Saturday, October 03, 2009

What Data?

The latest Climate Audit harumph reminds Eli of the original one. A great screech for the DATA. Eli seriously wants to know WHAT data. If it is the actual cores, or their images, that belongs to the Russians who did the coring, not Briffa who they loaned the data to. If so McIntyre is whining up the wrong tree. They are the intellectual property of the Russians and Briffa cannot give them to a third party.

Another point is that the "new cores" were taken by a different group at different times in a different (although perhaps slightly) place. You can't just graft them into the Yamal record as McIntyre tried to do.

Tom P (all credit) has appeared to demolish most of what McIntyre was claiming by doing reconstructions with and without various parts of the Yamal record that McIntyre criticized.


Deep Climate said...

Here's my take on the McIntyre-Briffa controversy. Wading through McIntyre's tiresome posts and comments was tough, but I figured somebody should do it. It should put paid to suggestions that McIntyre did not make explicit accusations of cherry-picking. And, yeah, I think Tom P. got there in the end.

Some sample quotes (the last one is the best one, but you'll have to go to the post to read it, since it's definitely a spam filter risk).

One doesn’t expect Team adjustments to leave even small scraps on the table and this proved to be the case here as well – the added data substantially increased 20th century values and substantially lowered 1150-1250AD values, thereby altering the medieval-modern differential in favor of the 20th century.


I’d be inclined to remove the data affected by CRU cherrypicking but will leave it in for now.


Jacoby, D’Arrigo, whatever other faults they may have, use the entire crossdated population from a site. (They cherry pick sites, but don’t cherry pick trees within a site.)

I’m assuming that CA readers are aware that, once the Yamal series got on the street in 2000, it got used like ***** ******* by paleoclimatologists.

CapitalistImperialistPig said...

OK, maybe I'm stupid, but Eli's second paragraph and tom P's post are opaque to me. Ditto deep climate. Would you mind trying to clearly formulate the point you are making? If the data McIntyre uses is no good, or he uses it in an unreasonable way, why so? somebody asked Gavin the same question and he just referred them to two even more obscure posts.

I ask the same questions in slightly more detail on my own blog.

PSC said...


There's a lot of trees one could use to build a temperature reconstruction with. Some trees show good correlation with temperature, some show poor correlation. There's a whole field of study (textbooks, courses, etc.) on picking good trees.

You want to use the good trees to build a temperature proxy.

Briffa has a technique for picking trees which are correlated with temperature called RCS. I can't comment on how good a technique it is, but people seem to think it's ok.

McKintyre has picked some of Briffa's trees and another bunch of trees from elsewhere and made his own proxy. It's not clear what methodology he used to pick his trees. He used this to create his own graph. He removed some particular trees he did not like. It's not clear what the criteria for removal was.

Tom P went built a a collection of graphs by methodically removing collections of younger trees from Briffa's data. He showed that Briffa's result did not depend on the data getting overwhelmed by very young trees. It's not clear to me that McKintyre was claiming it was getting overwhelmed by young trees. But then it's not at all clear what if anything McKintyre was claiming.

PSC said...

Scrub the above - McKintyre was using RCS; still not clear how he picked his trees.

EliRabett said...

The instrumental record is maybe 140 years long, and seldom even that long at the remote sites where these trees are. You need to use a local instrumental record to get any sense out of the cores as a temperature proxy.

The aim of methods such as Briffas is to chain tree cores together to cover centuries (the older cores are taken from buried logs in this case). That means you have to align the cores from different trees, more an art than a science.

To add to their woes, there is a detectable CO2 fertilization effect, not nearly as large as the Tim Curtins of the world think, but it has to be corrected for in recent years. Too many young trees can overwhelm the match to the local instrumental record while not joining very well onto the rest of the cores, you want the cores that are used for calibration to extend back as far into your proxy as possible.

It's like fitting a function but worse. You don't want too many data points at any one value of x.

Nick Stokes said...

The lead Russian author, Hantemirov, was also a co-author on Briffa2008.

Anonymous said...


My post was in part demonstrating that McIntyre had made various explicit and implicit charges of "cherry-picking" against Keith Briffa and other dendro researchers.

Some, such as RP jr, have not even commented on McIntyre's viscious and baseless attacks. Instead RP jr attacks Gavin Schmidt for calling a spade a spade.

Apparently RP jr hasn't read all of Steve's comments.

EliRabett said...

Given that the data was archived this year that makes sense.

Anonymous said...

(from CIP)

Thanks all, for the clarifications. OK, I think I understand the part about Briffa'a results not being sensitive to exclusion of some of his data, but (if I understand correctly) it does seems to be sensitive to the inclusion of McI's data, in which case the validity and relevance of that data becomes central. I haven't seen a specific, detailed critique of that from anybody yet - not Briffa and not Gavin.

If such a critique exists, it should be front and center. If not, then perhaps jets should be cooled until it does.


luminous beauty said...

For those on Macs who wish to see TomP's results:

As to McCawDitz near decade long whining about being unable to find Briffa's 2000 data, it took me all of ten minutes to locate this:

luminous beauty said...