Foreword: I had originally planned to post a story on this, but Steven Goddard of the UK Register sends word that he has already done a comparison. It mirrors much of what I would have written. There is a clear discrepancy between the two data sources. What is unclear is the cause. Is it differing measurement and tabulation methods? Or, is it some post measurement adjustment being applied. With a 30 percent difference, it would seem that the public would have difficulty determining which dataset is the truly representative one.
UPDATE: The questions have been answered, see correction below – Anthony
Arctic ice refuses to melt as ordered
Published Friday 15th August 2008 10:02 GMT – source story is here
Just a few weeks ago, predictions of Arctic ice collapse were buzzing all over the internet. Some scientists were predicting that the “North Pole may be ice-free for first time this summer”. Others predicted that the entire “polar ice cap would disappear this summer”.
The Arctic melt season is nearly done for this year. The sun is now very low above the horizon and will set for the winter at the North Pole in five weeks. And none of these dire predictions have come to pass. Yet there is, however, something odd going on with the ice data.
The National Snow and Ice Data Center (NSIDC) in Boulder, Colorado released an alarming graph on August 11, showing that Arctic ice was rapidly disappearing, back towards last year’s record minimum. Their data shows Arctic sea ice extent only 10 per cent greater than this date in 2007, and the second lowest on record. Here’s a smaller version of the graph:
The National Snow and Ice Data Center (NSIDC)’s troublesome ice graphThe problem is that this graph does not appear to be correct. Other data sources show Arctic ice having made a nice recovery this summer. NASA Marshall Space Flight Center data shows 2008 ice nearly identical to 2002, 2005 and 2006. Maps of Arctic ice extent are readily available from several sources, including the University of Illinois, which keeps a daily archive for the last 30 years. A comparison of these maps (derived from NSIDC data) below shows that Arctic ice extent was 30 per cent greater on August 11, 2008 than it was on the August 12, 2007. (2008 is a leap year, so the dates are offset by one.)
Ice at the Arctic: 2007 and 2008 snapshotsThe video below highlights the differences between those two dates. As you can see, ice has grown in nearly every direction since last summer – with a large increase in the area north of Siberia. Also note that the area around the Northwest Passage (west of Greenland) has seen a significant increase in ice. Some of the islands in the Canadian Archipelago are surrounded by more ice than they were during the summer of 1980.
The 30 per cent increase was calculated by counting pixels which contain colors representing ice. This is a conservative calculation, because of the map projection used. As the ice expands away from the pole, each new pixel represents a larger area – so the net effect is that the calculated 30 per cent increase is actually on the low side.
So how did NSIDC calculate a 10 per cent increase over 2007? Their graph appears to disagree with the maps by a factor of three (10 per cent vs. 30 per cent) – hardly a trivial discrepancy.
What melts the Arctic?
The Arctic did not experience the meltdowns forecast by NSIDC and the Norwegian Polar Year Secretariat. It didn’t even come close. Additionally, some current graphs and press releases from NSIDC seem less than conservative. There appears to be a consistent pattern of overstatement related to Arctic ice loss.
We know that Arctic summer ice extent is largely determined by variable oceanic and atmospheric currents such as the Arctic Oscillation. NASA claimed last summer that “not all the large changes seen in Arctic climate in recent years are a result of long-term trends associated with global warming”. The media tendency to knee-jerkingly blame everything on “global warming” makes for an easy story – but it is not based on solid science. ®
Bootnote
And what of the Antarctic? Down south, ice extent is well ahead of the recent average. Why isn’t NSIDC making similarly high-profile press releases about the increase in Antarctic ice over the last 30 years?
The author, Steven Goddard, is not affiliated directly or indirectly with any energy industry, nor does he have any current affiliation with any university.
NOTE OF CORRECTION FROM STEVEN GODDARD:
The senior editor at the Register has added a footnote to the article with
excerpts from Dr. Meier’s letter, and a short explanation of why my analysis
was incorrect.
To expound further – after a lot of examination of UIUC maps, I discovered
that while their 2008 maps appear golden, their 2007 maps do not agree well
with either NSIDC maps or NASA satellite imagery. NSIDC does not archive
their maps, but I found one map from August 19, 2007. I overlaid the NSIDC
map on top of the UIUC map from the same date. As you can see below, the
NSIDC ice map (white) shows considerably greater extent than the UIUC maps
(colors.) The UIUC ice sits back much further from the Canadian coast than
does the NSIDC ice. The land lines up perfectly between the maps, so it
appears possible that the UIUC ice is mapped using a different projection
than their land projection.
Click for larger image
Because the 2007 UIUC maps show less area, the increase in 2008 appears
greater. This is the crux of the problem. I am convinced that the NSIDC
data is correct and that my analysis is flawed. The technique is
theoretically correct, but the output is never better than the raw data.
Prior to writing the article, I had done quite a bit of comparison of UIUC
vs. NSIDC vs. NASA for this year. The hole in my methodology was not
performing the same analysis for last year. (The fact that NSIDC doesn’t
archive their maps of course contributed to the difficulty of that
exercise.)
My apologies to Dr. Meiers and Dr. Serreze, and NSIDC. Their analysis,
graphs and conclusions were all absolutely correct. Arctic ice is indeed
melting nearly as fast as last year, and this is indeed troubling.
– Steven Goddard
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
If there is a flaw in the methodology, it is not at that point. It would have to be tied to an inaccuracy in the UIUC map or my understanding of that map.
In which case why are you expecting Drs Serreze or Meier, of the NSIDC, to deal with this for you? Why are you not in contact with UIUC (why, for that matter, were you not in contact with both as an obvious aspect of basic research before writing your article?). If you have determined that any inaccuracy is down to the UIUC image, why are you not correcting your article, which explicitly states that the NSIDC graph “does not appear to be correct”?
I’m certain that the NSIDC calculations were done carefully and diligently.
In which case, why have you not withdrawn your allegation of “a consistent pattern of overstatement related to Arctic ice loss”, referencing their recent graphs as well as their press statements?
You seem to be suggesting that you will retract and apologise if the NSIDC can explain issues you have with a UIUC visualisation, but not that you will retract and apologise for the statements I have quoted, even though you now assert your certainty that the NSIDC has been diligent and that any inaccuracy is not to do with them!
I had another look at the distortion factor as noted at (01:28:52). Now I get +5% for 20080811 and +4% for 20070812 for a 2007 to 2008 increase of 31.6%.
Still for view height of 10000 miles.
I would still not be betting on the accuracy of my calculation, but for these images and projection it seems no big deal anyway.
Anthony,
I really like the option of seeing the higher resolution color maps, and wish they had those archived from last year. The increased detail would be useful in comparing vs. satellite images.
BTW – I need to throw in here that all my comments on this forum are my opinion only, not those of The Register.
Looking back at earlier posts, a few things caught my eye which I might be able to clarify:
1) The north pole issue: Back in June, there was some coverage about the possibility of the North Pole being ice free by the end of this summer. This was based on recognition that the area around the north pole was covered by firstyear ice that tends to be rather thin. Thin ice is the most vulnerable to melting our in summer. I gave it a 50/50 chance. Looks like I’ll lose my own bet and Santa Claus will be safe for another year. There was indeed some coverage a some years back of an open north pole (and I was interviewed). This opening, however, was pretty clearly causes by unusual winds breaking up this ice, and not from melting out.
2) The uptiicks/downticks in our updated time series of ice extent. The time series, as presented on our “Sea Ice News and Analysis Site” is based on 5-day averages. This is done to smooth out short-term “blips” that can occur from noise in the data (basically weather effects and surface melt effect that contaminate the passive microwave retreivals). Speciicially, for a given day N,
the 5-day average is: (N-2 + N-1 + N + N+1 + N+2)/5 . You run into a problem for the current day (e.g., August 20 as I write this) as we don’t yet have data for N+1 and N+2. Similarly, for August 19 (yesterday), we don’t yet have data for day N+2. So, for the current day and the day before, we do a modified 5-day mean by projecting values forward by 1 or 2 day based on the slope over the past few days. This 5 averaging procedure is why the slope for the last few days or our graph shows these somewhat puzzling upticks/downticks. It turns out we actually changed procedures slightly right around August 1. Before then, the five day average for the current day was (N-2 + N-1 + N + N + N)/5, which really gave too much weight to the current day. Nothing nefarious about making the change – it was a decision based on simply trying to improve presentation of the results.
3) Antarctic Sea Ice: There seems to be a great deal of misconception about why sea ice in Antarctica is not declining (and indeed does show an increase) and why Antarctica (except for the peninsula) is not strongly warming like the Arctic. One has to realize that a much slower and subdued response of the Antarctic to greeenhouse gas loading is something that was projected by even our earlier climate models over 25 years ago. There’s a number of reasons for this involving the much different nature of the ocean and atmospheric circulation, but the fact that it is the Arctic which is changing rapidly and not the Antarctic is no surprise. Indeed it’s exactly what we expected. The recent sea ice ingrease in Antarctica turns out be be a weak trend in a noisy time series, largely drivent by changes in atmospheric circulation. Please go to the “frequently asked questions” on our “sea ice analysis and news” site where these issue ae discussed.
4) I see no point in further discussion about malfeasance on the part on NSIDC. We stand by our data and results. My colleague Walt Meier stated the issues quite clearly. Some do not like the story that our data tell. I don’t like the story either, but I’ll continue to tell it as I see it.
There are many projections used to represent the earth’s surface on a flat surface and they all make compromises of some sort. The most common projection used for world map sis the Mercator which was designed for navigation and has the virtue that lines of constant bearing are straight lines which is great for navigation but distorts area as you go away from the equator (Greenland looks much larger than it really is). There are several projections used for views from the pole of which the polar stereographic is commonly used (e.g. by Uni Bremen and NSIDC), the projection onto a secant plane at 70ºN being chosen in this case to reduce the distortion of the Arctic.
The CT images apparently use a different projection, Goddard claims that it’s the view that an astronaut would have from 10,000 miles above the pole (a perspective projection), simple geometry tells us that it can’t be that, if indeed it is that type of projection it’s more like 3,500 miles.
Whichever projection is used it means a pixel on the map represents a different area depending on position and this must be allowed for in calculating the area from the map. In the absence of the actual data this is all you could do, however when the data is available it’s simpler and more accurate to use it (as several have shown here). I first pointed this out to Goddard early in this thread (on the 15th) and he repeatedly persists that it has no effect!
Pioneer stock like me eyeball weather. If we don’t, our crops don’t get put up and either freeze, rot or dry up in the field. So eyeballing the two Arctic images of last year and this year, this same date, we have lots more ice up there. Nuf said. Cold winter coming and soon. Don’t bother with a 3rd cutting. Irrigate like hell so what is left can be used to extend feeding till pastures run out or get snowed out and we have to start using expensive bailed hay. Plant winter wheat again now and hope to gawd it don’t freeze when it comes up like it did this year. Spring wheat won’t get in the ground in time to harvest and besides, you don’t get much for spring wheat. Winter wheat always sells higher. Buy heating fuel or cut wood now. And buy/cut at least enough to last 9 months.
wattsupwiththat (07:40:23) :
UPDATE:
Cryosphere Today has changed their color scheme presentation very recently.
See the new color scheme here:
http://arctic.atmos.uiuc.edu/cryosphere/
It appears the archived images are unchanged and remain the same color scheme. I can’t find any announcement for why this has been done, if somebody knows of one, pass it along please.
It appears to be more than just a color scheme but higher resolution too (longer to load 🙁 ). The ‘old’ style data can still be chosen (old (SSMI)), from comparison it seems to me that the new data is from the higher resolution AMSR-E imager as opposed to the SSMI imager (from JAXA, see also Uni-Bremen). A different color pallet is also used, most notably it appears to be logarithmic rather than linear.
“But the Cryosphere today images, such as these here: http://igloo.atmos.uiuc.edu/cgi-bin/test/print.sh?fm=08&fd=15&fy=2007&sm=08&sd=15&sy=2008
are PNG and totally lossless. So I would expect accurate pixel counts between them.”
Have you looked at those images closely? Although the image pairs are distributed as PNGs, at some point the imagery has been through lossy compression. The errors are most obvious along the ice margins, visible as a speckled pattern within 8 by 8 pixel blocks. Try selecting a single color in the color bar with the magic wand tool in Photoshop, or a similar image processing application: it doesn’t select a single horizontal line, which you would expect in a clean image.
REPLY: Simmon, good catch. the web page presentation compresses the image size, but if you do a right click and pull up the URL for the image pair that was generated, you get the full size image which is 1709 x 856 pixels.
I did that and copied it into my imaging program and zoomed in to look for JPEG artifacts, and indeed they are there. So yes at some point they did use JPEG.
This begs the question as to why would they take JPEG images, then save them as PNG? It defeats the purpose of lossless compression! It makes me wonder if UIUC is truly cognizant of what they were doing graphically.
The new color scheme images, such as this one
http://arctic.atmos.uiuc.edu/cryosphere/NEWIMAGES/arctic.seaice.some.000.png
Don’t appear to have been JPEG compressed. I don’t see the edge artifacts common to JPEG compression.
According Steven Goddard’s graph of the differences between 2007 and 2008 sea ice areas,
http://spreadsheets.google.com/pub?key=pj0h2MODqj3jAfVXaaKHbCQ&oid=5&output=image
on Aug. 4 there was a 41 percent difference and on Aug. 5 there was a 19 percent difference. Since it would seem unlikely during August for the sea ice area to significantly increase from one day to the next, this implies that the sea ice area between Aug. 4 and 5 in 2008 decreased by around 18 percent. An 18 percent decrease should be visible by eyeballing the maps. Such a large change should be obvious by comparing the maps on the CT site. Why isn’t it?
There are several other days in which the sea ice area in one year or the other would have had to have changed by more than 10 percent in a single day. Why don’t these large changes show up in “recent ice area” graph on the CT site?
Based on the numbers on the “recent ice area” graph on the CT site, there would have had to have been at least 3 days with sea ice area decreases of more than 350000 km^2 during a 19-day period. Does this seem at all plausible? I
AndyW
“Has UIUC been emailed to see what they say on the topic? ”
I emailed UIUC way back on Sunday about how they create their images, but I haven’t received a reply.
“It does beg the question though that if the data is right, and having read all this I still think it is, then why are the “popular” graphs at Cryosphere seemingly so far out of kilter?”
As has been pointed out several times, the “popular” graphs, NSIDC’s data, the map at NCEP, and the Bremen maps all agree. Only Steven’s results don’t fit. What is it about his data that makes it seem more credible?
Jeff (09:09:00) says:
Only Steven’s results don’t fit. What is it about his data that makes it seem more credible?
As I noted above, after some initial difficulties with colour distortion in the images, I was able to reproduce his results. I also checked my calculation visually by masking out the counted pixels on the original image and it looked OK.
What contrary evidence do you have that would diminish the credibility of this calculation?
dipole
“What contrary evidence do you have that would diminish the credibility of this calculation?”
The graphs on the CT site agree with NSIDC’s data. The sea ice map on the NCEP site agrees with NSIDC’s data. The sea ice maps on the Bremen site agree with NSIDC’s data. The large jumps in differences from day to day in Steven G.’s graph require changes in ice area that aren’t apparent by looking at UIUC’s maps, nor by looking at the maps on other sites, nor by looking at any of the graphs on the UIUC site. So far no one has mentioned another data source that agrees with Steven’s results.
I’ve posted the actual areas that I calculated from NSIDC’s data. If someone is willing to post the areas derived from the UIUC maps, it might add something to this discussion.
Pamela and others:
Here’s map of the world which I’m sure you’re all familiar with:
http://www.creativeforce.com/productimages/World_MER_2.gif
‘Eyeball’ it and tell me which has the biggest area, Greenland or Australia.
Talking of images, the direct NW passage is either nearly open or pretty well closed depending on what you look at.
http://iup.physik.uni-bremen.de:8084/amsr/arctic_AMSRE_nic.png
http://www.seaice.dk/iwicos/latest/amsr.n.ice.20080819.gif
http://ice-glaces.ec.gc.ca/Ice_Can/CMMBCTCA.gif
http://manati.orbit.nesdis.noaa.gov/ice_image21/D08232.NHEIMSK.GIF
There’s not too much consistency there to me to be honest. Are all these just pretty to look at in the main?
Actually they’re all very consistent as well as CT and NSIDC with the exception of the canadian site which appears to have a problem in that area (it doesn’t even agree with their own local radar which does agree with all the others).
http://ice-glaces.ec.gc.ca/prods/WIS56CT/20080818180000_WIS56CT_0003925512.gif
A lot of credibility, as far as the impartial observer (if that’s possible) is concerned, has been hung on the argument that artic ice would not revert back to 2007 melt levels. It’s looking bad at this point. What a differnce 5a week makes. I’m really rooting for the skeptics to be right overall. More science, less politics/wishful thinking.
Steven Goddard’s graph contained 18 1-day intervals. In 9 of those intervals, the difference increased and in the other 9 the difference decreased. If you make the very generous assumption that the 2007 sea ice area only decreased on days when the difference increased and that the 2008 sea ice area only decreased on days when the difference decreased, i.e., for each year there was NO melt on half of the days, and plug in the percentages from the graph, you find that the ice area would have had to have decreased by 38.6 percent between Aug. 1 and 19, 2007 and 40 percent between Aug. 1 and 19, 2008. More realistically, there would be a decrease on most days so that the graph implies that about half of the ice must have melted in the 18-day period in each year. The UIUC graph shows a decrease somewhere in the 20-25 percent range for Aug. 1-19, 2008. Can someone explain this apparent contradiction?
Jeff,
“Can someone explain this apparent contradiction?”
Hmm…well, you know, the blogosphere’s “tendency to knee-jerkingly blame everything on “global warming” [alarmism] makes for an easy story – but it is not based on solid science”, to borrow the words of our investigative reporter?
Jeff (10:39:21) :
There are 3 things here –
1. The UIUC/CT images
2. Steven Goddard’s interpretation of these
3. Everything else
You claim everything in (3) is consistent. I have persuaded myself (1) and (2) are consistent. If you want to argue with this I can only suggest you check it for yourself. Otherwise your criticism of (2) would be more accurately directed at (1).
Phil. (12:50:53) says:
‘Eyeball’ it and tell me which has the biggest area, Greenland or Australia.
I thought the discussion was way past ‘eyeballing’ and map projections.
dipole
“1. The UIUC/CT images
2. Steven Goddard’s interpretation of these
3. Everything else
You claim everything in (3) is consistent. I have persuaded myself (1) and (2) are consistent. If you want to argue with this I can only suggest you check it for yourself. Otherwise your criticism of (2) would be more accurately directed at (1).”
-NOT. Several people have described problems with Steven Goddard’s method. If his interpretation of the UIUC/CT maps is correct, then UIUC/CT’s interpretation of UIUC/CT’s maps is not correct, because their graphs indicate a 10-15 percent difference between the sea ice now and a year ago.
Steven Goddard’s results require implausible changes in the sea ice area from 1 day to the next. Do you really believe that there was an 18 percent (or more) decrease between Aug. 4 and 5, 2008? That’s over 800000 km^2. When the sea ice is decreasing by an average of around 1 percent per day, what could possibly cause it to decrease by 18 percent in one day? If there is a deliberate attempt on the part of AGW supporters to trumpet the decline in arctic sea ice, don’t you think that they would have made a big deal out of the loss of 18 percent of the ice in a single day? And made a bigger deal out of the loss of half the ice in less than 3 weeks?
Implausible results should lead to questioning the method. At least that’s my belief.
Jeff (18:55:09) says:
Do you really believe that there was an 18 percent (or more) decrease between Aug. 4 and 5, 2008?
My own pixel count gives a +1% increase between those days. Previous to that I was comparing 20080811 and 20070812 where my pixel counting gave the same approx 30% difference that Steven G. claimed.
So I don’t support the 18%, wherever it came from. If Steven G. got that figure he should certainly check his computation. But where does this leave your case against pixel counting?
I hereby coin the word ‘pixicle’ for a map pixel representing >85% sea ice coverage. It appears as a username on Google, but no previous usage in this context, I think.
Phil, I didn’t compare Australia to Greenland, and as a teacher, I am WAY past understanding map distortion. That concept is 5th grade stuff. But regarding Arctic ice, I looked at the Arctic picture from last year, same date, compared to this year, same date. More ice. Gonna be cold this winter. This isn’t rocket science.
dipole
“My own pixel count gives a +1% increase between those days.”
Did you also compare the corresponding days in 2007 (which might be 3-4 rather 4-5)?
Jeff (20:47:06) :
Did you also compare the corresponding days in 2007
Pixicle counts
20070805 30947
20070806 30967
20080804 36278
20080805 36574
dipole,
Based on your results, I’m guessing that Steven Goddard compared days straight up rather than offsetting by 1 day for leap year as he did for his The Register article. Otherwise, your results don’t agree with his.