Seems that there’s some excitement about a new paper A Statistical Analysis of Multiple Temperature Proxies: Are Reconstructions of Surface Temperatures Over the Last 1000 Years Reliable? to be published in Annals of Applied Statistics. Their reconstruction appears to be closest match to a hockey stick shape yet seen:



Using our model, we calculate that there is a 36% posterior probability that 1998 was the warmest year over the past thousand. If we consider rolling decades, 1997-2006 is the warmest on record; our model gives an 80% chance that it was the warmest in the past thousand years.


Update: Deep Climate on M&W is worth a read.

[Update 2: Eduardo Zorita.


  1. #1 Ian Forrester
    September 6, 2010

    Walsh once again shows that he does not have any understanding of the English language.

    Walsh said:

    to find that Ian, not surprisingly, failed to back his accusation that I made even one disparaging, much less libelous, remark about scientists. Wonder why?

    Ian Forrester said:

    I think the majority of posters on this blog know exactly what kind of a person he is.

    I’m finished with his nonsense.

  2. #2 Chris O'Neill
    September 6, 2010

    Bill Walsh:

    find the posts I made where I state either directly or implicitly that I know “more about some aspects of climate science than 97% of climate scientists.”

    I don’t have to look very far:

    I do find it more than a little interesting the amount of time and energy this paper seems to be taking from the AGW devotees

    97% of active climate scientists are “AGW devotees” and presumably you’re not. That means you think you know more about the AGW aspect of climate science than 97% of climate scientists.

  3. #3 Marion Delgado
    September 14, 2010

    Natural climate variability is not well understood and is probably quite large. It is not clear that the proxies currently used to predict temperature are even predictive of it at the scale of several decades let alone over many centuries

    umm. Proxies are used to infer temperature, not predict it. Models make the predictions. And it’s easier by far to predict long-term (centuries) trends than trends even over several decades.

    This is bats__t insane. I think The Annals of Applied Statistics does not do a good job editing or peer reviewing.

  4. #4 Wow
    September 14, 2010

    > I, on the other hand, am not wedded to any one idea and am able to see that there are merits to both sides of the discussion.

    And then ignore them.

    This is not seeing the merits.

    E.g. in your insistence that since people are complaining about this paper that there MUST be a pony in there somewhere.

    What you have demonstrated here, and in all your other posts is a holier-than-thou god complex.

  5. #5 Wow
    September 14, 2010

    You know, from the POV of a nutcase, all those sane people are a hell of a long way off.

    Aren’t they, Bill…

  6. #6 Bill Walsh
    September 24, 2010

    Just had to see if it was ever addressed. Of course, nothing was…

    Ian @496,

    You seem confused Ian, and ironically, you seem to be the one with language issues. Again, please back up your pathetic assertion that I in some way made libelous claims about scientists. Or even disparaging ones. You can’t, so you deflect. And you do a shitty job of it. But that is your internet MO so I am not surprised.

    Chris @497,

    Nothing about that comment directly or implicitly states that I know more than anyone. I simply found it interesting how much this subject got you folks so twisted up. Amusing. Nothing more. Care to try again Chris? If that is the best you can do, fail.

    Wow @ 499,

    I am the one with the “holier than thou god complex?” Now that is funny coming from you of all the people I see who post here. Given that I have yet to see you do anything more than spout your opinions as fact regardless of the topic, i would suggest a quick look in the mirror if you are looking for people with a “god” complex.

    You guys do make me laugh I will say that. And it isn’t because you are trying to be humorous. That would actually take a sense of humor.

    I’ll check back again later to see if Ian has scraped together a better answer to my request that he put up, or shut up. I suspect he will do neither.

  7. #7 jakerman
    September 24, 2010
  8. #8 Chris O'Neill
    September 24, 2010

    Bill Walsh:

    If that is the best you can do, fail.

    No Bill. You fail. Your pejorative comment:

    “AGW devotees”

    does imply you think you know more about the AGW aspect of climate science than “AGW devotees” and I pointed out the fact that “AGW devotees” includes 97% of active climate scientists.

  9. #9 Lotharsson
    September 24, 2010

    Meanwhile, as [others have pointed out]( there are two [new comments on M&W]( [PDF] accepted for publication.

    If the comments hold up, they suggest that the key claims made by M&W are totally indefensible. And some of the grounds for asserting indefensibility are precisely those pointed out much earlier here. Ignoring for a moment the trenchant criticism that these guys seem to be botching paleoclimatology and relying on novel but unvalidated methods, there’s this:

    > Assessing the skill of methods that do not work well (such as Lasso) and concluding that no method can therefore work well, is logically flawed.

    Well, duh! 😉 That’s what was being said here almost immediately, despite others arguing that it was not kosher to criticise before publication.

    If anything, the pre-publication critique may do some real good if M&W decide to withdraw their paper before it is published – one less badly flawed peer-reviewed paper for denialists and propagandists to exploit with their “peer-review means correct if it supports my position, but groupthink if it opposes it” double standard.

    Paging Judith Curry – does the paper still “look the real deal” – and if not what does it say about your ability to discriminate between “the real deal” and crap?

    Paging Anthony Watts – does the paper still “squarely hit the target”?


  10. #10 Wow
    September 25, 2010

    > I am the one with the “holier than thou god complex”

    Well I’m glad you now admit it. You can now either heal yourself or at least avoid subjects where your complex will become a problem.

    > Given that I have yet to see you do anything more than spout your opinions as fact

    Which opinions have I made as fact?

    And, given that this thread is about the opinions of people about this paper, you cannot merely point to something I’ve given as *opinion* without also showing that I’ve put it as FACT.

    But you can’t do that, can you bill.

    You know, with all your complaints and all this effort you’re making to belittle me or cast doubt on my veracity surely means there’s *something* to what I’m saying…

  11. #11 Anthony David
    December 13, 2010

    [Annals of Applied Statistics]( will be publishing the paper in the context of a discussion and commentary including a supportive discussion from M&M and an editorial outlining the process of the discussion. Statistics is on a list of “things I need to understand better”. That being said, Berliner points out that the authors assume a linear relationship between proxy observations and climate variables, suggesting the relationship is not tenable.

    For me, it is important that statisticians be involved in data analysis and model development discussions, but there is also a responsibility to understand the system being analysed.

  12. #12 Bernard J.
    January 31, 2011


    As Drongo has gone quiet on the Open Thread 58, I thought that I’d venture into the Bog to see if he’d retreated to beating his chest on the thread that Marohasy has erected in his honour. No sign of him, but it seems that Marohasy is back in business and giving Codling a run for her money.

    One thread that had me shaking my head in disbelief is [divorce lawyer cohenite and his mate David Stockwell claiming that it’s the sun, stupid](, and in doing so using liberal reference to McShane and Wyner:

    Even if we accept the NASA GISS temperature record as accurate the important issue is whether CO2 caused the trend. CO2 alone can’t have. CO2 has been increasing during the 20th Century at a constant rate but the temperature anomalies show many periods with cooling. Is it “cherry-picking” to focus on these cool periods?

    No. And for two reasons. Firstly, some of the cooler periods are longer than 30 years and so represent a climate period. It is legitimate to regard a cooler period as a contradiction to AGW and that some other factor is affecting the climate.

    Secondly, Lewandowsky has not considered that there may be a better statistical explanation for the temperature record; choosing an inferior explanation is hard to justify. In fact there is a better explanation than CO2.

    The better explanation is the Sun. One measure of the Sun’s influence on the Earth’s climate is called the Total Solar Irradiance [TSI]. TSI is a measure of the absolute intensity of solar radiation, integrated over the entire solar irradiance spectrum incident on the Earth’s atmosphere, that is, the sunlight reaching the atmosphere. Proxies of TSI such as sunspots go as far back as 1600. Figure 6.5 of the last IPCC report, AR4, shows the derived variations in TSI over this period according to a number of studies.

    How can a comparison of whether TSI is a better explanation than AGW for the temperature anomalies of NASA GISS be done?

    It can be done through statistics. Dr Jeffrey Glassman has compared the correlation between TSI and the temperature record over the 20th Century and found a 90% correlation. By comparison meteorologist Joseph D’Aleo calculates a statistical correlation between CO2 and temperature of only 42%.

    Both the science of the Sun and AGW are supposedly well established yet the amount of money being invested in AGW is vastly greater than that being spent researching the Sun with less than half the statistical justification. This may be good business but it is not good science.

    Lewandowsky says that for statistics to be effective one should use “All the data, for the entire globe, and for all available years.”

    But Lewandowsky has not met his own criteria of “All the data”. The NASA GISS record is NOT the complete temperature record of AGW. The complete record is much longer.

    Take Michael Mann’s hockey-stick. The hockey-stick is a 2000 year temperature record based on dendro-climatic or tree-ring data. The hockey-stick purports to show an even temperature until the 20th Century when temperature increases at the same time as CO2.

    We have already seen that TSI is a better statistical explanation for the 20th Century but what is the best explanation the rest of the 2000 years?

    The best explanation is not CO2. A new paper which was published with the honour of occupying the entire edition of a major research journal, The Annals of Applied Statistics, shows this. The paper is by two expert statisticians, Blakeley McShane and Abraham Wyner, who show the hockey-stick is based on flawed statistics and the Medieval Warm Period [MWP] was as warm and probably warmer than today.

    In the spirit of Lewandowsky’s study we showed his graphs to David Stockwell’s school-age daughters who also thought the trend would be up. However, when the girls were shown a 2000 year temperature record consistent with McShane and Wyner’s analysis, they extended the graph downwards. The conclusion? Perception of trend direction depends on the duration examined.

    More importantly, when asked if this was a good way to predict the future global temperature of the planet they said “Of course not!” Smart girls.

    McShane and Wyner’s study is a major contradiction to AGW. Real data, and a variety of statistical methods, including those relied on by AGW proponents, fail to show anything unusual about the present temperature; and that includes Lewandowsky’s record extremes which the IPCC has shown to have occurred nine times over the last 1000 years.

    Paul Nurse would weep.

New comments have been temporarily disabled. Please check back soon.