Climate Science Glossary

Term Lookup

Enter a term in the search box to find its definition.


Use the controls in the far right panel to increase or decrease the number of terms automatically displayed (or to completely turn that feature off).

Term Lookup


All IPCC definitions taken from Climate Change 2007: The Physical Science Basis. Working Group I Contribution to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, Annex I, Glossary, pp. 941-954. Cambridge University Press.

Home Arguments Software Resources Comments The Consensus Project Translations About Support

Bluesky Facebook LinkedIn Mastodon MeWe

Twitter YouTube RSS Posts RSS Comments Email Subscribe

Climate's changed before
It's the sun
It's not bad
There is no consensus
It's cooling
Models are unreliable
Temp record is unreliable
Animals and plants can adapt
It hasn't warmed since 1998
Antarctica is gaining ice
View All Arguments...

New? Register here
Forgot your password?

Latest Posts


Dear Heartland, Stop using Arthur Robinson's Trick to Hide the Incline

Posted on 18 May 2012 by Mark Boslough

Climate change is debated in letters to the editor of hometown newspapers all over the world.   In the Las Cruces, New Mexico, Sun-News, one reader recently cited "a 1996 paper by Kiegwin (sic) in Science which showed that, despite the present having a CO2 concentration of 388 PPM, the present temperature is cooler than the average of the last 3,000 years, and that it was considerably warmer than today during the Medieval Warm Period, the Roman Warm Period, and the Holocene climate Optimum.”    A few months later another reader asserted that “Keigwin, Science, 1996, shows present temperatures aren’t much different from the 3,000 year mean.”

Did the Keigwin paper really say that?  And how is it that two non-scientists from a mid-sized New Mexico city would be so confident that a scientific paper published a decade-and-a-half earlier supports their belief that the world was warmer during Medieval times? 

First, let’s review Keigwin (1996).   The title “The Little Ice Age and Medieval Warm Period in the Sargasso Sea” might provide the first clue that it isn’t about global temperatures, but about one location on Earth:  the Sargasso Sea.  What Keigwin did was to use oxygen isotope measurements in plankton skeletons from sediment cores as a proxy to reconstruct the sea surface temperature (SST) of the past 3000 years.   The cores were collected in 1990, and were divided into 50- to 100-year segments.  In the absence of mixing or bioturbation from below, the mid-point of the most recent 50-year thick sample, whose value would represent the most recent paleotemperature, would be 1965.  In a perfect world, the bottom of the segment would date to 1940.  However, sediments in the real world are never completely undisturbed.  It is very likely that the most recent segment contained shells from the early 1900s or even from the previous century.  That means the last paleotemperature is really an average that probably includes values from before automobiles and light bulbs were invented.

Keigwin published a graph, as Figure 4b (K4B), of his best estimate of the resulting time series. 

He also included several years of modern instrumental measurements at hydrographic station “S” in Bermuda, starting in 1954.  The modern year-to-year temperatures fluctuate significantly, but the mean is well above the 23°C average of the entire proxy record.

It is unlikely that the Las Cruces letter-writers ever read this paper, or they would have known it wasn’t about global temperatures.  It is highly cited in the contrarian literature as evidence against human-caused global warming, and turns up in many blogs and editorials without reference to the Sargasso Sea.  How did this happen?

The misuse appears to have started in the late 1990s, when Arthur P. Robinson of the Oregon Institute for Science and Medicine (OISM) started the so-called “Oregon Petition” to collect signatures of people opposed to the Kyoto Protocol.  With his son Zachary and two associates from the conservative George C. Marshall pressure group (Sallie Baliunas and Willie Soon), he self-published a paper called “Environmental Effects of Increased Atmospheric Carbon Dioxide” designed to look like a peer-reviewed article from the Proceedings of the National Academy of Sciences (US).  It was mailed out with the petition to many thousands of engineers, dentists, veterinarians, and even some scientists.  In January, 1998 it appeared in a periodical published by the Association of American Physicians and Surgeons (AAPS), a political advocacy organization with a stated mission to “fight socialized medicine and to fight the government takeover of medicine.”  The executive director of AAPS is also member of Robinson’s OISM.    In it was their Figure 2, a modified version of Keigwin’s K4B.  

Robinson and coauthors made several changes in representation and labeling.  First they inverted the axes so time runs from left to right, but they were unaware that when paleoclimate data are plotted “years before present”  means “years before 1950” so their data is shifted by about 50 years.   Second, they removed the data from hydrographic station “S” which showed that recent temperatures are above the long-term average.  Third, they neglected to label it as being a record for the Sargasso Sea.  Fourth, they called it a global temperature in the text, saying, “For the past 300 years, global temperatures have been gradually recovering. As shown in figure 2, they are still a little below the average for the past 3,000 years.”

This paper became the basis for statements in two influential Wall Street Journal opinion pieces.  The first, in 1997, was by Robinson and his son Zachary, called “Science Has Spoken:  Global Warming Is a Myth.”  They stated,

During the past 3,000 years, there have been five extended periods when it was distinctly warmer than today. One of the two coldest periods, known as the Little Ice Age, occurred 300 years ago. Atmospheric temperatures have been rising from that low for the past 300 years, but remain below the 3,000-year average.

The second editorial, with son Noah, was called “Global Warming is 300-Year-Old News.”  In that one, they stated that “Earth temperatures are now near the 3,000-year average and clearly not unusual.”  Their Oregon Petition figure was re-drafted with different temperature units, but the time scale was still wrong, and the current thermometer measurements were still missing.  Despite the misrepresentation in the text as “Earth temperatures” the graph this time was labeled “Temperature of the Sargasso Sea from 1000 BC to 1975 AD.”  The source of the year 1975 as the endpoint is unclear and did not come from Keigwin’s paper.

This Wall Street Journal version of the graph appears to have become the new “primary source” for those who argue that temperatures are actually lower now than they were in the past.  Award winning editorial cartoonist John Trever redrew it for an ironic dig at climate scientists who claim otherwise (including—with true irony—Lloyd Keigwin, the original author of the figure).

In 2004, I was asked by my management to review the original sources of these claims, and I wrote several messages to Arthur Robinson asking for some clarifications.   Because I was planning to write a report, I wanted to give him the courtesy of responding with any clarifications or corrections.  Among other questions, I included the following,

I'm wondering what is your basis for the statement, "During the past 3,000 years, there have been five extended periods when it was distinctly warmer than today." (Robinson & Robinson, The Wall Street Journal (Dec. 4, 1997).  I've seen this quoted by others (often without attribution) but it looks like you were the first to say it.

Robinson responded,

I note that the Sargaso (sic) Sea curve shows five earlier periods where the temperature was above the mean and therefore warmer than today. This is probably the source of the statement.

Significantly, business advocate Raymond Keating—in testimony to House Small Business Committee (June 4, 1998)—said, “During the past 3,000 years, there have been five extended periods when it was distinctly warmer than today.”

I proceeded to ask Robinson some more difficult questions,

I can see that your Figure 2 was taken directly from the 1996 Keigwin paper, but with the post-1954 instrumental "Station S" SST data removed.  Was there a reason you took the directly measured temperature off?  What method did you use to calculate the 23 C mean?  Did you derive it from the original Keigwin data or was that simply an estimate to the nearest degree?

You incorrectly represented the graph as global temperature.  You stated, "For the past 300 years, global temperatures have been gradually recovering (11). As shown in figure 2, they are still a little below the average for the past 3,000 years."

I plan to include these observations in my final writeup.  If you care to respond, I would be happy to include your comments.

To which Robinson replied,

Regarding the world data. We clearly labeled this data location. Since virtually all other available dats (sic) from other locations (see Soon and Baliunas) is similar, providing this example was entirely ethical.

It is too bad your employers could not find an objective scientist for this task. I will not be providing any additional comments, since I am quite sure they would not be presented in their enirety (sic) to your employers, any more than will those I have already written. You are clearly devoted to lifting selected things from their context.

Do not waste your time with additional email. It will be shunted to the unopened file here.


For the record, I provided all his responses to my management in their entirety.

The most recent chapter of this story began when the periodical of the AAPS re-published an edited and colorized version of the paper in 2007 under a different author rotation (Baliunas was removed, and son Zachary was replaced by son Noah).  Perhaps because of my 2004 criticisms, an instrumental data point was added for year 2006, and the mean temperature was shifted.

The paper explained the source of the 2006 temperature thusly, “A value of 0.25 °C, which is the change in Sargasso Sea temperature between 1975 and 2006, has been added to the 1975 data in order to provide a 2006 temperature value.” 

Unable to reproduce this temperature with the data I had, I wrote to Willie Soon in 2010 and asked for the source of the data.  He cordially responded and sent me the table, telling me,

…also about the most recent point at "2006"---sorry that I could not be more certain, but I am sure Noah has carefully included this updated SST series from station S that Dr. Keigwin sent me around July of 2007 (which as you can see from the file name was obtained by Dr. Ruth Curry of WHOI).

Graphing the Station S data with their data point for 2006 (blue) demonstrates that the 2006 is about a degree too low in Robinson et al. (2007).

If they had plotted the data they had, the way they said they did, it would have looked like this:

On Nov. 2, 2010, I presented this to a large audience at the Geological Society of America (GSA) meeting in Denver.  Always wanting to give others the benefit of the doubt, I wrote to Noah Robinson several times at his OISM address.  On Oct. 23, 2010 I wrote:

Willie & Noah,

Attached is a draft of a couple slides I plan to present, which strongly suggest that your team fabricated the 2006 data point to hide the increase in Sargasso Sea surface temperature.

You plotted your 2006 point too low by more than a degree C.  If this was an honest arithmetic mistake or silly drafting error, now would be the time to explain it and correct it.  If you let me know before my presentation, I will be happy to include your explanation.

Best regards,

Mark Boslough

I did not get a response.

The Heartland Institute, a fossil-fuel-funded political pressure group, reprinted a distorted version the latest OISM the graph in their advocacy report, “Nature, Not Human Activity, Rules the Climate” (S. Fred Singer, editor).  This was published for an organization called the Nongovernmental International Panel on Climate Change (NIPCC), and was sent to American members of Congress and other policy makers. 

In 2011, I submitted an abstract, coauthored by Lloyd Keigwin, to the Third Santa Fe Conference on Global and Regional Climate Change, so that we could present these findings.  According to the conference summary, it was “to focus on climate change and variability from observational and modeling perspectives.”  The chair of the conference, Petr Chylek is affiliated with one of the sponsors (Los Alamos National Laboratory) as well as the Heartland Institute (as a “Heartland Expert”).  With many speakers affiliated with the Heartland Institute (it turned out to be at least nine) it seemed like a good opportunity to provide feedback.  Unfortunately, Keigwin and I received a rejection letter from Chylek, who told us,

This Conference is not a suitable forum for type of presentations described in submitted abstract. We would accept a paper that spoke to the science, the measurements, the interpretation, but not simply an attempted refutation of someone else's assertions (especially when made in unpublished reports and blog site).

This was a puzzling rejection given that the “unpublished report” was the NIPCC document released by the Heartland Institute, which is widely cited by the many Heartland-affiliated speakers invited by Chylek.

Nevertheless, I was able to have a conversation with Fred Singer, the editor of the NIPCC report, about the Sargasso Sea graph.  At first, he told me that I should take it up with Robinson, but ultimately he assured me that it would be corrected in the next edition of the NIPCC.  He also revealed that he had been the one who had sent Keigwin’s paper to Robinson in the first place, back in the ‘90s.

In a final attempt to get my feedback directly to the Heartland Institute before writing this guest post, I offered in April, 2012 to give a presentation at their annual meeting in Chicago in May.   My offer was rejected by Heartland’s Director of Communications, Jim Lakely, who told me in no uncertain terms, “you will not be getting an invitation to speak.”

I look forward to their response to this article.

0 0

Printable Version  |  Link to this page


Comments 1 to 18:

  1. It isn't the first this specific form of distortion has been documented here. I recall essentially identical treatment of greenland temperature (last year?), right down to the 1-2 punch combo of 'make-local-global'/'make-past-present', in an attempt to floor that most pernicious of opponents, accuracy. I'm very grateful to the continued dissection of these types of presentations. Particuarly in regards to the methodology of such paleotemperature studies. It's pretty incredible that modern thermometer data is sieved so finely, month by month, when 50 year average data slabs containing noise from Churchill's childhood is accepted verbatim...
    0 0
  2. Would it be churlish of me to point out there is a self anointed pseudo auditor who has taken it upon themself to examine all the paleo misrepresentations, yet seems to have missed this one?
    0 0
  3. A fabulous look at the AstroTurf manufacturing process.
    0 0
  4. A very interesting and convincing history of the evolution of a false talking point, and the "telephone game" of increasing distortion that comes along. I am extremely impressed with the diligence and rigor with which you have sorted this out. One possible grammatical nit: I might be mistaken, but would a comma after Heartland in the title be the more correct way to address a statement to them? When I first read the title I thought of it like a newspaper article where the colon would indicate Heartland was saying the statement that followed (though obviously not a direct quote). Lastly, as a part time tone troll, I would submit that it is actually pretty rare for people to intentionally fabricate data (though this does occur), despite what the WUWT crowd might say. I'm also pretty sure the email quoted above to Willie and Noah asserting fabrication wouldn't have been the best way to get an answer as to why the 2006 point was plotted where it was, even if it had been an "honest mistake." I would wager that there is an alternative, confirmation-biasy, or slipshod-type explanation for the 2006 misplot, and that this might have been discovered with a "softer approach". No less horrible for our future, no less important to uncover, but possibly a bit less evil... Now don't get me wrong, an honest broker would print a retraction about the error once pointed out, and of course not use it again. Since it doesn't seem like such a retraction is forthcoming I would not consider Heartland or NIPCC honest brokers. I also cannot overstate how impressed I am at your heroic efforts to nip a distortion in the bud. I just wish we knew the definite real-life-human answer to why that 2006 point was a degree off, hence my tone trolling, for which I fully apologize. I also fully admit there could have been someone such as Noah who sat down and said "ooh, that 2006 point doesn't look good there, let me just move it down." But I bet he did something more subtle, human, and interesting than that, and figuring that out to me, holds greater hope for getting us out of the mess we are in...
    0 0
  5. For the non-scientist (and I count myself as one), this article is an excellent case study. You cover both the inner workings of the scientific review process (asking for clarifications) and the process of commenting on others work (I'm going to speak on your findings and thought I would give you a chance to comment). What's important for a site like this is we can see the attitude that the denier community plays when asked about the science: "It is too bad your employers could not find an objective scientist for this task. ... Do not waste your time with additional email. It will be shunted to the unopened file here." That, plus the fact that you're not welcome at their conference, speaks loudly to the anti-science slant so many legitimate scientists face when the offer serious, legitimate criticisms of scientific work that isn't intended to further scientific discovery. A legitimate scientist whould say why the question was off base, say why. These guys don;t have a leg to stand on.
    0 0
  6. #1 70rn , this does appear to be a common way for 'skeptics' to analyse palaeoclimate stuff. I covered a different record here a while ago. I mentioned 3 'tricks' used to 'hide the incline' which are changing the temperature scale, ignoring/hiding/deleting recent measurements and picking one region and implying (or even just saying) that it's a global record. That's what a number of 'skeptics' did with Hubert Lamb's paper tracking UK temperatures up to the 1930s, and with the Sargasso Sea reconstruction it sems that they've added 'including made-up data' to their repertoire. Although I agree with Utahn, people make mistakes and confirmation bias is a more likely reason than just fabricating the data.
    0 0
  7. Robinson and co. is not the only miscreant in this farce. We saw Soon and Balliunas had a hand in it here.
    0 0
  8. It looks very likely to me that they plotted the 2004 data point and mislabelled it 2006, possibly by accident. This would make it an egregious example of cherry-picking rather than fabrication. If they were willing to simply fabricate the data, why would they have held themselves to a single point?
    0 0
  9. I will definitely bookmark this post for future reference as an example of what constitutes denialism versus true skepticism. I hope that your work here will find it's way into the curriculum of some "methodology of science" type courses. You have provided an excellent and well documented example of how science should not be done. Or should we just call it anti-science, or in a nod to Seinfeld - science in the bizarro world?
    0 0
  10. > >Regarding the world data. We clearly labeled this data location. Since virtually all other available dats (sic) from other locations (see Soon and Baliunas) is similar, providing this example was entirely ethical. Has someone followed up on this? [I haven't.] I wonder if doing the average on these studies would produce a mean similar to the pictured one. I think many MWP proxies do not line up temp with years, leading to wide error bars and lukewarm means. If that is the case here, then taking this one sample might not be "ethical" or accurate as concerns the implied depicted suggestion that the temp has clearly been higher in the recent past.
    0 0
  11. Very cool investigation, Mark. I love this kind of in depth stories about climate myths. I commend you for your honesty in contacting the people involved and reproducing their comments (albeit marginally helpful in this case). I'm always amazed with people's obsession in using local records to imply global temperature and their simultaneous reluctance to use *actual* global averages.
    0 0
  12. Yup, another great informative article. well... and yes... I couldn't pass up bootlegging a copy onto whatsupwiththatwatts.blogspot That "embed code" is the coolest. Also I started a thread over at SkepticSocietyForum that might interest some of you. ~ ~ ~ ~ ~ ~ ~ " compared to" _________________________________________________ Post #1 Postby citizenschallenge » Wed May 23, 2012 1:05 am "So you have problems with my sources and I have problems with your sources" ~ ~ ~ "I know some claim is no different from "I have been told: "it's just a different perspective, you choose to believe and I choose to believe WUWT.” But, is it as simple as that? How do we decide on the respective veracity of each? . . . " ~ ~ ~ ~ ~ ~ ~ {and so on and so forth...} Keep up the good work :-) cheers, peter
    0 0
    Moderator Response: TC: voracity --> veracity. Let me know if I've misunderstood, for I am pretty voracious ;)
  13. Thanks for the heads up, I love writing. . . hate my misspellings and typos. Incidentally, I polished up on the whole of it this morning and made it a stand alone post. compared to
    0 0
  14. Good work. I appreciate your documentation of efforts by AGW Denialists to silence the expression of skeptical viewpoints. (They commonly, and with little actual justification, accuse the mainstream science community of the same offense.) The key issue here is that the present day Sargasso Sea SSTs are significantly higher than indicated in the diagram by Robinson et al. (2007). This is fairly clear, but I'm still uncertain regarding a few minor details. The following graph is an overlay of parts of three of your figures. It's a bit of a mess, but I hope it will serve. The base graph is the 6th one (above). Your 1st diagram includes SST data from Station "S". These appear on my overlay as the squiggly line centered around 23°C, at approximately 1996. It is my understanding that this represents data from 1954 up to the time the paper was published in Keigwin (1996). These same (??) data appear as the black dots on the base diagram, beginning in 1954, extended up to the present. The green line presumably represents a linear best-fit (although this is not specified). Can you explain the apparent discrepancy between these two representations of the data? Your 7th diagram indicates a "2006 temperature based on stated method", slightly higher than 24°C, but I didn't see where you state the method. This point differs from the "best fit" value indicated on the base diagram, although it does seem to fit the surrounding data from the base diagram. Finally, you refer to the Heartland Institute as "a fossil-fuel-funded political pressure group". The recently leaked documents from Heartland indicating their funding sources did not appear to me to support this comment. Can you indicate what you base it on? Thanks.
    0 0
  15. Your comments on Robinson's temperature curve were very helpful as I have been asked to critique his paper by some global warming agnostic friends. Equally puzzling, however, is his plot of total solar irradiance (Figure 3). Have you analyzed the origin of that? The 1980 to 2000 portion looks very different than the PMOD data shown elsewhere on the Skeptical Science website.
    0 0
  16. Sorry for my slow response. To CoalGeologist: Very good observation. The station "S" data that Lloyd Keigwin plotted in his 1996 paper was averaged in a different way than the data that Willie Soon sent me (which is what the Robinsons had). The hydrographic data were obtained every two weeks, but there are a lot of dropouts, especially in the '50s (and I think there was a funding lapse in the '70s with a long dropout). In one case each measurement anomaly was given equal weight, which is not really a proper way to do a time average. In another case the anomaly was calculated for each calendar month, and then the months were averaged. Still not a very good way to do it. But that's the data that was given to the Robinsons. I obtained the raw data (individual measurements) and did a time-weighted anomaly average just to see what it would look like. That reduces the scatter a lot, but there is still a similar increasing trend. It is worth noting that using calendar years is arbitrary. If we used climatological year, or some season-based year, we'd get different annual anomalies but the long-term trend would be the same. With regard to the 7th diagram, the "stated method" was “A value of 0.25 °C, which is the change in Sargasso Sea temperature between 1975 and 2006, has been added to the 1975 data in order to provide a 2006 temperature value.” But if you look at the 1975 temperature (the value that the Robinson's had) you can see that this is not what they actually plotted. To oldfueler: I didn't bother to analyze the solar irradiance graph. As my friend David Morrison once said, “Pseudoscience is like spoiled food; you don't have to eat it all to know something is badly wrong. Just a few bites will do."
    0 0
  17. Mark Boslough's forensic paper is very interesting, as are the charts he discusses, which include in their timescales the Medieval Warm Period and the Roman Warm Period. These periods are often noted in discussions about today's land temperatures in particular. We do have historical accounts, but what data do we currently have that can provide more information about the temperatures of those times for various locations? How can we quantify those historical accounts?
    0 0
  18. Past temperatures have to be derived from proxy data of various kinds. Historical (pre-thermometer) records are of much use for evaluating the significance of these proxies. ie is the temperature record in proxy consistent with historical accounts or is it too localised to be of use). See for example Ljungquist 2010 for proxies.
    0 0

You need to be logged in to post a comment. Login via the left margin or if you're new, register here.

The Consensus Project Website


(free to republish)

© Copyright 2024 John Cook
Home | Translations | About Us | Privacy | Contact Us