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)

Thursday, July 23, 2009

If She Weighs the Same as a Duck, She's Made of Wood

Update 7/28/09: This rather long exposition was written on my first exposure to the paper. One doesn't know what to expect, and of course tries to keep an open mind to new ideas especially after they have passed peer review. I have gotten a fair amount of attention from this but in retrospect I find it a bit balky, and wished I had made the point more succinctly.

Both my point (focusing on the tortured logic) and Tamino's (focusing on an obvious and demonstrable weakness of the tortured logic) are made very effectively in a brief comment on the RealClimate article by Ron Taylor which I quote here in its entirety. For most purposes this is all you really need to know about the new paper, Influence of the Southern Oscillation on tropospheric temperature, (McLean, de Freitas and Carter).
I have been reading and rereading this post and the comments, but still find the whole thing strangely puzzling. It seems to me that any good first year calculus student should be able to quickly find the fatal flaw in the methodology as a way of explaining temperature change (as opposed to temperature variability, since it discards any secular component). So I thought they were being really clever by making a valid correlation of SOI with temperature variability, then subtly changing the language to temperature variation. Many people, including journalists, would conflate temperature variation and temperature change, with no further effort required on the part of the authors. It looked like an example of: “If you can’t convince them with facts, then dazzle them with footwork.”

But then they claim in the press release that it actually explains temperature change, and they do so with no apparent embarrassment. Unless I am really missing something it seems incomprehensible that anyone in the scientific community would take this paper seriously.



Original posting begins:

The denialists are making a big deal about the new paper, Influence of the Southern Oscillation on tropospheric temperature, (McLean, de Freitas and Carter).

It's a data analysis paper, and pretty much a no-brainer. It shows that global mean temperature variance is dominated by the El Nino/Southern Oscillation index (which most climatologists would have said was obvious) and that the lag is about 7 months, with ENSO leading the temperature (a modestly useful result).

The odd bit is toward the end:
[33] Chapter 3 of the Working Group I contribution to the Fourth Assessment Report by the Intergovernmental Panel on Climate Change (IPCC) [2007] notes the strong relationship between the ENSO and various climate phenomena, including surface temperature. Chapter 8 reports “considerable model skill out to 12 months for ENSO prediction.” Chapter 9 raises the question of ENSO response to continued anthropogenic warming, although causal connections are not clear. However, nowhere does IPCC [2007] mention the delay between a change in ENSO and a corresponding change in GTTA, which we have shown here to be 7 months. Climate modelers acknowledge that their models do not adequately hindcast average global temperatures from 1950 to 1990 and apply a human influence factor to make up the deficit. The strength of the time lagged relationship between ENSO and GTTA, as demonstrated here, suggests that variation in the poorly modeled ENSO may account for the deficit and may be the cause of a large part of the observed warming since the midtwentieth century. The sequence of the lagged relationship indicates that ENSO is driving temperature rather than the reverse. Reliable ENSO prediction is possible only to about 12 months [IPCC, 2007], which implies that accurate temperature forecasting beyond that period will only be possible after improvements in ENSO prediction.

...

[36] Since the mid-1990s, little volcanic activity has been observed in the tropics and global average temperatures have risen and fallen in close accord with the SOI of 7 months earlier. Finally, this study has shown that natural climate forcing associated with ENSO is a major contributor to variability and perhaps recent trends in global temperature, a relationship that is not included in current global climate models.
This builds loosely on

[29] The Great Pacific Climate Shift of 1976 [Guilderson and Schrag, 1998] appears to have had an effect on the SOI and therefore on the GTTA. The abruptness of this change suggests a major physical event, but no definite cause has been established. McPhaden and Zhang [2002] report a significant decrease in the estimated cold-water upwelling since that time. This implies a warmer Pacific Ocean and a bias toward El Niño events, but it is unclear whether this is a cause or consequence of the shift. Figure 4 shows that the change in SOI, in February or March of 1976, was followed by a corresponding change in GTTA in the last few months of the year. The granularity of the RATPAC-A GTTA data prevents clear identification of the time of change but it appears to be consistent with the time lag reported here.

[30] For the 30 years prior to the 1976 shift (i.e., 1946–1975) the SOI averaged +1.93 but in the 30 years after 1976 (i.e., 1977–2006) the average was −3.06, which represents a shift from a La Niña inclination to an El Niño inclination. The standard deviations for the two periods were 9.48 and 10.40 on monthly SOI averages, and 6.56 and 6.35 on calendar year averages, which indicates consistent variation about a new average value. Only the RATPAC-A data are available for lower tropospheric temperatures both before and after this shift, and even then we are limited to 17-year periods for our analysis of RATPAC-A data because monitoring did not commence until mid-1958. From 1959 to 1975 the RATPAC LTT averaged −0.191°C and from 1977 to 1993 it averaged +0.122°C. The standard deviations on the seasonal data were 0.193° and 0.163 C°, and on monthly data 0.162°C and 0.146°C. We have already illustrated the close relationship between SOI and GTTA, but this description of the respective changes before and after the Great Pacific Climate Shift indicates a stepwise shift in the base values of each factor but otherwise relatively consistent ranges of variation.

[31] The findings presented here are consistent with the Southern Oscillation being a major driver of temperature anomalies, not only in the tropics but also on a global scale. In passing we also note that according to the relationship between SOI and GTTA, 1983–1984 would likely have been about as warm as 1998 if not for the cooling influence of El Chichón.

Now we see the authors making claims like this in informal venues:

"The surge in global temperatures since 1977 can be attributed to a 1976 climate shift in the Pacific Ocean that made warming El Niño conditions more likely than they were over the previous 30 years and cooling La Niña conditions less likely" says corresponding author de Freitas.

"We have shown that internal global climate-system variability accounts for at least 80% of the observed global climate variation over the past half-century. It may even be more if the period of influence of major volcanoes can be more clearly identified and the corresponding data excluded from the analysis.”

...

Bob Carter, one of four scientists who has recently questioned the justification for the proposed Australian emissions trading scheme, says that this paper has significant consequences for public climate policy.

and

"The close relationship between ENSO and global temperature, as described in the paper, leaves little room for any warming driven by human carbon dioxide emissions. The available data indicate that future global temperatures will continue to change primarily in response to ENSO cycling, volcanic activity and solar changes.”

“Our paper confirms what many scientists already know: which is that no scientific justification exists for emissions regulation, and that, irrespective of the severity of the cuts proposed, ETS (emission trading scheme) will exert no measurable effect on future climate.”

and

"When climate models failed to retrospectively produce the temperatures since 1950 the modellers added some estimated influences of carbon dioxide to make up the shortfall," says McLean.

"The IPCC acknowledges in its 4th Assessment Report that ENSO conditions cannot be predicted more than about 12 months ahead, so the output of climate models that could not predict ENSO conditions were being compared to temperatures during a period that was dominated by those influences. It's no wonder that model outputs have been so inaccurate, and it is clear that future modelling must incorporate the ENSO effect if it is to be meaningful."

So was the purpose of the paper to extract a useful correlation (7 month lag)? Or was it to construct a polemic? Did the reviewers not catch the little propaganda digs at the end? Are those digs justified by the paper? What about the broad conclusions being quoted in the denyosphere?

Here's how the argument goes as I see it:
  1. We identify the lag between ENSO index and temperature, and reassert the obvious fact that ENSO accounts for most of the variance in the global temperature record
  2. We point in the direction of an abrupt shift in Pacific temperatures in 1976, and assert that this correlates with a shift in the ENSO index, and that the causality is not well known.
  3. We acknowledge that IPCC has always said the obvious fact that ENSO dominates temperature variance, and make a few random apple pie statements about ENSO predictability
  4. Glommed onto the above paragraph we assert that CO2 sensitivity has been used as a fudge to account for unforced climate models failing to reproduce the temperature signal! ("Climate modelers acknowledge that their models do not adequately hindcast average global temperatures from 1950 to 1990 and apply a human influence factor to make up the deficit.") ??
  5. ENSO leads global temperature on a seasonal time scale, therefore this "suggests that variation in the poorly modeled ENSO may account for the deficit and may be the cause of a large part of the observed warming"! ??
  6. Accurate temperature forecasting beyond ENSO predictability horizon is impossible (well, on ENSO time scales, obviously... or...?)
  7. If ENSO forces temperatures and GCMs can't predict ENSO, therefore GCMs can't predict temperature. "natural climate forcing associated with ENSO is a major contributor to variability and perhaps recent trends in global temperature, a relationship that is not included in current global climate models" ? "perhaps trends" ?
Note that 6 and maybe 7 are reasonable statements in the context of ENSO predictability but not in the context of policy time scales. These are at the ends of the paragraphs and might well pass the eye of the ENSO-centric reviewer.

Points 4 and 5 though are pretty much meaningless in the ENSO context, but are huge whoppers in the context of policy time scales. Points 4-7 can easily be read as if they were a coherent counter-theory of global warming, and in retrospect were obviously intended to be so read. But points 4 and 5 are sort of buried where the hassled ENSO expert might glaze right over them.



I suggest that points 4 and 5 should not have passed peer review. The reviewers, unfamiliar with the deniosphere or the pseudocontroversy can be forgiven for missing the double meanings of points 6 and 7. And who expects a little piece like this of ulterior motives?

But points 4,5,6,7 can be strung together to make a case that the modest evidence presented in the paper doesn't remotely support. The actual result is really not involved at all save as camouflage for this "disproving" of "global warming". Have I missed something?

(Update: yes, I missed something, but it wasn't in the article's favor! The data are bandpassed and the trend is totally erased in preprocessing! Tamino)

We are going to be stuck with this stinker for a while. This is new and different; it seems the authors were aiming to score propaganda points as much as to publish a modest result. If I read that right it's very unfortunate. I hope this sticks to the reviewers, but of course it won't. There are no consequences for bad reviewing. That's a core principle of science journals. Otherwise the free labor would go away.

Update: Kiwi press coverage: "Climate Change Down to Nature"

Update: See also this follow-up. I argue that the denialists are making a mistake by latching on to this publication.

Update: Drudge points to the Washington Examiner on the story. Still no major media coverage, which I think is appropriate.

Update: Thanks to RealClimate for linking here. Please don't miss Tamino's article pointing out that the McLean paper can't possibly have anything to say about trends since the trends are removed in the processing.

Update: What is going on is a bit subtle, and this was written in haste, so it;s little wonder a lot of folks are missing the point.

The paper identifies the correlation lag between a bandpassed correlation of ENSO index and global mean tropospheric temperature. The methodology is dubious but that us secondary. The reason this is news is that they try to connect this result with a global warming trend.

Tamino offers a proof that the connection is meaningless.

My contribution was to attempy to explain how the design of the paper is cynical. I see that it is not going to be easy to explain this very well.

The ordinary result about El Nino is merely a vehicle. The payload is a bunch of unjustified conclusions toward the end, which are written in such a way as to resemble the usual bread and butter commentary of the El Nino community with a bit of modest speculation thrown in. The reviewers only let this pass because they did not notice the possibility of drawing on this to make controversial conclusions.

It's science by double entendre.

The trick being undertaken here is subtle, and explicitly draws upon the gap between the scientific community and the rest of the world, and specifically on the obsessions of the El Nino prediction community.

It doesn't look accidental to me. I see it as a successful attempt to slip something past the reviewers. Compare the modest claims of the abstract (which the reviewers would be thinking about) with the concluding paragraphs.

The abstract concludes:

"That mean global tropospheric temperature has for the last 50 years fallen and risen in close accord with the SOI of 5–7 months earlier shows the potential of natural forcing mechanisms to account for most of the temperature variation."

which is what any physical climatologist would have expected with or without the present paper. In other words "this is normal science, nothing controversial here, these are not the droids you are looking for".

The claim in the abstract has nothing whatsoever to do with the global warming trend, but only with the interannual variance.

Only a little bit of sloppy speculation in the paper even leans in the direction of talking about the trends. (Tamino's analysis shows how unjustified that speculation is). What I'm trying to say is that the paper was presented in such a way as not to emphasize that sloppy speculation at all, until it got published, when it was suddenly all over the press.

This is not just a lousy paper. It's something much worse. It's cleverly manipulative.

Update 7/26: Investors' Business Daily takes the bait.

25 comments:

Penguindreams said...

I think they slipped a couple things past you, which suggests, I guess, how it got past the reviewers too.

You describe it, as they usually do, as correlating SOI and global mean temperature. If that were what they did, it's same old same old, and the correlations are a lot smaller than they report. While SOI is the leading term in explaining global mean temperature variance, it doesn't explain much.

What they actually correlate is the processed SOI and processed global mean temperature. That processing being to take the 12 month delay difference of the 12 month moving averages, except when (see paper) they consider a period to be volcano-influenced. No justifications or consideration of effects was made of the averaging period, the differencing period, or the way they treated volcanoes.

The processings they apply amount to a spectral filter that (I think, to be established yea or nay in my blog note to come) is extremely predisposed to correlate any two time series, regardless of content. The moving average trims out the high frequency, the delay differencing trims out the low (but in a highly frequency-dependent way). Any series with anything happening in the intermediate frequencies will correlate pretty well.

Tomorrow my introduction to time series analysis will appear at my blog (nothing that's likely to be news for Michael, but I'll advertise anyhow). Probably Monday will be the application to this paper.

Michael Tobis said...

So much the worse, then.

But even if the signal processing isn't cooked, there is absolutely no way to go from this temperature phase lag to "CO2 was slapped in to cook the books" and "ENSO suffices".

They don't even lift a finger to argue that the Pacific shift was unforced. It might well not have been.

The whole thing is very shabby. It could and definitely should end up much more a blow to the publishing system than to mainstream climatology. But I expect the propaganda effect will be as intended.

Dan Satterfield said...

I think I get it. It's right out of Orwell's 1984.
Sudden change of enemies.

"It's the Sun stupid" is all wrong.

It's El Nino Dummy.

Yeah that's the ticket...

David B. Benson said...

Might care to look at Tamino's demolishment of this paper on Open Mind.

Michael Tobis said...
This comment has been removed by the author.
Michael Tobis said...

"More to the point, they don’t compute correlation between temperature and SOI, but between the time derivatives of temperature and SOI. They estimate time derivatives by taking the difference between 12-month running averages, 12 months apart."

mt (slapping forehead):

This was staring me (and you) in the face. I actually thought to look for trend removal, since one would, actually, do that in a case like this, if one were interested in the phase relation of a correlation of an oscillation and a response to it.

Somehow I missed that they just highpassed it to zero in the derivative step. D'oh!

Anyway, the rest of it was stupid enough.

Dano said...

First of all, I did a drive by and skimmed this post, and I saw the de Freitas name. This is the first cause for alarm, as this has been done before. IIRC de Frietas went on a Heritage Victory Tour a là Baliunas and Soon after they wrote a seemingly robust paper. Again, this is after some good Merlot tonite.

Then I see the third author is Bob Carter. Well, h-e-double-hockey-sticks, boah. That feller's loonier'n a Bantam near a fox bitch lookin ta feed her kits. But I appreciate your taking the time on it and acting as if it were serious, at least for a referent in the whack-a-mole game.

Best,

D

(wordpress agrees, as she says 'picle')

Michael Tobis said...

The blogosphere has to act fast on this, in the hopes that the press not be too quick to take it any more seriously than it deserves, which is not at all.

JGR is supposed to be a real journal. Words fail me. What a travesty.

Scruffy Dan said...

The words of John Mashey seem particularly relevant:

"The publication cycle of the most credible peer-reviewed journals is long enough that a non-expert should be prepared to be wary of any paper only 1-2 years old, especially if it has novel implications counter to mainstream established science."

Though given what you, and Tamino (Penguin I look forward your your debunking) has shown my money is that we wont have to wait 1-2 years to see how this plays out.

Scruffy Dan said...

My money is also on the fact that I am a terrible proof-reader

thingsbreak said...

De Freitas is obviously aware of what he is doing, and more or less outright admits his goal is to prolong the "debate" [from MT's stuff.co.nz link]:

De Freitas said the research would cause a stir in the scientific and environmental communities.

The paper had been peer-reviewed over six months and accepted by a top academic journal.

"This will have an effect. This is scientific research, not an opinion," de Freitas said.

"There will be people who will be forced to correct me, no doubt, but that is what science is all about it's all about robust debate."

guthrie said...

RObust debate? Can I expect you good fellow who work in the climate area to deluge this 'journal' with letters and alternative analyses showing why they are wrong and are cooking the books?

Michael Tobis said...

They didn't even look at the books they cooked!

Please read my summary carefully. It is not technically difficult. Please also consider Tamino's argument, which likewise should be accessible to an intelligent person with a high school education.

Overall, there are two possibilities: 1) naysayers have difficulty getting published because the scientific community is uniformly biased; 2) naysayers have difficulty getting published because they are demonstrably wrong.

The present paper is clearly a point in evidence in the latter class. Unfortunately, this one is so dreadfully wrong that it says bad things about science as a whole, not about bias but about standards.

One of the authors, at least, has been involved in comparable controversy before. The climate statistician Hans von Storch, who is not by any means a climate change alarmist, found himself resigning as editor of a journal which would not withdraw a de Freitas publication.

So, far from it being too hard for messages espousing a point of view to get published, it is much too easy. It's clear von Storch has a point.

guthrie said...

Yes, I did read your summary, and Tamino's, and the maths is beyond me but I do agree that the denialists are incompetent.

All I want to see (apart from mass media ridicule of the authors and the journal for printing it) is published refutations, so that when I next run into people who say "oh but its El nino that does it" I can say no its not, here's the papers saying why and if you don't understand any of it I can talk you through it to help you understand it. (Oddly enough no one whom I have offered to help in such a way has ever responded positively)

MAss media ridicule takes a bit longer to sort out, but if you guys who are involved directly in the science can give us the hammers, we science lovers can go out and bang some nails in.

Marion Delgado said...

The energy collective site and the repost are very neat.

Steve Bloom said...

It seems impossible that the JGR editors and the reviewers for this paper wouldn't have had any inkling that they should pay extra attention to anything submitted by de Freitas or Carter, let alone both together. I conclude that the publication of this paper is a feature, not a bug.

Philip Machanick said...

Welcome to the Monty Python school of scientific methodology. See another example here where I debunk junk someone quoted from a popular book in Australia.

Michael Tobis said...

That's an inauspicious if amusing coincidence, isn't it?

Thanks for the reminder of the Hungarian phrasebook sketch, one of my all time favorites!

Obviously we agree on the nature of the "debate", but how on earth are we going to convey it to the public or to decision-makers?

Horatio Algeranon said...

ENSO it goes

Hank Roberts said...

What I said at Tamino's:



I agree with MT as well.

1) Get paper published in reputable journal
2) Make baseless claims in press conferences
3) Defend those claims on others’ blogs
4) Make vague retraction in blog comments
5) Do not call press conference for retraction
6) Do not correct published work
7) ….?

Steve Bloom said...

Michael, I don't think IBD should be described as "taking the bait." They are a bait shop.

King of the Road said...

I have to say that, from a political/public relations point of view, the McLean, de Freitas and Carter paper strategy is brilliant.

They subtly got their paper published in a peer-reviewed journal. They then misdirected the public regarding the results of their study in press releases, etc., thereby allowing the conservative blogs sites and talk radio stations to claim that the "real science" is finally coming out in the primary literature just as the "warmies" demand, in such journals. I'm hearing and reading about it all over, in a "see? we told you..." context.

There's no way out, even were the Journal to publish some sort of clarifying information, the skeptical camp would present it as an example of political pressure from the "liberal AGW conformists," leaving the scientific community to look like thugs.

It's an admirably clever strategy.

Michael Tobis said...

Except that by actually feeding the press with press releases, they expose their strategy. This is necessary, of course. The press usually misses important science. You have to hit them over the head with it.

This in turn is why it is necessary to respond to stuff like this fast. The no-such-thing press will have already picked it up (that is where I noticed it in the first place) but the idea is to alert the mainstream how fishy it all is and prevent them from running with it. So far we have succeeded in that, I think. But yes, we are playing defense, and science is being gamed for nonscientific goals.

It may be brilliant, but it shows that nothing is sacred to those people.

The Septic Tank Blog said...
This comment has been removed by a blog administrator.
The Septic Tank Blog said...
This comment has been removed by a blog administrator.