So, Briffa has responded to McIntyre and there's another riposte up at Real Climate. I'll try to explain what is going on.
Let's first remind ourselves of the guts of McIntyre's argument. This is that Briffa had an very small set of tree ring cores in the latter years of his Yamal series and that when you removed this and replaced it with a somewhat larger set of data from the same region, the uptick in the hockey stick shape disappeared. Therefore Briffa's results weren't robust.
Now we'll look at Briffa's response.
Cherrypicking
Firstly he says McIntyre is implying that he, Briffa, cherrypicked uptrending series so as to get a hockey stick. In reply, McIntyre quotes what he said in his early post:
It is highly possible and even probable that [Briffa's] selection is derived from a prior selection of old trees described in Hantemirov and Shiyatov 2002...
and also a comment he made on another of the Yamal posts
It is not my belief that Briffa crudely cherry picked.
This seems to refute Briffa's accusation that McIntyre was implying malfeasance.
Standardisation
From a scientific perspective, this part of the debate has moved us forward slightly, in that Briffa has now confirmed that the selection of the 12 cores from the much larger population available was due to the Russians. What the reason was for their only using 12 cores remains a mystery. Briffa's response has, however, opened up a new part of the debate that I've not touched on before - this concerns standardisation of the raw tree ring data.
During its lifetime, a tree does not grow at a uniform pace. Tree rings are generally wider when a tree is young than when it is older. If you are using a set of tree rings in climatology, therefore, unless you do something about it, your "treemometer" would always show declining temperatures, regardless of what is going on in the outside world. Standardisation is the process by which this fix is applied, and it involves removing a kind of "average growth curve" from the record to adjust for these changes in growth rate. There are various ways of doing this, the details of which are beyond my ken, but as I understand it, the Russians used the "corridor" method. This works well when you have small numbers of tree cores so it was presumably a suitable choice.
The problem with the corridor method is that it tends to obscure long-term trends in the data, which is precisely what you're interested in when you are doing paleoclimate work. Because of this, when Briffa picked up the seventeen cores for use in his version of Yamal, he applied a different standardisation procedure called RCS, which is better suited to the retention of long-term information.
My application of [RCS] to these same data was intended to better represent the [long-term] growth variations ... to provide a direct comparison with the chronology produced by Hantemirov and Shiyatov.
This is problematic. RCS is not suited to dealing with small numbers of cores (I recall reading somewhere that it is not considered suitable with less than 50, but I'm not swearing to that). I also wonder about the nature of Briffa's paper. It strikes me that if the purpose was to "provide a direct comparison with the chronology produced by Hantemirov and Shiyatov", then there can be no arguing with the use of the same data. However, review of Briffa's original paper from 2000 suggests that intercomparison of standardisation methods was not part of his purpose at the time. The paper is a review of developments in paleoclimate and the calculation of a new temperature reconstruction using some of the new data. This being the case, the logical thing to do would surely to have used as much data as possible.
Robustness
Briffa's other concerns are with McIntyre's sensitivity test - replacing the 12 Briffa cores with the Schweingruber 34 - different cores taken from the same area. This is how he puts it:
The basis for McIntyre's selection of which of our ... data to exclude and which to use in replacement is not clear... He offers no justification for excluding the original data.
McIntyre's comeback on this is that he was very clear about the reasons for excluding the Briffa 12, namely that the number of cores was small. He wanted to test the robustness of the answer by swapping in a larger dataset that had not been used by Briffa.
As well as seeing what happened when the Briffa 12 were swapped for the Schweingruber 34, McIntyre also did a slightly different calculation to see what would happen when both were put in the mix. Briffa says that when McIntyre did this, he underweighted the 12 (i.e. making the loss of hockey stick shape more marked than it should have been). McIntyre has pointed out that Briffa has equally underweighted the Schweingruber 34 by not using them at all, and that debate about the weights doesn't affect the main point, which is that using the Schweingruber 34 makes the hockey stick shape disappear.
Briffa continues
Whether the McIntyre version is any more robust a representation of regional tree growth in Yamal than my original, remains to be established.
This one has been doing the rounds for years. McIntyre has been clear from the start that he is not creating an alternative reconstruction. He is testing "official" studies for robustness.