In a sudden new development, your correspondent has learned that Norfolk Constabulary have decided that climategate is too big for them to handle. According to an un-named source, they intend to hand over the inquiry to another force.
Anthony got this story before me. Greg Laden updated a post writing
I’ve decided to update this blog entry (20 Dec 2011) because it occurs to me that certain things could be misinterpreted, in no small part because of the common language that separates us across various national borders, and differences in the way debate and concepts of free speech operate in different lands.
I want to make it clear that I do not think that the blogger “TallBloke” a.k.a. Roger Tattersall has broken British law.
Could be? I think it’s rather obvious that if Laden did not mean to accuse TallBloke of being a criminal then what he wrote clearly was misinterpreted by many. That list includes people who cannot in anyway be labeled “climate denialist” by the definition used by anyone, anywhere. For example:
You utter moron. The police raid does not imply there is any evidence that any of these people were involved in the hacking.
Posted by: James Annan | December 15, 2011 3:32 PM
and
Wow. You managed to temporarily reconcile hardcore climate truthers with James Annan and Dhogaza?
Congrats Greg. I guess. 🙂
Posted by: toto | December 15, 2011 6:13 PM
Donna Laframboise writes about the confusion arising from the similarity between “nofrakkinconsensus” and “noconsensus”. The article closes with
Donna Laframboise is the author of the recently published exposé of the Intergovernmental Panel on Climate Change, The Delinquent Teenager Who Was Mistaken for the World’s Top Climate Expert. She blogs at NoFrakking Consensus.com.
Couldn’t the post write the url correctly with no space? Oh well….maybe we should check amazon to see if this resulted in additional sales!
As easy as it is to confuse Tallbloke’s site “noconsensus.wordpress.com” with Donna’s “nofrakkingconsensus.com”, it’s perhaps easier to confuse it with “noconsensus.org”, which is also Donna’s domain.
Based on TallBloke’s update, it could mean simply that the local police decided they lack the tech skills to properly investigate and they called in some help from branches that do.
Laden said:
So the leader of a band of thieves isn’t a criminal? My,my, Shaw was right about America and Britain as being two people separated by a common language. Maybe the definition of what “his is” is like the definition of what “is is”.
That’s JeffId’s wordpress.com site. 🙂
I’m not surprised members of the press got confused about which sites were frozen.
BarryW-
I think most American’s would interpret that Laden quote as calling Roger a criminal.
“George Tobin (Comment #87655)
December 21st, 2011 at 12:22 pm
Based on TallBloke’s update, it could mean simply that the local police decided they lack the tech skills to properly investigate and they called in some help from branches that do”
If the course of an investigation of a crime the police discover another, serious or more serious crime, they spin off the investigation to another unit or force.
The Norfolk Constabulary have had a long time to go through all the saved emails they obtained from UEA.
Actually it is kind of humorous reading the original thread. Greg Laden is unrepentant, even after being informed by folks like James Annan that he was way off base. Probably changed his tune after talking to a lawyer. Still, I wonder how Laden could be so disconnected from reality as to imagine making that kind of accusation was OK. Had he not backed off, he would at least have had a reasonable case to claim an insanity defense. 🙂
“hardcore climate truthers”. Never heard that one before. Maybe sometime toto would be so kind as to explain what that description is supposed to mean.
It’s not self-referential?
SteveF
Assuming he ranted when he got home, a wife who is either a) diplomatic or b) really nervous and emotional might have gotten him to change too.
I got the impression that Laden may have mistaken the fact that sometimes a court will accept a certain statement as hyperbolic or metaphorical and that can sometimes persuade someone a statement was not libel, that they will in all cases. In this case, where Laden was blogging about 6 policeman serving a warrant in the course of an investigation of something that — at least hypothetically- could involve a criminal violation, it’s not unlikely a court would interpret the use of the wort “thief” or “criminal” as conveying their literal meaning.
In contrast: I don’t know what courts are going to take of the “State Penn” vs. “Penn State” usage, but it’s at least possible to interpret that as a joke rather than a literal accusation that someone is a criminal.
Lucia,
.
Or maybe his wife just said something like James Annan, and then added “I’m not going into bankruptcy over your stupidity… issue a retraction tomorrow!” These kinds of things can happen.
As Eli said there is likely something on the Emails to Tallbloke and friends that points to big fish, almost certainly involved in other serious stuff. The CRU hack is only the hook.
Eli, it doesn’t look like the security experts were particularly impressed with your arguments.
Hi, Lucia, testing access from the UK. Please delete after reading, as they say in top climate circles 😉
“If the course of an investigation of a crime the police discover another, serious or more serious crime, they spin off the investigation to another unit or force.”
Is there actually a British law that has been broken that an investigation can point to in order to say that they are investigating in order to find out who it was that broke such a law?
I’m sure that there is a law having to do with unauthorized computer access. But that would apply to any hacker breaking into any site and would not be relevant in any special way to climategate emails. At this point, I doubt that the Norfolk Constabulary is even able to say that there has been such an unauthorized access.
Eli’s No Such Agency musing is kinda funny. If they are on the hunt for climate terrorists which side gets the free gitmo vacation?
>I don’t know what courts are going to take of the “State Penn†vs. “Penn State†usage, but it’s at least possible to interpret that as a joke rather than a literal accusation that someone is a criminal.
The accused had the opportunity to claim that Mann is a public figure and his statement was ‘fair comment’ but instead chose to defend himself on the grounds that he was speaking the truth. It’s not clear if he has to prove that he spoke the truth, but that is the basis of his defense. By using this defense, he can also make discovery claims against Mann, and request documents that might prove his case. Given the reaction to Cuccinelli and others, I think this case will be dropped, but only after some humorous motions from Mann.
Since I hadn’t seen it here, here’s a copy of Mann’s legal complaint.
The entirety of the complaint is the statement made in an interview, where Ball says “For example, Michael Mann at Penn State should be in the State Pen, not Penn State.”
He’s also being sued in Vancouver, which has very different legal statutes than the US does when it comes to libel laws.
Go freedom of speech! (No really, I mean “go!”).
Lucia (#87665)
That’s what I get for making a last-second edit! 🙁 [I’m feeling pretty silly now.]
Yes of course you are right. The sites were confused even on a Climate Audit comment.
“Could be? I think it’s rather obvious that if Laden did not mean to accuse TallBloke of being a criminal then what he wrote clearly was misinterpreted by many. That list includes people who cannot in anyway be labeled “climate denialist†by the definition used by anyone, anywhere.”
Once again, an ambiguous statement is used to attack people. If something is linked, it can mean no more than that. There is a link. When you link up evidence, that evidence is not necessarily someone being criminal, but someone involved in the chain of something being done, as Tallbloke apparently is. I would expect James Annan to apologize myself.
The Schadenfreude around here is so thick you could cut it with a knife.
Can I ask an O/T question?
Does anyone familiar with Woodfortrees know how long it usually takes for the Giss and Hadcrut data to update? I seem to recall the satellite data being added the same day it was published.
Is the delay usual? Christmas? Impatience on my part?
Suppose the access and capacity to leak the climateosophy emails suggest broader access to the EAU system? Suppose there was something really choice on the system which may or may not have anything to do with climate? And suppose the “authorities” have no way to know whether it got loose?
Read synopsis for Charlie Varrick movie (1973) here:
http://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&sqi=2&ved=0CC8QFjAA&url=http%3A%2F%2Fwww.imdb.com%2Ftitle%2Ftt0069865%2F&ei=3zHzTti7GYbW0QGjz9S4DA&usg=AFQjCNEGetmX6bO_tGU_Cbaa37Ba8mZxoA
Bugs,
” I would expect James Annan to apologize myself.”
Thanks for that comment, I needed a good laugh.
Another thing these sites have in common that might cause confusion: they don’t know what “consensus” means.
It’s likely a reference to the more conspiracy-minded in the skeptic community. Think the average Glenn Beck/Rush Limbaugh listener. Or Alex Jones listener.
Laden’s lame “because it occurs to me that certain things could be misinterpreted…” is equivalent to the standard politician’s non-apology: “I’m sorry if I offended anyone…”
Actually, I think it is a reference to a common progressive meme where you declare anything too terribly uncomfortable to your worldview a conspiracy so that you don’t have to face the implications. Because as we all know, conspiracies don’t exist so anything unilaterally cast as a conspiracy must be ridiculous.
This works for maximum comedic effect if immeadiately afterward, you mention the mind numbing effects from those big oil or big tobacco operatives or hint at powerful shadowy forces (such as Eli does above) who are out to scuttle the agenda of the righteous activist.
Schadenfreude? Really? What horrible misfortune are we getting joy from? As far as I can tell nothing horribly happened to Laden. He wrote something stupid and people told him it was stupid. He posted a rather lame rewrite– and people are observing that. But he hasn’t been drawn and quartered, made into a pariah, lost his job, lost a limb, sued, lost money …. nothing. So… Schadenfreude? Really? Come on.
Of course conspiracies exist. We know this because we have uncovered evidence. The climate conspiracy theorists–much like the chemtrail theorists, the truthers and the birthers–have no evidence and generally make fools of themselves in discussions of reality. This is not to say that every skeptic is a conspiracy theorist, but there are a lot of them out there. Around here they are mostly named “Andrew.” 🙂
Boris,
Can you please give us your definition of a climate skeptic? I just want to know where you think I stand (in or out of the group).
Boris:
It could as easily be applied to the Michael Mann types and their “well-funded disinformation campaign” fevered imagination.
Related article by Bjørn Lomborg:
Going to Extremes
“How activists and the media distort climate reports to make global warming sound scarier than it is.”
Which side is the one with the well-funded disinformation campaign, and is it equivalent to being a “truther” to notice?
Carrick #87741,
I am relieved that the definition doesn’t highlight the dishonest mud-slingers, and the usual deception they engage in.
Glad to see that Carrick is highlighting Lomborg’s appeal to follow the IPCC report. If the argument is that the media and activists should pay more attention to the science, I’m all for that.
Of course, Lomborg’s conclusion’s are themselves at odds with the IPCC report. One example:
versus the IPCC’s conclusion on Africa:
(WGII, Chapter 9, p. 434)
and on freshwater resources:
(WGII Chapter 3, p. 175)
It’s hard to reconcile what Lomborg is writing with what the IPCC report says–and that’s the report he says that we should listen to. But he’s misrepresenting it and not citing where he’s getting his numbers from. Find that odd? I do.
Boris:
I didn’t have any trouble following Lomberg’s arguments nor did I see anything inconsistent between what the IPCC wrote and Lomberg’s characterization of it.
In particular “But more precipitation from global warming will most likely bring the actual number down to about 1.7 billion” is not inconsistent with “Climate change will aggravate the water stress currently faced by some countries, while some countries that currently do not experience water stress will become at risk of water stress (very high confidence).”
(Emphasis mine.)
If anything you’ve just demonstrated the alarmist tendency to feed on just the fear-driven and extremist rhetoric, at the expense of what the science actually says.
No Around here they are mostly named “Boris†😉
Okay, Carrick, where in the IPCC report is Lomborg’s claim? He is saying that the IPCC is what journalists should be using for reporting, so that is obviously what he would use. Please use your awesome non-alarmist pure as snow powers and find the source for his claim.
Also, way to ignore the chapter three quote. And more from the technical summary:
Even if there is some cite for his claim, why has he ignored the passages I’ve highlighted here? Seems like cherry picking to me.
After reading the BBC’s report on the “investigation” I am inclined to believe that the Norfolk police really are technologically incompetent.
The “investigation” appears to be driven by a vague presumption that anyone who questioned the doings in Phil Jones’ little nest of secrecy is a possible criminal.
Apparently no one in authority knows the actual means by which the emails were removed from the server(s) on which they resided. The fact that they are looking for logs on third-or fourth party user hard drives suggests that Team Clouseau may not be terribly familiar with server or internet architecture.
Also, given that there is no evidence of a hack from an external source, the crime that justifies this excursion into private homes appears to be the distribution of emails that Phil Jones thought he had successfully deleted or otherwise persuaded others to remove from lawfully required public scrutiny.
Boris:
The IPCC regional predictions are utter crap. Pielke Sr. and others have pointed out that the model’s skill at generating global average temperature change is pretty low but spinning that general estimate off into specific regional climate change outcomes is pure speculation, not science.
Lomborg argues against (a) the media and activist ideologue desire to exaggerate and spin disaster narratives while (b) commending the IPPC where they have made more measured and bettr grounded projections. Your view is that the failure to endorse the IPCC report in every detail somehow makes Lomborg a hypocrite for saying nice things about them.
Lomborg’s view that climate change is more likely to produce regional winners and losers (which could even be positive on balance) is more reasonable in every respect that the pathetic attempt to find uniform bad news (with “high confidence”) to sell the agenda the IPCC was created to support.
The IPCC narrative impact on water resources is of the same quality as similar dire predictions on hurricane frequency, malaria outbreaks and the rate of glacial melt and should be given the same degree of deference. The fact that Lomborg looked for elements he could praise rather than point out the bias and silliness of much else is an odd thing to attack.
Do you not get that Lomborg is saying we should listen to the IPCC report? Your criticisms of the IPCC are not important to the point I’m making. Lomborg complains about the media misrepresenting the IPCC report then he…misrepresents the IPCC report.
TallBloke on BBC TV on YouTube
http://www.youtube.com/watch?v=10lUiNvpOYM
“All IPCC regions show an overall net negative impact of climate change on water resources and freshwater ecosystems. Areas in which runoff is projected to decline are likely to face a reduction in the value of the services provided by water resources. The beneficial impacts of increased annual runoff in other areas is likely
to be tempered in some areas by negative effects of increased precipitation variability and seasonal runoff shifts on water supply, water quality and flood risks (see Figure TS.5) [3.4, 3.5].”
The devil is in the details. This excerpt means little unless you accept the word of the IPCC as your Bible.
Adrian–
Well, one thing police raids do is trigger real-media news stories!
Kenneth Fritsch:
Which obviously Lomberg is not doing, Boris’s claims to the otherwise, since if Lomberg were arguing to treat the IPCC as the Bible of the Later Day Global Warminologists :-P, he wouldn’t have started out by pointing out that *gasp* the IPCCC reports contain errors.
I think he would say, and I would agree that the IPCC is a good compendium of data and research papers, if you approach it wisely. And he certainly wouldn’t advocate approaching it with the type of complete naivety that Boris seems to be embracing here by assuming that every assessment statement is equally as valid as the less heavily spun data that accompany it.
As a researcher, I’ve always found the most reliable part of reports to by the analytic part–where the data are included—and the least reliable part the conclusions they draw from it.
Why this is the case is pretty obvious with a little reflection: People who are experts on e.g. hydrology aren’t experts on global models, (assuming the models can be trusted which is another problem). So when they try and merge what they are well trained on with things that they aren’t, they often muff it.
In this particular case, regarding Boris’s chapter 3 quotes—because I don’t buy the part about higher temperatures equals greater variability. It’s actually exactly the opposite. We see much larger diurnal temperature swings in winter, greater variability of weather, more dangerous extreme weather events etc. (There’s this nice Heartland Institute report on winter weather. Let’s hear it for “fossil fuel supported disinformation campaigns”, especially those that are just tables of facts along with supporting references.)
The tropics, where temperatures are warmest, have, except for a few days of the year, very little variability from day to day. Tahiti upcoming weather this week (12/23/2011) is 85/76 (±2 on the low end). Fairbanks upcoming weather by contrast is “+9°/0° 4°/-19° -15°/-24° -13°/-33°”. (And winter precipitation in Fairbanks is both disruptive and potentially life threatening, except in very rare exceptions that is not the case when precipitation events occur in Tahiti.)
And I would claim the underlying physics for this are well established. Moist air has a lower lapse rater than dry air. That means the amount of convective energy available is lower, unless you are in a mixed region, like the central plain states in the US in spring, where high level air is coming from a cold region, and low level air from the Gulf of Mexico.
(Another news flash here, in spite of being in the most violent tornado belt in the entire world, the central US also happens to be one of the most productive regions agriculturally in the world.)
Although it’s a mantra of the CAGW community that warmer weather equals greater variability, people who are experts on fresh water haven’t necessarily sat down and judged the credibility of the claims regarding the question of increased temperatures = greater variability.
Regarding flooding, runoff and water quality, I thought that was just politically-spun dreck. Lomberg ironically even addressed some of those issues in his short commentary.
You can look at the rainfall in the US. I’m pretty sure the change in rainfall from the central plains to the SE US is much greater than the change expected by any reasonable global warming scenario.
This figure shows mean rainfall increase together with the 95% variability bounds. Interestingly the variability in rainfall does not reflect the increases suggested by Boris’s quoted text.
Where I live we get 48″ of rain, and yet we aren’t at a higher flood risk than say people in the Missouri river basin. The reasons are straightforward, and they are ones addressed by Bjorn Lomberg.
We’ve reestablished flood plains, a water control system, and even alternative water ways to reduce the load from heavily spring flooding.
We had a terrible devastating flood in 1927, which was transformational in how the US addressed water control issues. This book gives a good summary of it and changes in land usage up to it. For anybody who wants to do something besides just bloviate on this issues, this should be required reading.
By changes in land usage, I mean of course things like the encroachment on the river by communities, the loss of flood lands, and the ever higher river levees that acted to reduce the capacity of the river to expand from its banks. In other words, we had basically reverse adapted ourselves into a circumstance where flooding would become more frequent and more devastating with each abnormal rain period.
The modern flood system in the eastern US has to a large degree ameliorated the risk of flooding, oddly again this syncs well with Lomberg’s comments. This is a case where the modest increased risks associated with greater precipitation can easily be handled by improved adaptive strategies.
I think it’s also the case that many countries currently are at a heightened risk of flooding not due to changes in precipitation patterns from GW but due to changes in land usages that brings them to a higher risk level.
Boris: “This is not to say that every skeptic is a conspiracy theorist, but there are a lot of them out there.”
So, Boris, let’s say that a few scientists, Jones, Mann, Trenberth, etc. work for a common cause that they all understand – as is clearly shown by the emails. Is what they are doing a conspiracy? If I point out the evidence that they are working for a common cause, am I a conspiracy theorist? Or has the use of the term “conspiracty theory” simply become a tool for smearing people in place of actually dealing with the points that they are making? I almost never see skeptics declare that there is a conspiracy. But I constantly hear warmers declare that skeptics are declaring a conspiracy? Would you call the Republican party or the Democratic party a conspiracy? Do those parties work to a more or less common agenda. If someone points out the common elements that Republicans or Democrats are working towards, are they conspiracy theorists?
Tilo:
I think this falls under the rubric of “conspiracy by consent”. It explains how minority groups remained locked out of certain types of jobs, without requiring secret boardroom meetings to ensure it doesn’t happen. Winks, nods and deleted emails work just as well here.
Carrick, I agree with what you say here and I think if all sides of the AGW issue would see the IPCC from this veiwpoint we would avoid the waste of time conversations about fraud claims on the IPCC on one side and using it like a Bible on the other.
I can learn a lot from reading bad climate science papers and/or papers that overly extent the evidence presented to the conlusions. Mann (08) is a prime example, in that a close read shows the problems with the Schweingruber, Instrumental and Tijlander series and at the same time clearly shows, and at least noting, the divergence problems with dendro and non dendro series. The paper also showed that the hockey sticks of the earlier Mann papers have changed significantly to something less HSish. Items from which we can learn can be overlooked if too much personality gets in the way.
“I think he would say, and I would agree that the IPCC is a good compendium of data and research papers, if you approach it wisely. And he certainly wouldn’t advocate approaching it with the type of complete naivety that Boris seems to be embracing here by assuming that every assessment statement is equally as valid as the less heavily spun data that accompany it.
As a researcher, I’ve always found the most reliable part of reports to by the analytic part–where the data are included—and the least reliable part the conclusions they draw from it.”
Re: Kenneth Fritsch #87785,
An Anon left me a link to a 2011 defense of Mann08 by Schmidt, Mann, and Rutherford, attempting to rebut McShayne and Wyner (2011). Discussion of: “A statistical analysis of multiple temperature proxies: Are reconstructions of surface temperatures over the last 1000 years reliable?”
It’s embarrassing to read. These guys just can’t see what they can’t see. E.g. Schweingruber, Divergence. They can’t even be forthright about what they kinda sorta can see.
Emphasis mine. The italicized words are misleading, to the extent that they aren’t flat-out wrong.
This is followed by the claim that their findings confirm IPCC AR4 assertions. The turgid prose:
This first grudging acknowledgement of the existence of such a thing as “posterior probabilities” is progress, I suppose. Since the entire multiproxy reconstruction enterprise is an unwitting exercise in the pitfalls of post hoc analysis. Pitfalls which turn their quantitative estimations — 67%, 99%, 90% — into mush.
Perhaps the editors and referees at The Annals of Applied Statistics could benefit from some lessons in… statistics?
Boris: try repeating your point another four or five times. You never know, at some point the reflexive IPCC-bashing might attenuate a little and one of the posters might actually read what you’re writing (as in, realizing that your point is utterly independent of how much one trusts the IPCC).
.
.
I missed this particular talking point. What’s it about?
Schweingruber has a number of proxies developed. I’m not sure what specific point AMac is making. With Yamal, McIntyre found nearby Schweingruber sites that were not combined, that showed a different result. Briffa had done a similar combination at a different site.
toto:
The part about it being a canonical text, like the Bible, but for AGW birthers?
Or maybe just be hopelessly naive about the “lack” of bias and policy influence in the summarizations?
(It might be useful if you actually repeated once the point you thought Boris should repeat five times. Us Luddites are left guessing.)
Amac, perhaps you could expand on the Schweingruber issue?
Nobody here appears to know about it, so toto seems to have some other meaning of the word “talking point” than the normal one. (I think he’s using it in the sense “I brush off this argument without ever hearing it by dismissing it as a talking point.”)
Sure:
Boris is not talking about the accuracy (or lack thereof) of the IPCC report. He’s talking about the accuracy (or lack thereof) of Lomborg about the IPCC report. Whether the IPCC report is correct or not is irrelevant to the point.
In the end, Boris misread or misrepresented Lomberg, who did not misread the IPCC, but merely disagreed with some of its conclusions.
The irony is rich here.
Well, it was mentioned in Kenneth’s message as well. Hence my hunch that it is a talking point that I have somehow missed.
Actually, it was more like “I’d like to know what it is before I dismiss it (or not)”. Some people like to do things in that order 🙂
It just seemed a bit odd to start off with the assumption it was a talking point, especially coming from Amac.
“Talking point” sounds like something that would generated in the bowels of the ginormous nefarious well-funded climate disinformation network.
No. Lomborg specifically says he agrees with the IPCC report and that journalists should base their reporting on the IPCC report. He never says anywhere that he disagrees with the IPCC (other than the obvious errors that have been pointed out.)
Quoting:
And this:
Boris:
Since he specifically says this, you can provide the exact quote instead of paraphrasing, right?
This certainly doesn’t say that:
Fairly reliable means something entirely different than we should treat it as a bible, or even just take everything it says on face value.
Beyond that, you’re just cherry picking sentences to try and change the plain meaning of the essay. You simply don’t take selected sentences from an essay without considering the context in which they were written. Not and retain any credibility with people who do understand the mechanics of reading comprehension.
It’s as simple as this: If Bjorn thought the IPCC report was error free or was making the buffoon-like suggestion it should be read uncritically, he wouldn’t have started with examples where errors were made. [It’s also easy to work out the make topic of the essay was the sensationalization of climate science by the main stream media and activists: “How activists and the media distort climate reports to make global warming sound scarier than it is,” and not that “Lomborg specifically agrees with the IPCC report and that journalists should base their reporting on the IPCC.”]
Like with any source, one must be an “intelligent consumer.” And academic like Bjorn would certainly be aware of that. I’ll note that this figure I generated above showing projected future precipitation under the SRES A1B scenario was made using the IPCC model output.
Is it inconsistent to use part of a report and not rely on another?
No. Not at all.
I’ll end by repeat my own view on this,
> Amac, perhaps you could expand on the Schweingruber issue? Nobody here appears to know about it
Sorry for the delayed response — Christmas party.
I recalled Kenneth Fritsch noting that proxies put together by Prof. Schweingruber and used in Mann08 had actually been composited — that is, turned into a mix of tree-ring and instrumental temperature record. Seems absurd; perhaps I was gullible and need to eat my words?
Well, after a few minutes Googling, it does seem that this is another instance of non-intuitive treatment of proxy information in Mann08. There doesn’t seem to be any procedure that’s too silly for these authors to employ.
An Oct. 5, 2008 post at the Air Vent provides an explanation of the compositing, as well as a visual depiction of its effects. MXD Tree Removal Service (Choppin Wood off A Hockey Stick).
Jeff Id quotes Steve McIntyre (here) quoting Tim Osborn and the authors of Rutherford05 on this Schweingruber MXD data as follows:
The website currently appears to include this text:
But are we talking about the same series that Mann08 used? Jeff Id has looked more closely at Mann08 than anyone else, to my knowledge, and he thinks so. tAV has time-series graphs of these particular Schweingruber latewood MXD series, showing the difference caused by the post-1960 compositing. In the Oct. 3, 2008 post here, with further discussion at the already linked post from two days later. “Mann08 used 95 of these series in the final 484 or 20% of his data,” Jeff says.
So the exercise is to find a correlation between temperature and temperature, then use the result as validation of the temperature-sensing utility of MXD tree-ring data series. That’s a bizarre procedure.
If these claims have been rebutted by defenders of Mann08, please add a link. I didn’t find any. I’ll still eat my words if the evidence points convincingly in another direction.
Thanks, Amac. I’ll have to sit down in the morning over some coffee and read through your comments.
By the way, regarding the Mann 08 proxies, I think I understand now Mann’s justification for using so many that are obviously not in “temperature limited growth areas”.
I believe these are actually precipitation proxies, and the assumption is there is a positive correlation between temperature and precipitation (but not necessarily “local” temperature, hence the necessity for the use of a “climate field reconstruction” when using these proxies).
Regarding Schweingruber, I found those totally unsatisfying, and wouldn’t use them in a reconstruction were I ever daft enough to attempt one.
Carrick [@87800]
Given the time of year, I would suggest you should ask Santa for a copy of Donna Laframboise’s recent book on the IPCC. It might cause you to change your rather benign views of the outfit, its practices and the reliability of its “science”.
Lomborg -like the proverbial armadillo- has become the victim of his own need to fairly walk the middel of the road…
Give me a break, Carrick. I’ve actually presented evidence that Lomborg thinks the IPCC is a good source of information. Your response is “out of context, meow chow.” without even bothering to supply the supposed missing context. Then you supply your subjective view of the IPCC to conclude that:
So let’s review: Boris uses actual quotes from Lomborg and gets accused of cherry picking and changing context, and then Carrick comes along and uses his imagination to suppose what fantasy-Lomborg really meant.
Then, along with your helpful suggestions for reading comprehension you suggest that I think Lomborg thinks the IPCC report is a “Bible” and that it should be read “uncritically.” Instead of conjuring up fantasies of straw-Boris, maybe you should pay close attention to what I’m saying. Here it is in numbered list form:
1. Lomborg criticizes the media and “activists” for exaggerating the risks of climate change.
2. Lomborg suggests in example after example that the IPCC report is a good source, and in fact the source the media and “activists” should be using.
3. Lomborg himself minimizes the risks of climate change by selectively quoting the IPCC report.
Of course, you are stuck on number 2 and don’t believe it’s true. For which you have provided zero evidence from the essay–except the fact that Lomborg points out errors and thus can’t think the IPCC is a good source even though he says so in plain language.
I’m sorry you accidentally posted a source that says the IPCC is doing a good job. Just admit it and we can move on to more interesting subjects.
“Fairly reliable” vs “good source”.
Meh.
Boris, take that one to a holiday party and see how much interest you can generate.
As a skeptic, I’m perfectly comfortable having people read the entire piece by Lomborg and coming to their own conclusions.
Obviously, from the comments at Slate, most of the “correct thinkers” weren’t too happy with it, which I take as a good thing.
AMac,
After reading the Jeff Id post I CAN’T BELIEVE MANN ISN’T IN JAIL!!!
But seriously, instrumental temps were actually part of the validation in Mann08, through the Lutebacker proxies:
I’m not sure if Shweingruber was part of the Luterbacher proxies? (I’m thinking no.) Jeff Id has not provided any evidence that the instrumental temperatures were used for validation. If there were evidence for that–and it was shown to change the results–then that would be interesting.
Ah, thanks for the background, Boris. I’ll look into it in a bit. Humble pie coming up, or not… stay tuned.
Boris, you do yourself no credit with this over-the-top interpretation of Lomborg’s article that you are trying to foist on us coupled with what amounts to cheap rhetorical tricks designed to discredit what you obviously see as an “enemy”.
The evidence that your interpretation is wrong is contained within the article itself–indeed the apparent contradictions that you point to either suggests that Lomberg is a complete id*ot, or your criticisms are just wet.
If the point of the essay is we are supposed to take every word of the IPCC literally (which is the only way anything you say makes any sense), what do you suppose is the point of leading it off with “How activists and the media distort climate reports to make global warming sound scarier than it is.” Then following it with “Such claims merit close scrutiny. A 2007 report by the U.N. Intergovernmental Panel on Climate Change received considerable media attention. But it was discovered two years later that some of the IPCC report’s key claims—for example, that global warming would cause the immense Himalayan glaciers to disappear by 2035 or halve African crop yields by 2020—were based on statements made in appeals by environmentalist organizations and were backed by little or no evidence.”
Is this not a blatant warning that one can’t interpret the IPCC report as if it had “biblical accuracy”?
As to point 2, I think my version is much closer to what an academic would have said. But generally, regarding point 2:
“2. Lomborg suggests in example after example that the IPCC report is a good source, and in fact the source the media and “activists†should be using.”
Generally I agree with this, the only stipulation is that you be “an intelligent consumer” when reading the information provided by the IPCC and recognize that the parts that are more analytic (e.g., data or theory driven) are going to be less prone to personal bias than parts that are summarizations of the literature.
Given that Lomberg is an academic, my expectation is he will use an academic approach to interpretation of any scientific work, which is what I described above, not your uncritical regurgitation of everything you are told. Given that you are not an academic, I’m not surprised this “literalist” interpretation is the one you advocate.
In the end, this article was not on the academic method, you cannot claim it was, it was clearly labeled to be on “How activists and the media distort climate reports to make global warming sound scarier than it is.”
I think you can fault him if you choose for not being clear about how the media should interpret the IPCC and other scientific literature… but that would be a separate essay wouldn’t it?
Regarding three: 3. Lomborg himself minimizes the risks of climate change by selectively quoting the IPCC report.
No, he makes his own assessment based on the data from the IPCC and comes to different conclusions than the policy driven portions of it.
It is generally doing a good job and I do rely on it for information. I don’t find some of it’s conclusions credible, but that’s no different, than any other academic research.
As a layman, I don’t expect the nuances of academic methodology to come naturally to you. (You were what, a polysci major? Hardly a field that is particularly rigorous in the academic method.)
But if you’re going to put forward criticisms of people who are academics, you will just make yourself look foolish, as you have done here, by what is obviously reactive thinking on your part towards somebody who is not “helping the cause”.
Maybe you should write him and ask him if he would write another essay on “being an intelligent consumer of scientific research”. Some how, I think you’d be happier if he and everybody else who doesn’t support the cause is either censored, or when censorship fails, ridiculed.
Because that’s how you types roll.
tetris, Laframboise’s book is not likely to change my view, because I already knew how the IPCC working groups were put together. Nor is she saying that the results are all flawed, so much as there are flaws in how the process worked.
The 3 proxy series in question for Mann (08) are important to me for the following reason:
Mann(08) uses a selection/screening process whereby the correlation of the proxy response and temperature in the instrumental period are used. In that exercise Mann(08) notes that obtaining those correlations by chance would be statistically improbable. He uses a model for the series in questions.
In the series that are selected in the screening process, 3 standout as proxies that should have never been used and these proxies amount to a large number of data points.
Lutebacher is a long series but becomes an instrumental series in the period of interest here, i.e. the instrumental period.
Schweingruber is the MXD series that in Mann (08) is cut off at 1960 (because of divergence) and replaced with something that, from the paper, is not well explained.
The Tiljander series has been thoroughly discussed before and is used arbitrarily upside down.
If you remove these series from the screening process and use a model with long term persistence, one can easily obtain the correlations of proxies to instrumental temperatures by chance.
All these issues have been well covered by Jeff ID and Steve M at CA. You just cannot legitimately defend what was done in Mann (08) with those three proxies and the claims about the correlations not occurring by chance. I am more concerned about the lack of attention to these matters by the climate science community than I am that the paper was published.
A careful read of the paper indicates that: yes, we can obtain the measured correlations between proxies and the instrumental record by pure chance and on the other side that proxies/reconstructions clearly diverge and it is not unique to the dendro series. Finally it is apparent that the Mann reconstructions get less hockey stickish as time passes. This is all good information.
Kenneth, I’m to the point where I think the only way the proxy reconstruction can be trusted now is if a completely independent reconstruction (ala BEST) were performed with minimal interaction with the existing cadre of research groups.
Most of us have known for while that the “professional” global mean temperature time series can be reliably replicated. Proxy reconstruction seems like a much harder problem, partly because the data aren’t nearly as well archived here, and because the available meta data is much more varied in quality. (With GHCN, we all start with the same central archive, here the various data sets seem strewn around the web, with different versions appearing in different places.)
Mann in particular, has this really nasty habit of inventing data to fill in gaps, without any explanation or even adequate justification (in many case without any notice) that he is doing so. And a nasty habit of creative statistical processing, which would be OK if he actually knew how to put his own creative methods through a testing phase before using them in published papers.
[Edit: This comment was written prior to reading Kenneth Fritsch #87813 or Carrick #87814.]
Boris —
Flipping from Lomberg back to Schweingruber:
In Comment #87808, you offered text from page 2 of Mann08’s SI. Here’s a more complete quote:
While I look into this issue, I have a question for you.
In your opinion, is the way that Mann08’s authors handled the composited Schweingruber proxy issue an instance of Good Scientific Practice? Is it an example of Acceptable Scientific Practice?
At this early stage, I already have my answers: no, and probably not.
Here’s a reason why: “separate validation experiments and reconstructions back to A.D. 1500 were also performed without using these predictors.”
In other words: “We’ll perform our main validation experiments and reconstructions using proxies that we know are unsuitable, but we’ll also exclude certain unsuited proxies, and also present these alternate results — say, somewhere in the S.I.”
This brings to mind Wall Street sharks burying key investment information in the prospectus’ footnotes. You’d best read through the whole thing, with great care. More fool you, should you fail to discover the warnings buried in the small print. And heed them.
Tilo makes a valid point: The believers now add “conspiracy theory” to ‘denialst’ as words to stop conversation and to demean those they wish to silence.
Carrick #87814 —
As I see it, there are two big issues with these multiproxy reconstructions.
The first is nicely illustrated by Mann08, and now again by Schmidt11 (Comment #87786). The folks doing this work are in way over their heads, and they’ve become committed to their mistakes for reasons of ideology, pride, career, or whatever else. Norm Wolpert described this phenomenon in a different context — the deadly and persistent mistakes of the medical establishment in approaching breast cancer therapy.
Sound familiar?
The second is that proxy selection and reconstruction-building are classic instances of post hoc analysis (Wikipedia). When post hoc conditions apply, one can always still calculate statistical measures. Alas, they don’t mean what careless authors and unwary editors, reviewers, and readers think they mean.
It would not necessarily be impossible to generate climate recons with statistical validity. But it would be difficult, and the error bars would be wide. The current generation of thought leaders has gorged for so long at the proxy table (1,209 delicious entrees!) that the notion of such a crash diet-and-exercise program is obviously very unappealing.
Why do you recreate this straw man after I have repeatedly pointed out the inaccuracy of your interpretation of my points?
Why are you putting your own erroneous interpretations in quotation marks here? I’m sure I can find any number of consensus scientists who would approve of the IPCC report but warn of the errors inside it. That you think this is some winning point is telling.
Yes, and my point is that Lomborg distorts the exact same reports. So it’s okay for him to distort the report because he doesn’t believe them, even though he says he does trust them, but you just KNOW he doesn’t, deep down?
The fact that you need to continually refer to my academic background in this discussion just reinforces the fact that you are having trouble with the logical points of the argument. Is this one of the “cheap rhetorical tricks” you are bemoaning? Oh, wait the rhetorical trick I used was direct quotation of the source in question. Silly non-academic!
More fantasizing from Carrick! I’m sorry I missed on the rigorous super-academic courses of daydremaing and making shit up that you apparently passed with distinction.
Oh, I left the best part for last:
Maybe you are thinking of Bjorn Lomborg, you know the guy who got his PhD in political science from the University of Copenhagen?
Boris, if you think my comments are strawmen, there isn’t any further purpose to this converstation.
I’ll write this off as you being too stubborn to admit you attacked Lomberg on political grounds and to stubborn to realize just how hollow and transparent your reasoning is here.
Lomberg is in the school of business at his current academic position by the way. He deals with substantive real-world problems. I would have expected you to know that.
I see you are really thin skinned about your weak credentials. Your problem.
Was there a significant difference in skill?
Not at all. Mention it anytime you think you need the help.
The point of an academic background is also missed on Boris. Not to be unexpected, not because he’s not an academic, but because he’s at heart anti-intellectual in his worldview.
What it means is people who are academics share a common set of methodologies for how we treat a problem, and a common attitude about how we treat other scientific literature, and in general a common intellectual culture. So certain things don’t need to be explained to a fellow academic, because “we just get it”.
Lomborg’s essay would be a “fail” if the purpose of it were to explain how (in the academic view) one should read the IPCC document. As has been pointed out ad nauseum (and hence this is my last attempt on this), the obvious and stated purpose of the Lomborg article was “How activists and the media distort climate reports to make global warming sound scarier than it is.â€
Had it been about something else, it should have been written differently.
Re: Boris (Dec 24 12:13)
> Was there a significant difference in skill?
No, the SI says there was not.
It’s much better science to do things the right way. It’s also better to keep things as simple as possible (there’s even an acronym to remind folks of this).
The Tiljander saga abundantly demonstrates the pitfalls of presenting a complex, multi-part experimental strategy where some parts are claimed to be analyzed correctly over here, whereas other aspects are said to be done properly over there… and so on.
One effect of all this complexification has been to befuddle many scientifically-literate people who have commented on Mann08. The extent to which the authors themselves are confused, I couldn’t say. Though it doesn’t seem unlikely.
In your opinion, is the way that Mann08′s authors handled the composited Schweingruber proxy issue an instance of Good Scientific Practice? Is it an example of Acceptable Scientific Practice?
Looking at individual proxies is most revealing and comparing the hills and valleys in those series to series generated to show long term persistence allows one to appreciate how easily it is to obtain a rising proxy response in a period the length of the instrumental record and at the end of the series. One can also obtain as readily a downward sloping response at the end but then we might expect that those proxies and their responses would not see the light of day in a reconstruction or be written off to divergence. (This observation also points for the need of an a priori and physically based set of criteria for selecting temperature proxies). Further, if one throws in the divergence phenomena that often is obscured in temperature reconstruction papers and one recognizes that the instrumental record must be tacked onto the end of the series (as if it the equivalence of the instrumental and proxy record had been previously demonstrated) to obtain the hockey stick appearance, the weakness of the case for proxies as thermometers can be seen.
I do think that there is some long term persistence in proxy responses that might not be in the long term temperature record and particularly in the late instrumental where we know there is a deterministic trend. Proxy responses can also show to some degree a qualitative coherence in response to something common to the proxies that changes over time, but these responses are not quantitative. To the extent we see coherence in the instrumental period between temperature and proxy response, at least qualitatively, would indicate that there is some response of proxies to temperatures or something co-varying with temperature. That observation does not, however, show that the response is sufficiently quantitative to allow proxies to be used as thermometers and particularly in the past.
AMac, thanks for the comments on the hockey stick.
This whole problem seems fantastically complicated compared to the usual problem of temperature reconstruction from instrumental records (which itself is already a challenge).
Regarding this question:
It’s certainly not the way I would have handled it, but I’ve been told I have a lot higher standards than a lot of people do. I certainly wouldn’t start everything in the confrontational style that Mann handles all of his criticism either, or posture myself thee way he attempts to by painting himself as an authoritative figure lofting above us unwashed masses.
Kenneth Fritsch:
This seems like there is a use for tree rings as their own measure of climate change:
It’s certainly not a one-to-one relationship between temperature, or precipitation and tree-ring growth, and certainly not a functional relationship between the various met variables and the proxy response (“many to one relationship”). But it still seems to have some use in quantifying changes in climate over time.
The simplest and most obvious interpretation of tree-ring proxy would be something like “equitability of climate for tree-ring growth”.
AMac, the one thing I absolutely would never do is use the 1850-1900 period of the instrumental temperature record for validation…
At least not without a huge amount of work to make sure that the very spotty temperature records available during that period were reliable enough to be used for validation purposes.
Anthony Watts posted some emails in which further insider reservations about hockey stick reconstructions appear. At this point, defense of Mann is simply a gesture of warmist ideological loyalty rather than serious discourse.
Carrick:
I started to respond to Boris earlier but I realized that he missed the point I made about how he missed the Lomborg point and recognized that engaging him merely invited more mindlessness.
Boris’ point is that a positive acknowledgement of any part of any IPCC report (offered solely as a contrast to the kind of exaggerated climate porn that media types, bugs and Boris love) must somehow be the logical equivalent of a blanket endorsement of the IPCC.
Given that Lomborg has an extensive collection of writings in which he critiques the alarmist project, only an idiot would interpret the article as an invitation to make a blanket endorsement of the IPCC. The fact that most of the posted comments were predictively kneejerk negative attacks on Lomborg by Alarmist True Believers, it appears that across the entire political spectrum Boris stands alone as the one guy who did not get the point.
Carrick — Yep, 1850-1900 is one more problem.
As a matter of interest, the derived CRUTEM3v temperature anomaly for the 5 degree by 5 degree gridcell that covers Lake Korttajarvi in southern Finland shows the expected rising trend over the Mann08 period of calibration and validation. However, the temperature record of the closest weather station with a long history does not — inspection of Tiljander03’s Figure 2 with the Mark I Eyeball reveals no particular trend, 1880-1995.
Which temperature history is more relevant to the
fourtwo Tiljanderproxiesdata series? Well, the more important question is, how should one anticipate and resolve such issues so as not to wind up inadvertently performing a post hoc analysis?The continuing failure of the authors of papers such as Mann08, Mann09, Kaufman09, and Schmidt11 to address such issues is powerful evidence that these folks are in over their heads. In my opinion.
Editors, peer-reviewers, and the broader scientific community are supposed to promote a return to good practices, when things go off the rails like this. But that’s not happening. Again, my opinion.
C’mon everybody. Let’s play nicely. After all, lurkers (our secret target audience) seem to like it better that way.
Get personal, and our hostess might be tempted to leave a lump of CO2-emitting black carbon in your stocking!
Off to do Eve-y things now, so Merry Christmas Carrick, Boris, George, Kenneth, Hunter, and everybody farther back on this thread!
Personally, I would not use cru gridded temps.
Much better would be to take the proxy location and do
a indepentent temperature estimate using one of the better
methods, tamino, stokes.
That might be interesting.
George:
I’m sorry that I continue on sometimes, George. It feels at times like tilting at windmills.
Yep. This is part of the “in context†sensible reading interpretation I was appealing for. I don’t think it’s at all reasonable to criticize Lomberg for not addressing issues that weren’t relevant to his essay (like why he obviously disagrees with some parts of the IPCC AR4 assessment on fresh water), though I would admit it would be interesting to hear his reasoning.
AMac thinks we should be nice, so I’ll pretend for tonight, for his sake. >:-}
Steven Mosher:
There’s some real contorted reasoning they appear to use on this point.
The problem is that local temperature often completely fails to correlate with the so-called temperature proxy, but does better when correlated against regional scale temperature, what they call in the field “climate field reconstruction”.
(I’ve not seen a systematic comparison of the two… this begs to be done independently, meticulously and with surgical skill on the part of the researcher.)
I think this approach makes some sense when the proxy is a precipitation proxy that you are trying to correlate with temperature.
I think though the first order of business needs to be a “good” set of well-characterized and understood proxies. No more taking proxies concluded to measuring CO2 and use them as temperature for example.
I’m afraid if we tried to do that, we wouldn’t end up with many proxies left.
Like the tree in siberia knows the weather in moscow.
merry christmas
Here is the McShane & Weaver’s Rejoiner in case people have missed it.
Steven Mosher:
If you restrict yourself to trees in Siberia the same latitudinal band as Moscow, and only calibrate using very long-period (> 30 years) variability, it might actually be true.
Kenneth:
Hmmm… Here’s Mann et al. 2008 SI Text:
Sounds pretty explicit to me.
toto, here is the challenge for you, if you think it gives enough information: Successfully replicate what he did without writing to him for details.
“Because of the evidence for loss of temperature sensitivity after ~1960 (1), MXD data were eliminated for the post-1960 interval. The RegEM algorithm of Schneider (9) was used to estimate missing values for proxy series terminating before the 1995 calibration interval endpoint, based on their mutual covariance with the other available proxy data over the full 1850–1995 calibration interval. No instrumental or historical (i.e., Luterbacher et al.) data were used in this procedure.”
The SI is not the paper. But I’ll not engage in a sidetrack discussion and only reiterate that cutting off a proxy series and replacing it with some averaging process from parts of the remainder of the proxy series used is not good science or statistics for that matter.
Also, toto, please note that if you dig into the proxy data that was used in Mann(08) you will not be able to see what the replacement of the post 1960 period of the Schweingruber series was, i.e. you will not find the replacement data. Neither will you be able to determine from the paper and SI what the Schweingruber series data looked like after the 1960 period. Neglecting to show that information is even more evidence of poor science.
Not only is the SI not the paper, because it is not archived at the journal, it can disappear or be modified at any point in time.
Mann’s SIs have been known to go through modifications over time, which is especially mysterious, since the SI is supposed to represent supplementary information related to the original paper.
If you want to publish a corrigendum-like extension of you paper, it needs to be so labeled and kept separate from the original SI, whose contents should be left intact and not changed over time. While the SI is useful, it cannot be viewed, as handled by most journals, with the same providence as the original publication.
I think this shapeshifting ability of Mann once again reflects the poor quality of his general scholarship, and unsurprisingly to me, Mann has abused the fact that he modified his SI to excoriate people who were criticizing his original published work by pointing to a version of his SI that was apparently modified after the criticism had been received.
Re: toto #87853,
> “MXD data were eliminated for the post-1960 interval. The RegEM algorithm of Schneider (9) was used to estimate missing values for proxy series terminating before the 1995 calibration interval endpoint, based on their mutual covariance with the other available proxy data over the full 1850–1995 calibration interval.”
I’m familiar with Mann08’s use of RegEM-based estimation of missing proxy values, because these authors used this procedure for the
fourtwo Tiljanderproxiesdata series, for the years 1986-1995. Tiljander03 only reported up through 1985, so Mann08 used a make-your-own-data algorithm to extrapolate the data for the remainder of the calibration/validation period.“Naw, they wouldn’t simply make their own! And use the invented time points for calibration and validation purposes!”
Well, they would, and they did. More than that, the editors and reviewers green-lighted this practice, to the extent that they were paying attention.
This isn’t Mann et al. pulling a fast one–they described what they did. To them, this is quite okay.
Toto, if I could ask you a variation on what I asked Boris upthread:
In your opinion, is the way that Mann08′s authors handled RegEM-based extrapolation of proxies in the calibration/validation period an instance of Good Scientific Practice? Is it an example of Acceptable Scientific Practice?
AMac, why do you think they extended the proxies with infilled data? The answer to that question might help you out.
Got to get back to glazing the ham. No, that is not a euphemism.
Lubos Motl to a type like Boris: “I leave it up to you whether you learn just the hard data or also their bizarre interpretation”
Boris:
Actually, they didn’t extend the proxies. They truncated them, and replaced that part of the series with what you are calling “infilled” data.
The term for this in science is called “fudging your data.”
(It’s part of why I advocate against using Mann’s versions of the MXD series.)
carrick – reason is wasted on Boris…he reminds me of the Moonies…once a normal person but he attended a Moonie service and his brain was washed
You guys… “fudging”? “make your own“? For a linear combination of the other data series? Seriously?
Amac:
You bet they did. And on top of that no one bothered to submit even a comment on this supposedly fraudulent method. Doesn’t that make you pause for even a second?
Carrick: couldn’t agree more with you about the annoying “SI is the new Methods” trend. Please direct your ire to where it belongs – namely to the journals and their ridiculous page limits, which basically force authors to shift any kind of precise description of methods and secondary tests to the SI. At least PNAS is nice enough to format and edit the SI, as opposed to just dumping the authors’ PDF.
toto, Mann didn’t like the MXD series past 1960 so he invented his own.
The fake astonishment on your part that somebody would think this is an example of “fudging” is duely noted.
Toto:
I agree about what you say here to a point.
However, there are issues related to responsible handling of your data and your own documentation as well. I think it’s a crock that the authors should be required to retain the SI and be responsible for it.
That said, it’s inappropriate for me or any other researcher to edit or modify any notes or documentation that I have that support a published manuscript, whether that documentation is considered to be “published” or not, without leaving an audit-able trail of those changes.
And it certainly is not an example of responsible conduct of research to modify what is supposed to be supporting documentation for a manuscript, without having some sort of “change notes” that alerts the readers to changes post publication of the manuscript.
Enough said, I shouldn’t have to teach ethical behavior here.
By the way, toto, you used the phrase “supposedly fraudulent method”, not anybody else.
Mann freely admits to the manipulation of the data in his series, so whatever it is, it isn’t fraud.
However he does so in the SI, which may or may not have been available to the reviewer: You are making two assumptions here: 1) that it would have been and 2) that it has remained unmodified between then and the “current” version on Mann’s website.
As somebody who peer reviews research papers, I know as a fact that the SI is not always made available to the reviewer, and we certainly know as a fact that Mann has modified his SIs after publication in the past.
Carrick #87871 —
To clarify: the Mann08 S.I. has been continuously available at PNAS in its original form, on this page. (The article itself is now freely available under PNAS’ open access policy.)
The authors of Mann08 and their supporters caused the rise of this issue by citing figures that are — seemingly — from the S.I., linking URLs at Prof. Mann’s Penn State website, instead of at pnas.org. At least one of those PDF’ed figures — Fig. S8a — had had substantial alterations and corrections made to its two “gray” versions at PSU. The confusion came from the “stealth” manner in which the changes were made and then introduced into discussions.
Boris (Comment #87859) —
> AMac, why do you think [Mann08’s authors] extended the proxies with infilled data?
I can offer an answer that’s rather obvious. If you’re angling for something subtle or ironic, you’ll likely have to try again.
As far as I can see, they extrapolated certain proxies because (1) they believed this was an appropriate step to take, and (2) doing so provided support for the thrust of the paper. A thrust that these authors thought at the time contributed significantly to our understanding of the climate of the past 2,000 years. (They haven’t changed their thinking on this.)
I’ve already reviewed why I think analogous practices are not considered acceptable in other subdisciplines, i.e. why paleoclimatology is an outlier. It would be boring to repeat myself; please check upthread.
Also: In your opinion, is the way that Mann08′s authors handled the composited Schweingruber proxy issue an instance of Good Scientific Practice? Is it an example of Acceptable Scientific Practice?
AMac:
That is news to me thanks.
I’ll just point out we still don’t know whether the reviewers were aware of the RegEM extrapolation (it’s not technically just an infill, since you are padding off the end of the series up until the year 1995), it’s difficult to be too tough on them on this point.
I’d like to point out that one thing the manipulation of the MXD data did do… was delete a the portion of a series with a negative slope (MXD) and replace it with a synthetic (“fudged”) data series that had a positive slope.
I’m not sure that even Mann was aware of the consequence of what he had done via this manipulation, especially in the case of the gridded MXD data.
Toto (Comment #87866) —
> You guys… “fudging� “make your own“? For a linear combination of the other data series? Seriously?
Yeppers. Seriously. I have Tiljander varve data to hand, so here is an example from that series: the final 10 years of “Lightsum” data archived by Tiljander03 author A.E.K. Ojala (1976-1985), followed by the final 10 years of the Mann08 calibration/validation period (1986-1995). Varve thickness (mineral component) in micrometers, to three significant figures.
Ojala — as submitted by A.E.K Ojala to the NOAA FTP.
Mann08 — per in the Mann08 archives submitted to NOAA (main page).
Year / Ojala / Mann08
1976 / 2,720 / 2,720
1977 / 1,300 / 1,300
1978 / 1,120 / 1,120
1979 / 970 / 970
1980 / 2,170 / 2,170
1981 / 2,110 / 2,110
1982 / 1,810 / 1,810
1983 / 2,810 / 2,810
1984 / 2,400 / 2,400
1985 / 1,560 / 1,557
1986 / no-data / 1,700
1987 / no-data / 1,320
1988 / no-data / 1,700
1989 / no-data / 1,130
1990 / no-data / 1,110
1991 / no-data / 1,260
1992 / no-data / 1,080
1993 /no-data / 750
1994 / no-data / 1,140
1995 / no-data / 1,060
Links to archival material, Excel files at Bitbucket.org, etc. are here.
> no one bothered to submit even a comment on this supposedly fraudulent method.
As Carrick pointed out already, you are raising the issue of “fraud,” not him or me. In no way is this fraud — it’s laid out in the Methods and S.I. for editors, reviewers, and readers.
> Doesn’t that make you pause for even a second?
Look, I’m making a strong statement: The Lightsum figures of 1,700 um for 1996, 1,320 um for 1997, and so forth do not refer to any actual measurement or calculation of any actual varve — they are simply extrapolations of pre-1996 trends.
Show me that I’m wrong, and enjoy watching me eat my words. I’ve linked to all the resources at my disposal, so this should be child’s play.
Right?
In your opinion, is the way that Mann08′s authors handled RegEM-based extrapolation of proxies in the calibration/validation period an instance of Good Scientific Practice? Is it an example of Acceptable Scientific Practice?
Carrick (Comment #87791) (way upthread)
> Amac, perhaps you could expand on the Schweingruber issue?
I’ve learned quite a bit about the Schweingruber issue thanks to the latter part of this thread. Specifically, that there appears to be grounds for challenging how Mann08 handled certain Schweingruber MXD Latewood tree-ring proxies in building, calibrating, and validating their recons. The heart of the matter lies in how Mann and coauthors handled these data series in the 1960-1995 period, when, apparently, “divergence” became an issue.
Did they use the MXD series as-is, as would typically be appropriate? Apparently not, for the most part (some SI figures appear to show such a use).
Did they use the MXD series as a composite with the instrumental temperature record? Possibly, but it seems not.
Did they truncate the MXD series at 1960, and graft on a RegEM-generated extrapolated pseudodata series from 1960 to 1995? For the most part, apparently they did.
I followed the “easier” conversations, and haven’t drawn any firm conclusions at this point, on the central question. I likely won’t get to the needed background reading for a while 🙁
Further insights from Boris, Carrick, toto, Kenneth, and others would be welcome, as always.
Amac:
I don’t have a problem in infilling missing data, I don’t even have a problem with a (limited amount of) extrapolation of a series past the end of the series. I have a huge problem with removing part of a data set and replacing it with a “science fiction” version of the original data set that tells a completely different story than the original version of the data.
Doing what Mann did and having it reverse the sign of the slope, that is fudging the data, pure and simple.
Interpolation or infilling is when you are missing a few interior data points from a series, and you use some method to “fill in” the missing values with what information you have available.
Extrapolation is when you are missing points at the end, and have to project into the future the behavior of the series. This is more dangerous than infilling, but if the extrapolation is short enough an interval, it’s usually an OK practice (certainly much better than zero filling or filling with the last value in the series, as was done in previous Mann reconstructions).
Extirpolation (extract + interpolate) is when you have a few errant points, that are obvious outliers, and you remove them and replace them with interpolated values. This is OK as long as you have an objective criterion for removal of the point, and if the removal doesn’t dramatically change the interpretation of the dataset (e.g., usually it’s done to reduce the noise floor not enhance the signal amplitude).
Fudging is when you take a series that says one thing, and transform it so it says another thing entirely. In this case, you aren’t reducing the noise floor, you are artificially inflating the amplitude of the signal.
I can’t think of a single example in science where a similar manipulation of data as was done with the MXD data has been permitted.
Why do you suppose it was allowed, and accepted, here?
Amac,
I think that you need to be clearer on what Mann actually did with this extrapolation. It isn’t creating data, in the sense of presenting them as results. They don’t present post-1960 data as results. A starting point for clarity would be to not describe it as a calibration/validation period. It is calibration, not validation.
In calibrating, you compare some statistical measures of the population of proxy results with some measures of temperature results to assign values to parameters used in the reconstructions. You are not specifically concerned with pointwise data. When they use their RegEM process, they are not adding to or subtracting from their knowledge of the statistics of the population. I presume their need for the extra points is to facilitate some comparison of the populations (proxy vs instrumental) expressed through the pointwise values.
There is a genuine problem, which is that they are short of data, and the calibration will be less effective. But adding more data consistent with the population measures is not changing that, either better or worse. as I think they recognise. It is just a convenience for calculation.
1 is correct (and the journal editors and reviewers presumably agreed) and 2 is wrong. It is all about their procedure. As Nick notes, Mann08 likely needed data for all years, so they need data simply as placeholders for the years where data is missing. Now, what method would you use to infill missing data to accommodate your procedure that would have the smallest effect on your result?
As for removing the post-1960 data, that is the divergence problem, which we have discussed before.
No one serious thinks this is “fudging” and those who describe it thusly are hoping that loud outrage will substitute for sound criticism.
Carrick (Comment #87877) —
Your arguments make sense, yet I find myself strongly disagreeing. Still, I confess up front that this is an intuitive perspective. It could be that when one works through the numbers, you are vindicated and I am in error.
I recall the old joke about the economist saying, “Let’s assume that we have a can-opener” (near the bottom of this page). Folks working in, say, mobile-phone digital processing can start off assuming that the proxies they’re working with — electromagnetic waves — contain information, and that such information pertains to sound waves that constitute a conversation. Does a particular infilling (etc.) procedure aid in reconstituting real information and not lead to spurious results? Engineers can check at every step, refining promising procedures and discarding dead-end ones.
By contrast, consider the subject of interest here. What was the worldwide average temperature anomaly’s actual trace over the past 2,000 years? There’s no independent source of knowledge beyond the sketchy thermometer record and, before that, proxies. So, nobody knows.
If data series X is indeed a proxy for temperature and if it has the characteristics that we think it does, then we might be justified in undertaking the manipulations you describe. But — what if these assumptions are partly or entirely invalid? In that case, we’ve unwittingly sailed into terra incognita.
As a for-instance, I think the evidence that Tiljander varve series are temperature proxies is very weak. In my opinion, they are more likely to be reflective of three things, pre-1720: (1) precipitation, (2) extent of ground cover in the watershed, and (3) extreme precipitation events (downpours, spring thaws). By manipulating the series, Mann08 inadvertently made the record take on more of the appearance of a temperature proxy. That’s a mistake.
As discussed upthread, ad hoc procedures highlight the issue of post hoc analysis — another pitfall. And it should go without saying that it’s a further error to fail to document such changes as are made.
Nick Stokes (Comment #87878)
> [Mann08] don’t present post-1960 data as results. A starting point for clarity would be to not describe it as a calibration/validation period. It is calibration, not validation.
Nick, sometimes I wonder if your purpose in writing stuff like this is to waste my time looking things up. From the paper (page 13254, column 1):
Clearly, the 1960-1995 interval is a large portion of the late validation period, and a smaller but still significant portion of the late calibration period.
> There is a genuine problem, which is that they are short of data
Agreed.
> adding more data consistent with the population measures is not changing that, either better or worse… It is just a convenience for calculation.
This is handwaving on your part. I assume the absence of compelling links or strong arguments on this point means that you don’t have any to offer.
Boris (Comment #87879)
> Mann08 likely needed data for all years, so they need data simply as placeholders for the years where data is missing. Now, what method would you use to infill missing data to accommodate your procedure that would have the smallest effect on your result?
“Oui, monsieur, but that is not my dog.” (PDF).
> As for removing the post-1960 data, that is the divergence problem, which we have discussed before.
Yes, we have discussed it. At this point, it highlights another highly unsatisfactory aspect of Mann08; see upthread.
> No one serious thinks this is “fudging†and those who describe it thusly are hoping that loud outrage will substitute for sound criticism.
Handwaving.
Also: In your opinion, is the way that Mann08′s authors handled the composited Schweingruber proxy issue an instance of Good Scientific Practice? Is it an example of Acceptable Scientific Practice?
Re: AMac (Comment #87880)
I liked the second joke on that page (http://www.nd.edu/~jstiver/jokes.htm)–the one where the punchline is “what do you want it to equal?” It may be even more appropriate here.
“There is a genuine problem, which is that they are short of data, and the calibration will be less effective. But adding more data consistent with the population measures is not changing that, either better or worse. as I think they recognise. It is just a convenience for calculation.”
We can always attempt to segment the discussion in order to defend something that when looking at the bigger picture is no longer defensible. What Nick states above has some very important statistical implications in that the degrees of freedom in such an endeavor must change and not in some easily understood manner. You normally would not be allowed, for the convenience of a calculation, to arbitrarily average the available data and then apply it as a new entry.
We have a series (Schweingruber) that obviously diverges from what was expected given the instrumental record and it probably diverges earlier than the other proxies in the Mann(08) reconstruction – which also diverge as shown and noted in the paper. We therefore have a major problem of divergence in a reconstruction that Mann (08) notes but fails to deal with in the manner that it obviously requires given its importance and implication for the validity of the reconstruction in the past. Do you Nick or the other defenders of this know why the date 1960 would be chosen for the cut-off of all the Schweingruber proxies. Why would you attempt to change the discussion to the need for infilling of data for a more convenient calculation when the issue is why would the authors arbitrarily cut-off data from proxies that do not agree with the instrumental record. If one cannot even conjecture why the series is “good” at 1960 and “bad” thereafter surely the whole series should be tossed.
The reader of Mann (08) who is not looking merely to defend but rather to understand the implications of the paper would see that what Mann (08) brings home very clearly is the need to use a priori criteria for selecting proxies based on good physical rationale and then using those proxies in the reconstructions – without modification.
Further the critical evidence stated in Mann (08) that the correlations of the proxies and the instrumental record could not happen by chance is based on the extensive and adulterated Schweingruber series, the later part of the Luterbacher series which is instrumental and the Tiljander proxies.
Even given the changes to the Schweingruber series and recognizing the important implications to the paper’s conclusion it is certainly remiss for the authors of Mann(08) not to have shown in the SI or even references from the SI what the removed data was and exactly what was the data replaced with.
In no way am I interested in getting into segmented discussions with Nick Stokes or other defenders, but rather what I think is important is to attempt to inform willing listeners whose prime purpose is not to defend.
I suspect that many who do not carefully read these papers and even those who have listened to some pretty good analyses coming out of SteveM and Jeff ID do not fully understand the extent that some of these papers like Mann (08) go to paper over some critical issues.
Maybe this would mean something if you had actually attempted to think about the issues instead of automatically assuming that Mann has done something wrong. A few people on blogs might have problems with this method, but Mann’s paper has been published for three years, cited over 200 times and there has not been some groundswell because he decided to infill data. This alone should tell you something. But if you are still convinced that Mann’s practice is poor and you want others to agree with you, then you need to show this in a convincing way, ideally to people who can make a difference (the editors at PNAS or perhaps a respected scientist.) This is just a echo chamber with the usual suspects throwing around personal insults and nodding in unison as if they are free thinkers when it is abundantly clear that they have a vested ideological interest in questioning the science of AGW.
Science isn’t about point-scoring on narrow, inconsequential grounds. If you have an improved method, show it your own work. If you have found a significant error, provide the correction. Everything else is just noise.
Boris,
I’d like to acknowledge reading your Comment #87885. I agree that it’s important to think about the issues being discussed. Nothing should be automatically assumed by people desiring to contribute to a technical thread, such as this one has become.
I’m also in accord with your notion that personal insults shouldn’t be thrown around. That makes it harder to constructively engage on the science. Besides, life’s too short.
“Maybe this would mean something if you had actually attempted to think about the issues instead of automatically assuming that Mann has done something wrong.”
Boris,
Mann presented his work and people have found things “wrong” with it. You seem to have something other than detached observation regarding this.
“Mann’s paper has been published for three years (irrelevant appeal to a number), cited over 200 times (another irrelevant appeal to a number) and there has not been some groundswell (irrelevant appeal to a subjective concept) because he decided to infill data.”
Andrew
Nick Stokes:
This is perhaps one of the most confused arguments I’ve seen in my entire life. What the heck, Nick.
We’re discussing deleting a portion of a data record that had the wrong slope and replacing it with invented data that has the “right” slope.
The problem wasn’t that the data wasn’t there. The problem is that the data inconveniently did not behave like a temperature proxy during that period of the calibration.
Boris:
One also shouldn’t assume that because it was cited 200 times it was correct either.
And you’re still using “infill” to incorrectly describe the operation that was actually done that is being objected to.
Secondly, there are obvious problems with the paper, that he nor his supporters will accept. One is the use of Tiljander, which wasn’t even a temperature proxy, so it’s beyond any reasonable question it’s in error.
The other was an antiquated version of the Sheep Mountain strip-bark proxy which not only was used in complete odds to the guidance provided by the NAS report, but neglected to use an updated version that did not have the hockey stick shape of the less complete version of the record.
It’s been shown if you drop both Tiljander and the obsolete version of Sheep Mountain are removed from the reconstruction, it has a substantial impact on the reconstruction (as in destroys the “robust” hockey stick shape.)
Actually the errors with this paper have been reported, are well documented and replicatable. (It’s an amusing fact that the reconstruction actually has zero skill before the year 1800 for example.)
The fact this paper has gotten so many citations just indicates to me the poor quality and low standards for scholarship this field embraces. Really it’s an indictment of the field.
Boris:
Anybody worth their salt will accept it is fudging data.
It is taking data that are doing one thing and replacing it with synthetic data that are doing another.
If that isn’t an example of fudging, what would it take before you would call it “fudging” data?
AMac:
I found myself unclear which part you were strongly disagreeing with. Perhaps you could state that more clearly?
Amac,
What Mann is saying is that they sometimes swap the periods around. But he used this extrapolation for calibration, not validation. How do we know?
Because you can’t use it for validation. In that phase you compare reconstructed proxy-based temps with instrumental. You’ve already used the numerical proxy data to get there.
Carrick, #87889 – confused? No, the problem is that you are making a lazy argument. As Boris says, it’s the way things go here. You don’t follow through to see what use Mann has actually made of these numbers. You do half an analysis and say, gosh that looks bad. It must be, it’s Mann. But no, it’s you.
Yes, Mann deemed the post-1960 data to be unsatisfactory, because of divergence. We know that – heard a lot about it. OK, it’s gone. NA. Doesn’t matter what it’s slope was.
So it’s NA. Need to do something to make the calibration formulae work, and like most of the world in this situation, some kind of infilling is used. Now everyone knows that infilled data is not real information, and if you find yourself using it as such, you’ve gone wrong. That’s what you have to show here.
As Kenneth says, there’s a dof problem. It arises from the actual lack of data, and is made neither better nor worse by the infilling.
Carrick (Comment #87891) —
> I [Carrick] found myself unclear which part you [AMac] were strongly disagreeing with.
Here’s my paraphrase of your Comment #87877. You reviewed a number of possible treatments of data series: infilling, extrapolation, removal-and-replacement, and “extirpolation.”
You deny that wholesale removal-and-replacement is a legitimate technique. This is what Mann08’s authors appear to have done with certain MXD Latewood series, 1960-1995. I agree with you. It’s hard to see how this can be a controversial stance, but apparently it is.
In the context of analyzing possible temperature proxies for multicentury recons, you can see infilling, extrapolation, and extirpolation as reasonable strategies to ease analysis and tease signal from noise.
In my view, there are two problems with this. One has to do with changes in the degrees of freedom, and thus with the interpretation of the statistical measures used to evaluate the experiment. Kenneth Fritsch has commented on this point, e.g. Comment #87884 (I am far from a stats whiz, but that DoFs would change as the result of such a step seems intiuitively correct.)
The second has to do with the nature of “signal” and “noise” in a data series such as a candidate temperature proxy.
A proxy might be influenced by one or more of the following: air temperature, lake temperature, severity of insect infestations, recent forest fires upwind, average summer humidity, the number of strong wind events per year, rapidity of snowmelt, extent of farming in the watershed, technology of farming in the watershed. Among other things. If we are focused on “temperature,” all but the first one or two might strike us as “noise.” But if we were interested in, say, the history of forest fires, our view of what constitutes noise in the proxy data would be quite different!
It seems to me that the more one manipulates a data series (e.g. by extirpolation), the more confidence one is displaying in one’s sense of what this series is about. “Of course this MXD series/varve series/d18O series is a temperature proxy, so of course I can bring out more of the temperature-proxy signal from amongst the noise with a touch of conservative manipulation.”
Can you?
What if the series in question isn’t a temperature-related series at all? In that case, you’ve succeeded in making a not-temperature series conform to the expected norms of a series that does contain a temperature signal.
For instance, in extirpolating the super-high anomalous value for the year 1326 of the Tiljander data series, Mann08’s authors have made it look more like a temperature proxy, and less like an extreme-weather-event proxy.
I would expect that such added confidence should be countered by widened error bars in the final analysis. But how would one go about this task?
Nick, sorry, I’ve been looking at this paper off and on for at least two years now, so your attempt at calling me lazy is shall we say weak at best, and a total fail with respect to your Jedi mind reading skills.
The problem with the MXD data is of course that it has a divergence problem. The question is what to do about it. My solution is simple–the data are invalidated by their divergent behavior. The hypothesis that they are temperature proxies fails.
Fudging the data (artificially inflating the amount of signal present post 1960) may be an appropriate solution for some people. But it’s wrong headed to not acknowledge that this what has been done here.
That said, your description of removing real data that aren’t missing but are behaving “improperly” with respect to the assumed casual variable (temperature) and replacing it with synthetic data that do, if not confused, is just wrong because you describe it as a “missing data problem”.
Whatever it is, it certainly is not that. Confused is a kind term for that inappropriate description of the series of operations that were actually done.
(I note with bemusement in an excoriatve commentary aimed at my intellectual laziness, you still choose to use the term “infilling” to describe delete followed by extrapolation of data series.)
Carrick #87888
“The problem wasn’t that the data wasn’t there. The problem is that the data inconveniently did not behave like a temperature proxy during that period of the calibration.”
In Amac’s example, I think this isn’t even true. As I see it, they had and used data to 1985. They lack data (inconvenient or otherwise) from 1985 to 1995, which they infilled.
AMac:
Actually all three of these are well studied and well controlled in the sense the uncertainty associated with the operations can be estimated and included in the final confidence level bounds.
The problems with what Mann did (deletion of a real data series, replace it with a synthetic one), is (as far as I know) there is no way to obtain an estimate of the effect of this replacement of series on the confidence level in the final reconstruction.
The main reason for this is somewhat subtle: We don’t really know the origin of the divergence of the gridded MXD series. In the absence of a model, we can’t say the reason for the failure of the MXD series after 1960.
Of course if we had a model, we might be able to correct the series for the effect that was causing it to diverge from a pure temperature proxy. In the absence of a model, we can’t even say with certainty that the MXD series were ever temperature proxies (as opposed to another quantity that covaried with temperature over say 1900-1960.)
Nick, I was discussing the MXD series collated by Briffa, not AMac’s example. From Mann’s SI (and via Boris above):
Had they not eliminated the MXD data post 1960, it would not have been “missing”.
Nick Stokes (Comment #87895)
> In Amac’s example…
Although we have been discussing the Schweingruber MXD Latewood series, that example (Comment “#87875) concerned Mann08’s extrapolation of the Tiljander Lightsum series past its end point of 1985. I was illustrating two points that seemed to be in contention.
1. That the extrapolated data is synthetic, or constructed, or artificial, or pseudodata, or non-real — whatever you want to call it. The RegEM-derived points don’t refer to any actual varve or tree-ring.
2. if a series is extrapolated past its end, it is not interpolated. If one insists on using the term “infilling” to describe this procedure, one should be clear that the obvious sense of the term doesn’t apply. This caution applies equally to RegEM as applied to Tiljander and to Schweingruber.
It’s possible to read your this in “I think this isn’t even true” as referencing a belief on your part that the Tiljander Lightsum series does “behave like a temperature proxy during that period of the calibration.†I assume that’s not what you meant.
Nick: “You do half an analysis and say, gosh that looks bad. It must be, it’s Mann. But no, it’s you.”
A lot of people are prone to that fallacy. In my my opinion Carrick and Amac are among the least affected by it. I’m sure you know. It is true, however, that this fallacy is like glued to certain people.
.
Sometimes it has the opposite direction.
Amac,
“if a series is extrapolated past its end, it is not interpolated.”
The distinction is overrated. What you do is decide on a predictive model (sometimes polynomial fitting, here RegEM) and create extra points for numerical convenience consistent with the model. It doesn’t matter where they are (inter or extra); they don’t create extra information, and you shouldn’t be relying on them doing so.
As to the issue of MXD vs Tiljander, I think Carrick is misreading the extract he cited:
It seems to me that the second sentence is not referring to the 1960-truncated series, but “proxy series terminating before the 1995 calibration interval endpoint” (there’s that calibtartion word again). That is, to situations like the 1985 example.
So “this isn’t even true” just refers to the question of whether they ever did infill the post-1960 MXD region.
“Yes, Mann deemed the post-1960 data to be unsatisfactory, because of divergence. We know that – heard a lot about it. OK, it’s gone. NA. Doesn’t matter what it’s slope was.
So it’s NA. Need to do something to make the calibration formulae work, and like most of the world in this situation, some kind of infilling is used. Now everyone knows that infilled data is not real information, and if you find yourself using it as such, you’ve gone wrong. That’s what you have to show here.”
This is utter nonsense. Mann (08) has a stated problem with divergence and not just with dendro proxies like Schweingruber but with non dendro series also. Schweingruber is not merely a proxy that is causing a minor inconvenience to be solved by infilling but rather it is a major issue with the entire reconstruction. Mann (08) decided to get rid of the divergent post 1960 data in the 105 Schweingruber series. If you look at the referenced Schweingruber series you will not see any data post 1960 – all the series end at 1960. Mann (08) replaces it with a data referenced to by an infilling algorithm. There is not infilling here but rather wholesale replacement of a block of data. We never get to see what was replaced and what it was replaced with. In fact many people, including myself, have attempted to go back to Schweingruber’s original archived data and determine which series correspond with those in Mann and without success.
Interesting side light here but LaymanLurker found that some of Briffa MXD series diverge (and converge in you go far enough back in time) from the around 1900. He subtracted the NH instrumental temperature series from the reconstruction response and looked for what could be classified as deterministic trends in the difference series. I found breakpoints (see link below) with the pre 1960 Schweingruber series used in Mann (80) at 1883 and 1944. So why did Mann (08) cut off the series at 1960? And why do some climate scientists talk about divergence as a later time phenomena than LaymanLurkers analysis showed?
If you take to spending full time defending these papers you never get around to asking the right questions about them.
http://img33.imageshack.us/img33/5700/schweinminusnhtempplotc.png
Nick–since you are imploring people to not be lazy, you should check this for yourself, but they did replace the eliminated MXD series (as with all series that didn’t terminate in 1995) using RegEM, and the result of this replacement was to remove a series with a negative slope with one with a positive slope.
If you are saying I am misreading something else, I’m not sure what it is, this is the sole issue I’m commenting on here.
I’ve almost given up on trying to get you to stop misusing the word “infill” in this context. When you are adding data points to the end of a series beyond where there is data in the series, regardless of the algorithm, this is not “infilling”.
Kenneth
“Do you Nick or the other defenders of this know why the date 1960 would be chosen for the cut-off of all the Schweingruber proxies. ”
That is a question many of us have asked from the beginning.
What objective criteria was used? If they merely selected 1960 because it was the last point in the proxies before they turned south, then I think thats a rather suspect choice. More interesting is what happens if you change that date to 1965, 70, 75, etc.
That study gives you some indication of how important the decision is. That’s just fundamental data analysis. That gives you insight into the structural uncertainty.
Further were all proxies put through a similar test? And does the test consider proxies that overespond to temperature? For example, tree that show 6 sigma growth spurts. I suspect not.
a one sided test that culls out under responders and negative responders, will decrease the variance and over estimate the certainty.
“It seems to me that the second sentence is not referring to the 1960-truncated series, but “proxy series terminating before the 1995 calibration interval endpoint†(there’s that calibtartion word again). That is, to situations like the 1985 example.
So “this isn’t even true†just refers to the question of whether they ever did infill the post-1960 MXD region.”
Nick, go to the Mann (08) SI and the first Excel link towards to end of the SI. In the Excel table go to line 910 which is where the Schweingruber MXD series
begins its listing. There are 105 series listed which go to line 1014. Look in columns that are headed with Start Year (for recon) and End Year (for recon) for the lines 910 through 1014 and you will see the End Year entries (for recon) all have 1995 listed -and not 1960. Look at the next two columns to the right and you will see Start Year and End Year which I assume gives the actual start and end years for the proxies – none of these ends before 1960.
when was the last time Nick Stokes ever admitted error….quatloos for the tqakers…
I know it’s not his fault, because you can’t pick your parents, but Nick Stokes is missing the Feynman gene.
============
Carrick,
“result of this replacement was to remove a series with a negative slope with one with a positive slope”
This is the laziness that I object to. The prosecution should make its case. Firstly show that they did it (I’m agnostic at the moment for the 1960-truncation), and then say why it was harmful.
The implication left hanging is that they want to show warming. But they aren’t doing that. We don’t need proxies to tell us about post-1960 temps, and they don’t use them for that. They (may) use them for calibrating. That’s a different issue. If they really were trying to tip the scales toward warming, I suspect that would actually have the reverse effect. And it’s really important to them to get the calibration right – otherwise they just fail validation.
In reference to my previous post at this thread:
To further show that the Schweingruber series where used in the screening test and for what period look at the next columns to the right of Start Year and End Year and you will see three columns labeled Pass screening over 1850-1995(r) , Pass screening over 1896-1995(r) and Pass screening over 1850-1949 (r) and you will see the entries for r for the Schweingruber series.
You guys must all live where the weather is not so good. Here in Florida: 23C, gentle breezes from the east. Tomorrow, much the same. Too many things to do outside.
“The implication left hanging is that they want to show warming. But they aren’t doing that. We don’t need proxies to tell us about post-1960 temps, and they don’t use them for that. They (may) use them for calibrating. That’s a different issue. If they really were trying to tip the scales toward warming, I suspect that would actually have the reverse effect. And it’s really important to them to get the calibration right – otherwise they just fail validation.”
Absolutely not the case. The problem is that divergence puts into doubt the capability of a proxy to act as a thermometer in present and past time. In fact tacking on the instrumental record on at the end is misleading not because in might be incorrect but rather that before doing that the instrumental and proxy thermometers have to be shown as equivalent. To get the best picture show the proxy series without the instrumental record and if the proxies divergence, so be it.
Nick,
I think if you go back to briffa orginal work, he argues that
if you dont truncate the series the MWP will be higher and that it will not correlate well with other recons.
He never considers ( as esper does) that the divergence may be from getting the local temps wrong. So, esper showed that if you use unadjusted temps there is no divergence.
In terms of proof, its the person cutting the data off who needs to supply the proof
1. proof that the cutting makes a difference
2. proof that all series are subjected to the same test.
3. proof that the data is anomalous and not the science itself.
“You guys must all live where the weather is not so good. Here in Florida: 23C, gentle breezes from the east. Tomorrow, much the same. Too many things to do outside.”
Here in IL the weather has been very good, for IL in Dec that is. It could change tomorrow. Weather in Florida appears to repeat itself day after day- kinda like this thread.
“He never considers ( as esper does) that the divergence may be from getting the local temps wrong. So, esper showed that if you use unadjusted temps there is no divergence.”
It would take a lot of diverging local temperatures to change the Schweingruber series and the Briffa reconstructions, for that matter. Steven Mosher, please link to the Esper paper.
Final proof that the Schweingruber series were “extended” to 1995 would be combining the following excerpt from the Mann (08) SI (below) with the listing of the correlations in the first column noted above under: Pass screening over 1850-1995(r)
“Although 484 (_40%) pass the temperature screening process over the full (1850–1995) calibration interval, one would expect that no more than _150 (_13%) of the proxy series would pass
the screening procedure described above by chance alone. This observation indicates that selection bias, although potentially problematic when employing screened predictors (see e.g. Schneider (5); note, though, that in their reply, Hegerl et al. (10) contest that this is actually an issue in the context of their own study), does not appear a significant problem in our case.”
The 484 number must include the Schweingruber series to be correct and the excerpt says that these series pass the temperature screening test over the full (1850-1995) calibration period.
For me, this conversation has elements of deja vu all over again. It’s akin to where I started with Tiljander-in-Mann08.
To review, defending Mann08 partisans poured scorn on the notion that the paper could possibly have [X], as Bad People had claimed. Or if it did, that the implications could possibly be [Y], as those Bad People had also asserted.
In the event, I read the paper, read the SI, pulled Tiljander03, located the NOAA archives, unzipped the Mann08 and CRUTEM3v files, formatted the data series in Excel, graphed them, and inspected them.
The [X] claims (upside-down, contaminated) turned out to be true. In consequence, so did the [Y] implications. And a series of [Z]s have arisen, casting further doubt on Mann08’s methods and conclusions.
Now, when people make claims in defense of Mann08 that seem to violate principles of common sense, lack logical arguments, and are bereft of supporting links, I’m much less inclined to credit them. I guess I’ve become more skeptical.
If this is a common response among scientifically-literate laypeople, that would be evidence that this Defend-Mainstream-Science strategy has costs, as well as benefits.
AMac, I would not be concerned about the defenders or even consider that their situation is unique with climate science. You look at the data, you do your analysis, you blog your results, you avoid the partisan traps and you let the chips fall where they may. You learn, we learn and without proselytizing.
SteveF,
I live even further south, where we value coolness at this time of year (and it’s OK). But my daughter arrived in Miami yesterday (escaping NVC), so I’m glad to hear it.
Kenneth,
Yes, undoubtedly divergence is a worry for proxy validity, and has been much worried about. But I’m curious whether anyone has thought through what a change of slope post-1960, in calibration would do. It seems to be taken for granted that it would artificially enhance warming, But would it? How?
Nick Stokes (Comment #87920)
> But I’m curious whether anyone has thought through what a change of slope post-1960, in calibration would do. It seems to be taken for granted that it would artificially enhance warming, But would it? How?
Demonstration of enhanced warming may not be the most important part of the story. Mann and coauthors have placed great store in demonstrations of the validity and skill of their reconstructions, and scorned competing recons for their supposed deficiencies on this point. Some people — myself included — speculate that the stated uncertainties in Mann08, etc., are about an order of magnitude lower than they would be, were more sources of error properly accounted for.
Mann08 and similar papers are based on the notion that the selected proxies correlate monotonically to temperature. “The divergence” — the post-1960 decline in the Schweingruber MXD series, in this case — lead directly to doubts about the validity of this dogmatic claim.
Nick:
I already explained why.
You either didn’t read it.
Or you don’t understand it.
And you have have the training you need to figure this out yourself.
It’s not my responsibility to think for you. And you object to me being lazy. What a crock of sh*t.
Anyway, why this fixation on “prosecution”? Do you have some fantasy that you’re a defense lawyer?
For most of us, these conversations are see as exploratory in nature. Your constant need to turn these conversations into an adversarial process gets in the way of fact finding. And that is something I personally find objectionable.
We shouldn’t have to go to court to discuss scientific issues and express our views and opinions.
Nick:
Sorry, but if I wanted to say something, I’d just say it. I wouldn’t leave any implication “hanging”, trust me.
The primary issue for me (and I think AMac and Kenneth and steven) is whether the methodology being used is sound.
There are other issues, but I try not to engage in guessing other people’s motives, let alone impugning them, such as you have been doing on this thread.
SteveF:
Actually I was on the road today to Florida. Northern part, not Southern.
AMac:
Actually it appears to be worse than that.
McIntyre found
(Emphasis mine.)
Lightsum is an inverted proxy which should not have been used. So if Mann had avoided the extrapolation, his results would have been not as bad.
However, in the CPS version, they simply calculate a correlation level. They should just take the correlation for the smaller dataset.
For me, this conversation has elements of deja vu all over again. It’s akin to where I started with Tiljander-in-Mann08.
This is kind of a rewriting of history to support your own storyline. No one claimed that the proxies weren’t contaminated. Mann even said they might be in his paper. And the implications were also presented in the paper.
Carrick,
“Anyway, why this fixation on “prosecutionâ€? Do you have some fantasy that you’re a defense lawyer?”
I think you have it about right. My impression is that Nick feels some obligation to defend everything that has ever happened in climate science, even when what he defends is obviously indefensible. If you listen to Nick, there has never been a significant lapse of judgment, never a violation of FOI laws, never a significant error in any climate science paper that suggests very high climate sensitivity, and there is no reason to imagine that the political/philosophical views of climate scienctists has ever influenced their thinking on ‘the science’. What is more, even a clear, direct statement about the need to exaggerate consequences and stoke fear in the public, in order to make certain the public supports reductions in fossil fuel use, never said what the plain meaning of the words stated.
.
The best parallel does seem to be a criminal defense lawyer in the courtroom. I find it too tiresome and predictable a tactic to take seriously, which is why I seldom bother to interact with Nick any longer.
Before the subject of the current line of discussion gets changed I have two issues pending for which I would like to hear replies.
Nick Stokes, based on what I showed in the Mann (08) SI regarding the “extension” of the Schweingruber proxies to 1995, what say you now?
Steven Mosher, do you have a reference for your comment on Esper and the use of local unadjusted temperatures. And by the way if we found that unadjusted temperature data gave significantly different and “better” results for reconstructions would not that cast doubt on the entire adjustment processes for temperatures.
“This is kind of a rewriting of history to support your own storyline. No one claimed that the proxies weren’t contaminated. Mann even said they might be in his paper. And the implications were also presented in the paper.”
That is the fun part of what we do here, Boris. I agree that much of the weaknesses in these reconstruction papers is noted and sometimes buried in the papers or SIs. What we do is not defend these papers, but rather hone in on those niggling doubts and analyze the background data and bring those sometimes vaguely worded doubts front and center for all to discuss and learn from.
I do not envy the duties of the defenders. It cannot be much fun but rather like defending the status quo and whatever it deals us and never moving forward.
“But I’m curious whether anyone has thought through what a change of slope post-1960, in calibration would do. It seems to be taken for granted that it would artificially enhance warming, But would it? How?”
If you had a divergence that covered most of the calibration and instrumental period, I would suppose that it might not ever be labeled a divergence. We base divergence on the measured instrumental record and divergence from it over the record period. If you had a divergence that was not recognizable because it covered the entire instrumental period and it was related to anthropogenic effects that endured mainly in the instrumental period than I would assume that calibrating an artificially low response to temperature increases and applying that model to past temperatures would inflate the past temperatures.
I really think that what LaymanLurker revealed in subtracting the instrumental temperature from the proxy response and noting that the divergence for the reconstruction he analyzed went back to 1900 and appeared that it could be accounted for as a deterministic trend in the difference series, deserves some considered attention. I think the divergence issue is handled so generally by climate scientists, who appear to me to be eager for it to go away, that they have not really dealt with it any detail other than attempts to show that it has an anthropogenic origin – and thus saving the validity of the reconstructions. Of course, an alternative explanation is that the divergence is what would be expected from a proxy response with long term persistence that does not respond faithfully to the temperature as a thermometer would. Anyway, if divergence actually varies from proxy to proxy and goes back further in time than what is currently assumed, I would think that there would have to be some major rethinking of its source/explanation.
Boris:
The bigger problem is Mann should never have included the Tiljander proxy in his series, “might be contaminated” was never an issue. Since you are insisting on historical accuracy, they were documented as contaminated when they were improperly used.
It’s interesting to consider the distortion that the incorrect inclusion of Tiljander has on the science done in this paper. It allows Mann to claim he can produce a hockey stick without using a tree ring proxy. It allows him to do a sensitivity study where he can remove other proxies (while retaining Tiljander) and show his results are “robust” against their removal.
Never mind Tiljander is being misinterpreted. The bigger error is methodological. Including Tiljander is little different than correlating against any other meaningless variable, such as temperature versus number of fire engines in San Francisco.
Erroneous methodology such as this allows you to do what is arrive at a misleading and wrong statistical inference about the significance of your result. Adding Tiljander is the empirical equivalent to dividing by zero during a derivation. It was a nonsensical thing to do, and unfortunately invalidates the entire inferential process that follows it.
Rumpole would resign from the case if he had as much admission of guilt as Nick has from his indicted scientists.
===============
Kenneth Fritsch:
Since, the trend of unadjusted temperature is practically meaningless, if the proxies are correlating better with them than the adjusted ones, IMO it just tells you they aren’t temperature proxies.
re: Kenneth Fritsch (Comment #87933)
Steven Mosher, do you have a reference for your comment on Esper and the use of local unadjusted temperatures. And by the way if we found that unadjusted temperature data gave significantly different and “better†results for reconstructions would not that cast doubt on the entire adjustment processes for temperatures.
Sharp comment Kenneth. it is interesting that this point should come up. A few days ago I looked in on Roman’s blog and saw that there were a few new comments on an old thread comparing adjusted GHCN to unadjusted. I read re-read Roman’s post and found this graph to look very familiar. Here is my graph showing the difference between NH observations and the climategate email (non-truncated) version of Briffa01.
I just don’t have much time right now but I will be exploring this further and post any interesting findings.
Sorry, the first paragraph above should have been blockquoted.
LL, thanks for graphs. Interesting that the divergence in the case of your analysis of the Briffa reconstruction would probably at least partly go away if unadjusted temperature data were used. I think Carrick makes the cogent statement that I was hoping someone else would make -the adjustments probably make good sense and thus if unadjusted data is used to make a point it probably is not valid.
By the way, it took me a while to appreciate what your analysis was doing exactly when you posted it (not your problem but rather mine as it turned out). I say this because it adds to the point Carrick made about analyses at these blogs (minus the incessant defending of papers) and discussing information in the light that none of it is final or finished but rather to promote further thinking on it and analyses.
Trends and uncertainties in Siberian indicators of 20th century warming
Of course I mentioned the possibility that temperature records may cause the divergence in our book, noting that Briffa could not question the records that Jones created. He could not question the science. he could only, therefore, delete the data.
When tree rings diverge from temperature you know almost nothing.
you know a little,
1. The science may be shot
2. the data may be shot
3. the temperature may be shot
Or ANY combination of 1, 2 and 3.
Esper:
“Estimates of past climate and future forest biomass dynamics are constrained by
uncertainties in the relationships between growth and climatic variability and uncertain-
ties in the instrumental data themselves. Of particular interest in this regard is the boreal-
forest zone, where radial growth has historically been closely connected with temperature
variability, but various lines of evidence have indicated a decoupling since about the
1960s. We here address this growth-vs.-temperature divergence by analyzing tree-ring
width and density data from across Siberia, and comparing 20th century proxy trends with
those derived from instrumental stations. We test the influence of approaches considered
in the recent literature on the divergence phenomenon (DP), including effects of tree-ring
standardization and calibration period, and explore instrumental uncertainties by em-
ploying both adjusted and nonadjusted temperature data to assess growth-climate
agreement. Results indicate that common methodological and data usage decisions alter
20th century growth and temperature trends in a way that can easily explain the post-1960
DP. We show that (i) Siberian station temperature adjustments were up to 1.3 1C for
decadal means before 1940, (ii) tree-ring detrending effects in the order of 0.6–0.8 1C, and
(iii) calibration uncertainties up to about 0.4 1C over the past 110 years. Despite these large
uncertainties, instrumental and tree growth estimates for the entire 20th century warming
interval match each other, to a degree previously not recognized, when care is taken to
preserve long-term trends in the tree-ring data. We further show that careful examination
of early temperature data and calibration of proxy timeseries over the full period of
overlap with instrumental data are both necessary to properly estimate 20th century long-
term changes and to avoid erroneous detection of post-1960 divergence. “
Steven Mosher:
Well, not quite true. Thermometers are well understood, simple, generally linear, measurement devices designed specifically to measure temperature.
Trees are complex organizations whose growth depends in general on a host of different stimuli. Through a series of somewhat convoluted and nebulous arguments, a tenuous theoretical connection between tree ring growth in some proxies and temperature (either local scale or regional, depending on the proxy) is posited.
I would expect, sans changes like TOBS and station moves, local temperature measurements, averaged over a month, to be incredibly reliable—compared to tree ring proxies in any case. While it is possible that somehow thermometers don’t work as well as claimed, this seems to me to a much less plausible argument than simply that the tree ring proxies in question don’t behave (over relatively long time periods) as temperature proxies.
Carrick,
I guess my point would be that the divergence would cause me to recheck everything, and report the results of the rechecking.
the fact of the divergence tells you nothing about the cause.
you might argue that it is more likely to be the science or the rings rather than thermometers, But that is evidence brought to the decision. If you get that distinction.
The evidence, divergence, does not determine the explnation.
that is, nothing in the data itself suggests one explanation over
another.
Steven Mosher:
Unless you think about the difference in the nature of a thermometer temperature measurement versus the proxy measurement from e.g. tree rings.
I agree it’s still interesting to think about…what would be required in terms of a breakdown of the temperature measurements for the proxy to be a more reliable measure of temperature.
But without invoking some really exotic mechanism, I am really doubtful you can explain a global-scaled divergence problem as simply an instrumentation problem. (H*ll, tree-ring chronologies from the same tree can be wildly different…this just exposes how difficult it is to get reliable proxy measurements.)
Boris (Comment #87930)
Boris, I agree with you and Kenneth Fritsch (#87933) that “much of the weaknesses in these reconstruction papers is noted and sometimes buried in the papers or SIs.”
Prof. Mann has standards with respect to the science he publishes, and he holds to them. While it’s easy to lament that his papers’ disclosures aren’t as full or complete as some of us would prefer, he deserves to be praised more often for what he has done in this regard.
As Carrick pointed out in #87935, that Mann08 made a partial, oblique disclosure doesn’t make up for the absence of due diligence with respect to the inclusion of the
fourtwo Tiljander03proxiesdata series. Nor does it remedy the severe methodological defects of the paper with respect to Tiljander.As far as the claim that I am rewriting history to support my own storyline, that offers readers a well-worn alternative — the one that runs, “Who are you gonna believe, me or your lyin’ eyes?” The risk is that your narrative might engage readers, and they might decide to check for themselves. Google or Bing will serve, or they could browse the resources I’ve compiled at my blog (linked posts at Pielke Jr.’s blog, Stoat, and Collide-a-scape are particularly germane).
Boris, the offer of a guest post on this or any other Tiljander-related issues you wish to discuss still stands. This is subject to editing only for issues related to vulgarity and libel (though I don’t expect either would arise). Or send me a link so I can feature it as a post; that removes me entirely from the editorial equation.
Carrick (Comment #87937),
“Since, the trend of unadjusted temperature is practically meaningless, if the proxies are correlating better with them than the adjusted ones, IMO it just tells you they aren’t temperature proxies.”
See Hansen et al. 2001 page 3 and especially Figure 1 in the appendices: http://pubs.giss.nasa.gov/docs/2001/2001_Hansen_etal.pdf
This is not the whole story but a good start.
Like so many claims made in these climate science papers, the claims sound rather substantial until one looks further at the details such as SteveM has done in the link below.
I doubt that divergence would really hold up in detailed studies of it as a problem from approximately 1960 onward even though that selected date jibes better with an anthropogenic cause related to increasing levels of GHGs.
Also it is good to remember that proxies in the form of tree rings have been shown to over respond to temperature in the recent warming period as well as under respond as in the divergence problem. We have the Yamal 6 sigma response. Remember also that in Mann (99) the North American TR series PCs were adjusted for showing unrealistic strong growth in the latter parts of the series. The adjustment was based on another TR series. The rationale for the adjustment was that NA TR PC showed growth that was conjectured to be caused by CO2 fertilization in trees at high altitudes. That growth leveled off in the latest part of the series and this was attributed by the authors to a “saturation” effect. The authors ended up stating that regardless of the exact mechanism the growth was unnatural and with anthropogenic causes.
From Mann(99) Northern Hemisphere Temperatures During the Past Millennium: Inferences, Uncertainties, and Limitations:
http://www.meteo.psu.edu/~mann/shared/research/old/mbh99.html
“A number of the highest elevation chronologies in the western U.S. do appear, however, to have exhibited long-term growth increases that are more dramatic than can be explained by instrumental temperature trends in these regions. Graybill and Idso [1993] suggest that such high-elevation, CO2-limited trees, in moisture-stressed environments, should exhibit a growth response to increasing CO2 levels…
.. As there is no a priori reason to expect the CO2 effect discussed above to apply to the NT series, and, furthermore, that series has been veri_ed through cross-comparison with a variety of proxy series in nearby regions [Overpeck et al, 1997], it is plausible that the divergence of the two series, is related to a CO2 inuence on the ITRDB PC #1 series.
The residual is indeed coherent with rising atmospheric CO2 (Figure 1b), until it levels off in the 20th century, which we speculate may represent a saturation effect whereby a new limiting factor is established at high CO2 levels. For our purposes, however, it suffices that we consider the residual to be non-climatic in nature, and consider the ITRDB PC #1 series \corrected”‘ by removing from it this residual, forcing it to align with the NT series at low frequencies throughout their mutual interval of overlap.”
From SteveM at CA on the Esper paper:
http://climateaudit.org/2009/11/03/esper-et-al-2009-on-west-siberia/
“The cutline in their abstract declares in effect that the divergence problem is not as “bad as we thoughtâ€:
Despite these large uncertainties, instrumental and tree growth estimates for the entire 20th century warming interval match each other, to a degree previously not recognized, when care is taken to preserve long-term trends in the tree-ring data. We further show that careful examination of early temperature data and calibration of proxy timeseries over the full period of overlap with instrumental data are both necessary to properly estimate 20th century longterm changes and to avoid erroneous detection of post-1960 divergence.
However, it doesn’t seem to me that this is supported very convincingly by their data analysis.”
“(H*ll, tree-ring chronologies from the same tree can be wildly different…this just exposes how difficult it is to get reliable proxy measurements.)”
That is another potential surprise to those who want to go back and look at TRW from the same trees. It was for me.
Hmm
Im not being clear Carrick. I don’t think it is soley down to a thermometer question. I think that Briffa could not look at that
possibility. Esper could. In the end I don’t know what portion
of the divergence is down to
1. funky trees
2. funky thermometers
3. funky science.
What I do think is that any approach that doesnt consider all three is incomplete and suspect. What I do think is that merely lopping off all the data post 1960 is fooling yourself. Tossing all the science out the window is also a bit extreme. Put another way, you or I would look at that divergence as a huge challenge and a opportunity to do something unique. Squirreling the data away, hiding the decline is last thing a curious type would do.
that’s the thing that puzzles me the most.In some ways you’d kill for data that goes the wrong way.
“that’s the thing that puzzles me the most.In some ways you’d kill for data that goes the wrong way.”
Exactly, but the new lines of inquiry lead far away from public interest, spotlight and probably funding.