IPCC Third Assessment Report and the hockey stick

By Judith Curry | Climate Etc. | April 29, 2014

Back in April 2011, I had a post on The U.S. House of Representatives Hearing on Climate Change: Examining the Processes Used to Create Science and Policy.  John Christy’s testimony is worth revisiting, in two contexts:

• problems with the IPCC process, most recently highlighted in context of WG3 [link]
• the Steyn versus Mann and Mann versus Steyn lawsuits [link]

John Christy has a unique perspective on how the hockey stick became the icon of the Third Assessment Report (TAR) – he served as a Lead Author (along with Michael Mann) on Chapter 2 Observed Climate Variability and Change. Relevant excerpts from Christy’s testimony:

In simplified terms, IPCC Lead Authors are nominated by their countries, and downselected by the IPCC bureaucracy with help from others (the process is still not transparent to me – who really performs this down-select?) The basic assumption is that the scientists so chosen as Lead Authors (L.A.s) represent the highest level of expertise in particular fields of climate science (or some derivative aspect such as agricultural impacts) and so may be relied on to produce the most up-to-date and accurate assessment of the science. In one sense, the authors of these reports are volunteers since they are not paid. However, they do not go without salaries. Government scientists make up a large portion of the author teams and can be assigned to do such work, and in effect are paid to work on the IPCC by their governments. University scientists aren’t so lucky but can consider their IPCC effort as being so close to their normal research activities that salary charges to the university or grants occur. Travel expenses were paid by the IPCC for trips, in my case, to Australia, Paris, Tanzania, New Zealand, Hawaii, and Victoria, Canada. Perhaps it goes without saying that such treatment might give one the impression he or she is an important authority on climate.

As these small groups of L.A.s travel the world, they tend to form close communities which often re-enforce a view of the climate system that can be very difficult to penetrate with alternative ideas (sometimes called “confirmation bias” or “myside bias”.) They become an “establishment” as I call them. With such prominent positions as IPCC L.A.s on this high profile topic, especially if they support the view that climate change is an unfolding serious disaster, they would be honored with wide exposure in the media (and other sympathetic venues) as well as rewarded with repeated appointments to the IPCC process. In my case, evidently, one stint as an L.A. was enough.

The second basic problem (the first was the murkiness of our science) with these assessments is the significant authority granted the L.A.s. This is key to understanding the IPCC process. In essence, the L.A.s have virtually total control over the material and, as demonstrated below, behave in ways that can prevent full disclosure of the information that contradicts their own pet findings and which has serious implications for policy in the sections they author. While the L.A.s must solicit input for several contributors and respond to reviewer comments, they truly have the final say.

In preparing the IPCC text, L.A.s sit in judgment of material of which they themselves are likely to be a major player. Thus they are in the position to write the text that judges their own work as well as the work of their critics. In typical situations, this would be called a conflict of interest. Thus L.A.s, being human, are tempted to cite their own work heavily and neglect or belittle contradictory evidence (see examples below.) In the beginning, the scientists who wrote the IPCC assessment were generally aware of the new responsibility, the considerable uncertainties of climate science, and that consequences of their conclusions could generate burdensome policies. The first couple of reports were relatively cautious and rather equivocal.

In my opinion, as further assessments were created, a climate “establishment” came into being, dominating not only the IPCC but many other aspects of climate science, including peer-review of journals. Many L.A.s became essentially permanent fixtures in the IPCC process and rose to positions of prominence in their institutions as a side benefit. As a result, in my view, they had a vested interest in preserving past IPCC claims and affirming evermore confident new claims to demonstrate that the science was progressing under their watch and that financial support was well spent. Speaking out as I do about this process assured my absence of significant contribution on recent and future reports. Political influence cannot be ignored. As time went on, nations would tend to nominate only those authors whose climate change opinions were in line with a national political agenda which sought perceived advantages (i.e. political capital, economic gain, etc.) by promoting the notion of catastrophic human-induced climate change. Scientists with well-known alternative views would not be nominated or selected. Indeed, it became more and more difficult for dissension and skepticism to penetrate the process now run by this establishment. As noted in my InterAcademy Council (IAC) testimony, I saw a process in which L.A.s were transformed from serving as Brokers of science (and policy-relevant information) to Gatekeepers of a preferred point of view.

A focus evolved in the IPCC that tended to see enhanced greenhouse gas concentrations as the cause for whatever climate changes were being observed, particularly in the 2001(Third Assessment Report or TAR) which was further solidified in 2007, (the Fourth Assessment Report or AR4.) The IAC 2010 report on the IPCC noted this overconfidence when it stated that portions of the AR4 contained “many vague statements of ‘high confidence’ that are not supported sufficiently in the literature, not put into perspective, or are difficult to refute.’” (This last claim relates to the problem of generating “unfalsifiable hypotheses” discussed in my recent House testimony.)

My experience as Lead Author in the IPCC TAR, Chapter 2 “Observed Climate Variability and Change”, allowed me to observe how a key section of this chapter, which produced the famous Hockey Stick icon, was developed. My own topic was upper air temperature changes that eventually drew little attention, even though the data clearly indicated potentially serious inconsistencies for those who would advocate considerable confidence in climate model projections.

First, note these key points about the IPCC process: the L.A. is allowed (a) to have essentially complete control over the text, (b) sit in judgment of his/her own work as well as that of his/her critics and (c) to have the option of arbitrarily dismissing reviewer comments since he/she is granted the position of “authority” (unlike peer-review.) Add to this situation the rather unusual fact that the L.A. of this particular section had been awarded a PhD only a few months before his selection by the IPCC. Such a process can lead to a biased assessment of any science. But, problems are made more likely in climate science, because, as noted, ours is a murky field of research – we still can’t explain much of what happens in weather and climate.

The Hockey Stick curve depicts a slightly meandering Northern Hemisphere cooling trend from 1000 A.D. through 1900, which then suddenly swings upward in the last 80 years to temperatures warmer than any of the millennium when smoothed. To many, this appeared to be a “smoking gun” of temperature change proving that the 20th century warming was unprecedented and therefore likely to be the result of human emissions of greenhouse gases.

I will not debate the quality of the Hockey Stick – that has been effectively done elsewhere (and indeed there is voluminous discussion on this issue), so, whatever one might think of the Hockey Stick, one can readily understand that its promotion by the IPCC was problematic given the process outlined above. Indeed, with the evidence contained in the Climategate emails, we have a fairly clear picture of how this part of the IPCC TAR went awry. For a more detailed account of this incident with documentation, see http://climateaudit.org/2009/12/10/ipcc-and-the-trick/.

We were appointed L.A.s in 1998. The Hockey Stick was prominently featured during IPCC meetings from 1999 onward. I can assure the committee that those not familiar with issues regarding reconstructions of this type (and even many who should have been) were truly enamored by its depiction of temperature and sincerely wanted to believe it was truth. Skepticism was virtually non-existent. Indeed it was described as a “clear favourite” for the overall Policy Makers Summary (Folland, 0938031546.txt). In our Sept. 1999 meeting (Arusha, Tanzania) we were shown a plot containing more temperature curves than just the Hockey Stick including one from K. Briffa that diverged significantly from the others, showing a sharp cooling trend after 1960. It raised the obvious problem that if tree rings were not detecting the modern warming trend, they might also have missed comparable warming episodes in the past. In other words, absence of the Medieval warming in the Hockey Stick graph might simply mean tree ring proxies are unreliable, not that the climate really was relatively cooler.

The Briffa curve created disappointment for those who wanted “a nice tidy story” (Briffa 0938031546.txt). The L.A. remarked in emails that he did not want to cast “doubt on our ability to understand factors that influence these estimates” and thus, “undermine faith in paleoestimates” which would provide “fodder” to “skeptics” (Mann 0938018124.txt). One may interpret this to imply that being open and honest about uncertainties was not the purpose of this IPCC section. Between this email (22 Sep 1999) and the next draft sent out (Nov 1999, Fig. 2.25 Expert Review) two things happened: (a) the email referring to a “trick” to “hide the decline” for the preparation of report by the World Meteorological Organization was sent (Jones 0942777075.txt, “trick” is apparently referring to a splicing technique used by the L.A. in which non-paleo data were merged to massage away a cooling dip at the last decades of the original Hockey Stick) and (b) the cooling portion of Briffa’s curve had been truncated for the IPCC report (it is unclear as to who performed the truncation.)

In retrospect, this disagreement in temperature curves was simply an indication that such reconstructions using tree ring records contain significant uncertainties and may be unreliable in ways we do not currently understand or acknowledge. This should have been explained to the readers of the IPCC TAR and specifically our chapter. Highlighting that uncertainty would have been the proper scientific response to the evidence before us, but the emails show that some L.A.’s worried it would have diminished a sense of urgency about climate change (i.e. “dilutes the message rather significantly”, Folland, 0938031546.txt.)

When we met in February 2000 in Auckland NZ, the one disagreeable curve, as noted, was not the same anymore because it had been modified and truncated around 1960. Not being aware of the goings-on behind the scenes, I had apparently assumed a new published time series had appeared and the offensive one had been superceded (I can’t be certain of my actual thoughts in Feb. 2000). Now we know, however, that the offensive part of Briffa’s curve had simply been amputated after a new realization was created three months before. (It appears also that this same curve was apparently a double amputee, having its first 145 years chopped off too, see http://climateaudit.org/2011/03/23/13321/.) So, at this point, data which contradicted the Hockey Stick, whose creator was the L.A., had been eliminated. No one seemed to be alarmed (or in my case aware) that this had been done.

Procedures to guard against such manipulation of evidence are supposed to be in place whenever biases and conflicts of interest interfere with duties to report the whole truth, especially in assessments that have such potentially drastic policy implications. That the IPCC allowed this episode to happen shows, in my view, that the procedures were structurally deficient.

Even though the new temperature chart appeared to agree with the Hockey Stick, I still expressed my skepticism in this reconstruction as being evidence of actual temperature variations. Basically, this result relied considerably on a type of western U.S. tree-ring not known for its fidelity in reproducing large-scale temperatures (NRC 2006, pg. 52).

At the L.A. meetings, I indicated that there was virtually no inter-century precision in these measurements, i.e. they were not good enough to tell us which century might be warmer than another in the pre-calibration period (1000 to 1850.)

In one Climategate email, a Convening L.A., who wanted to feature the Hockey Stick at the time (though later was less enthusiastic), mentions “The tree ring results may still suffer from lack of multicentury time scale variance” and was “probably the most important issue to resolve in Chapter 2” (Folland, 0938031546.txt). This, in all likelihood, was a reference to (a) my expressed concern (see my 2001 comments to NRC below) as well as to (b) the prominence to which the Hockey Stick was predestined.

To compound this sad and deceptive situation, I had been quite impressed with some recent results by Dahl-Jensen et al., (Science 1998), in which Greenland ice-borehole temperatures had been deconvolved into a time series covering the past 20,000 years. This measurement indeed presented inter-century variations. Their result indicated a clear 500-year period of temperatures, warmer than the present, centered about 900 A.D. – commonly referred to as the Medieval Warm Period, a feature noticeably absent in the Hockey Stick. What is important about this is that whenever any mid to high-latitude location shows centuries of a particularly large temperature anomaly, the spatial scale that such a departure represents is also large. In other words, long time periods of warmth or coolness are equivalent to large spatial domains of warmth or coolness, such as Greenland can represent for the Northern Hemisphere (the domain of the Hockey Stick.)

I discussed this with the paleo-L.A. at each meeting, asking that he include this exceptional result in the document as evidence for temperature fluctuations different from his own. To me Dahl-Jensen et al.’s reconstruction was a more robust estimate of past temperatures than one produced from a certain set of western U.S. tree-ring proxies. But as the process stood, the L.A. was not required to acknowledge my suggestions, and I was not able to convince him otherwise. It is perhaps a failure of mine that I did not press the issue even harder or sought agreement from others who might have been likewise aware of the evidence against the Hockey Stick realization.

As it turned out, this exceptional paper by Dahl-Jensen et al. was not even mentioned in the appropriate section (TAR 2.3.2). There was a brief mention of similar evidence indicating warmer temperatures 1000 years ago from the Sargasso Sea sediments (TAR 2.3.3), but the text then quickly asserts, without citation, that this type of anomaly is not important to the hemisphere as a whole.

Thus, we see a situation where a contradictory data set from Greenland, which in terms of paleoclimate in my view was quite important, was not offered to the readers (the policymakers) for their consideration. In the end, the Hockey Stick appeared in Figure 1 of the IPCC Summary for Policymakers, without any other comparisons, a position of prominence that speaks for itself.

So, to summarize, an L.A. was given final say over a section which included as its (and the IPCC’s) featured product, his very own chart, and which allowed him to leave out not only entire studies that presented contrary evidence, but even to use another strategically edited data set that had originally displayed contrary evidence. This led to problems that have only recently been exposed. This process, in my opinion, illustrates that the IPCC did not provide policymakers with an unbiased evaluation of the science, whatever one thinks about the Hockey Stick as a temperature reconstruction.

Judith Curry comments:  Christy’s assessment, when combined with the University of East Anglia emails, provides substantial insight into how this hockey stick travesty occurred.  My main unanswered question is:  How did Michael Mann become a Lead Author on the TAR?  He received his Ph.D. in 1998, and presumably he was nominated or selected before the ink was dry on his Ph.D.  It is my suspicion that the U.S. did not nominate Mann (why would they nominate someone for this chapter without a Ph.D.?)  Here is the only thing I can find on the U.S. nomination process [link].  Instead, I suspect that the IPCC Bureau selected Mann; it seems that someone (John Houghton?) was enamored of the hockey stick and wanted to see it featured prominently in the TAR.  The actual selection of Lead Authors by the IPCC Bureau is indeed a mysterious process.

The IPCC process is clearly broken, and I don’t see anything in their recent policies that addresses the problems that Christy raises.  The policy makers clearly wrought havoc in context of the AR5 WG3 report; however there is a more insidious problem particularly with the WG1 scientists in terms of conflict of interest and the IPCC Bureau in terms of stacking the deck to produce the results that they want.