Monday
Apr262010
by 
Bishop Hill
Bishop Hill A good trick to create a decline
Apr 26, 2010 Climate: Mann Climate: Statistics Shortly after the Climategate emails broke, an guest article was posted at Climate Audit. The article is important, but was rather overlooked by the sceptic community in all the excitement over the emails. “The Hockey Stick and the Milankovitch Cycle” uses some of the Climategate files to solve one of the remaining mysteries of the Hockey Stick.
This is my attempt to put the post into layman's language. It's rather long so you may want a cup of coffee to keep you company.
Read it at the link below.
Reader Comments (36)
Nice cup of coffee...
And you have been busy!
Are you planning a sequel to 'The Hockey stick Illusion' yet?
Thanks for this, and also for your excellent book, which I have just finished. By the way, have you sent copies to any of the climate experts in the cabinet?
I know of several MPs who have discussed and/or read it, but I would have thought the cabinet were a lost cause.
I see another book on the horizon.
Well done.
Just before the "Finlandia" heading, you have this paragraph:
"The problem was that when McIntyre used co2detrend.f to process the tree ring and
carbon dioxide data, it didn’t turned out that it wasn’t the adjustment used in the Hockey
Stick paper. Quite what it did and where it was used remained a mystery."
I can understand it with a few passes, but you might want to revise the grammar to make it less ambiguous.
Conclusion: The hockey stick papers were not and are still NOT replicable - therefore they are NOT science.
That is damning.
Good job Andrew, nice summary.
I can imagine Mann in his office trying all available techniques to get the graph he always dreamed of to present "a nice tidy story of unprecendented warming" (Briffa)....
Andrew thanks for reminding me of Jean's post on Climateaudit earlier this year and your clear, layman's language, rendition of it.
BTW I read your "The Hockey Stick Illusion". I thoroughly recommend it to anyone wanting a lucid & highly readable account of this saga.
Your thread title `A good trick to hide the decline` might also be applied to the rubbish policies we are offered by the three main political parties and the Greens. The trick here is the drive towards CO2 reduction, the decline being to our economy and standard of living.
I have just finished watching Andrew Neill question Ed Miliband, Greg Clark, Simon Hughes and the man from the Green party on climate and environment issues. It was a dispiriting experience. All four are equally deluded supporters of the Climate Change Act. The politicians all hide behind `the science` to justify their positions. From this you will correctly conclude that not only am I a `flat earther` (per Brown) but also a `nutter` (per Clegg).
All four calmly accepted that reducing CO2 will cost us all £18 billion pa. In addition Simon Hughes says he wants us all to fly less and pay an extra £3.5 billion pa for the privilege (apparently video conferencing will achieve this for us all!) and also wants to spend £100 billion on wind farms in lieu of nuclear power stations.
Neill observed that to build 4000 windmills the size of Blackpool tower - the Labour party`s plan - would require one to be built every day! Ed Miliband was unable to confirm if one would be completed today.
The only slightly redeeming feature was the information that many of the prospective new Conservative MPs are sceptics. This was an object of much criticism from the other parties to which Clark had no answer.
The course we in the UK are on is utter lunacy. I shall continue to try to persuade my future MP to cause a rethink of the Climate Act. I urge others to do the same.
This is a bit off topic, but stick with me.
I don't know if you've heard about the Orlando Figes scandal, or not? But there is something very interesting that came out of it the other day which has a bearing on the climate gate scandals.
For those not in the know, Orlando is a very famous history professor whose specialist area is Russia. He has been found guilty -- after many twists, turns, and outright denials of manipulating Amazon and other online reviews to denigrate the work of his rival historians, and promoting his own.
This has been quite widely reported in the media. His career is effectively over.
Yesterday evening, I heard an interview with Robert Service, one of the specialist Russian historians whose work Orlando had been denigrating in online reviews. Robert Service, whilst happy that Orlando's activities have been exposed, is determined that an extensive review will now take place of all academic work that Orlando Figes has come into contact with. Robert Service's attitude is that any such work is now potentially tainted by association with Mister Figes, and must be treated as such.
This interview was conducted on the BBC. At no point did the interviewer express surprise that any such enquiry into academic malpractice should take place. They did not, at any time, attempt to defend Mister Figes.
Compare and contrast with the behaviour of the mainstream media, and the BBC in particular, to the issue of academic malpractice involving Prof Jones, Trembath, Mann et al.
Andrew, maybe a dumb question, but how was the graph of CO2 levels going back to 1400 determined? Some type of sediment carbon dating process? Need to know this to judge whether Mann's subsequent scaling and shifting were justified.
Reading about Orlando Figes chimes with the (very few) bad reviews on Amazon of Bish's book. Most reviews on Amazon seem genuine enough, but it would surely be a simple matter for them to restrict reviews to those with a purchase record?
I know that wouldn't guarantee that they'd read it, and it would exclude one or two who had bought it elsewhere, but it would cut down on the trolls and render the published opinions that much more reliable.
UKIPer
Steve's history is here.
It looks as though he took the CO2 data from Mann's UVA website, but he says he's not clear on the provenance.
I read your book, I have just read this article forgive my profanity, what the fuck is Mann measuring?
Stuck-record
The irony of your comparison between the criticism of Figes and the serial whitewashes of Jones and Mann is that, as it happens, Figes' infantile behaviour has no bearing on the validity of his historical work, which should stand or fall on its own merits, while by contrast the behaviour exposed by Yer Grace's book and the Climategate files goes to the heart of the science practised by the team.
Sorry to double-post, back on topic this time. Can anyone explain why Mann's error bars don't get any wider as his series goes back in time, even though the number of trees involved reduces and the age of the rings increases? I have an explanation in mind, but it is not very flattering to him. Surely the error bars should look like a section of a cone on a dog with fleas.
While this is interesting in itself and historically, the theory of a CO2 fertilizer effect on bristle-cone pines is pretty much passe. The fact is that it's the strip-bark pines which show the rapid increase in "ring" width and the cause of this isn't either fertilization or temperature, but a artifact of the strip-bark process itself. I've been convinced of this perhaps before Steve Mc was, but if you look at his posts on the Starbucks Coffee experiment, you'll see that after the injury which caused the strip-barking of a tree, it takes a while to heal, but then there's a long period of much faster than normal growth which can be caused by a smaller crown being able to use the remaining roots which can access all the water and/or soil resources near the tree but only creating a linear rather than circular wood growth. Eventually the strip bark overrides the old stem and resumes a circular pattern, and the apparent growth slows down. Since in recent decades we only see the initial linear growth spurt, the statistical effect is to create a hockeysticklike pattern which fits well with the instrumental record and thus is very highly rated by Mannian methods. But it has no relationship with Mann's desire to find a ringwidth / temperature correlation.
There is mention of the Milenkovich Cycle but virtually no meaningful and quantitative discussion on how the Earth's temperature varies due to this phenomenon.
Have I missed something somewhere?
There is mention of the Milenkovich Cycle but virtually no meaningful and quantitative discussion on how the Earth's temperature varies due to this phenomenon.
The Milenkovich Cycle is a complex cyclic variation in the tilt of the Earth, distance from the Sun, and a bunch of other things that effect the amount of Sunlight impinging on the surface of the Earth. It is, in short, both cyclic and not man-made and has nothing to do with CO2.
Read about it here. For some reason, it does not google directly.
http://en.wikipedia.org/wiki/Milenkovich_cycle
Can anyone explain why Mann's error bars don't get any wider as his series goes back in time, even though the number of trees involved reduces and the age of the rings increases? I have an explanation in mind, but it is not very flattering to him.
You already understand.
Oh, OT, well sort of. Mann is huffing and puffing about suing M4GW. I hope he does.
Read all about American libel law here -- and why Mann will fold on his threat.
http://www.foxnews.com/scitech/2010/04/26/climate-scientist-heated-satire-threatens-lawsuit/
Thanks for this, Your Grace. I'm not a scientist (and I don't even play one on the Internet), but I'm a little confused by the following on P. 5.
"McIntyre’s first attempt at the regression is shown in Figure 4:"
but the caption on Figure 4 reads:
"Figure 4 McIntyre’s second attempt"
[BH adds: It was his first attempt at a rescaling, but his second attempt. It could be clearer, couldn't it?]
Thanking you for my evening read your grace
Though climate related and "Connelly" appearing in the [edit] "history" the item appears fairly neutral - google "Milankovitch"
@David S.
Coz it's data points, not error bars...
3X2 --
Gracias. La edad es una cosa terrible!
No Benjamin you are wrong there: I quote from the note to the diagram in MBH 1998: "uncertainty limits are shown by the light dotted lines".
For parameter after parameter in AGW climate ‘science’, the most desirable raw data seems to consist of suitably dubious shotgun clouds of red noise, providing the most suitable blank canvas for true artistic creativity.
The reality is that the "Hockey Stick" has become the Piltdown Man of climate science.
It took over 50 years before the scientific community declared Piltdown Man a fraud. The 'trick' being to combine the skull of a human, with the teeth of a chimp and a jawbone of an orangutan at a crucial time when the theory of human evolution was of significant scientific interest.
Even though many people declared their scepticism when it was first discovered it did, however, fool the British Museum, the Geological Society, the College of Surgeons and many other august bodies. A scientific consensus was established that Piltdown Man did represent the 'missing link' between humans and apes. Over 250 scientific papers referred to the importance of its discovery long before it was discredited as a fraud.
The fraud actually hurt science in general and greatly impeded the progress of the study of human evolution.
The same fate awaits the Hockey Stick and again it will have profound consequences on the public's perception of science.
Cup of tea in hand, I enjoyed the article very much and understood it for a novice. Pat on the back and now its written pour yourself a beer before carrying on pushing the AGW stone up the hill. Hopefully with more and more people coming to help push it you and Steve and company will get ot the top and have the joy of pushing it down the other side watching it run over all the Cliamte scientists (aarrgghh!!!, sorry always scream when i mention this bunch of deluded men and women).
Benjamin & David S
MBH 1998 in Fig 5b used plus & minus 2xStandard deviation (SD), which does not vary according to sample size. (Actually they use the symbol small sigma to represent SD, which is actually the population parameter estimated by SD from sample data.)
On the other hand 2xSE or 95% confidence interval, which are also often used in plots, are inversely related to the square root of sample size.
...but in MBH99 (the paper referred to by the BH piece), Fig 3a uses SE, perhaps explaining why the error bars are much larger pre-1600 (as they are based on less data).
Q: MBH99 does not use SE. In fact, the error bars in MBH99 are still somewhat unsolved mystery although several people have tried to figure them out.
David S & others: MBH98 (and MBH99) error bars do get (a bit) wider back in time. "Confidence intervals" in MBH9X are derived from (stepwise) calibration residuals, and as the calibration fit is worse in early steps, CIs are also wider.
Politics:- This perhaps before Steve Mc was, but if you look at his posts on the Starbucks Coffee experiment, you'll see that after the injury which caused the strip-barking of a tree, it takes a while to heal, but then there's a long period of much faster than normal growth which can be caused by a smaller crown being able to use the remaining roots which can access all the water and/or soil resources near the tree but only creating a linear rather than circular wood growth. Eventually the strip bark overrides. He has been found guilty after many twists, turns, and outright denials of manipulating Amazon and other online reviews to denigrate the work of his rival historians. It was a dispiriting experience student aid. All four are equally deluded supporters of the Climate Change Act. The politicians all hide behind `the science` to justify their positions. From this you will correctly conclude thanks medical school.
It's late, so it may be my eyes have given out on me, but Figures 11 and 12 look identical to me. What did I miss?
jorgekafkazar: They are very similar but not identical. See for example the part before AD1080: in the true hockey stick (Fig 12) no year reachers zero anomaly but in Fig 11 there are many of those. However, Fig 11 has very poor RE "scores" (which Mann used as a goodness measure), which seems to be the only thing making it different from the true hockey stick. In other words, like BH is saying, it very much looks like the "handle of the hockey stick" was unscientifically "crafted" with "trial and error" approach.