Friday
Jun082012
by 
Bishop Hill
Bishop Hill Gergis paper disappears
Jun 8, 2012 Climate: MWP Paul Matthews has just drawn my attention to the page for the Gergis et al paper at the AMS Journal website, which now displays a notice as follows:
The requested article is not currently available on this site.
Is this significant I wonder?
Update on Jun 9, 2012 by
Bishop Hill
Bishop Hill
Steve M reports that the paper has been "put on hold".
Reader Comments (279)
"Abstract still up at noaa"
Just took a look at this abstract and noticed that the time around 1300 was warmer than today.
And where is "Australasia?" Is that located in the Southern Hemisphere? If so, I thought the MWP was mainly in the North?
I'm coming to the conclusion that we need to start thinking in terms of some Bayesian filtering of climate-science papers.
Given sufficient (and sufficiently accurate) data, one could systematize the weighting of, say:
- the publication history of the authors and their known public statements, past affiliations, pressure-group memberships, arrest history, etc
- the completeness of archived data, metadata, and algorithms
- the uniqueness of the mathematical and statistical methods chosen
- the relative quality of the datasets used and the documentation history of 'adjustments' to those datasets
- the citation history of the authors in particular the ratio of 'circular' references to references outside the authors' 'family'
- etc
in order to come up with a probability that a particular paper requires an extra degree of rigor in the review process.
Richard Drake
"This is commonly referred to as ‘research’."
Yup, and all too commonly accepted as such.
Time will tell, enjoy the wait!
Oh where, oh where has my little paper gone?
Oh where, oh where can it be?
With its proxies cut short and detrended all wrong.
Oh where, oh where can it be?
http://climateaudit.org/2012/06/08/gergis-et-al-put-on-hold/
Dear Stephen,
I am contacting you on behalf of all the authors of the Gergis et al (2012) study ‘Evidence of unusual late 20th century warming from an Australasian temperature reconstruction spanning the last millennium’
An issue has been identified in the processing of the data used in the study, which may affect the results. While the paper states that “both proxy climate and instrumental data were linearly detrended over the 1921–1990 period”, we discovered on Tuesday 5 June that the records used in the final analysis were not detrended for proxy selection, making this statement incorrect. Although this is an unfortunate data processing issue, it is likely to have implications for the results reported in the study. The journal has been contacted and the publication of the study has been put on hold.
This is a normal part of science. The testing of scientific studies through independent analysis of data and methods strengthens the conclusions. In this study, an issue has been identified and the results are being re-checked.
We would be grateful if you would post the notice below on your ClimateAudit web site.
We would like to thank you and the participants at the ClimateAudit blog for your scrutiny of our study, which also identified this data processing issue.
Thanks, David Karoly
Print publication of scientific study put on hold
An issue has been identified in the processing of the data used in the study, “Evidence of unusual late 20th century warming from an Australasian temperature reconstruction spanning the last millennium” by Joelle Gergis, Raphael Neukom, Stephen Phipps, Ailie Gallant and David Karoly, accepted for publication in the Journal of Climate.
We are currently reviewing the data and results.
paul_in_ct - no worries, I am apt to fall into the same trap now and then but I have a "reminder app" on my phone to read a couple of articles by Richard Black or Geoffrey Lean and that brings my cynic meter back into focus straight away. Painful but necessary.
Kramer, the 'Australasia' they refer to is all of Australia, New Zealand and surrounds, including a lot of the Indian and South Pacific Oceans. Paper is still available at
http://web.science.unsw.edu.au/~sjphipps/publications/gergis2012.pdf
So yes, evidence for a general MWP.
Oh Eee Ohh Eee oh wee ice ice ice
Raisin' sea levels twice by twice
We're scientists, what we speak is True.
Unlike Andrew Bolt our work is Peer Reviewed... ooohhh
Who's a climate scientist..
I'm a climate scientist..
................ usw.........ad nauseam..........
Reading between the lines, it seems that Dr Karoly, one of the co-authors, has taken over communications with SM at least. His note to SM is very different in tone from the last effort by the principal author!
Has this become some kind of a gate yet?
"which also identified this data processing issue."
So CA only gets an also?
Shub - Ok, I'll start - "Southern Hemisphere And Gergis" gate.
The significance is the precedent of net-based crowd-sourced review leading to a reaction normally based on formal comment or a paper in reply. This has shortcut that process, to the good of all. Kudos to the authors, or at least those of them who have acknowledged the procedural problem with the method. This is indeed the way science ought to work.
Now it remains to be seen how the AGW blogs and press will treat the episode, not to mention the IPCC.
What about 'Research'gate?
Well, it was accepted for publication in the Journal of Climate ... just wondering ...
Rhoda - nice thought, I suspect we'll all be looking forward to Richard Black and the Guardian headlining with this story.
BTW it's not Klapp is it?
Breathgate.
In memory of all those victims holding their breath until the first warmist news organisation publishes a correction.
oakwood, I also posted a brief and polite comment at realclimate that Gavin has chosen not to print.
JohnBuck
we can't have a SHAG Gate, thats just plain rude
I suggest
Australasian Reconstruction Series Evidence -Gate
"This is a normal part of science"
-David Karoly
It is impossible to remain equanimous with these people.
EternalOptimist - Whoops sorry! We'll go with yours then - hey wait.....
'This is commonly referred to as ‘research’.'
Becomes
'This is commonly referred to as 'audit'.'
Poxyproxygate ?????
So, farewell then Gergis et al,
You were quite the trend
for a while.
Now you have been
detrended.
Significant? YES!
http://wattsupwiththat.com/2012/06/08/american-meteorological-society-disappears-gergis-et-al-paper-on-proxy-temperature-reconstruction-after-post-peer-review-finds-fatal-flaws/
So has this paper's decline been hidden?
Strange how Gavin and his little helpers at RC have been orgasming over this paper since May 22nd - and none of their self-declared statistics genii spotted anything wrong with it (hello Tamino!).
"We are currently reviewing the data and results"
I suspect they are happy with the results, it's just the data that needs "work".
Think about Karoly's supposed timeline. Assume they discovered the problem before Jean S mentioned it at CA. They let a whole lot of people do a whole lot of work without saying a thing. It wouldn't have taken much in the way of courtesy to say -- Hey folks we're working on the problem already. This is what happened. Thanks for your interest. They could even have explained all the stuff they left out of the study and gotten some input from some really bright folks.
And it defies all logic to think they were not aware of the conversation unfolding at CA.
this might be the very nfirst time that a pro-CAGW paper has been held up because of identified deficiencies. Normally, the papers are right even when the methods and calculations are false - eg Mann seriatim.
Congratulations to Jean S.
Comments 36 & 37 at RC are delicious:-
I think, in honour of dbostrom it ought to be .......
............................... juvenilecoupgate
Foxgoose - McIntyre was absolutely fastidious about the civil tongue.
Until he, and everyone else, discovered just how civil the discourse was among 'climate scientists' when they thought no one was looking.
dbostrom can eliminate into a headwind.
Suppose this study had actually been published and cited in AR5 (it was cited in the first draft), and then JS & SM examined it at Climateaudit? Would this paper have been withdrawn at that point, or would the wagons have been circled for defense? It seems that when a study makes it into the IPCC circles, it tends to linger for awhile.
I'm waiting for the editor to resign...
I'm waiting for the editor to resign...
That, LevelGaze is an interesting observation, for sure if there had been a different bias I am sure there would be a clamour for the editor to resign. No come to think about it, he would have already been handed his choice of demise.
The difference in standards is staggering
The paper's lead author has thanked Steve at Climate Audit for "also" finding the error.
http://climateaudit.org/2012/06/08/gergis-et-al-put-on-hold/
"Thanks, David Karoly"
And thank you right back.
I trust if (and I do mean if) the selection method you've reported results in a Mann-like selection of proxies "upside down" from their original interpretation, you will continue to behave in the manly (not Mann-ly) mode you exhibit here.
FOI request to Uni of Melbourne currently inbound.
@JEM 9:52 PM: "I'm coming to the conclusion that we need to start thinking in terms of some Bayesian filtering of climate-science papers."
Bit off-topic: I was thinking for some time about applying Bayes' theorem to FakeGate; I just I gave it a try. I never did Bayesian statistics before, so please be cautious with my attempt. Comments are welcome.
Define the following events:
GF - Gleick (or an accomplice) faked the Heartland strategy memo himself
GNF - Gleick (nor an accomplice) did not fake it; someone else must have done it
GM - Gleick was mentioned in the forged memo
GC - Gleick's typical commas appear in the forged memo
GW - Gleick committed wire fraud to obtain the other documents from Heartland
We want to establish the probability P(GF|GM&GC&GW) that Gleick did the forgery, by binging in the evidence GM, GC and GW in three steps. GM, GC and GW seem mutually independent to me.
Let's say that, with no evidence given, there is one chance in 10 thousand that Gleick faked the memo.
Thus the first prior P(GF) is 0.0001
Now consider the evidence that GM occurred. Consider the conditional events:
GM|GF - Gleick would mention himself in the memo, given that he wrote the memo
GM|GNF - Someone else who forged the memo would mention Gleick therein
Let's estimate the probabilities; e.g., take P(GM|GF) = 0.5 and P(GM|GNF) = 0.01
The ratio between these two conditional probabilities is essential here: 50.
Now P(GM) = P(GM|GF)*P(GF) + P(GM|GNF)*P(GNF) = 0.010049
And according to Bayes' theorem:
P(GF|GM) = P(GM|GF)*P(GF)/P(GM) = 0.00497562
I.e. the posterior is roughly 50 (the earlier ratio) times as big as the prior.
Next we consider the additional evidence that GC occurred. With the same ratio 50 between P(GC|GF) and P(GC|GNF), the posterior would become 0.2 (40 times as big as the previous posterior, not the full 50 since we start with a bigger prior).
And with the third evidence GW (again ratio 50) the posterior becomes pretty close to 1: 0.926.
The ratios could well have been taken higher. With value 100, the three posteriors become: 0.01, 0.5 and 0.99.
I would say Gleick did it.
The hate-filled alarmist bile and vitriol allowed through from commenters on RC is worthy of the show trials of Stalin. If these people had unfettered power, it is almost certain that a real life anti-skeptic pogrom would be instituted. And Gavin, as the ringmaster of that skeptic-censored circus, should be ashamed of himself. It seems to me that while Hansen and Mann cop most of the flak, being transparently what they are, much more attention should be focused on Schmidt's central role in the perpetuation of the CAGW myth.
Re-Gergis-Gate?
The Gergis et al methodology also infringes stock exchange rules on reporting of investment fund results, which would result in prosecution if the managers only reported the results of their holdings that increased in value over the year and excluded all those that did not. Steve M is right, ALL the records must be included not just those showing what was being looked for.
Copied from RC the expanded et al "by Joelle Gergis, Raphael Neukom, Stephen Phipps, Ailie Gallant and David Karoly". And to that we add the peer reviewers, is that typically 3? So 8 people, plus or minus, have already reviewed the paper but the most fundamental analysis of all, the proxy selection, wasn't scrutinized? That must be true, as even if it does turn out fine they wouldn't need to put the paper on hold if they knew the answers now.
EternalOptimist
Perhaps you'll just have to settle for 'Gergis Kan't".
Please note from the maps that the whole of mainland Australia did not produce a single site in this paper. There were a couple of mentions of off-shore reefs and some old work from Tasmania. The land mass of Australia is similar to that of the USA 48, yet not a mention of data except in a vague aggregate sense.
In May 2012 the Bureau of Meteorology released a new high quality temperature set named ACORN-SAT. To the extent that it contains adjustments of about 1 deg C over several decades at some stations, it would be scientifically prudent to hold back on the publication of Gergis et al 2012 until the new data were digested by the authors and changes made as or if needed.
...................................
As an Australian, I resent having this work described in terms such as the most important recent contribution to the climatology of the Southern Hemisphere. (Indeed, the authors venture into the Northern Hemisphere, figure 1.)
Even if the statistics errors are corrected, there will still remain some questions:
1.The statistics are correctly calculated numerically, but are they the appropriate choice?
2.The authors use part of the year, Sept-Feb, to represent SH summer, then use some NH points and others close to the equator, where the summer selection is almost meaningless. In the Tropics, the sun is directly overhead at any Earth point twice a year.
3.What novel and imaginative field work was conducted by the authors, to justify publication?
4.Why did the reviewers fail to find errors?
5. Apart from statistics, there seems to be more errors, or bad calls, such as the use of very low correlations and indeed negative and positive correlations from sites close to each other
6.Earlier papers by non-main authors such as Ed Cook raise difficulties that are not addresed in this paper.
7.As others noted before me, why has there been some indication of selective culling of candidate sites, for example ice cores from Law Dome Antarctica, when Vostok, 1,000 km further away, was included?
..................................
Who gave permission to name me - and indeed most other Australians - as agreeing with the suitability of this paper to represent Australia before the IPCC? We had no say. That's not democracy.
The shameless bunch who performed the circular analysis, and couldn't even get that right, better release the data - the whole of it now.
Imagine this. If they fix their detrended method, their results would match what they''ve previously had, only if they incorporate circularity of inference in some way or the other.
And this passed peer review? What kind of peer review does this journal perform?
The whole editorial board should resign.
Shub, see my post above. If they won't release data and code, their reasons why will be on the record for all to see.
"We will not be entertaining any further correspondence on the matter" will not cut the mustard.
http://climateaudit.org/2012/06/08/gergis-et-al-put-on-hold/
I'm willing to bet a Mars bar that the "also" and "independent" discovery of the cocked up process was a consequence of being alerted by the direction of posting at CA ... the alarm bells chased them to cheque their processes otherwise like all good alarmista they'd have just pushed on with publication.
To me the "also" is an indication that when they did check their processes they also discovered the anomolies that SM, Jean S and others had expressed issues with.
Given Gergis' strutting about like a rutting peacock over her misplaced and rude sense of self importance, this will be a huge boot in the derriere ... I wonder whether she will dare show her face around the 'traps' for the foreseeable future ? As for Karoly, he's only got off his derriere to avoid being carpeted again by Australian sceptics and bloggers.
I love this story. I am sure the authors will pull something out of this and try cobble a new Hockey stick - it will surely be too embrassing to ditch altogether? However this episode is almost designed like a small sketch to give an idea of the whole picture - the typical interaction of the technical sceptics versus the keepers of the faith. Like the self similarity of a Mandelbrot pattern at a smaller scale this is the history of climate science ;)
I mean look at the initial reaction at RealClimate. Like a bunch of teenage Justin Bieber fans, with them all swooning and cooing over how much the study fits their expectations. These guys are supposed to be the proper scientists remember?
Then the Climate Audit crowd come along - Ooh gosh! But they are all horrible cigar chomping brutes who looked at the study askance and made some pretty unflattering observations!
Steve McIntyre gets an arrogant brush off which ironically starts the ball rolling for attention and...
Which group of people end up getting a grudging credit from the orginal author?
This is a sterling example of the problems with climate science; there is no real pressure on it to improve except from the vilified sceptics. Whenever sceptics find these problems and they come out in to the open, you only get back the boiler plate about self correction.
However I don't believe for a second that these problems (now acknowledged by the authors) would have been addressed otherwise. I mean why were the authors re-addressing the paper that had been peer reviewed, had all the media love, and completed the required story? There is no other explanation than the criticism appearing on Climate Audit.
Lets be clear here: what Climate Audit has done here is considered bad, forget the praise through gritted teeth, everything will be done to engineer the back history to create the impression they were minimal in their effect and the "science" corrected itself. The study will be propped up again and the data will better hidden with finer obfuscation. The only self-correction going on here is the methods of PR, spin, and wagon circling.
Let's also be clear here a paper that had all this love and media attention wasn't about to be ruined by any further attention from the authors and faithful, hence the initial snotty rebuff to McIntyre telling he ought to be able to do the "research". The paper is clearly not motivated by normal science - as Ben Pile often says: The politics are prior the science.
I notice Steve McIntyre is telling people to hold back on the triumphilism. Which is fair enough, and typical of his normal measured responses, but I think the rest of us technical layman sceptics can easily justify keeping this episode as a reference point when telling the uninitiated how and why we can dismiss the "science".