"The 'hiding' concerned the decision to leave out a set of tree-ring-growth data that had stopped reflecting local temperature changes. That alteration in growth pattern is strange, and unexplained, but eliminating it is not sinister."
and:
" 'I’ve just completed Mike’s Nature trick of adding in the real temps to each series for the last 20 years (ie from 1981 onwards) and from 1961 for Keith’s to hide the decline.' Trickery associated with Dr Mann was catnip to the sceptics. But Dr Jones has clarified that 'The word trick was used here colloquially as in a clever thing to do. It is ludicrous to suggest that it refers to anything untoward.' "
My concern about the tree-ring data is that, on face, it seems like tautological reasoning. Please correct me if I'm misrepresenting what's going on.
The tree ring data, after a certain point in history, is no longer a reliable indicator of local temperature changes. There is no explanation for why trees no longer reflect local temperature changes. We know that they no longer reflect local temperatures, presumably, because we have other more reliable measurements available now (measurements that are not available further back in history).
This leaves us with a conundrum: is there any period in which alternative measurements and tree-ring measurements correspond with one another? Is the divergence between our measurements and the tree-rings something we can trace? Is it sudden or gradual? If there is no period in which there is a correspondence between our measurements and the tree-rings, what does this say about our assumption that tree-rings have ever accurately captured local temperatures?
It seems that the implication is that changes in the environment have rendered tree-rings unreliable. But if we're arguing that tree-rings are unreliable because of environmental change (i.e. the chemical changes to the composition of the atmosphere and water cycle that contribute to global warming), and we're seeking to explain environmental change, we're begging the question (assuming the explanandum in the premise of our argument).
Until a causal mechanism for the divergence between tree-rings and other measured temperatures can be specified, how can they legitimately be discarded as evidence, without invoking tautological reasoning?
If you don't have accurate data for a certain time period, it's preferable to have approximate data than none, as long as you take that lower accuracy into consideration.
And this is why people who don't know science should ask questions first.
The tree ring data, after a certain point in history, is no longer a reliable indicator of local temperature changes. There is no explanation for why trees no longer reflect local temperature changes. We know that they no longer reflect local temperatures, presumably, because we have other more reliable measurements available now (measurements that are not available further back in history).
There is NO SINGLE EXPLANATION for the discrepancy. This is NOT the same as there being no explanation.
What happens instead is that if you look in specific localities you can find specific explanations. Increased acidity here. Drought conditions there. Insect infestations in another. And in yet another the correlation with temperatures remains.
What happened is that after 1960 due to large scale changes in environment in many localities the correlation broke down, so the overall correlation broke down. But there is no single explanation because in different areas the specific reasons differ.
This leaves us with a conundrum: is there any period in which alternative measurements and tree-ring measurements correspond with one another? Is the divergence between our measurements and the tree-rings something we can trace? Is it sudden or gradual? If there is no period in which there is a correspondence between our measurements and the tree-rings, what does this say about our assumption that tree-rings have ever accurately captured local temperatures?
Actually there is about a century of comparative data before that which shows a good correlation. That is why we are willing to project temperatures back farther using that record.
Now there is the possibility that there were previous periods which saw a similar breaking of the correlation. This question is a question of scientific debate.
It seems that the implication is that changes in the environment have rendered tree-rings unreliable. But if we're arguing that tree-rings are unreliable because of environmental change (i.e. the chemical changes to the composition of the atmosphere and water cycle that contribute to global warming), and we're seeking to explain environmental change, we're begging the question (assuming the explanandum in the premise of our argument).
There is no begging the question when we are also studying the mechanisms involved.
Until a causal mechanism for the divergence between tree-rings and other measured temperatures can be specified, how can they legitimately be discarded as evidence, without invoking tautological reasoning?
You are making two large assumptions here. The first is that there is a single causal mechanism. The second is the assumption that we don't actually understand the mechanisms involved. Both assumptions are wrong.
Look. Scientists aren't exactly stupid people. When they spend a lifetime studying something, you should assume that they've asked the obvious questions and try to find out the answers they have found rather than assuming that they are stupid and you're the first to notice the glaring issues.
Unfortunately climate scientists are also humans who are under constant assault from people who can't and won't be bothered to actually learn the record. People who feel they are under constant assault engage in defensive behavior. And people who are being defensive tend to circle the wagons and have difficulty in remaining open with outside groups.
When you encounter such defensive behavior I strongly recommend trying to understand what the defensive person's position is rather than immediately piling on and validating their defensiveness. If they are wrong, this approach leaves you in a better position to figure that out. But if they have a point, you'll be more likely to figure that out as well.
Simply because a heckuva lot of smart people have studied an issue for decades does not mean -- at all -- that somehow they have made more progress than some schmuck reading about all of this for the first time in Newsweek. You do not get a free pass by being smart and asking a lot of questions.
(A side note: I find among academia the meme that somehow intelligence, hard work, and admiration of peers equates to progress. These are all great traits to have, but in themselves they don't automatically translate into solid value for posterity. In fact, in some fields there may actually be an inverse relationship)
Having said that, you are absolutely correct in saying that you should ask questions first. There's no point in getting everybody on the defensive.
But if the reply is "Look -- it's all very complicated, and we've thought about it a lot and we're way ahead of you here and this is the best plausible way to settle several internal arguments... etc, etc" then you have not received an adequate answer.
I applaud your appeal to being inquisitive and polite. I find your defense of this particular case, however, sorely lacking.
You've mentioned several mechanisms: acidity, drought, insects, and presumably more. This raises two comparative questions in my mind.
First, changes in acidity, droughts, and insect infestations are certainly not events that are unique to the post-1960s environment. Do local temperature data from tree rings from prior time periods also show discontinuities during known periods of drought or insect infestation?
Second, there have been large-scale changes to the environment in the past, some for a short period (the "year without a summer") some for a longer period (the "little ice age"). Do tree rings correlate with known large-scale temperature changes, or do they show discontinuities of the type we are witnessing in the post-1960 era?
Look. Scientists aren't exactly stupid people. When they spend a lifetime studying something, you should assume that they've asked the obvious questions and try to find out the answers they have found rather than assuming that they are stupid and you're the first to notice the glaring issues.
I agree, but as a PhD student, I've found that highlighting errors in methodology, even in peer reviewed articles, is actually pretty easy. You sound like you're familiar with research methodology, so you should know just as well as I do. When taking classes on methodology, one of the principal methods of learning is critiquing the research designs, assumptions, questions, data sets, etc. of other scientists who have been published. It's not hard to find methodological errors, or at least questionable assumptions and questionable glosses over complex issues, when reviewing even the published work of others.
Unfortunately climate scientists are also humans who are under constant assault from people who can't and won't be bothered to actually learn the record.
Because climate interaction is non-linear and operating on a macro- and long-term scale, it will obviously have a hard time standing up to Popperian standards of falsifiability, and the public is going to have a harder time gaining confidence in climate science's conclusions. Just because climate scientists are under assault from wingnuts who see them as a nefarious player in the Culture Wars doesn't mean that they are excused from answering questions about their methods, assumptions, and data.
If the tree-ring growth data doesn't match directly measured temperature at significant intervals of time (necessitating what amounts to arbitrarily changing data sources mid graph) why on earth would you trust it for periods where its the only data source.
I think the evidence for AGW extends beyond this particular incident, but I think its highly unethical and very poor science.
That's a fair criticism, but it would be worth seeing the figure caption for it as well. If the caption had within it that the source was switched part way due to a known problem, that would be fine.
Out of curiosity, anyone know a journal reference for that particular graph?
Comments on the OP suggest it was cover art. Not sure what to make of that without more context.
Making it clear the data changed stops it being unethical, but I think the criticism of concatenating two data sources remains. Especially when the point of the graph is the sharp rise at the end - just as the data source changes. You really want to make sure the data sources correlate highly to make that kind of argument.
I don't know what ESR's politics looked like before, but he's been in deep teabagger territory for a while now. I would take anything he says on the subject of politics with… Well, I'd just not take it. He can pretend to be objective, but the use of winger slang (“AGW” and “MSM” are right-wing dog whistles in the same way as “Democrat party” or “socialism”) is a giveaway.
Here are some of his prognostications from the election last year. They're delightfully wingnutty, with a crunchy glibertarian shell:
“The Obama campaign smells of defeat”: http://esr.ibiblio.org/?p=449
“Democrats’ polling league in a generic contest has collapsed. Control of Congress is in play. Wow. Just wow. Considering how hard the Republicans have been fucking up and alienating their base, this is astonishing. EPIC FAIL.”
“Sarah Palin, American Centrist”: http://esr.ibiblio.org/?p=481
“So when I tell you that I think Sarah Palin’s religious position is pretty near dead center in the American spectrum, you can be pretty sure I’m not fudging to make that position look good.”
“The Obama campaign smells of defeat: II”: http://esr.ibiblio.org/?p=490
“The attempt to smear, discredit and delegitimize Palin has steadily become more intense and more damaging — to the Democrats.”
When it comes to politics, dude simply does not know what the fuck.
You're just attacking ESR's character. It's fallacious to discount his blog post just because you don't like his character. Please, can't we have a rational discussion on the topic?
He is certainly attacking ESR's character, but I think that the old mantra that we should always address an argument and not the person delivering it isn't entirely fair. The strongest reason for this, in my opinion, is that most complex systems can't be thoroughly analyzed with any objectivity, because it's impossible to consciously process all the available data; there's just too much going on. So whenever somebody argues about a complex system, there's a very implicit part of their argument that involves which data they choose to present. A bystander who's trying to follow (and validate or invalidate) a logical argument can't always tell if the choice of data is unbiased, and of course looking at all available data takes the bystander as much time as it took the arguer to formulate the argument in the first place.
So USUALLY what an unbiased observer will do is proceed with caution: see what the arguer's point is, see how it fits with the observer's intuition (because this is the same as appealing to the observer's subconscious, which in most cases has processed far more data relevant to some issue than has their conscious mind), and see whether the data seems overwhelming enough that its conclusion couldn't be the result of biased data selection on the part of the arguer. But this method often fails. Without formulating your own opinion by immersing yourself in data, you just can't be sure whether you're looking at a legitimate picture of an issue. That's why whenever I hear a convincing argument about something complex, I only provisionally believe whatever I've just been convinced of. And that's why an argument's author actually matters if the argument relates to a sufficiently complex system.
Now, arguments about simple (by which I really mean "completely specified") systems aren't victim to this weakness -- mathematical proof, for instance, can actually be really verified by an observer without the observer doing the same work that the discover did, as can many logical arguments in the hard sciences.
But what ESR is trying to show here is that some particular set of scientists committed scientific fraud. This is a) a value judgment and b) depends substantially on context.
So I think that it is not only reasonable but very much a good thing that ieure pointed out that ESR is a nutjob. Now I know that even if I find his post compelling, he has no intellectual honesty and the data he's showing me have been carefully selected from the unimaginably vast amounts of data in the world because they happen to support his worldview. Maybe he's still right -- but ieure's post is still very much relevant.
"So I think that it is not only reasonable but very much a good thing that ieure pointed out that ESR is a nutjob."
That was a very sophisticated formulation of "I don't listen to people whose politics I disagree with."
That is exactly the problem of the entire Republican party, and you seem to be saying that the principle is sound (just as long as it is directed at "them" and not "you").
It saddens and surprises me that you and your up-voters seem to be endorsing ad hominem as sound debating practice. Of all places, I would hate to see Hacker News descend to this level.
> That was a very sophisticated formulation of "I don't listen to people whose politics I disagree with."
It wasn't. It was an argument for believing that "in many cases it matters where the arguments that I hear come from."
Let me be clear: I don't like the fact that I think this policy is right. I just think it's right.
And let me also be clear that in many or most cases it makes sense to disregard the source of an argument. In most exchanges on HN, the people involved all implicitly agree on the data that's being argued from, and in those cases I tend to think that we don't ever need to think about people's biases.
But the fundamental belief underlying the "we can always disregard an argument's source" movement is that sufficiently logical and dispassionate people can validate or invalidate an argument very easily (compared to the work it took to formulate the argument in the first place). I really really would like this to be true. I just don't think it is, IN GENERAL.
To reiterate why: I think it's perfectly possible to follow a logical train of thought without actually re-deriving that train of thought -- and so I think that we can untangle argument and arguer when arguments are purely logical -- but if an argument is based on data from a complex system I don't think it's possible to determine if that data was chosen in an unbiased manner without actually looking at all the available data and essentially re-deriving the arguer's conclusion.
Combine that with two hopefully-obvious (and hopefully-true) premises: first, that a large data set on any sufficiently complex system contains individual data points that support any given belief about that system,
and second, that most really complex systems can't be considered completely logically, because there's so much going on, and must instead be approached by looking at data, and I think it follows that you need to think about the source of any argument about a complex system when you're deciding whether you (provisionally) believe it.
Incidentally, I find it genuinely ironic that your post is actually a not-particularly-sophisticated example of "I don't listen to people whose politics I disagree with," since you ignored my argument based on my dissenting conclusion.
"...but if an argument is based on data from a complex system I don't think it's possible to determine if that data was chosen in an unbiased manner without actually looking at all the available data and essentially re-deriving the arguer's conclusion."
But this is exactly what I come to Hacker News for. In this case, there are definitely HN readers who can understand the code ESR is talking about, and pretty quickly figure out if there are flaws in ESR's reasoning regarding the code in question.
Once we have the actual flaws in his reasoning, his tea-baggery or whatever become completely irrelevant, because we have much more solid reasons for rejecting his conclusions. If this were a different forum, with no participants well versed in technology, mathematics, and the scientific method, maybe you need to consider the motivation of the arguer when considering the validity of the argument. HN is not that forum.
I think a better formulation is "I don't listen to people who've shown that they're not going to argue in good faith". If ESR has really been in deep teabagger territory, then the fact that he's smart really has nothing to do with it -- he's not seeking truth, he's seeking confirmation, so who cares about his conclusions.
I was not aware that one could only upvote a position one agrees with. I upvoted because it is an interesting argument cogently made, not because I agree with the premises or conclusion.
But we're not. The original poster's point is that ESR shouldn't even be admitted into the debate.
Personally, I'm ignoring everything I read about this story unless it there's evidence that it comes from someone who has invested some serious effort into the subject.
There are far more relevant points, such as the fact that the fudged data series isn't even used anywhere in the rest of the program. The code fragment is indeed pretty useless without the entire program. It seems like ESR has stopped looking at the code once he thought he had found what he was looking for. The adjusted data is never referenced in the rest of the program: http://news.ycombinator.com/item?id=964896
Ideally we would have access to the entire program + data and could re-run it to see whether the graph was produced with or without the adjustments enabled. It is very well possible that the program was altered after the graph was plotted and before that version of the source became public.
No, what I’m doing is pointing out that his opinions are a) wrong and b) delivered with right-wing spin. As such, you should not take what he says about matters political at face value.
You seem to be making some unsubstantiated leaps from what he's written, to "right-wing", to "wrong".
Just because you don't agree with what someone has written doesn't mean you can dump them in a camp you don't like, and just because you do so doesn't mean that you can discount him entirely on all such matters.
Considering the variable produced by code in question is never used, as shown by two comments in this thread, indicates to me that ESR is wrong.
Which is odd, because ESR is an expert programmer, so one would assume he'd check to see if this was a case of dead code. He does not do this, and instead prattles off wingnut talking points on climate gate.
ESR a programmer? Do you ever read the mailing list of the major FLOSS projects out there? Read them and search for his name. It was even the theme of a webcomic in the past: http://geekz.co.uk/lovesraymond/archive/show-them-the-code
It's much more fallacious to discount the overwhelming scientific consensus on global warming because of one scientist's character.
A virtual army of teabaggers combing over this data and all they could find was one marginal case of cooking the data? I'm more confident than before.
Also -- it's not fallacious to point out historical bias by the author when discussing a partisan issue. It's additional information to be taken into account.
EDIT: Partisan downrate? Please, people, go to redstate.com or something if that's how you feel.
With all due respect, if you're going to take ESR to task for being politically biased and point out his use of right-wing catchphrases, you slightly undermine your point by using the term "teabagger".
...not to say that ESR isn't detached from anything resembling reality, but responding in kind isn't terribly helpful.
From the CRU code file osborn-tree6/briffa_sep98_d.pro , used to prepare a graph purported to be of Northern Hemisphere temperatures and reconstructions.
;
; Apply a VERY ARTIFICAL correction for decline!!
;
yrloc=[1400,findgen(19)*5.+1904]
valadj= [0.,0.,0.,0.,0.,-0.1,-0.25,-0.3,0.,-0.1,0.3,0.8,1.2,1.7,$
2.5,2.6,2.6,2.6,2.6,2.6]*0.75 ; fudge factor
if n_elements(yrloc) ne n_elements(valadj) then message,’Oooops!’
;
yearlyadj=interpol(valadj,yrloc,timey)
This, people, is blatant data-cooking, with no pretense otherwise. It flattens a period of warm temperatures in the 1930s--see those negative coefficients? Then, later on, it applies a positive multiplier so you get a nice dramatic hockey stick at the end of the century.
1. The fudge factor is flat after 1970. The hockey stick we care about is after 1980 or so, sharply increasing in the 1990s and 2000s.
2. The fudge corresponds to problems with recent tree-ring data that were discussed in previously published literature (I don't have the citation). The phrase "very artificial" almost certainly means it's a crude estimate based on some model that's simpler than the author would prefer -- hence, the flat line for recent values.
3. If this was an open-source project, you'd see a reference to the article that explains where the fudge came from, but it's not. Natural scientists tend to not really believe in software engineering, they just write a procedure and munge it until it works, leaving dead code and mysterious comments in their wake. (Source control? That sounds like fluffy industry stuff...) As JCG pointed out, yearlyadj isn't used in the copy of this script that we have, although it might have been used in an earlier draft of the script.
Natural scientists tend to not really believe in software engineering, they just write a procedure and munge it until it works, leaving dead code and mysterious comments in their wake. (Source control? That sounds like fluffy industry stuff...)
Not the ones I've worked with. While they may be a little behind the times (CVS instead of hg, fortran instead of c), most academics aren't stupid.
Physics and math software is usually fine, sure, since the performance and reliability of the code matters in the experiments. I mean traditional biology and the less computer-oriented sciences, where they tend to just use scripts to filter data into a usable form, crunch some numbers, and generate visualizations. There, unless the software is the point of the research, writing code doesn't get the same priority (or respect) as traditional experiments. A lot of those professors just don't care about writing maintainable code; I'm guessing Dr. Briffa is one of those.
They are used in the next version, see http://di2.nu/foia/harris-tree/briffa_sep98_e.pro (promoted from a comment by jgrahamc on his blog which I would think should get added to his main blog post).
How do all the "best" climate prediction models have squiggly lines after 10 years from now and not a broad range or something realistic. Their algorithims smell of BS and way too many variables go into it to make them have anywhere near their plotted accuracy for 100 year projections. From wikpedia:
Well, the graph you link to doesn't include error bars or any other indication of uncertainty. Leaving out error margins and plotting only the mean is unfortunately common when presenting scientific results to general audiences; but that doesn't mean the actual scientists don't have the full data.
I'd be interested in that too. I wonder if they include things like predictable levels of solar activity, volcano eruptions, weather cycles and the like to get this. Or, worse, it might even be stochastic or "random walk"-like which would make the variations meaningless (see diagram on http://en.wikipedia.org/wiki/Random_walk )..
So maybe we haven't found a smoking gun yet, but the general gist of the documents and emails in that package is not very encouraging.
You can find reviews of IPCC's book that are full of very strident critiques. You can find emails discussing the inherent inaccuracy of their work. You can find internal dismissal of the tree ring data, stating how useless it is. You can find a scientist openly glad that he's not being held accountable so he can push his climate interests.
After spending a couple hours pouring through the repository, I haven't found much that makes this work seem stable enough to base global lifechanging legislation on.
I'm not necessarily attempting to justify any data-cooking whatsoever, but why is it such a big deal? I mean, the climate "skeptics" do it all the time and it hardly makes the news.
We're never going to have a decent public discourse on this issue until both sides are honest with the public. Maybe I'm just a pessimist, but I don't see this happening any time soon.
"The 'hiding' concerned the decision to leave out a set of tree-ring-growth data that had stopped reflecting local temperature changes. That alteration in growth pattern is strange, and unexplained, but eliminating it is not sinister."
and:
" 'I’ve just completed Mike’s Nature trick of adding in the real temps to each series for the last 20 years (ie from 1981 onwards) and from 1961 for Keith’s to hide the decline.' Trickery associated with Dr Mann was catnip to the sceptics. But Dr Jones has clarified that 'The word trick was used here colloquially as in a clever thing to do. It is ludicrous to suggest that it refers to anything untoward.' "
http://www.economist.com/opinion/displaystory.cfm?story_id=1...