Another GISS miss, this time in Iceland

Ágúst Bjarnason, it is not a question of whether the Icelandic Meteorological Office is reliable or not. Nobody is doubting that. The question is whether the rush to judgement exhibited above is justified. As it stands, and allowing for Nick Stokes comment that GISS do their own adjustments, we have four sets of data for Reykvavik. We have the Icelandic Met Office (IMO) data and the GHCN unadjusted data, which essentially agree. We also have the GHCN homogeneity data and the GISTEMP UHI adjusted data (described by Hansen et al 2010 http://pubs.giss.nasa.gov/docs/2010/2010_Hansen_et al.pdf), which also essentially agree but disagree with the first two. That is all that is established by Paul Homewood.
At that point you can apply a decision tree.
Is the IMO/GHCN-QCU data more accurate a record of Reykjavik temperatures than the GISTEMP?GHCN-QCA data?
If yes, is their a good reason for the adjustment based on the different uses for the data (as Nick Stokes plausibly suggests)?
If no, is algorithm used to make the adjustment valid, ie, unlikely to make an error except rarely, and in unusual cases? (Again, Nick links to information that plausibly suggests this is the case.)
If yes, is the situation in Reykjavick such that the algorithm made a mistake their due to unusual circumstances?
It is only when you get to the end of the full decision tree (rather than the schematic above) that that there can be any plausible suggestion of wrong doing by GISS. Yet without discussion of any point other than that there is a difference in the records, comentors above are all set to call in the lawyers. The proper conclusion is that the justice they are after is lynch mob justice.

Too right, DirkH. And let’s not forget that a warming trend in the Arctic/sub-Arctic is twice as valuable to CAGW proponents as one elsewhere. I doubt this adjustment would have been made back in the early 80s when climatologists thought the Antarctic would warm at least as fast as the Arctic (those climatologists who didn’t believe the ice age was imminent, that is).

Tom Curtis:
This is a time when you throw the “paper” out the window and look at facts, then look at why the error occured.
I am sure most of the data is done by computer.
GISS should be thankful that someone was diligent about local/regional conditions and is bringing this to GISS attention so that the reliability of their temperature projection has credibility.
As it stands, it is extremely obvious to anyone that there is something very wrong with how the algorithims have been developed.
The next feat will be to see if this regional error is extrapolated over the Arctic as a whole and other areas of the world.
We need reliable temperature sets. At this time GISS is not one of them and should be.

The suggestion from Tom Curtis and Nick Stokes that these adjustments arise from a homogenisation correction does not survive a moment’s scrutiny.
Just look at the unadjusted GHCN data graphs at
ftp://ftp.ncdc.noaa.gov/pub/data/ghcn/v3/products/stnplots/6/
Click on the eight icelandic stations (ones starting with 6200…) and you’ll see that all six of the stations that go back far enough show a warm period 1930-1940.
Maybe somebody could put these 6 raw-data graphs together on one page?
The GHCN adjustments are wrong and the GHCN adjusted data should not be used.

Paul Homewood, it is entirely possible that the sudden temperature changes in 1939 and 1941 confused the algorithm. It is also possible that those sudden changes are due to some extraneous factor resulting in high readings at the time. Which of the two is not clear from the data that you present. What is clear, is that this situation should be looked at more closely to either determine that the algorithm is acting properly; or that it fails in this circumstance, and therefore not use the algorithm in this instance. On that point I am not criticizing your post at all. I am, of course, strongly criticizing the rush to judgement exhibited by some of the comments.
By the way, you may be interested in the IMO’s comments on 2010:
“The average temperature in Reykjavík was 5.9°C, 1.6°C above the 1961 to 1990 mean, and has only once been higher, 6.0°C in 2003. It has twice before been equally high as now, in 1939 and 1941. The temperature has now been above average (1961-1990) for 15 years in a row. Nine months registered above average temperatures, three were below. January had the largest positive anomaly, almost 3°C, October the coldest, about 0.5°C below average.”
http://en.vedur.is/about-imo/news/2011/nr/2112
Clearly the IMO is using the unadjusted figures and considers them sound. Clearly also the temperatures in the 21st century are warmer than those in the 1930’s/40’s in Reykjavik even on unadjusted figures, although only just.

Caleb says:
January 26, 2012 at 2:14 am
It is interesting that the GISS map shows it as being so “warm” up in the Bering Straits in December, but when you look back at the maps of how the arctic ice built up this fall and winter, the Bering Straits were the one place with more ice than normal.
________________________________
Caleb,
Total Arctic ice isn’t dependent as much on temperature as it is on “storminess”. December ice was at the highest levels seen in years, but has experienced precipitous loss twice since, due to large storms breaking up the mass and ejecting slush out through the Fram Straits, etc.
Wind and ocean currents rule Arctic ice, not temperature.

@Tom Curtis:
Please give us just one clear example where the same algorithms have resulted in a cooling trend instead of a warming trend. Because every adjusted dataset I have seen either has no effect or results in a more pronounced warming trend. Remember, skeptics don’t have to prove anything but we have already shown that these algorithms produce unacceptably skewed datasets. It is up to Team AGW to prove that these erroneous corrections are exactly cancelled by erroneous corrections of the opposite trend – or simply run their data through a better algorithm.
Alternatively we sit and wait. Because the data of 10 years ago is now fixed and there can be no justification for further “corrections”. So we wait to see what these electronic thermometers in their new sites next to major airports will be telling us in 20 years time. So far they have been flatlining since the “great adjustment” was made – I have a suspicion they will still be flatlining in 20 years time despite China pumping out enough CO2 to terraform Mars.