“GISTEMP is High“. Steven Goddard discovers that temperatures measured at ground-level are not the same as temperatures measured in the troposphere. Well, learning is a good thing I suppose.
He also discovers that different data-sets have a different geographic basis! The GISS data-set attempts to include high Arctic value, while the HadCRUT3 data-set (which Steven calls “Had-Crut”) excludes those values. Shocking.
The dastardly GISS data-set estimates 2010 to be the “warmest ever“, while HadCRUT3 expects 2010 to be the “fifth warmest ever“. It seems that Steven has been reduced to trying to cherry-pick the data-set that shows the least warming. Whatever happened to the claims of cooling? An inconvenient memory apparently.
Wott's Up With That? 
HadCRUT IS LOW
GIVEN: “Current GISS Global Surface Temperature Analysis,” Hansen et al(2010) Section 7
Click to access gistemp2010_draft0319.pdf
This is the paper that Steve links to. GISS extends the shore temperature anomalies to cover the Arctic areas without measurements.
HadCRUT just excludes these areas. The net effect of this is to give the HadCRUT Arctic the mean hemispheric temperature anomaly. This results from the eventual simple averaging of the two mean hemispheric anomalies, as if they covered equal areas.
Obviously the GISS shore anomalies estimate is more representative, than the HadCRUT in-effect-estimate, of the mean hemispheric anomaly. And, it explains why GISS is higher than HadCRUT.
What would denier’s do? Would they really exclude the region with the greatest warming from the global averages, by withholding any form of area weighting? And call that representative?
In the same section (7) Hansen has provided them a sample of how this would turn out. Figure 13 shows masked (‘GISS anomalies limited to HadCRUT area’) results, and they are similar to HadCRUT’s. Then 1998 becomes the warmest year…and I suppose the earth returns to a cooling phase.
GIVEN: The ECMWF (European Centre for Medium-Range Weather Forecasts) study, as quoted at Climate Progress (Feb 25, 2010).
“The ECMWF analysis shows that in data-sparse regions such as Russia, Africa and Canada warming over land is more extreme than in regions sampled by HadCRUT.”
http://climate progress.org/2010/02/25/met-office-re-examine-of-climate-data-temperature-record/
As I understand it, the ECMWF analysis just takes the GISS arguments about HadCRUT’s Arctic exclusions, and applies them to HadCRUT’s other exclusions, in Russia, Africa and Canada.
GIVEN; A NCEP re-analysis discussed at Real Climate (18 Nov, 2008) also looks at the hole-in-the-Arctic.
http://www.realclimate.org/index.php/archives/2008/11/mind-the-gap/
GIVEN: “…the low Arctic sea-ice extent over the last summers are independent evidence of high Arctic temperatures.” And, the low sea-ice volume…
(Sorry this is so long. I’ve just had too much time to work on it.)
[Well said. – Ben]
One More…GIVEN: “If It’s That Warm, How Come It’s So Damned Cold?” Hansen et the same al 2010. This is a popular version of some of the above GISS, Section 7.
Click to access 20100115_Temperature2009.pdf
“HadCRUT just excludes these areas. The net effect of this is to give the HadCRUT Arctic the mean hemispheric temperature anomaly. This results from the eventual simple averaging of the two mean hemispheric anomalies, as if they covered equal areas.”
Exactly, so by citing HadCRUT3 as global the skeptics are implicitly infilling the arctic as you mention.
Additionally, notice that RSS data doesn’t go above 82.5N
So yep, despite Steve Goddard’s opposition to “extrapolation”, he has “extrapolated” tropospheric temperature above 82.5N from RSS 60N-82.5N data.
You just can’t make it up.
The main thing that bothers me though is how such skeptics argue extrapolation is wrong with nothing more than pointing out it means estimating temperatures. They don’t bother doing any analysis of temperature maps to see whether such estimation works. They don’t even bother looking into Hansen’s original work on verifying the extrapolation method in GISTEMP.