[see bottom of post for an update]
Steve McIntyre, chief blogger and workhorse at the blog ClimateAudit, has a recent post which is grabbing a lot of attention across the web and being trumpeted by some as a triumphant unmasking of the fraudulent behavior in the preparation of the IPCC Third Assessment Report (TAR).
Science/science policy blogger Roger Pielke Jr. covers Steve’s post, with a post of his own, under the title The “Trick” in Context. However, I think the post should more accurately have been titled The “Trick” in “Context.” For the “context” is one supplied by Steve McIntyre. My read of the relevant emails surrounding the incident in question doesn’t lead me to the same conclusions as Steve.
Context must be supplied in this case. For as anyone who has looked through any of the leaked/stolen Climategate emails (available here) quickly realizes, most of the email threads are not complete from start to finish, and, as is typical of most conversations, they assume the participants already know a lot of what is being discussed, including the context. To outside parties peering in, often the context must be derived, inferred, or guessed at.
The topic in question has to do with development of the discussion and graphics to be included in the paleoclimate section of Chapter 2 of the IPCC’s TAR. (The IPCC has since published its Fourth Assessment Report [AR4], which dedicates an entire chapter to paleoclimate rather than a brief section of a single chapter.) The discussion taking place in the emails highlighted by McIntyre is about what the best scientific understanding (at the time) of what the earth’s temperature behavior was during periods that pre-date the widespread direct temperature measurements made by thermometers (in this case, the past 1,000 years or so).
This is relevant to the issue of anthropogenic climate change because quantifying the degree of “natural” variability of the earth’s temperature helps to understand how unusual our current warmth might be. To some people, it is also important because they want to use it to try to argue that the earth’s temperature has gotten to be as high as it is now solely because of natural process. But this view is almost certainly wrong.
The Temperature ‘Trick’ Revisited
McIntyre has dedicated a phenomenal amount of time and energy into trying to decipher just how paleoclimate researchers have come to assemble their millennial temperature reconstructions—which are necessarily built upon uncertain data, relationships, and interactions—and then trying to determine whether the methods (and thus the ultimate results) are robust. By and large, he has had to do this without the assistance and/or blessing of the paleoclimate community (to put things nicely—look through the Climategate emails to see opinions of Steve in less nice terms). Steve’s effort has basically introduced the idea of research “audits” to the climate community, as well as forwarding the idea that all data and methodologies in the peer-review literature should be archived and open. This alone is a praiseworthy and significant contribution that will (hopefully) have lasting positive influence on how climate science is documented.
But, in this particular case, I think that he is trying too hard to attach something nefarious to actions that occurred nearly a decade ago.
To understand the scenario that Steve is putting forth, you have to visit his site and read through his line-of-reasoning. Basically, he thinks the authors responsible for the relevant portion of the IPCC TAR Chapter 2 were being “tricky” to an extent unjustified by the scientific knowledge at the time, in order to try to show that the earth’s temperatures during the latter part of the 20th century were unprecedentedly high (furthering the evidence that human greenhouse gas emissions were driving the climate warmer).
However, when I look through the emails and try to follow Steve’s thread, I don’t arrive at the same conclusions (nor do some other folks). Steve imparts a context to snippets of emails that I believe does not accurately reflect the true context. Steve sees the authors trying to “hide the decline” (i.e., trying to conceal one particular reconstruction: the one published by Briffa et al. in 2000 that shows temperatures in the latter half of the 20th century declining), while other reconstructions–and the observations themselves–show that temperatures were rising. But I see them trying to understand why the available temperature reconstructions are different from one another. Steve sees actions that are highly questionable; I see fairly normal scientific discourse.
I can agree with Steve on one thing: the IPCC presented some data in a way that was different from how the data was originally presented in the peer-reviewed literature. Primarily, the reconstructed temperature as described by Briffa et al. (2000), which originally extended from the years 1402 to 1994, was truncated to end in 1960. However, this truncated version of the Briffa et al. (2000) reconstruction was supplied to the IPCC by one of the co-authors of original research, with the explanation that they did not have very much faith in the accuracy of the reconstruction past that date (because the reconstructed temperatures diverged from the observed temperatures) (see email 951763817.txt).
That the IPCC authors choose to accept and use this version of the Briffa dataset doesn’t seem to me to be much of a “trick.” If I were going to use some data, and the researchers who had derived the data informed me that some of it was quite possibly wrong, I probably would try to avoid using it—or at least the questionable portions.
Perhaps the only thing vaguely “tricky” was that Briffa was also a contributing author to this IPCC chapter. So perhaps it could be argued that he was allowing his dataset to be used in a way other than how he originally published it in order to further the IPCC’s desire to show that the recent temperatures were unprecedented (and humans were the cause).
However, three things argue against that proposition: 1) that in Briffa’s peer-reviewed publications (prior to the IPCC review) describing his reconstructions, he openly recognized the post-1960 divergence problems between his data and the observations and admitted it needed further explanation; 2) in one of the emails he stated that he did not believe that there was sufficient evidence to conclude that the temperatures in the recent decades were unprecedented in the past 1,000 years (see email 938031546.txt); and 3) based on the personality that Briffa displays throughout the Climategate emails, he certainly doesn’t act like one who rolls over under pressure. So it seems unlikely that Briffa volunteered to truncate his dataset in order to satisfy the IPCC’s desire (or at least the desire of the other authors of that particular Chapter) to show that recent temperature were unprecedented. The only person who knows for sure is probably Keith Briffa.
Now perhaps, you may argue that if Briffa was uncomfortable about any parts of his analysis the prudent thing would have been for him to have requested that it not be included at all. And maybe that would have been the correct thing to do and then there would have been no apparent “trick” at all. But that is not what Briffa decided—and that decision 10 years ago has led to this discussion.
I totally agree that several (or many) IPCC/TAR authors wanted to push the idea that the warmth of our most recent decades was unprecedented in the history of earth’s temperatures during the past 1,000 years. And those authors who wanted to do so were incredibly successful in actually doing so (look no further than Figure 1b on page 3 of the Summary for Policymakers [SPM]). And perhaps the science at the time did not justify this. But, SPM Figure 1b (and its creation) is not the figure under discussion by McIntyre.
Don’t get me wrong: I think that there is a lot of scientific misbehavior evident in the email collection, some of which I commented on in an earlier post. But I also think there is a lot of normal scientific discourse. I want to be careful not to confuse the latter for the former. In this specific case, I come down on the side of the latter.
[update 12/14/2009: Upon further reflection and consultation of the (at the time) extant Briffa peer-reviewed publications, I think the best course of action would have been not to include the Briffa series in IPCC TAR Figure 2.21. The problems with it were not fully resolved and it was not an apples-to-apples comparison with the other data series in Figure 2.21 anyway. The IPCC TAR authors acted outside the peer-review literature when they decided not to include the Briffa series post-1960. If they were intent on doing so, their decision should have been clearly justified in the text. There was a brief description in Section 188.8.131.52 (TAR p. 135) of a problem with tree-ring temperature reconstructions in recent decades, but in no indication that the TAR authors took this into account in the preparation of Figure 2.21. In fact, a reader would have been confused if they had turned to Figure 2.21 in order to confirm the description in Section 184.108.40.206 that the “tree-ring density variations have changed in their response to temperature in recent decades.” There was no evidence whatsoever of this in Figure 2.21.
The TAR authors should have included a direct acknowledgement in the caption of Figure 2.21 that the Briffa series had been modified. But they did not. Whether or not this was done in a coordinated IPCC effort to “hide the decline,” or whether it was a decision that rested only with Briffa being uncomfortable with the that portion of his reconstruction, cannot be determined from the associated email content (in my opinion). -Chip]