In 1997, the 121 m Lomonosovfonna ice core was drilled in Svalbard. As of mid-2009, when Hu McCulloch and I wrote CA posts on this core, nothing had been published on
O18 values prior to AD1400 nor had any Lomonosovfonna data been archived, even for the post-1400 period.
Both Hu McCulloch and I, in separate CA posts here and here, speculated that the withheld O18 values prior to AD1400 would elevated values. A digital version of the pre-1400 data became available this week in connection with Hanhijarvi et al and confirmed our surmise, as shown below.
The figure below shows Lomonosovfonna O18 values back through the MWP as calculated from data in the SI to Hanhijarvi, Korhola and Tingley. As Hu and I had surmised, the O18 values prior to AD1400 (denoted by dotted line) were elevated.
Figure 1. The Hanhijarvi et al SI provided reconstructed temperature at Longyearben using Lomonosovfonna. The post-1400 portion was a linear transformation to O18 data at NOAA (used in Tingley and Huybers 2013). The O18 plotted here used the linear transformation to recover the underlying O18 values.
Lomosovfonna O18 values after AD1400 had been shown in a number of articles by Isaksson and coauthors.
Hu’s and my interest in pre-1400 Lomonosovfonna data had been prompted in 2009 by the publication of a multiproxy reconstruction of sea ice by Macias Fauria and coauthors, including Elizabeth Isaksson, claiming that sea ice was unprecedented since AD1200. In the post, I observed that a related article by the coauthors (Grinsted et al 2006) showed that “washout indices” at Lomonosovfonna were at much higher than modern levels prior to AD1200, but, by sheer coincidence, the reconstruction began in AD1200.
The reconstruction used d18O measurements from Lomonosovfonna, Svalbard, but Macias Fauria refused to provide the data (see CA post here). Macias Fauria said that the data had to be obtained from Elisabeth Isaksson, disregarding the fact that Isaksson was a coauthor of the article.
Hu McCulloch took up the issue in a CA post entitled Svalbard’s Lost Decades, forcefully challenging Grinsted, who had commented on other CA threads, to explain the omission of this portion of their record. Grinsted responded in comments, but failed to provide a satisfactory explanation. Nor was Grinsted more responsive to requests for pre-1400, also pointing the finger at Isaksson.
Lomonosovfonna was again used later in 2009 in a multiproxy reconstruction by Kinnard et al. Kinnard et al provided an excellent archive, including a digital version of post-1400 Lomonosovfonna that is a slightly smoothed version of the data now available.
The post-1400 version of Lomosovfonna was also used in 2009 by Tingley in the first draft of what was recently published as Tingley and Huybers 2013.Tingley and Huybers 2013 discuss a post-1400 reconstruction using ice core O18, with a network almost identical to Kinnard’s, including the post-1400 portion of Lomonosovfonna. I’ll return to this reconstruction in a separate post.
In May 2011, Isaksson archived the post-1400 data at NOAA (see here), citing Macias Fauria et al – making me wonder whether Macias Fauria used Lomonosovfonna data in their reconstruction between 1200 and 1400.
In July 2011, in a little noticed article (Divine et al 2011 here), Isaksson and coauthors showed pre-1400 data in the following graphic (but did not archive the data):
Figure 2. From Divine et al 2011. This corresponds to Hanhijarvi SI version.
Isakkson and coauthors observed:
However, both the reconstructed winter temperatures as well as indirect indicators of summer temperatures suggest the Medieval period before the 1200s was at least as warm as at the end of the 1990s in Svalbard.
Later in 2011, I reviewed the comprehensive Arctic O18 data from Kinnard et al 2009 here. I observed that the Kinnard Hockey Stick was not observable in the long O18 data and therefore had to come from something else, illustrating this with the following graphic.
Figure 4. CA comparison of long O18 data from Kinnard et al 2009 to their reconstruction.
In their SI, Tingley and Huybers 2013 show a post-1400 reconstruction using Arctic O18 values, but draw opposite conclusions. I’ll return to this in another post.
Postscript: The dating of the bottom portion of the core appears to have changed dramatically between Macias Fauria et al 2009 and Divine et al 2011, increasing the earliest dates by about 400 years (from the 12th century to about AD770). Because of the exponential thinning of ice cores, early dates are tremendously sensitive (particularly in short cores) to the estimate of the thinning parameter, as shown in the following graphic from Divine et al 2011:
Figure 3. From Divine et al 2011. Because of this redating of the bottom portion of the core, it appears that the version now available is not the same version as the one used in Macias Fauria.
Climate Audit 
63 Comments
Reblogged this on Climate Ponderings and commented:
GOOD READ!
I don’t understand what it means when you say recon (inverted) in Fig. 2. Can you explain what that means?
I would suppose the reconstruction is in temperature, and so is there a scaling factor in the O18 values, or in the temperature values?
Ed — Sea Ice Extent is an indicator of coldness, while d18O is an indicator of warmness, so that one of them has to be inverted in order to compare them.
Steve,
First, I’m a fan, and really appreciate the time and energy you spend trying to make the science solid.
I realize you are a bits and bytes kind of a person (i.e., let the data show the way), but I’m wondering whether you have a position statement on your general views. At one time, I think I saw that a comment that you agreed with the basic science behind C02 as something that warms the planet (some amount). I also recall a statement that you had seen many “hockey sticks” in your career, and they almost all were too enthusiastic. However, I don’t know much beyond that. It would be great to have a simple statement of what you are trying to achieve, and why you are expending the energy. If you have it somewhere, I think many would be interested.
I personally think you are a truth-seeker, or perhaps an amazing BS detector, and are using the years of your experience to dig through the natural bias of people, as even scientists are biased (as I’ve read in a number of Nature articles with the bio-sciences).
In any event, I do not understand why the climate scientists do not embrace your work. It seems to me with like minded careful scientists, and I’m certain there are these folks even in the Paleo world, that you could put together some really solid analysis. The general sense I get from reading your work is the Paleo people are off chasing a specific hockey stick result. Wouldn’t it be great if instead they were using your statistical experience, to produce truly clear results? A lot is riding on this question, as I’m sure you appreciate.
Ed.
Why is the “about” page not good?
I don’t see a position statement there. Did I miss it?
There are two charges as to why some other climate scientists don’t embrace this kind of work (perhaps 3)…
1. SM does not, as the firstfruits of his writing, assume the path of least nefariousness on the part of his targets, nor give ‘proper’ credit/homage to ‘large majority of things that the IPCC does well, has gotten right, works really hard at, etc.’ (in the above case here the implication being that the authors ‘withheld’ contrarian ice core data)
2. SM does not pursue work that ‘furthers the science’, but instead seeks to ‘muddy the waters’ or ‘increase doubt’. For example, instead of figuring out what temperatures were in paleo-times, he works at trying to expose why ____ et al isn’t correct. The logical end-conclusion is that eventually nothing is correct and nothing can be known.
3. (in light of the other 2 I suppose) for whatever his detractors impute the worth of his skills, they will follow it up with the claim that he’s only interested in applying them to mainline scientists, leaving ‘obvious’ problems in contrarian publications off the menu.
etc. etc.
Fortunately, the charges Salamano lists are not true, as Steve’s posts during the almost-10-year history of this blog show pretty clearly.
In particular, Steve’s long-time efforts to get climate scientists to publish their data and code along with their conclusions are one of the best ways to increase clarity and transparency, thus un-muddying the waters and reducing doubt.
Armand,
While I agree that the list Salamano is not objectively true, I can imagine it being subjectively true from the POV of those who have been the focal point of Steve’s enquiries.
It must be very difficult to embrace the work of one who does not readily embrace the work of your team mates, advisors and professors.
Jeff, scientists are supposed to be objective rather than subjective. That’s sometimes tough, but other fields such as molecular biology and physics manage to do it, so I don’t think climate science has a credible fig leaf here.
Thanks for that little screed Salamano.
I needed a good laugh this morning and your observations provided quite a good laugh.
Pure comedy gold from the get go. Please keep up the great work and you will soon be in the Mann-Tamino et al Hall of Fame.
Proper analysis muddies the waters? Interesting.
Jeff is on the right track.
Any regular reader of ClimateAudit, RealClimate, and Tamino over the past 7-8+ years can probably see the above charges leveled against Steve McIntyre many times and instances. Many calls for him to publish his own reconstruction given his knowledge/conclusions about temperature; many calls to ‘audit’ contrarian work with the same vigor as he does with the typical “et al” crowd, etc. etc.
One of the most frustrating things for progressive-based Climate advocates is that Steve is not explicitly linkable to “astroturfed” organizations in which their funding extends back to someone affiliated with an industry or individual set to materially benefit from the obfuscation or rejection of any sort of policy conclusions that arise from catastrophic climate projections. Many other contrarian scientists fail this check, which avail advocates the ability to recommend the rejection of the arguments by that appeal. It’s ironic though, considering there’s just as much astroturfing on the part of advocacy groups, as Greenpeace, George Soros, and Green Energy industries have just as much a interrelation and material benefit industry as the Koch Brothers, Fossil Fuel industries, and the Heritage Foundation (to name a few). There is some sort of insulation of governmental organization and “pro-Earth/Peace/Green” groups from charges that their scientific proclivities are linked to their economic pursuits (and therefore must be rejected). Unfortunately, this means Steve probably has to stay self/citizen funded (and probably poorly so) for the forseeable future, and not end up being a “fellow” or otherwise linked to any of these organizations, lest they be able to leap to this ridiculous and reflexive logic appeal.
However, I do think it’d be important that the following try to occur:
1. When there is a Marcott et al publication that comes out, that there be at least some effort to show what the publication will conclude when proper statistical methods are applied, or perhaps moreso do what needs to be done to show how Marcott can be reasonably improved. Instead of only showing why something is doubtful or less rigorous, getting down what IS more rigorous or probable. As it currently stands, more of what Steve can publish is most likely for Statistics Journals than climate, even though it’s clear that Climate Science can benefit greatly from statistical acumen that people like Steve and Grant put forth.
2. Don’t you think there’s enough knowledge out there, enough discussion out there over the last 10 years for an actual reconstruction publication to be made? I mean there’s now a lot of archived paleo data (to Steve’s credit), and there’s a lot of statistical muscle inserted into the paleoclimate world (much to Steve’s credit) — It’s high time that all the published science minds at this and other blogs get together and publish some reconstruction (with whatever error bars the statistics requires), from all the sources that are now publicly available? It seems to me that this can/should happen. Why not?
My personal conclusion from my research on tree growth is that extending a climate reconstruction farther than a couple of hundred years back is not possible due to confounding of temperature and precipitation (and other reasons). Thus this proxy is not useful and should not be used. I think boreholes have unresolvable problems of inverse inference. Other proxies seem to have weird things going on with discontinuities, sudden changes in behavior, confounding, etc. So I don’t see the time is ripe for a reliable reconstruction yet. Which is why I have not revisited this topic since 2007.
I think the problem is that few are going to accept that “nothing” can be known from these proxies when it comes to temperature. That would plausibly challenge the whole world of paleoclimatology in its current principle pursuit (reconstructing temperature back beyond the instrument record). Especially since there’s a whole house of cards (for lack of a better term) built from the set of assumptions/conclusions that appears to fit together regarding temperature reconstruction.
Isn’t it possible that someone could publish using the existing framework, adding an increasing measure of statistical prowess, and come up with a reconstruction that is incrementally improved on the previous without guaranteeing “this is absolutely what I believe to be the final story”? In this way, no one can criticize your results without also criticizing theirs, unless the criticism is something like “the verification is better against the modern temperature record using these (controversial proxies, annual New York Yankees team batting average, etc.) therefor it SHOULD still be used”– a criticism that might as well be permitted to exist.
Re: Salamano (Apr 14 12:09),
I think you may have answered your own question:
It seems unreasonable to ask more of others who have already spent so much time and done so much work on a voluntary basis. I think it is more reasonable to ask that the well-funded, publicly-supported, career climate scientists be required to meet some set of rigorous external and well-defined objective standards in researching, archiving, analyzing and reporting findings. Funder’s oversight and peer review doesn’t seem to fulfill all of those requirements. It is my understanding that this sort of thing is the object of much of Steve’s work and with reportedly increased and better archiving, it seems that he has seen a measure of success.
How well proved the common refrain of unprecedented 20th century anthropogenic warming is, or at least should be, of enormous concern to all of us. I don’t think it is too much to ask that the people behind these claims be held to a very high standard of proof given the consequences of government(s) policy and spending decisions based on this research. I am also given to understand that there was a publication goal tending issue in that the goal posts were somewhat obstructed to the opposition so that the most direct way of scoring points, so to speak, was to observe and report on own goals through self-publishing. Perhaps it is different now. I don’t know, but would assume that this could have contributed to any disinterest in pursuing the formal publishing route.
I marvel at the amount of work that is done pro bono by interested citizens like Steve and the many other contributors to this discussion in order to promote further understanding. I have learned so much from all of you – even those who don’t necessarily appreciate independent blog-based science and analysis, so thanks.
I hope you keep doing what you are doing as I’ve much to learn and I find the subject fascinating. Without blogs, it would be much more difficult for non-academics, like me, to explore and gain knowledge and understanding.
PS – I apologize in advance if I’ve misread the intent and thus am speaking out of turn.
You seem convinced that proxies point the way and we all should strive to find a way to use them to an answer worth considering. You mention all the data available to analyze. There’s far more data on securities market moves in the DOW or S&P 500. The issue is not lack of data. The issue is making sense of it. The data may be unreliable or there may be relationships between data and the reality it reflects that we cannot interpret accurately. Perhaps like climate scientists studying inverted Korttajarvi sedimentation, we’ll soon see quants studying securities markets and inverting the sign of market movements as a means to assure return on investment (or Big Pharma treating larger death rates as a justification to approve a new drug). The proxies may be dependent on six or ten or fifteen variables, and we may not be able to ascertain how the proxy reacts to each variable separately, therefore forever rendering us unable to draw conclusions from amalgamated inputs. Climate Audit hits the nail on the head – just publish your raw data and publish your code for the world to see. Then your assumptions are there for everyone to see, as are your statistical techniques with respect to the data. I understand PCORI is requiring all medical researchers to publish not only completed study data, but any data collected before payment of the associated grant. That way, all the data gathered is available for every researcher going forward. If you want conclusions, then press for Bona-Churchill data to be published less than 10 years after it is collected. We don’t have an obligation to make the proxies prove something regardless fo how well we understand them; you are starting from the premise that all is lost if we cannot use them even as a very faint Ouija Board.
Re: Jan
You are correct, and I was being somewhat rhetorical. Nevertheless, there are (and have got to be) regular readers that are climate scientists that are publishing in the field. They are either scoffing at the prospects of eliminating the ‘clear’ climate signal entirely by applying statistical rigor from one field that is not necessary for this field. At times, there is a declared uniqueness to climate science that sets it apart from things like engineering, though Tamino has lately been discussing work in astrophysics that may well be applicable. Anyway, if Steve is (truly independently) working to augment statistical standards as it applies to paleoclimatology, then there should also be a clarion call for someone to try to put it to work with published material (but must have an answers for the aforementioned pitfalls they are up against within prevailing climate scientific understanding). Otherwise, would not arguments of “doubt peddling” have a non-zero level of merit?
Re: Douglas
I am re-stating what the prevailing climate understanding is– the one that has now become that which you must overcome/explain within your possible publication (rather than the idea that all proxy data is open to limitless exploration with no underlying expectations/hypotheses). Simply put, the instrumental record exists, and that is pretty-much the best validation metric we have across which to grade proxies, yes? All attempts to rule a particular proxy in or out can end up as the word of the one who first attains/publishes it (or simply about it). Some success has been achieved in working to get particularly specious proxies ruled out at the source (original author corrections, etc.) which should naturally flow to the reconstruction papers themselves (though hasn’t for some of them). But, I think that declaring (without demonstrating through publishing) that the main pursuits of paleoclimatology are fruitless and worthless are not going to be received well. There’s already a lot of work out there to the contrary that would need to be answered. They have the ball and the possession arrow.
Re: Salamano (Apr 14 12:09),
Sometimes, the best that science can do is “doubt peddling”, i.e. sometimes our best understanding is: “sorry, we’re clueless.”
It is completely UNscientific to create models that have little or no predictive value. And right now, the climate models aren’t doing so well. Lots of heat about why… but normally scientists would be fascinated when someone with expertise in stats, and in exposing fallacious data models, has something to say. But that’s not generally the case in climate science.
I am NOT saying the following to argue the case in this thread. The implication of correct or incorrect uncertainty is simply this: if it is true that 2C temp rise due to man-made warming is imminent, then that means we can (and should) correct it. But if we don’t know that, then we don’t know that we have the ability to correct whatever warming may be coming, and we probably ought to be finding ways to survive the warming (or cooling), while studiously working to better understand our planet.
So I would suggest it is very very good science to do a better job of understanding our uncertainties.
Perhaps you should spend time reading about the wrongfully convicted where the evidence pointed to one outcome and other evidence that didn’t fit was dismissed.
Everything Steve has done has furthered the science. The idea that leaving bad science alone makes science better is misguided.
Salamano, your description of Steve’s approach would be a perfect description of the auditors for my company. “All” they do is correct the financial statements of other people’s companies.
Perhaps the next time thier audit finds an error in the balance sheet of my company, I should ignore their suggestions and ask them why they don’t go and start their own company.
“In any event, I do not understand why the climate scientists do not embrace your work. It seems to me with like minded careful scientists, and I’m certain there are these folks even in the Paleo world, that you could put together some really solid analysis.”
The recommendation that climate scienists embrace professional statisticians in their work was made in the Wegman Report some 7 years ago, so if there are paleos out there willing to take on board the help they’re keeping a very low profile, If you want the answer to your question ask it on realclimate.org, but wear a crash helmet and body armour when you do.
“The general sense I get from reading your work is the Paleo people are off chasing a specific hockey stick result.”
The soubriquet they’ve given themseles is “The Hockey Team” so there’s a clue in there somewhere.
Sorry should have been realclimate.com. Pay a visit anyway to compare how the proceedings are managed by “real” scientist Gavinn Schmitt. Try it this way. Tell them climateaudit has just posted an article showing the pre-1400 Lomonosovfonna data had elevated temperatures. Then duck.
I’ve posted on realclimate before, and what I’ve noticed is that they get the last word. They often set up a comment as if it were a straw man, and then tear it down. They refuse to post the rejoinders. So I’m not going to do that anymore. It’s not an open forum.
Ed, a good place to start with the McIntyre oeuvre is the Ohio State presentation 6-8 years ago.
What’s funny about Real Climate, that’s where I first read about Climate Audit. That was before Steve became “He Who Must Not Be Named” over there.
Bernal,
It is important to remember that Climate Audit was started as a result of how Steve’s (and Ross’s) questions were mistreated at Real Climate. IIRC, they applied the very same immoderation skills on Steve’s question and comments that they employ today. Team mythology tries to paint it the other way round.
Steve: CA was started in direct response to RC where much of their early work was directed at us. However, it’s not correct to say that it was started because of the way questions and comments were treated. I vehemently objected to censorship and stalling of my attempts to comment there later in 2005, but by that time both blogs had been started.
Conceded Steve, memory too is subjective. Climate Audit was not created because of Real Climate, but some of your earliest posts were made in response to your inability to respond adequately at Real Climate to incorrect interpretations (your “reconstruction”) being touted at Real Climate.
Regardless Real Climate was started to defend the hockey stick interpretation of climate prehistory.
Steve: this isn’t right either. Climate Audit was started because of Real Climate. Many of my earliest posts were direct responses as you observe. But I wanted to set out my version on my own location. Their censoring of my comments was a later issue.
As I recall, Climate Audit started after Steve and Ross published MM03, which revealed the grotesque core of Mann’s MBH98 reconstruction.
Sonja Boehmer-Christiansen thought the paper was sufficiently important to feature it on the top page of the E&E site. So far as I know, that had never happened before, or has since. Sonja B-C was right, though. That paper was a true watershed event, and it attracted quite some attention.
The response at RealClimate was to launch a violent attack against the message of MM03, and in defense of Michael-Mann-the-much-maligned. Steve started CA as a way to defend and elaborate his work and himself. I remember that early version of CA. Steve was, ummm, a little less personally reticent to call out a spade in those heroic beginning days.
Steve: Ross and I had webpages in 2004. However Climate Audit did not start until 2005, more or less concurrent with publication of our 2005 GRL and E&E articles in late January, which were widely covered. RC started in December 2004. Its early attacks were intended to preempt our forthcoming publications. For the most part, their attacks were accepted by the community, which paid little attention to our actual articles. Their attacks were based on our 2004 submission to Nature, rather than our actual 2005 articles which altered focus, in some places considerably.
Okay, stop. Stop. Stop there – stop there. Stop. Phew! Ah! … Climate Audit was started because of Real Climate…blah blah blah. Cardinal, read the the tree rings.
Yay! That means I have been a reader right from the start!
I found emails in my sent items folder, where I circulated “Backgrounder for McIntyre and McKitrick “Hockey Stick Project” January 27 2005.
I sent it around to friends who work in risk management of a large bank. My comment way back then stands the test of time:
“Pay particular attention to the obstuctionist tactics of Mann et al. They are acting like traders who have mismarked their books.”
Ed that’s my point. I don’t want to know Steve Mc’s motivation because he just deals with the science, and in my experience (after nearly seven years of reading this blog) it’s just about that. Visiting realclimate.com was totally different. If I had Steve’s knowledge of statistics and his experience my motivation would be to show up the finagling that’s going on in these papers, I haven’t (got his knowledge and experience that is) and don’t know his motivations, just that he’s doing a service to science.
I’ve posted on RC before. If there is any serious results that offend the gate keepers, your post is deleted. Less sensitive ones are relegated to the “bore hole”.
About 5 years back, Tamino (F. Grant) & I tangled on his CEL analysis. In it, my analysis, using Fourier Convolution, or spectral analysis, showed a plateau forming over the next few years.
Suddenly the topic was closed, discussion ended.
Strange, that DOD & NASA accepted such analysis for manned space flights, but not OK for RC.
Just when you though you knew when yesterday was …
I like the nearly horizontal part at the bottom, where as you point out, the sensitivity to assumptions goes through the roof.
w.
It would be interesting to see how the series looks as the assumed compression rate is altered. In fact, I wonder if one might get an estimate of the compression rate by running a non-linear regression of the varve series on an ice core series from neighboring Greenland where the temperature proxy is dO18 instead of thickness.
All of Lonnie Thompson’s core also have very pronounced exponential thinning. Thompson has made almost nothing available on the dating of his cores – the recent Quelccaya archive has a little data, but there’s almost nothing for Dunde or Guliya. I seriously question Thompson’s dating of Kilimanjaro. AR4 editors (remarkably) accepted this review comment.
Poking around the NCDC archive, found this which used volcanic and climatic markers to match canadian ice cores against greenland cores. The tables showing the match points really highlights the compression.
Steve: I wouldnt say that this example is really on point. the phenomenon is much more apparent in tropical and temperate cores. There’s a core in Chile which thins out in a couple of hundred years (Cerro Tapado by memory, but need to check)
Off topic, but is it not time for some updates on Climategate 3?
Perhaps AG would revisit 1200.
========
I think the point of the posting was to show that a graph was published which again showed warming. However when you had the previous 1000 years data its doesn’t. In fact its totally outrageous that this stuff is being allowed to be published. Maybe Ed above is asking SM to be more forthright ie: “Yes, by now I realize this AGW/climate science is total XXXX!”
“I do not understand why the climate scientists do not embrace your work. It seems to me with like minded careful scientists, and I’m certain there are these folks even in the Paleo world, that you could put together some really solid analysis.”
On the contrary Ed Barbar. It is exposure of the lack of solid analysis that is clearly unsettling the cosy consensus club and gaining more converts back to the original null hypothesis that our climate is always changing and we’re only just beginning along the path to understanding why.
Under the circumstances you could hardly expect the self appointed King’s fine coutouriers to readily accept they’ve got the King decked out in his underpants, what with all those years of drilling ice cores and mucking about with sediments and tree rings now could you? As for the King he doesn’t want to be a laughing stock either, given all the gold and imprimatur that he’s flung at such pursuits. With no CO2 induced warming for 15+ years it’s really looking that disastrous for them all if they can’t ‘shoot’ the messenger.
“withheld O18 values prior to AD1400 would elevated values.” should read “withheld O18 values prior to AD1400 would have elevated values.” ?
you wrote “..claiming that sea ice was unprecedented since AD1200”
What about sea ice was unprecendented? The loss of? I don’t understand.
Salamano Posted Apr 14, 2013 at 4:31 PM
“Isn’t it possible that someone could publish using the existing framework, adding an increasing measure of statistical prowess, and come up with a reconstruction that is incrementally improved on the previous without guaranteeing “this is absolutely what I believe to be the final story”? In this way, no one can criticize your results without also criticizing theirs, unless the criticism is something like “the verification is better against the modern temperature record using these (controversial proxies, annual New York Yankees team batting average, etc.) therefor it SHOULD still be used”– a criticism that might as well be permitted to exist.:”
This might be just me, Salamano, but without more details this sounds like utter BS to me.
Maybe that’s why it hasn’t happened yet.
A reconstruction that doesn’t use Bristlecones or upside-town Tiljander or whatever else gets rejected by reviewers because it doesn’t verify against any sort of calibration period (like the instrument record) better than a reconstruction that does use them all.
But there’s got to be something published either to challenge the earlier author’s contention that proxy X is ‘good’ or ‘bad’ for temperature (otherwise future authors will just defer to them as previously declared), or to assemble a reconstruction without all the ‘bad’ ones and deal with the fact that it won’t validate/calibrate well. Perhaps it will involve new statistical applications? (That’s how MBH broke ground, yes?)
Salamano Posted Apr 14, 2013 at 6:37 PM
Salamano, I think, in your apparent need to salvage something from these reconstructions or the proxies developed for that purpose, you have failed to consider the proposition that the basic selection process of proxy responses to temperature is flawed. The basic premise for proper selection of proxies is to develop a well thought out criteria that is made prior to selecting proxies and avoids the statistically unsupportable post facto selection based on how well the proxy response is to modern temperatures. The selection criteria should also have a reasonably based physical rationale.
If you have viewed sufficient individual proxy responses you will note that they appear to meander through history and while most probably are affected by temperature that affect is only one of many factors affecting the response. You can find these series with upward trending finishes emulating the modern temperature trends , others are downward and some are rather flat. This some directional trending can be shown to occur with ARIMA series with either larger AR1 or long term persistence. Selecting a series ending proxy response trending upward after the fact is not a difficult task and a trend that occurred more or less by pure chance.
It can be argued that proxies can respond to a known and well-dated event that would be predicted to change the global or regional temperature in a dramatic fashion, e.g. a volcanic event. Unfortunately a closer look at several of these same proxy species will most often reveal that the response magnitude to that event varies greatly amongst the individual proxies.
I would hope that better temperature proxies can be developed in the future, but that development would require a prior physically based selection process. I would judge that best proxies would be based on well understood chemical and/or biological processes that would allow precise measurements.
Don’t you think this is challenging the entire paleoclimate industry too much? I get that there are statistical acumen for engineering-level significance, and there are acceptable tolerance levels in CNC machine-work that climate science isn’t meeting…
I just don’t think the whole arena of dendrochronology (and others) are just going to roll over and consider their work too uncertain to continue, especially considering the work they have done already to get to this point. Is it possibly a matter of no one even noticing the field until as they worked to develop/justify their specialized methodology and only later jumping on them once they finally start applying it for results?
You gotta think for a proxy researcher gazing across the sea of possible proxy members and seeing a lot of them responding a certain way that just happens to look like the instrument temperature record that they’re going to say they’re on to something while everything else is contaminated. Saying nothing can be known yet is just not going to make it into the next IPCC report, so there’s got to be new ground to publish in the arena of why its not so if it’s not so.
I like the idea of concentrating on areas/regions in selection criteria rather than weighting specific trees within a sample of an accepted area. Has anyone published reconstructions that have attempted to do this to see what results? The other thing I was interested in was adjusting selection criteria to screen in proxies by slopes rather than magnitudes (ie., to have members that only go up/down for the RWP, MWP, and LIA/ModernTemp record, etc.) …Rather than just the Yamal-ian uptick at the very end.
A few years ago, a couple of very senior people in the field told me – under drop-dead promises of anonymity – that we had pretty well destroyed the basis of Mannian and similar proxy manipulation and that the only way forward was the development of properly calibrated proxies – a project that they thought might take 10-20 years. Other than IPCC and politics, that was probably a good guess.
I viewed my objective as simply trying to see if these guys had proven what they claimed to have proven i.e. an audit. I’ve avoided presenting my own assembly of proxies because I’m conscious of the problems.
Having said that, my advice to people trying to move forward is to work on careful reconciliation of all available information in specific regions. For example, until Ababneh’s bristlecone results were reconciled with Graybill’s, they shouldn’t use Graybill’s results. Instead, Ababneh’s results have been ignored. Same with Yamal versus the decline.
IMO people in the field spend far too much time on fresh multivariate methods and far too little time on the properties of the proxies themselves.
Steve, on that note, I recently got into a good amount of debate (on a forum frequented by those in my profession) about the Hockey Stick after Marcott was published when I attempted to convince the masses there that there are no satisfactory multiproxy reconstructions. I linked to MM2005, a convincing argument for MBH98’s invalidity, and attempted to explain that the problems there are present (in some form or another) in virtually all studies on this topic.
The general reaction among those arguing against me was disbelief that so many studies could repeatedly make the same errors. This much is understandable; I think it is hard for anybody to believe what has been going on in this field for over 10 years. They also wondered why you (or others) had not published formal responses to these studies if they were so obviously wrong, and argued that one paper (MM2005) + people posting stuff on the internet does not hold as much weight as the tens of multiproxy studies not directly addressed in the literature.
I wasn’t sure how to explain that you’ve issued as many comments as possible, but that the multiproxy studies churn out faster than it is possible to decompose and formally reply to them all.
So, my thought was, is there any practical way to write a paper that directly names and calls into question (substantially all of) the multiproxy studies that the IPCC relies on? I don’t see any other way to address the huge bulk of erroneous published papers without such a shotgun approach. Maybe such an article, with many authors and papers named, would make the field really think about whether they are doing this the right way.
Like you say, so many examples of science-abuse in this literature body, it’s hard to respond comprehensively on the playing field of the peer review journals.
I can’t speak for Steve but it seems most effective to try to produce a site for exploration of selected examples of these, as he has. I and clearly many many others really appreciate these and have learned a lot from this forum and his deconstructions.
Besides, there is no journal I’m aware of that specializes in publishing pieces on ‘errors and omissions’ or the like. Rather, to rebut on their playing field, one must submit a Letter to the Editor for each specific abuse.
I recommend you try it sometime at NATURE. You will be prompted immediately to submit to their $200 paywall.
To my mind, it’s now idiocy to wait for peer reviewed journals to self police, and given their hurdles above, it’s perhaps idiocy to try to convince them of anythng wrong with their products. On the other hand, if new, better work is emerging, then perhaps these same journals will drop the current line of thought like last night’s bar pickup.
@Salamano
If one gets the calculations and therefore the tolerances wrong in a CNC machined part, it may well fail and possibly cause the failure of the construct of which it is a part. In the case of e.g. a coffee grinder that’s an annoyance. In the case of an airliner’s wing box, the same type of error and failure would be catastrophic.
The core problem [bad pun, I know] with the entire field of proxy based temperature reconstructions is the reality that policy decisions with far reaching economic and social consequences, involving by now in excess of a trillion dollars in OECD countries, have been made based on reconstructions and projections derived from what is essentially an amalgam of “normalized” junk science. With attendant catastrophic outcomes, as evidenced by the very real energy poverty on display these days in e.g. Germany and the UK.
Salamano Posted Apr 15, 2013 at 2:57 PM
Salamano, I really am not much interested in the political and emotional problems of doing the reconstructions correctly, but I have thought a bit about why it is so difficult for otherwise intelligent people to see the errors in their approaches – as I noted over at TAV today.
I compare what I see in climate science’s handling of proxy selections and reconstructions to what I have experienced with individuals using investment strategies based on in-sampling testing and not being aware of what that portends without out-of-sample verification. I have seen people who are scientists and in the hard sciences not able to come to grips with these statistical problems in investing and climate science.
I see two reasons. The first is that these individuals want to believe that the system works and for whatever reasons. Secondly I think that hard scientists, and engineers also, work from theories and based on theories might do some initial in-sample testing looking for some confirmation. Now in investing and in climate science out-of-sample testing is difficult to do since you have no control over the experiment while the hard scientist can run a confirming experiment with control over the conditions. Out-of-sample tests in investing and climate science have to either await for time to pass in order to confirm their in-sample tests with data that they could not possibly have had for in-sample testing or scrupulously set aside data and not look at it before finishing the in-sample tests. Setting aside that data and not looking at it is a tricky proposition.
Without out-of-sample testing one has to use procedures in temperature reconstructions where the selection criteria are well established a priori and where the individual doing the reconstruction once the proxies are selected cannot throw out data without a very obvious, rational and detailed reason.
Jason:
“The general reaction among those arguing against me was disbelief that so many studies could repeatedly make the same errors.’
It’s happened in other fields has well. There were (IIRC) 39 “independent” studies that showed that post-menopausal women on hormone replacement therapy had better health outcomes than those who did not. However, these were all observational studies with selfl-selected cohorts.
When a properly controlled double-blind study with randomly selected cohorts was finally done, the results were so dramatically in the opposite direction that the people running the study felt they could not ethically continue it for the entire planned period and ended it early.
An enormous increase in the value/accuracy of trend estimates derived from tree ring size (i.e. ring width) can be had by an enhanced field sampling effort, by returning to sites already sampled, and specifically sampling the small and medium sized trees that were under-sampled by the original samplers. This will go furlongs toward alleviating the bias inherent in the RCS detrending process, which arises because there is some residual bias present in the “regional curve” that is due entirely to the fact that the full age/size structure of the stand is not typically sampled, for reasons that have historical roots.
Related point there is that ring density is a different beast altogether, being both less influenced by age/size effects, and less inherently variable for a given age/size, than is ring size.
The problems of non-linear responses remain however. Improvements are possible there too however, wherein widened confidence intervals on historic climate estimates could potentially be obtained, by calibrating non-linear relationships between ring response and climatic var (sensu Loehle 2009), and using them to predict bimodal or other non-normal confidence probability functions for times past. In a weirdly ironic way, this makes use of the divergence phenomenon in a positive way, because you’re specifically looking for sites where the decadal+ trends diverge between the rings and the instrumental data. However, you have to solve problem in first P above first, or this is moot.
And some other possibilities exist as well–too complicated to go into here.
Most definitely don’t want to throw the baby out with the bathwater. Improvements are possible.
That was in response to Salamano at 2:57 btw
Re: Salamano (Apr 15 14:57),
Per Steve McI, Kenneth Fritsch, and others, the issue is for paleoclimatology to catch up with every other discipline and avoid post hoc selection. However, when the Thought Leaders are ever-more-determined to keep digging, it becomes very difficult to implement the First Law of Holes. The pleasures of groupthink versus the unappetizing prospect of a career-endangering break.
But see dendrochonologist Jim Bouldin’s remarks of April 16, 2013 at 8:53 PM below. Also check his blog.
It seems to me that the entire problem around building a paleo temperature record resides with the drive to show that current temperature changes are excessive, therefore the earth and all that lives in and on it are doomed. Without that drive it would be much easier to look at ice cores, tree rings and sediments objectively and assess their usefulness. This is how I see SM working, he has no barrow to push. I can imagine that if these data did (statistically soundly) show accelerating recent temps he would be as happy to accept that as he would the alternative.
along the same lines, has anybody seen this one ?
http://www.nature.com/ngeo/journal/vaop/ncurrent/full/ngeo1787.html
it is atm featerurd in the science section of our prime newspaper “De Standaard”
Palmer Peninsular Phenomenon.
===================
Going forward it would seem to me that the problematic science we wittness could be improved fairly quickly if the public funding for climate research was made genuinely contestable. In other words the funders of the science should not be biased towards particular outcomes, the breadth of research to be considered was not linked to pre-determined outcomes and the pool of applicants was not pre-selected by existing membership in any particular camp or political grouping.
It would also be useful if public funding was targetted at the collection of useful data that could be achived directly and immediately for free access (before any conclusions or opinions are derived from it).
Dreaming and/or p—ing into the wind?
Steve — I felt that Aslak Grinsted was very forthcoming in his comments on my 2009 post on “Svalbard’s lost Decades”. He stated at the time that they were very confident about the dating back to 1259 AD (from a volcanic eruption), and were willing to extrapolate it a few decades back before that (to 1200), but not clear to the bottom of the core (for which their best guess at the time was 1123 AD). Now it appears they have pushed that back to before 800 AD even, which is a lot farther than I would have guessed at the time.
Aslak controlled the isotope washout ratio data, and was unwilling at the time to use it before 1200 AD because of the dating uncertainty. Still, it shows warmth somewhere before 1200 AD, even if we don’t know exactly when, so I think that this should have been acknowledged at least qualitatively in the papers.
Aslak had no control over the d18O data, which was controlled by Elisabeth Isaksson, as he pointed out in his comments, so that it is unfair to blame him for its unavailability until now. Indeed, the graph you show indicates continued warming clear back to 800 AD.
Should be ionic washout ratio.
One Trackback
[…] linking CO2 rise with declining ice levels……again 6a010536b58035970c017eea5c93d4970d.png https://climateaudit.org/2013/04/13/n…t-on-svalbard/ Reply With […]