Prove Me Wrong

We have comments insinuating that my graphs are wrong. Prove it.

Data is here.

ftp://ftp.ncdc.noaa.gov/pub/data/ushcn/v2/monthly/9641C_201208_F52.avg.gz

ftp://ftp.ncdc.noaa.gov/pub/data/ushcn/v2/monthly/9641C_201208_raw.avg.gz

***"fully adjusted temperature series"***

  - "9641C_YYYYMM_F52.avg.gz" if you want the average of fully-adjusted monthly mean maximum and minimum temperatures (with estimates for missing values).

***"raw (unadjusted) temperature series"***

    - "9641C_YYYYMM_raw.avg.gz" if you want the average of monthly mean maximum and minimum temperatures that are not bias adjusted.

ftp://ftp.ncdc.noaa.gov/pub/data/ushcn/v2/monthly/readme.txt

The correction is much larger than they admit.

ts.ushcn_anom25_diffs_urb-raw_pg.gif (650×502)

It is interesting reading the comments below. Typical alarmist strategy. They come here and boldly lie about every aspect the discussion.

About stevengoddard

Just having fun
This entry was posted in Uncategorized. Bookmark the permalink.

87 Responses to Prove Me Wrong

  1. Dave N says:

    Who needs to look at your graphs? Even without your comparison, what is the justification for a continuing and increasing upward adjustment?

  2. Billy in NZ says:

    And lots of it.Only thing that pisses me off is,it’s OUR FKN MONEY.

    • Whatever says:

      Why would that produce money? An why would anyone pay for that?

      • miked1947 says:

        Whatever:
        Your questions are to easy. However if you can not see the answers, nothing anyone else can say will allow you to see them. See your proctologist for an optical exam

  3. Billy in NZ says:

    Don’t let up Steve,they are crumbling.Thanks for all your good work.

  4. Robin (in NZ too) says:

    Good comment Billy. Dont you just love our ets, not. Keep up the good work Steve.

  5. sod says:

    you are wrong, as always.

    for a start, the two plots show different time periods. Much of the “difference” will vanish, when you restrict your graph to 1900-1999.

    The NOAA graph gives a range between -0.1 and +0.6 over this period. Your graph in that period is between -1 and -0.1 (so there obviously is a different zero point)
    The small remaining difference is most likely based on some small data handling problems.
    but the big difference is by your choice of plotting a different time range!

    any other error i can help you with?

    • Me says:

      Maybe, I guess you didn’t see the links provided above.

    • It is the entire USHCN data set. Every station. Every month. Every value. You are lying.

    • Your description of the differences in the two graphs above defies the reality of what I see when I look at them (and you haven’t questioned Steven Goddard’s presentation of the second, NOAA graph, so I assume these are the two graphs you tried to compare, but so poorly). The NOAA plot shown above hugs 0.0 F (actually, about -0.02 F) between 1920 and 1940, then rises to plateau at 0.52 F after 1990, while Steven Goddard’s plot goes from -1.0 F, to a plateau of about -0.05 F, but his plateau is reached only after 2000. So your description is totally incompetent, the scales of the two plots are off by roughly a factor of 2, and stretching Goddard’s plot out to home in on 1900-1999 will not change either that or the 10-years’ misplaced plateau. The variations around 1950 and 1970 are clearly different in the two plots also. Anyone can look at the graphs and verify what I have just written. So I think it’s best to conclude you are a liar, and so deluded you think the scientific readers who pass this way will not notice. Incredible.

    • tckev says:

      “The small remaining difference is most likely based on some small data handling problems.”
      You mean you computer model is broken.
      Would that be as small as the percentage CO2 in the atmosphere and it’s nonexistent effect on climate?

  6. Ibbo says:

    Errm SOD.

    They are quite clearly two completely graphs illustrating two completely differing set of adjustments, and not meant for comparison with each other.

    Can you read ?

  7. Richard T. Fowler says:

    So, it does matter (to the scientific conclusions that are implied) that there are huge, spurious “adjustments” made to the raw data after 1999?

    • sod says:

      no.

      the huge jump in one year will turn out to be an artefact.

      this is, why real scientists adjust raw data. a lot of it has errors.

      and it is just dishonest, to post two graphs that look very different, when the real difference is mostly a different time period chosen!

      • Richard T. Fowler says:

        You’re still talking about 0.6F. How much of the warming up through 2012 disappears if you back out the “adjustments” that are not explained by the creators of those adjustments?

        RTF

      • Josualdo says:

        When an error always goes in the same direction, it’s no longer an error. Real scientists know how real error distributes.

      • tckev says:

        “why real …. …adjust raw data. a lot of it has errors. ”
        The model is broken.

    • Nick says:

      Neither you nor Steven have made the case that the adjustments are “spurious”. Or unexplained.

      • It is not up to anybody to prove they are spurious. It should be up to USHCN to justify them.

        The first stage must be for them to actually publish the full differences between RAW and FINAL, which they did in 2000, and if they are significantly different to Steve’s version, to explain why.

      • sod says:

        no.

        if you claim, there is an error, them you have to show the error first.

        there is tons of material about the adjustments freely available. why don t you simply read it?

        • Are you completely daft? The second graph is from their explanation and shows that the adjustments are much larger than they acknowledge.

          Can you make a feeble attempt to have an honest discussion?

      • Stephen Richards says:

        Adjusting raw data is scientifically unacceptable. You MUST publish the raw data and put your uncerntainties out there with them. NOT ADJUST THEM. That’s what real scieintists do. Only fraudsters adjust things.

      • Peter Ellis says:

        Stephen: They don’t adjust the raw data. They publish the raw data – Steve just linked to it! They also publish papers saying what errors and biases they can prove to be present, and finally they publish the best-available version with the known biases removed (and indeed they publish several of the intermediate steps along the way). So what precisely is your problem?

      • Sod

        if you claim, there is an error, them you have to show the error first.

        Don’t be ridiculous. If a govt agency makes changes to real data, they need to

        1) Be totally transparent and explain what effect the changes have made.
        2) Justify in detail everything they have done.

        Don’t you wonder why they have not done this?

      • Stephen Richards says:

        Peter Ellis

        Don’t be a clown. I am a qualified physicist (form ChPhys) and engineer. You do not adjust data. GET IT. Steven has explained in words of few syllables and English why they are wrong. Unless you have something intelligent to add go to RC. The data they make publically available is adjusted.

      • Peter Ellis says:

        Steven: You’ve claimed it, repeatedly, but you have not proven your claim, To prove your claim (that the adjustments are much larger than they acknowledge), you would have to replicate their published methods and show that the results are different from those shown in their graph. You have not done so.

        • USHCN1 shows no new adjustments after 1990. USHCN2 shows exponentially increasing adjustments after 1990. The stations aren’t moving. There is no legitimate explanation for this.

  8. sod says:

    Sorry Steven, but you are plain out wrong.

    NOAA has been plotting annual data from 1900 to 1999.

    you place it next to a graph that shows monthly data of a different time period.

    this does in no way support your claim, that “adjustments are much larger than they acknowledge”.

    Instead it is a misleading trick on your part.

    there are a couple of monthly values which are bigger/smaller than the average annual value. Bravo for showing that!

  9. OldOne says:

    sod appears to be the typical CAGW zealot who will absolute refuse to acknowledge anything that goes against their belief system. Facts and data don’t matter to them. They will look at black and call it white. It’s a common behavior of those brainwashed in a cultic belief system. Their ‘faith’ trumps facts.

  10. suyts says:

    Lol, Sod, I don’t know if you’re being pedantic or simply diverting. The fact is, none of what you’re saying changes the fact that nearly all U.S. warming seen in the temp record is seen only by adjustments.

    Now, the NOAA graph is fascinating. It’s wrong, but hey, I’ll go with it. As seen with the discussion around Watts et. al. much of the adjustments done is from TOBS adjustments. If that’s the case, then looking at NOAA’s graph, the thermometers TOBS, and siting were just about right until about 1960, then our thermometer reading skillz went to crap!

    There’s deception, and then there’s deception.

  11. PeterMG says:

    Just for a second image you are and engineer designing the wing of a 747, and the wind tunnel come up with some data that doesn’t fit with your theory??? Do you adjust the raw data or is it back to the drawing board? As an engineer I find this adjusting of data appalling and fraudulent. Sod you have no idea what so ever of the importance of being not only accurate, but also completely transparent when presenting data on long term climate trends. (well actually 100 years is but a blink in terms of climate)

    Data that is completely acceptable for day to day forecasting is often useless for climate science, especially as we are all aware of the changes to not only the number of stations now reporting compared to the past, but also the on going changes to the environment that many have experienced (urban sprawl)

  12. Eric Webb says:

    It is a new AGW troll everyday, lol, keep ’em coming, they’re just making themselves look like idiots.

  13. RobJ says:

    Any cretin can see from the graphs that the bogus adjustments are easily TWICE as large as the ones that NOAA admits for 1900-1999. Perhaps Sod needs to prove that he has taken a third grade level chart reading class before he is allowed to post here anymore.

    After looking at Steve’s graph I can see why NOAA cut off their chart at 1999. They will have a hard time justifying nearly 2 degrees of “adjustments” for the last 12 years. Unbelievable.

  14. pyromancer76 says:

    I have noticed, Steven, that trolls/attackers seem to be multiplying. When they don’t listen to replies to their inaccuracies, when they do not offer their proofs, their reasoning, when they keep repeating the same idiotic thing (government agencies can adjust anything, anytime, anyway they want and can destroy historical data) they should be banned. They are attempting to destroy the scientific method — probably their “assignment” from the beginning. Your blog is magnificent. Your evidence for history and science presented with great depth. Thanks.

  15. RobJ says:

    Steven, have you or anyone else out there actually confronted NOAA directly with the evidence that they are only admitting to less than half of their bogus “adjustments”? If so, what was their response?

  16. AMA says:

    Steven,

    I think sod’s main complaint is that your graphs are of slightly different time frames (even though it makes no difference).

    Please post a final-raw graph but only showing 1900-1999 like the range of the graph from the NOAA site. This should make it clear to him.

    • Peter Ellis says:

      Remember to grid the data before calculating the monthly/yearly means, and thus the effect of the adjustments on those means. I’ve only had a quick look, but it looks to me as though you didn’t grid the data. The USHCN graph you link to used a grid of 2.5 degrees X 3.5 degrees

      • It is a station to station apples to apples comparison. Griding has nothing to do with it.

      • Peter Ellis says:

        Gridding is important if you want to compare your graph to the USHCN graph (which was gridded).

        Your graph (without gridding) shows the impact of adjustments on a station-by-station basis. The USHCN graph shows the impact on the final measured temperature of contiguous US.

        To see why, consider a case where many nearby stations all require a particular type of correction to be applied (maybe they were upgraded from liquid-in-glass to MMTS as a batch, say). This has a large effect when you look at the station-by-station average, as you’re doing. However, the impact on US temperatures as a whole will be much smaller, because all the adjusted sites fall in the same grid cell.

        This seems to be a likely reason why your estimate of the impact of the adjustments differs from USHCN’s estimate (and in any case when you compare across the same time frame, there’s not as much difference between the two as you imply).

        • The comparison is the average of all adjusted temperatures at all USHCN stations minus the average of the raw temperatures of the same set of stations.

          It tells us the average adjustment, very precisely. Gridding is not appropriate.

      • Peter Ellis says:

        Gridding is not appropriate if you want to know what the average adjustment per station, I agree. Gridding is appropriate if you want to know what effect the adjustments are having on the overall calculated temperature.

        More to the point, you CANNOT compare the ungridded result to the gridded result and use that to claim that USHCN are lying about the magnitude of the adjustment – as you do repeatedly. Stop it.

    • If USHCN published a more recent graph, I would be happy to publish it. Can you guess why they haven’t?

      The main problem is the change since 1990.

    • The trouble is NOAA have not (to the best of my knowlege) issued a more upto date graph since the 1999 version. When we have that, we can make upto date comparisons.

      The fact they have not suggests they have problems with transparency.

  17. Bob Johnston says:

    My experience is that your typical CAGW believer thinks it sufficient just to call me a denier, as if that’s the end all, be all of argument enders. When I ask them if they can actually define the skeptical argument the result has always been silence.

    These people operate solely on emotion, there’s no sense in debating them, unless of course you simply want to sharpen your debating skills against a feeble opponent.

    • Well you can’t convince them because they don’t deal in logic. However, you can observe their rhetorical techniques. They do not have many so after a brief analysis you can grasp the overall pattern. Their most popular technique at the moment is misdirection. Pull some criticism out of thin air and post it here, even if it completely misunderstands the point of the argument. Make it sound a little technical so that fence sitters think there are legitimate points of debate. Unfortunately, Alarmists overall are not very bright and mathematically illiterate, so the success of their efforts tend to be mixed at best.

  18. sod says:

    sorry Steven, i can understand that you have “no patience for people arguing math who have no math skills.”.
    But as your comments seems to address me (at least indirectly), i can assure you that this is not the case. (university math major, i finished with an exam in numerics and group theory)

    I did a small example with woodfortrees data:

    http://www.woodfortrees.org/plot/hadcrut3vgl/from:1917/to:2000/plot/hadcrut3gl/from:1940/to:1990/mean:12/offset:-0.4

    the graphs look very different, but the reason is not that one of them is using unadjusted (with a -04 offset to get a better picture) and the other adjusted Hadcrut3 data. (you can switch the data source to see that the difference in this case is tiny)

    Instead the difference is mostly caused by the different time period and by the choice of annual vs monthly data.

    can anyone tell me by looking at those graphs, how much larger the correction is?

    • No the problem is that they are cheating.

      In USHCN1, the adjustments went flat after 1990, and in USHCN2 they went exponential in 1990.

      This should be obvious to a math major. What is the problem?

      • sod says:

        i didn t look at the data, but just from a visual inspection of your graph, there are too few “dots2 in the decades around 1900 and after 2005. (should be 12 per year. that is a lot of dots, if they are spread out a little and not cover nearly the exact same place)

        i guess that this is caused by a data problem.

        so if we don t have all the data from the last about 5 years, and some of the data we have are massive outliers (like the dot above 1.5), there might have been a problem with the last 5 years (which would be the last part of the 1990 to 2000 decade) when NOAA made their graph.

        i am still curious about how the graph would look restricted to those years up to 1999 and with annual data. you have already worked with the data, so it would be great if you could add the graph…

        • The USHCN V1 graph (bottom graph) went completely flat after 1990. Menne et al specifically said that the TOBS adjustment has not changed since V1, yet USHCN 2 shows huge changes after 1990 in both TOBS and their other adjustments.

          I have studied and remarked on this extensively, while you haven’t – and instead showed up spamming us with your half-assed non-analysis.

    • Josualdo says:

      ? They superimpose nicely — to me — when you remove the offset.
      http://www.woodfortrees.org/plot/hadcrut3vgl/from:1917/to:2000/plot/hadcrut3gl/from:1940/to:1990/mean:12
      But then, I was just a biochemist 🙂

      • Excellent point. Simply remove Sod’s idiotic -.04 “offset adjustment” and you will see the data is in agreement and therefore different start or end periods have nothing to do with it. Strangely Sod’s just refuted his own argument… this takes dumb to a whole new level.

      • sod says:

        Steven obviously has the skills to produce a graph. I wonder what we will see, if we superimpose the two, over the same time period and both with annual values.

        If Steven keeps using the big square symbols for the dots in his part of the graph again, i expect we will see pretty little of the NOAA graph….

      • You just had your own claim destroyed but right away you repeat the same nonsense again. Amazing.

      • Sod is either just deliberately wasting everybody’s time, or is a few pennies short of a shilling. Which is it?

      • There are dumb people who know they are dumb. There are dumb people who think they are clever.

      • Josualdo says:

        I’m led to believe that sod’s point (I’d rather say misinterpretation) is all about sliding (moving) averages and the such.

        If we were talking about a moving average with a wide (for the sample amplitude) averaging window, then the x-interval of the samples would count. The same sample, with different intervals, with very wide averaging windows, would produce different lines in the extremes of the x-interval.

        AFAIK it is not the case. The plots are just plots of data points, Therefore this argument about sliding averages doesn’t apply. But I’m not going to check if this is the case or not that the graphs above are simple “dot” graphs without any sliding transform.

        If it is the case, either sod mislead himself misinterpreting the graphs production method or willfully popped up with several red herrings.

        It is also possible that he didn’t understand the nature of the “adjustment” made to data. After all, adjusting data to fit our idea of how they should be is a scientific deadly sin — if you’re caught at it, you should be fired and charged with fraud. He made the graphs above from the same, already adjusted data, one unsmoothed, the other smoothed, hence they may only diverge very little. Again, it is not this what we’re talking about.

  19. CoRev says:

    SOD has been around for years, and also has been banned from many, many blogs. He will come back using sock puppetry. He is incorrigible.

  20. Jimmy Haigh says:

    sod says:
    August 20, 2012 at 9:57 pm

    “I didn t look at the data,’

    Well done Sod. You have just written every “climate scientist’s” epitaph.

  21. Jimmy Haigh says:

    Sod says: “(university math major, i finished with an exam in numerics and group theory)

    Being an anonymous troll, how do you expect us to believe that? Prove it by coming out of the closet so we can check your academic qualifications.

    Your written English skills are also appalling.

  22. sod says:

    It is very interesting to see, that commenters think that it is a good thing, that adjusted and unadjusted hadcrut3 data does “superimpose nicely” without the offset i put in.

    problem is, that this is completely contradicting the reasoning behind this topic. (increasing adjustments being an important source of the rising temperature)

    so why do multiple attempts of dataset analysis (BEST being the latest) not show a big difference between adjusted and unadjusted data, while Steven has “shown” above that the corrections are much bigger than NOAA admits?

    the simple answers have been given in these discussion: Steven is using a different time period, monthly instead of annual data and he seems to be comparing gridded to ungridded data.

    If we only fix some problems, as Steven did in the graph overlay, we already remove the 0.6F that Richard T. Fowler was talking about above to something closer to 0.1F.

    • Josualdo says:

      I see what you mean. I cannot answer, as I would have to compare with the raw data (and I have a better life than that 🙂 ). I have the feeling that “unadjusted”, for woodfortrees, is no longer raw, jus unsmoothed. It just means it’s not variance-adjusted, etc. We’re talking “homogenization” etc here.

      So your example is between two sets, with different domains and basically the same data, one smoothed, one unsmoothed and using different smoothing windows; which is why they agree nicely, being the same values. We’re talking about what happens before all that. Upward and downward ‘a sentimento’ adjustments of raw data to fit the theory.

      Raw data is never wrong. In case you can prove it wrong, you don’t adjust it according to your views of correctness of data, you stop using it (it’s wrong!) then correct the measuring instrument/method. You don’t transmogrify data sets.

    • Stop being daft HadCRUTv3 means “variance adjusted”

  23. Josualdo says:

    Yet, refocusing, you said “Instead the difference is mostly caused by the different time period and by the choice of annual vs monthly data.”

    Therefore I’m glad to see there is no noticeable difference in your demonstrative example. I’m not saying your argument is entirely wrong, I’m saying that it is so in this example (sliding adjustment window size vs. domain as well as domain limits). which is why the introduction of a useless offset is hmm perplexing. It looks like you expected a difference and got none, and so introduced an offset.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s