Cherry picking confirmed

Earlier I suggested that surfacestations.org was cherry picking by showing a station with warming as an example of a “bad” station, and a station with no warming as an example of a “good” station. Of course, it could have turned out that I was wrong, and those were the temperature trends of typical “bad” and “good” stations. But now they’ve classified one third of the stations and you can see that the cherry picking has been confirmed — the trends are the same for “good” stations (in red) and “bad” stations (in green).

i-ba2a3299ec8ddbbecc4eb5431667fa00-crn12_crn5_giss.png

BigCityLib has more.

Comments

  1. #1 sod
    September 16, 2007

    the real problem is this:

    for climate research, the only important information is a change in TREND. a “type 5″ station (error>=5°C!!!), if unchanged, will provide accurate information. in contrast a “type 1″ station could be situated downwind from some heavy growing UHI effect and give a completely wrong picture.

    the definition of the station types is taken from a french guy, who s work nobody actually read.

    it is still completely unclear, what sort (average? maximum?) of “error” he is talking about.
    wether the error of a type 5 station is positive or negative is unclear as well.

  2. #2 bigcitylib
    September 16, 2007

    What’s hilarious is the “bad stations” seem to mostly err slightly on the cool side.

  3. #3 Markus Massmuenster
    September 17, 2007

    Hopefully c-audit doesn’t find any more cherries, right? :-). Seriously. I’m partly colorblind, so please correct me if ever. The green stations (bad) show an additional increase of 0.2 C on a 120Y time scale as well as throughout the last 50 years. But I agree, during the last 30 years, when most of the warming occurred worldwide, there is no deferring trend. (C-audit argues that type 5 stations were not equally distributed, mostly in the West). Some of the “bad” stations are located 10 m above the ground. It remains pretty cool up there. The standard is 2 m above the ground. I think sod also has a point with the downwind argument. But that goes only for the winter when it’s cloudy. In the Swiss city where I grew up, the downwind side of the city usually has slightly less snow. I’ll look further into this because a new Swiss study was released on this subject. Then, there is the population growth (1.7 Billion in Y1900 and 6 Billion in Y2000). That’s new urban heat all over. On the other hand, the fingerprint of the GHG effect is mostly visible in dry cold air (Siberia, Canada, Alaska, Arctica in winter), right? Sigh – it’s complicated. Isn’t it that the consensus on AGW is defined as “human activities” and not “just” fossile fuels?

  4. #4 guthrie
    September 17, 2007

    Markus, if you read the IPPC reports (Yes it will take some time, I havn’t read them all myself) you will find that land use changes, methane emissions, and chlorofluorocarbons are also taken into account when estimating the warming.
    So yes, you could say that “Human activities” is the issue here, although fossil fuels are by far the biggest contributor.

  5. #5 Ian Hopkinson
    September 17, 2007

    …and don’t forget man-made black carbon, there was an interesting paper in Science about this a week or so ago. They had automated the analysis of Greenland ice-cores and proposed that the warming of Arctic regions in the early part of the last century was driven by this.

    (This is the link to the news item: http://www.sciencemag.org/cgi/content/full/317/5843/1333
    and this is the link to the paper itself:
    http://www.sciencemag.org/cgi/content/full/317/5843/1381
    )

  6. #6 Eli Rabett
    September 17, 2007

    The neat thing is John V did exactly what everyone has been telling the CA crowd should be done and is done for claims in the scientific literature. This is a major refutation of the CA approach. More credit to John V.

  7. #7 Markus Massmuenster
    September 17, 2007

    Yes, let’s move to Antarctica. No soot there. Black carbon sure has accelerated glacier meltdown in Switzerland, not well-filtered unleaded. What about x-rays and electromagnetic frequencies? I believe CO2 from fossile fuels is the star culprit because it is THE new currency. Yes, and with this new hot potato traded, people like John V do a great scientific job.

  8. #8 Vernon
    September 17, 2007

    Your so funny here. JohnV gets credit for doing some initial work and you wankers just have lie about it.

  9. #9 Hans Erren
    September 17, 2007

    What do the graphs look like if you take 1901-1930 as the reference period?

  10. #10 Eli Rabett
    September 17, 2007

    Who lied about what Vernon? We’ve given John V full credit. True your whining is ingrained, but please leave some at the door.

  11. #11 Vernon
    September 17, 2007

    Eli, there has been no cherry picking. First it was a picture cannot show anything. But then we find that it can. Then it was where is there proof of the quality of the sites. But then it becomes apparent that showing which NOAA class a site is can be shown. Then it was, there is no bias, the errors or random, but now we see that most the bias is a warming bias. Now it is the raw data matches GISSTEMP.

    Eli, your the biggest fraud here. You pretend to be a scientist but you flee any discussion that has to do with facts and logic.

  12. #12 Dano
    September 17, 2007

    Shorter Vernon:

    Waaaaaahhhhhh! Where’s my binky?!?!?!?!?!?

    Best,

    D

  13. #13 Lee
    September 17, 2007

    re 9, Hans,

    They would look like you had cherry-picked the reference period to maximize the apparent difference.

  14. #14 Lee
    September 17, 2007

    vernon, are you back again? Do you LIKE getting spanked?

  15. #15 Eli Rabett
    September 18, 2007

    Tim, what is it about Deltoid that attracts the cream of the
    idiocracy?

  16. #16 guthrie
    September 18, 2007

    Obviously Eli, you have never met AFDave, or Dave Springer, or indeed William Dembski. They are so far above the idiocracy as to render it 2nd class.

  17. #17 Vernon
    September 18, 2007

    Lee, I like dropping in to see how the true believers like to ignore facts in support of their dogma.

  18. #18 guthrie
    September 18, 2007

    Make that Dave Scot, not Dave Springer.

    Vernon, what facts have been ignored in this case here?

  19. #19 Dano
    September 18, 2007

    There’s interesting little bits in this “phrase” “uttered” by “Vernon”:

    true believers like to ignore facts in support of their dogma.

    1. Dano finds a clue in the use of ‘true believers’ by commenters wishing to deny the fact that scientists may have a fact or two.

    2. Dano finds projection in the use of ‘true believers’ and ‘ignore facts’ and ‘support of their dogma by commenters wishing to deny the fact that scientists may have a testable hypothesis or two that turned into theory.

    3. Dano finds confirmation bias in the inability of denialist commenters to support their ideology with the use of one or two septic testable hypotheses, validated models, good ideas, robust data, clever arguments, journal articles, cute fossil records, scribbles on a napkin, anything.

    Best,

    D

  20. #20 Hans Erren
    September 18, 2007

    re 9 Lee

    It would show the difference in the warming trends since 1901 more prominently wouldn’t it?

  21. #21 J
    September 18, 2007

    Hans wrote:

    “What do the graphs look like if you take 1901-1930 as the reference period?”

    For most of the 1900s and 1910s, the CRN12 line (the “best” stations, right?) falls almost exactly on or slightly below the GISS line. For the 1920s, it appears to average slightly above the line. Net result of using 1901-1930 as a reference period – probably zilch.

    You can pick better cherries than that. Why not suggest normalizing to the year 1898?

    In any case, during what Tamino refers to as the modern global warming period (1975-present), the GISS trend differs from the CRN12 trend only microscopically.

  22. #22 Hank Roberts
    September 18, 2007

    When I took statistics, the prof. told us a fundamental rule was that we were not permitted to go back and try to rearrange our data then do the numbers again to get a better answer.

    If the denial guys don’t have to obey that rule, then the next move for them would be to look for any way at all of grouping the stations so, when groups are compared, they get the difference in results they want.

    Maybe “Hibachi or AC, months beginning with J, elevations below 300′ and installed by Hibernians” gives a cluster of warm sites, for example. So we throw out all stations installed by Hibernians.

    How’re they doing so far?

  23. #23 Sean Egan
    September 19, 2007

    I agree with 22 that “corrects” to readings decades old does not create confidence . However all the major time series are repeatedly “corrected”. Just look at the fall and rise of 1998 compared to 1934 in the NASA series for the 48 states.

  24. #24 Markus Massmünster
    September 19, 2007

    Bingo, Sean, regarding revised data. I just had to revise time series myself and it DID create confidence to explain that the accounting standards changed. Audit is normal in the business world, why shouldn’t it become a standard in the measurement of temperatures, especially with this new currency CO2 traded?

  25. #25 bigcitylib
    September 19, 2007

    #24 “Audit is normal in the business world, why shouldn’t it become a standard in the measurement of temperatures…”

    Because Science is not Accounting, and we wouldn’t want to turn Scientists into Accountants?

  26. #26 Markus Massmünster
    September 19, 2007

    To bigcitylib: I am referring to specialists who record, analyse and interpret temperature data. Well yeah, isn’t it a sort of accounting, to record temperature data? Somebody else tests them (audit), another specialist interpets them, then it takes many scientists to even attempt a projection into the future. Is there any meteorogist around us who could tell us if audit in temperature recording is being done on a – say – 100 year time scale? Is Steve the prototype or just a “cherry picker”?

  27. #27 richard
    September 19, 2007

    “Is Steve the prototype or just a “cherry picker”?”

    Not much of a prototype, I’d say. Auditing would have to be done using a scientific methodology with a certification process. In this case, the assessment of stations has been done in what appears to be an arbitrary and amatuerish way that won’t add much value. Was the objective to certify the results or simply to cause confusion? Based on the approach used, I’d say the latter.

    In any event, has this ‘review’ improved the datasets or altered the observed trends. Not significantly it appears.

  28. #28 bigcitylib
    September 19, 2007

    #27 Actually, I think the purpose was to demonstrate that, once data from the “bad” stations had been dropped, there would be a divergence from the official record as compiled by Gisstemp. Once this was demonstrated NOT to be the case, Steve and Anthony immediately backed away from their analysis and started raising this nonsense about applying accounting principles to the gathering of temperature data.

    They’ve tried to change the subject, in other words.

  29. #29 jb
    September 19, 2007

    If Watts et al have applied “accounting/auditing methods” to their photo-shoot, I’d like to know what those are.

    Some of the photos include measurements to things like AC’s and parking lots and some of them don’t. I’m not even sure how watts is “classifying” sites without such measurements.

    To take Anthony Watts’ classification of these stations at face value is just ridiculous. He may claim to be following a classification system established by climate scientists, but that does not mean he has correctly applied the scheme and that his classifications are correct.

    I think its premature to be re-calculating results based on Watts’ classification of sites before that has been validated by climate scientists.

    The real irony here is that the “Climate Auditors” have been as loose with their standards as they are claiming NOAA and others have been.

  30. #30 Eli Rabett
    September 19, 2007

    Hell, Watt, Pielke Sr. and McIntyre KNEW that whichever way you sliced the stations, the trend was the same. There are a zillion papers out there showing it to be the case.

  31. #31 sod
    September 19, 2007

    i kept telling them, that the station categories are not a good way to judge those stations, as the only important factor is a CHANGE in station environment.

    i kept asking 3 simple questions on the stations types:

    1. what is the meaning of error >= 5°C

    2. how much reading of Leroy did you guys do?

    3. do you agree, that a type 5 station can produce better CLIMATE TREND results, than a type 1 station does?

    then a nice uy from meteo france showed up and provided the answers:
    http://www.climateaudit.org/?p=2069#comment-138764

  32. #32 dhogaza
    September 19, 2007

    To take Anthony Watts’ classification of these stations at face value is just ridiculous.

    Clearly what’s needed is a Climate Audit Audit project :)

  33. #33 Hank Roberts
    September 19, 2007

    Can’t fault their enthusiasm.

    The thing about science is, it’s apt to draw in even those who have strong beliefs about what’s real, because the real world does keep intruding and asking to be counted and measured — twice, or thrice, and again.

    Far better than if they were reading theology books for answers, eh?

  34. #34 Markus Massmünster
    September 20, 2007

    Yes, Roberts, religion without science is blind, but science without religion is lame (Albert Einstein). This science is exciting and i just LOVE asking questions. I am very happy to learn a lot in this blog and comments thread. Thanks, Hans, j, Sean, Richard, bigcitylib, jb, Eli, sod, and even dhogaza and Roberts for your valuable imputs.

  35. #35 Hans Erren
    September 23, 2007

    Hans wrote:

    “What do the graphs look like if you take 1901-1930 as the reference period?”

    something like this…
    http://home.casema.nl/errenwijlens/co2/crn5-crn12rev_giss_0130.gif

    Kudos to JohnV for the data.

  36. #37 Dave Dardinger
    September 24, 2007

    And in case anyone here doesn’t get Han’s point, we’re looking at trends in the graph at top, not absolute temperatures (just as they are in the info Jones or Hanson put out). As Han’s graph shows quite clearly there’s a .4 deg C difference between good and bad sites. Now admittedly it’s the #4 sites which are the most common and therefore they should be compared with the #1&2 sites to get an idea what the true bias in the surface temps are, but it’s very likely not negligable.

  37. #38 John Cross
    September 24, 2007

    Dave: I am not so sure I agree with the starting point that Hans picked, however what I am struck by is that after reading about how bad the class 5 stations are we see that the bias is actually pretty small.

    As some on here have been very eager to point out in the past, the class 5 stations are those supposed to have a minimum error of 5 degrees C. The fact that the error is not just smaller but so much smaller shoots holes in just about every argument I have read over the past month.

The site is undergoing maintenance presently. Commenting has been disabled. Please check back later!