Skip to content

GISS Caught Tampering With Temperature Record In Minnesota

February 15, 2012

By Paul Homewood  (h/t Mostly Harmless)




Reader Mostly Harmless commented

Just a few minutes ago, I had displayed the GHCN temperature chart for Rochester, Minnesota, among others. With a dozen browser tabs open, I didn’t realise that the chart was still displayed, so I reselected it from the GISS station selector, and displayed it again. Unbelievably, the data had changed by a few tenths of a degree (both for early and late years) within ten minutes! Some kind of official investigation is needed of this unjustifiable shenanigans.


The above graph was the original one he was looking at. It then changed to this :-




When I checked the datasets here and here, it was clear that the the small increase for 2011 was the result of November and December data being added in, which had been provisional before. However, much more intriguing was the fact that 1948 and 1949 had been wiped from the record. As the graphs show, these two years were comparatively warm, certainly much warmer than the following years.

Maybe errors had crept into the readings for these years to make them unreliable or a station move was responsible for a step change? Well, not according to the NCDC. The Climatological Summary for Rochester in 1950 shows the 1948 mean temperature was 44.9F (7.17C) and 1949’s was 46.0F (7.78C), in line with GISS’s original figures. The NCDC summary also shows that 1950 was much colder, again as GISS showed.

The summary also confirms that there had been no station moves or any other changes which would have affected the reliability of the record. So once again we see GISS/GHCN rewriting history to create an artificial warming trend.

Whether this corruption of the data is intentional or not, it is apparent that the temperature records of both GISS and GHCN are rapidly running out of credibility. At the very least GHCN’s Version 3.1, the current one introduced last November following major system rewrites by a Google Summer Student, needs to be withdrawn pending a full and independent enquiry.



It is amazing what you find when you are not looking! NCDC’s Climatological Summary, mentioned above, conveniently lists all annual temperatures back to 1909 (except for a gap from 1921-28). This naturally raises the question why GISS don’t use them. Could they be a bit inconvenient?

The report confirms that the station has been at the same site since 1932, at Rochester Airport. Prior to that it was no more than two miles away.

So, splicing the NCDC temperatures up to 1949 onto the GISS ones, what do we find?




Surprise, surprise! The warming trend, that the cut off GISS data produces, suddenly disappears and temperatures in the last decade are pretty much what they were in the 1940’s and 1950’s. (And remember this is an airport site, so there ought to be UHI effect here).

Ah, but you say, after stirring and homogenising, the true picture will mysteriously appear! Well, the NCDC trends for Minnesota seem to show rising temperatures.




I wonder which one is more reliable?

  1. DirkH permalink
    February 15, 2012 8:48 pm

    “At the very least GHCN’s Version 3.1, the current one introduced last November following major system rewrites by a Google Summer Student”

    So you’re an agency full of government funding. You let a Google Summer Student (What is that even, I guess a pre-graduate bachelor student?) tamper with one of the products you deliver. YOU FAIL AT ANY ATTEMPT AT QUALITY CONTROL. Well, you probably don’t have quality control. That would be boring and tedious.

    Can someone shut them down and reclaim the money? They’re not even useless; they’re harmful. They took a fine recording from Iceland and turned it into something false. Greek accountants couldn’t have done worse.

    • February 15, 2012 10:19 pm

      Reto Ruedy of GISS has already effectively said it has nothing to do with them, as it’s GHCN’s data they use.

      If Ford sold me a car that crashed and blamed it on faulty brakes they had bought in, I don’t think I would be very impressed.

  2. February 16, 2012 3:47 am

    But note the temperatures since 1998 on that NCDC Minnesota plot.. Winter temperatures across the States (with very few exceptions) show statistically declining temps since 1990, and an obvious downward trend since 1997/8. I’ve posted NCDC winter plots for 1990 to present for all 48 contiguous states, with no comment. The charts speak for themselves.

    I’ve also noted that several datasets downloaded from GISS on 7th and 8th of this month has since changed. Not by much and only in parts, but it’s changed. I thought only totalitarian regimes re-wrote the past?

    I called out Minneapolis-based meteorologist Paul Douglas recently for claiming that “Winters in Minnesota have grown milder since 1998”. NCDC shows otherwise, as does GISS data (hence my re-checking Rochester, MN). I emailed him to ask where I could see data which contradicted NCDC. I got no reply (surprise!), but he’s just blogged on winter temps, so perhaps THAT’s my indirect reply?

  3. February 16, 2012 6:26 pm

    This latest data deletion was by GHCN, not by GISS, to be precise.
    The recent change is a result of the change to the new ‘improved’ version v3.1.1 that I mentioned on the most recent iceland thread.
    If you look at the graphs at

    you can see that the the GHCN ‘adjustment’ algorithm deletes the data for 1948 and 1949.

    I still have the files for the previous version from January and in that version both sets of data, adjusted and unadjusted, go back to 1948.

    (I guess this is the same data as in the files Paul H links to , but his links don’t work for me).

  4. February 19, 2012 8:13 pm

    Latest note From NASA webpage:

    February 12, 2012: The reported December 2011 data for the stations LIEPAJA, ALEKSANDROVSK, and ST.PETERSBURG were replaced by corrected reports and the strange Dec 1991 report from MALAKAL is no longer part of the adjusted GHCN v3. The corresponding entries in the GISS list of suspicious data were removed. The analysis was redone on Feb 17 after learning from NOAA/NCDC that the operational version of GHCN v3 was only made available that afternoon.

    • February 19, 2012 8:32 pm

      That’s remarkably quick, seeming as how I pointed out a month ago that the Iceland and Greenland adjustments looked wrong.
      GHCN are still “investigating”. (In other words, hoping the problem goes away). And Reto Ruedy of GISS says” We have to use GHCN data, as nothing else is available”

      • GreatHoax permalink
        February 20, 2012 4:44 pm

        Paul, Last week a colleague and I twittered the news of the Iceland GHCN temperature adjustments. We cc’d NASA Goddard Space. They sent us a reply. I have highlighted what I found to be interesting. Please, share your thoughts on this statement with me and the validity of my follow up questions. Those have not bee answered yet…I await with great anticipation.
        Here is their response:
        GISS’s Reto Ruedy says “First – the data originated from NOAA/NCDC, not NASA/GISS. The supposed manipulation is actually a side effect of NCDC’s automatic homogenization. That procedure has been developed and tested for several years and will undoubtedly go through further modifications; NCDC is investigating the effect it had on 3 stations in Iceland.

        But no matter how much this procedure will be improved, there will always be occasional exceptional situations that it cannot distinguish from artificial discontinuities. However, since these errors are random, they are liable to cancel out when computing global means. We started using GHCN v3, the version that includes that adjustment, only in December 2011 and we illustrated at that time the effect of that change on the web page.

        Differences are visible but insignificant; trends e.g. do increase by 0.03 degC/century, but the uncertainty of the result is at least twice as big (0.05-0.10 degC/century). Part of that increase is due to a documented artificial discontinuity for many Australian stations that affected GHCN v2 but was corrected in GHCN v3.

        Discontinuities can be as simple as moving a weather station from one level to a building roof, or to another place within the sampling area – much like snowfall measures can change depending on where you stick your yardstick.

        They could also have to do with changes in the way daily averages are computed, e.g. High-low v. hourly. Discrepencies in data from one station are practically negligible when averaged into the global picture.”

        Occasional exceptions…v3 throws out “suspicious data” and the warming in the 30s & 40s disappears???? Your thoughts, Paul????

      • February 20, 2012 6:41 pm

        Reto is saying pretty much what he said to me – effectively he knows the adjusted data is wrong, but hey don’t worry because it’s probably cancelling out other errors.

        I have never accepted the logic that if all adjustments balance down to zero, then everything is OK. There may be good reasons why adjustments should, for instance, reduce temps for UHI or other effects.

        The fact remains that Iceland/Greenland are symbolically important because of the Arctic connection, and of course have a big influence on projected Arctic temps.

        I would only be happy if GISS showed raw temperatures in their database alongside adjusted ones.

  5. March 11, 2012 11:55 am

    Reblogged this on Climate Ponderings and commented:
    GISS – Intentional corruption of data?


  1. More Magic Marker Mania from Hansen over at GISS - Page 3 - US Message Board - Political Discussion Forum

Comments are closed.

%d bloggers like this: