAXA shows more or less the same thing, about roughly equally extent. So I think you have the wrong story here.
What I’m wondering about is that the SSMI data shows 2010 with clearly more extent than 2007. This raises some questions about the reliability of the ice cover datasets prior to 2002 and whether data from the newer and older instruments can be reliably compared.
_____________________________________________________________
When NSIDC switches sensors, as they have to do every several years, they overlap the two sensor data sets by at least a year, you can look back at their sea ice news archives, they switched sensors ~2 years ago, and have several articles there showing their procedures for doing so.
JDN says:
October 30, 2010 at 1:42 pm
Isn’t thes sea ice charts all processed information? I’ve asked repeatedly if anyone has ever had primary data from these satellites so that people can check up on the results. Is the community now going to accept secondary data without complaining about lack of transparency? No point in worrying about it or really discussing it seriously if there is no way to check up on the results.
_____________________________________________________________
You can always work with the gridded dailies from NSIDC, but their in HDF format.
You want raw sensor data?
Well then it takes several passes and you need to know the sensor’s orientation relative to what your looking at and do a whole heck of a lot of geometery and land masking, etceteras.
Good luck with doing that yourself.
I haven’d done so, simply because it’s too muck fraggin’ work!
R Gates: You purport to be a scientist. I am not, but I know enough about statistical analysis to comprehend that 30 years worth of satellite images without any baseline or norm for comparison is interesting but insignificant. As I understand it, no climate trends or conclusions can be determined from 30 years worth of computer generated images derived from satellite sensor data. Isn’t this type of information considered anecdotal ?
You say: “Arctic sea ice is still some 1.4 million sq. km. below the longer term (30+ year) running average and has not been above that average since 2004.”
So what? Even if true, the computer generated images are daily snapshots of but one metric for regional weather conditions. There is no rational basis, much less any evidence, to support a claim that 30+ years of Arctic ice extent images show conditions beyond the scope of natual varibility.
tallbloke says:
October 30, 2010 at 2:08 pm
“The SSMI data seems to be telling another story:
http://arctic-roos.org/observations/satellite-data/sea-ice/ice-area-and-extent-in-arctic“
Interesting TB, DMI also showing the line intersecting 2009, not 2007!
http://ocean.dmi.dk/arctic/plots/icecover/icecover_2010.png
Pay attention everyone. This is pretty simple thermodynamics and the mechanics of heat dissipation.
Ocean surface temperature below 60N is plunging. Spencer just gave us the latest AQUA sat SST measurements yesterday or the day before.
Q: Where is the SST heat going?
A: It’s being radiated out into space at the pole.
The albedo of the pole is insignificant. Water is a reflector when the sun is low in the sky and it’s always low near the pole. If the ice melts in fact it improves the radiative performance as ice is an insulator.
When the surface at the pole is warmer there is a greater delta T between it and the cold of space (a constant 3 degrees above absolute zero). That indisputably (unless thermodynamic laws stop working) means more heat is escaping into space.
No mystery here. The only mystery is why so many people can’t connect the dots. It doesn’t seem like it should take an engineering genius to figure this out.
It is really good to see a climate blog with a post about climate, congratulations Anthony!
To help us understand what is happening at the moment it would be good to know what is happening with salinity. I presume that all the fresher water has frozen and that the warmer/more saline water is now slowing the freeze-up. Obviously there is also still ice moving through Fram Stait. Your reference page is excellent Anthony, but can you access salinity data and ice pack movement data as well?
Conversely as tropical oceans cool and the arctic ocean warms the delta T between tropics and pole decreases. This delta T is what provides the energy that drives the heat pump moving excess heat to the poles. When water and air are being moved from one place to another that’s work being accomplished. Heat gradients provide that energy. Very simple mechanics here. Everyone should probably take as much time as it requires to learn and understand the operating principles underlying things like Stirling engines, the Carnot cycle, and things of that nature. Without that familiar knowledge little in the way of atmospheric and oceanic physics is going to make any sense.
I agree. The problem is that in this battle for the truth the Warmists do send press releases regarding the “weekly”……. “downs” which you referred to. Never the ups. Have you notice this? Warmists by the way are not allowed to send press releases indicating that Arctic sea ice extent, during the Holocene, has been lower than 2007.
Correction:
Sceptics by the way are not allowed to send press releases indicating that Arctic sea ice extent, during the Holocene, has been lower than 2007.
Low quality comment:
Oh, the ice bears will note have as much ice to crunch and eat this year as last year if the development goes on like this! How will they survive?
Stephen Wilde says:
October 30, 2010 at 1:26 pm
“We’ve seen two years recovery of sea ice at minimum (then a small step back due to an El Nino and residual warmth in the Atlantic from 30 years of positive PDO) but then most likely further sea ice recovery to come.”
——————————————————————
Look at the NSIDC plot of September minima from 1970 to 2010 (http://nsidc.org/images/arcticseaicenews/20101004_Figure3.) and please tell me where you see any recovery whatsoever.
You seem to inhabit a fantasy world.
Sea ice in the Gulf of Finland on 28 Oct 2010?
Are you sure about that?
SMHI Sea Surface Temperature Issued 2010-10-28
[googlemaps http://maps.google.com/maps?f=q&source=s_q&hl=en&geocode=&q=Gulf+of+Finland&sll=37.0625,-95.677068&sspn=33.984987,56.513672&ie=UTF8&hq=&hnear=Gulf+of+Finland&ll=60.072633,26.44557&spn=2.675009,7.064209&t=h&z=7&output=embed&w=425&h=350%5D
Guys, I don’t know what the ice situation is in the Arctic right now. But here in Ottawa, exactly 5000 km south of the North Pole, it’s below freezing and snowing hard.
3 years ago, the first snow fell on November 16th. 2 years ago, in 2008, the first snowfall was November 20th. Last year, the first snow fell on December 1st. This year…October 30th.
Yeah, yeah, I know the warmist line – when it’s hot it’s climate, when it’s cold it’s weather. We could sure use more climate and less weather up here.
Walt, you need new glasses.
Dave Springer, thanks!
Sorry about that Googlemaps posting glitch.
Here is the NIC snow & ice map for 28th Oct
Seems to be still mild in Southern Finland tonight.
Tie 7 Box (Helsinkiin)
IJIS and Uni of Bremen are both showing the same profile as NSIDC. ROOS is a bit higher but still shows the same slow-down towards the 2007 extent for this time of year.
The corroboration seems to be robust, as it does for the aerial concentration blink graph. Having found nothing wrong, seems a bit churlish to tick off Walt Meier about something else entirely, particularly considering the time he took to make a post for this site recently.
And lest we forget, we’re talking about weather anomalies here. I suppose if the graphs showed rapid increase instead, we’d be reading comments about how the ice is ‘recovering’, as if it’s a climate issue.
EFS_Junior says:
October 30, 2010 at 3:32 pm
What is the HFS format? There are too many acronyms with HFS for me to find it. Is there a simple reader or file convertor?
As for raw data, I don’t want to convert every pass. What I’d like to do is assign the raw data to approximate GPS position over open ocean to correlate it with ocean imaging from planes or ships which might be in the area. Basically, I’d like a spot check on the performance of the satellite and its ability to identify % of ice cover.
NSIDC are continuing to update as the ice coverage increases. ‘skeptics’ update if there’s something to hang their preferred agenda on.
You seem to be suggesting that a tit-for-tat meme-war is necessary, rather than a proper, comprehensive education on the issue. The last time we had that here was in a post from Walt Meier at the NSIDC, while articles like the above continue to propagate, educating no one, but spieling the narrative that gets ‘skpetics’ cheering along.
Boy, have you got that backwards. The ‘press release’ (a blog post) on Holocene minima completely distorted the study it was based on, which was of an area <10% of the Arctic. To find a reference to the possibility of mid-Holocene ice coverage being lower than today – one that doesn't distort scientific papers – you can visit… NSIDC.
That page hasn’t been updated since Polyak 2010 came out, which indicates Holocene sea ice concentration is likely to be at the lowest currently.
Note the qualification ‘likely’. Think about that, and then consider what you wrote. The study your conception is based on was by no means definitive, never mind that it’s conclusions were blown 900% out of proportion.
BTW, Jimbo, if what you say is true, there should be a number of ‘press releases’ about the slow-down in ice growth leading to sea ice extent similar to the record 2007 for this time of year. By all means demonstrate your ‘observation’ by linking us up to newsy items about that.
This all just navel gazing until we see a full 60 years to account for the PDO (or even longer if we want to factor in the AMO).
The Pacific NW is looking more like the early 1950’s than the 1990’s of late. Even the koolade drinkers are getting a bit antsy about global warming theory now.
For what good they are, charts based on JAXA 15% to 29th:
http://img513.imageshack.us/img513/8949/13day20101029.png
http://img832.imageshack.us/img832/4182/all20101029.png
http://img714.imageshack.us/img714/6170/minmaxtime.png
http://img401.imageshack.us/img401/3756/hyst20101029.png
The cycle is clearly as hysteretic as hell, constrained by geography and zenith angle, attenuated by SSTs, weather, etc. It’s a big noisy heat pump. The important bit is the size of the latent heat suck/blow and how that energy dissipates; about which extent figures tell us nothing reliable. Nit-picking over a mere 30 years or even a century of extent data and drawing trend lines from which to make scary predictions is therefore IMHO half-baked nonsense – the stuff of nut jobs.