Thursday, June 28, 2007

Consensual Pruning

Steve McIntyre at Climate Audit continues to point out the state of the Emperor's Clothes concerning the "science" underlying the IPCC's claim of consensus:
As an IPCC reviewer, I

Show the Briffa et al reconstruction through to its end; don’t stop in 1960. Then comment and deal with the “divergence problem” if you need to. Don’t cover up the divergence by truncating this graphic. This was done in IPCC TAR; this was misleading. (Reviewer’s comment ID #: 309-18)]

In response, IPCC section authors said:

Rejected – though note ‘divergence’ issue will be discussed, still considered inappropriate to show recent section of Briffa et al. series.
What an odd choice of words. "Inappropriate" - not "incorrect", not "inaccurate", not "unsupported", just "inappropriate".

Perhaps it's a "scientific" faux pas to display incongruent data?

Lysenko had it so much easier.

15 comments:

Luther said...

"Inappropriate" indeed.

Great comparison, Lysenko. The path has been well tread. And they were taking notes.

chuck said...

Briffa's 1999 paper is worth reading if you like to see how these things are done, although the graphs are poorly reproduced in the PDF. I didn't think it was bad science, but there is a divergence between the temperature obtained from the tree rings and the observed temperature after 1960, no doubt due to the fact that the tree rings were calibrated against temperature over the period 1881-1960. The divergence for the later period when temperature data is available naturally calls into question the derived temperatures for earlier times where there is no independent temperature record. It also indicates that there may be factors other than temperature determining the width of the rings. These facts may account for the omission of the later data from the IPCC report. I will have to read the paper more closely before saying more.

Rick Ballard said...

Chuck,

On page 2932 Briffa explains the omission of data thusly:

"We used simple linear regression, fitting the the regression equations over the period 1881 - 1960, or over the total available period prior to 1960 when the instrumental record was shorter. The period after 1960 was not used to avoid bias in the regression coefficients that could be generated by an anomalous decline in tree density measurements over recent decades that is not forced by temperature."

How does that not fit into the "if data don't fit, we just omit" mold? Why not present the actual data and then explain how regression analysis is a more perfect tool than compilation of actual data? Why is regression analysis more accurate for the period of 1881-1960 than direct observation with more accurate instruments from 1961-1991?

I certainly don't disagree concerning the existence of an anthropogenic component to warming but I'm unimpressed with the predictive value of studies that trim off 30 years of the "best available" data with a "could" tied to an "anomaly".

chuck said...

Rick,

In defense of the paper, Briffa et al do show all the data on the plots, so the anomaly is pretty clear. On the other hand, the anomaly itself looks like the most interesting bit of data there. It's in tracking down the cause of the bits that don't fit that interesting science occurs. I would also be curious to see what the results look like when the anomalous data is included in the calibration.

I suspect tree ring data may turn out to be a poor proxy for temperature, but it is papers like Briffa's that expose anomalies and the argueing that follows that will clarify the issue. The problem here is the intrusion of politics into the IPCC report, short circuiting the scientific process before it has worked itself out and scientist shave finished yelling* at each other.

*Scientists do fight, sometimes viciously and loudly. My favorite story along those lines is of the two physicists at SLAC who stepped outside to settle their differences the old fashioned way.

Barry Dauphin said...

The warmers seem to worship the goddess of the tree rings, as tree ring proxies have become the postmodern version of immaculate conception (conceived without original sin). And they can be used to reveal the sins of mankind. To many of the warmers, tree rings are unassailable. Maybe if they study the tree rings carefully enough, they'll find the fossilized fingerprints of Adam and Eve.

Rick Ballard said...

"I would also be curious to see what the results look like when the anomalous data is included in the calibration."

That's all I would want. Put in a separate presentation with a "we rejected inclusion because..." and let people scratch their heads.

The big deal with me is the IPCC reply using "inappropriate" as a "reason".

Was that according to commonly accepted Druidic dogma or a direct revelation from Gaia herself?

Luther said...

"inappropriate", again, which I think squares the circle. How does that word appear in scientific discourse? Not that I know a damn thing about science. But really. How audacious.

vnjagvet said...

I hate to put in a lawyer's two cents into this erudite scientific discussion -- BUT [as he does it anyway] I wonder if the data from the tree ring proxy could ever be admitted in evidence in our Federal Court system under the rules set out by SCOTUS in Daubert.

Sounds like the kind of voodoo science the Justices are trying to exclude.

Rick Ballard said...

Vnjagvet,

What "erudite scientific discussion"? Chuck could do one of those but I sure couldn't. I think Briffa would be allowed to testify under Daubert but if Luboš Motl was called as a rebuttal witness, I think his view would prevail.

Unless, of course, Edwards was handling the plaintiff side. By the time he got done channeling poor Gaia in closing the jury might be swayed to the point of bankrupting mankind in service of a "higher" lie.

That Motl piece is pretty amazing. The "ever diminishing impact" portion would make a good post here - by someone that actually understood the physics involved.

[Broadest of hints, Chuck]

buddy larsen said...

Nice discussion.

Heard a blurb on tv that "big law firms" are already "busy forming GW dep'ts", presumably in anticipation of fresh new torts sprouting across the deliberately-damaged-by-deep pocketed-GW-miscreants human landscape. This is scary to me, how 'bout you?

Also blurbed on tv yesterday, this better news (another one bites the dust?).

Rick Ballard said...

Vnjagvet,

I'd like to take an opportunity to 'revise and extend' my remarks re Briffa being accepted under Daubert. McIntyre published another piece today with some claims of cherry picking that, if true, would make Briffa et al dubious candidates for expert status under Daubert.

There's too much "if data don't fit, we must omit" going on. Almost as if receiving continuing grant money were contingent upon showing the "right" results...

vnjagvet said...

That is my impression too, Rick.

Yet we are supposed to make rational political and economic decisions on "science" that cannot pass the elementary red face test.

Ain't it great?

chuck said...

Rick,

As far as I can tell, the cherry picking here consists of three types.

1) There are many series of data and not all of them are correlated with the known temperature records. So the trick is to use those records that *do* correlate with temperature in the hope they also tracked temperature in the past. Making that selection naturally raises the possibility of bias, which, when properly analysed, should be reflected in the estimated error of the result. However, to correctly estimate that error, one also needs to know how well the series used correlated with temperature as well as how small the selection was relative to the whole. The first is where the talk about "red" noise and degrees of freedom come in: there aren't as many *independent* temperature data points as there are measurements. That is, because of the slow trends in temperature, it is possible to make a pretty good prediction of near term temperature given the trend at any time. The ideal temperature series would be random. Estimating the fraction the available data chosen relates to the next point.

2) The data in the literature has already been selected for correlation with temperature. Because the same data keeps getting reused, the different published studies don't independently validate each other, rather, they reproduce whatever errors are present.

3) There was additional selection for results in addition to the two previous points. In addition, the anomolous temperature predictions for the period after 1960 that were Briffa's earlier paper has disappeared in the most recent version. If one accuses Briffa of excising that part of the result in order to sell a conclusion, then one is essentially accusing him of fraud. I would be careful of making that accusation until more information came in.

The cure for the first two problems is the same: more data series and more exacting computation of the errors in the extrapolated series. The cure for the third problem will follow from these.

Science is actually tough. Figuring out how to measure the temperature 1000 years ago and assign an error to the result is not a trivial thing. Such results need a lot of inspection, arguing, and criticism, and to some extent that is what we are seeing. What is disturbing to me is that policy is being made long before the outcome is settled and, as as a result, there are attempts to short circuit the necessary process.

loner said...

Climate is what we expect, weather is what we get.

—Robert A. Heinlein, Time Enough For Love

Rick Ballard said...

"If one accuses Briffa of excising that part of the result in order to sell a conclusion, then one is essentially accusing him of fraud."

As you pointed out, Briffa included the post '60 data in the graph in his paper - it was the IPCC that exceised the data when they reproduced the graph. As to an accusation of fraud, McIntyre is coming very close to that in the second piece that I highlighted in comments. That piece uses the strongest language by McIntyre that I've seen.

I know science is tough and I admire those competent to engage in searching for answers. I'm just not sure that's what's going on in the climate field.