By Steven Goddard
GISS has explained their steeper temperature slope since 1998 vs. Had-Crut, as being due to the fact that they are willing to extrapolate 1200 km across the Arctic into regions where they may have no data – whereas Had-Crut prefers to work with regions of the Arctic where they actually have thermometers. WUWT reader “Wren” suggested that I compare GISS Arctic trends vs other sources to see how they compare. GISS has been showing Arctic temperatures rising very fast, as seen below.
However, GISS Arctic temperatures have been rising much faster than other data sources. The graph below shows the difference between GISS and RSS (GISS minus RSS) Arctic temperatures.
And the same graph for UAH.
Conclusions: GISS explains their increases vs. Had Crut as being due to their Arctic coverage. Their Arctic coverage is poor, and they rely on extrapolations across large distances with no data. Comparisons with other data sources show that GISS extrapolations across the Arctic are likely too high. In short, GISS trends over the last decade are most likely based on faulty extrapolations in the Arctic, and are probably not reliable indicators of global or Arctic temperature trends during that time period.
AnonyMoose, you are correct, they are making recent measurements though it’s been more than 1 year of them. Nevertheless I shouldn’t have said “more”. It is a sign of warming however to have the methane being released.
Craig Moore says:
May 20, 2010 at 9:58 am
James Sexton says:
May 20, 2010 at 9:39 am
“I don’t think coy non-answers do much for someone who writes for this blog. Credibility is important.”
Lol, I wasn’t trying to be coy, I was simply restating what was said. He said he’s not the same guy. What else is there to say? But, now that you’ve brought it up, if I were to discover a fundamental truism that had otherwise gone unnoticed that I wouldn’t have any credibility because I lacked the proper amount of published material or the proper amount of letters in front and behind my name? Isn’t it possible to study and critique assertions based on the assertion itself rather than the reputation of the person making the assertion?
Phil Clarke
I can assure you that there is very little (if any) correlation between the climate/weather of California and R Gates house in Denver. Some years California is flooding, with drought in Colorado – and vice-versa.
JDN says:
May 20, 2010 at 9:40 am
How does false extrapolation cause warming? I view these datasets as works of fiction on several grounds; lack of coverage, lack of consistency in siting and equipment, willingness to arbitrarily adjust temperature upwards, lack of UHI correction, willingness to remove inconvenient stations, lies, lies, lies, etc. But the fact that the arctic isn’t covered well should not produce any more bias than is already present. I’m not sure you’ve uncovered what is going on.
________________________________________________________________________
You forgot two other points. The change in reporting – use an M instead of a minus sign
and the “infilled data” (WAG – wild donkey guess)
If a reading is high due to a dropped minus sign in the real data it is smeared over several grids. Therefore ONE dropped minus in Canada can have a major effect with no balancing bias.
see
“GISS & METAR – dial “M” for missing minus signs: it’s worse than we thought” http://wattsupwiththat.com/2010/04/17/giss-metar-dial-m-for-missing-minus-signs-its-worse-than-we-thought/
jeff brown,
Re NSIDC Arctic stations:
“Temporal coverage varies by station, with the earliest record in 1913 and the latest in 2002.”
R gates
“acceptable” by whom, your cheating friends?
stevengoddard says:
May 20, 2010 at 5:04 am
nick
GISS doesn’t have any consistent thermometers north of 80 degrees. They just make extrapolations across huge distances.
Instead of talking about RSS and UAH “Arctic” temperatures as though that means the entire Arctic, why don’t you break down where the NOAA satellites are actually measuring ?
UAH, RSS and UW (University of Washington) all use the NOAA satellite data, but process it differently (issues concerning orbital decay, subtracting stratosphere signals from the troposphere returns, etc). The NOAA satellites since 1978 were in polar orbits, but exactly what were/are their inclinations ?
The very first satellite in this data series, TIROS-N, launched Oct. 24, 1978, had an inclination of 98.9127 ° (no data above 81.08 ° N)
NOAA 12: inclination was 98.7 ° (no data above 81.3 ° N)
NOAA 16: inclination was 99.1 ° (no data above 80.9 ° N)
NOAA 17: inclination was 98.6 ° (no data above 81.4 ° N)
NOAA 18: inclination was 98.74 ° (no data above 81.26 ° N)
AQUA: inclination was 98.14° (no data above 81.86° N)
All these satellites were used in the multi-decadal “satellite temperature dataset”, as analyzed by UAH for instance:
http://vortex.nsstc.uah.edu/data/msu/t2lt/readme.13Apr2010
Look at a map of the Arctic – notice the 80° N latitude circle:
http://clasticdetritus.files.wordpress.com/2009/07/arctic-svalbard-map.jpg
Here’s the Arctic circle:
http://upload.wikimedia.org/wikipedia/commons/thumb/3/31/Arctic_circle.svg/1000px-Arctic_circle.svg.png
The satellite temperature record ignores roughly 9° latitude at each pole.
GISS makes a “best guess” (extrapolations/interpolations from the closest measurements) for the temperature anomalies (not temperatures) for these polar unmeasured areas because they are comparing global temperature anomalies from their GCM models to the entire globe.
Your graphs (above) comparing GISS to UAH and RSS “Arctic” temperatures does not mention that the satellite data has a huge, empty hole in the Arctic, which they make no attempt to estimate. The satellite “trends” which you show, above, are just for part of the Arctic, unlike the GISS trends.
GISS would gladly stop their extrapolation/interpolation from measurements within 1200 km if the Arctic was fully measured, either on land or by satellite.
Steve M. from TN says:
May 20, 2010 at 7:31 am
Wren:
I agree that differences in anomalies can be meaningless if the same things aren’t being measured, but not necessarily. The temperature probably is different on the sunny and shady sides of my house, but the trends should be the same. So if similar things are being measured, we may be able to do meaningful comparisons of their anomalies.
I doubt the trends are the same on both sides of your house. Do you think the trend on the sunny side stays the same depending on the clouds? I would think cloudy days vs sunny days would change the trend from the sunny side to the shaded side.
=====
Over a span of several years I don’t know why that would make a difference in the temperature trends for the two sides of my house unless there was a cloud trend. I would be very surprised if the two trends diverged over a decade or two.
So the question is there Phil old boy, can you tell the temperature (anomaly or absolute) in Paris from one thermometer in Bedfordshire England 400km away?
You can make an estimate provided you know the average or baseline temperature in both locations. If it is 1C above average in Bedfordshire, the research shows it is also likely to be above average in Paris. As in all thing there are uncertainties and these are understood and quantified. 1200km was chosen as the point at which the correlation drops below 0.5. This really is fairly fundamental, uncontroversial and built into the GISTEMP code. Did you read the paper?
Or perhaps one could use a barometer, based on our new understanding of the influences that atmospheric pressure has on surface temps? 😉
Alexej Buergin says: May 20, 2010 at 11:37 am
There is the Amundson-Scott station at the south pole.
I wonder: How would one proceed to have a station in the vicinity of the north pole?
Just ask Mr. Medvedev to attach a buoy to the titanium flagpole that Russians planted there few years back
BTW An archive of material that belonged to a member of Captain Scott’s doomed expedition to the South Pole including a journal, pictures and photographs is to be sold at auction.
http://www.telegraph.co.uk/news/worldnews/antarctica/7745181/Pictures-from-Captain-Scotts-expedition-to-go-on-sale.html
Alexej Buergin says:
May 20, 2010 at 8:53 am
“R. Gates says:
May 20, 2010 at 7:04 am
I await your next Arctic Sea Ice update, as temps remain high in the arctic region (as they’ve been for the whole winter and spring) and the extent for 2010 year-to-year data has now fallen below 2007, 2008, 2009, 2005, & 2003.”
Please learn that there is more than one organization estimating arctic ice extent. Nansen shows it “normal” for 2 month now and practically at the 2009 level, clearly higher than 2007 and 2008. Your scientific credentials might improve if you mention what you are citing, a good practice in any case
___________
I use IJIS/JAXA Data right from their site, just for consistency. Other’s bounce around from site to site, but I find this data to be the most immediately accessible, and they don’t use any too outrageous of algorithms as I think others do.
http://www.ijis.iarc.uaf.edu/en/home/seaice_extent.htm
Still, I think the single best graph on Arctic Sea ice anomaly is here:
http://arctic.atmos.uiuc.edu/cryosphere/IMAGES/sea.ice.anomaly.timeseries.jpg
It is the best long term graphic display of the “pulse” of how arctic sea ice is doing in terms of year-to-year extent.
skye says:
May 20, 2010 at 7:57 am
Steve/Anthony, when are you going to mention the Arctic sea ice? It has dropped below 2007 and is at the previous record for this time of year set in 2006. Why are you remaining so quiet about this? Can you do a post as to why this is occurring?
I don’t think it has dropped below 2006 yet ( but the data is 1 day old, so maybe…)
http://www.ijis.iarc.uaf.edu/en/home/seaice_extent.htm
Zooming in, you can see 2010 is dropping at a steep angle but has not yet reached 2004 and 2006, but probably will within a couple of days.
And don’t forget the University of Washington Arctic sea ice volume estimate is at an all time low since the beginning of satellite data (last updated 5/13/2010)
http://psc.apl.washington.edu/ArcticSeaiceVolume/images/BPIOMASIceVolumeAnomalyCurrent.png
Expect even less Arctic sea ice volume come the September minimum.
Why are you remaining so quiet about this?
I’m sure they’ll get to it – so many topics, so little time.
Phil Clarke
The whole idea of rapid climate change is that the climate is changing rapidly. GISS is showing a 10C/century trend for the Arctic, which would invalidate any baseline climate assumptions.
Feel free to use lots of words like “fundamental” and “uncontroversial” though. They make you sound very authoritative. ;^)
Vuk etc. says:
May 20, 2010 at 12:44 pm
Next auction: Consensus scientists for the dozen. “They can make all the models you need to enforce your climate related local policies/businesses”
[snip – this is all off topic, all speculative, and all your own opinion about skeptics and conspiracy theory. It’s not relevant to the discussion at hand, which is differences between GISS and satellite data sources. Look I get what you are trying to do, but don’t drag threads off topic. When one comes up about this subject, then you can post something like this. – Anthony]
Oops, sorry, wrong post.
Phil Clarke says:
May 20, 2010 at 12:38 pm
Why is it that the size of the hot spots I see on global graphs appear to be inversely proportional to the number of actual temperature measurement points within that area?
Secondly “a correlation of 0.5 “. You have got to be kidding. Why not stop at a correlation of 0.9.
Thirdly I would expect the distance versus correlation coefficient to vary with latitude and the 0.5 distance to be much smaller as you approach the poles due to the variation of TSI with latitude.
Gail Combs says:
May 20, 2010 at 11:53 am
Gail, perhaps you should actually bother to read some of the papers that discuss the factors contributing to Arctic sea ice loss. Then you would realize how foolish your statement is. Your beloved negative AO state this winter certainly isn’t helping you out much now is it? Guess the really warm air temperatures in the Arctic all winter that continue today have absolutely nothing at all to do with the current rate of decline being observed. But hey, if you’re right that it’s all about the ocean temperatures, then Wieslaw Maslowski’s predictions of no summer ice by 2012 might be spot on.
Also Gail, since the winds and sea ice correlate at 0.3 and the air temperatures and sea ice correlate at 0.7, which do you think is more important? Oh but wait, I know better than to say that because statistical correlations while they may help illuminate relationships do not prove cause and effect.
Sorry Phil that is completely lame. A correlation of 0.5 between two sets of data that are subject to the same underlying conditions – they are on earth, in the same hemisphere and at a similar latitude – is just about meaningless. How on earth can you or any of those you are defending think that you can use it to project temperature anomalies across vast swathes of the globe that are devoid of thermometers? Climate science has truly been debased if people are prepared to defend such SWAG methods.
And yes, I looked at NSIDC, and as posters have said their data are restricted to Greenland and Iceland and stopped in 2002. Have a look at this and be afraid.
http://blog.qtau.com/2010/05/dude-where-is-my-thermometer.html
I wish some of the very capable persons commenting here would wake up and actually start fighting science with proper science. These reported temperatures are tainted by the selected few station chosen out of some 10,000+ stations available, we know that.
Take an alternate set of stations but still use the homogenizing algorithms and 1200 km extrapolation rules that GISS uses to create its temperature maps. GISS ignores some 9600 stations, let’s just ignore another different 9600 satations. Possibly even picking 3 or 4 different sets of stations that conform to GISS adjustment rules and see what the differences appear. Maybe a network such as SurfaceStations.org.
In that manner nothing can be claimed “unscientific” or improper because the exact same rules are being used to create the sets as GISS uses to create their own set. Get it peer-reviewed by proper peers.
Maybe then some real progress could occur, or, am I missing something that would prevent that from becoming a reality, or, would there never be a difference at all (but I can’t imagine that)?
Anu, skye,
No need to wait:
http://wattsupwiththat.com/2010/04/28/disconnected-computer-modeling/
Phil Clarke:
In that case, what on earth is the point in attempting to extrapolate the temperatures across the Arctic? If you’re only interested in the anomalies, and the anomalies are likely to be the same, why not simply leave the Arctic ‘data’ out altogether?
Erm, as the whole greenhouse theory has radiative balance at its core, and as blackbody radiative energy is proportional to the forth power of absolute temperature, I would say that the local absolute temperature is very far from being irrelevant.
nedhead: Greenland was warmer seventy years ago. This inconvenient truth was again ignored when some climate researcher told us on Norwegian TV a couple of hours ago that the melting of the Jakobshavn glacier was proof of global warming.
stevengoddard says:
May 20, 2010 at 1:39 pm
Steve, Huh? how does that relate to the questions posted by Anu and Skye? Honestly, you were quick to jump on the fact that the ice was approaching “normal” this winter, but you are very quiet about the ice now approaching a new record low. Come on, be unbiased and talk about both.
In that case, what on earth is the point in attempting to extrapolate the temperatures across the Arctic?
Equally, why throw away information? I am bemused that this is problematic. Imagine two idealised thermometers at the same location, clearly they will show a correlation in their readings of 1. Now move one away; as the distance increases, so the correlation will, on average, drop. Analysis of real world locations shows that the correlation drops to 0.5 at about a distance of 1200km. And yes, that is an average figure and yes it does vary with latitude – read the paper.
Given that we know this – why not use that information to interpolate an estimated temperature where coverage is sparse? That is what GISTEMP does, and it is all that GISTEMP does. The arguments about using a thermometer in Paris to tell the temperature in London or whatever is just a rather obvious and apparently ignorant straw man.