Skiphil writes:
Andy Revkin of Dot Earth/NY Times blog is inviting questions to be submitted to the authors of Marcott et al. (2013). Since Revkin is one of the only journalists who might have a chance of getting the study authors to be responsive, this is a good opportunity.
Specifically, he’s asked for someone to prepare one list of questions which are “perceived as unanswered.”
Folks could start a list here at WUWT to post at Dot Earth, or simply post questions/points at Dot Earth until we have a good list.
submit questions on Marcott study to Dot Earth/NY Times blog
Andy Revkin Dot Earth blogger
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
LEAVE IT TO STEVE. I’ve read some of the comments over at Revkin. Inept. All on the wrong tack.
Dear Dr Marcott et al
You went to a lot of trouble to produce a set of temperature readings for a long time period.
Then you “Mean-shifted the global temperature reconstructions to have the same average as the Mann et al. (2008) CRU-EIV temperature reconstruction over the interval 510-1450 years Before Present.” (Your statement on RealClimate blog)
What was the size of the “Mean-shift” adjustment?
Why do you have more confidence in Dr Mann’s work than in your own?
1) This question concerns the uncertainty about the scale normalization of Marcott et al. proxy temperature anomalies (calculated initially relative to 5500 – 4500 YBP) to those of Hadcrut4 (calculated relative to 1961-1990).
Shakun writes
Is the net effect of this process a simple linear increment of the reconstruction by ~0.3C? If yes – what is the error on this offset value ? If no – how was the renormalization achieved ?
2) To justify a statement that 20th century warming is unprecedented since the last Ice Age it must be demonstrated that any previous climate excursion of a similar magnitude would have been observed in Marcott’s data. Hence the importance of Tamino’s work. I performed an independent study which concludes that it would not be possible to detect any such ~1C excursion lasting less than ~400years because of proxy timing uncertainty. This is supported by this statement on realclimate:
Do the authors agree that this age uncertainty implies that the reconstruction are indeed insensitive to excursions lasting only 200 years ?
Leave it to Steve. His questions will go to the heart of the matter. Don’t set up some easy straw men for these guys to knock over instead.
What is the new scientific practice used in Climate science that allows data to be changed to fit the conclusion?
I’ve tried Wiki but that’s no help as usual.
Does he consider re-dating proxies in order to force a hockey stick shape to be scientific fraud?
I would be interested to know if the MS was rejected by Nature, and whether the reconstruction resembled the thesis version or the Science version if so.
I have seen a response somewhere that the resolution of the proxies is 120 years, but the answer was crafted rather carefully.
Does the reconstructed record have 120 year resolution along its whole length, or does the resolution vary? If the latter, how poor does the resolution get?
JF
I agree with statements saying “Leave it to Steve”. This is very much a matter of statistical methods of analysis, about which he is an expert, and a matter in which he is already engaged. Go get ’em Steve!
What I find so disappointing in the comments for the original Revkin article is the number of posters that say that the original Hockey Stick is robust and has been replicated.
Mann’s own work prior to the year 2000 proves it is absolute crap.
The lack of a MWP and the changes to his own 20th century temperature Records shows a complete lack of Integrity.
Add to that the claim of “Unprecedented 20th century warming” shows it to be absolute nonsense.
What a really sad day for SCIENCE.
There is only one question: “How is it possible to take the same set of data and come up with two different graphs?”
To prevent a flood of questions both good and bad, like a couple of other commenters here I figure the most productive option is simply to let Steve McIntyre pose questions on behalf of skeptics. As far as I can see no-one else is quite so deeply into this paper and its problems than Steve. And his clinical approach means maximal trust from almost all in the skeptical community.
Jean S has it laid out clearly @ur momisugly ClimateAudit.
=============
There’s only a handful of people whom questions should be taken from. Steve number 1 of course, and Clive Best has done some good work on this. Even with the others still best to go through Steve so duplicates aren’t submitted.
Yes. Mr. Marcott, are you truly the Son of Mann?
I agree with Moshpit and others: Let people who can see through the statistical illusions ask the hard questions (Steve Mcintyre, AMAC, Brandon, and Just The Facts). But be sure to include the following:
Please provide a list of all reviewers.
Please provide a list of everyone involved with the statistical analysis of the data.
Please provide a list of the editors at Science who approved this paper for publication.
Please provide the code.
Exactly when and where will all of the above information be made available?
I don’t expect any real answers, I predict we’ll get very carefully parsed, obfuscating, weasel-worded statements prepared not by the authors, but by Peter Clark, Gavin Schmidt, Mike Mann, and other members of the Team.
Judy Curry says::
“Please don’t… inadvertently send them to the RealClimate refugee and training camp.”
Too late. Marcott & Shakun have already entered the ‘House of the Rising Sun’.
After reading all of the analyisis of this at Climate Audit, Steve has all the questions to be asked.
Go Steve !
Steve Mc’s comment makes it clear that he is ready to communicate with Revkin on the high priority questions he has, so it’s better if most of us stay away from Dot Earth on the subject and let Steve assemble the questions he thinks are most significant. If there is a barrage of assorted questions then the authors will cherry pick the ones they want to answer and pretend that they’ve been responsive.
Perhaps it is a useful exercise to brainstorm here and refine some questions in case Steve et al. find any useful suggestions. Just to prime the pump, here are some questions which Jean S already posted at Climate Audit:
http://climateaudit.org/2013/04/04/marcott-monte-carlo/#comment-410510
oh, I haven’t even had time to ask all the questions and I’m sure there are more to come later. Some questions (apart from those obvious relating to the re-dating and the “upstick”), not in any particular order:
1. Where’s the code? Where’s the code for Shakun et al., Nature (2012) as promised in the corresponding SI?
2. In your temperature perturbation MC analysis for alkenones and Mg/Ca proxies, why did you account only for the model uncertainties?
3. What is the rationale for using two independent normal (symmetric) variates for perturbation in the exponential model (Mg/Ca proxies)?
4. What exactly is the uncertainty described for the ice core proxies (SI, p. 6 (f))?
5. Why did you join the instrumental temperature record (CRU) with the EIV reconstruction into a single curve without any indication of this in the case of Mann et al (2008) EIV-CRU?
6. Why did you use EIV-CRU (land only) as the reference? Wouldn’t it be more natural to use, e.g., EIV-HAD (land+ocean) from the same reference?
7. Why did you use the early portion of EIV-CRU (510-1450 yr BP) for the reference? Wouldn’t it be more natural to use the later part of the overlap (or the full overlap), where the Mann et al (2008) reconstruction is more reliable? How sensitive are your conclusions regarding comparision with the modern temperatures to the choice of the reference time interval and the reference series?
8. What are the uncertainties for the (EIV-)CRU record post 1850?
9. Why do you use 1\sigma uncertainties for your reconstruction, but 2\sigma uncertainties for Mann et al (2008)?
Since it was my comment which Andy Revkin responded to, and after reading reactions at a number of locations, let me put forward a suggestion which may be acceptable to all parties. It is clear that any questions should be framed with only one intent, to extract answers to facts which should have already have been largely answered in supporting data supplied as part of the published paper. It is therefore clear that it would be far preferable if Marcott et al make available ALL code and any other supporting data immediately. It would improve their credibility to do so.
In the event they are unwilling to do this, I suggest that a maximum of four questions to be asked by EACH of Mr S Mcintyre, Jean S, Mr S Mosher, Nick Stokes, and two others considered non skeptics who may have questions.
Supplimentary questions should be allowed where the original question has not fully answered.
A video of the questions and answers should be posted by Andy Revkin on his Dot Earth site.
Leave it to Steve. Andy Revkin has a soap box, but not more. Steve has the facts, and a world wide following.
I have taken a different tack. Wrote a concise short request to Science editors that the article be substantially revised in light of abstract sentence 1 and supporting figures 1a and 1b being directly contradicted by S18c and now the FAQ, or retracted entirely since the core top redatings were unnoted and unjustified although proven in some cases to be in direct contravention of published core top dates and/or age controls. The editors have acknowledged the communication, and requested time for due process under Dr. Smith’s supervision.
If Science does not act in some appropriate manner in a reasonable period of time, I will provide Anthony with the archived communications for an expose.
I am more interested in the Research Management.
Chapter 4 of the Marcott PhD Thesis was the basis for the Marcott et al paper. The PhD thesis did not show any uptick in the 20th century.
* Why was it decided to revisit that thesis and adjust the data times (core tops) changing the output?
* What was the reasoning behind the redating of each of the proxy data series?
* Who was the proponent for the alteration in the proxy data series and timing?
My local paper is using Marcott for alarm today. http://pressrepublican.com/0205_columns/x1916524832/Climate-change-A-clear-and-present-danger
@jim McCulley – don’t be a wallflower. You have the tools at your disposal here at WUWT to write a letter to the editor. If more skeptics spoke up about these issues publicly, the media might think twice next time – Anthony
Just for the record I happen to think Andy Revkin does a much better job of reporting on highly technical climate and other issues than most reporters.
Never fear I am all over it. I have been writing on this subject for years. And had Bill Mckibben laughed out of the Adirondacks with his step up for carbon campaign. 🙂
Adam says:
April 7, 2013 at 12:05 am
“Thus, the 20th century portion of our paleotemperature stack is not statistically robust, cannot be considered representative of global temperature changes, and therefore is not the basis of any of our conclusion…”
Why did you elect to not exclude that period which was “not statistically robust”………………….
——————————————
The published Abstract appears to be 100% truthful although it leaves a false impression by omission.
Without their hockey stick a cooling problem materializes.
1. Was the Marcott hockey stick blade included on the released graph?
2. Why was the not statistically robust statement not in the Abstract?