Bishop Hill AR5: dead in the water?
Jun 1, 2012 Climate: IPCC FOI Greens Steve McIntyre's latest post seems to me to be of huge importance. The refusal by Joelle Gergis and colleagues to release data behind their paper follows on behind similar refusals from authors in the same clique - principally Raphael Neukom. This stonewalling of reasonable requests represents yet another blow at the credibility of paleoclimate. To make things worse, the credibility of the Gergis paper is shattered by the revelation that it is based on circular reasoning - a fallacy that has been repeatedly noted in paleoclimate papers, yet one which is constantly given the seal of approval by peer reviewers in the field.
Despite the refusal of authors in the Gergis-Neukom clique to release data, as thing stand the IPCC will allow their work to be cited in the Fifth Assessment Report. This seems to me to be a ringing endorsement of pseudoscience.
If this were not a bad enough indictment of the IPCC, take a look at this report at Haunting the Library, which reveals Neil Adger, the man in charge of the AR5 chapter on "Human Security", as a member of an eccentric group called the Resilience Alliance:
Whilst the organisation might sound like a convention gathering from Star Trek, their aims are far, far, more serious and wide-ranging than that. Indeed, their ultimate goal, as stated on their website, is nothing less than what they term “Panarchy” in accordance with “nature’s rules” of “unpredictable change”. An important part of the philosophy of panarchy is that national governments are increasingly sidelined in favour of multi-jurisdictional institutions.
Is is possible to argue in these circumstances that the IPCC remains a credible organisation? I would say not. I wonder if the Fifth Assessment Report is dead in the water already.
Bishop Hill
Frank in the comments points out that the sorting was to the detrended temperatures. If so then this would mitigate against the circular reasoning criticism. I note Steve M's comments:
Gergis et al 2012 say that their screening is done on de-trended series. This measure might mitigate the screening fallacy – but this is something that would need to be checked carefully. I haven’t yet checked on the other papers in this series.
Reader Comments (69)
Harold:
I went and read the comment you linked to. I disagree and Steve M's fallacy is upheld. The argument posed seems to first of all state that autocorrelation is a factor. It isn't, the spurious correlation that arises will arise whether the time series is autocorrelated or not. It is the selection process that creates the hockey stick. Whether you take red noise, or autocorrelated noise, or noise based on a temporal correlation defined by a variogram or some other spatially or temporally correlated function makes no difference (actually as I understand it, the chance of finding a spurious correlation will increase with an autocorrelated series as opposed to random data with, say, a red noise spectrum).
The problem is the selection. If you take random series (non-conditional in the jargon), whether autocorrelated or not, and compare them to the temperature curve and only retain the ones that correlate then you are selecting random noise on the basis that it accidentally agrees with your prior assumption. The absolute confirmation of this is seen in the hockey stick itself: by selecting on random data that matches the temperatures in the modern period, the noise cancels out in the earlier period, hence the flat handle of the hockey stick. If the hockey stick produced the modern uptick plus the Medieval Warm Period it might be believable, but because everything cancels out in the earlier period it suggests something is wrong: the selction is on random noise and so the reconstruction is meaningless. Craig Loehle's reconstruction is the most believable in this regard, first because it contains MWP and LIA which are independently corroborated by anecdotal and contemperaneous records and secondly because it has no tree rings in it.
The final nail in the coffin is the Briffa "hide the decline" divergence problem. In the modern period the tree rings do not follow temperature, in fact the opposite: they are inversely correlated. So tree rings either (a) are not thermometers (and this was already a very doubtful possibility given the very low/no significance values of reconstructions) or (b) the response of tree rings to temperature is multi-valued, in other words a given tree ring response could have one of two quite different temperatures associated with it: high or low. In both cases an attempt at temperature reconstruction from this data is of no value.
Harold, Dr. Bouldin believes that if you find a good correlation on a subset of trees, then there is causation, thus these trees are a good proxy. The problem is that he looks at a subset of tree series and if you look at some other series, you may have a "divergence problem". The problem and bias is in the screeening for "good" proxies, but if other series of the same kind show to be "bad" proxies, one should be very wary of using that species as a temperature proxy.
About the use of detrended temperatures (and series) for the screening: you omit the trend "screening fallacy", but as shubb said, you introduce a high frequency "screening fallacy". It is not because there is a good correlation with temperature at interannual periods that there is a good correlation with the long term trend, one even can have opposite trends...
Take e.g. tree rings: these show a quite good correlation for year by year temperature (and precipitation) variations at any temperature level, but the longer term trend is an upside down U curve around an optimal temperature...
Tim Osborn should also jump in to criticize the withholding of data,since he included this recommendation: 'conventional scientific method of checking". I assume checking a paper would require access to all of its data.
climatebeagle,
I think Osborn's reference to 'conventional scientific method of checking' should be read to agree with Jones' vow to redefine the meaning of peer review if that was necessary to present his 'approved' version of science. I believe that this is what is considered 'conventional' in climate science.
There is definitely something rotten in research and science publishing, but it's not just in climate science. Check out this report on a study of cancer science published articles http://www.reuters.com/article/2012/03/28/us-science-cancer-idUSBRE82R12P20120328
The study attempted to reproduce the results of 53 landmark cancer science studies. 47 of them could not be reproduced. The money quote:
Part way through his project to reproduce promising studies, Begley met for breakfast at a cancer conference with the lead scientist of one of the problematic studies.
"We went through the paper line by line, figure by figure," said Begley. "I explained that we re-did their experiment 50 times and never got their result. He said they'd done it six times and got this result once, but put it in the paper because it made the best story. It's very disillusioning."
From the Ecclesiastical Uncle, an old retired bureaucrat in a field only remotely related to climate with minimal qualifications and only half a mind.
Bishop, I fear your ‘dead in the water’ question may be bad tactics,
A potted history will, I hope, provide a comprehendible basis for my concern.:
In an alter ego, you describe the creation of the IPCC in the opening pages of the HSI I have also found a fuller account at www.spiked-online.com/index.php/site/printable/3540/ .It seems that conferences prior to the IPCC’s establishment were captured by those who understood the potential that unpredictable changes in climate then presented. Risk was exaggerated and governments were urged to take expensive precautionary measures that included funding the advance of climate science. At the same time, the cold war was ending, the left was vanquished and the right found itself the uneasy victor, reliant on theories that, devoid of the old enemy, seemed insufficient to provide the solid theoretical foundation that would enable it to withstand attack. . So governments worldwide, all now (comparatively) on the right, were ready to adopt the precautionary policies advocated at the conferences- green policies. – because the public, then enjoying a cyclical peak in prosperity, were readily persuaded that the world might indeed be harmed by their unprecedentedly profligate consumption and that restraint would be appropriate. Adoption of these policies restored to governments the moral high ground hitherto provided by comparisons with the evident failings of the far left. The impetus was there and the IPCC was created.
In creating the IPCC, governments’ actual purpose was, of course, to promote current (green) policies rather than merely to advance climate science. Clearly, in those heady days, the existence of a gap between the two was not recognized. Governments were led to expect that new climate science would conclusively prove that green policies were prudent. Perhaps this explains why the actual purpose is not enshrined as the IPCC’s overtly stated aim. This, its website now states, is ‘to provide the world with a clear scientific view on the current state of knowledge in climate change and its potential environmental and socio-economic impacts’.. However, the actual purpose might be written ‘to provide the world with a clear scientific view on the current state of knowledge in climate change in order to facilitate minimization of iits undoubted adverse environmental and socio-economic impacts’. This is what the IPCC was actually created to do and what it has always done.
Warmists will say that in practice there is no difference between the two; skeptics that the gap is too wide to be bridged.
Prior to the issue of your update, it was easy for me to deplore raising the issue that AR5 might now be ‘dead in the water’ in the form of a question. Of course it would be ‘DitW’: the IPCC’s temporary bureaucracy, would have to ensure that AR% conforms with the organisation’s actual objectives in the same way as has always previously been the case, now notwithstanding interference from the IAC.. In addition, for me, this is not an isolated occurrence: I am often surprised that scholars on this and other blogs seem prepared to argue about details of IPCC publications without observing that this or that falsification is a necessary concession to the organisation’s actual purposes. And more, in the wider field of arguments between warmists and skeptics,, why is it not always pointed out that the ‘warmist would say that, wouldn’t he?.
Turning back to current concerns, since the issue of the update I cannot be so sure. If the de-trending legitimizes the paper (which I understand is questionable) then the paper moves into the common ground between the IPCCs actual and overt purposes and you.cannot be faulted for raising the issue as a question. I wonder if the matter will ever be resolved to the satisfaction of both warmists and skeptics: warmists have their reputation with the IPCC and their funders to defend and skeptics will bask in the bath of scientific rectitude.
Roll on, roll on!
Thanks Harold for the link. I read the Dr Bouldin thread, without being able to follow all the statistics, but half-persuaded by his point about high correlation within a particular subset. Then I thought about the garden, where the flowers are all blooming in one patch, while a few yards away the same species are struggling. Any measure of their bloomingness will clearly correlate with a lot of different but interrelated variables.
Now I must get back to the weeding.
Bishop I read your book so you should know the answer to this already ;-)
The Pearson correlation coefficient by construction is insensitive to linear trends.
What I said on Steve's blog:
Red noise is a problem and it does cause hockey sticks. Jeff ID has (to my purposes) satisfactorily demonstrated that data with red noise generates hockey sticks when you preselect using correlation. (In fact, you can recover any shape in the training period, including crazy functions like sinusoidal temperature functions.)
@ Judy's Gamesmanship thread I've claimed that with 1,760 PS3s I can write the Summary for Policymakers for AR7.
==============
Carrick, 'recover any shape' sounds like it could be a very useful tool. Now which end of it do I use?
===============
"@ Judy's Gamesmanship thread I've claimed that with 1,760 PS3s I can write the Summary for Policymakers for AR7." Kim
I don't need even one ps3, I can do it now, it will say the situation is worse than we thought, and we need more global governance ie coercion from the UN.
RKS on June 1, 2012 asks: "Has anyone heard what Richard Betts has to say about this?"
...and I repeat: "Has anyone?" because I have some faith in Richard and believe any comment here would have value.
Richard, can you spare us a dime... um... time?
Jun 1, 2012 at 10:36 AM | Mike Jackson Past adjustment rigor
Add in correction of moves from thermometers to more sophisticated devices and the difficulty of making the new instuments continue the old trend - then maybe you have it. You would indeed have it right if the error bands were properly calculated, for then essentially the whole thermometer record would be indistinguishable from "No change".
There seems no difference in kind between the methods used by Gergis et al and a drug company basing trials' results on successful outcomes while disregarding and suppressing those of patients who suffer adverse side-effects.
Am I missing some subtlety?
Jun 1, 2012 at 11:57 AM | Simon Anthony
...........................................
Yes. Death of patients.
I was rejected from a "blind" study for a new migraine treatment because I had a pacemaker and the designers told me that it would mess up their numbers if I died during the trial.
"This reads like a beginner's guide to the infiltration of scientific institutions such as the RS or the APS, not to mention UEA. Or am I going too far in my conspiracy theory?"
John in France
I think it's certainly true that groups, political or otherwise, who wish to influence events often try to get their people into positions of power. That way a very small number of people can subvert any sort of democracy when their lack of numbers (and inability to convince the masses of their cause) would otherwise fail to have any impact at all.
In many cases a body with a potentially powerful influence has a fairly complacent leadership and membership. Once a small number of activists have achieved influential positions they are able to marginalize any dissenters, browbeat the weak, ease the passage of more of their colleagues into positions of power, play the system and generally wear everyone else down because most of the people who don't share their views are less committed.
Of course, this doesn't mean that there is one centrally-controlled conspiracy operated from a hollowed out volcano, rather a number of loosely-affiliated groups and individuals (many of whom are relatively clueless cannon-fodder) with a more or less common goal which they see as a noble cause.
There are any number of bodies which spring to mind as having been co-opted in this way from the IPCC down to local council bodies.
IMHO:
The IPCC AR5 was already dead in the water. When they seemed to go and recruit a fairly large number of unknowns with dubious science qualifications from a number of Third World Countries. This time around.
Latimer:
My primary school teachers, some nice and others not, ALL insisted that I 'show my workings'. their colleagues, further along in my education in any subject, were equally insistent that I not only showed my workings, but also attributed my quotations. Gerjis appears to consider herself immune from the requirements of the academic conventions that confer credibility; the only conclusion I can reach about her work is that it is utterly without value and credibility.
They tell us that global temperature has increased by 0.8C since the end of the Little Ice Age.
Seems like it's in the right direction, why aren't they pleased about that?
Bish may want to further update this post.