NAS report

I posted up at RC a dictionary definition for "plausible", and strangely enough, that didn't make it past the censor ! My concise dictionary gives "apparently reasonable". I am tempted to think that the NAS used the word that they meant to use. Kind of leaves you wondering if "plausibility" is an adequate yardstick for describing scientific proof...

As you say, none of the M&M methodological complaints make it into the summary; but there are an awful lot of them that seem to be adopted in the depth of the report. And it does seem that the NAS panel has an entirely different view of the MBH temperature reconstruction from 1000-1600, compared to the original MBH paper, with its 95% confidence intervals.

Pielke says this is a "near-complete vindication". What do you think, William ?
cheers
per

I guess if you look at it purely as a PR battle, that's great that the MM criticisms weren't in the summary. If you look at it as an assessment, you'll notice that they did say that the use of bristlecones was a concern, that the pre-1600 results were so uncertain that they could not quantify them, and that the PCA method of Mann was in error.* To me, it's more important what their opinions are then what made it into the summary. But I'm sorta like that.

* albiet minimal in impact. BTW, I have faulted Steve for not being clear and mathematical and just plain organized in describing how much of an impact acentric PCA has. I also fault Mike very much (more) for not admitting that acentric PCA has been proven to bias results and is basically a mistake. I don't mind if Mike retreats to the "minimal impact" defense. Heck, he can quantify it also. But the first step, needs to be a very technical one of admitting that it is just incorrect equation-wise to use acentric PCA.

I'll for once agree with per. Roger has a Humpty Dumpty style with words, they mean what he wants them to mean. Redefinition is a polemical trick which is especially effective if you don't tell people you are doing it. Lest you think I am simply bitter, look at this little exchange both at my place http://tinyurl.com/hmuxl and his http://tinyurl.com/jxtzf. And yes, you do have to read chapters 9 and 10.

This looks like a clearly written report. To be honest many of the graphs such as http://darwin.nap.edu/books/0309102251/html/56.html and http://darwin.nap.edu/books/0309102251/html/58.html don't look as shocking as the usual exponential-type curves of temperature rise that are presented for the last century.

Taken over the last thousand years, the current trends are just another statistical fluctuation.

[Um. Did you actually read the report? The current trends are exceptional in the reported record. If you believe the report, the record is less certain pre-1600 than previously thought - but that hardly supports your statement -W]

In polar conditions there is no isotopic evidence of global warming at all over long time scales in graphs like http://darwin.nap.edu/books/0309102251/html/66.html

I'm curious about the cause for the Little Ice Age of the 1850s when the Thames in London regularly froze over regularly for a decade or so.

The report just compiles piecemeal evidence. It does not produce a analysis by evaluating all the evidence with a complete mathematical model for all the variables.

To understand temperature variation data over the past 1000 years you need to know or plausibly model how the gulf stream flowed over that time, how rainforests responded to temperature and atmospheric gas variations, etc. In particular, you need a good model for what happens to carbon dioxide converted into CaCO3 by sealife that precipitates to the seabed. Some of that gets locked up for a long time until released in volcanic processes.

[Ah, the familiar "we don't know everything so we know nothing so lets do nothing" argument... -W]

what is your ref for the Thames freezing regular in the 1850's for a decade or so? A simple "google" shows that 1814 was the last complete freeze:

http://www.bbc.co.uk/dna/h2g2/A970733
http://en.wikipedia.org/wiki/River_Thames

of course things such as bridge building & more importantly the creation of the Victoria & Albert Embankment (which narrowed & increased the velocity of the flow) have more to do with the lack of freezings post "Little Ice Age!"

Anyway, my take on the report is that the M&M & their cheerleaders (who are mostly right-wing hacks and not even in the climate science field) basically dragged the reputations & careers of MBH through the mud for very little (although they're crowing it's a "victory"; sort of like Chemical Ali in Baghdad). And all done in their zeal for a "smoking gun", and to stroke their egos (I guess it gets lonely at the "University of Guelph" or in doing "mineral engineering"). It seems obvious from reading the CA blog that at some point long ago it just became far more about egos and reputation (M&M trying to gain a reputation; MBH trying to preserve theirs); and not about science or linear algebra.

Hans van Storch, a climate scientist whom I greatly respect & trust, and is not on either side of the "climateaudit vs realclimate binary debate," has a nice commentary of the NAS report up. Unfortunately the site seems to be down now, but a google search should find it.

Basically it goes: AGW is a real concern (which will anger the M&M & Cato Institute types), there are some errors in MBH98 that don't invalidate the whole hockey stick or AGW, MBH shouldn't have been so secretive & proprietary in their methods etc. Overall I'd say that the pendulum swings back a tiny bit from what I guess "realclimate" would have ideally wanted, but in the grand scheme of things (AGW) this is not that huge a deal. So a lot of nastiness done ostensibly in the name of science.

I'm more interested out of this how scientists can make sure the peer review process works optimally -- is it as simple as just hand over the data to everyone who comes along (even if they're jerks? ;-) Should peer reviewers tediously go through the complete dataset and do entire reconstructions instead of trusting the analysis of the authors? The latter is probably not so bad if we're talking about a small ensemble of climate models, but I recognize that the MBH98 was a large set of some dozen or so separate multiproxy studies etc.