Saturday, 10 August 2013

The Past is Getting Colder

Wanted: a detective. Mission: to track down Iceland's temperature records from a century ago.

These records would have been written in ink on paper. They may still exist in an archive somewhere. If they can be found, and images of the originals can be made public, then a scientific fraud can be exposed.

Last year I wrote a piece called GISS - Strange Anomalies, where I showed the gradual lowering of historical temperatures in the Arctic by the NASA division GISS - the Goddard Institute of Space Studies headed until recently by global warming fantasist James Hansen. It has got worse...

I presented screenshots from GISS in which they publish the "historical data". But the past is changing - or rather it is being changed by these scoundrels. Rather than becoming blinded by a blizzard of data I chose - in a tight focus - the first 3 months from 1900 at the Icelandic place of Teigarhorn. There's more detail at the GISS - Strange Anomalies page, so please do visit it. In particular it examines the basis for the warmists' claim that the Arctic is warming faster than where you live.

In November 2011 GISS were reporting Teigarhorn's temperatures around 1900 thus:

Year       Jan       Feb      Mar

     The temperature at Teigarhorn in Jan 1900 was 0.7C

They then changed their minds and in March 2012 reported thus:
     Year     Jan       Feb     Mar

The temperature at Teigarhorn in Jan 1900 was -0.2C


Today, at August 2013, late-Victorian Iceland has got colder still:

  Year       Jan         Feb        Mar

The temperature at Teigarhorn in Jan 1900: -0.5C
 One begins to worry about those poor people in Edwardian times. The past is getting colder fast!
You can take a look yourself at GISS's constantly shifting historical data. Try here:

If that link has broken then try here instead:
Then click on "Teigarhorn" and at the foot of the resulting page a link invites you to "download monthly data as text", and you should see the latest version of the matrices above.
What might their motive be for - ahem - restating historical data? This data, we may be sure, was once written down in unambiguous digits. Why has it been 'got at'?

My working hypothesis (which I will be happy to ditch given evidence refuting it) is that people within GISS are attempting to perpetuate the myth of Global Warming, and since today's temperatures stubbornly refuse to shift (flat since 1998!), they are 'making the past cooler' by fraud.


  1. It might be worth approaching the Iceland Met Office for a photocopy of the original records.

    Iceland seems to have a favourable attitude to FOI requests.

  2. Suggest you check when/how the data was converted from F to C - no field thermometer in 1900 could read a fraction of degree F, let alone degrees C

    1. The original Icelandic readings would have been in centigrade, as Denmark, the governing power in Iceland at the beginning of the 20th century, along with the rest of mainland Europe, used and still uses the SI system of measurement.

      There is no reason a field thermometer cannot be read to 0.1 C if sensitive and the scale covers a small range. You don't give the instrument makers of the time credit for the skills they actually possessed.

    2. You don't need to read to fractions. The fractions arise when working out monthly averages.

    3. In that case they need to report the precision. The decimal value is not a valid statement of the temperature if it arose from 1 degree or 1/2 degree original numbers.

      Averaging a number of readings allows one to state with some confidence where the actual average lies, but does not increase the precision of the data set or the result.

      It would appear that lowering the temperature by 1.2 Deg C is fraudulent as that amount exceeds the precision of the readings. It should still have error bars on it.

    4. Why not just ask GISS for an explaination rather than assuming missconduct?

  3. I think that the issue of units and instrumental error is not relevant here. The temperatures written down in 1900 - well or badly - are a reference point. That source data is being got at. How do we know? Because its modern-day guardians are reporting a single historical fact - that which was written - with forked tongue.

  4. Look, if the real world data disagrees with the models then there's only one thing to do...

  5. Is NASA trying to estimate what the temperature would have been in 1900 if the infrastructure and instrumentation were the same as it is today? If they are, and I'm not saying they are, then direct comparisons between then and now could be done. Because there is more development around the site, an historical temperature of 0.7 would equate to a -0.5 today at the current location. If they are doing this, then they must add uncertainty around the "corrected" value. The model by which the temperature is updated to reflect current conditions must be inexact.

    1. I can't agree ... "0.7C" in 1900 is still the same "0.7C" when viewed today. If the temperature in the same location in 2013 is measured at 0.7C it does not justify changing history to get context for UHI, etc., ... the 'modern' temperature data must be adjusted downward to reflect the environmental conditions of 1900 at that location, not the historical temperature data.

      However, what you describe is exactly (incorrectly) what is being done ... the past is being cooled in relation to the environmentally contaminated modern temperature record.

    2. If you wanted to adjust historic temperatures to harmonize them with modern UHI corupted data, you would have to increase historic temperatures by adding a UHI fudge factor. You defenetly would not lower historic data. That only makes the UHI problem worse.
      Unless, you goal is to show increased warming....

  6. I've looked at two other data sets for the same station.

    The NCDC gives 0.7 °C for data downloaded in 2008 and 2012.

    The CRU gives 1.0 °C for for CTUT3 and CRUT4.

  7. On Bishop Hill 'TinyCO2' has posted the following link.

    This seems to be the same the CRU version.

  8. The idea is to slightly cool the past (while also minimizing the LIA) to make the temp upswing off the later 20th look greater.

    I caught the NCDC doing the same thing with data for the lower 48, which reduces the warmth of the 20's and 30's while also attempting to make this decade look warmer and show that there has not been a pause

  9. Talk to Paul Homewood if you have not already.
    He has done loads of work on this.
    Paul Matthews

  10. For what it is worth here below is a link to the the English version of front page
    to the "official" temperature time series from the Icelandic weather bureau (IWB).
    ( I assume pre-computer era data are the "uncooked" data taken from the written records. though I can not vouch for it ).

    Look for a subheading saying

    "Longer series for selected stations"

    Data for Teigarhorn can be accessed through the selection menu
    in the line below that heading, there is also a handful of continuous
    monthly temperature series from other stations going back to around
    1930. And one other that goes all the way back to 1823 ( Station :
    Stykkishólmur ), but I think that the oldest part (from 1823 to 1845)
    of that one is a reconstruction, If that matters the man to ask at IWB is
    mr. Trausti Jónsson ( his work email can very likely be found somewhere on the
    IWB website, staff section).

    Now the giss numbers for Teigarhorn sure look a bit different from
    the ones recorded at IWB.

    Here is the part for January, February and mars from the IWB record
    for the period 1899 - 1908 ( in Celsius degrees)

    Jan Feb Mar
    1899 -1.5 0.2 -3.0
    1900 1.0 -2.2 -1.6
    1901 1.8 -0.2 0.0
    1902 -4.3 -2.0 -1.7
    1903 -1.0 -1.1 -0.6
    1904 0.4 -0.3 1.2
    1905 -1.3 -2.2 1.9
    1906 0.6 -2.0 -1.6
    1907 0.1 -2.3 -0.9
    1908 0.2 -1.0 1.0

    and tabulated differences from the very "august" record of the
    same by giss version 2013.( also in °C)

    diffs: Giss - IWB
    Year Jan Feb Mar
    1899 -1.4 -1.5 -0.9
    1900 -1.5 -1.5 0.2
    1901 -1.3 -1.0 -0.3
    1902 -1.2 -1.5 -1.1
    1903 -1.3 -1.2 -1.3
    1904 -1.5 -1.4 -0.9
    1905 -1.4 -1.3 -1.0
    1906 -1.3 -1.1 -0.7
    1907 -1.4 -1.2 -0.8
    1908 -1.5 -2.5 -0.6

    My guess is that computers at NOAA/Giss are overfeeding some
    of their databases with those magic TOBS-pills that are
    absolutely necessary to predict the recent past ( hard work that
    is,we all know) with any certainty.

  11. Every day I try to download data from GHCN.
    Here is some result for Teigarhorn:

    Data file from: January 1900
    2012-07-12 -0.16
    2012-08-13 -0.19
    2012-09-08 -1.36
    2012-09-24 -1.35
    2012-11-28 -0.90
    2012-12-09 -0.30
    2013-05-24 -0.33

    All I know data from Theigarhorn should show 0.70 degree C January 1900,
    but GHCN change the data every day. I have documented GISTEMP for a while.
    They use GHCN data set. I have never seen to same series from Nasa.
    Every Month they change global temperature.

    Have a look:
    Sorry Norwegian text but the graphs is in english.

    Knut Aasen

  12. In this case, since GISS treats Teigarhorn as a rural station and does not further adjust the GHCNM values, it is NOAA/NCDC rather than GISS which is responsible for these shifting values. GISS takes GHCNM data at face value, seemingly rarely questioning the validity of that data as input for further analysis. Also, although knowing that there are serious mistakes in the GHCNM station inventory file, they still use that file to find radiance values for stations, to determine whether the stations are urban or rural, pleading lack of manpower to check the entries in that file, and as a result misclassifying a substantial number of stations.

    Your values for November 2011 come from GHCNM v2, while those for March 2012 and August 2013 come from GHCNM v3. And even within GHCNM v3 there have been other substantial shifts in these values. If you had taken the October 23rd 2012 version of the GHCNM adjusted (qca) data file used by GISS you would have found:

    Year Jan Feb Mar
    1900 -1.36 -4.56 -2.26
    1901 -0.46 -2.06 -1.16

    and there may be other even greater shifts - I have only grepped the data from the data files currently decompressed on my hard disk, and not the many others still compressed.

    The corresponding unadjusted (qcu) GHCNM v3 data remains unchanged for all the corresponding decompressed files:

    Year Jan Feb Mar
    1900 0.70 -2.50 -0.20
    1901 1.60 0.00 0.90

    (which you may also recognise as the GHCNM v2 data)

  13. I took a peek at the Jan,Feb,Mar Giss-v2 and Giss-v3(agu. 2013 edition) monthly temperatures at Teigarhorn for the years 1899-1908 ( most of which are visible in the first and third screenshots in the article above).
    For fullness sake below is the table of all the 2011 giss-v2 data for this period.
    the v3-data is fully visible in the 3. screenshot up above.

    Year Jan Feb Mar
    1899 -1.7 -0.1 -2.7
    1900 0.7 -2.5 -0.2
    1901 1.6 0.0 0.9
    1902 -4.3 -2.3 -1.6
    1903 -1.1 -1.1 -0.7
    1904 0.1 -0.5 1.4
    1905 -1.5 -2.3 2.0
    1906 0.5 -1.9 -1.1
    1907 -0.1 -2.3 -0.5
    1908 -0.1 -1.3 1.5

    and I noticed that the the diffrence Giss-v2 - Giss-v3 is = -1.2°C for all those months exept the 1908 Mar entries where the v3 value is ¨only¨ -1.1°C down from the v2-value.
    Wonder what the explanation can be for this. Anybody know how the rationalization
    goes?. I am to lazy right now to try to go and chase it down myself.

  14. Who controls the present, controls the past, Who controls the past, controls the future...

  15. Raw data is here on a blog run by the iceland met office

    In fact it is worth reading the whole blog starting from the first entry in January 2012 , which was roughly when Paul Homewood started drawing attention to these erroneous adjustments.

    As oneillp says, it is not GISS doing the adjustment, it is GHCN/NOAA/NCDC.

    Paul M

  16. Same thing in Texas. The downward adjustments published in 2011 for years around 1900 were increased, indeed doubled, in the 2012 publication.

  17. Update July 2014. GHCN's latest efforts have dipped a further 0.1C. The Jan 1900 temperature at Teigarhorn is now -0.6C. Here's the link:

    My enquiries to the Icelandic Met Office yielded a good honest result. They have transcribed the original pen-and-paper record. It shows that the temperature recorded at the time was +1.0C.

    Thank you Onelip and Paul M for pointing out that the fraud is being conducted by GHCN and not GISS.

  18. Update Oct 2014. The Jan 1900 temperature is now claimed to be 0.2C. Try as one may to understand that some form of "homogenisation" may be needed to give a true picture or to apply a little common sense, for historical temps to yoyo around on a scale of a half-degree here or there is inadmissible. It's data tampering. The whole Global Warming debate hinges on a few lousy tenths and here they are mucking about with the historical record on a scale twice that. Villains!

  19. For the record, here is the Icelandic Met Office's admirable reply to my query in Aug 2013.

    It begins:

    Dear Brent.
    We have the original observation journals from Teigarhorn at
    hand. The calculation of the monthly means are not quite as straightforward as
    some people tend to think and is not done in the same manner all over the world.
    In 1901 the observations were made with instruments from and according to the
    guidelines from the Danish Meteorological Institute (DMI). There were three
    fixed-time observations every day at 8am (local time), 14pm and 21pm. The
    average of each was calculated and then the average of the average, adding an
    adjustment depending on the location and the season. In the case of February
    1901 the averages where: 8am: -0.9°C, 14pm: +0.9°C, 21pm: -0.6°C. The average of these is -0.2°C. As the observation near midday (14pm) is used this direct mean
    tends to be higher than the "true" average. This is especially evident during
    the summer. Therefore the seasonal adjustment is needed. In February it is very
    small, the DMI used -0.1°C for that month. The published value in the yearbook
    for 1901 is therefore -0.2-0.1=-0.3. The formula used by the DMI is usually
    called the Copenhagen-adjustment formula. In addition to the monthly average,
    the daily observations for Teigarhorn are listed in the yearbook, as well as the
    daily maximum and minimum temperature and the averages of these. In many
    countries the monthly average is calculated by averaging the daily maximum and
    minimum of the month. This is not done in Iceland and has never been done. An
    average of the daily max and min in February 1901 is -0.5°C, i.e. slightly lower
    than the published value.
    The Icelandic Meteorological Office took over the observations in Iceland in 1920. The Copenhagen-adjustment was in use (with
    slight changes at many stations) until 1956 when another average calculation
    variant took over (see: and Most of the
    earlier temperature means (prior to 1956) have now been recalculated so that the
    means for the whole period of observations at Teigarhorn in use now by the IMO
    are slightly different from the original published values. The value used for
    February 1901 is -0.2°C.
    In January 2009 the farm at Teigarhorn was abandoned - so there are no manual weather observations after that. An automatic
    station was put in operation in late 2001 so there are about 7 years of
    parallell observation at the two station types. The annual mean temperatures
    during the whole perod were exactly equal (the automatic station 0.02°C colder
    than the manual one), but there is a systematic seasonal bias cycle in the
    differences. This could be caused by the screen differences - or that the old
    seasonal diurnal adjustment constant for the manual station was not quite
    correct in all months.
    It is likely that the high resolution automatic observations will be used to calculate new seasonal adjustment factors back to the beginning of the observations, but when this will be done is not known at this time.
    Most of the early Icelandic temperature averages got into the present international market via the Smithsonian World Weather Records publications. The early Teigarhorn series found there is calculated by a direct average of the temperature of the three daily observations. These figures therfore differ from the original ones published by the DMI and have a daytime bias that must be corrected before a meaningful long-term series for Teigarhorn can be constructed. It is therefore entirely natural that there is a difference between the originals, the WWR tables and more recent adjustments.
    I have not taken a look at the GISS-adjustments.

  20. Second part of the IMO reply:

    I have not taken a look at the GISS-adjustments.

    It is very important to note that there will be no final version of any temperature series, but it is as important that new adjustments are always made by the use of the 8am, 14pm and 21pm means but not by an iterative adjustment of earlier adjustments..

    Last year I collected the original published values of monthly average temperatures in Iceland prior to 1961. They are to be found in an attachment (viðhengi) at this

    The values of course include the Copenhagen-adjustment during the relevant

    The original published daily records for Teigarhorn during 1873
    to 1919 is found in the DMI-yearbooks. The relevant links can be found

    At the blog

    more relevant information regarding the Icelandic temperature averages can be found. Please navigate to earlier pages by using the list of months in the leftmost column of the blog - or by using the paging at the bottom of the page (næsta síða = next page, fyrri síða = earlier page).

    Please note that in the DMI-yearbook Teigarhorn goes by the name Berufjord.

    Best wishes,
    Trausti J.


    How refreshing to receive such a clear and direct communication from this excellent institution; a rare ray of sunshine!