Hockey Stick Redux

Before anyone reads any further, I want to offer up an extraordinary disclaimer: What I’m about to say should be taken as the view of a relative outsider.  I’m approaching the following issue as a philosopher, hopefully impartially, and looking at the reasons that are offered by (at least) two parties.  I want to discern who has the burden of proof, who has to demonstrate what, given the scientific state of affairs.  In particular, I want to respond to growing excitement in the community of those who are skeptical about the data on climate change (a.k.a. the Skeptics, Deniers, and Denialists).  I don’t want to introduce a discussion about the climate science itself, since I’m not really qualified to do so, except as a well-educated non-specialist.  Further, as a well-educated non-specialist, my tendency is to defer to established scientific record — which in this case is the IPCC AR4, as well as most of the climate science following from that — and to make judgments about the reliability of challenges to this record based on my understanding both that shenanigans happen in the peer review process and the scientific record is open to amendment.

I can’t give the whole back-story on the issue I’m responding to, as it’s way too convoluted.  Best to refer to my colleague Roger Pielke Jr. for his take on things, and then to refer to McIntyre’s original post, followed by Bishop Hill’s relatively straightforward, albeit extremely entertaining, accusation that some very famous climate data was cherry-picked (or insufficiently and unjustifiably employed, depending on your preference).  When you’re done with that treat, follow it up with a heavy shot of RealClimate.  Roger thinks RealClimate is unjustifiably bristly.  I’m not so sure.

Point being: I’m really stepping into the thick of things here.  Not sure why.  Sadism, perhaps.  Stupidity, maybe.  Too much wine, most likely.  This issue has just caught my attention.  Anyone who would like to fillet me should feel free, though they should know that even with my extraordinary disclaimer, my skin ain’t that thin and I can usually hold my own.

Now, on to it.

As I’ve said, I’m not familiar enough with the history of this conflict, but from what I’ve gathered, my immediate reaction is that McIntyre has a pretty steep hill to climb.  What has to be determined are Briffa’s reasons for selecting the data that he selected (and by “Briffa’s reasons,” I don’t mean “reasons that he actually held,” nor necessarily “reasons that he believed,” but rather, “reasons that best support and explain his actions”).  Scientists like Schmidt at RealClimate like to ask for evidence — that’s what they do, they’re scientists — but what Schmidt is asking for (or, more aptly, demanding) is a smoking gun that suggests that Briffa had bad reasons for selecting the data that he did.  That’s much harder to demonstrate, and I think McIntyre basically can’t demonstrate such a thing.  People, even scientists, don’t always air their reasons, particularly if they’re not challenged up front.  There’s nothing duplicitous about this.  That’s the way all actions are taken.  I have an infinite number of reasons for doing the things I do, for believing that my next step will not be the step during which my shoe spontaneously combusts and catches my hair on fire, for instance.  I don’t air these reasons.  They’re just part of the fabric of doing things.  To ask me to air such a reason places too heavy a demand on me as an actor.

Given McIntyre’s story, we don’t know what Briffa’s reasons were; and it is likely that if he had bad or duplicitous reasons, he’d keep those bad and/or duplicitous reasons to himself.  RealClimate offers some pretty compelling reasons for using the data that Briffa did, to my mind, so I think we can infer that perhaps those were also Briffa’s reasons.  (Their reasons, so far as I can tell, are that the Yamal Climate Record was already standard by the time Briffa got to it; and that trees in Schweingruber alternative were too small to detect a centennial trend without introducing noise.  But again, I’m not a climate scientist, so I think it’s wise to let the big kids play in that sandbox.  I have no reason to doubt RealClimate’s claim about the smaller trees and about the use of the Yamal record; and I equally don’t have a reason to doubt McIntyre’s numbers, so I’m just going off of where I think the reasons fall.)

In any case, it seems to me that Schmidt is defending Briffa against the implication that he somehow engaged in research misconduct by deliberately cherry-picking data.  That’s a minor misfire on Schmidt’s part, since it’s not the deliberateness in question, as Roger points out (though it’s also not unreasonable to assume that the cherry-picking charge is embedded in the original claim, so it makes sense for Schmidt to debunk it up-front; let’s just not get confused).  McIntyre, on the other hand, is charging that the reasons that Briffa had for selecting the data that he did, whether malicious or not, were unjustified.  He hasn’t yet demonstrated that.  He also hasn’t demonstrated that the Schweingruber alternative is a real and reasonable dendro-climatological possibility.  Just because it’s data doesn’t mean it’s not noisy or unreliable data.

I think the burden of proof is really on McIntyre to demonstrate why the broader data set is the best data set, at least if he wants to show that the currently accepted position is faulty.  Given McIntyre’s arguments, it’s not clear why one would necessarily include the data in the Schweingruber set.

Having said this, given the volatility of this topic, I think there’s a similarly heavy burden of proof on Briffa (or maybe those at RealClimate) to explain why the data used in the Briffa model are more reliable than other data…simply because McIntyre’s concern, by being raised, has become an issue.  But this burden of proof is primarily political: only if they want to further reinforce the robustness of the hockey-stick model, which it seems to me is already sufficiently robust.  (Again, not a climate scientist, just going off these graphs.)  Even if Schmidt is super snarky in his blogpost, which seems to bother Roger, I don’t think there’s any reason to focus on the snarkiness.  As I’ve mentioned before, I ♥ snarkiness, even from people with whom I vehemently disagree.  Snarkiness is just noise, much like, it seems to me, noise introduced by the alleged Schweingruber alternative.