Scrutinizing the consensus numbers

Whether scientists actually think humans cause global warming

Why study the consensus? Because in a world full of people trying to manipulate your beliefs, it’s the only way to tell what scientists really think.

Climate science is an enormous field with thousands of scientists publishing research, not all of which agrees. It’s complicated enough that anyone with an agenda can mislead, large enough that anyone with a bias can find a couple of papers supporting their position, and nuanced enough that it won’t be fully settled for many years. However, we have to decide now what to do about CO2, because it lingers in the atmosphere for centuries,¹ and our emissions have been rising exponentially.

You may have seen smaller numbers than 97% or even 90%. That’s because there are so many to numbers to choose from:

Source

Yes, Obama was wrong that “97% percent of scientists agree”.

There is no consensus among “scientists”. There is only a consensus among scientists that study Earth’s climate. With all the political messages flying around, you can’t really expect scientists with no knowledge of climate science to buy into AGW. Novice climatologists often don’t buy it, either.

There’s a metastudy reviewing the various studies of consensus. The consensus numbers steadily increase among those with more experience in climate science, reaching between 90% and 97% among experienced climatologists, depending on which study you choose to believe and who you call “experienced”.

Source: PBL Netherlands Environmental Assessment Agency based on Verheggen’s 2014 study.

To understand these studies, be aware that working scientists typically write (or co-write) one or more papers per year, so if an author has 10 “publications”, they probably have less than 10 years of experience. The most experienced authors have over 300 “publications”, and every study asked for number of publications, not years of experience.

I reviewed several individual studies personally. Before I discuss their findings, please note that there is no question that global warming is happening: almost no one actively working in climate science denies that.

• Up to 97.5% of the top 200 “climate researchers” agreed with the IPCC in 2010 that “Anthropogenic greenhouse gases (GHGs) have been responsible for ‘most’ of the ‘unequivocal’ warming of the Earth’s average global temperature over the second half of the 20th century” according to Anderegg 2010, but only 90% of 908 researchers who had their names on at least 20 papers appeared to agree. In Footnote [3] I suggest that on closer inspection, the consensus may only be 88% .
• 88% of 124 AMS climate scientists who mainly publish in climate science agreed in 2014 that roughly half or more of warming over the past 150 years was caused by human activities, including 78% who agreed that the cause of global warming over the past 150 years was “mostly human” (Stenhouse 2014). The other 12% had a variety of opinions, the most common being that humans have had some effect, but we need more evidence to determine how much. Agreement was lower for for climate scientists publishing mainly in ‘other areas’ (81%). In this study, notice that the question says “150 years”. Most climatologists agree that warming after 1950 was mostly caused by humans, but they also agree that less than half of warming before 1950 was caused by humans. In fact, the study says, “six respondents sent e-mails to notify us that their answers would have been different if we had asked about the most recent 50-yr time frame”.
• Excluding indeterminate answers (‘I don’t know’, ‘Unknown’, ‘Other’), 84% of survey respondents in Verheggen 2014 said that more than half of “global warming since the mid-20th century can be attributed to human-induced increases in atmospheric GHG concentrations”. This rose to 91% among the 25% (467 scientists) with the most “self-declared publications”. This survey was among the largest, with dozens of questions and 1868 respondents. 218 scientists from “signatories to public statements disapproving of mainstream climate science” were contacted and about 4% of all respondents were in this group. The study notes that “by also soliciting responses from signatories of public statements who are not necessarily publishing scientists, it is likely that viewpoints that run counter to the prevailing consensus are somewhat magnified in our results.” Thus 84% may be an underestimate of the consensus, but the study doesn’t say how much of the 4% did not publish in climate science. Individual survey results were kept anonymous. The amount of warming attributed to increased greenhouse gases was higher among experts in the physical science basis of climate change (WG1 group), especially self-described experts in clouds/aerosols and attribution of the causes of climate change (Figure 2). In my opinion, there’s a flaw in the survey’s design, due to which the indeterminate answers are difficult to interpret.
• 93% of “working PhD Earth scientists” surveyed in the American Association for the Advancement of Science (AAAS) in 2015 agree that “Earth is warming due mostly to human activity.” This 2015 Pew survey also asked scientists about other controversial issues like GMOs and nuclear energy, so you may find the results interesting for other reasons.
• 75 of 77 respondents to the Doran 2009 survey of mostly North American scientists “who listed climate science as their area of expertise and who also have published more than 50% of their recent peer-reviewed papers on the subject of climate change” agreed that “human activity is a significant contributing factor in changing mean global temperatures.” However, the total number in the category was 79. A reasonable interpretation of this data is that when asked “do you think that mean global temperatures have generally risen,” two said no and then didn’t answer the question about whether “human activity is a significant contributing factor” and therefore weren’t counted. This interpretation of the data suggests a 95% consensus rather than 97%.
• Of 11,944 climatology papers published between 1991 and 2011, 4014 expressed or implied an opinion about AGW. Of those, 97.1% of those papers implicitly or explicitly endorsed AGW, 1% expressed uncertainty about human causation, and 1.9% rejected or “minimized” AGW, implicitly or explicitly (Cook 2013). Also, endorsement of AGW has increased slightly over time. After criticism by Richard Tol, some abstracts were re-reviewed — producing a new count of 97.2%. When counting the authors of the papers instead of the papers themselves, the agreement was 98.4%.

This last study (Cook 2013) is probably the most publicized of all the consensus studies, so I think it’s worth discussing a notable criticism of it: that the numbers 97% and 98.4% are in large part based on comparing papers that implicitly agree with the consensus with those that implicitly disagree — any acknowledgement that humans cause any warming was counted as an endorsement of AGW, unless the paper uses language that “minimizes” that endorsement or offers an alternate explanation. The most famous contrarians like Spencer and Christy were counted in the 1.6%, but contrarians taking softer positions, such as Judith Curry, were in the 98.4%.[2]

Cook 2013 detailed results ; *author survey received 1200 responses (14% response rate)

At face value, Cook 2013 only offers weak support for the consensus. In fact, Cook 2013 found only 64 papers that “explicitly state that humans are the primary cause of recent global warming”, compared to 9 papers that “explicitly state that humans are causing less than half of global warming”. Some would say this means the consensus is 87% (others have claimed this means there is a 0.5% consensus by dividing 64 by the total number of papers. By that same logic, we could divide 9 by the total number of papers and say there is a 99.9% consensus.)

However, part of the Cook 2013 study was an author self-ratings survey: all authors were emailed to classify their own papers. After criticism about the “only 64 papers” thing, the SkepticalScience rebuttal responded with more data from author self-ratings, data that wasn’t mentioned in the original study (note that John Cook is the founder of SkepticalScience.) The data showed that authors were much more likely to classify their own papers as explicitly supporting the consensus than Cook’s team was. Although the author survey had only a 14% response rate (1200 papers), 224 papers were self-rated as explicitly endorsing the view that humans are the primary cause of global warming, with only 9 explicitly rejecting that view: that’s 96%.

So I would summarize Cook 2013 like this:

There are no “uncertain” classifications in the top-right category. It looks like Cook 2013 did not gather that information; “uncertain” papers and “unrelated to the topic of AGW” papers were put in the same category. This probably occurred because the Cook team did not initially realize the importance of distinguishing “uncertain” papers from “no position” papers. The 40 “uncertain” papers are actually just an estimate based on a late-stage review of 1000 “no position” papers, which found ~5 papers that suggested the cause of AGW was uncertain. Since there are 7970 “no position” papers in total, ~5 of 1000 papers was extrapolated to 40 of 7970.

Based on all the information, I conclude that the consensus for humans causing “at least half” of the warming since 1950, is between 84% and around 90% for all scientists who have published a peer-reviewed paper about climate science, and between 88% and 97% for experienced climate scientists who publish in that field as a career. [Edit: however, based on Footnote 3, plus the caveats I mentioned above on Cook 2013 and Doran 2009, I think the number 97% is too high as a quantity of climate scientists. 96% might be correct as a quantity of peer-reviewed papers, but that’s based on a survey with a low response rate. I think therefore that the issue deserves further study; however, the Cook 2013 study was done by a group of unpaid volunteers, so I would guess that not a lot of grant money is available to study the consensus.]

For me, Verheggen 2014 is the most interesting study because it asked the most questions. One of the things it found was that some scientists say human greenhouse gas emissions are responsible for more than 100% of warming. “more than 100%” means that if it weren’t for our GHG emissions, the Earth would have been in an overall cooling trend since the mid-20th century. In fact, this minority was larger than the minority saying “50% or less”. (Edit: actually, it is thought that much of this “cooling” is also caused by humans; when fossil fuels are burned, they produce non-GHG byproducts that can reflect sunlight and cause cooling. Notably, these “aerosols” are removed from the atmosphere fairly quickly, unlike CO2.)

Verheggen 2014 showed one of the weaker levels of consensus among the studies, but it also asked the scientists how confident they were of their answers. It turned out that the “under 50%” crowd tended to be less certain:

Source: PBL NEAA. Interestingly, the contrarians were bimodal: there’s an extra hump (top right), a small group (2% of the total) who are “virtually certain” GHGs don’t affect the climate much, unlike the majority who were less certain. Perhaps these “true believers” form the backbone of the denier community.

When contrarians were asked why they disagreed with the consensus position, they gave a variety of answers, but none of them broke the 50% mark, i.e. there is no consensus among contrarians.

More than one answer was allowed, and on average those disagreeing with the IPCC gave 2.45 reasons why. While “natural variability” was the most common answer, it is also pretty vague, as it includes numerous different phenomena. Source: PBL NEAA. Question 3c (Figure 6) in the original Verheggen study also suggests that contrarians do not agree on what, if not greenhouse gases, causes global warming.

This article was adapted from a larger one.

Footnotes

[1] Joos et al. (2013) gathered estimates from several different models on the question of how long it would take for a sudden addition of carbon dioxide (100GtC or 366 gigatons of CO2) to leave the atmosphere. They found that 25±9% will still be present in the atmosphere 1000 years later; the ocean will have absorbed 59±12% with the rest absorbed by land (16±14%). See graph.

[2] You can use the Consensus Project database associated with Cook 2013 to check other names. Note: the database only has last names and initials.

[3] I misunderstood Anderegg 2010 the first time and so the text has been corrected. I implied that 97% of 908 climate scientists who have their names on at least 20 papers supported the consensus, but in fact the number 97% is associated with only the top 100 scientists (and 97.5% with the top 200). The paper doesn’t give a percentage for the whole 908, but this number can be inferred by the size of the groups: 817 in the “convinced by the evidence” (CE) group and 93 in the “unconvinced by the evidence” (UE) group. This gives us a “90% consensus”.

The paper says three scientists were in both groups, which can’t affect the number 90% as long as we round it off, but it leads me to a point of concern: why didn’t Anderegg investigate what it meant for 3 people to be simultaneously convinced and unconvinced?

Anderegg’s largest group of “convinced” scientists were the 619 authors who helped write the IPCC WG1 report (the Physical Science Basis of Climate Change). He appears to have assumed that anyone who helped write the WG1 report must agree with its main conclusion, but the IPCC is an inclusive organization, and it seems plausible that a few scientists helped write the report but still disagreed with that conclusion. Perhaps the 3 unnamed researchers fit that description, and there could be additional dissenters. Verheggen 2014 noted that “four respondents tagged as AR4 WG1 authors chose the ‘26–50%’ option and, as such, disagreed with AR4’s attribution statement.” Presumably not every dissenter responded to the Verheggen survey; given the overall response rate of 29% we could extrapolate this to 14 dissenters out of 619, with some uncertainty. 14 dissenters would lower the Anderegg consensus to a little over 88% of the 908 experienced scientists.