Over the last couple of months the IPCC’s Fourth Assessment Report (published in February 2007) has been at the centre of a media storm. Revaluations about exaggerated or groundless claims have called into question the reputation of an organisation that has assumed a mantle of scientific invincibility during the last three years.
Alarmist predictions about the future of Himalayan glaciers, the Amazon rain forests, agricultural production in Africa, increasing devastation caused extreme weather events and rising sea levels have been shown to be based on evidence that at best is anything but robust and at worst is no more than hearsay. Worse still, it seems that the authors of the report were aware of the shortcomings of the evidence they were relying on but used it anyway.
Publication on the Internet of over a thousand emails from the Climatic Research Unit at the University of East Anglia, now known as Climategate, has added to disquiet about the IPCC’s activities. They suggest that Professor Phil Jones and many other leading climate scientists have attempted to subvert the accepted standards of their profession in order to protect their research findings from criticism. Many of those involved have been extremely influential within the IPCC process and the emails reveal an unhealthy culture of hostility towards anyone who questions the orthodox view of climate change that this organisation represents. It is questionable whether objective scientific research can take place under such circumstances.
The effect on the IPCC’s reputation, and that of its chairman Dr Rajendra Pachauri, has been devastating, but at every stage of this scandal we have been assured that the core science underpinning concern about anthropogenic climate change has remained unscathed. The IPCC and its supporters have been able to undertake this damage limitation exercise because attention so far has focused on only one of the three sections of the most recent assessment report: Working Group II: Impacts, Adaptation and Vulnerability. This deals with the symptoms and perceived consequences of climate change. The core scientific evidence that the climate is changing and that human influence is playing a part in this is contained in another section of the report, Working Group I: Climate Change 2007: the Physical Basis. But can we be confident that the same problems of sloppy authorship and exaggeration do not extend to this part of the IPCC’s assessment too?
On page 8 of the Working Group I: Summary for Policymakers there is a table (SPM.2) that has the following snappy title:
Recent trends, assessment of human influence on the trend and projections for extreme weather events for which there is an observed late-20th century trend.
http://www.ipcc.ch/pdf/assessment-report/ar4/wg1/ar4-wg1-spm.pdf
Extreme weather events are a particularly potent weapon in the battle to win hearts and minds for the crusade against climate change. Hurricanes (cyclones), droughts, floods and heat waves all feature as dramatic news stories regularly; they are the stuff of which editor’s dreams are made. A combination of dramatic pictures and human interest assures them a place in the headlines, and on the front pages, whenever there is an excuse for publishing them. From the point of view of those who wish to promote the idea of anthropogenic global warming such stories present a wonderful opportunity. If extreme weather events can be linked to catastrophic climate change in the public consciousness then the message that humans are influencing the climate, with appalling consequences, is going to be reinforced repeatedly because hardly a month goes by without some kind of weather related disaster being reported.
This is what Table SPM.2 says:
[A]
Phenomenon aand direction of trend
[B]
Likelihood that trend occurred in late 20th century (typically post 1960)
[C]
Likelihood of a human contribution to observed trend b
[D]
Likelihood of future trends based on projections for 21st century using SRES scenarios
[1] Warmer and fewer cold daysand nights over most landareas Very likely c
[>90%]
Likely e
>60%
Virtually certain e
>99%
[2] Warmer and more frequenthot days and nights overmost land areas Very likely d
>90%
Likely (nights) e
>60%
Virtually certain e
>99%
[3] Warm spells / heat waves.Frequency increases overmost land areas Likely
>60%
More likely than not f
>50%
Very likely
>90%
[4] Heavy precipitation events.Frequency (or proportion oftotal rainfall from heavy falls)increases over most areas Likely
>60%
More likely than not f
>50%
Very likely
>90%
[5] Area affected by droughtsincreases Likely in many regions
since 1970s
>60%
More likely than not
>50%
Likely
>60%
[6] Intense tropical cycloneactivity increases Likely in some regions
since 1970
>60%
More likely than not f
>50%
Likely
>60%
[7] Increased incidence ofextreme high sea level(excludes tsunamis) g Likely
>60%
More likely than not f, h
>50%
Likely I
>60%
For ease of reference I’ve made a few additions to this table in square brackets. The columns are now labelled A-D and the rows 1-7. I’ve also added numerical values for probabilities (‘likelihood’). These are explained briefly in a footnote on page 3 of the Summary for Policymakers, and in more detail in Box TS.1 in the Technical Summary of Working Group I as follows:
Likelihood Terminology Likelihood of the occurrence/ outcome Virtually certain > 99% probability Extremely likely > 95% probability Very likely > 90% probability Likely > 66% probability More likely than not > 50% probability About as likely as not 33 to 66% probability Unlikely < 33% probability Very unlikely < 10% probability Extremely unlikely < 5% probability Exceptionally unlikely < 1% probability
If we look at row [3] in Table SPM.2 for example, which deals with heat waves, we find the following:
Column A describes a phenomenon and a trend: heatwaves becoming more frequent.
Column B assesses the ‘likelihood’ that this trend has been confirmed by observation: is there empirical scientific evidence that heatwaves have become more frequent?
Column C introduces a hypothesis: if a trend in the frequency of heat waves has been observed then human activity is contributing to that trend. The likelihood of this being true is assessed.
Column D considers a prediction about the trend – increasing frequency of heat waves – continuing during the rest of the 21st century. This prediction is based on a range of scenarios set out in the SPM from Page 18 onwards. All of these envisage a world in which the fraction of Co2 in the atmosphere is growing as a result of human activity with a consequent rise in global temperatures exceeding any natural variation that can be expected.
In terms of the scientific method, we start with an assumption in Column [A]; heatwaves are increasing. This is then tested by observation in Column [B]: is it possible to detect an increase in the number of heatwaves, particularly during the last fifty years? A hypothesis follows in Column [C]: if it can be shown that there has been an increase in the number of heat waves, then anthropogenic warming has contributed to this. Finally, in Column [D], there is a prediction that depends on the preceding columns.
If we now look at the levels of ‘likelihood’ that the IPCC have assigned in each column we find something really quite remarkable. There is only a 60% – 89% chance that the frequency of heat waves have in fact increased; this leaves significant room for doubt that any such trend exists. There is even less confidence (50% – 59%) that, if an increase in heat waves has occurred – and we’re not sure it has – then human activity has something to do with it.
The conclusion that the IPCC draws from this is that, although there is a significant level of uncertainty as to whether the frequency of heat waves has increased during the last half century, and there is even more uncertainty as to whether, if the frequency has in fact increased, this can be attributed to human influence, a prediction can be made that heatwaves will increase during the next ninety years as a result of anthropogenic global warming. The ‘likelihood’ assigned to this is of 90-94%. Therefor according to the IPCC, confidence in the prediction is higher than confidence in either the observations or the hypothesis that the prediction is based on.
This makes no sense to me, but then I am not a scientist, let alone a climate scientist. It would be very interesting to hear the views of researchers from other disciplines, not on the merits of the scientific evidence, but as to whether this table does in fact defy logic.
Scanning the levels of ‘likelihood’ expressed in the rest of the table shows a remarkable consistency. The ‘likelihood’ expressed in Column [B] is higher than that in Column [C], but Column [D] is higher than in Column [C]. In some instances the assessment of ‘likelihood’ in Column [D] is higher than in either Columns [B] or [C] . So in each case the IPCC are saying that they have greater confidence in the predictions than in the hypothesis on which it is based, and in some cases that confidence in the prediction is even higher than in the observed trend.
The Working Group I Summary for Policymakers is intended to present a transparent and objective assessment of whether anthropogenic global warming is taking place, and do so in a way that is accessible to laypeople who would be unable to draw their own conclusions from the thousands of pages of scientific references in the main report. To a great extent this requires that policy makers must trust the scientists who write the IPCC’s reports, and they are likely to do so. We are repeatedly assured that the IPCC’s findings are based on the most carefully reviewed scientific research and therefore there is no room for argument about the facts. This point of view is epitomised by the oft-repeated mantras, ‘the science is settled’ and ‘you can’t argue with the science’.
The take-away message for anyone reading Table SPM.2 is that the IPCC attaches a high level of confidence to predictions that anthropogenic warming will cause extreme weather events to increase during the current century. They do not need to add that droughts, floods, hurricanes and heat waves cause loss of life and wreak economic havoc and misery on those who experience them. Regular news converge of such events leaves us in no doubt about this. Such events are the most visible and dramatic symptoms of our restless climate and we are well aware of their consequences.
The IPCC’s message to policy makers is clear: reduce greenhouse gas emissions now or more people will get hurt, and it will be your fault. So it is very important that those who are in a position to allocate billions of pounds, dollars or euros to fighting global warming should be fully aware of the methodology that the IPCC is using to reach this conclusion, and particularly just what confidence can be placed in their predictions.
The footnote on Page 3 of the Working Group I Summary for Policymakers that describes how confidence in scientific understanding and predicted outcomes is assessed says this:
In this Summary for Policymakers, the following terms have been used to indicate the assessed likelihood, using expert judgement, of an outcome or a result …
My emphasishttp://www.ipcc.ch/pdf/assessment-report/ar4/wg1/ar4-wg1-spm.pdf
Therefore the predictions of ‘likelihood’ in Table SPM.2 are not derived from any specific research contained in the assessment report but rely on the views of the authors of the report. So we are not talking about confidence levels that have been arrived at mathematically here, but the opinions of scientists. And reliance on their opinions is far reaching.
A minuscule footnote (f) to Table SPM.1 says:
Magnitude of anthropogenic contributions not assessed. Attribution for these phenomena based on expert judgement rather than formal attribution studies.
This refers to the predictions in Column [C] rows 3,4,6 and 7. It would appear that so far as these categories of extreme weather events are concerned there are no studies cited in the report that attribute human influence to the phenomena concerned, and the claims made in the table are based entirely on the ‘expert judgement’ of the authors.
Looking at the notes accompanying Table SPM.2 we find that it summarises the findings of Chapter 3 of the Working Group I section of the report: Observations: Surface and Atmospheric Climate Change . The coordinating lead authors of this chapter were Dr Kevin E Trenberth of National Center for Atmospheric Research (NCAR), and Professor Phil Jones, at that time director of he Climatic Research Unit (CRU) at the University of East Anglia.
The CRU emails leaked on the internet before Christmas made Professor Jones the unwilling star of what has been described as the biggest scientific scandal in living memory. He has now stepped down as director of the CRU pending inquiries into Climategate.
Dr Trenberth was also one of the correspondents featured in the Climategate emails and he has this to say (12/10/2009) in a message to Professor Michael Mann of Hockey Stick graph fame concerning the present decade long standstill in global warming;
Well I have my own article on where the heck is global warming? We are asking that here in Boulder … [Colorado, home to NCAR]
The fact is that we can’t account for the lack of warming at the moment and it is a
travesty that we can’t. The CERES data published in the August BAMS 09 supplement on 2008 shows there should be even more warming: but the data are surely wrong. Our observing system is inadequate.
This suggests that Dr Trenberth’ reaction to observed data that fail to match predictions is to reconsider the observations rather than question the skill of the predictions; something that I think most research scientists would find alarming.
(This message was copied to Stephen H Schneider , Myles Allen , Peter Stott , Phil Jones , Benjamin Santer , Tom Wigley , Thomas R Karl , Gavin Schmidt , James Hansen and Michael Oppenheimer. Although it is outside the scope of this post, it is worth noting that three of the recipients are involved in compiling global surface temperature records: Jones (CRU), Karl (NOAA), and Hansen (NASA GISS)).
In 2005 Dr Trenberth organised a major ‘media event’ at which he informed the world’s press that global warming was causing increased hurricane activity. This led to the resignation from the IPCC of hurricane expert Chris Landsea who was the contributing author responsible for the hurricanes section of Working Group I Chapter 3 of the Fourth Assessment Report. Prior to the press conference, Landsea warned Trenberth – who was not a hurricane expert – that there were no credible research findings that supported this view, but his warning was ignored. When Landsea complained to the governing council of the IPCC that making claims without sound scientific evidence would prejudice the organisation’s credibility their response was to defend Trenberth’s behaviour.
The extent of the scandal revealed by the CRU emails is beyond the scope of this post, but one quotation from a message written by Professor Jones would seem to be particularly relevant so far as the matter of ‘expert judgement’ is concerned:
As you know, I’m not political. If anything, I would like to see the climate change happen, so the science could be proved right, regardless of the consequences. This isn’t being political, it is being selfish.
Does this suggest a frame of mind in which one can reasonably expect a scientist to exercise his ‘expert judgement’ objectively? There is plenty of circumstantial evidence in the CRU emails suggesting that Jones may have behaved improperly both in connection with his own research and his contributions to the IPCC. Climategate, and Jones’ conduct, are now the subject of inquiries by both the University of East Anglia and The House of Commons Select Committee on Science and Technology.
As well as being the Coordinating Lead Authors on Working Group I Chapter 3, both Trenberth and Jones were Draft Contributing Authors for the Working Group 1 Summary for Policymakers. It is inconceivable that they would not have been involved in the creation and inclusion of Table SPM.2 in that summary.
One of the most worrying aspects of the flaws in the IPCC’s Working Group II report, which have already received so much publicity, is that they have only come to light because they were publicised by bloggers. As I said earlier in this post, I am not a scientist and there may be perfectly reasonable explanations for the apparent inconsistencies that Table SPM.2 reveals that I have missed. So far as I am aware there is no one involved with the IPCC who I can ask about this in the expectation of getting a fair and objective response. If the Working Group I report is to be scrutinised, then this too is only likely to happen if bloggers persistently ask awkward questions.
At the beginning of this week Professor Jones gave evidence before the House of Commons Science and Technology Committee. Andrew Orlowski of The Register ended a very perceptive report of the proceedings by concluding that the committee may have come to the conclusion ‘that rotten scientists perhaps mean rotten science’. At the time that the IPCC’s Fourth Assessment Report was published there was no one outside the sceptical bogosphere who was likely to entertain the idea that this vast and heavily hyped document might contain egregious errors. All that has changed now with the revelations concerning Working Group II. It is high time that the same detailed analysis is applied to the findings of Working Group I, the part of the report that is supposed to provide the real evidence for anthropogenic global warming. Furthermore, where the behaviour of scientists as revealed by Climategate is in doubt, their input to the IPCC reports should be scrutinised, otherwise the IPCC’s claims that none of the revelations so far have weakened the essential evidence of anthropogenic global warming will be credible. The address headers of the CRU emails read like a list of the IPCC hierarchy.
Here is one last word about the IPCC use of ‘expert judgement’ in assessing the probability of scientific understanding and likelihood’s. The United Nations Environment Programme is one of the parent bodies of the IPCC, with the World Meteorological Organisation (WMO). UNEP handled the publicity for the February 2007 press release of the Working Group I Summary for Polichymakers. This is the headline that they used:
Evidence of Human-caused Global Warming “Unequivocal”, says IPCC
The media worldwide dutifully ran headlines and stories that said exactly that, but the term ‘unequivocal’ only appears once in the Summary for Policymakers in the following context:
Warming of the climate system is unequivocal …
The supposed human contribution to this warming, which amount to significantly less than 1oC is assessed as ‘very likely’, which falls a long way short of ‘unequivocal’, so it appears that UNEP were prepared to ‘sex up’ this report from the outset.
In the IPCC’s Third Assesment Report, published in 2001, the probability that humans are contributing to climate change was assessed as ‘likely’. It would be very interesting to know whose ‘expert judgement’ led to the probability being increased to ‘very likely’ in 2007, and on what considerations this decision was based.
Lol- I noticed their arguments always boiled down to intimidation that science was squaling like a pig for deliverance.
Excellent article. Don’t apologise for not being a scientist. I’m not one either, which doen’t stop me arguing with scientists. (It’s a dirty job, but someone’s got to do it).
On your point about the apparent inconsistency of greater confidence in the forecast of long-term trends than in estimates of current trends or of human responsibility:
I imagine they’d argue that the methods for establishing levels of certainty are logically different. The percentages quoted for future trends are based on the estimated errors of the data fed into their models, and therefore can claim some statistical validity (as long as you accept the initial assumptions); estimates for the probability of human responsibility are little more than hunches, so they rightly give them a lower probability rating. Of course, the whole process is intellectually flawed, not only in terms of the science, but in more general epistemological terms. Apart from a few fields, like the prediction of eclipses, claiming to know with near certainty what will happen in a hundred years time is just not a rational activity. Normal human beings don’t do it.
Reading your article made me realise for the first time to what an extent the IPCC programme is totalitarian in its very nature, independently of the quality of the science. Midway between a five-year plan and a thousand-year Reich.
On the specific point about European heatwaves, I had a go at this in a comment at
http://www.climate-resistance.org/2010/02/ipcc-wgi-ch10-projecting-alarm.html
An analysis of the relevant chapter (IPCC AR4 WGII 8.2.1) shows yet another case of the IPCC announcing conclusions which are not supported by the evidence.
In normal probability you simply multiple the probability along a row to get the probability of the “accumulator bet” at the end.
If you need to roll a six then flip heads on a coin then the odds are
(1/6) * (1/2) = (1/12)
As you can see the probability gets smaller as you move to the right.
The IPCC uses some kind of “post-normal” thinking. Their probability percentages (90%, etc) are not the result of a statistical test. Instead they use the lookup table to turn phrases like “very likely” into numbers. The phrases are the output of “expert commentary”.
A generous interpretation of this words-into-numbers process is to help non-english speakers who often struggle with “escalating adjectives” like good < great < < fantastic < awesome.
A less generous view is fake science like an astrologer opening an office at CERN, wearing a lab coat and giving her horoscopes on a computer printout with a confidence level: “Jupiter is ascendant which means a 90% confidence that a dark stranger will …”
TonyN
Excellent report!
SPM 2007 Table SPM.2. (which you cite) is a disaster, as you point out.
An “observed trend” with “>66% likelihood” of having actually occurred and “>50% likelihood” of having had a “human contribution” of some undefined magnitude (not based on “formal attribution studies”, but rather on “expert opinion”) provide the support for predictions of “future trends” with “>90% likelihood”
This is not “science”. It is pure propaganda.
Using Jack Hughes’ statistical approach, we would have had (assuming the “human contribution” represents a high >50%):
Likelihood of recent human-caused phenomenon = >66% * >50% * >50% = >16.5%
Likelihood of predicted future human-caused phenomenon = somewhere below 16.5% (due to uncertainties always inherent in predicting the future), NOT >90%
But this is just the tip of the iceberg.
SPM 2007 and AR4 WG1 need to be independently audited in detail, to highlight all the other instances of this sort of rubbish.
The audit should start with SPM 2007 and the most suspect chapters of AR4 WG1:
Chapter 9: Understanding and Attributing Climate Change
Chapter 6: Paleoclimate
Chapter 8: Climate Models and their Evaluation
We were bamboozled into believing that IPCC represented the “gold standard” in climate research, yet the latest revelations are showing us just how flawed the IPCC process, scientific bases and conclusions really are.
Max
TonyN
Thanks for this noble effort. I think Geoff has identified why there is a discrepancy in the columns.
The columns are not derived from the same type of information so the probabilities are not multiplicative.
Columns B and C are based on observations. That means real recorded data have been used for statistical analysis. So for example increased frequency of precipitation events since 1960 can only be detected in the records with a probability of 60%. This in itself is daft because in such a short period some records will undoubtedly show an increase and some a decrease. With cyclic phenomena such as the PDO, climate over 40 years will show localized trends.
Column C is interesting because they are saying that it is really a 50:50 proposition whether man is contributing to the 60% likelihood of increased precipitation in Column B. In other words they don’t know one way or another whether AGW is a factor. The “greater than” 50% only just tips it towards “more likely than not”.
So now we get to Column D where all the high probabilities are contained. The results in Column D do not use the Column B and C numbers derived from real data. Column C is entirely based on the future climate modelling by simulating a series of SRES scenarios. Because they have so many model runs they can assign probabilities to the outcome based on the model results. The models create what is known as a “spaghetti graph” which shows all the traces of the model results and it is easy to do the arithmetic to determine probabilities of different outcomes.
But as Geoff points out, the models are only correct as long as you accept the assumptions and the way the models are formulated. So for precipitation intensity, Table SPM.2 is entirely hanging on the validity of the models. The recorded data do not provide conclusive evidence for increased floods and droughts so the WGI findings on “climate catastrophe” are solely based on the modelling. And as I have pointed out on another thread, the models do a very poor job of replicating the current climate let alone 100 years into the future. Furthermore if you read the details in the modelling chapters in IPCC 2007 you will find that the models do not claim much “skill” in modelling precipitation anyway!
A usefully detailed and timely reminder, Tony – thanks.
I’ve wondered for years why the much cited IPCC report and, in particular WG I, Table SPM 1, doesn’t get far more attention. It was reading it over two years ago (I’ve just checked) that turned my early misgivings about the AGW scare into serious scepticism. To my mind, the fundamental question is this: to what extent did human CO2 emissions contribute to 20th century warming? To find out, I turned to the IPCC report (not to some sceptics’ blog) and immediately I discovered that even the “Summary for Policymakers” is not as clear as is so often claimed. It says, “Most of the observed increase in global average temperatures since the mid-20th century is very likely due to the observed increase in anthropogenic GHG concentrations”. Hmm – only “since the mid-20th century”. But what about the 1910-1940 warm period (the one Phil Jones now says was “not statistically different” from the recent warm period)? Moreover, what does “most” mean? It’s not defined, and could mean anything between 51% and 99%. Sloppy science – in my view. Nonetheless, our political leaders tell us that the Western world’s already weakened economy must be further undermined to “address” this sloppily defined “problem”.
Ah no, we’re told – by people like Peter Martin – the proper scientific references underpinning the “problem” are to be found in Working Group I. So, to see what that said about human causation, I looked at it. And very quickly found the extraordinary Table to which you refer. What it said was an eye opener – and that was going no further than your column [C]. For example, re “the frequency of warm spells / heatwaves over land areas”, we’re told that a “human contribution” is “more likely than not”. Note that: contribution, not causation – a contribution could be no more than 1%. Even worse (as you point out), a footnote (f) notes that, “Magnitude of anthropogenic contributions not assessed. Attribution for these phenomena based on expert judgement rather than formal attribution studies”.
So, on the crucial matter of human causation, it seems that Peter Martin’s “mainstream science” amounts to little more than vague guesswork. And even that only says that there is a better than 50% chance that humans might have “contributed”. As a guide to policy, it’s worthless.
Yet Al Gore still insists (NYT 27 February) that, because of AGW, the world faces an “unimaginable calamity requiring large-scale, preventive measures to protect human civilization as we know it.”
Geoff: I take your point that the probabilities in the three operative columns are probably derived differently, but they are presented as though they are contingent on each other. This appeared in a summary of the main report, and the message to policy makers is a clear one: we know what we are talking about and the chances of catastrophe are very high. Yet Table SPM.2 is not in fact summarising anything but expressing varying levels of confidence in the opinion of the authors.
Jack: I think that your reference to post normal science is a shrewd one. Let’s not loose sight of the fact that no claim is made in this table that any of the probabilities are derived from statistical analysis. All are based solely on expert judgement. Indeed so far as Column C is concerned there is an admission that this is not even based on attribution studies.
Max: I would agree with what you say. The thinking that has led to this table being included in the SPM has more to do with propaganda, or persuasion, than with scientific research, which is meant to be what the IPCC’s reports are restricted to.
Potentilla: If the probabilities are based on expert judgement then there can be no question of their being multiplicative, as you say, which is why I didn’t consider the implications of this being the case when I wrote the post. On the other hand, even if the methodology is different in Columns B-D – and therefore I am being unfair to the authors in what I said – this is no way to use science to present information to policymakers. A question remains about how valid Column D can be if there is greater uncertainty in the preceding columns. This table is clearly intended to be read from left to right with the logical payload being cumulative and Column D carrying the payload.
Here is another take on expert judgement:
A week ago, Beddington, Slingo and Watson appeared before the HoC Select Committee on Scinience and Technology in the role of experts on climate science.
Early on in the proceedings they were asked if we could still be confident in the scientific evidence for AGW. All were unequivocal in their opinion that we could, with only Slingo admitting to there being some uncertainties which had no implications for the overall conclusions.
Later, Beddington was asked whether Climategate has being damaging to climate science in general and the reputation of UK science in particular. He explained at some length that this was not the case and the other two endorsed his opinion.
In both cases the scientists were conveying to the committee their expert judgement. None of them seem to have wondered why they were sitting in a HoC committee room being grilled (very gently) by members of parliament if Climategate is not a gigantic and very damaging scandal.
Presumably, the members of the committee were prepared to rely on these experts’ opinions about the validity of the evidence for AGW, but why should they do so when their judgement about the impact of Climategate was so clearly at fault? It would be comforting to think that the committee noticed this and wondered if self-deception could extend to the experts’ response to both questions or whether their expert judgement could be trusted on the first question if it clearly could not be trusted on both.
What is required is a means by which expert judgement on AGW, and matters relating to it, can be subjected to forensic examination. At the moment there seems to be no way in which this can happen.
Tony N:
Your spotting of the break in the chain of logic betwen the IPCC summary for policy makers and the UNEP press release is a major discovery. The whole IPCC process is riddled with similar breaks, in my opinion.
On the IPCC table: your observations about how it’s meant to be read are spot on. It’s the presentation that counts, because the whole report is a strategy document. We were all too overawed by the claim of 100% peer -review to look closely.
When the chain of evidence was revealed to be: Indian scientist’s phonecall > Fred Pearce in New Scientist > WWF > IPCC, Donna Laframboise simply looked up WWF and found a dozen other dodgy dossiers. Then someone spotted the Canadian tourist brochure, the advice on cleaning your snowboots, etc, and the whole house of cards started to collapse.
Of course, that’s not how the media or the public see it – yet.
On the CRU hearings: you wonder whether the committee members noticed the inconsistency of relying on the experts’ opinions about the validity of the evidence for AGW, when their judgement about the impact of Climategate was so clearly at fault. There was some interesting body language at the end of the hearings, when the beardy who had given Lawson and Peiser such a hard time was exchanging glances with Stringer, who had so successfully shown up Jones.
Beddington, Slingo and Watson clearly hadn’t seen the previous witnesses, and couldn’t know what a pitiful showing Jones and his Vice Chancellor minder had made, so obviously didn’t realise how out of place their breezy confidence seemed.
TonyN:
What amazes me is that this analysis has been around for a long time. You may recall that, in April 2008, you and I both had posts about it on CA. And I know I’ve posted the burden of my #6 elsewhere – indeed, I cut/pasted much of it from my original notes, now two years old.
It goes, I suggest, right to the heart of the catastrophic AGW hypothesis. Yet no one seems interested – even the CA thread is no longer available.
Some of you may recall that almost exactly two years ago I set up a thread on the (sadly now defunct) Climate Audit Message Board called
“Errors, distortion and exaggeration in IPCC AR4”.
Tony, Robin and especially Max were major contributors.
I have the entire original thread if anyone wants it.
I have distilled my interpretation of the contents here, which I have more recently updated to include the widely publicized examples from WGII. It is curious that the media have not yet caught on to all the bias in WGI. As Tony says, it is ‘high time’ to reconsider this, in the light of the climategate emails, and now that the media are at last beginning to listen. It is notable that some of the most blatant false claims are in chapter 3, written by Jones and Trenberth, such as
– the false claim of warming since the TAR (2001)
– the false comparison of short trends with longer ones
– the false statement that temperature rose ‘more stongly’ late in the 20thC than earlier
– the false claim of increased warming in the last 25 years.
Tony mentioned his ‘more confidence in predictions than observations’ point about table SPM2 in that thread. I confess I don’t entirely follow Tony’s logic (though it is clearer than the muddled logic of the IPCC). From their ‘We experts are sure that disaster is just around the corner’ perspective, it is possible for column D to be higher than columns B or C.
There is an item in my list about the Landsea resignation over hurricane exaggeration, and the UN’s misquotation of its own report. There is an email at the end of the list for comments.
PaulM
Thanks for you summary of the errors, distortion and exaggeration in IPCC AR4.
Yes. I would be interested in the original CA thread (have looked for it a few times, but it has disappeared).
Regards,
Max
PaulM
Tony’s reference to Table SPM.2. as an example of fuzzy logic in the IPCC presentation is valid (see my post #4 for the reasoning).
It basically boils down to assigning a much higher probability to future predictions than to the past physical observations upon which these predictions were based.
Max
potentilla
To IPCC Table SPM.2. you logically state that the “likelihood of future trend” is based on model simulations (using the SRES scenarios) rather than directly on “likelihood that trend occurred in late 20th century” plus “likelihood of a human contribution to observed change”.
But the logic here is fuzzy, because the “model simulations” have to be based on observed past trends to be meaningful. If they have no observational basis they are worthless.
Therefore, giving a future projection (with all the unknowns the future holds in any case) a significantly higher level of likelihood than a past observation or past causation upon which this future projection is directly (or indirectly) based is not logical (and Tony’s plus Robin’s conclusion is right).
Max
TonyN:
Table SPM.2 is definitely misleading, and I suspect deliberately so. You rightly say:
It comes back to the total reliance of the AGW proposition being based on climate models. Even if you believe the models are properly formulated and calibrated (which they are not) there is a huge disconnect, even within the table, between the observed climate (Columns B and C) and the model projections (Column D). Despite increases in carbon in the atmosphere in the past 10 years, there has been no increase in average global temperature and no observable increase in catastrophic events. (Not counting earthquakes!)
One conclusion you could draw from this table is that the Climategate issue is largely irrelevant. The IPCC is drawing its major conclusions on the basis of the climate models which have very little to do with CRU. So we are really being asked to believe in the climate models as a basis for significant changes to our economies. And based on the evidence so far, that belief is being extremely tested.
You are right to draw attention to the conclusions of WG1 for they seem to be as flakey as the conclusions of WG2.
Manacker:
We are all in agreement (perhaps disturbingly so) regarding the lack of logic in the table. My #5 was an attempt to clarify the sources of infomation in the table. You say:
My understanding is that the models are set up to replicate past increases in global mean temperature. Nothing more nothing less. They are not calibrated for temperatures and precipitation at specific locations nor do they simulate past floods and droughts. Though complex in their formulation and ambitious, in that they simulate the entire globe, they are very simplistic with regard to the complexites of real climate and ocean processes. So simplistic that the conclusions regarding climate catastrophe are based on conjecture from the model results rather than directly from the models. The models claim to project future climate average values. They do not claim (even in the IPCC reports) to model extreme events.
Robin and Geoff:
When I originally put a comment about Table SPM.2 on the CA Message Board, rather than on my own bog, it was because I was by no means sure of my ground and I wanted to test reaction in a forum where science was the normal topic. There were quite a few people, as I remember, who seemed unable to see a justification for what the IPCC had done but they were unwilling to accept that I was likely to be right; presumably because of the assumed infallibility of the assessments at that time. Himalayagate, Pachaurigate and all the other ‘gates’ that have now become attached to the IPCC process seem to have dented that image considerably.
Perhaps people will now be prepared to consider that WGI contains egregious errors that would make it a laughing-stock if the subject were anything other than AGW. That is why I brought the matter up again now, and as I wrote it a few other pieces that have come to light since then fell into place too.
The problem is that it took me nearly 3000 words to explain what the trouble might be, and even then it is necessary for the reader to have at least a rudimentary knowledge of the scientific method and statistics. How does one package such a story for the Sunday Times, let alone the Daily Mail?
potentilla
We are in agreement.
In fact, it goes a step further.
If, as you suppose (and I must agree) the model simulations (based on the cited “SRES scenarios”) simply project “future temperatures”, but are not able to carry this to the next step of projecting “future extreme weather events”) then Table SPM.2. is even more of a fraud.
Compounding all this is the fact that the two most severe “SRES scenarios” [A2 and A1F1] invoke atmospheric CO2 levels that are physically impossible to ever reach from human CO2 emissions, because there are not enough optimistically estimated fossil fuel reserves on our planet to generate that much CO2.
Scenarios B2 and A1B involve annual increases in atmospheric CO2 (CAGR – %) that are almost twice what we have seen over the past 50 (or 20) years, so these are also unrealistic.
This leaves use with scenarios B1 and A1T, which have temperature increasing by 1.5 to 2.1°C higher than today by year 2100 (see my post 9841 to PeterM on the NS thread), even with all the exaggerated assumptions on “positive feedbacks”.
It is hard to see how very much “extreme weather” will “very likely” result from such a small projected increase in temperature.
Max
PaulM
Tony mentioned his ‘more confidence in predictions than observations’ point about table SPM2 in that thread. I confess I don’t entirely follow Tony’s logic (though it is clearer than the muddled logic of the IPCC). From their ‘We experts are sure that disaster is just around the corner’ perspective, it is possible for column D to be higher than columns B or C.
That is precisely the perspective that I am questioning; expert judgement dressed up to sound and look like the product of attribution studies and an impartial assessment process.
I agree that whoever determined the probabilities in that table might have done so on the basis that Column D was independent of the other columns. But surely it must have occurred to them that the result, when considered in context, looks pretty odd. And surely the models have some foundation in observation, if any are available, rather than just creating a virtual world based entirely on theoretical processes.
I seriously think that the time has come for the questions you gathered together in the CA MB tread to be explored further. As I remember it, not all the comments that turned up were winners, and some were pretty contentious, but among them were ones that posed perfectly valid and very interesting questions. I’ll have a closer look at the listing on your site at the end of the week.
Incidentally, on Friday a BBC reporter mentioned that an announcement from the IPCC was expected later in the day about their inquiry, but that he understood that instead of being a review of the organisation’s past performance it was now going to concern itself only with revising procedures prior to the start of preparations for AR5. Has anyone heard any more, and why the unexpected delay I wonder?
Manacker:
That is correct except the models also predict future average rainfalls. Because of my professional interest and responsibility, I went through the IPCC reports very carefully to determine the origin of the projected increase in extreme rainfalls. As far as I could determine the projection is based on the models only to the extent that the modes predict higher average temperatures and average rainfalls. An atmosphere with a higher temperature can hold more water vapour so it is thus conjectured that intense rainfall events will be more frequent.
So it is that same old game again with a questionable chain of logic and sleight of hand. A casual reader would infer that the models show increased rainfall intensity so if the models are right then the projection is soundly based. But the projection of more intense rainfall events is conjecture based on the model results, and is not in the actual model results. It is rather like Thompson and his conjecture on the impact of loss of the Himalayan glaciers. The conjecture was buried in the paper and was based on the research but the research did not address the impact directly.
Here’s a special task for the non-scientists posting. Is there a word for this type of logical fallacy? I see there is a book by Engel on the Chain of Logic but not much available on his writings on the web – unless Geoff can take a break from snow shovelling and apply his French to useful effect.
[…] I’m interested to hear your thoughts on it. And so is the author.Read the complete article: Harmless SkyNo related […]
Nice, an excellent piece of work. Thorough, balanced and insightful. My only criticism is here:
“The CRU emails leaked on the internet before Christmas”
Please, don’t stoop to using carbon-alarmist type spin. Whenever I read about this on Pro AGW websites the emails were always hacked and I have yet to see any evidence to support either claim.
Might I humbly suggest that the emails appeared on the internet before Christmas. Apart from that I can only admire your analysis.
Excellent post. A “keeper” as I call these types of articles. They come in handy in debates with alarmists.
I think I can guess what the alarmist response would be. If I may play devils advocate for a moment…
Imagine you have bet on the outcome of a coin toss. You bet on heads. The chances that heads will be the outcome is 50%. If heads indeed is the outcome, your chances of winning is 100%.
So the key is how the question is framed.
If you’re asked “what are your chances of winning?” The answer is 50% If the question is “what are your chances of winning SHOULD heads be the outcome?” the answer is 100%.
So then, which is the “correct” way to ask?
In the context of CAGW, I guess I can answer that with another (extreme) example.
What are the chances earth will be hit by a large asteroid in the foreseeable future? One in many millions I guess.
What are the chances we will be wiped out if such an event takes place? Probably 100%
So is it more relevent that we ask “what are the chances we will be wiped out by an asteroid in the foreseeable future?” or “what are the chances we will be wiped out “should” an asteroid strike?”
For policy makers, it can only be the former.
So you are correct (even though you’re not a scientist lol) In the table provided, row 1 column D should read >53% not 99% (99% of 60% of 90% = 53%)
I hope I made sense.
TonyN:
At #16 you said “it took me nearly 3000 words to explain what the trouble might be… How does one package such a story for the Sunday Times, let alone the Daily Mail?” I saw that as a challenge and produced the following 470 word overview:
Comments?
Max: you say that SPM 2 is “a fraud”. Well, it may be hopelessly misleading about future events. But I trust it’s not a complete fraud – after all, a careful reading (see above) shows that the IPCC does not really think mankind is responsible for climate change.
This post is now being discussed at http://www.climategate.com and http://www.reddit.com, so the word is beginning to spread.
Can I ask anyone who is commenting on other blogs to consider giving it a mention in the hope that we can get a discussion going in other parts of the blogosphere. With luck the MSM will eventually take an interest.
Robin:
You summary is good but might benefit from a little more elaboration on the future projections. It is clear from Tony’s analysis that the IPCC put more weight on the model results than on observations. To most logical people this is bizarre but nevertheless this is what they have done. I think it it because it is very difficult to “tease out” significant trends in extreme events from the data and even more difficult to assess the human contribution. Models, however, can be made to provide nicely packaged answers and the results can be analysed to provide beautiful statistics.
The overwhelming belief in the validity of the models is exemplified in a comment by Vicki Pope of the UK Met Office some months ago. She was commenting on the series of relatively cold years since 1998 and the fact that the models did not predict them. She said that in many ways we know more about the climate 100 years from now than we know about the climate next year. To anyone who has done modelling such a statement is incongruous. It sort of makes sense though if you keep in mind that all they are modelling is 30-year averages of temperature and precipitation. So she is saying we know more about the 30 year averages , 100 years from now than we know about the climate next year. This really demonstrates the limitations of the climate modelling, particularly for extreme events and it is disturbing that so much reliance is placed on them.