The only thing we can be sure of about the future is that it will be absolutely fantastic. So if what I say now seems to you to be very reasonable, then I have failed completely. Only if what I tell you appears absolutely unbelievable, have we any chance of visualizing the future as it really will happen.

- Arthur C. Clarke (h/t Brin)

Friday, August 22, 2008

Pielke, Part of the Problem



This question of "speaking up" can cut the other way, though.

I received a bit of denialist drivel in email, pointing among other things to the infamous bogus CO2 record, to give you an idea of the quality of the correspondence.

However, I while judging someone by the company they keep may have some value, it is a mistake to judge someone by the company that keeps them. So, in the same message, there is a pointer to a recent article in Prometheus linking to a recent summary in a libertarian blog of the hockey stick story. This item deserves some attention, if not especially on the basis of merit, as it seems that the Bishop Hill article (and its catchy appelation of a "Jesus Paper") will have some resonance among the opponents of timely action on global change issues.

A particularly intransigent comment to Pielke's above-referenced Prometheus article from William Eschenbach reads as follows:
The problem is not the behavior of the few. A few people will always do wrong. The problem is that the behavior of the community as a whole has been just what you said. They have not stood up to oppose the bad science done in their name. They have not clamored for an investigation into the bad science. They have, in large part, done absolutely nothing in response to this abysmal situation. Nothing. No public statements. No behind-the-scenes maneuvers. Nothing. Zip. Zero.

Instead, by and large, they have in your words “stayed out of the limelight” … and now you are claiming that they are the victims in this case?

In my opinion, they have no one but themselves to blame for the fact that they are being tarred with the same brush as the miscreants.
First of all this assumes the existence of "miscreants", which goes far further than even the excessive Wegman report (which in its own excesses raises some uncomfortable questions about the conduct of modern science) does. Second, it places an onus on a "community" to police itself in a way that provides an unrealistic model of the community. The IPCC, even constrained to WGI, provides summaries of the positions of a wide range of loosely connected communities, among whom dendrochronology forms but a tiny corner.

One wonders who is expected to "speak up" and when. And how, in the light of limited resources and competitive funding, one is expected to find the time to work out the details. The concept that some oceanographer or satellite engineer or icthyologist has some obligation to "get into the limelight" about something as narrow as tree rings doesn't ring true to anyone actually working in the field.

There are real issues with the conduct of science, but the question at hand is how important they are. It is absolutely crucial to note that no responsible party, neither Wegman nor McIntyre himself denies that the millenial temperature curve will likely turn out hockey-stick-like once enough data is collected and analyzed. In fact, a contrary result would be quite surprising!

They argue that the statistical methodology for obtaining these results is inadequate, and stake out a position of defending the integrity of science. It is hard for me not to sympathize with these claims. Few close to modern science will deny that the process has important flaws, but fewer still are in much of a position to address them.

The problem of the conduct of science pales in importance to the problem of bringing the human impact global environment into stability, though. The tragedy is that these quibbles are inflated onto accusations of such spectacular dishonesty specifically in order to color the policy debate.

Regarding the hockey stick, I have to line up with Pielke in shrugging and saying I haven't spent the effort to figure out how the science shakes out and I don't really plan to. It seems likely to me that the temperature really did follow a hockey stick pattern (as so many things do nowadays). The scientific question is only the extent to which the data confirms that expectation, not whether in fact the hypothesis has been shown invalid. And in the grand scheme of science this particular question has very modest importance, despite the political weight placed upon it.

The denialists are not especially interested in the question as to whether the hockey stick is real, it turns out. They are mostly interested in what it reveals about the IPCC process. And here, it is hard to say they don't have something.

The accusation, removed of acrimonious ranting, is 1) that IPCC knows in advance what result it is delivering, and 2) that the process for delivering its report is too informal and that papers are rushed into print in order to meet the IPCC deadlines. On the first point, this only amounts to an accusation in the event that there is no consensus. Since, in fact, Pat Michaels notwithstanding, there is one, there is little basis for worry on the first point. It's simply tautological. If you are asked to report on a matter on which you are convinced and your reader is not, obtaining the result you hold to be true is not itself evidence of bias.

On the second point, one can make a case that papers are rushed into print specifically in order to be referenced by IPCC. Since "getting the runs in time for IPCC" is the driving force of climate modeling these days, and this distorts the software engineering process, I can actually state confidently that there is some truth to the complaint.

As usual, the forces of truth and justice are caught between a rock and a hard place, though. In demanding a formal process to justify the nontrivial changes in social structure and international relations required by the state of things, people resisting such changes have a solid point. However, they proceed further by also resisting the massive changes in the scale and scope of earth end environmental sciences that would be required by such a process.

Does it matter, though? That depends on whether the "conspiracy" is drummed up or real. It is usually possible to reinforce ideas of conspiracy when there is a segment of the public inclined to believe in one. Whatever error may or may not be involved in selecting certain trees for inclusion in a dendrochronology may constitute malfeasance if one is in a particularly judgmental frame of mind. What, then, is the moral status of quibbling about tree rings when the radiative balance of the atmosphere is being forced at a rate without remote precedent in the entire history of mammals.

In the end, science is an imperfect instrument, and we must nevertheless make decisions based on what we know. By stressing the former and not the latter fact, by fertilizing the ground where others are happy to plant wild conspiracy theories, McIntyre and now Pielke do an enormous disservice.

As such, they are ironically part of the very problem they identify, placing more attention to the advancement of their own reputations and positions than on the advancement of knowledge and governance.

It's literally tragic that they are recycling this endless quibbling about bristlecone pines rather than stepping back and looking at the balance of evidence. There is simply no way to formalize the process entirely. Human judgment is easily derailed, but we will have to collectively judge this issue and come to difficult and necessarily imperfect decisions of major consequence, soon.

If somebody wants to talk about "malefactors", let's talk about the people who are working so hard to skew this matter away from the big picture. It's not about publication records and tenure cases. It's about survival. It's about whether or not to extract so much value from the world that the world itself becomes valueless.

Bristlecone pines or not, the carbon has to stay in the ground.

17 comments:

Dano said...

Willis Eschenbach is a knob. I got into a long, long series of comments with him on Chorus, Amen regarding future agricultural yields and ecological side effects. It all stemmed from him plotting a trend line for ag productivity and calling it good, and gosh, why didn't the moron agronomists think of this, geez scientists are stupid, they should be engineers like me look how smart I was to plot a trendline!!!!

After not being able to make him explain why he wasn't sharing this massive, Galileo-like agronomics breakthrough with the world, I finally opened it up for a bet, to anyone and everyone. Anyone submitting an agron manuscript using this technique and getting it past first cut got $100.00 from me - there was no limit on submissions and the potential for numerous journals to accept for review made it modestly lucrative indeed.

As was to be expected, all I got for a response was a bunch of excuses as to why no one could submit such a manuscript, all along the lines of 'I have to wash their hair for the next 6 months or so'.

----------

The tragedy is that these quibbles are inflated onto accusations of such spectacular dishonesty specifically in order to color the policy debate.

Not good enough, Michael.

Much better is:

The tragedy is that these molehills are inflated into accusations of such spectacularly high mountains of dishonesty specifically in order to color the policy debate.

or

The tragedy is that these few crumbs are trumpeted into descriptions of such spectacularly huge feasts specifically in order to color the policy debate.

etc. Must frame tactics properly.

----------

In demanding a formal process to justify the nontrivial changes in social structure and international relations required by the state of things, people resisting such changes have a solid point.

Yes, because this would delay things further.

IOW: further processizing and bureaucratizing (odd that the anti-gummint types want more bureaucracy, no?) is a delaying tactic.

----------

Bristlecone pines or not, the carbon has to stay in the ground.

- Say, Uncle Morty: DUIs or not, the bottle has to stay on the shelf.

- Cooking meth being dangerous or not, you have to stop doing crank - it's bad for you.

IOW: this is bad framing for addicts, Michael. The oil addicts (whether for cheap energy, easy lifestyle, or big profits) don't want to hear what you - or anybody - has to say if it means they can't have their parenthesized.

Best,

D

tidal said...

Maybe a bit O/T, but speaking of "accusations of spectacular dishonesty"... Yale's Gary Yohe, a lead author of Lomberg's Copenhagen Consensus Project's principal climate paper says this today in The Guardian:

"I can say with certainty that Lomborg is misrepresenting our findings thanks to a highly selective memory.

Lomborg claims that our "bottom line is that benefits from global warming right now outweigh the costs" and that "[g]lobal warming will continue to be a net benefit until about 2070." This is a deliberate distortion of our conclusions...

... In short, we never advocated research into new technologies as a stand-alone way to fight climate change, nor did we accept Lomborg's dismissive attitude toward the threat climate change poses.

The negotiators in Copenhagen will need credible, accurately reported analyses upon which to base their discussions. This is not the time to deny the scope of the problem or belittle efforts to implement solutions. We need all options on the table. This was the message of the Copenhagen Consensus Challenge paper, and even a sceptical environmentalist should understand that."

David B. Benson said...

NOt only stay in the ground, but lots of it needs to be put back in the ground.

About 500 GtC.

bi -- International Journal of Inactivism said...

Michael Tobis:

"Since `getting the runs in time for IPCC' is the driving force of climate modeling these days, and this distorts the software engineering process, I can actually state confidently that there is some truth to the complaint."

I'm quite surprised you'll think that. The problem with this reasoning is, is there any reason why the `no global warming' folks can't also get their own climate model runs in time for inclusion in the next IPCC report? Does UAH's Roy Spencer not have enough resources at his disposal to do his own climate modelling? So this accusation amounts to saying that there's some sort of over-arching Warmist Conspiracy in which editors of scientific journals purposely rush papers predicting global warming into print, and purposely delay or suppress papers which predict no global warming.

* * *

Interestingly, Energy and Environment has indeed been rushing garbage into print for the express purpose of attacking the IPCC -- see the `journal paper' titled "The IPCC Emission Scenarios: An Economic-Statistical Critique".

I'm quite confident that Pielke, instead of "speaking up" against this, will dodge and weave and find some flimsy pretext to avoid talking about it.

duffandnonsense said...

"One wonders who is expected to "speak up" and when."

Well, we could start with James Hansen when he is asked, politely, to provide the codes from which his adjusted figures are produced.

That would be a good start, I think!
David Duff

Anna Haynes said...

Thank you Michael. Especially for these -

"What, then, is the moral status of quibbling about tree rings when the radiative balance of the atmosphere is being forced at a rate without remote precedent in the entire history of mammals[?]"

and

"...by fertilizing the ground where others are happy to plant wild conspiracy theories, McIntyre and now Pielke do an enormous disservice."

Michael Tobis said...

David, that seems to conflate two separate things.

I frankly believe that all public sector computing should be on a 100% open source and open data stack, scientific software or otherwise; some very narrow exceptions for security related work conceivably apply. (It definitely applies to voting machines, though. Many security experts also prefer open source for everything except the passwords.)

You won't find me defending people hiding their calculations. Alas, things don't work that way at present and there are plenty of people motivated to keep it that way for plenty of reasons.

I don't think the buck stops quite as clearly with Hansen as many people propose. It's not as if his job were actually to be chief of software QA.

Atmos found some funky aspects to some GISS data if I recall right. That doesn't mean he subscribes to a grand conspiracy hypothesis.

But all of this is entirely a red herring. The demand was for the community to stand up and be counted in their outrage about Mann et al, and that is just a bizarre idea. If you'd like to defend their position, do so. If you'd like to change the subject, give it a try. But please don't do the one under pretense of the other.

Michael Tobis said...

Frank, you have fallen into the denialist frame!

When I say that the five year cycle "distorts the science" I do not mean that the five year cycle misrepresents the global mean temperature sensitivity, and certainly not that it introduces a bias in that quantity or similar results. (In fact, I suspect the prevailing bias is to understate risks, though I think the most likely outcome is on fairly solid ground.)

I do not think any further refinement of the climate prognosis will come from existing model strategies. Specifically, we need meaningful regional predictions, and we are a ling way from getting them, and the current setup is almost ideally contrived to prevent substantial progress.

Maybe it's a coincidence that the decline of significant progress in computational climatology coincides almost exactly with the emergence of the the IPCC process. I am starting to wonder, though.

This isn't to say it is possible to create a climate model comparable in fidelity to mid-1990s CGCMs which has small sensitivity. Of course nobody thinks that it is possible, or somebody would have done it. I think that is Frank Bi's point, and I completely agree.

I find it interesting that I could be read as making the opposite point.

The existing climate modeling institutions do have turf interests in their climate centers, though. The state of science is not a thing of beauty.

Would fixing it bother the denialists much? Probably not. Conspiracy theories can always be spun out of the thinnest of evidence.

I guess my point here is to acknowledge that while there are a lot of problems with the scientific community, prominent critics of existing climate science are as likely as anyone else to base their behavior on these distorting influences.

Michael Tobis said...

Dano, thanks for your response. I'm not exactly sure which part you wnat me to take up. I think the most interesting part is this:

Me:

In demanding a formal process to justify the nontrivial changes in social structure and international relations required by the state of things, people resisting such changes have a solid point.

Dano:

Yes, because this would delay things further.

IOW: further processizing and bureaucratizing (odd that the anti-gummint types want more bureaucracy, no?) is a delaying tactic.


Of course this is to some extent right. You will find strong support among denialists for the idea that the earth contrives to neatly cancel out global warming with albedo changes (as if this would mean there were no problem), and yet remarkably little support for launching the already built Triana/DISCOVR instrument that could put this idea to the test.

However, it is a mistake to assert either that all opposition to mitigation policy is unanimous. There are some people who simply want a reliable chain of evidence and who aren't convinced they are getting one.

So it's true; we have people eager to point out a problem who tend to be very reluctant to do anything about it.

At first blush, answer to them seems simple enough: pony up the funds and build the structures you would trust. Meanwhile, you'd best go with the best we can do now.

You can install your night vision camera later, but your car is on the foggy mountain road right now. If you don't slow down you won't have anything to attach your fancy new gizmo to later on.

There is no reason to deny that it would still be better if the fog were to lift a bit.

bi -- International Journal of Inactivism said...

Michael Tobis:

"The existing climate modeling institutions do have turf interests in their climate centers, though. The state of science is not a thing of beauty."

I don't know much about the ins and outs of the climate science community, but how different is this from what's going on in the rest of the scientific community, nay, the rest of academia? To me it just looks like the normal rush to publish as soon as possible, period. (Unless there's some special status accorded to papers which are cited in an IPCC report, that is.)

John Mashey said...

I'm of two minds about the effects of schedules on simulation:

it's too bad if bad software engineering happens to try to make a date...

... but on the other hand, having shipped / helped ship / managed the shipping of many software products, having predictable schedules like trains has merit.

At SGI, we went from "we'll ship a new OS whenever some new hardware comes and we have all the new features in" to:

"there will be a regular release every quarter, and if your stuff isn't ready, it will wait for the next train."

Michael Tobis said...

John, I'm all for incremental releases and triage of features in most kinds of software.

Alas, the whole nature of a climate model is tight coupling. The whole advance in software design methodology is loose coupling.

It would make for a damned interesting computer science problem if computer scientists were foolish enough to take on how to properly decouple the code and built and test large families of legitimate models. The efforts to do so (and here I refer primarily to ESMF) seem to be massive train wrecks in progress, largely because of politically expedient and technically ludicrous design decisions made in early phases.

(I would like to take it on myself, in the right company and under the right circumstances. That isn't to say it would be easy or low risk.)

The whole development environment is peculiar. Many lessons from other sectors don't apply directly.

Consider that the IPCC requires two years of RUNS as the end product. The software isn't the focus of the process; it's the output of the software that is the focus. (And of course nobody has ever heard of Hamming's law in these circles; the purpose of the exercise is numbers.)

This means that forty percent of the time at a minimum development is frozen. Plus the runs have to be documented in published literature, which leads to the odd dance that the McIntyreans have caught on to, of papers that will eventually be published being rushed into press to serve various institutional goals.

So now you are down to a 50% duty cycle. Much of this is, as you point out, about new architectures.

In this next round, a great deal of effort has to be put into "compliance" with ESMF standards, which would be one thing if they started from a clean slate but is another thing entirely when it's jammed into extant codes.

Working backwards from that there is a lot of time that needs to be carved out for tuning and CGCM so it's even remotely stable for multicentury runs. (Is the result consequently overly stable? Hmmm.)

The end result is that the physics to be included in the next generation NCAR model targeted at the 2012 time frame is being finalized in the next few months. You may argue that this yields plenty of time to work on 2017, but of course nobody is doing that because the institutional momentum of climate research centers is so severely damaged as a result of the interest the work draws in political circles.

So NCAR's models have some qualitative resemblance to MS Office, with lots of features nobody uses bolted on, bizarre bugs, inadequate documentation, and little substantial improvement in utility or stability since the first functional releases.

I don't know as much about other modeling centers but I think the same pressures operate.

David B. Benson said...

ESMF?

Michael Tobis said...

Time to start a technical thread, methinks.

Google has "Earth System Modeling Framework" as the first hit for "ESMF" for me, though.

John Mashey said...

For any software I've ever seen, there's a "natural" release timing that is short enough to be responsive and far enough apart to actually get some work done. I've seen several years apart for major releases of commercial software (like SAS, which had major correctness/QA issues across platforms).

Is there a proper release interval for climate models/IPCC?

bi -- International Journal of Inactivism said...

Michael Tobis, John Mashey:

The way I see it, part of the problem is that "research" and "development" are mushed together in the case of climate modelling. Is there a way to separate out the development of new kinds of climate models ("research") with the running of these climate models to produce results with practical usage (which more properly falls into "development")?

Marion Delgado said...

Out in the non-net world I have said a lot of what you said here, Michael. It's well-expressed.