Matt Ridley’s Third Test

Where’s Bart Verheggen when I need him?

In Matt Ridley’s recent essay published by the GWPF, he says he needs to have 10 questions answered before he will think that policies to address global warming make any sense. I have tried to address his first two questions here and here. Matt was kind enough to leave some comments on my second response.

The third issue he needs answered is, “Nor am I convinced that sulphate aerosols and ocean heat uptake can explain the gap between model predictions and actual observations over the last 34 years. Both are now well understood and provide insufficient excuse for such an underperformance. Negative cloud feedback, leading to total feedbacks being modest, is the more plausible explanation.”


The role of aerosols is a matter of debate in climate science. They block the sun. They can contribute to cloud formation . Many scientists are convinced that they mask global warming and that if we actually got rid of them, warming would increase.

Skeptics have often expressed the opinion that some climate modelers use aerosols as a ‘plug’ in their models to help them hindcast past temperatures. Modelers vigorously dispute this characterization as a slur on their integrity.

And we are in way over my head. I don’t know how to answer Mr. Ridley’s point. I haven’t studied this in any depth, nor have I seen plain English explanations that make me nod my head in vigorous agreement or shake it in equally vigorous opposition.

So I hereby ask for your help and am making this a dedicated thread to discussion of aerosols and whether or not there is a sufficient body of evidence to present to Mr. Ridley as an answer to his question.

(As for global ocean heat uptake, I think we’re all just hanging around and waiting for enough Argos observations to make some definitive points either way. There’s sufficient blustering on both sides of the issue for anyone to pick an argument that suits–but we just need those buoys to keep rising, falling and measuring a little longer. Feel free to pitch in on this one, too.)

‘Bout time I made you all do the heavy lifting, anyhoo. I got a Super Bowl to get ready for.

Update: As the conversation is getting a bit heated, let’s reset the terms. I really would like this to be a place where those on either side of the fence can have a civil–almost friendly–discussion. That requires effort on both sides of the ball (why am I using a sports metaphor? …oh, yeah, that’s why…)

Please provide answers to the following questions and discuss in a civil manner:

1. Is Matt Ridley correct to be concerned about this issue? Is this a ‘deal-breaker’ for climate science and its former conclusions of .2C per decade for the early part of this century?

2. Is the volume of aerosols adequate to produce the effects claimed for it?

3. Are some of the properties of aerosols producing more than one climatic effect (e.g., blocking sun and lowering heating, nucleating clouds with unknown effects, etc.)?

4. Is the treatment of aerosols in the literature proportionate to their real or potential effect(s)? Are they, for example, handled appropriately in AR4 (or AR5 if anyone has read the relevant sections in the leaked draft)?

5. Are estimates of aerosols treated appropriately in the employment of GCMs?

I really want your opinion on these issues. I am really not interested in your negative opinions of others on this thread–especially as some of you who are fussin’ and feudin’ are really good people who just are both commenting and reading too quickly.

I speak as someone who often commented in such a fashion in the past. It got me nowhere in the conversation–I know what I’m talking about regarding commenting.




112 responses to “Matt Ridley’s Third Test

  1. Hi Tom,
    This is an area that I don’t follow as closely as I might. Aerosols have a shorter lifetime than greenhouse gases in the lowest part of the atmosphere and are generally washed out by rainfall. The main sources of Aerosols are highest in highly populated regions of the world, usually where coal is burnt . Aerosols are usually associated with poor air quality and we know dirty emissions have been reduced in Europe and the US in recent decades. Since aerosols cool climate, a reduction in aerosols in order to clean up air quality, could lead to warming of climate. Our efforts to clean air in the latter part of the 20th century seem temporally associated with the small increase in global temperature during that time period. Any good studies on this?

  2. Well, I guess that the good news today is that Pennsylvania’s most trusted climatologist predicted an early spring.

  3. Hi Tom,
    ” Where’s Bart Verheggen when I need him?” Warning- snark. He may not have recovered from the total smack-down given to him by the mysterious Visiting Statistician (VS).

    • Bob, he was just hosting that discussion. After the first ten comments I think he left the field to the stats guys.

      • Hi Tom,
        I with you on the 49ers. That smack-down was a total evisceration of all, including Bart. At the end of the day, the data generated by climate scientist are no different than any other field. Numbers are numbers regardless of discipline and how you treat them statistically is what separates good scientists from mediocre ones. If you read that session closely you will see that Bart retreated because he could not follow the discussion. Most of the great controversies in climate studies, e.g. Mann, Steig, etc. arise when climate scientists refuse to collaborate with statisticians, and as a result produce shoddy work. Sorry Tom, but Bart received the equivalent of a scientific concussion.

      • Not to belittle my own powers of persuasion, but I’m just doing what Bart did–but I’m doing it before I embarrass myself…

      • Tom, for a good discussion on the statistical handling of sensitivity studies you should read this BH post ( Apologies if you have already seen it.

    • Please don’t say something accusatory like that when Bart’s not here to defend himself, and then not even give a reference so others can check your interpretation. I just wasted 15 minutes trying to track down whatever it is you’re referring to, and failed.

      And by the way, statisticians are not Gods and science isn’t purely a matter of statistics etither. You have to actually know something about the physical system you’re studying before you start quibbling over which statistical method is better or worse than another. This fact is lost on more than one statistician.

      • Actually I found it on my own and I think you meant to refer instead to this post:

        I remember that thread and I get the time (Ha!) I will try to read through it again. However, for those of us time limited folks who have to pick and choose carefully what’s worth reading, “VS” puts a pretty big damper on reading what he has to say when he starts off with:

        “Actually, statistically speaking, there is no clear ‘trend’ here, and the Ordinary Least Squares (OLS) trend you estimated up there is simply non-sensical, and has nothing to do with statistics.”

        which has to rank as one of the stupidest comments in this debate that I’ve ever read. OLS trend lines have nothing to do with statistics eh? Okey Dokey brother. He then follows that gem immediately by referring to first ADF tests on data from either 1850, or 1881, to 2009, whereas Bart’s graphs and discussion were focused primarily from 1975 to 2009, the period of greatest GHG increases. Complete and utter shift of the goalposts by “VS”.

        So why am I not too interested in reading any further what “VS” has to say on the topic?

      • Hi Jim,
        As Eli would say, some, but not me of course, would say you would you might have cherry-picked one of VS’s comments and referred to it out of context. It is probably true that between VS, Steve Jewson, and Nic Lewis, many real questions have been raised about the validity of all the sensitivity discussions in the IPCC. Jewson admonished them and we will witness their behavior in the not to distant future.

      • I have read the threads on Bart’s blog repeatedly. Let me close the conversation on that thread with this coda: Those thousands of comments were mostly people on the sidelines with pre-formed opinions watching someone who clearly understood statistical theory but was new to climate science (and did not know the extent of the literature and discussion of the issues) debate several who were far more aware of the literature but might have been surprised to discover that statistics had explored some of the consequences of decisions made by scientists in treating the data they had gathered.

        The discussion got so vehement so quickly that many teachable moments were wasted. I don’t want that to happen here.

      • So you’re accusing me of cherry picking then, do I have that right? Even though I clearly stated that, oddly enough, I started at the top of the comments and read the first one by “VS”. This is exactly why the various self appointed critics, such as yourself, make enemies as fast as they do–you accuse people of things you don’t have any basis for.

        And we weren’t discussing Nic Lewis or Steve Jewson, who pretty clearly are several steps ahead of “VS”. We’re discussing your accusations against Bart Verheggen w.r.t. “VS”. Then when I do a cursory check on who this “VS” is to validate your claims, I find in his first comment that the guy has no idea what he’s talking about. Now maybe you want to read through the rest of what he says, but I think I have better things to spend my time on brother.

      • Hi Jim,
        Apologize for getting your blood pressure up. I stand by what I said. You put Nic Lewis and Steve Jewson several steps ahead of VS – I ask you Jim, how would you know? You might of course be right but I maintain you are not the person to make that judgement. Praising Nic lewis and Steve Jewson is the right thing to do because they are first rate mathematicians and statisticians, but it certainly doesn’t help your cause. I understand that you choose not to bother with further reading. In your case I think that is wise.

      • Thanks pal, but I don’t really need your opinion on what does and does not “help my cause” OK?

        And save whatever apologies you might have in your tank for Bart Verheggen, he’s the one you defamed. I’m just the one who called you on it.

      • Hi Jim,
        A couple of loose ends. You say, ” Please don’t say something accusatory like that when Bart’s not here to defend himself”. Since when is it my job to invite or compel Bart to show up. And more thing, when you say about VS, “I find in his first comment that the guy has no idea what he’s talking about”. Astonishing – have you had discussions with Professors of Stats about that comment trail. I have. For you to make a comment as such says a lot about you Jim. I know it is uncomfortable when you meme and mantra are unraveling.

      • Hi Jim,
        Do you know what defamed means when you say, “And save whatever apologies you might have in your tank for Bart Verheggen, he’s the one you defamed”. I pointed to discussion and simply gave my opinion. You need to get a grip.

      • Bob, Jim’s a really good guy. Back off–I think you are, too. Let’s get back to the real topic.

      • Whatever you say brother.

      • Jim, VS jumps back to that comment later. He basically says his statement was wrong, but that he was just highlighting his point that there is no stochastic trend, to speak of a trend in the data is wrong, because of the existence of a unit root.

      • As for a smackdown, Bart never engaged VS too directly. Indeed, VS left because he felt Bart wasn’t following what he was saying.

    • Bob,
      For most of us numbers are numbers and amenable to the same rules and procedures.
      But clearly not to climate science. Accuracy, proper stats, complete data, open review, consistency of data are all up to the lords of climate and are heavily defended by their faithful many.

  4. Tom, have you followed the recent discussion at RC, where statistician Steve Jewson suggest that numerous climate papers should essentially be invalidated.

  5. The question – like everything Ridley writes on climate – is stupid. In this case, because there is no gap. For example

    • Great way to open a conversation–call the opponent’s questions stupid. How… very like you.

      • Wm missed the Memo: The Empire is Dead and England has not produced an original thinker since Henry VIII got a boner for Anne B and eschewed Catholicism. After that, the Scots took over, then moved to the States.

        et vous pret partez

  6. Speaking of Super Bowl, everyone has written off the SF already, it seems.
    I think SF may have something to say about that.
    But I have to be discreet, since our Superbowl party hostess is from Baltimore, and we are required to wear purple, lol.
    As to aerosols / sulphates / carbon black – It is clear these are very important in climate dynamics. It is also apparent our alarmist friends have not accounted for these factors. It is one of the multiple damning bits of evidence as to why alarmists are not basing their extreme claims on evidence that actually exists.

    • What is “apparent” is that you are wrong, which even a cursory examination of the literature and the syntheses thereof would show you.

      • Hi Jim,
        Perhaps you can tell us then – How much of the warming that occurred from 1978-1998 is a result of reducing regulatory anthropogenic aerosols?

      • Ain’t what I was objecting to Bob, which was this:
        “It is also apparent our alarmist friends have not accounted for these factors”.

        However I’m sure Bart can discuss the question you’re interested in, since he actually studies that topic. I’m sure he can also address your subsequent question, which would certainly be “and how much lower would temperatures have been if we also had not increased CO2 concentrations by 40%”

        After that we can discuss why VS has no idea what he’s talking about when he says that there’s no trend in the global mean temperature.

      • Jim, unlike you I am taking Tom Fuller’s advice. You get to heated to have a civil discussion with.

    • Hiya hunter–the KaeperKid will make purple once again the color of royalty–deposed royalty…

      Hunter, just from reading what Bart has written on his blog I think you are exaggerating (for effect?). There is extensive literature treating aerosols and sulphates–I see no signs that it has been ignored. I’m not sure the answers they have arrived at are adequate.

      Let me ask you this–if we find a suitably accurate ‘story’ to present to Matt Ridley, would it be adequate to address some of your concerns?

  7. wmconnolley,
    IPCC credibility is reasonably questioned by people who look at evidence and accuracy and ethical behavior as a guide to making informed decisions. I wonder why you would dismissively reject even asking the questions Ridley asks, and then look even more reactionary and uninformed by using the IPCC as your source of authority?

    • Unfortunately, extremely few are able to rationally question the IPCC because doing so usually requires a fair bit of expertise. You are clearly not in a position to make any claims like yours about the IPCC since you are not knowledgeable enough.

      Also, to defend the blogger for once, he didn’t dismiss anything. He actually asked for help on this one.

      But I expect that all he will get is a bunch of science-opponents with strong opinions and very little to back them up.

      • ss,
        >sigh< your version of the true believer dodge is not nearly as good as others I have seen. Whydoncha get back when you can actually address the issues and not just rely on infantile circular reasoning?
        One sign of faith-based irrationality is that the believer is comfortable repeating their chosen dogma, even though they have no actual training or education, but demand critics of the dogma have "expertise". And then of course the true beleiver's next defense is to define "expertise" as "agrees with me".
        Good luck ss. The good news is you have room for improvement. Lots of room.

  8. Ridley gets it all wrong. His claimed “gap between model predictions and actual observations over the last 34 years” is a fantasy. This is like a creationist asking why there’s no crocoduck.

  9. Hi snerkersnerk,
    I see you are up from the basement.

    • Now there’s a helpful comment

    • For my money, ignoring snerkersnerk is the best strategy.

    • Hi Bob,
      I see you are being a hypocrite again. Typical ****.

      Snerkersnerk, we are rapidly coming to a parting of the ways. I am nearing the point where I will ban you. Not for the peccadillo of derogatory language–we are all grown-ups here and it’s easy to delete your folly. But you are becoming boring and interfering in the conversation of sane people.

      There will be no more warnings. If I judge your behavior to be out of bounds I will mark all your comments as spam.

  10. Tom, I think you’re going to have to make some decisions on whether you’re going to allow people like Bob and hunter to derail the discussions and turn this site into the typical science- and scientist-bashing that is so common elsewhere.

    • I’ve just updated the post above with what I hope are better terms of reference for this discussion and a plea for civility. I’ll bring it down here as a separate comment.

    • Tom, I think Jim is right about the quality of comments. I am sorry but I thought you wanted a little give and take. I make a comment about one of the best climate statistical blog discussions of all time and Jim doesn’t like the tone, especially when he makes inane comments. He is essentially asking you to calm “people like Bob and hunter” down so that we don’t ask him questions he is ill-equipped to answer. In other words, Tom, tell them to shut up or I will take my ball and go home. Gosh, I hope all eco-minded plant biologists are not that sensitive. Tom, from here on I will honor your request.

      • “from here on I will honor your request”
        read: “but first let me get a couple of digs and distortions in, then I’ll honor your request”.

        Y’aint bluffin’ anybody but yourself pal. Here’s what this is about: You thought you could come here and bash Bart Verheggen, **without even giving a reference** for your outrageous claims for others to check on. Reason: because your kind is so used to getting away with scientist-bashing on other blogs, that’s why. You think it’s your birthright. It’s ingrained in some of you all and you think you can just come here and turn this blog into a clone of Watts’s blog or Curry’s or whomever, where that kind of thing is not only tolerated, it’s expected. Well I’m here to tell you that I’m going to get right in your face about it pal. Don’t like it? Then don’t spout things you can’t defend. Got that?

        You don’t have any idea why VS is completely off base in his arguments, in that thread you cite. If you think atmospheric CO2 increase is a “random walk” like he claims, and that therefore it and global T increase are amenable to Dickey-Fuller and similar type co-integration tests, then you have screws loose on the issue, period. Atmospheric CO2 increase results from a “random walk” process in the same way that rising water level in a sink with the faucet on and drain plugged is a “random walk”. [Guess what–a cumulative sum will be highly auto-correlated and therefore exhibit characteristics of random walks if you ignore the physical evidence that tellls you that it ain’t]. Not to mention other fatally flawed statements and arguments.

      • Jim, in all honesty you don’t sound like a sensible Ph.D., but rather like a street thug.

  11. Update: As the conversation is getting a bit heated, let’s reset the terms. I really would like this to be a place where those on either side of the fence can have a civil–almost friendly–discussion. That requires effort on both sides of the ball (why am I using a sports metaphor? …oh, yeah, that’s why…)

    Please provide answers to the following questions and discuss in a civil manner:

    1. Is Matt Ridley correct to be concerned about this issue? Is this a ‘deal-breaker’ for climate science and its former conclusions of .2C per decade for the early part of this century?

    2. Is the volume of aerosols adequate to produce the effects claimed for it?

    3. Are some of the properties of aerosols producing more than one climatic effect (e.g., blocking sun and lowering heating, nucleating clouds with unknown effects, etc.)?

    4. Is the treatment of aerosols in the literature proportionate to their real or potential effect(s)? Are they, for example, handled appropriately in AR4 (or AR5 if anyone has read the relevant sections in the leaked draft)?

    5. Are estimates of aerosols treated appropriately in the employment of GCMs?

    I really want your opinion on these issues. I am really not interested in your negative opinions of others on this thread–especially as some of you who are fussin’ and feudin’ are really good people who just are both commenting and reading too quickly.

    I speak as someone who often commented in such a fashion in the past. It got me nowhere in the conversation–I know what I’m talking about regarding commenting.

    • Why don’t you put a leash on your ***** attack dogs, then? They are apparently free to attack anyone who accepts the science, while those who accept the science are consistently censored and harassed on your blog.

      (Intriguing introduction of a theory that the Iliad was not in fact written by Homer, but by another Greek writer with the same name, reluctantly removed by blog administrator–h/t to RAH)

  12. I have another question or three–do aerosols mix as quickly and thoroughly in the atmosphere as GHGs? If not, can regional analysis be brought to bear on their effects? Has it been tried?

  13. You still need to present evidence that there is a “gap”. I’ve presented evidence that there isn’t.

    • From your link we get this statement from the IPCC’s AR4: (

      Some potentially important forcings such as black carbon aerosols have not yet been considered in most formal detection and attribution studies.”

      And this from the same source: “However, uncertainties still exist in the magnitude and temporal evolution of estimated contributions from individual forcings other than well-mixed greenhouse gases, due, for example, to uncertainties in model responses to forcing.

      And this: “Uncertainties remain in estimates of natural internal climate variability. For example, there are discrepancies between estimates of ocean heat content variability from models and observations, although poor sampling of parts of the world ocean may explain this discrepancy. In addition, internal variability is difficult to estimate from available observational records since these are influenced by external forcing, and because records are not long enough in the case of instrumental data, or precise enough in the case of proxy reconstructions, to provide complete descriptions of variability on decadal and longer time scales.”

      Thanks for the link. I don’t believe it settles any arguments here.

      • You’re missing the point. The graph there shows no evidence of a gap. If you think there is a gap – well, feel free to present evidence for it. But so far you haven’t. If you want an explanation for something, you have to tell us what that something is.

        For all I know this is such a common meme is the septic-o-sphere that MR feels no need to explain it, but for the rest of us out in the real world we need to know what it is you’re talking about.

      • tom,
        You have gotten to the nub, and we see the believer response: deny (ahem) the evidence when quoted ver batim.
        I believe you ran into this in the response (or lack of same) in your Crutape Letters. When confronted with in-full quotes in context, the true believer will still ignore the issue, blame the one bothering their faith experience and dissemble, distract and call names.
        I just read your request. I will focus on the issues and civility. I am certain the true believers will not do the same.

      • A gap between modeled predictions of temperature rises and actual observations? I think all you need to do is get over to Lucia’s excellent website The Blackboard and you can find ample evidence that there is a gap. You might start here:

        “As long as I’m showing the 2001 graph, I’ll discuss what this communicates. When the analysis using the assumptions discussed briefly above is performed, a trend equal to exactly 0.2 C/decade falls outside the ±95% confidence intervals for the trend consistent with the NOAA/NCDC data observations. This means that based on the assumptions of this analysis, if we select a confidence level of 95%, we should treat the assumption that the trend is 0.2 C/decade as false.”

        You might then continue here:

        ” Specifically, trend analysis beginning with start dates prior to 1980 all indicate we should reject the multi-model mean trend as reproducing the observed trend.”

        And here:

        “Are any of you wondering how the model-mean and multi-model mean trends from AOGCMs compared to observations? The main answer is: About as poorly than the trends from individual realizations discussed Tuesday.”

        I actually think a lot of people besides Mr. Ridley understand that there is a gap between models and observations. I find it surprising that you do not.

      • Errm, are those the results that MR and you meant? It would have been nice to say so at the start in that case. But that doesn’t make sense. MR talks of the last 34 years, the pic you’re now using is only since 2001. Using T trends over such a short period, and using some unclear method of calculating uncertainty limits, won’t do.

        is better, because it goes further back. However, the obscure use of MEI in the regression means I don’t trust her error bounds.

        Also, I don’t trust her model-obs errors. If you look at you’ll find a totally different picture. The uncertainties are much larger, and destroy any stat sig disagreement, but also the absolute errors are smaller.

      • I do trust her calculations. She has run them for many different timeframes to avoid the charge of cherry picking. Feel free to investigate on your own.

        I am not holding Ridley to that particular timeframe.

        I think it is clear from what you write (and your choice of Open Mind as source) that you are not in a position to judge Lucia’s work. But maybe Ian Joliffe will weigh in…

        Important to this discussion is, were climate models created to provide realistic temperature projections? I see no indication that they were and some discussion around the point that they were not. General trends and movements, yes–and I think they’ve done a creditable job of that–they caught the overall rising trend and caught the Pinatubo cooling quite well.

        I would quote Kevin Trenberth from Nature blog on this: “The IPCC instead proffers “what if” projections of future climate that correspond to certain emissions scenarios. There are a number of assumptions that go into these emissions scenarios. They are intended to cover a range of possible self consistent “story lines” that then provide decision makers with information about which paths might be more desirable. But they do not consider many things like the recovery of the ozone layer, for instance, or observed trends in forcing agents. There is no estimate, even probabilistically, as to the likelihood of any emissions scenario and no best guess.

        Even if there were, the projections are based on model results that provide differences of the future climate relative to that today. None of the models used by IPCC are initialized to the observed state and none of the climate states in the models correspond even remotely to the current observed climate.” (

        Now, you like Tamino’s Open Mind. I prefer Lucia Liljegren’s Blackboard. I think it is common knowledge that there is a distinct gap between modeled temperatures and observed results. You evidently do not.

        I also think that that is a red herring that does not impact an honest discussion of either climate change in general or aerosols in particular.

      • I was with you right up to the end, then you lost me. We disagree over whether there is a gap that needs explaining. You prefer one set of data, I another. But then you say “is a red herring that does not impact an honest discussion of either climate change in general or aerosols in particular”, which is incomprehensible. MR’s entire point is predicated on the existence of a gap; if there isn’t one, his point falls (though you can repurpose this to a discussion of aerosols, but I think that would be confusing).

        You quote Trenberth, who correctly points a number of things that we all know, I hope. The models were made for a variety of purposes, simply understanding the climate being one of them. They are intended to provide projections that are useful in considering the future but do not (as we all know) track year-to-year changes in an accurate manner.

      • I’m sorry. I should have been clear that my suspicion is that it is Mr. Ridley that is employing this gap as a red herring, although it is equally possible that he thinks models are supposed to be able to provide accurate predictions.

      • “There is no estimate, even probabilistically, as to the likelihood of any emissions scenario and no best guess.”

        If you read the story lines they are dependent on some difficult policy areas. The IPCC being just a one fish in the world of geopolitics is not in a position to resolve those policies issues.

        I.E. Scenario A1T depends on technological integration. Hindrances to technological integration include proliferation and safety concerns. This is an IAEA and security council matter…not an IPCC matter.

        If we read the Cancun resolutions..’technology transfer’ was high on the list of things to do.

        The IAEA additional protocols have been working their way thru the diplomatic world for more then a decade.

        So a simple minded person like myself can actually come to an almost certain conclusion as to what scenario is the ‘preferred’ scenario….IPCC A1T…what I can’t determine is how long the delays will be due to various concerns that have nothing to do with economics.

    • Wm

      you still need to produce evidence that is actually evidence, not some statistical whet cream.

    • “The multimodel average tropospheric temperature trends are outside the 5–95 percentile range of RSS results at most latitudes. The likely causes of these biases include forcing errors in the historical simulations (40–42), model response errors (43), remaining errors in satellite temperature estimates (26, 44), and an unusual manifestation of internal variability in the observations (35, 45). These explanations are not mutually exclusive. Our results suggest that forcing errors are a serious concern.”

      Click to access 1210514109.full.pdf

      Santer was heavily cited by the Warmist community when he concluded that warming trends were consistent with GCM’s. (Although in his prior analysis he stopped near the top of a super el nino event.) Now that he has concluded GCM’s are no longer consistent, Santer is off the xmas card list.

      • Here’s something written in a language Nitschke might actually understand:

        Now, will he stop trying to misrepresent science in the future? I doubt it.

      • Ah, this is a much better ref than Lucia’s stuff. Why didn’t TF start here? This is trop temp; Lucia uses sfc, I think; and MR doesn’t specify. Nevermind, its valid to talk about either.

        I can’t really provide a good rebuttal to the Santer paper. But I’ll have a go. Note that its mostly interested in fingerprints, which is why it starts with strat cooling: The lower stratospheric cooling trend is nearly 10% larger in the more realistically forced O3+V subset than in the BASE subset, and it is closer to the observational result. Handwaving about cloud feedback doesn’t help there, so MR is understandably silent about the strat trends.

        For the trop, we have In the lower troposphere (Fig. 2), model TLT trends over the satellite era vary from 0.121 °C to 0.392 °C per decade. Those most certainly do overlap the observed values. Which fits in what I’ve been saying for ages (and I even have a paper implying it): that just averaging all the AR runs together isn’t a good idea, because some of them are very poor indeed.

        But the key thing you want is “There is no overlap between the 5–95 percentile ranges of the RSS trends and the multimodel average TLT trend results”. I’m not quite convinced this is what you think it is. The RSS bounds appear to be generated from their “percentile realizations” and these represent ““sampling error, premerge adjustments to each individual satellite, and the merging procedure” but not, it would seem, the trend error to be expected from decadal noise. If that isn’t accounted for, then there is a gap in your gap. Furthermore, “and the multimodel average TLT trend results” isn’t clearly defined. Defined as the range, then clearly there is overlap. So presumably they mean mean+/- SD? But then that, too, doesn’t account for decadal noise.

        > Santer is off the xmas card list.

        Grow up.

      • @Connolley

        I don’t think anyone is asking you for a rebuttal. You appear to have moved the goal posts. Twenty one researchers including Santer worked on this peer reviewed paper… You dismiss it within a few hours of seeing it, by means of expressing your opinion in a blog post. Of course, you’re welcome to do that, although it strikes me as a farcical exercise. You requested evidence and I provided it. If you don’t like the evidence and wish to think up reasons to dismiss it, well warmists and sceptics play that came every day, and it’s largely a waste of time. If you think the paper’s conclusions as I’ve cited them are wrong, you’d be better served to get your point of view across by getting a comment published in a journal.

      • > You dismiss it… the paper’s conclusions as I’ve cited them are wrong…

        No, on both counts. I don’t dismiss it; I think its conclusions are quite likely to be valid, and as science its certainly valid. But you could try reading what I wrote, which was to doubt your facile interpretation.

        Note that, contrary to what you say, you haven’t cited its conclusions. Those are in the “Conclusions” section. You’ve cited (actually, quoted) the “Zonal Mean Temperature Trends” section. That’s because the paper is actually about fingerprinting, mostly.

        In fact, you’re the one “dismissing” it because you’re cherry picking just a small part of its conclusions. Or would I be correct in assuming that you accept its conclusion that “Tropospheric warming is mainly driven by human-caused increases in well-mixed greenhouse gases (16, 29). The multidecadal cooling of the stratosphere and warming of the troposphere, which is evident in all satellite datasets and simulations of forced climate change examined here, cannot be explained by solar or volcanic forcing, or by any known mode of internal variability.”

      • @Connolley

        The quote I cited here was directly related to your specific claim that there was no evidence for model inconsistencies. Obviously, the authors of the paper identify inconsistencies with trends and regard them as serious. Refer to my quotes. You disagree. That’s fine. But I will stick with the claims in the paper over your opinions. My apologies if you’re offended.

  14. Hi Tom,
    When you quote Nathan as saying ” This means that based on the assumptions of this analysis, if we select a confidence level of 95%, we should treat the assumption that the trend is 0.2 C/decade as false,” your readers should know that statisticians generally do not refer to things as false or true. You either accept or reject the null hypothesis at your pre-determined confidence level.

  15. Advance notice to all–starting in the very near future (half hour?) I will be turning my attention to far more important subjects related to the performance of one of my city’s corporate surrogate athletic clubs in what is the last game of the season. I am hopeful that they will match the performance of another one of my city’s surrogate athletic clubs and will report back on the outcome for those of you possessed of common sense or living in a country that sensibly ignores such events.

  16. Tom, my prediction – SF by 3, 27-24, for what it is worth.

  17. There are several Wikipedia pages that demonstrate there is a significant gap between the climate model forecasts and the observations to date.

  18. Aerosols and TSI are in a certain sense climate model “plugs”. (TSI less so.) But you can’t just change a few variable assignments to get a different tuning result. They are too complex for that. But one clue to the fact that they *are* ‘tuned’ is to look at how well they hindcast and yet how poorly they forecast. Attempts at building models designed to predict very complex phenomena, such as stock market movements, have also exhibited similar patterns of behaviours. Or in other words, their hindcasting was very good, but their forecasting was of no value.

    Also, if you accused a climate modeller of using “plugs” or tuning his model arbitrarily he would no doubt be offended by the accusation, because it’s far more complicated than that.

  19. Well, my team was vanquished, despite a valiant effort in the latter stages of the game. They are, however, still high in my regard. Perhaps another year.

    • God has granted you the Giants’ crowns in two of the last three years, don’t get greedy.

    • Your coach who is a superior tactician and never makes mistakes made uncharacteristic mistakes in the big game. I’m also a bit surprised he didn’t go for it on 4th and 6 down 22 even if it was the 3rd quarter.

      Patriots had the same 1st and goal from the 7 in their undefeated season, and scored too quickly. Wish the 49ers had just run run run there.

      • I actually think John outcoached Jim big time. The 49er Harbaugh didn’t have a contingency for the safety play. Three straight to The Crustacean Sensation was going to the well once too often. Niners didn’t run one trick play. They always run one trick play, going back to Walsh.

        That said, you give up a special teams touchdown, throw a pick and cough up a fumble, it’s a long road ahead of you.

    • The third half of the game- the one that started when the light ballasts warmed back up- was one of the best half’s of football ever played.

  20. Pingback: Matt Ridley’s Fourth Test | The Lukewarmer's Way

  21. Just discovered this thread and I can’t let Jim Bouldin’s comments go unchallenged. He desribes the perfectly sensible statement by VS “Actually, statistically speaking, there is no clear ‘trend’ here, and Ordinary Least Squares (OLS) trend you estimated up there is simply non-sensical, and has nothing to do with statistics.” as “one of the stupidest comments in this debate that I’ve ever read”. The point is simply that while it is possible to fit an OLS trend to any data this is a purely alegrbaraic operation. The importatn question is whether the trend is significant. To find that out we need to make sure that the equation is properly specified. The ordinary significance claculations require that the residuals from the equation are effectively random noise (otherwise thereis structure in the data which is not being taken into account). If you just fit a simple linear time trend to temoperature data by OLS the residuals will not be well behaved in this sense and the conventional significance calculations will be meaningless.

    • Yeah, but if we’re going to beat Bart’s for the record longest thread, I’m gonna need more coffee.

    • There’s more to what “VS” was claiming than that– a lot more–but I’ll be putting up a post on my blog on exactly the issue you raise in the next 24 hours, because it seems to be a point of great confusion, and it is on exactly that point that he/she is wrong. I was going to try to work in issues in Bayesian approaches also, but I’m going to have to drop that.

  22. Jim,
    Of course there was a lot more to the VS thread. All I was saying was that the quote you chose was perfectly sensible, not stupid. Look forward to your post.

    • mike, if I say things like “OLS estimates are nonsensical and have nothing to do with statistics” and “atmospheric CO2 increases are explainable as a random walk process” then I’ve made really stupid statements. Even if the trend is best explained as a random process–which it is NOT–it’s just an empirical fact that it’s still a trend regardless of the cause. What he/she meant to say was that it’s a *spurious* trend caused by auto-correlation in the data, but again, even this argument is wrong.

  23. Jim, I think you will find if you look at the context that VS was objecting to a simple ols calculation using “standard” calculations of significance with no attempt to provide any discussion of the residuals. The random walk discussion got a little confused because, while a random walk has a unit root, it is not the only stochastic process with a unit root. I don’t follow your discussion of trend. A random walk is often described as a stochastic trend, but it is different from a deterministic trend and there are tests to distinguish them, that’s where your augmented Dickey-Fuller test comes in.

    • mike,
      On the trend terminology question–I agree and your terminology is best. But call it “stochastic” or call it “spurious” –same underlying idea–I was just saying VS should have at least used some qualifier.

      I know there are tests for assessing whether a stochastic process can “explain” a time series, and Dickey Fuller is one. The issue is bigger than that: you **have** to bring in external data to be able to truly discriminate between hypotheses. Dhogaza tried to tell him this repeatedly. But even if you *don’t* do that, you can still make a much more confident statement, based on likelihood ratios, which he did not do that I’ve seen.

    • I’ve been trying to wade through that discusssion at Bart’s and even just rudimentary skimming of each comment is taking forever. That post and thread are fascinating from several angles. As near as I can tell, “VS” has no real concept of what maximum likelihood estimation is all about and how it would seriously challenge his basic thesis. It’s certainly not referred to in his argument. He instead seems completely hung up on what he believes to be a certain set of strict, foundational principles and tools to use when conducting an analysis, but which are either very questionable or known to be wrong or at best, overly simplistic or generalized. It’s very clear that he is heavily influenced by economists’ approaches to statistical analysis, where much less is understood about cause and effect than in the physical sciences, and in which there is therefore a large emphasis on the kinds of pattern indication tests that he refers to over and over. Those things are not proofs of anything–they are one type of evidence, not a be-all and end-all. He seems not to understand that physical data have to be referred to as the ultimate arbiter, even though he says otherwise. If you take his arguments to their natural limits, you will conclude that you can never prove anything about cause and effect, anywhere. The idea that he somehow took everybody to school on these issues is completely ludicrous.

      Still, I’ll finish up what I started and post it since I’ve invested some time in this and have a different argument, though it won’t be as soon as I’d hoped.

      • Jim, I swear I’m going to scrape that thread and save it somewhere, in case Bart ever deletes his blog. That whole thing is a piece of work.

      • Yeah it’s absolutely incredible Tom, never seen anything like it. Bart deserves some serious respect for that one post alone. Also, although I’m clearly critical of VS’ statistical statements, he does make a number of good comments in there about the conduct of the discussion and whatnot. A lot of people do actually.

  24. Tom, please forgive me for ignoring the fascinating gladiatorial combat above and attempting to answer part of your questions.

    The Pinatubo cooling was the result of sulfates ejected into the stratosphere. Nathan Myrhvold’s company Intellectual Ventures, the modern counterpart of Edison’s patent empire, takes sulfates so seriously that they are working on a “hosepipe to the sky” to spray sulfur dioxide into the stratosphere.

  25. Jim, It seems to me perfectly clear in context (and in popular usage) that the trend VS was criticising was a linear deterministic trend. He said that just fitting an OLS trend to a series with a deterministic time trend without considering the character of the residuals has nothing to do with statistics, and he was right.
    As for the rest, your accusations are insufficiently precise for me to be sure what you are saying. Perhaps some background is in order. Time series statisticians and econometricians have known for some time now (indeed the first papers were published in the 1920s, though the full analysis happened in the 1970s and 1980s) that there are problems in analysing relationships between non-stationary variables. The most famous example is the Granger and Newbold 1975 demonstration that two random walks will exhibit very high degrees of correlation a lot of the time, even though by construction they are quite independent of each other. There is therefore a danger of spurious correlation. Therefore the first step is an analysis these days is to see if the variables to be analysed look as though they are generated by a stationary process or not. This is normally done by breaking the variables down into a deterministic bit given by their own past history and/or a deterministic time trend and an iid residual. There may be more than one such description and there are issues about how to choose between the descriptions. The fact that there may be a random component does not mean that the variable is governed by quantum type non-determinateness, merely that the description is incomplete though only out by random factors. Co-integration does not come in at this stage, and the VS thread never seemed to reach discussion of co-integration. If we have a good description of a variable in these terms then that is a fact that requires explanation, not an impossibility.

    • Hi Mike, thanks for taking the time to elucidate your thoughts, necessary if we are to avoid talking past each other, which we still are but getting closer I think.

      I’m not really all that hung up on his “nothing to do with statistics” statement. He still shouldn’t have made it as an opening statement, but that’s *not* my main point. I understand exactly your points, and his, regarding spurious cross-correlation due to auto-correlation within each of two different time series that actually have no cause and effect relationship. I get that, I do, and I’m even going to demonstrate it in my post. As you increase the lag-1 ac coefficient, you increase the chance of spurious correlation between stochastically trending data series, because the chance of such a trend within each series increases, greatly, as you approach a random walk where ac1 = 1.0, and it’s almost guaranteed when > 1.0.

      The problem here is that just because D-F (or other) tests indicate that your series have unit roots, and are therefore non-stationary, does **not** mean that either of the trends, or their correlation with each other, are *necessarily* stochastic in origin. It’s entirely possible there *is* in fact a real (i.e. deterministic) trend in one or the other series. And the only way to evaluate the relative possibilities of these two hypotheses (stochastic vs deterministic trend) is to assess their relative weights of evidence, by arguably the most important concept in all of statistics: maximum likelihood estimation. But he didn’t do that, at least within the first 25% of the comments, which is as far as I got, where he most certainly should have.

      When, on top of that, he stated that atmospheric CO2 increase is itself a random walk, that was a dead giveaway, because there is absolutely no way you can make such a claim, given what we know about the science on that. The scientists are not fishing around for variables to see what might correlate with what. They have a definite and constrained set of possible mechanistic explanations. It’s not like economics, where that’s often not the case.

    • Too frazzled to write the whole thing up at once, but a first post on the issue is now up. The most important stuff will come later.

Leave a Reply

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

You are commenting using your 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