It is time to stop quivering in our boots in pointless fear of the future and just roll up our sleeves and build it.
- Ray Pierrehumbert

Tuesday, September 28, 2010

More Top Notch Stuff

Perhaps most importantly:

Please and thanks.

There's also two more important climate meta-science pieces:

and Stoat has the latest installment on


Doug said...

Michael, as a person of judicious temperament and better than average knowledge on the topic, do you have any suggestions for journalists about when it may actually be permissible to mention climate change in connection with an extreme weather story? As usual, MIT's Science Journalism Tracker wags the finger at a reporter for straining the patience of readers by connecting climate change with LA's recent historically high temperatures.

Surely at some point it's necessary to set some context? When? The record here was broken by some 6 degrees fahrenheit. Would it have ok to bring up if had reached 10 degrees?

This is becoming an annoying tendency on the part MIT ScJT.

Michael Tobis said...

Doug, this is a topic I've taken up recently. The issue is called "rolling a 13", coined (I believe) by Prof. Steven Sherwood of the University of New South Wales in Australia.

See this article which is pretty much where my thinking is on that subject at present.

I think the MIT article is fine as far as it goes, but it fails to address the point that we can be absolutely sure there would be no steam bubbles in the pot at all if the burner under the pot were turned off. So although you cannot attribute any specific bubble to the intensity of the heating within a certain range, once the range gets broad enough you get new phenomenology altogether. So the argument is valid for small changes and invalid for changes beyond a certain threshhold.

As an example, we have the retreat of Arctic sea ice, which can reasonably be attributed to human forcing.

Another ything to add is that a pure statistical attribution is impossible. Statistics never proves or disproves hypotheses at all. In some circumstances it can support or weaken hypotheses so strongly as to make little difference, but not always. The absence of statistical "proof" (as if such a thing existed) is very different from a statistical refutation. Statistical thinking is a useful tool but not everything in science can be reduced to statistics.

Earth and planetary sciences must rely on different sorts of reasoning in many cases because we only have one planetary history. For events with few or no analogs, it's impossible to apply statistics.

Statistics are mute and moot on the summer of 2010. We haven't seen anything like it. Is it a "3000 year event" or is it something that could not happen at all in a natural state? We don't know yet.

The claim that this is not a meaningful question is, in my opinion, wrong. On the other hand, you should not be of the impression that I am speaking for the consensus on this point. I suspect it is still a minority opinion.

jstults said...

On the technical issues Wegman and North substantially agreed (yes, I've read both, rather than "analysis" written by someone trying to get me to write my congressman). The upshot is it's probably a bad idea to roll your own PCA code when there are so many good open source alternatives available.

My goodness, Wegman sent the report to teh statisticians he knew !!eleventyone!1! Someone gen up a social network graph, quick!

jstults said...

Also, it is sad to see Only In It rate as "top notch" what is a silly Glenn Beck style political hack.
Is this guilt by association or is there really guilt regarding the Wegman Report?
Gee, so subtle...

dhogaza said...

"On the technical issues Wegman and North substantially agreed (yes, I've read both, rather than "analysis" written by someone trying to get me to write my congressman). "

Strange, I don't remember North claiming that CO2 sinks and forms a layer close to the surface of the earth ...

jstults said...

Red Herring.

Both reports are interested in methodological critiques:
The Chairman of the Committee on Energy and Commerce as well as the Chairman of the Subcommittee on Oversight and Investigations have been interested in an independent verification of the critiques of Mann et al. (1998, 1999) [MBH98, MBH99] by McIntyre and McKitrick (2003, 2005a, 2005b) [MM03, MM05a, MM05b] as well as the related implications in the assessment.
Wegman Report

The NRC asked the committee to summarize current scientific information on the temperature record for the past two millennia, describe the main areas of uncertainty and how significant they are, describe the principal methodologies used and any problems with these approaches, and explain how central is the debate over the paleoclimate temperature record to the state of scientific knowledge on global climate change.
North Report

Both reached the same conclusions about flubbing up the methodology (which happened to be PCA, see). None of this probably matters a great deal except for the political hay making (which results in some pretty entertaining buffoonery all around, so thanks for that).

Doug said...

Thank you, Michael.

My fear is that as we're looking at a statistical feature only noticeable to us on a scale of perhaps one to two decades, Petit's prototype chiding (regardless of who delivers it) over referring to climate change when journalists cover eyebrow-raising individual weather events will forever dominate.

In a strange way it reminds me of people who scoff at mandated use of life-preservers. On any one or boat ship we're unlikely to see a life saved by such regulations, but if you're the person sitting behind the desk where industry statistics arrive, the picture is suddenly very clear.

Perhaps of more relevance to this issue, It's highly conventional these days for newspaper accounts of injury traffic accidents to note if vehicle occupants were wearing seatbelts, regardless of what role restraint devices played in a particular accident. Nobody seems to find this objectionable.

Of course, local coverage is necessarily myopic. Almost simultaneously with LA's experience, a few hundred miles up the coast Bella Coola, BC received over twice the average monthly rainfall for September, in 36 hours another historical record. Meanwhile here in Seattle, we've been experiencing evening temperatures far above normal for a couple of weeks. Local necessarily coverage can't capture this sort of regional, synchronous strangeness.

frank -- Decoding SwiftHack said...
This comment has been removed by the author.
frank -- Decoding SwiftHack said...


So when Wegman and his students engage in blatant plagiarism and shoddy fact-checking and referencing, and MT draws our attention to these things, it's actually MT who's a "political hack"?

(Then surely plagiarism is the William Shakespeare of Liberal Fascism! Just as the White Man is the Jew of Liberal Fascism.)

-- frank

jstults said...

Of course you're right frank. I forgot that shoddy scholarship implies conspiracy in the climate blogosphere.

So many self-righteous indignorami are convinced that Mann's mistakes of statistics and Wegman's mistakes of climate are proof of vast conspiracy. Could it be that getting a multi-disciplinary problem right is hard? I mean, just look at the cast of thousands that had a hand in that North report compared to the number of authors on the Wegman one.

I certainly wasn't calling MT a hack (read what I wrote more carefully). I was simply appealing to his better angels. I know he's got some because some of what he writes rises above mere hackery.

adelady said...

Doug, words for journalists?

They should all have a little cue card. **Although scientists are reluctant to say that any given weather event is caused by the warming climate, they do say that we can expect these events more often as warming continues.**

Once they've mastered that, they can use their writing skills to reframe for their particular audiences.

John Mashey said...

Shoddy scholarship and conspiracy are relatively orthogonal.

I have seen plenty of shoddy scholarship for which there was no obvious evidence of conspiracy.

However, in this the evidence of conspiracy is very strong, and is enumerated at length.

The simplest one was the claim that the Wegman Report was independent, and yet Yasmin Said wrote that the Barton staffer Peter Spencer (who has attended several Heartland conferences) sent them a daunting amount of material over the next 9 months.

Now, if someone wants to believe that

a) having a Barton staffer choose the research material is not conspiracy

b) that Spencer did the research himself

c) that Spencer spent the effort to know all about bristlecode pines and proxy combinations

d)that the thinktanks folks and MM, who'd been working on this for years suddenly stopped...

then I offer a fine selection of bridges for sale, not just that old Brooklyn Bridge.

"Conspiracy theory" justifiably has a bad reputation, but occasionally, real conspiracies do exist ...

and oddly, the very social networks analysis that the Wegman team mis-used has been very useful to me, albeit in informal form. [It's actually been in all 4 reports I've done, although mostly on the last page of the Monckton/Oreskes one.]

Of course, there was much in CCC, including the delicious emails in A.9. What else does one think is happening when CEI's Myron Ebell has copies of letters to MBH, to send to the White House, *before* one or more recipients?

Rattus Norvegicus said...

John, I say you offer the Tacoma Narrows bridge :-)

Michael Tobis said...

Josh S, John Mashey has sufficient credibility and seriousness of purpose that when he makes assertions like this I feel confident asking the press to look into it. I have not taken a position myself, but I think it sufficiently likely that if I were to look into it in detail John's story would hold water that I will try to contribute to getting this story broken out of the blogosphere.

It is certainly the case that most accusations against scientists with the misfortune to have ended up doing annual-resolution temperature reconstructions have been grotesquely out of proportion with any real issues that the field or individuals within it face.

So the fact that there is at least an informal collaboration if not a full blown conspiracy to defame them is hardly news. That Wegman's report may have deliberately fed the fire rather than accidentally so is plausible. Mashey's allegations should be further investigated.

jstults said...

MT: Mashey's allegations should be further investigated.
Sure, you may be able to generate some smoke and heat with this (it may give you something to write about for a while, and the social network analysis is kind of neat, I won't lie). But at the end of the day what do you have when you find what you expect? "I caught him red handed sending references to Wegman!!!"

It's still Bubkis, because you're in fundamentally the same position as the folks calling for more investigations of UEA's CRU.

dhogaza said...

"But at the end of the day what do you have when you find what you expect? "I caught him red handed sending references to Wegman!!!""

No, it's not Bupkis.

1. Senator Barton commissioned Wegman to do an *independent report* to inform his committee.

2. The "independent report" depends almost entirely on information sent to Wegman et al by Barton's staff.

3. Therefore, it is not an independent report, and representations to that effect are nothing other than lies.

Yes, the cynical among us aren't surprised and many wondered at the time. But guessing isn't the same as documentation ...

jstults said...

John Mashey:
Of course, there was much in CCC
Thanks for the link. Some of what you wrote is interesting.
I have avoided delving into the statistical minutiae, because they really do not matter (A.10.4). In fact, the approach used in the WR is classic: take an argument into terminology that looks impressive, but that most people cannot possibly follow, but creates doubt and confusion.
Ah, that's a shame, but conspiracy is more titillating than linear algebra and statistics. Every equation you add drops the readership, right?

Climate models are boundary-value problems, not initial-value problems like weather prediction. They provide ensembles of results to model large-scale and longer-term average behavior. They worry about 20-30-year trends, long enough to see the signal amidst the noise.

It is simply incredible that competent Physics PhDs would not understand all this. Two of them demonstrably know enough climate science to know how wrong this is.

Not really (this fairly recent, main-stream work shows that the initialization matters by comparing a random initializations with realistic initializations). What would actually be incredible is if climate scientists had found functionals of the earth system that weren't chaotic, or didn't depend on initial conditions. Then you'd have a "statistical mechanics of climate", and I think that would be quite a break-through. Twenty to 30 year trends are climate, but 10 or 11 years are not, OK; what about 50 years?

I like your tribute to Tukey in that report. It's also been my experience that statisticians often complain about not being consulted enough, and they're usually right. Also, thanks for your work on Unix.

Ha-Ha, Tacoma Narrows, you've got to be a really uniquely informed guy to have heard of that, good one...

Mal Adapted said...

JStults: Ha-Ha, Tacoma Narrows, you've got to be a really uniquely informed guy to have heard of that, good one...

I heard about "Galloping Gertie" as a child, so Rattus isn't uniquely informed. For that matter, I think you'll find most of the commenters here are at least as well-informed as you are, about nearly everything. You might keep that in mind when writing your comments.

jstults said...

Mal Adapted,

I am in no way being sarcastic in giving you a "Thank You" (as big as my previous, hearty "Ha-Ha") for the concern you express for my future comments. Since it's so famous that you heard of it as a kid, maybe I will buy that bridge after all.

Whoooshfully yours,

Hank Roberts said...

Here’s North:
video of his presentation given to Dessler’s class — explaining the Hockey Stick and Wegman hearing reports.

Primary source. Don’t miss it.

David B. Benson said...


Michael Tobis said...

In my opinion it's babkes, but you have to make compromises to your audience. There's also a case for "bobkes".

"bubkes (also spelled "bupkis"): emphatically nothing, as in He isn't worth bubkes (literally 'goat droppings', possibly of Slavic origin; cf. Polish bobki 'animal droppings')"

Mal Adapted said...

JStults, for my part I'll keep in mind that humor doesn't always come across the blogs in the way we expect 8^).

jstults said...

Yeah, I should have used an emoticon

John Mashey said...

re: HR's mention of North:

Yes, that is well worth seeing, not just for the discussion of Wegman, but for the first-person description of how NRC panels really work. that talk was an *extremely* useful source, in fact, it was a key to an entire Appendix, A.1.

Also, it should be a good reminder that Texas not only has folks like Joe Barton, but some *very good* climate scientists.

David B. Benson said...

Ah yes. Bupkis.

Doug said...

Not to begin routinely popping up here like a jack-in-the-box, but this piece by Richard Rood in the AMS policy blog seems quite in keeping with an arc Michael has been following of late: what's the place of scientists when it comes to informing policy (a mystery to some, apparently):

As the article is titled "If Lady Chatterley’s Lover, then …", here's a suitably provocative tease:

"At this point we have established that, going forward, “climate change” is a political issue, and it is subject to both the well founded and the pernicious aspects of the political process. This is nothing new; it fact it is ancient. Scientific investigation has challenged, with dire consequences to scientists, that the Earth is at the center of the universe and many other tenants of nations, religions, and corporations. The ramifications of their investigations rarely enter the minds of young people motivated by the scientific process. Therefore, not only are scientists not well positioned to participate in the realm of bitter ideological confrontation, scientists are, I assert, by both training and predisposition, easy foils for savvy political strategists.

This leaves the scientist in a lose-lose situation. They are required to defend themselves, but their self-defense perpetuates and amplifies the political confrontation. The confrontational process is not one, as one of my readers more eloquently stated, where we are looking for knowledge-based reconciliation of an issue. Knowledge-based reconciliation is the scientific instinct."

The rest of the story:

David B. Benson said...

Why is the notch always at the top? Easier to reach?