Tuesday, November 24, 2009

The Medieval Warm Period and the CRU E-Mails

Recently hackers broke into the University of East Anglia Climate Research Unit and posted about 62 megs of internal e-mails about the CRU's work.* The e-mails have caused a scandal because they appear to suggest that the CRU researchers have been burying and manipulating data to create the illusion of man-caused global warming.

For a layman reading the e-mails, it is hard to tell -- sometimes in private conversations academics have short-hand ways of talking about things that could be misconstrued if overheard. I'm going to contact a friend much more expert in such things than me to get his take, to see if these are innocent remarks, evidence of sloppiness, or outright nefariousness.

As regular Wordhoarders know, I've been very critical of the treatment of the Medieval Warm Period (MWP) by the National Oceanic and Atmospheric Administration (NOAA) in this country. In the past, NOAA has posted summaries of its research on their paleoclimatology page that I found misleading and at odds with the actual research cited. Until last year, NOAA's page suggested there there was no Medieval Warm Period, but they quietly began to acknowledge it (along with the Little Ice Age), but they fudge it by writing, "In summary, it appears that the late 20th and early 21st centuries are likely the warmest period the Earth has seen in at least 1200 years." Incidentally, elsewhere NOAA still calls it the "so-called Medieval Warm Period."

In any case, some of the commentary about the CRU e-mails has been regarding the Medieval Warm Period, but much of it has been redacted to just include the dirty stuff, so here I offer you the more complete context from an e-mail exchange started on June 4th, 2003:

[W]hat I had in mind were the following two figures: 1) A plot of various of the most reliable (in terms of strength of temperature signal and reliability of millennial-scale variability) regional proxy temperature reconstructions around the Northern Hemisphere that are available over the past 1-2 thousand years to convey the important point that warm and cold periods where highly regionally variable. Phil and Ray are probably in the best position to prepare this (?). Phil and I have recently submitted a paper using about a dozen NH records that fit this category, and many of which are available nearly 2K back--I think that trying to adopt a timeframe of 2K, rather than the usual 1K, addresses a good earlier point that Peck made w/ regard to the memo, that it would be nice to try to "contain" the putative "MWP", even if we don't yet have a hemispheric mean reconstruction available that far back [Phil and I have one in review--not sure it is kosher to show that yet though--I've put in an inquiry to Judy Jacobs at AGU about this]. If we wanted to be fancy, we could do this the way certain plots were presented in one of the past IPCC reports (was it 1990?) in which a spatial map was provided in the center (this would show the locations of the proxies), with "rays" radiating out to the top, sides, and bottom attached to rectanges showing the different timeseries. Its a bit of work, but would be a great way to convey both the spatial and temporal information at the same time. 2) A version of the now-familiar "spaghetti plot" showing the various reconstructions as well as model simulations for the NH over the past 1 (or maybe 2K). To give you an idea of what I have in mind, I'm attaching a Science piece I wrote last year that contains the same sort of plot.

Anyway, there you have it in fuller context for you to form your own opinion. Obviously, the dirty part people are talking about here is where the writer discusses trying to "contain" the Medieval Warm Period by consciously employing the logical fallacy of cherry picking a data set.

*I hesitated at first to post something that was illegally hacked, but I decided to press forward because 1.) the e-mails are already now fully public, and 2.) now that people know they are there, they've also been made public through legal means as well, and 3.) when in doubt, it's better to talk about the facts. My apologies to any Wordhoarders who find this ethically questionable.


  1. Thanks for posting on this. Outside of the illicit way in which these emails were obtained, I find two elements of the story to be particularly distressing:

    (1) The appearance of impropriety in the way research experts handle and present data. We just don't live in a world right now where people trust science and scientists, who are cast as untrustworthy 'elites' following a dastardly agenda against traditional values. The mere appearance of impropriety in this one case will have ramifications across disciplines, I predict.

    (2) The willingness of commentators to make broad assertions about all climate researchers and all climate data. It's hasty generalization and ad hominum time, folks, and I've seen lots of it in just a few days. It's a sad state of affairs for those of us doing and teaching critical thinking.

    The common thread in these two items concerns the ethics of data, or the ethics of using data to make reasonable inferences/interpretations.

  2. Anonymous8:19 PM


  3. Out of my depth here, but I think that further context may put Mann's e-mail, part of which you quote, in a better light. First, here is a link to one site where that e-mail has been posted. On that page, it is the last in a chain that started the previous day. In the exchange, Mann and his colleagues agree to provide a short position paper to the journal Eos, mainly in response to paper (published in a different journal) that is skeptical of anthropogenic global warming and that they feel is 1) bad science and 2) getting a lot of popular attention: the Discovery Channel, for instance, has reported on it. Their effort is explicitly intended to be not a full scholarly paper reporting any newly obtained results: the Eos editor advises that it "can be as long as 1500 words, or approximately 6 double-spaced pages. A maximum of two figures is permitted. A maximum of 10 references is encouraged...". It is, to some degree, an op-ed piece (albeit one with figures and citations) meant to convey what they feel is the current understanding of the science to a broader audience.

    It appears that when Mann expresses a desire in the e-mail to "contain" the Medieval Warm Period in one of the two figures by showing trends for two rather than one millennia, he means that the beginning of the MWP would not be shown in a diagram for 1001-2000 A.D. He wants to give it some context. However, there was at the time of the position paper apparently no global or hemispheric temperature reconstruction for the period 1-1000 A.D. that incorporated all the different temperature measurements that he believes shows the best picture -- although he notes that he has one being reviewed for publication. So he proposes instead to use other data sets. From the end of that e-mail:

    "In addition to the "multiproxy" reconstructions, I'd like to Add Keith's maximum latewood density-based series, since it is entirely independent of the multiproxy series, but conveys the same basic message. I would also like to try to extend the scope of the plot back to nearly 2K. This would be either w/ the Mann and Jones extension (in review in GRL) or, if that is deemed not kosher, the Briffa et al Eurasian tree-ring composite that extends back about 2K, and, based on Phil and my results, appears alone to give a reasonably accurate picture of the full hemispheric trend."

    Note Mann's desire to give a "reasonably accurate picture" using his various sources, and his interest in "independent" work that verifies other study.

    As for his use of the term "putative" to describe the MWP, I gather that while climatologists acknowledge that temperatures in and near the north Atlantic were notably warm at times in that period, they also have records indicating that was not true elsewhere. This [small pdf] is the paper in Eos to which that e-mail discussion led. Note the chart showing different medieval temperatures at different places.

    Additionally, Gary Schmidt in comments at his "Real Climate" blog indicates that this chapter [very large pdf] eventually grew out of the position paper. You may find the subsection on the MVP of interest.

  4. I've got an expert looking at the e-mails and the subsequent papers published. He'll need a few days, what with the holiday and all. I'll post his comments when I get them.

  5. Anonymous9:26 AM

    I can't read that extract as cherry-picking. He's talking about providing extra data to diminish the stand-out effect of the MWP graph, if anything, but as the Brigand above says, it's chronological span that's being expanded, not datasets that preach the climate change gospel being buried. The point, for either side, is that, for example, Greenland doesn't tell us about global climate, it tells us about Greenland (the knock-on problem being that the same is true of Arctic ice-cores, or, well, anything really).