Is the scientific confidence on climate change greater than 99% or less than 1%? And does the IPCC truly have confidence in its own conclusion? At first glance these questions may seem trivial and pointless. Even a bit embarrassing. Yet, upon closer examination, it turns out that only 0.6% of peer-reviewed scientific papers explicitly endorse the IPPC’s central position – namely, that there exists a consensus that human activities, especially by the emission of greenhouse gases, are the dominant cause of recent global warming. Yes, there is a general consensus that humans influence the climate, but only in an explicitly unquantified sense and probably rather small. And that is something quite different.
The IPCC deserves credit for indicating in most of its assessments, the degree of “likelyhood” of their statements and the degree “confidence” the author’s have in their own conclusions. However, those reported levels of likelyhood and confidence are notably low, and often fall below what might be considered appropriate for statements presented with scientific authority. It seems that for most of the author’s of the IPCC Assessment reports the science is not settled.
You probably don’t believe this right away. So please read the article below. It is largely adapted from the paragraph’s 1 and 3 of Chapter 3 of my book “Crisis or Hoax”. Published by Bookbaby (printed) and Amazon (e-book). An earlier version of this article was published on the Dutch website “Climategate”.
Consensus, likelyhood and confidence
1. A consensus of 97% or more?
On May 16, 2013, U.S. President Obama tweeted, “97% of scientists agree. Climate change is real, man-made and dangerous.”
The tweet became extremely important and may have been the most quoted tweet ever. His successor, Twitter fanatic Donald Trump did not even come close. At first glance, it seems an odd time for such a tweet. In May 2013, the average global temperature had barely risen for 14 years. But Obama wasn’t reacting to the weather or the climate either; he was reacting to an article by John Cook (et al.) that had appeared the previous day (!), on May 15, 2013, in the peer-revied journal Environmental Research Letters It was entitled “Quantifying the consensus on anthropogenic global warming in the scientific literature”. (J.Cook et al, 2013). The lead author was an assistant professor of communication sciences.
It is hard to imagine that Obama, or one of his closest aides, had little to do that day in May, for amusement picked up the previous day’s Environmental Research Letters and spontaneously decided that this hitherto totally unknown author deserved the president’s support. Obama had his own reasons. Nearly half of the American population at that point did not yet have a position on climate change and the scientists were divided among themselves. This could hurt both the negotiations on the new Paris Agreement, which Obama was eager to sign, and the chances of a Democratic candidate in the upcoming presidential election. Support for a study that seemed to show that virtually all scientists were in favor of the IPCC consensus therefore looked like an excellent chance, if chance it was. As of May 2013, no discussion of climate would be the same. Any skeptic or denier who still dared to open his or her mouth could be confronted and defeated with the 97% argument. The tweet became one of the foundational texts of the “consensus” on climate change.
In 2013 the climate alarmists were in desperate need of some scientific support. Many thousands of alarmistic peer reviewed articles appeared yearly in the scientific journals. The various IPCC reports claimed more and more self-confidence and called their own arguments more and more convincing. But it were the same arguments, and it just didn’t get any warmer. Despite the many billions for climate research (Weart, 2008), spent for the most part on alarmist research, no direct evidence for the greenhouse gas theory had yet been found. Many Americans did not believe in climate change. The alarmist world decided to focus on “consensus”.
Of course, they realized that consensus was not a scientific proof. Science advances primarily through the falsification of hypotheses, not through consensus. But they hoped to get away with it. If they could demonstrate that an overwhelming majority of scientists (say over 95%) would support their alarmist theories, who would dare challenge it? Who would dare risk the fate of the earth against the opinion of more than 95% of scientists?
They were largely proven right. The magic word consensus became a world hit and the term would become and remain highly influential. Consensus was a major factor in negotiating the Paris Agreement in 2015. Obama got the green light and was able to sign the treaty in 2015 and ratify it the same year. The IPCC paradigm was completely back, the alarmists had won.
How had it come to this? Of course, the concept of consensus was not invented in 2013. After getting rid of skeptical naysayers in its own circle between 1992 and 1996, the IPCC quickly promoted consensus in the scientific community. Skeptical voices no longer penetrated the rapports except by “mistake,” as in AR 3 in 2001. But the researchers knew about the criticism and the many Petitions and Open Letters, and it did not sit well with them. Their response was as unscientific as the concept of consensus itself, and as effective. They invaded and occupied the scientific institutions and the scientific journals.
Immediately following the release of the IPCC Assessment Report 3 in 2001, the National Academies of Science of the U.S., along with 17 more Academies of Science from all major countries (the Netherlands did not participate), would issue a Joint Statement in which they emphatically supported the tenets of the IPCC reports and discouraged criticism of them. The Joint Statement was printed in an editorial of Science, the journal of the American Association for the Advancement of Sciences. This publication would for skeptics deny access to most major journals on climate change. Some journals, such as Science, would clarify and publish their intended censorship, but most journals did it silently, through the instrument of “peer reviews”. (More about this in Chapter 4.)
How often would it have happened that so many Academies of Sciences consented to a relatively recent theory, which had never been proven nor experimentally tested, and de facto prohibited criticism of it? This was not science. It was politics. And it came from very high up.
The National Academies’ Joint Statement did not stop there. Declarations of support followed from many quarters. In 2005, the European Geosciences Union; in 2006, the Geological Society of America; in 2007, the International Union of Geodesy and Geophysics and the Royal Meteorological Society. In 2008, the European Federation of Geologists and in 2012, the American Meteorological Society. By 2007, all major scientific institutions – with the exception of the American Meteorological Society – had explicitly expressed support for the IPCC consensus. But the positions taken were the opinions of the “institutes”, the presidents and administrators. What about the members, the real scientists ? That was far less clear.
To see how many scientists actually supported the so-called “consensus” there have been a fairly large number of special studies. Important and often cited was the study by Naomi Oreskes, a highly respected professor at Harvard University and a true alarmist. She analyzed the Abstracts of 928 articles published between 1993 and 2003 relating to human activities warming the Earth’ surface. She found that: “Remarkably none of the papers disagreed with the consensus position”. Her study was repeated several times. Almost all of these studies indicated a scientific consensus of 85 to 100 % on the existence of human caused global warming. However, these results were also hotly contested. John Cook’s study from 2013 allegedly suggested a 97% consensus between the scientific papers and the IPCC consensus, while critics found, in the same paper, only a consensus of 0,6 %. This of course cries out for further explanation.
As said, by far the most cited and influential consensus study was that of John Cook (et al.) from 2013. They examined 11,944 articles written by 29,083 authors. Cook concluded that in the “abstracts” – the short summaries – a 97% consensus could be found on the existence of human-induced warming. He said that the number of articles that explicitly or implicitly rejected “anthropogenic global warming (AGW)” was a minuscule fraction of the published literature.
Besides the explicit support of many hundreds of authors and the implicit support of a large majority of climate scientists, press and politicians, it also received many criticisms. Sometimes only about the statistical processing of the data, sometimes much more fundamentally. Legates stated, “It is mind- boggling that any journal can publish an article that notes a 97% consensus on climate change, when from the author’s own figures it must be concluded that the true consensus is below 1%”. (Legates. 2015)
How is that possible?
In March 2012, John Cook and a small team searched the ISI Web of Science using the search terms “global warming” and “global climate change”. The period examined spanned from 1991 to 2011. The search yielded 12,465 articles of which 12,418 provided an abstract. Cook (et al.) worked only with abstracts.. Then the articles that were not peer-reviewed were deleted followed by those articles that were not about climate; despite the search terms. Over remained 11,944 articles written by 29,083 authors, in no less than 1980 different journals. Purpose of the study was “to determine the level of scientific consensus that human activity is very likely causing most of the current GW (anthropogenic global warming”.
Cook analyzed the abstracts and classified them via a questionnaire into seven categories, roughly from explicit and quantified support for the theory of human caused climate change to explicit rejection of that consensus.
Category 1. Inclusion in Category 1 followed if it could be inferred from the abstract that the authors explicitly agreed with the statement: “Humans are the dominant cause of recent global warming.” (IPCC AR 4, and later also AR 5 and AR 6) )
Category 2 Involved explicit support, without further quantification, for the theory of anthropogenic warming: “Humans are causing global warming“; but we don’t know how much.
Category 3 Implies humans are causing global warming, without explicitly stating humans are the cause.
Category 4 (By far the largest) involved abstracts that according to Cook had no position on warming.
Category 5 An implicit rejection of the theory of AGW .
Category 6 Was an explicit rejection without quantification.
Category 7 Was an explicit rejection with quantification.
In his article, Cook did not publish the underlying data of his research. But they can be looked up, with quite a bit of effort, in the various appendices.
Divided into the 7 categories, the results were as follows:
Category 1: 65 hits
Category 2: 934 hits
Category 3: 2,931 hits
Category 4: 8,261 hits
Categories 5, 6 and 7: had 53, 15 and 10 hits, respectively.
It must have been a very disappointing result for John Cook. Only the choice for Category 1 implies that humans are a dominant cause of global warming and supports the IPCC position on climate science. Only 65 papers could be found that supported his Category 1! That’s only 65 out of 11,944 and that’s 0.6%: “well below 1%”. (Legates, 2015). Not the 95% he had hoped for.
Cook’s study clearly falsified the “scientific consensus” and gave no support for the IPCC hypothesis as formulated in AR 4 and in AR 5 (published less than a year later but available in concept) . It was definitely not the result that politics (with Obama himself) and the National Academies had expected.
What to do?
Cook combined the categories 1, 2 and 3 and thus came up with 2,930 hits. He disregarded the 8,261 hits of category 4 (no opinion) and concluded that 97% of all abstracts supported the theory of human influence on climate change. Putting it this way leads to a very serious danger of misinformation. Cook changed the definitions. The 97 % consensus claimed by Cook is a consensus of all three categories combined and relates to the existence of “some unquantified human cause of global warming”. But this existence of “some” human effect on the climate is an open door, virtually all climate scientists, inclusive the skeptics, accept this. We know that a relatively large number of authors who consider themselves skeptics did not object to their articles being categorized as 2 or 3. Virtually all researchers, including most skeptics, are willing to acknowledge that humans have (some) influence on climate. But some influence is not a dominant cause as is required for Category 1 and is very different from being counted among the supporters of the IPCC consensus.
Did Cook explicitly say that there was a 97% consensus with the IPCC’s criteria, and was he therefore lying? Not really. He said, “We examined a large number of abstracts to establish the level of a consensus that human activities were very likely the cause of more than half of the current global warming”. He did not explicitly say how high that consensus with IPCC criteria actually was. It was only 0.6% and Cook knew that. But by his combination of categories 1, 2 and 3 and neglecting category 4, he managed to suggest a 97% consensus with the IPCC consensus.
By switching definitions Cook turned an open door (humans have some influence on climate) into a highly relevant but completely false suggestion of a consensus with the IPCC views and a supposed scientific consensus. It is incomprehensible that he got away with that so easily. How many people would have seriously read his article? How many had read his article and decided not to react?
By the way, how can we know the division into the categories 1, 2 and 3? It is not made easy. The article itself does not mention it. In the attached statistical analysis, the figures for categories 2 and 3 are given, but the data for category 1 are missing. You really have to go through all 12,000 articles cited in the Data File to find category 1 voters! I did. Indeed, there are only 65 ! (By the way, AI has made it much easier. With the right questions you can get all the data in a few minutes.)
Of interest is the large group of 8,261 respondents who did not answer the questions posed by Cook. What about them? Most of them will have shared the statement that the earth has warmed since 1850 and humans were partly to blame. But how many belonged to category 1? It seems fairest to divide the 8,261 non-voters proportionately among categories 1, 2 and 3. That will result in an additional 170 votes for the first category of Cook. It then becomes 235 voters (instead of 65) who support the IPCC consensus. Not 0.6% but about 2%. It’s still an almost negligible number! And we can’t make it any better.
We cannot but conclude that Cook’s study is hopelessly flawed in terms of its design, questions and further analysis. There was and there is no consensus ( certainly not of 97%) with the IPCC views as published in AR 4 (and later AR5 and AR6. It was much, much less than Cook indicated. But Cook’s study did become the most cited and influential publication on consensus in climate science. Legates article was ignored, he himself was sacked.
Still, I think Cook’s study is very relevant and very important. Cook shows clearly that in the period from 1991 to 2011, there was virtually no support in the scientific literature for his category 1: “Humans are the primary cause of recent global warming.” There was no support for the IPCC consensus!
Cook’s study was not the first and also not the last. Interestingly all authors came to roughly the same conclusions as Cook. And everyone made the same mistakes; they saw any warming as evidence for a dominant anthropogenic cause.
But what was the consensus in the post-2011 period? In October 2021, Environmental Research Letters published a follow-on article by another staunch alarmist, Mark Lynas. The title was “Greater than 99% consensus on human-caused climate change in the peer-reviewed scientific literature.” (M.Lynas , 2021).It replicates Cook’s (2013) research, but now for the period 2012 to 2021. Using virtually the same search terms as Cook, he finds 88,125 articles on “climate change.” Lynas rightly considers this an unworkable number and uses a representative sample of 3,000 abstracts.
It is a sympathetically written article. Unlike Cook, he is very open about the results he found. Also about the number of voters on category 1; there were only 19 out of 3000! Again about 0,6 %!
Lynas uses virtually the same 7 categories as Cook used in 2013 and gets the same results.
His category 1 is explicit support for the proposition that,
“Human actions are the primary cause of recent warming.” 19
Category 2 and 3 refer “to some human influence” 873
Category 4 without a position 2104
Category 5,6,7 human influence denied 4
Again, we see that the vast majority of abstracts, nearly 70%, make no statement about human impact on climate. Lynas suggests they should be considered as category 2. But that does not change the conclusion. Again, we see that the consensus-between the opinion as expressed in the abstracts and the IPCC’s conclusion (AR 4) is very low! Category 1 has only 19 hits out of a sample of almost 3,000. Almost the same as in 2013. The consensus with the IPCC conclusions is not 99.6% , but less than 0.6%!
Just like Cook did, they added categories 1, 2 and 3 together, ignored category 4 and concluded to a 99.6 % consensus. They had not learned anything, maybe did not even know Legates’ article, or seen the many skeptic Letters, Petitions and Declarations.
Lynas says, “Our analyses demonstrate a more than 99% agreement in the peer-reviewed scientific literature on the primary role of greenhouse gas emissions from human activities in driving modern climate change. on warming since 1850.” It is utter nonsense. Only 0.6% of the abstracts support a primary role for humans. I don’t think Lynas is lying; he is too forthright for that. He and his editor believe firmly in his analyses and results. But that doesn’t make it any better.
This means that over the last 30 years there has been no, I repeat no, consensus between the opinions of actively publishing climate scientists in peer-reviewed journals and the IPCC consensus.. On the contrary, the proportion of abstracts that explicitly supported the IPCC conclusions is between 0.6% and 2%. Of course this is highly sensitive and alarmists are very eager to forget it.
Figure 1. Only 19 out of 2,718 papers (0.6%) support the IPCC consensus.(M. Lynas, 2021)
The above mentioned conclusions were not formally challenged in the scientific literature, but of course there were many different voices.
A well-known author on consensus was James Powell. According to him, all, I repeat all, relevant publications from the first 7 months of 2019 (there were 11,602) would support the IPCC’s anthropogenic warming theory. (Powell, 2017, 2019). To me, this is incomprehensible, inexplicable and unacceptable. As mentioned earlier, there have been many thousands of scientists who emphasized in Petitions and Open Letters that they did not endorse the IPCC consensus. And there have always been publications by skeptics in peer-reviewed journals. Far fewer of course than the alarmist articles but certainly more than zero. Longhurst’s impressive book is full of critical statements from skeptical publications from peer-reviewed journals (Longhurst 2015) . The same goes for Don Easterbrook’s book, published in 2916. Many thousands of scientists signed skeptical Appeals and Declarations.
In September 2019, the Dutch foundation CLINTEL, on behalf of 500 international scientists, sent a highly skeptical letter to the UN Secretary-General and the Secretary of the UNFCCC, emphatically rejecting the IPCC consensus. The number of signatories has since reached more than 2000.
I myself receive articles, book reviews and other announcements almost daily, unsolicited, from alarmist as well as skeptical scientists. The number of skeptical books published in the last few years must run into the hundreds; the number of articles is even greater. Of course, I do know that the number of books and publications by alarmist scientists is very much larger. But that is not what matters. What matters is that the criticism of the IPCC consensus has received emphatic support from scientific circles, and it is not appropriate to further ignore those voices and speak of a 100%, a 99% or a 97% consensus. Articles that do so are highly biased and answer to political “overlords”.
We should not confuse a political consensus with a scientific consensus. Societal understanding -shaped by media narratives and political rhetoric – often simplifies or exaggerates the scientific findings, turning complex uncertainties into moral or ideological positions. In sharp contrast the real scientific support for the IPCC consensus expressed in articles published in the peer-reviewed journals over the last 30 years has been remarkably low.
2. Likelihood and confidence in the IPCC reports
Some six Assessment Reports and quite a few Special Reports have been published since the IPCC’s inception in 1988. The message therein, as presented by the Summaries for Policy Makers, has been largely consistent. But there has always been disagreement, also within the IPCC itself.
You can see it in the little confidence the IPCC seems to have in its own views. The IPCC almost never speaks of “proof.” Almost always it refers to “evidence”, to indications.
The IPCC indicates margins of uncertainty in many of its statements. It is often praised for this, rightly so in my view.
The IPCC distinguishes between “proof” and different degrees of “likelihood,” understood as the most objective statement possible, based on factual materials. In most of the scientific chapters “proof” is virtually non-existent . ‘Virtually certain’ has a probability of 99-100%. ‘Extremely likely’ is 95-100%. ‘Very likely’ is 90-100%. ‘Likely’ is 66-100%. Only a few statements are “extremely likely”. A majority of statements and propositions is likely or very likely.
This means that according to the IPCC itself, the majority of its views should be assigned a likelihood of over 66%, sometimes over 90%, rarely 95%, very rarely 99-100%.
It is very honest, yet also a surprisingly low percentage. In most social sciences, to make a scientifically sound statement, a likelihood of at least 95% is required. In the exact sciences it is quite a bit higher still, close to 99,99%. But in climate science, we work from a 66% probability and make decisions costing many thousands of billions on that basis.
In addition, the IPCC uses the term “confidence“; this is the degree of confidence the researchers have regarding their own conclusions being correct. The term was introduced in AR 4 and used increasingly frequently thereafter. The new AR6 only mentions the “confidence”. (Terje & Renn 2015).
“Very high confidence” means that in a research group, authors estimate that the chance of their being correct is at least 9 out of 10. High confidence is support from the group of about 8 out of 10 of being correct. Medium confidence is about 7 to 5 out of 10. Low confidence 4 to 3 out of 10. Very low confidence is less than 2 out of 10. A very high confidence is rare. Usually it is a high or medium confidence. Even low occurs regularly. Confidence rating can be a way of dealing with different opinions in one’s own ranks. A “medium confidence” can means that half the authors and reviewers disagree with each other. If confidence is “low,” the lead author has pushed his way through against most of the rest. In some controversial issues confidence can be “very low”. How is the level of confidence determined? Often it is simply done with a show of hands. It is to be expected that social peer pressure can play a major role. Confidence is often overrated.
In the AR 6 the IPCC points out that “confidence”, i.e. confidence of the research groups in their own conclusions, has risen sharply in recent years. In AR6 (2021/2022) it was much higher than it was in 2013 (AR 5). The IPCC gave no explanation for this, and it is still rather low. It falsifies any statement about a scientific consensus. For the scientists of the IPCC the “science is not settled”.
“Confidence” in IPCC Reports AR5 AR6
Very high confidence 6 6
High confidence 36 56
Medium confidence 37 32
Low confidence 20 6
Fig. 2 Confidence of IPCC in the Reports AR5 and AR6. ( Coren 2021)
The figure 2 shows that in the AR5 the confidence of the majority of the authors (57 %) was medium or low. This means that when the UNFCCC drafted the Paris Accords in 2015, it knew that its conclusions could not be based on confidence or consensus of the authors of the 2013 IPCC report. They did not care.
The AR 6 shows a higher confidence than the AR5 but the high (56) and very high (6) confidence is too low to justify the many claims of an unprecedented and catastrophic climate change.
We cannot but conclude that the IPCC promises to its “policymakers” a level of confidence that is remarkably low. Much lower than normally required for science based conclusions. We must also conclude that the policymakers did not care.
Despite lots of references to supporting literature, beautiful images and dozens of modeling studies, the IPCC reports are still reminiscent of R. Feynman’s description of “cargo cult-science“.(Caltech, 1974). Outwardly it looks a lot like science but actually it isn’t, the drive for finding and communicating the truth is lacking. I think a good number of scientists, including many alarmists, are aware of this. Skeptical researchers attach little importance to the abundant alarmist literature and detailed reports. Some see that abundance primarily as a smokescreen, a veil that hides the view of the big picture. There is a crisis of confidence; something that has been pointed out before. It seems that there has always been much more doubt about its own views within the IPPC itself than in the alarmistic outside world. Unfortunately, the IPCC has never clearly communicated that and the international politics (the UNFCCC) has not taken notice or did not care.
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.