Eric Raymond on "Hiding the Decline, Part 1"

http://esr.ibiblio.org/?p=1447

by skmurphy • 17 years ago

72 54 17 years ago

esr.ibiblio.org

skmurphy2• 17 years ago

opening paragraph comments on a code fragment:

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.

indy• 17 years ago

jgrahamc mentioned in an earlier comment (and also on his blog: http://www.jgc.org/blog/2009/11/about-that-cru-hack.html) that these variables aren't used.

waterlesscloud• 17 years ago

Great, so now all we need is to run this in the released data and verify if the graph was produced with the code commented in or out.

If there was released data, that is.

The point being we only know how the code looks in that file, we don't know when the commented out sections were commented out.

So nothing is proven here either way.

defen• 17 years ago

Check out harris-tree/briffa_sep98_e.pro - http://di2.nu/foia/harris-tree/briffa_sep98_e.pro

You'll see that it is used in that version of the program.

skmurphy2• 17 years ago

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).

etal2• 17 years ago

Notes:

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.

yummyfajitas• 17 years ago

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.

Even by academic standards, this code is fugly.

martincmartin• 17 years ago

As The Economist points out:

"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...

greendestiny3• 17 years ago

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.

jsz0• 17 years ago

Was it the only data source in other time periods? or was it collaborated by other data?

Perceval2• 17 years ago

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?

algorias• 17 years ago

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.

btilly• 17 years ago

And this is why people who don't know science should ask questions first.

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.

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.

There is no begging the question when we are also studying the mechanisms involved.

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.

jacoblyles• 17 years ago

While I'm glad people are digging through the leaked source code, I doubt they find anything so blatant as outright fraud.

j_baker2• 17 years ago

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.

DomesticMouse• 17 years ago

Science that relies on cooked data isn't science. It's snake oil sales.

alex_c• 17 years ago

The code is pretty damning, but before I get really outraged, a bit of context is always useful.

Has anyone figured out where this graph has been used?

jgrahamc• 17 years ago

No. And, also, the interpolated value yearlyadj isn't used. It's only referenced in commented out code.

http://www.jgc.org/blog/2009/11/about-that-cru-hack.html

hop• 17 years ago

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:

http://en.wikipedia.org/wiki/File:Global_Warming_Predictions...

petercooper• 17 years ago

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 )..

camccann2• 17 years ago

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.

azgolfer• 17 years ago

Anyone who says they have a good computer model of the climate is either ignorant, stupid, corrupt or some combination thereof.

ieure2• 17 years ago

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.

javert• 17 years ago

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?

ieure2• 17 years ago

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.

trominos2• 17 years ago

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.

voidpointer• 17 years ago

Well put.

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.

akamaka• 17 years ago

If we were in a formal debate, you'd be right.

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.

jbooth2• 17 years ago

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.

camccann2• 17 years ago

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.

1gor• 17 years ago

   he's been in deep teabagger territory

What has ESR's political views ('right wing' in your opinion) have to do with his argument in this case?

Shoud we disregard everything you say just because you seem to be a 'left-wing treehugger' (using your language)?

Let's not turn HN in an echo-chamber for a particular political grouping.

yters• 17 years ago

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.

Joeboy• 17 years ago

Can anybody tell me what, if any, published work this data has made it into?

Eric Raymond on "Hiding the Decline, Part 1"

Related Stories