Skip to content

Temperatures Altered Yet Again In Alice Springs

June 24, 2012

By Paul Homewood

 

image

 

The poor folks of Alice Springs must be getting terribly confused. The clowns at GHCN have made wholesale changes to the town’s temperature record yet again.

Just to recap, the true temperature record below, as originally published by GISS and confirmed by the Australian Bureau of Meteorology, shows that temperatures have been pretty stable, with the last couple of years unusually cold.

 

image

Original

 

In March, we discovered that GHCN’s new software, (written by a Google Summer Student, Daniel Rothenberg, and introduced last November), had made large temperature adjustments to many stations worldwide, including Alice Springs, where a hockey stick miraculously appeared. (Full story here).

 

image

March

 

But then in May the new software appeared to go berserk, changing the picture of a warmer Alice Springs to a much colder one.

 

image

May

 

It will therefore come as no surprise to find out that a month later the May version has been changed back to the previous version.

 

image

June

It must be remembered that the global surface temperatures, as calculated by both GISS and NOAA, are derived from these adjusted temperature records. We have already seen that large and unexplained adjustments, such as these, have been made at a number of stations in many areas including the Arctic and South America, as well as Australia. Nearly all of these adjustments have made the past cooler, thereby creating a false warming trend. It is clear that the current GHCN Version 3.1 temperature record is unfit for purpose and should be replaced by the earlier 3.0 Version, which existed before Rothenberg rewrote the programs.

Perhaps, in future, when they decide to employ a boy to do a man’s work, they should get him to make the tea instead.

9 Comments
  1. June 25, 2012 3:29 am

    The clue is in the name – Alice SPRINGS.
    Temperatures spring up, then down, then up again.

    Perhaps the RBS group used a summer student to upgrade their banking software, which could explain the spectacular FAIL in one of their core systems, which left thousands without cash, and with direct-debit cards rejected at tills across the UK.

    It’s unclear to me (with past experience in large commercial systems) why they ended up in this mess. Shit happens, but thorough testing and checking of test output should have revealed such a fundamental defect (not crediting customer accounts with incoming payments).

    Thank gawd that GHCN isn’t in charge of sea level data, or they’d have London under 2 metres of water, then dry, then…..

  2. Kip Hansen permalink
    June 25, 2012 2:16 pm

    Mr. Homewood — the personal attack on the summer intern is unseemly, unprofessional and inappropriate. It was adequate for your purposes to note that the work was done by an intern.

    • June 25, 2012 4:22 pm

      I am sorry if it comes across as this. I have no criticism of Rothenberg himself. I am highly critical, though, of the decision to employ him in the first place.

  3. Kip Hansen permalink
    June 25, 2012 5:03 pm

    Paul,

    Somehow I doubt that you know enough, if anything, about Rothenberg to criticize. Had you said such a thing about me, I would have challenged you to a duel! (Well, symbolically speaking.) All you know is that the results of the code he wrote seem to be wonky.

    It is perfectly proper to criticize the decision to have such an important task performed by a summer student. National and International policies are being formed based on the results of his code, and apparently, no one bothered to run any kind of quality check or verification on it.

    If things are as you state, then you, or someone, needs to raise a bigger stink than this one blog entry on this issue.

    There is a major journal paper in this for you –> go for it.

    • June 25, 2012 6:11 pm

      Kip

      I have been in contact with GHCN since February, when these sort of discrepancies started to come to light (in Iceland). Despite promises from Jay Lawrimore that these issues are being reviewed, nothing more has been done.

      You’re right though. We do need to find a way to ramp up the pressure on them.

  4. Paul Matthews permalink
    June 25, 2012 5:18 pm

    The June version is in fact not quite ‘back to the previous one’.
    The step down in 2009 was largely eradicated in the March version but not the June one. So the March numbers are roughly a degree lower than the June ones before 2009.
    I have drawn a graph with the raw data and four different forms of the adjusted data.
    Notice that the historical temperatures vary by over three degrees between the different versions.

  5. June 27, 2012 2:07 am

    Here are the Alice Springs values for 1976 which I have on disk:

    qcu (unadjusted): stable

    pattern: [501943260001976]
    files: [.\ghcnm.tavg.v3.1.0.2012*.qcu.dat]
    ghcnm.tavg.v3.1.0.20120309.qcu.dat 501943260001976TAVG 2380 W 2510 W 2240 W 1790 W 1170 W 970 W 1030 W 1230 W 1730 W 2040 W 2370 W 2800 W .
    ghcnm.tavg.v3.1.0.20120317.qcu.dat 501943260001976TAVG 2380 W 2510 W 2240 W 1790 W 1170 W 970 W 1030 W 1230 W 1730 W 2040 W 2370 W 2800 W .
    ghcnm.tavg.v3.1.0.20120320.qcu.dat 501943260001976TAVG 2380 W 2510 W 2240 W 1790 W 1170 W 970 W 1030 W 1230 W 1730 W 2040 W 2370 W 2800 W .
    ghcnm.tavg.v3.1.0.20120321.qcu.dat 501943260001976TAVG 2380 W 2510 W 2240 W 1790 W 1170 W 970 W 1030 W 1230 W 1730 W 2040 W 2370 W 2800 W .
    ghcnm.tavg.v3.1.0.20120412.qcu.dat 501943260001976TAVG 2380 W 2510 W 2240 W 1790 W 1170 W 970 W 1030 W 1230 W 1730 W 2040 W 2370 W 2800 W .
    ghcnm.tavg.v3.1.0.20120515.qcu.dat 501943260001976TAVG 2380 W 2510 W 2240 W 1790 W 1170 W 970 W 1030 W 1230 W 1730 W 2040 W 2370 W 2800 W .
    ghcnm.tavg.v3.1.0.20120521.qcu.dat 501943260001976TAVG 2380 W 2510 W 2240 W 1790 W 1170 W 970 W 1030 W 1230 W 1730 W 2040 W 2370 W 2800 W .
    ghcnm.tavg.v3.1.0.20120528.qcu.dat 501943260001976TAVG 2380 W 2510 W 2240 W 1790 W 1170 W 970 W 1030 W 1230 W 1730 W 2040 W 2370 W 2800 W .
    ghcnm.tavg.v3.1.0.20120611.qcu.dat 501943260001976TAVG 2380 W 2510 W 2240 W 1790 W 1170 W 970 W 1030 W 1230 W 1730 W 2040 W 2370 W 2800 W .

    qca (adjusted): well, see for yourself

    pattern: [501943260001976]
    files: [.\ghcnm.tavg.v3.1.0.2012*.qca.dat]
    ghcnm.tavg.v3.1.0.20120217.qca.dat 501943260001976TAVG 2268 W 2398 W 2128 W 1678 W 1058 W 858 W 918 W 1118 W 1618 W 1928 W 2258 W 2688 W .
    ghcnm.tavg.v3.1.0.20120309.qca.dat 501943260001976TAVG 2267 W 2397 W 2127 W 1677 W 1057 W 857 W 917 W 1117 W 1617 W 1927 W 2257 W 2687 W .
    ghcnm.tavg.v3.1.0.20120317.qca.dat 501943260001976TAVG 2267 W 2397 W 2127 W 1677 W 1057 W 857 W 917 W 1117 W 1617 W 1927 W 2257 W 2687 W .
    ghcnm.tavg.v3.1.0.20120320.qca.dat 501943260001976TAVG 2267 W 2397 W 2127 W 1677 W 1057 W 857 W 917 W 1117 W 1617 W 1927 W 2257 W 2687 W .
    ghcnm.tavg.v3.1.0.20120409.qca.dat 501943260001976TAVG 2604 W 2734 W 2464 W 2014 W 1394 W 1194 W 1254 W 1454 W 1954 W 2264 W 2594 W 3024 W .
    ghcnm.tavg.v3.1.0.20120412.qca.dat 501943260001976TAVG 2604 W 2734 W 2464 W 2014 W 1394 W 1194 W 1254 W 1454 W 1954 W 2264 W 2594 W 3024 W .
    ghcnm.tavg.v3.1.0.20120510.qca.dat 501943260001976TAVG 2602 W 2732 W 2462 W 2012 W 1392 W 1192 W 1252 W 1452 W 1952 W 2262 W 2592 W 3022 W .
    ghcnm.tavg.v3.1.0.20120515.qca.dat 501943260001976TAVG 2602 W 2732 W 2462 W 2012 W 1392 W 1192 W 1252 W 1452 W 1952 W 2262 W 2592 W 3022 W .
    ghcnm.tavg.v3.1.0.20120521.qca.dat 501943260001976TAVG 2604 W 2734 W 2464 W 2014 W 1394 W 1194 W 1254 W 1454 W 1954 W 2264 W 2594 W 3024 W .
    ghcnm.tavg.v3.1.0.20120611.qca.dat 501943260001976TAVG 2380 W 2510 W 2240 W 1790 W 1170 W 970 W 1030 W 1230 W 1730 W 2040 W 2370 W 2800 W .

  6. July 4, 2012 8:01 am

    Mostly;
    “thorough” checking would hardly have been necessary. Even a single pass with each major type of transaction, electronic or other deposits surely being one of those, would have uncovered it. I think they had to have tried hard in order to achieve that level of negligence and carelessness.

Comments are closed.

%d bloggers like this: