Cherry-picking in Yamal?

[skierinvermont]

The 'Yamal controversy' surrounds a global temperature reconstruction by Keith Briffa published in 2000:

http://www.climateau...fa.2000.qsr.pdf

The reconstruction uses tree ring chronologies from many sites around the world, one of which is Yamal. In other words, Yamal is one region of many used to create the global reconstruction. The controversy surrounds this single region. The controversy surrounds the reconstruction of post-1950 temperatures in this region. It has nothing to do with temperatures the other 1,950 years or in any of the other regions used to create the global reconstruction.

In the following climate audit post Steve McIntyre shows that the Yamal tree ring chronology is based on only 12 trees post 1988. Prior to 1988 there are many other trees that were used. The full Yamal chronology is formed by stitching together 100s of fossilized trees. But the modern portion of the chronology uses comparatively few trees.

http://climateaudit....climatologists/

In his next post Steve McIntyre goes on to show that when you use another set of data (34 trees) from Yamal published by Schweingruber a completely different result is found than when using the 12 trees used by Briffa. Instead of warming, these 34 trees show cooling. Many have interpreted that Briffa intentionally ignored these 34 trees and cherry-picked the 12 that show warming (whether McIntyre intended this to be implied or not is questionable).

http://climateaudit....rgence-problem/

Keith Briffa has responded to these posts directly.

He accurately sums up the accusation against him:

My attention has been drawn to a comment by Steve McIntyre on the Climate Audit website relating to the pattern of radial tree growth displayed in the ring-width chronology "Yamal" that I first published in Briffa (2000). The substantive implication of McIntyre's comment (made explicitly in subsequent postings by others) is that the recent data that make up this chronology (i.e. the ring-width measurements from living trees) were purposely selected by me from among a larger available data set, specifically because they exhibited recent growth increases.

He goes on to explain that the data for the 17 living trees used (12 post 1988) were provided to him by Rash*t Hantemirov and Stepan Shiayatov. Briffa was attempting to perform an RCS analysis on the same data used by them. This is why the other 34 trees, which come from different researchers, were not used.

This is not the case. The Yamal tree-ring chronology (see also Briffa and Osborn 2002, Briffa et al. 2008) was based on the application of a tree-ring processing method applied to the same set of composite sub-fossil and living-tree ring-width measurements provided to me by Rash*t Hantemirov and Stepan Shiyatov which forms the basis of a chronology they published (Hantemirov and Shiyatov 2002). In their work they traditionally applied a data processing method (corridor standardisation) that does not preserve evidence of long timescale growth changes. My application of the Regional Curve Standardisation method to these same data was intended to better represent the multi-decadal to centennial growth variations necessary to infer the longer-term variability in average summer temperatures in the Yamal region: to provide a direct comparison with the chronology produced by Hantemirov and Shiyatov.

They also go on to explain that the 34 trees used by McIntyre come from a single location (KHAD), whereas the 12 trees they used come from 3 locations (POR, YAD, JAH).

They agree that the KHAD trees should be included (and given appropriate weighting) but note that "However, we simply did not consider these data at the time, focussing only on the data used in the companion study by Hantemirov and Shiyatov and supplied to us by them."

In other words, they did not cherry pick it. They used the data that was available to them provided by Hantemirov and Shiyatov. When you include the Schweingruber (gotta love these names) data the result is changed somewhat for the post 1950 Yamal series. It still shows moderate warming post 1950.

http://www.cru.uea.a...iffa/yamal2000/

http://www.cru.uea.a...iffa/yamal2009/

They have provided a reworked Yamal chronology which includes the KHAD data from Schweingruber brought to their attention by McIntyre. As you can see, when the Schweigruber data is included, the series is ever so slightly cooler in the 1970s and 1980s (black vs blue or red).

[skierinvermont]

And there I was thinking certain posters actually wanted to discuss this.

[WeatherRusty]

The lack of interest in this thread confuses me. Next time we hear the phrase "hide the declined" mentioned as referring to global temperature we will know why. As if we didn't already!

[PhillipS]

And there I was thinking certain posters actually wanted to discuss this.

Well, my opinion is that the certain (im)posters you mentioned realize that this is a technical discussion that they would just emabarrass themselves on. You and WeatherRusty aren't easily impressed by cut and paste nonsense, or links to papers that have nothing to do with this topic. They are much happier posting on threads where they can ignore requests for supporting evidence and can freely attack anyone who disagrees with them.

Have you noticed that tree ring proxies have come up serveral times in recent weeks, typically as a strawman or red herring to the topic at hand? Each time the actual facts are pointed out the pseudo-skeptic dropped the issue. They hold no cards in their hand so if a bluff doesn't work, all they can do is fold.

On a personal basis I found your post to be very informative. I don't know much about dendrochronology but I know a lot of research is going on today. I thnk the controversy over the Mann98 paper stirred up a lot of interest in what was a quiet area of investigation.

Regards - Phillip

[radarman]

Just curious, what does the shaded blue area in the back of the graphs represent?

[PhillipS]

Just curious, what does the shaded blue area in the back of the graphs represent?

If I understood the linked paper, the gray area on the plots is the tree sample counts for the black composite data series. The scale is on the right edge of the plots. And, as you probably noticed, the lower plot is just an expansion of the rightmost portion, 1800 to today, of the upper plot. Also, the upper plot "shows the data smoothed with a 40-year low-pass cubic smoothing spline".

[Strongbad]

Just so that we can have both sides of the story, attached is McIntyre's response:

http://climateaudit.org/2009/10/04/gavins-guru-and-rcs-standardization/

[Sunny and Warm]

Just so that we can have both sides of the story, attached is McIntyre's response:

http://climateaudit....tandardization/

CA must be down. I can't get the link to work, and when I go to CA directly, the site isn't working either.

[Sunny and Warm]

The fact of the matter is that tree ring proxies are problematic for several reasons. Briffa's claim is that the data represents temps changes. Yet, there are other factors which can lead to changes in tree ring growth, such as rainfall, disease, insects, soil nutrients, canopy influences, etc, etc. I'm not a dendro guy, but using tree rings for temp analysis appears iffy.

[PhillipS]

Just so that we can have both sides of the story, attached is McIntyre's response:

http://climateaudit....tandardization/

Did you by any chance notice that the CA blog column was posted before the Briffa paper Skier linked to? And that McIntyre is criticizing an earlier 2008 Briffa paper? If McIntyre feels he has solid concerns about any paper's methods he should submit his analysis to the journal in which the paper was published. Journals have that process for catching errors that slip through peer-review. That's how science works. Of course, the critical comments themselves get reviewed for accuracy before they are accepted, which eliminates ad hom attacks and statistical nit-picking that doesn't change the paper's conclusion.

But McIntyre prefers long rambling posts on his blog to actually participating in the scientific method.. Wonder why that is.

[PhillipS]

The fact of the matter is that tree ring proxies are problematic for several reasons. Briffa's claim is that the data represents temps changes. Yet, there are other factors which can lead to changes in tree ring growth, such as rainfall, disease, insects, soil nutrients, canopy influences, etc, etc. I'm not a dendro guy, but using tree rings for temp analysis appears iffy.

Gosh, a post of yours I fully agree with. You are correct, tree ring proxies have always had limitations, including the ones you mention. The only one I'd add is that they are, at best, regional, not global. This year's tree rings in dought-stricken Texas will look very different than those of trees in, say, Kansas. That's why there are large uncertainty bars on paleoclimate reconstructions.

But tree ring proxies, iffy though they may be, are better than no data at all.

[tacoman25]

Did you by any chance notice that the CA blog column was posted before the Briffa paper Skier linked to? And that McIntyre is criticizing an earlier 2008 Briffa paper? If McIntyre feels he has solid concerns about any paper's methods he should submit his analysis to the journal in which the paper was published. Journals have that process for catching errors that slip through peer-review. That's how science works. Of course, the critical comments themselves get reviewed for accuracy before they are accepted, which eliminates ad hom attacks and statistical nit-picking that doesn't change the paper's conclusion.

But McIntyre prefers long rambling posts on his blog to actually participating in the scientific method.. Wonder why that is.

I don't think that's a fair accusation. I haven't really gone to CA much recently, but from what I've heard and seen, McIntyre seems to use scientific analysis pretty rigorously. He also does not accept or go along with some of the silly/non-scientific tangents that the WUWT crowd seems to.

[tamarack]

Gosh, a post of yours I fully agree with. You are correct, tree ring proxies have always had limitations, including the ones you mention. The only one I'd add is that they are, at best, regional, not global. This year's tree rings in dought-stricken Texas will look very different than those of trees in, say, Kansas. That's why there are large uncertainty bars on paleoclimate reconstructions.

But tree ring proxies, iffy though they may be, are better than no data at all.

As I noted in a thread long expired, my short term measurements (in a moist and temperate climate, therefore of limited applicability) show tree diameter growth to be affected far more by drought than by temperature. Ideally, dendrochronology (for temperature reconstruction) should utilize trees from locations where moisture is rarely limiting but temperature often is.

Even there the results may not be all that clear. Most tree species will survive, and even grow well, in a climate far too cold for them to mature a seed crop. Therefore, the continuing multi-generational presence of a tree species probably means that the temperature regime (and moisture as well) lies well within the range where an established individual tree could survive and grow.

Fuzzy data is usually better than none, as long as the data is not so skewed as to mislead. In that earlier thread, there were responses to my post (about drought vs temp) which gave some confidence that the data has sufficient validity to be useful, not so much as to be the main support of one's position.

[PhillipS]

I don't think that's a fair accusation. I haven't really gone to CA much recently, but from what I've heard and seen, McIntyre seems to use scientific analysis pretty rigorously. He also does not accept or go along with some of the silly/non-scientific tangents that the WUWT crowd seems to.

Well, I should have made it clear that I was only expressing my opinion. But take the CA post that S&W linked to - by my count there are about 30 paragraphs, of which only 5 or 6 are directly related to the topic of whether younger trees should be excluded from tree ring proxies. The rest of the column is an uninformative rant against Gavin Schmidt and a commenter called Tom P. How many folks really care if his feelings were hurt by something Gavin wrote? We want the science, dammit!

Few, if any, people question McIntyre's statistical expertise. He knows a lot. It is just my opinion, but I feel that McIntyre would reach a larger audience if he adopted a more focused approach when auditing a researcher's work. Something along the lines of:

Here's what they did

Here's what they could (should) have done

Here's how it affects their conclusions.

He should leave out 95% of the background verbiage and 100% of the personal disputes, It would also be refreshing if he was an 'equal opportunty' auditor and used his laser focus on skeptical papers, too. By just focusing on mainstream papers he makes it too easy for people to lump him with the denialist fringe.

[tacoman25]

Well, I should have made it clear that I was only expressing my opinion. But take the CA post that S&W linked to - by my count there are about 30 paragraphs, of which only 5 or 6 are directly related to the topic of whether younger trees should be excluded from tree ring proxies. The rest of the column is an uninformative rant against Gavin Schmidt and a commenter called Tom P. How many folks really care if his feelings were hurt by something Gavin wrote? We want the science, dammit!

Few, if any, people question McIntyre's statistical expertise. He knows a lot. It is just my opinion, but I feel that McIntyre would reach a larger audience if he adopted a more focused approach when auditing a researcher's work. Something along the lines of:

Here's what they did

Here's what they could (should) have done

Here's how it affects their conclusions.

He should leave out 95% of the background verbiage and 100% of the personal disputes, It would also be refreshing if he was an 'equal opportunty' auditor and used his laser focus on skeptical papers, too. By just focusing on mainstream papers he makes it too easy for people to lump him with the denialist fringe.

Fair enough. Like I said, I haven't read CA much recently so I'll take your word for it. I do know he and Schmidt have had a long-running feud and both have personally attacked the other.

[Sunny and Warm]

But tree ring proxies, iffy though they may be, are better than no data at all.

I disagree. Proxies that are flawed end up leading people down the wrong path, and takes longer to get to the the real answers later on and influence grant money disbursement that could be better used elsewhere. Bad proxies pollute the sciences, and this science in particular is polluted with too many of them.

[Sunny and Warm]

Well, I should have made it clear that I was only expressing my opinion. But take the CA post that S&W linked to - by my count there are about 30 paragraphs, of which only 5 or 6 are directly related to the topic of whether younger trees should be excluded from tree ring proxies. The rest of the column is an uninformative rant against Gavin Schmidt and a commenter called Tom P. How many folks really care if his feelings were hurt by something Gavin wrote? We want the science, dammit!

Few, if any, people question McIntyre's statistical expertise. He knows a lot. It is just my opinion, but I feel that McIntyre would reach a larger audience if he adopted a more focused approach when auditing a researcher's work. Something along the lines of:

Here's what they did

Here's what they could (should) have done

Here's how it affects their conclusions.

He should leave out 95% of the background verbiage and 100% of the personal disputes, It would also be refreshing if he was an 'equal opportunty' auditor and used his laser focus on skeptical papers, too. By just focusing on mainstream papers he makes it too easy for people to lump him with the denialist fringe.

I agree in that both sides are guilty of this, but it has hurt the alarmist side more, thus the polling data ...

[skierinvermont]

Just curious, what does the shaded blue area in the back of the graphs represent?

If your referring to the graph that I included in my post and not one of the ones contained in the links, the shaded area (looks grey to me) represents the # of trees used for that portion of the graph. As you can see, once you get to 1980, the # of trees used drops off dramatically.

EDIT: just noticed Philip already responded

[skierinvermont]

Just so that we can have both sides of the story, attached is McIntyre's response:

http://climateaudit....tandardization/

This is not McIntyre's response to Briffa, this is his response to an internet poster named Tom P.

Briffa doesn't agree with Tom P anyways, and says that Schweingruber should be included and that when it is included, you see slightly less warming. He also says that he did not have Schweingruber available to him at the time and that McIntyre's accusation of cherry picking is false.

If somebody can find a McIntyre response to Briffa (instead of Ron P.), that would be of interest.

EDIT: I see Philip already pointed this out too haha

[skierinvermont]

The fact of the matter is that tree ring proxies are problematic for several reasons. Briffa's claim is that the data represents temps changes. Yet, there are other factors which can lead to changes in tree ring growth, such as rainfall, disease, insects, soil nutrients, canopy influences, etc, etc. I'm not a dendro guy, but using tree rings for temp analysis appears iffy.

Which is why they use trees from temperature limited environments, such as the northern edge of the treeline, and mountains.

[PhillipS]

I disagree. Proxies that are flawed end up leading people down the wrong path, and takes longer to get to the the real answers later on and influence grant money disbursement that could be better used elsewhere. Bad proxies pollute the sciences, and this science in particular is polluted with too many of them.

We may be talking past each other. My point is that if the choice is between poor data and no data at all, use the poor data and keep looking for better data to improve the analysis. I think your point is that if there is better data available, don't use the poor data at all because it will skew the results. If I've understood you accurately then I think we're on the same page.

I think we all agree that data can never be perfect, i.e. all data has measurement errors and uncertainty, and that proxy data is more uncertain that direct measurement data.