Are Scientists and Journalists Conspiring to Retract Papers?

Dr. Roger Pielke, Jr has published two posts recently having to do with extreme weather and the IPCC. In the first post, he expresses frustration over an apparently inevitable retraction of a paper on extreme weather. The paper is Alimonti et al 2022 (A22) published in The European Physical Journal Plus (EPJP)[1]. I became aware of this paper about a year ago; it was not a good paper and was not particularly influential; it received almost no attention except for a brief period of time when the usual blogs promoted it, and then after SkyNews in Australia publicized this study as demonstrating there is no climate emergency. In the second post, Pielke summarized what he believes the IPCC's position is on extreme weather, and here he refers again to this previous paper. He declares that he believes that popular and scientific treatments of extreme weather have become far mor extreme than the position taken by the IPCC. According to him, "with the exception perhaps of only extreme heat, the IPCC is badly out of step with today’s apocalyptic zeitgeist." He then speculates:

It may also help to explain why a recent paper that arrives at conclusions perfectly consistent with the IPCC is now being retracted with no claims of error or misconduct.

There's a lot going on here in this quote and in these posts. One the one hand, I think he's far too generous to this paper and far too critical of the scientists that he believes are colluding with journalists to get this paper retracted. I think it tells a story about how ideologies and tribal allegiances are shaping public discourse on climate change - that is, how we treat those we believe to be our "allies" and how we treat those who we believe to be our "enemies" in the climate debate. The paper itself (I believe) is largely trivial, whether it ends up being retracted or not. What it reveals about our allegiances in public discourse on climate change is far more important. I want to cover this as if answering three questions. First, is A22 actually consistent with the IPCC? Second, should A22 be retracted? And third, do we have reason to believe scientists and journalists are conspiring to get little-known papers retracted? Then I'd like to consider how it is that our ideologies and tribal allegiances may be disrupting public discourse.

Is A22 "Perfectly Consistent" with the IPCC?

The IPCC is unequivocal that AGW is already having an impact on extreme events. They consider it "established fact" that some weather extremes are becoming more frequent and/or more severe. Consider just the summary position articulated by the IPCC from AR6:

It is an established fact that human-induced greenhouse gas emissions have led to an increased frequency and/or intensity of some weather and climate extremes since pre-industrial time, in particular for temperature extremes. Evidence of observed changes in extremes and their attribution to human influence (including greenhouse gas and aerosol emissions and land-use changes) has strengthened since AR5, in particular for extreme precipitation, droughts, tropical cyclones and compound extremes (including dry/hot events and fire weather). Some recent hot extreme events would have been extremely unlikely to occur without human influence on the climate system

Admittedly, A22 had only consulted a draft version of AR6, but Pielke has had access to AR6 and claims to be very aware of its claims on extreme weather at the time of this blogpost. But the contrast between the position of AR6 and A22 is pretty obvious. The basic position of A22 is stated in the abstract. After acknowledging that there are "robust changes" in yearly values of heatwaves with insignificant trends in "heatwave intensity," they write that with respect to these other response indicators:

None of these response indicators show a clear positive trend of extreme events. In conclusion on the basis of observational data, the climate crisis that, according to many sources, we are experiencing today, is not evident yet.

These two positions are difficult to reconcile with each other. I suppose you could make the point that AR6 doesn't use the word "emergency," but let's be honest; these two positions are not really compatible. The conclusion of A22 makes the disparity even more clear. Here, A22 clearly claims 1) there is "deep uncertainty" about how climate will "play out" in the twenty-first century, 2) we need to prioritize adaptation over mitigation ("We need to increase our resiliency to whatever the future climate will present us"), and 3) our current climate situation is simply in a "warm phase." In contrast, the IPCC concludes that current warming is not a "warm phase" but in the midst of a persistent trend caused entirely by human activity (greenhouse gas emissions partially masked by aerosol pollution). This position is not reconcilable with the IPCC's position. 

A22 is essentially arguing that global climate is simply going through a "warm phase" and there's no way to know how the rest of this century will "play out." So far, they say, most extreme weather events have "no clear positive trend," with the exception of heatwave values unrelated to intensity. So we should just adapt to whatever future changes happen with little need for mitigation. We may possibly detect a nod towards the benefits of mitigation when they acknowledge we should "minimize our impact on the planet and to minimize air and water pollution," but this also consistent with saying we should focus on environmental policies not that do have carbon emission reductions as a goal.

In fairness, the issue may have more to do with Pielke's reading of the IPCC than this paper. In fact, Pielke seems to be nearly alone in his assessment of the IPCC's position on extreme weather. A22 explicitly disagrees with the IPCC, and it's under investigation and likely to be retracted. As we'll see later, journalists have asked climate scientists to evaluate this paper, and they are both critical of A22 and supportive of AR6. Even prominent contrarians now frequently denounce the IPCC's position on extreme weather as misinformation because they believe AR6 says something similar to what these climate scientists claim it says. The views taken by A22 are anything but "perfectly consistent" with the IPCC, and this will be even more clear in the next section.

Should A22 Be Retracted?

The fact that A22 disagrees with the IPCC on extreme weather does not mean that it's wrong. It's certainly possible it provides a critical appraisal of flaws in the conclusions made by AR5 or AR6. But signs are pointing towards its eventual retraction. Pielke argues that a retraction of A22 is unwarranted because it's "based not on any claims of scientific misconduct, but simple disagreement." He insists "there is absolutely no allegation of research fraud or misconduct here, just simple disagreement." Here I have to take issue with Pielke's criteria for retraction. There are very good reasons to retract a paper for reasons that don't amount to fraud or misconduct. And in other contexts, Pielke agrees. In another post about the COVID-19 "proximal origins" paper, Pielke advocates for a broader set of scruples regarding when a paper should be retracted, and he quotes approvingly of Nature's statements. Here are a couple important highlights:

Retraction is a mechanism for correcting the literature and alerting readers to articles that contain such seriously flawed or erroneous content or data that their findings and conclusions cannot be relied upon.

And again, 

Unreliable content or data may result from honest error, naïve mistakes, or research misconduct. The main purpose of retraction is to correct the literature and ensure its integrity rather than to punish the authors.

While I disagree with his position on the "proximal origins" paper, I do agree with Pielke when he acknowledges that a paper can and should be retracted for containing serious flaws that mean their conclusions are not reliable. It doesn't matter if the flaws are due to honest/naïve mistakes with no evidence of fraud or misconduct. The issue is the integrity of the literature. These are the scruples I'm going to use. 

After SkyNews reported on A22 late last year, both the Guardian and AFP ran stories about this paper, and each asked four scientists to read and evaluate the paper. Representative of the responses is Richard Betts' assessment was that "the paper gives the appearance of being specifically written to make the case that there is no climate crisis, rather than presenting an objective, comprehensive, up-to-date assessment." Others noted that key areas were largely omitted. For instance, there was no significant treatment of heatwaves. Michael Mann was perhaps the most scathing. He says this paper is

another example of scientists from totally unrelated fields coming in and naively applying inappropriate methods to data they don’t understand... Either the consensus of the world’s climate experts that climate change is causing a very clear increase in many types of weather extremes is wrong, or a couple of nuclear physics dudes in Italy are wrong.

This to me goes beyond evaluating the paper to a personal attack. Even if accurate, it wasn't necessary to evaluate the merits of the paper. But what's clear is that, the 8 scientists that evaluated the paper and at least some of the reviewers that later would investigate this paper (and an addendum) gave similar criticisms. This paper cherry picks many true things and leaves out a substantial amount of evidence that would demand different conclusions if it were included in the study. These scientists are not simply disagreeing with the paper; they're saying that it so flawed that its conclusions cannot be relied upon. In my own reading of the paper, I have to agree. Given the length of this post, let's just look at two aspects of this paper: the lack of coverage of extreme heat and the selective coverage of tropical cyclones. 

Heatwaves

There is less than a full paragraph covering heatwaves. The bulk of what they say is summarized as "The global analysis carried out by Perkins-Kirkpatrick and Lewis showed for the period 1951–2017 a significant increase in yearly values of heatwave days, maximum heatwave duration and cumulative heat, while the global heatwave intensity trends are not significant." Interestingly, though, Perkins-Kirkpatrick and Lewis[2] say a bit more than A22 tell you. The first sentence of the abstract begins, "Heatwaves have increased in intensity, frequency and duration, with these trends projected to worsen under enhanced global warming." Why does this paper say heatwave intensity has increased when A22 says it hasn't? What A22 says appears to come from a not-so-careful reading of the following:

Heatwave frequency demonstrates the most significant increase across almost all regions, with nowhere experiencing a significant decrease. While average heatwave intensity displays little change, cumulative heatwave intensity increases at a similar rate to heatwave frequency. Moreover, the most intense heatwave seasons as defined by cumulative intensity generally occur post 2000.

This study found that average heatwave intensity displays little change (statistically significant increases were observed regionally but not globally) but cumulative heatwave intensity has increased in step with heatwave frequency. More than this, the paper explains why it is that average heatwave intensity hasn't changed much.

Although mostly insignificant, trends in average intensity are smaller (or as mentioned above, even decreased) for most regions than global warming over the same time period (0.1 °C decade−1 since 1950). By itself, this result is not entirely surprising since the heatwave definition used here is based on a fixed threshold (see “Methods”); thus, average intensity is inversely proportional to the number of heatwave days. Therefore, as the number of heatwave days increase (Fig. 1b, Table 1), little or no change in average intensity can be expected. However, measuring heatwave intensity in this manner does not address the fact that more heatwaves mean more overall exposure to extreme temperatures. Indeed, other well-used measures of heatwave intensity such as the hottest heatwave day,—which generally have more significant trends than average intensity—also do not account for this. Our assessment of cumulative heat fills this gap. Consistent with global changes (Fig. 1h), regional changes in cumulative heat have increased since 1950.

By definition, as heatwave days increases, average intensity decreases, which is why average heatwave intensity doesn't change much. However, as heatwave frequency increases, exposure to extreme temperatures also increases. Cumulative heat is a better metric to assess overall changes in intensity.

Average Heatwave Intensity (a) and Cumulative Heat (b)

It would seem A22 did not spend sufficient time understanding the position taken by their source material and did not represent it accurately. This paper does not provide us with anywhere near an adequate treatment of extreme heat to justify its conclusion, and what it does say is a bit misleading. And this is not a minor component of AGW's effect on extreme weather. It is the essentially the primary way AGW affects extreme weather, but A22's coverage of it is trivial and inaccurate.

Tropical Cyclones

The treatment of tropical cyclones in A22 is likewise extremely superficial. The following quote summarizes the position of A22 pretty well: "To date, global observations do not show any significant trends in both the number and the energy accumulated by hurricanes." The entire section on tropical cyclone trends focuses on hurricane frequencies and difficulties assessing trends, concluding (correctly) that we can't be confident of any trends in hurricane frequency. However, they don't explain that climate scientists generally aren't expecting increases in hurricane frequency either. This is the least significant aspect of any analysis of hurricane trends.

For hurricane intensity, the quote above encapsulates everything the paper says - there's no trend in a particular metric called Accumulated Cyclone Energy (ACE). ACE is a metric that sums the square of a tropical cyclone's maximum sustained winds, measured every six hours. It's not specifically an intensity metric; it's value for a hurricane season is affected by cyclone frequency, intensity, and duration. A season with fewer but more intense storms may have the same ACE index as a year with more, less intense storms. A22 does not evaluate TCs at all using other metrics like Power Dissipation Index (PDI) or even the proportion of Cat 3+ storms, which has been increasing across the satellite era.

There are many aspects of tropical cyclones aside from frequency and intensity that to a greater or lesser degree may be affected by AGW, and these can affect the risk that coastal populations may experience from tropical cyclones. None of these are even mentioned. This paper doesn't address the likelihood of storms carrying more water and dumping more rain, increasing risk of flooding. As sea level rise continues, storms are likely to cause more damage inland. As global temperatures increase, there's a likelihood of the poleward movement of tropical cyclones. This paper also doesn't address risks associated with rapid intensification of storms of the possibility that storms may "stall" more frequently, causing more damage where they stall. There's no accounting for the effect of wind sheer limiting storm formation and duration. Pretty much all the most interesting aspects of tropical cyclone trends are missing from this paper.

A22 knows about a summary of recent literature on tropical cyclones published by NOAA. NOAA's treatment is very good and details much of what we know and don't know about tropical cyclone trends. While A22 quotes from it, it neglects everything in the NOAA summary except for statements that hurricane frequency hasn't increased. There is also a two-part meta analysis[3][4] that is an impressive synthesis of a large amount of TC-related data. A22 doesn't even cite these studies. They would have been a significant resource for addressing these issues above, but A22 doesn't show it has any knowledge of the existence of these studies. It would seem A22 chose a single component of tropical cyclones - frequency - for which there are no trends and then ignored other aspects of tropical cyclones for which there are detectable trends. With this kind of cherry picking, it's easy to conclude "none of these response indicators show a clear positive trend of extreme events." And Ken Rice uncovered similar issues with other response indicators. Other than a passing reference to and inaccurate summary of heatwaves, the response indicators A22 chose to cover didn't show clear positive trends because those response indicators that do exhibit trends are largely ignored.

Conclusion on Retraction

So should this paper be retracted? If we use a simple "no fraud and no misconduct" criterion, maybe not. But if a paper is allowed to be biased and misleading so long as its authors are not guilty of outright fraud or misconduct, that's a pretty low bar. But if we use Pielke's additional scruple that a paper should be retracted if it contains "such seriously flawed or erroneous content or data that their findings and conclusions cannot be relied upon," then I think yes, it should be retracted. This paper's conclusions cannot be relied upon - the research was too biased and superficial and insufficient to warrant its conclusions.

Consider a hypothetical paper investigating Drug X designed to promote weight loss. Let's say the paper concludes that there's no evidence that Drug X is unsafe when used as directed, with the most common symptoms being temporary headaches and nausea. Let's say there were 15 trials for Drug X, and this paper examined 5 of them. In those 5, the authors state truthful things that are in the trials but neglect to mention that Drug X was associated with increased rates of heart attacks among men older than 65. In the 10 studies the paper didn't examine, Drug X was associated with a large increases in spontaneous miscarriages among women. In 6 of the trials that were ignored, people gained weight with respect to the control group. Now let's say every sentence in the paper is true, even if highly selective, and the paper concludes, "based on the evidence we examined, Drug X is safe." Even though every sentence in the paper is technically true, should it be retracted? Since its conclusions can't be relied upon, absolutely. Without a retraction, this hypothetical paper could cost people's lives.

Are Climate Scientists and Journalists Colluding to Retract this Paper?

The most disturbing aspect of Pielke's treatment of A22 comes from the accusations he makes concerning the 8 scientists interviewed by the Guardian and AFP. He says the retraction is "politically-motivated" and the paper is being retracted simply because climate scientists "disagree" with it, rather than because of the flaws in the paper. He further says these "activist scientists teamed up with activist journalists to pressure a publisher – Springer Nature, perhaps the world’s most important scientific publisher – to retract a paper." His conclusion is that "Shenanigans continue in climate science, with influential scientists teaming up with journalists to corrupt peer review." Now, we've already seen that there are plenty of reasons why this paper might be retracted without any need to invent political motivations. But is there evidence that "activist" scientists are colluding with journalists "behind the scenes" to pressure Springer Nature to retract a paper? Does it even make sense that Springer Nature could be pressured by these two news stories? I don't think so. I have no idea why Springer would be pressured to bow to the will of positions shared in two news stories. Two of the scientists even argued against retraction, since doing so could "lead to further publicity and could be presented as censorship." How would Springer be pressured to retract by news stories that take both positions on retraction?

I looked back over the Guardian and AFP stories, and a couple things stand out as telling. First, this entire affair was initiated by contrarian news coverage, not by "activist scientists and journalists." This paper was published in January 2022.  Following this in February 2022 NTZ blogged about this and Pielke tweeted about it, but even Pielke admits it received little attention. Then about 8-9 months later, SkyNews reported about it. The publicity generated by contrarian media prompted WUWT to blog about it, and the Guardian and AFP wrote their stories in which they solicited evaluations from climate scientists. It was around this time that I became aware of this paper in online forums. I also found a very good response to the paper by Ken Rice. What is interesting is the degree to which climate scientists didn't even know about this paper or the journal it was published in. None of those interviewed by AFP had heard of the paper, and at least one had never even heard of the journal. I've not seen any evidence that any climate scientist (other than Pielke) was paying attention to it until it became an Australian news story. My best guess is that Pielke's accusation comes entirely from the title of the AFP article, "Scientists urge top publisher to withdraw faulty climate study," even though the text in the article doesn't support that. Two of the four scientists said it shouldn't be withdrawn.

From the evidence we have, this is the full extent of what these "activist" climate scientists have done. When either the Guardian or AFP asked them about the paper, they read the paper and gave the Guardian or AFP their evaluation of the paper, snippets of which were reported by journalists. While all agreed the paper was fatally flawed, some said the paper should be retracted; some said it shouldn't. Then, after Springer and EPJP editors read the Guardian and AFP stories and launched an investigation into the paper, EPJP editors contacted these scientists, requesting a scientific comment to the paper. That request was apparently met with silence. Pielke says, "The eight 'colleagues who expressed concern' via the media (and listed above) all apparently chose not to provide a scientific comment on Alimonti et al. and no further discussion of the comment was made in subsequent correspondence that I have seen." I've searched for other public statements from climate scientists on this paper and so far found nothing (unless you count Ken Rice above). I can't find any public rebuttals or any claims that any of them took any action other than responding to the requests from the Guardian and AFP.

Now it looks like the process between Springer and EPJP may be a complete mess, and Springer may  well be pressuring EPJP. Pielke does not have full information, and he has not shared all the information he has, so I'm not going to comment on the process between the authors, Springer, and EPJP following the investigation. But whether Pielke is right about problems within that investigation or not, he supplied absolutely no evidence that scientists are teaming up with journalists to get this paper retracted. For all we know, these scientists gave their opinions on the paper to the Guardian and AFP and then completely forgot about it. And more importantly, if they wanted to exert their influence to get the journal to retract the paper, they were invited by EPJP to do so and didn't. In what possible world does this count as "activist" scientists and journalists colluding behind the scenes to pressure a publisher to retract a paper? They didn't even agree with each other on retraction.

Conclusion

This whole thing is puzzling to me because Pielke Jr is a smart guy. I disagree with him here, but he's a published scientist in this field, and he is probably far more knowledgeable than I am on the influence of AGW on extreme weather. I've tried to understand his point of view here, but I cannot for the life of me figure out why it is that it appears like he's juggling multiple sets of scruples and applying them selectively in various contexts. So:

1. Why does Pielke apply the retraction criterion of a "paper so flawed that its conclusions can't be relied upon" to the "proximal origins" paper but not to A22?
2. Why does Pielke go to such great lengths to defend A22 despite its obvious flaws while simultaneously unleashing such venom towards these 8 climate scientists on the basis of so little evidence?
3. Why is Pielke accusing these scientists of trying to interfere in the review process while at the same time inserting himself into the review process? After all, he sent EPJP an an email asking them to justify their actions.

I think the clue to my answers may by in his phrase, "politically-motivated." Ideology can be a very useful tool, but it can also be a blinder. We're more likely to defend those that conform to our ideology and attack those who have a competing ideology, particularly if we become convinced that the controversy is really about the ideology and not about the evidence.

And it's not like all eight of these scientists are completely innocent here. Mann's response in the Guardian goes beyond an evaluation of the paper to a personal attack on the authors. And perhaps a reasonable case could be made that if they were going to go on record with their evaluation of this paper, they should also have been willing to participate in the post-publication review process and submit a scientific comment to the journal as requested by the journal's editors. Would that have helped the process between Springer and EPJP go more smoothly? I have no idea.

If I'm right about the issues surrounding this paper, perhaps it's Pielke's ideology that's blinding him to the flaws in A22 and generating the "scientists teaming up with journalists" explanation for the retraction of this paper. But I'd be foolish to think that my own ideology never affects me in similar ways. Could it be that I rush to judgment against those I disagree with or rush to defend those I agree with? Is my reading of this situation as flawed as I believe Pielke's is? The lesson I'm taking from this is that it's important for me to be humble about this possibility and take a step back to evaluate my own positions and my own motivations for either defending or criticizing the studies I read. Should that uncover something about me that made me miss something, I'll come back to this and reassess.

References:

[1] Alimonti, G., Mariani, L., Prodi, F. et al. A critical assessment of extreme events trends in times of global warming. Eur. Phys. J. Plus 137, 112 (2022). https://doi.org/10.1140/epjp/s13360-021-02243-9

[2] Perkins-Kirkpatrick, S.E., Lewis, S.C. Increasing trends in regional heatwaves. Nat Commun 11, 3357 (2020). https://doi.org/10.1038/s41467-020-16970-7

[3] Knutson, T., and Coauthors, 2019: Tropical Cyclones and Climate Change Assessment: Part I: Detection and Attribution. Bull. Amer. Meteor. Soc., 100, 1987–2007, https://doi.org/10.1175/BAMS-D-18-0189.1.

[4] Knutson, T., and Coauthors, 2020: Tropical Cyclones and Climate Change Assessment: Part II: Projected Response to Anthropogenic Warming. Bull. Amer. Meteor. Soc., 101, E303–E322, https://doi.org/10.1175/BAMS-D-18-0194.1.