When the Team is on the move, they can sometimes move with surprising speed, as you’ll see in today’s story.
Over the last month, we’ve seen multiple changes to the Mann SI, at first, without any notice. More recently, they’ve started to note the existence of changes, though, the changes themselves are typically not reported (in the sense of a list of before and after values) and the old files have been deleted when the change is made, even if the incorrect values were actually used in Mann et al 2008 (this happened with the lat-long reversal of the rain in Spain). Sometimes, the change notices themselves have not proved permanent: for example, the notice of the correction of the Schweingruber MXD locations, once on the website, has itself now been deleted.
Now that we have a little more experience with changes to Mann’s SI, I’d like to re-visit our experience with the first such incident, one where I got wrongfooted and both Gavin Schmidt and a couple of CA readers criticized me as a result. There’s an interesting backstory on exactly how I got wrongfooted, which may appeal to people interested in dating problems and arguably shows a surprising degree of coordination of the Team when on the move.
The dispute arose over my comments on Sep 4 and 5 about an inline response by Gavin Schmidt to RC comment #23 ( the comment timestamped Sep 4 9:18 am EDT, but the inline response does not have a timestamp), in which Gavin stated:
The raw data (before any infilling) is also available on the SI site, and so you can look for yourself.
As it happens, on Sep 4, between 11.33 and 11.35 am EDT, over 2 hours after the above comment was submitted, I visited the relevant directory at Mann’s website and did not observe the existence of non-infilled data at that time. Strangely enough, I have timestamped files on my computer which enable me to establish the exact time of this visit. Here is a screenshot from my computer showing timestamps of 3 downloads from Mann’s website (rtable1209, rtable1209late and itrdbmatrix), timestamped locally on my computer between 11.33 and 11.35 am EDT.
The screenshot of my directory also shows a subdirectory “proxy” timestamped Sep 5 9:11 am, which is when I downloaded the Sep 4 version of the “original” data. This directory now has a later timestamp on Mann’s website because Mann deleted the Sep 4 “original” data on the afternoon of Sep 5, inserting a new version of “original” data.
So when I read Gavin’s inline response to Comment #23 (some time on Sep 4), my assumption – and given my late morning inspection of Mann’s website, hardly an unreasonable assumption – was that he’d been wrongfooted by Mann and had simply been in error in his claim that non-infilled data was available at the SI. I had checked the SI late in the morning of Sep 4 and it wasn’t there. I certainly didn’t think that he was intentionally misrepresenting the situation; I simply assumed that he’d got the wrong impression – that sometimes happens.
By early the next morning (Sep 5), a reader had observed at CA that Mann had altered his SI and provided a link to a new directory which proved to contain non-infilled data. Although the timestamp was overwritten when later on Sep 5, Mann deleted the Sep 4 “original” data and replaced it with new “original” data, a CA reader had recorded the Sep 4 timestamp as 15:42, which I adopt here for my timeline.
Gavin subsequently asserted on Sep 5 that (1) he had verified that the non-infilled data was online prior to making his inline response to the 9:18 am comment, (2) he made his inline response to the 9:18 am comment when he approved the comment; (3) the approval and inline comment was some time after 9.18 am EDT. The inline comment was a finite time before 12:14 pm, because a comment on the RC thread at 12:14 pm refers to comment #25 (made after #23). Therefore, comment #23 had to have been approved and online long enough for the RC reader to have read the various new comments on the thread and composed a reply. The 12:14 pm comment is short and wouldn’t take much time to compose, so approval needn’t have been much before 12:14 pm.
There’s one more part of the puzzle – server time. When was 15:42? If it was EDT (as RC server time), this would be 3:42 pm and would have made it impossible for verification to have taken place prior to 12:14 pm. However, one CA reader hypothesized that Penn State server time was UTC and another CA reader was able to prove this. UTC which is 4 hours ahead of EDT and 15:42 converts to 11.42 am EDT.
Mann altered his data 7 minutes after I visited his site.
Perhaps critics will have a little more sympathy for my being wrongfooted on this. Surely even someone as diligent as me can’t check Mann’s website every 10 minutes to see if he’s changed it. And look at the near-military precision of the Team on the move.
|11.35 am||SM visits Mann website. Non-infilled data not there.|
|11:42 am||Mann inserts non-infilled data.|
|~12.00||High Noon: Frank Miller arrives on train, meets gang; gunfight breaks out between Gary Cooper and Miller gang. Grace Kelly returns to assist Gary Cooper. Gary and Grace defeat all 4 members of the Frank Miller gang|
|12.03||Schmidt verifies existence of non-infilled data at Mann’s website|
|12.05||Schmidt approves comment #23, inserting inline response.|
|12:14 pm||RC comment mentioning comment #25|
While the above timeline is pretty improbable, it does seem to be what happened. Based on this review, I’ve added inline comments to any of my earlier comments that now appear to have been incorrect.
In this case, I think that one can also conclude that Mann had emailed Schmidt notifying him of the alterations to the SI and directed him to the new location. And that Schmidt had information not available to the public when he visited Mann’s website. (It’s possible that it’s one more bizarre coincidence, but c’mon.) If Schmidt had received information about alterations to Mann’s website that was not publicly announced, then, in my opinion, he should have provided a forthright change notice in his inline response, agreeing that, yes, Mann’s original SI failed to include “original” data, but reporting that Mann had amended his SI on Sep 4 (referring to the date of the amendment) so that there was no uncertainty. In my opinion, Schmidt’s inline response falls well short of the form of notice that is appropriate for someone in possession of information not available to the public. (If Schmidt was not in possession of such information and this was one more coincidence, then this criticism would not apply.) Provision of a proper change notice, with an explicit reference to the date of the change, would have avoided most, if not all, of the subsequent misunderstandings.
At the time, Gavin also made an accusation that appears to be against me that I wish to deny.
An RC reader:
made the following polite request to Gavin Schmidt:
Gavin, I realize it’s not your responsibility to patrol the skeptic hordes, but could you offer a quick summary of how the data set has been updated and where these changes are recorded?… I think (hope?) that McIntyre would happily “move on” and apologize after a clear statement that you were acting in good faith.
Instead of providing a clear timeline (which might have helped), Gavin petulantly stated:
[Response: What is the point? The presumption will be that I’ve just made something up and even if I didn’t, I’m a bad person in any case. I have no interest in communicating with people whose first and only instinct is to impugn my motives and honesty the minute they can’t work something out (and this goes back a long way). Well, tough. You guys worked it out already, and I have absolutely nothing to add. If McIntyre was half the gentleman he claimed to be, we’d all be twice as happy. – gavin]
I presume that I’m supposed to be one of the people whose “first and only instinct is to impugn my motives and honesty the minute they can’t work something out (and this goes back a long way)”. I deny this allegation on a number of counts.
First, I don’t have a track record “going back a long way” of repeated incidents where my “first and only instinct is to impugn [Schmidt’s] motives and honesty the minute [I] can’t work something out “. On the contrary, I think that any CA reader has to concede that I’m very patient with working out paleoclimate studies. On an earlier occasion, I asked a reader critical of me to name one such incident and received no reply.
I also deny that “impugning” people’s motives and honesty is “my first and only instinct” when I can’t work something out. It is neither my “first instinct” nor my “only instinct”. My general practice is exactly the opposite – to refrain from speculating on author’s “motives”. It’s a policy that I try to maintain at the blog though not all readers observe the policy. I challenge Schmidt or anyone else to support this allegation.
This is not the first such incident involving Schmidt. As an IPCC Reviewer (SOD Review Comment 6-760), Schmidt impugned the integrity of Ross and myself, falsely accusing us “deliberate obsfucations [sic]” as follows:
M&M2003 is a non peer reviewed publication, and as such should not be referenced here. The points raised are almost invaraibly due to misunderstandings, errors in the archived data set (subsequently corrected at Nature) and deliberate obsfucations. [IPCC SOD Review Comment 6-740]
Schmidt has no basis whatever for accusing us to IPCC of “deliberate obfuscations”. That is an untrue and defamatory allegation.
On an earlier occasion, Mann had made a similar defamatory allegation to Natuurwetenschap & Techniek (and you’ll be amused at the question that occasioned this answer):
This claim by MM is just another in a series of disingenuous (off the record: plainly dishonest) allegations by them about our work.
Again, this is an untrue and defamatory allegation.
I must say that I’m surprised at the recklessness of Mann and Schmidt, both as individuals and as employees representing presumably responsible organizations, in making such comments.