On the subject of climate change

David T. C. Davies: I read recently that Tuvalu would be wiped off the face of the earth within 10 years if we did not do something about global warming. The only problem is that the article was 10 years old, and all those islands are still there. Is this not just more exaggeration from those who want higher electricity prices as a result of decarbonisation?

Caroline Lucas: I do not really know where to start to respond to such an ignorant intervention. I will not even bother wasting my time with it.

I think that one goes on points to Mr. Davies.

53 thoughts on “On the subject of climate change”

  1. From time to time, Tim, some of us try to apprise you of the likelihood that the whole thing has degenerated from a mixture of hubris and incompetence into straightforward fraud.

  2. Seems like drivel from Davies to me. If his only point is that the meeja frequently talks nonsense then yes, he’s correct, but its not an interesting point.

  3. Do you think Tim? Because to me, a question that asks me to speculate on the motives of an unspecified source really isn’t worth replying to. (Despite which, having expressed her frustration Lucas did reply, but you cut that.)

    I spent a minute trying to find an article like the one Davies claims to have read, but couldn’t. Here‘s a ten-year-old report on Tuvalu by the BBC:

    Scientists have predicted an 88cm rise in sea levels in the next century – if that is proved correct, Tuvalu, like other low-lying atoll countries in the Pacific Ocean, could find itself underwater within 50 years.

  4. I’d expect better out of our elected representatives that ‘Na, na, na, na, not listening! *sticks fingers in ears*’ but then, this is a Green MP we’re talking about…

  5. I’m pretty sure I read on Wikipedia that global warming is real, and in no way a massive scam to create jobs for marginally employable leftwing scientists named after famous Scottish comedians.

  6. I’m seeing climate change/global warming as a religion. A belief system with many who follow it.
    I don’t share that religion and tend to get annoyed at those who regard me as a sinner who needs to repent!
    Just turned the central heating off for a bit a half hour ago, hopefully won’t need it on until this evening.

  7. The failed Prof Cockend will be along shortly (if not already) to say, it was never said in the first place, even if it was it was the media, it does not matter, you are evil, and we are right all the time.

  8. Martin Davies,

    I’m seeing climate change/global warming as a religion. A belief system with many who follow it.

    Well, in terms of how they talk about it, most followers treat it as a religion.

    1) It has a priesthood “the scientists say so” (yet they ignore scientists that say we should use more GM and nuclear power).

    2) It has token sacrifice: recycling.

    3) It has a prediction of armageddon.

    4) It is ultimately about control.

    I’m not referring to the Lomborg/Worstall take on climate change here that is based on looking at a problem scientifically, but how many advocates (and especially watermelons) use it.

  9. Beats me why this massive sea level rise is overwhelming Tuvalu & completely undetectable in Essex. Must be the wrong sort of global warming.

  10. … greenery as a religion …

    5) they erect their symbol everywhere to reinforce their dominance – but their crucifixes revolve sometimes.

  11. Just turned the central heating back on. Not done it to annoy the religious climate people, thats just a beneficial side effect.

    New ice age was coming when I were a lad now its increase in temperature that should eventually reduce my gas bill, won’t have to have the central heating on so much.

  12. When someone (certified expert or not) predicts something and it doesn’t happen- e.g. no more snow in ten years, rising temperatures- I am inclined to doubt his expertise.
    When someone (certified expert or not) proposes a course of action that involves massive expense and a total reorganisation of society I want to see evidence- and should that evidence be inadvertently lost (Like the eminent professor Jones’s for example) I will be as impressed as I would be with the opportunity to invest in a gold mine where no assays were available.
    When the course of action proposed is excessive bearing in mind the problems predicted I begin to doubt the motives of the proponents

  13. Supercilious, arrogant, condescending – the remainder of the response from Lucas is, if anything even worse, so Tim did Her a service by not printing excerpts from it…..

  14. Paul Ralley,

    Most animals aren’t facing extinction because we have them in zoos and wildlife parks, or in private hands. Tigers are NOT facing extinction. They might be in the wild, but there are more than enough tigers being looked after.

    And honestly, in many cases it’s entirely logical. If I lived in a village near rhinos, and someone was going to come along and kill them I’d be pretty happy about it, considering how destructive rhinos are.

    It’s only the panda that seems to be a problem and honestly, if you’re going to be picky fuckers about what you eat and who you shag, you deserve to go extinct.

  15. “Climate change” is pure and total unadulterated bullshit, the shittiest bull that ever dropped from the prizewinning most shit-filled bull ever in a bullshitting contest on planet Bullshit in the Bullshit quadrant of the Bullshit galaxy.

  16. Lucas goes on to say:

    “the bottom line is that what we are discussing is literally life and death. People’s life or death is at stake today.”

    Just how fast does she think the water level’s going to rise in Tuvalu? Is there some reason why we can’t rescue people from sinking islands?

  17. Perhaps I can help with the difference between a scientific belief and a religious one.

    Science proceeds by the collection and analysis of data, and the development of theories to explain it. Work of sufficient quality is published in scientific journals, where anyone with sufficient technical grasp can read it and understand what has been done. If you come up with a significant improvement or correction to published work, then you can publish yourself, gaining the esteem of your peers. This creates a process by which scientific knowledge and understanding improve with time.

    The IPCC WG1 reports are essentially reviews of the current state of the scientific literature on AGW. They are of course imperfect, but problems with them can be identified and corrected by following their scientific sources.

    Religion, on the other hand, is simply a matter of faith. You believe what you believe, because of an inner conviction that it’s right. There is nothing other than the reliability of your instincts to relate your religious beliefs to fact. (Unless of course your beliefs are being guided by an actual deity: let’s not get into that.)

    Examples of religious beliefs would be “I


  18. Examples of religious beliefs would be “I[apostrophe]m seeing climate change/global warming as a religion” or “[quote]Climate change[unquote] is pure and total unadulterated bullshit”

  19. Squander Two,

    Is there some reason why we can’t rescue people from sinking islands?

    Population: around 10000. For less than a billion we could rehouse them somewhere else. Much cheaper than spending money on a problem that may not actually be as bad as predicted 20 years ago.

    Just out of interest: anyone know how much the sea level has risen around Tuvalu in the past decade? I presume it’s being measured and someone can check it against the predictions that PaulB has linked to?

  20. Squander Two: you might judge, albeit on the basis of emails taken out of context, that some small number of climate scientists have been partial in presenting their results. That would lead you to treat with suspicion anything presented by those scientists. It would, however, tell you nothing about the climate science corpus as a whole, which, in a non-religious analysis, is to be evaluated on its own merits.

  21. PaulB,

    Sorry, but what utter bollocks.

    First of all, the emails weren’t taken out of context. Their defenders have done good business claiming that quotes from them were being taken out of context — for instance, a lot of non-mathematicians quite wrongly jumped on the use of the word “trick” as implying culpability, when of course it doesn’t — but the leak was in fact an entire corpus of emails, not just snippets, and it was the context that made them bad. Some of the sound-bite quotes are mildly damning but maybe excusable. The context is unequivocally damning.

    Secondly, they weren’t merely partial in presenting their results. According to their own words, they can’t even reproduce their own results, they can’t get their software models to work without fudging them, they can’t debug properly, and they apparently oppose the scientific method itself — the bit where you’re supposed to publish your data precisely so that other scientists (yes, even [gasp!] ones who disagree with you) can try to find fault with it. They even conspired to break the law to prevent the release of their data. Referring to this wholesale fraud and premeditated criminality as “partial in presenting their results” is grossly misleading.

    Finally, that “small number” of climate scientists are the ones who have created the data on which the entire edifice of AGW rests. Without them, there might be a bit of warming, but there is no hockey-stick. And where are all the climatologists who have disowned the egregious actions of the UEA crowd and stopped using their flawed and compromised data but still claim that AGW is an urgent and real problem? Oh, there aren’t any.

  22. PaulB – tell me, have you ever investigated the global warming science firsthand? Done your own measurements?
    Or are you actually taking what someone else has published on faith?
    There are many scientific facts I’ve never seen for myself, never had a chance to see an atom even. Plenty of scientific theories get accepted as fact and later get disproved or junked in favour of the next theory by the scientific community.

    Plenty of instances of faith in this world. Some people go so far as to make a religion out of it.

  23. Squander Two: there were several investigations into the emails, which between them came up with no findings whatever of the fraud you allege. The reports did make some sensible and welcome recommendations to promote more openness.

    You’ve made a testable claim – that the “hockey stick” reconstruction depends vitally on unreproducible UAE results. Which results?

    Martin Davies: I’m not a climate scientist. If all you’re saying is that unless we conduct our own research into any particular scientific result we are trusting the rest of the world not to be engaged in a mass conspiracy to falsify the data, then yes. Not suspecting such conspiracies is no more a matter of faith than disbelief in Russell’s teapot.

    It’s normal for scientific theories to evolve. Occasionally they change completely. (But “plenty…get junked” is an exaggeration.) That’s the way science works, and that’s why we can be increasingly confident that the ones that survive represent quite good approximations to the truth.

  24. The parliamentary investigations into the emails missed the point entirely. Hardly any of the MPs involved knew a damn thing about science, unfortunately.

    Without getting into a long hard argument about the merits of the emails — I’ve made my point, you’ve made yours — I will just reiterate that I was responding to your claim that people who don’t believe the claims of the AGW catastrophists are subscribing to a religious belief. Those emails, regardless of to what extent you think they reflect on the field as a whole, are evidence. Making conclusions based on that evidence (and I did change my mind about a number of things when I saw it) is not a religious belief. You yourself have already conceded that point — you disagree about the extent of the implications of the evidence, but not that it is evidence, making further argument between us a wholly evidence-based argument, not a religious one — so perhaps you might now like to retract your slander.

    > we are trusting the rest of the world not to be engaged in a mass conspiracy to falsify the data

    No, no conspiracy required. Humans are perfectly capable of being wrong together without conspiracy.

    I believe that the UAE scientists honestly believe that you can generate data from computer models, honestly believe their own results, honestly believe that the consequences for the Earth are catastrophic, honestly believe that they know what are the best (or indeed only) political responses required, and honestly believe that it is therefore so important that government and the public be persuaded of their cause that it is OK to do a bit of data-fudging and law-breaking for the greater good. This, sadly, is how humans usually work, and is why Communism remains so popular no matter how many people it slaughters.

  25. “Squander Two: there were several investigations into the emails, which between them came up with no findings whatever of the fraud you allege.”

    None of the enquiries examined the validity of science, and most did not seriously examine the allegations, either. Certainly, there is no context or answers given.

    “You

  26. “Squander Two: there were several investigations into the emails, which between them came up with no findings whatever of the fraud you allege.”

    None of the enquiries examined the validity of science, and most did not seriously examine the allegations, either. Certainly, there is no context or answers given.

    “Youve made a testable claim – that the “hockey stick” reconstruction depends vitally on unreproducible UAE results. Which results?”

    He didn’t say “hockey stick”. The results being referred to are probably CRU TS2.6, as reported by ‘Harry’. But the comment could also refer to the HadCRUT homogenisation, or the Wang/Jones Chinese station metadata claims.

    The issue with the hockeystick wasn’t that it was irreproducible (McIntyre reproduced it eventually fairly closely, and the Ammann-Wahl paper did so more exactly) the problems were the mislabelling of data series, the short-centering error in the PCA, the fact that virtually all the weight was on a handful of corrupted data sets from one part of America, and most damningly of all, the fact that the reconstructed temperature had virtually no correlation with the actual temperature outside the calibration period. The cross-validation R^2 statistic was calculated by Mann (although he later lied about doing so), but the failure of this validity test was not published.

    But if you’d like to give me page and line number in the enquiry reports where all this stuff is addressed, I’d be happy to have a look, and review my opinions if required.

  27. Squander Two: You misunderstood me. I did not say that all objections to AGW theory are religious. Indeed, any scientific objections are welcome – it’s import to test the theory.

    However, I note that in your long comment you declined to answer the one scientific question I asked you in my short one.

    NiV: I can’t imagine why you think enquiries into alleged impropriety at UAE should have investigated temperature reconstructions. Suffice it to say that there have been very many reconstructions. AR4 cites twelve of them, with broad agreement between them.

  28. “I cant imagine why you think enquiries into alleged impropriety at UAE should have investigated temperature reconstructions.”

    Because that’s what some of the allegations were about?

    “Suffice it to say that there have been very many reconstructions. AR4 cites twelve of them, with broad agreement between them.”

    Classic! “The errors don’t matter, because my friends all got a similar answer.” That’s better than “The dog ate my homework.”

    The problem with your answer is that the rest of the twelve also used variants on the same corrupted sources, had many of the same flaws, and failed the same statistical tests. A system that will pass one broken hockey stick will pass lots of them.

    But the primary issue here is the response to bad science being discovered – which is not to throw it out and take steps to make sure it doesn’t happen again, but instead to try to argue it “doesn’t matter”. If you get the “right” answer by the wrong method, you think that’s still good science, perfectly acceptable, nothing wrong with that at all. That’s voodoo science.

    AR4 still cited the MBH98 Hockeystick, after it’s flaws had been revealed. The original was published and promoted by the IPCC after the author knew that it failed validation. And nobody else checked it. This one graph is just a symptom; the bigger problem is the process that allows this to happen, and condones it.

  29. The alleged flaws have been much examined, and shown not to be problems. But people motivated by considerations other than science repeat them anyway. e.g the “short-centering error in the PCA” you complain about is not an error at all, it’s a somewhat arbitrary choice in the statistical procedure which, if the procedure is implemented properly, makes no significant difference. And other statistical methods not using PCA at all give very similar results.

  30. “the alleged flaws have been much examined, and shown not to be problems.”

    Shown where? By who?

    “e.g the

  31. “the alleged flaws have been much examined, and shown not to be problems.”

    Shown where? By who?

    “e.g the “short-centering error in the PCA” you complain about is not an error at all, it-s a somewhat arbitrary choice in the statistical procedure which, if the procedure is implemented properly, makes no significant difference.”

    It is an error. It’s the wrong thing to do. It gives the wrong answer.

    PCA, roughly speaking, takes an ellipsoidal spread of data and calculates the axes of the ellipsoid. You first subtract the average of the data so you’re measuring things relative to the centre of the ellipsoid, and then you measure the size of the spread in various directions about the origin.

    If you subtract the wrong value, you are measuring the spread not from the centre of the ellipsoid, but from a point offset from it – in this case a projection of the centre onto one of the coordinate hyperplanes. This means that in the second stage the variation from the origin is mostly the offset, not the spread of the data, which also throws off the calculation of all the subsequent axes, and you get the wrong answer.

    It’s mathematically wrong. It’s nonsensical and meaningless. But some people are so fanatical about defending it at any costs, they’ll tell you any old rubbish and rely on their authority to sell it.

    If you do it properly, the hockeystick signal drops from the first principal component (the longest ellipsoid axis) to the fourth (i.e. the fourth longest axis), which was originally not included, causing the hockeystick shape to disappear. You then have to come up with some contrived excuse why you ought to take more components so you can include it. (As well as retaining all the other errors.)

    The origins of the hockeystick signal have been traced. There’s a relatively small group of trees in the US that spike dramatically upwards in the 20th century. They were already known not to be temperature proxies when they were sampled – it’s thought to be either a fertilisation effect or the results of physical damage to the trees. Local temperature records show no such spike. But they’re included in the data and so their signal appears low down in the PC4. It’s a significant signal, but not one related to temperature.

    And in any case we know it’s the wrong answer because in the critical early parts the result is not correlated with the measured temperature in the verification period. As was known before it was published, it’s not reconstructing temperature.

    But again, diverting attention to whether it changes the conclusions or towards other studies misses the point. There is an error in the method, and the climate scientists know it, but they still cite and use the result. They still excuse it and defend it. That’s bad science.

    The fact that they do it makes rational people wonder what else they might be doing it for. Why should we believe anything they say about anything else, when they’ll tell us with a straight face that blatant errors in the calculation “don’t matter”?

    If it was a real global emergency, do you think anyone would find that acceptable?

  32. Shown where? By who?

    Extensively in the peer-reviewed literature. Here’s
    Wahl & Ammann

    …the Mann et al. reconstruction is robust against the proxy-based criticisms addressed. In particular, reconstructed hemispheric temperatures are demonstrated to be largely unaffected by the use or non-use of PCs to summarize proxy evidence from the data-rich North American region. When proxy PCs are employed, neither the time period used to

  33. Shown where? By who?

    Extensively in the peer-reviewed literature. Here’s
    Wahl & Ammann

    …the Mann et al. reconstruction is robust against the proxy-based criticisms addressed. In particular, reconstructed hemispheric temperatures are demonstrated to be largely unaffected by the use or non-use of PCs to summarize proxy evidence from the data-rich North American region. When proxy PCs are employed, neither the time period used to [quotes]center[unquotes] the data before PC calculation nor the way the PC calculations are performed significantly affects the results, as long as the full extent of the climate information actually in the proxy data is represented by the PC time series.

    The simple reason why short-centring is not a problem is that the centring and standardization choices make no significant difference to the result so long as enough principal components are included in the analysis. (One simple test for ‘enough’ is to add the next PC and see if the result changes in a way that affects your conclusions). The error made by Mclntyre and McKitrick (your source, I suppose) was in using only two principal components, when this simple test would have shown them that several more were required by their implementation. The order of the PCs is not important so long as you include enough of them.

  34. I already addressed that point.

    Wahl and Amman are incorrect in describing it as “the full extent of the climate information actually in the proxy data” because as I noted above, it was already known not to be related to climate. If Mann had chucked in the FTSE 100 index and the global number of telephones, those would also have appeared as a significant signal in some low-down PC, and would likewise have contributed hockeystick spikes to his “temperature”. That doesn’t mean it has anything to do with temperature.

    And your test for ‘enough’ wouldn’t work. You’d put in the first PC, find no hockeystick, put in the second PC, find the conclusion unchanged, and therefore decide that was ‘enough’.

    But again, none of this addresses the main point. The calculation was definitely wrong, but all anyone can do is make excuses for it.

    I’m not complaining, mind. The more examples there are of believers defending mathematical errors as acceptable in science, the better the sceptic case looks. Had you simply acknowledged that the calculation was wrong and not used it again – as any real scientist should have done – we would have no case. Science is an imperfect, trial-and-error process. Dropping it would have demonstrated climate science’s integrity in quite a powerful and persuasive way. But when people defend, and show climate scientists defending, what must be one of the most comprehensively debunked error-ridden papers of modern times, and cannot be persuaded to stop, well, it’s absolutely brilliant. It’s like free beer and doughnuts every day.

    The simple reason why short-centering is a problem is that it is mathematically wrong. Like looking at the fraction 26/65 and cancelling the 6’s on top and bottom. If you don’t think that’s important, I can’t help you.

  35. The hockey stick result is in no way dependent on PCA. You can do the analysis using any plausible method and get a similar result, because that’s what’s in the data.

    I agree that you can’t help me, nor anyone else who’s interested in the truth. I recommend anyone who wants to explore this to study the literature.

  36. “You can do the analysis using any plausible method and get a similar result, because that’s what’s in the data.”

    Ljunqvist and Loehle.
    http://wattsupwiththat.files.wordpress.com/2010/09/loehle_v_fig1.png

    Or Mann’s ‘censored’ directory.
    http://climateaudit.org/2005/02/13/errors-matter-3-preisendorfers-rule-n/

    And once again, you avoid the issues of:
    1. It’s wrong.
    2. They knew it was wrong.
    3. It’s not measuring temperature.
    4. It’s based on a small and unrepresentative number of trees.
    5. The trees in question were already known not to be temperature proxies, the spike is not correlated with local temperature records, and this was reported in the paper that published them.
    6. This is not the only error in the paper.
    7 It doesn’t matter how many times you try to make excuses for it, mathematical errors are never acceptable in science.

    “I don’t think we can say we didn’t do Mann et al because we think it is crap!” “We both know the probable flaws in Mike’s recon, particularly as it relates to the tropical stuff.”

  37. It’s not wrong. You’ve failed to understand what the PCA is doing here.

    It’s helpful of you to link to McIntyre’s commentary in which he tacitly admits that it’s his analysis that’s wrong – he’d wrongly omitted significant PCs.

    There’s a comprehensive review of temperature reconstruction methods (as of 2006) here.

    You’re quite right that Loehle produced a reconstruction which is something of an outlier. But lumping Ljunqvist in with him is odd because Ljunqvist himself described his reconstruction as “surprisingly similar” to that of Mann et al (2008). (Not surprisingly, Mann has improved his methods and data since 1998.)

    Perhaps we could move on. No one, including Mann, would argue that the 1998 methods were perfect. Could we agree to disagree (albeit by a considerable distance) about the boundary between imperfection and error?

    We now have many temperature reconstructions using various proxies and statistical methods, almost all of them in broad agreement. Do you have a problem with this consensus?

  38. “It’s not wrong. You’ve failed to understand what the PCA is doing here.”

    It is wrong, you’ve given me no reason to think otherwise.

    I’ve explained what the PCA is doing, and why what Mann did originally was wrong. I’ve also explained why the ‘corrected’ version, with more PCs taken, is also wrong.

    Loehle and Ljunqvist are outliers in terms of what’s published, but they’re not outliers in terms of what’s reconstructible. It’s very easy to reconstruct non-hockeystick series, it’s comparatively difficult to get ones that are hockeystick shaped. There are about a dozen corrupted data series around with spikes at the end, that are in many cases known to be for other reasons than climate – like changes in drainage, shade, fertilisation, disease – and the reoccur in every published reconstruction. Leave that tiny handful out, as Mann did in his experiment, and the shape disappears.

    I linked the wrong diagram previously, this is the one I wanted. http://wattsupwiththat.files.wordpress.com/2010/09/loehle_v_fig21.png It doesn’t look similar to me. And Mann’s 2008 paper invented entirely new errors, like putting in series upside down, so the warm bits in the past cancelled.

    “Perhaps we could move on.”

    We can move on when everyone stops citing and supporting the MBH98 hockeystick as valid science. So long as people keep supporting it, I’ll keep knocking them down. It’s to my advantage, it makes the consensus look bad, so why would I want to stop?

    “Could we agree to disagree (albeit by a considerable distance) about the boundary between imperfection and error?”

    Oh, I agree that we disagree. That’s the point, really. That’s the root of the dispute.

    On one side, people see badly documented, unreproducible results with mathematical errors on corrupted data as merely ‘imperfect’, but still acceptable, so long as it comes to the right conclusion. On the other side, there are people who think that it is only by applying the greatest care and rigour that science gives us any better assurance of truth, and that unreliable results have to be detected and ruthlessly weeded out. It’s hard not to fool yourself.

    Different scientists have different standards. Many climate scientists are sloppy. And other scientists (like me) are unimpressed. It’s bad enough when it happens in some obscure academic backwater, but you can argue it doesn’t really matter there. But when it’s to do with people’s health and safety, or large amounts of other people’s money, you don’t mess around. When it affects the lives of hundreds of people you take it more seriously, and when it affects millions, more seriously still.

    It is claimed that this science is reporting on a potential “end of the world” scenario. I find it simply inconceivable that it’s still being run by a bunch of bumbling idiots with no understanding of data management, software quality, model verification and validation, transparency, proper statistics, or scientific logic. And who have an attitude that regards any attempt to impose such as offensive and obstructive. They’re moving fast at the cutting edge of science, and don’t have time to waste shoring up the tunnel behind them.

    I certainly recognise the attitude, but I don’t think it would have been allowed to last so long if people really thought this was serious.

    “We now have many temperature reconstructions using various proxies and statistical methods, almost all of them in broad agreement. Do you have a problem with this consensus?”

    Yes. The problem is you keep on looking at the conclusions while I keep on looking at the methods. What evidence is there that what any of them is calculating is related to temperature?

  39. It[apostrophe]s very easy to reconstruct non-hockeystick series, it[apostrophe]s comparatively difficult to get ones that are hockeystick shaped.

    You’ve written approvingly of Ljungqvist’s reconstruction. So have a look at Figure 4 in his paper here. The figure shows mean centennial temperature anomalies, averaged over various classes of proxy. Cover up the left half of the charts (before 1400). Then compare the right half with the “50-year lowpass” line in Figure 5b of Mann et al (98) (which starts in 1400).

    What you see – no one could possibly deny it – is the same hockey-stick shape in both. The hockey-stick is not, as you strangely assert, an artefact of Mann et al’s statistical methods. It’s obvious in the raw data.

  40. “What you see – no one could possibly deny it – is the same hockey-stick shape in both.”

    Ljunqvist drops by about 1SD between 1400 and 1650-1850. The MBH98 hockeystick doesn’t. That was the point – to get rid of the MWP/LIA.

    But sceptics don’t deny the similarity in shape…
    http://www.cartoonsbyjosh.com/tipping_point_scr.jpg
    🙂

    “The hockey-stick is not, as you strangely assert, an artefact of Mann et al’s statistical methods.”

    Very strange indeed, since I actually asserted just the opposite. (see “The origins of the hockeystick signal have been traced”.) The hockeystick signal comes from a small group of tree-ring series with giant spikes at the end that were already known not to be related to local temperatures. The statistical errors pick out and over-emphasise them in the result.

    What should have been done is to take only series for which there were external, validated reasons to believe that they were related to local temperature. (By external I mean not based on the data’s correlation with temperature, otherwise you get a circular argument.) Then you average them, weighted by area. Then the correlation with observed temperatures is determined.

    Or if you’re going to identify temperature sensitivity by correlation with observations (which is not ideal, for all sorts of reasons), then you exclude those that are definitely not related, you calibrate the rest of them using only part of the temperature observations, do your reconstruction, and then test the correlation with the (unseen) rest of the observations.

    Mann did the latter. The subset of the reconstruction going back to 1400 has a 0.018 squared correlation with observed temperatures. Ammann and Wahl listed the correlations (very grudgingly) in table 1S. The r^2 correlations for the segments are: 0.018, 0.010, 0.006, 0.004, 0.00003, 0.013, 0.156, 0.050, 0.122, 0.154, 0.189. (Guess which one of those is the only one Mann reported in the original paper!) The early part of the record has virtually no correlation – less than 2% – with observed temperature. It’s 98-99% uncorrelated noise.

  41. The hockeystick signal comes from a small group of tree-ring series

    No. The hockey-stick is obvious in Ljungqvist’s various groups of proxies, most of which don’t include American tree rings.

    It’s very easy to reconstruct non-hockeystick series, it[apostrophe]s comparatively difficult to get ones that are hockeystick shaped.

    On the contrary. The raw data are shaped like a hockey-stick. It’s therefore very easy to reconstruct a hockey-stick shape. It takes considerable incompetence to get anything else.

  42. Hmm.

    Do you know what people mean by ‘a hockeystick shape’ in the context of the climate debate?

  43. Yes. They mean a graph showing a reconstruction of historical temperatures which is roughly in the shape of an ice-hockey stick, with the blade representing a rapid rise in temperatures over the last century.

    As you really ought to acknowledge, the only way to do a reconstruction and not get such a shape is to get it badly wrong, like McIntyre and McKitrick.

    In your comment #47, you say:

    Ljunqvist drops by about 1SD between 1400 and 1650-1850. The MBH98 hockeystick doesn

  44. Yes. They mean a graph showing a reconstruction of historical temperatures which is roughly in the shape of an ice-hockey stick, with the blade representing a rapid rise in temperatures over the last century.

    As you really ought to acknowledge, the only way to do a reconstruction and not get such a shape is to get it badly wrong, like McIntyre and McKitrick.

    In your comment #47, you say:

    Ljunqvist drops by about 1SD between 1400 and 1650-1850. The MBH98 hockeystick doesn[apostrophe]t. That was the point

  45. I’m not doing very well here. One more try:

    In your comment #47, you say:

    Ljunqvist drops by about 1SD between 1400 and 1650-1850. The MBH98 hockeystick doesn[apostrophe]t. That was the point [dash] to get rid of the MWP/LIA.

    That’s absurd in itself – the Medieval warm period was over long before 1400. But the detail is wrong too. I refer the interested reader to table 2 in Ljungqvist’s reconstruction paper. It shows in his reconstruction of Northern Hemisphere temperatures a fall from the 14th to the 18th centuries of 0.1 degree, and, for comparison, a fall in Mann et al’s roughly corresponding reconstruction (2008) of 0.29 degrees.

    I take it that you’re an enthusiastic supporter of Mann’s 2008 work.

    Ljungqvist shows no fall at all (actually a very slight rise) between the 15th and 18th centuries. Eyeballing Mann et al (98) I’d say it gets the same result, but it’s not directly comparable because it’s a global reconstruction.

    It’s worth mentioning for comparison that temperatures have risen by about one degree over the last 100 years. We are discussing very small differences in the slope of the shaft when compared with the rapid rise in the blade.

    What this seems to come down to is that you think that Mann et al (98) was lucky in getting the right result using poor methods, whereas I think the methods were not bad, and deserved to get it about right. Which seems not to be a very interesting thing to disagree about.

Leave a Reply

Your email address will not be published. Required fields are marked *