Bishop Hill The ups and downs of the Marcott
Mar 11, 2013 Climate: MWP This is just a brief post to point to a few analyses of the Marcott hockey stick.
The National Journal has a science-lite round up of the paper including this quote:
To be clear, the study finds that temperatures in about a fifth of this historical period were higher than they are today. But the key, said lead author Shaun Marcott of Oregon State University, is that temperatures are shooting through the roof faster than we've ever seen.
"What we found is that temperatures increased in the last hundred years as much as they had cooled in the last six or seven thousand," he said. "In other words, the rate of change is much greater than anything we've seen in the whole Holocene," referring to the current geologic time period, which began around 11,500 years ago.
Meanwhile Rud Istvan, writing at Judy Curry's, notes there's just a small problem with the National Journal's case:
The proxy selection was deliberately weighted toward ‘low frequency’ resolution, since the entire Holocene was being assessed...there is no statistically valid resolution to the combined proxy set for anything less than 300-year periods...
Marcott neglected to tell NPR his methodology did not recognize ‘fast’ century changes at all–until recent thermometer records were spliced onto the 73 paleosites.
David Middleton, at WUWT has more.
Reader Comments (47)
I haven't check this myself but ....
"I discovered that only nine of the 73 proxies contained data that extended to 1950. Of those nine, only two contained data that extended to 2000."
"This new 73 proxy study has alarmists convinced that this is an independent verification and vindication of Mann’s hockey stick. It isn’t. The hockey stick blade at the end of the reconstruction is resulting from an adjustment of the proxy data to agree with Mann’s treemometer study. That, or it is an outright splice of Mann’s data directly."
http://suyts.wordpress.com/2013/03/10/the-hockey-stick-resurrected-by-marcott-et-al-2012/
Why am I not surprised that they're trying to tilt the pinball machine again?
I've gone to Climate Audit and gotten a bad, off-center homepage screen with no columns for recent posts or comments on the right. Another poster, Stephen Brown, at WUWT said he was getting the same thing.
Anyone else getting a screen where the text is off center to the left with at least a quarter of it off screen at Climate Audit? It's the only website I've gone to that is having that problem, so I don't think it's my Mac computer running Firefox.
@ theduke
Caused by global warming.
theduke
Same here, PC - Firefox, looks like the theme/format is screwed up. Wouldn't think it will take long to get sorted.
CA works fine for me on safari and Firefox??
Back on topic BUT could the reason be for the similarity between Manns hockey stick and this one us because they have used Manns data?
If they have used Manns data you can bet your left testicle that data will NEVER be released.
Mailman
@ Mailman
You haven't quite got the hang (ooh, er, missus) of the gambling thing, have you?
The blog discussions of this paper are a bit confused, but thats because the authors failed to explain how they averaged the proxies to get the HS. If it is a splice to instrumental data, thats amazing because the paper doesnt say so. If it is not, thats equally amazing as there is no HS in the proxy data.
Suyts is certainly right that hardly any proxy data extends up to now. So it may come from just one ir two data points in one or two proxies.
Guess we have to wait for you know who to sort it out.
Which all reminds me of the day the younger Klapp tripped over his tongue and coined the word statestical.
Newsnight: Chris Huhne's former special advisor Duncan Brack say's his fourth career could be in the environment sector.....pro bono is out of the question then.
CA is fine with Google as the browser.
Tony.
That "signal" does in fact appear, at least to this electrical engineer, to have two regions with vastly different characteristic times (or dampings). Any SINGLE naturally occurring low-pass-like smoothing (diffusion, etc.) of the early portion would not have let that recent spike through. Looked at another way, the early portions may well have had many spiky regions with large derivative, all low-passed to much smaller values. So I am agreeing with the original post here.
If they spliced the thermometer data onto the proxy data (as Judith Curry says), shouldn't they have put the temperature record at its mid-point at the splice point. That way the 150 year temperature record would be one point if averaged and would be properly centered. Instead it seems that used the lowest point of the temperature record to splice onto the proxy records, thus creating an exaggerated hockey stick.
Marcott is either being tricky or stupid.
Okay, CA now looks back up and running to me also. But for most of the morning into the early afternoon, that was not the case. (it's almost 5 pm here in California.)
Sorry for the interruption.
Can't wait to see McIntyre's demolition of this paper.
I find it difficult to understand why the paleo crowd don't run their papers past real statisticians before publication. Perhaps because it would leave them without a paper?
It's already been accepted by Climate Psyientists that traditional paleo-proxies such as tree-rings haven't performed to specification for the last half-century or so.
The reasons for this are steeped in highly complex post-normal calculations involving quasi-mechanical teleconnections; a branch of funding-mathematics that tranforms a political viewpoint into simple graphical outputs that can be displayed in Powerpoint by many and, it has been rumoured, even in Excel by a few highly-skilled eggheads!
The problem of how to fix the Anthropocenic proxy-problem occupied the waking and sleeping hours of some of the worlds brightest Nobel laureats-in-waiting for days on end until Dr Mann had his Eureka moment!
With hindsight his solution was both simple and brilliant, so simple that it didn't need explaining and so brilliant that it illuminates the way for current groundbreakers such as Marcott 2012 et al.
I refer, of course, to Mikes 'TRICC" (commonly misspelt as Trick) - an acronym for Transfer of Revenue Into Climate Coffers.
His thinking - I hypothesise - went a bit like this -
(A) The Paleo-proxies represent temperature.
(B) Recent proxies don't give us the correct temperatures.
(C) But we've got the temperatures we need without using proxies.
thus (and this bit is pure Genius)
(D) We'll use temperature records as a proxy for.. drumroll ... TEMPERATURE!
Dr Marcott et Al (izzat thanks for the Internet-thingy Al?) have indeed stood on the shoulders of a giant.
Brandon Schellenberger had an interesting comment at ClimateAudit and Steve responded:
Deja vu. It goes like this .....
Press release reveals new study finds recent warming is "unprecedented". It's all our fault. We're all gonna die.
The press relase is lapped up and quickly disseminated by the mainstream media and CAGW bloggers, and the study enters CAGW folklore.
The unwitting public takes note of the mainstream media reports. It's all our fault. We're all gonna die.
The study is paywalled and/or or data and models are not made public. However determined scientists/bloggers thoroughly debunk the new study showing it to be based on dodgy data manipulation and/or distinctly dubious scientific methods.
But it's all too late. The mainstream media reports none of the findings of the scientists/bloggers, so the majority of the public remains blissfully unaware. Various politicians and self interest groups quote the study ad nauseum to further their own agendas.
More insane policies are introduced to combat CAGW.
Outside, it's snowing again.
Mar 12, 2013 at 12:25 AM | Alex Heyworth
"I find it difficult to understand why the paleo crowd don't run their papers past real statisticians before publication. Perhaps because it would leave them without a paper?"
It would leave them without self-respect. They sense this.
SteveB: excellent summation but you forgot one thing. After "Outside, it's snowing again" you should have written, "which is attributed by climate scientists to global warming."
[Snip - venting]
"What we found is that temperatures increased in the last hundred years as much as they had cooled in the last six or seven thousand," ...
This claim seems historically illiterate to me. The increase in the Central England Temperature record shows a rise of about one degree centigrade in the 20th century and about half of that has already vanished in the past decade. The rate of increase shown in that data set does not seem any greater than the rate of cooling during the Little Ice Age.
http://www.metoffice.gov.uk/hadobs/hadcet/
I wonder what the rate of cooling was when the Norse settlements in Greenland declined and vanished?
We need an estimate of the MTBD for junk science papers in the climate area (arena?): Mean Time Before Demolition.
Thanks to the internet and what may well be a growing number of angry analysts, I'd guess the MTBD is of the order of 5 days, and dropping.
It took years for the farcical methods of the MBH hockey-stick to be exposed and thereby demolished. But nowadays, how long would it take? Three cheers for the inspirational lead of McIntyre and McKitrick and for the Bish's HSI which captures and communicates their achievement so well. M&M&M -Three Musketeers de nos jours!
"Why am I not surprised that they're trying to tilt the pinball machine again?"
Because they're smacking the penny-drop? [Actually, the last time I saw one, there weren't any machines that took less than 10p coins.]
A detailed piece by David Middleton. But I believe a debunking of this claim by the authors:
is more easily debunked, by simply citing BEST temperature estimates dating back to 1750. These include decadal warming periods in both the 17th and 18th centuries that extend to +/- 2.5-3.0 degrees farenheit (1.5 degrees celcius):
http://www.statschat.org.nz/wp-content/uploads/2013/03/Rplot001.png
Steve McIntyre has made two interesting comments at Judy Curry's place here and here.
A point on the implication of periods in the recent past (say last couple of thousand years) in which global temperatures were warmer than now...
First, to avoid misunderstanding, I should say that I think Mann et al's work on historical temperature records to be a travesty of science and mathematics, an offence against reason itself.
Having got that off my chest, suppose that you thought both of the following were true:
a) CO2 is responsible for the post-industrial warming;
b) The medieval warm period was as warm or warmer than current temperatures.
You'd therefore think that naturally occurring increases in atmospheric temps were at least as great as the recent man-made increase. You'd therefore also be concerned that a natural MWP-type warming might add to the man-made increase and potentially double or more the net effect.
Now I understand why many people want to "get rid of" the MWP so that recent warming looks very dramatic. But, if they looked at things more rationally, since their physical arguments for CO2 causing the recent increase don't depend in any way on whether or not there was an MWP, if they "rediscovered" the MWP, they could be even more alarming about the combined natural + man-made effect.
It would have the added benefit of consigning Mann and his acolytes and imitators to richly-deserved obscurity, which would be good for everyone involved, no matter what they think about the effects of CO2.
Simon Anthony,
I have wondered about the possibility of anthropogenic warming added to a nautural event (cycle of unrecognized periodicity perhaps?) yet yours is the first suggestion of this that I've recognized in the 5 years I've been looking at this stuff - thanks.
Jonathan, thanks for pointing to those posts of Steve's, which I would certainly have missed without you. He finishes:
I've heard of paralysis by analysis but I think we need a more precise phrase for the difficulty Steve is in. Paralysis by extent and depth of crap being analysed? I'm not saying it's pretty but neither is the difficulty, nor the so-called science that seems to delight in producing it.
Other analysts have already done a number, but if Steve McIntyre was on the prowl around a stats-laden paper that I had put together, I would be a tad nervous. Steve's comments suggest that there is an evisceration in the offing.
The other side of that, Simon Anthony, is that if considered so, we may be at or near the peak of the latest millenial scale change, and anthropogenic warming will only counteract whatever natural cooling would be happening. Evidence of gradual cooling throughout the Holocene adds ice to this swirling, now slightly more carbonated, drink.
Then, it's a little easier to understand that whatever Anthro warming we get, it will be adequate only to buffer our exit out of the Holocene.
==============
Kim,
that take i have seen (agw + cooling) - and it does make sense. But it's the other which seems infrequently contemplated, but then maybe the human provoked catastrophe seems sufficient for the alarmists without piling a 'natural' high onto it. Besides, "they" are more interested in controlling our activities, not much they can do about nature's.
It could also be that if they were to look at the sum of agw AND a natural high, it might be so awful without mitigating our contribution that there would be no point in us changing anything, and maybe that's why they leave it alone.
This sentence from Steve Mc at ClimateEtc stood out for me: "It looks like a real dog’s breakfast."
Too many unappetizing parts.
Thanks to JJ for finding the signal in the noise (ctrl-f mcintyre).
Steve is on fine form:
"If Marcott is right, the ability of early 20th century Northern Hemisphere societies to cope with the 1.9 deg C increase between 1920 and 1940 bodes well in my opinion for the prospects of adapting to the lesser temperature increases projected in the next 60 years in most climate models. Of course, it is also possible that the 20th century portion of the Marcott reconstruction is completely worthless."
He's referring to the numbers in the "regional stacks" on sheet 3 of the unpaywalled excel file.
The numbers for the NH go 0.27 -1.66 -1.87 ... at 20 yr intervals going back from 1940, so the 1920-1940 change is allegedly 0.27+1.66=1.93C.
Meanwhile in the SH the numbers are 0.82 0.62 1.22 ..., so apparently the temp there fell by 0.6C between 1900 and 1920.
Here is a plot of the Marcott proxies over the last 300 yrs. The time axis counts back from 1950, so 0->1950 and 50 -> 1900.
Each graph has had its mean subtracted.
You can see that there are no proxies that give a sharp uptick in the last 100 years, so their graph appears to be a fabrication. There are two slightly spiky ones near the top of my graph, one going up-down, the other going down up but they are proxies 25 and 46 in the SH so they can't be anything to do with the claimed NH spike.
Paul Matthews,
That is a remarkable graph. It is going to be "interesting" to find out the story on the "blade" of their hockey stick, since the actual proxies would not seem to offer any large uptick after 1900. They have to infill data, splice temperature records, or some other dubious procedure which has nothing to do with the proxies they claim to be studying.
I'm no stats guy, but the Marcott data seems to have a "divergence problem" for the entire 20th century.
A couple of things I have seen noted include a lack of high resolution proxies to say anything useful in comparisons between the 20th century and centuries in the distant past. Look at 1940-2000 in the temp. records to compare divergence during the past century with the impossibility of their data presenting any similar periods in centuries past.
I found this comment by Jos at Judith Curry's to be helpful:
http://judithcurry.com/2013/03/11/lets-play-hockey-again/#comment-301951
Reading all the comments and looking at Paul Matthews graph makes me wonder if there is not another Gergis et al debacle in the making.
Paul: Nasty chart - at least nasty for Marcott. Have you looked at Figure 5.7 in the second draft of AR5? I do not see a HS except for the imposition of HADCRUT3 NH.
Funny, Alarmists are quick to decry any of what they regard as politicization of a study they don't like. Yet here is a flagrant attempt to exaggerate, distort, and politicize a dubious study which DOES NOT SAY what the Congressman claims it does:
US Congressman gets Marcott conveniently wrong
[emphasis added]
So is this new sort of verification - disagreement with measured temperature data does not matter as long as the reconstruction is "statistically indistinguishable" from Mann's (debunked) reconstruction - the novelty of this paper ?
I wonder who the peer reviewers were?
Mailman
Thanks Skiphil bernie theduke. I wonder if someone - good sceptics that you are - would be so kind as to check that my graph is correct? The 73 proxies are all in the excel spreadsheet that is available on the journal website, each on its own sheet starting on sheet 5. In most cases the yr and temp columns are E and F but in a few of them it's different.
mailman, I think I can guess who one of the reviewers was!
Mailman--
I've graphed the 73 files using the published age and published temperature (usually columns E and F). Up on Dropbox below:
https://dl.dropbox.com/u/75831381/Marcott.SM.database.S1--law.xlsx
There is also a "Marine09" age with 1-sigma error (usually columns G & H). Possibly that should be used as the x-axis. It makes little difference on a graph that often extends to >10000 years but can make quite a bit of difference at x=0. Do you have a feel for which age we should use as the x-axis? Might be interesting to plot the horizontal error bars on the estimated ages, just to further the "dog's breakfast" interpretation of Steve McI.
Several of the later Excel sheets (IDs in the 60s) seem to include a second batch of values, apparently at lower depths, that sometimes result in double-valued curves or extremely sharp discontinuities. I am suspecting these are oversights that were meant to be removed.
There are also one or two cases where there are some zeros that probably were meant to be missing.
Whoops-- I meant to direct this to Paul Matthews
I've graphed the 73 files using the published age and published temperature (usually columns E and F). Up on Dropbox below:
https://dl.dropbox.com/u/75831381/Marcott.SM.database.S1--law.xlsx
There is also a "Marine09" age with 1-sigma error (usually columns G & H). Possibly that should be used as the x-axis. It makes little difference on a graph that often extends to >10000 years but can make quite a bit of difference at x=0. Do you have a feel for which age we should use as the x-axis? Might be interesting to plot the horizontal error bars on the estimated ages, just to further the "dog's breakfast" interpretation of Steve McI.
Several of the later Excel sheets (IDs in the 60s) seem to include a second batch of values, apparently at lower depths, that sometimes result in double-valued curves or extremely sharp discontinuities. I am suspecting these are oversights that were meant to be removed.
There are also one or two cases where there are some zeros that probably were meant to be missing.
oohhhh myyy ... we are a part of some "troll army"??? but are we "fully-mobilized"???
The journalist (sic) seems to be doing some extreme framing -- for instance it's not evident that Marcott used any of the phrase "charging into the fully-mobilized troll army" yet the article is written as though Marcott said something like that. The lack of quotation marks suggests that the journalist (sic) is putting his own spin on the conversation, though:
Marcott on Mann and skeptics
[emphasis added]
See CA, where Steve Mc is wading into Marcott et al. (2013). It seems that Marcott is not exactly forthcoming about his 20th century work except to say that it is "not robust".... yeah, we thought so:
Marcott Mystery #1
[emphasis added]
@Paul Matthews at 12:45 pm: I'm flattered that you might think I could give you a valid opinion on your graph, but I'm really not expert enough to do that with any confidence. My interest in this debate/controversy is mostly from a historical perspective of science and how it confirms something resembling irrefutable truth. I simply don't have the time to immerse myself in all the science and mathematics required to gain a full understanding.
I'm a pretty good bullshit detector, however. History is full of liars and their behavioral patterns don't vary that much.