By Steve Goddard
In recent articles, I have have been discussing GISS’ claim that their divergence from Had-Crut over the last decade is due to better Arctic coverage.
They use the two images below to justify their claims.
The GISS 2005 image is deceptive, because GISS uses 1200km smoothing. Their actual 2005 Arctic coverage (using 250km smoothing) is below, and shows that they have very little data north of 75 degrees.
GISS 2005 Annual Map – 250km smoothing
In order to make it easy to visualize the differences in coverage between GISS and Had-Crut, I mapped them onto 3D spheres, which eliminates the infinite distortion near the poles in this favorite GISS projection. I also threw in a recent NASA/NOAA/UAH equivalent map. None of these images have been cropped.
![[Image]](https://i0.wp.com/discover.itsc.uah.edu/amsutemps/browse/AMSU_A_15.latest.a_01.png?resize=481%2C273&quality=75)
NOAA satellite map used by UAH
The images were created by taking the maps, replacing the gray areas which “signify missing data” with black, scaling the images to 512×512, and mapping them on to an OpenGL sphere. Assuming that the original maps all reach 90N at their top, the images are an accurate representation of their coverage. That probably is not exactly correct, but is close. It appears from the NOAA/UAH map geography that they have coverage further north than 82.5°.
Conclusion: GISS claims of better 2005 Arctic coverage in their recent paper are not justified.
UPDATE: Steve has rendered this video for further discussion.
Nice work Mr. Goddard. Guess we can call off the code brown. (Monsters vs. Aliens reference. I got kids)
All this warming in the North Country has made for a lot of precipitation – unfortunately at this time of year it comes in flakes – our typical Alberta Victoria Day long weekend usually gets snowed on and it did not fail to disappoint this year (or the last three for that matter: http://www.theweathernetwork.com/your_weather/details/786/2602287/18/upload/
In spite of global warming, we still can’t really do a lot of planting before the long weekend in May except for hardy strains. “La plus la change, la plus la meme chose.”
The big red areas in the north don’t mean much if they mean it is 9 C below instead of 10 C below cause my hay won’t grow at either temperature.
899 says:
May 21, 2010 at 6:13 pm
Excellent post !! Cheers !!
O/T: Some weather fun.
(via Luboš Motl’s blog)
Hailstone Fury
/dr.bill
stevengoddard says:
May 21, 2010 at 8:38 pm
dr.bill
My purpose in including the UAH maps was just to show the spatial coverage, which is the vast majority of the planet.
I don’t really understand the color coding either. Dr. Spencer would probably be the best person to explain. It would be interesting to hear an explanation.
Because it’s the brightness temperature with an unknown emissivity, that’s why it’s not used. That’s why UAH and RSS use the microwave emissions of O2 to derive their temperatures (they are at altitude to minimize the surface contribution).
Because of errors associated with the direction of scanning as the satellite passes near the pole the method can’t make lower tropospheric measurements within 7.5º of the pole. You really shouldn’t discuss this material when you don’t understand how it’s done. You haven’t included the UAH maps, but I guess you’ll continue in your delusion.
The importance of all this lies in realising just how much we don’t know particularly in the context of claims of unprecedented Arctic warmth.
Steve, you do a great job creating nice looking graphics but I once more I think that you miss the mark in terms of relating what is really happening in the Arctic right now. Now even though 2010 Arctic Sea ice has fallen below even 2007 levels for this same date, this is not nearly as important as the Arctic sea ice volume anomaly, which is showing rapid negative declines. I would direct readers attentions to these two pages:
http://psc.apl.washington.edu/ArcticSeaiceVolume/images/BPIOMASIceVolumeAnomalyCurrent.png
and also here, to undestand how the sophisticated PIOMAS volume model is validated:
http://psc.apl.washington.edu/zhang/IDAO/index.html
Steve, et. al. here on WUWT were quite excited earlier this year when the Arctic Sea ice made a small late season surge, though myself and others tried to be realistic about this surge in pointing out that it was new very thin ice that formed from a series persistent low pressure systems over the Bering Sea mainly. Analysis of the winter however, with warmer than normal temperatures in the Arctic (as shown by GISS et. al) had failed to thicken up the multi-year ice in the same manner as might have been found in previous winters. Sure enough, the low volume of Arctic Sea ice as shown in the graph above is now being seen quite dramatically in a sea ice extent that has fallen a rate that one would normally expect for later in the heart of the melt season mid-June to Mid-August.
Steve’s insistence of harping on the GISS data techniques, which may or may not have data smoothing issues, which may or may not result in a few tenths of a degree change at most, when the arctic is showing several degrees of anomalies, seems spurious at best. and a distraction from the real issue– the rapid decline of sea ice volume in the Arctic. If the AGW skeptics doubt the PIOMAS volume model, then please (sincerely) let us know what issue you have with it. But as of right now, the Arctic sea ice is behaving exactly as one would expect if the volume is as low as PIOMAS says it is. We may not have lost as much ice through the Fram Strait this winter (due mainly to the very negative AO index), but the ice that remained did not get extremely thick, and in fact, has a very negative volume anomaly, which is exactly why it is melting so quick, even before the main melt season hits.
So winter at Prudhoe Bay, Alaska, has a common temperature of -35F and a wind speed of about 20mph from the west. If the anomaly was +30F it would mean temperature of a mere -5F. Yep, that heating will surely melt all the ice and flood the world….
Does it matter whether the Arctic temperature is -100F or -35F or -5F in the winter?
Another great post from Mr. Goddard!
@Phil
Is it just me or do all of the graphs at clearclimatecode.org
Look as if they have been run through the Mann Hockey Stick Program?
Maybe they are using “adjusted” data.
Interolation of data at 1200KM interval seems a stretch if I am understanding the concept!
dr.bill says:
May 21, 2010 at 9:56 pm
O/T: Some weather fun.
(via Luboš Motl’s blog)
Hailstone Fury
Yep, dr.bill, that’s where I live. Lucky again. We only found one lone 1″ diameter hailstone. That anvil-head was quite impressive and I remember many like that back in the 70’s & 80’s. That’s the kind you can get sprites and blue jets from I hear. Talk about a transfer of energy! I’m guessing forty thousands feet high+ and spread over some 200+ square miles by the jet-stream as the radar seemed to indicated. That’s where my mention of the dip of the horizon plays in. Tell me vertical height matters not in such cases. Have had my space heater on, again (never before in May), from the time of that storm until yesterday. It pulled the temperature departure way negative again for a solid week.
Unfortunately, that is just normal spring Oklahoma weather playing with the dry-line from the Texas panhandle, just weather, not climate. Perhaps a meteorologist from this area could fill in some missing gaps, I really didn’t pay that much attention to it once I could see the bulk was going to miss us by a couple of miles. I get really tired of the constant hype, we have always had weather like this and always will. For those who have just moved to Oklahoma, welcome, get used to it.
anna v says:
May 21, 2010 at 9:26 pm
Sometimes I think we are still in the middle ages discussing differences between archangels. We have accepted that average global anomalies have a meaning and reason merrily along on differences of representations.
Thanks anna, that needs to be re-raised regularly!
A complex story of biased or unsubstantiated corrections, alterations, assumptions and adjustments – this wouldn’t work in any other field of science, where this paper would just be an interesting theory. However, in the astrological world of ‘climate science’, this theory is portrayed as being hard facts.
This truly is “BS baffles brains” for the average person, let alone a politician.
I am reminded of the ancient guilds who controlled critical sections of the economy, where how to do or make something seemed obvious until you tried to duplicate it – then you would find it was impossible, unless you knew the deep secrets of the science of how it was really done.
O/T
In this web site Gavin tells us it’s not too late, calculates (economic also) impacts of polar melt down for any level of sea level increase.
http://www.scientificamerican.com/article.cfm?id=interactive-12-events select polar meltdown and look in the bottom for Dr. Gavin interview
Steve, a great visualization! Thanks!
Saying that GISS uses 1200 smoothing doesn’t describe it correctly, i think – smoothing implies something like interpolating, running a smooth filter like a Gaussian, but that’s not what they’re doing -they are filling an area with an anomaly value from a station up to 1200 km away; it’s a crisp cutoff at 1200km if i understand Hansen’s methodology correctly, nothing smooth about it. It’s a filling algorithm.
I’m just guessing here, but if your level of talent matches your level of arrogance, you must fairly glow in the dark. I bet it was hard to sneak back into the house when you were a kid, huh? Ease up man. The rest of us dummies are just trying to understand things.
/dr.bill
Just as a guide: 1200km (~750 miles) is approximately the distance from Land’s End to John O’Groats in the UK, or Dunkirk (Dunkerque) to the Pyrenees in France.
Phil,
There is something interesting in the UAH map, but not what you were expecting…
Hansen and Lebedeff’s 1987 paper outlining the rationale for as much as a 1200 km smoothing does not exactly provide a convincing argument for using it. It seems as though they already made their mind up to use it even though their own data does not really support it.
They show the Correlation Coefficients between stations at different distances (in several different latitude bands) and the numbers are very poor.
At 1200 km, the Correlations range from 0.60 to 0.35 (and everything south of 44N is very low).
If anything, the data indicates the smoothing distance should be kept to a minimum.
http://pubs.giss.nasa.gov/docs/1987/1987_Hansen_Lebedeff.pdf
The new high resolution NASA satellite temperature maps show a few hundred kms is probably okay but beyond that, the smoothing algorithm will introduce errors.
http://neo.sci.gsfc.nasa.gov/Search.html?group=67
The GISTemp anomaly for April was 0.73C at 1200 kms but only 0.60C at 250 kms.
R Gates: ‘2010 Arctic Sea ice has fallen below even 2007 for this same date’
Looking at current sea ice coverage, present day coverage for this time of the year seems to fall within a very narrow band which suggests this of all times is not a time we can make confident predictions about coverage.
You further say:
‘the low volume of Arctic Sea ice as shown in the graph above is now being seen quite dramatically in a sea ice extent that has fallen a rate that one would normally expect for later in the heart of the melt season mid-June to Mid-August.’
But again, given the very narrow band width of sea ice extents at this time of the year, suggesting the current extent is falling at a rate consistent with a June to August melt seems a trifle premature.
I think it’s too early to call.
stevengoddard says:
May 22, 2010 at 6:25 am
Phil,
There is something interesting in the UAH map, but not what you were expecting…
On the contrary it’s exactly what I was expecting, as I and others (Bob Tisdale, Anu, etc.) have been telling you UAH infills the pole from 82.5ºN on their maps.
dr.bill says:
May 22, 2010 at 5:02 am
I’m just guessing here, but if your level of talent matches your level of arrogance, you must fairly glow in the dark. I bet it was hard to sneak back into the house when you were a kid, huh? Ease up man. The rest of us dummies are just trying to understand things.
As we all are, it’s not helped by Steve making mis-statements and refusing point-blank to acknowledge that he has done so. Mixing up NASA surface brightness temperatures with the UAH MSU is a rather basic error for someone who’s been posting here as long as he has.
It’s not just on this thread:
Steven mosher says:
May 21, 2010 at 10:00 am
Stevegoddard.
You made a simple mistake. But in making that mistake you accused somebody else of hiding data. Bob and Zeke have both pointed out the error and without acrimony or innuendo about your motives. My suggestion would be that you admit your mistake and thank those who pointed it out.
The current behavior I see reminds me of Mann when he was caught out on simple errors. Two paths forward. choose.
REPLY: Mosh gives good advice, take it. – Anthony
Any CAGW’er needs to look very closely at the data for northern Africa. The original data designates one tiny area near the mid-eastern coast as what I’ll call warmer mid-red. The smoothed graphic shows ~60% of Northern as warmer mid-red, when that warmer temperature was non-existent in the original data but for one small spot. Then look at the differences near the USA-Canadian border (N. & S. Dakota, Mn. Ia. Wy. Ks. Ms. Ill. and parts of other states). Original data for 90% of that region has been smoothed to warmer mid-red. Come on Anu, explain that to me.
R. Gates, love that site you provided on volume models. Reminds me of the various scenarios for El Nino prediction models. Statistical versus dynamical. A very good site that we should all visit.
My bet is on the statistical natural climate models based on…wait for it…weather pattern variability and oscillations (IE if these weather system parameters exist, ice will do that), not the second one that includes projected warming scenarios from long-wave re-radiation as a result of increasing levels of greenhouse gasses over the Arctic (IE added warming temps to the natural variability Arctic system).
I noted that the explanations of statistical models were given a little “greenhouse gas” color commentary by the authors speculating that natural variability may be driven by greenhouse gasses. This is an unsupported statement and should have been made with proper references, which it was not.
Regardless, great site, and thanks for alerting us.
http://psc.apl.washington.edu/zhang/IDAO/index.html
I’d like to suggest that even if the world were not being torn apart by politicians, anarchists, and special interest groups around the planet over the issue of Man-Made Global Warming (and the trillions in dollars, euros, pesos, and gold so much at stake), that within scientific circles this issue would be as much debated as now – though with perhaps a “little” less of the understandably human emotional component. Life’s a beach, always different, always the same.
Brightness Temperatures:
For those like myself who didn’t know what a brightness temperature was, I dug through some of my books and determined what it means. To wit:
– An object emitting thermal radiation follows the shape of Planck’s Law. If you integrate that over all frequencies, you get the Stefan-Boltzmann result with the T^4 dependence. If the object isn’t a blackbody (pretty much always the case), then the emissivity “e” enters the picture, and the total emitted intensity is given by I = eσT^4 (in W/m²).
– When only a segment of the frequency range is measured, as is the case with the microwave sounding units on the satellites, the emissions from the object still follow the shape of Planck’s Law, and the brightness temperature “Tb” is defined as the temperature a blackbody would have if it were producing that amount of intensity in that frequency range. This quantity and the actual temperature “T” are related by the following expression:
1.00·[exp(hf/kT) – 1] = e·[exp(hf/kTb) – 1]
– In cases where the photon energies “hf” are much smaller than the thermal energies “kT”, which is the case with microwaves (where hf/kT is in the 1/100 range), the exponentials can be expanded in power series, and keeping only the first terms gives a much simpler expression:
T = Tb/e
– The emissivity “e” is always less than 1.00, and thus the real temperature will be higher than the brightness temperature, which, as I interpret things, is what the satellites can measure (but I didn’t dig that far). The emissivity, however, appears to be the big “if factor”, since it will vary widely across the planet, depending on a host of geographical and time-dependent factors. I have no idea how they find those values, but the actual temperatures derived from the brightness temperatures depend critically on them.
/dr.bill