At Lewandowsky’s blog here, conspiracy theorist John Mashey has challenged Deming’s “get rid of the MWP” recollection claiming that:
Nobody in 1995 would have been worried about “getting rid of the MWP,” although they might have wanted to dispel the idea the Lamb schematic was Truth. They had years before abandoned the schematic… Anyone citing this schematic as credible any time after 1992 might be likened to a flat-earther clinging to Anaximander.
Although the climate community now pretends that the IPCC 1990 graphic never existed or was a one-off, variants of this graphic continued in use by leading climatologists through 1996, because, as Crowley complained in a Climategate email, “there was no alternative”. This doesn’t prove that Deming’s recollection is valid, only that Mashey’s argument that “nobody in 1995 would have been worried about “getting rid of the MWP” is without support.
In today’s post, I’ll review two of the most relevant contemporary publications by the UC Global Research Information Office – a 1991 article by Bradley and Jack Eddy, an older contemporary, and a 1996 article by Tom Crowley, both using variants of the IPCC 1990 graphic. I’ll also review interesting 2007 Climategate correspondence in which Bradley and Crowley reflect back on these articles.
Bradley and Eddy 1991
In 1991, Eddy and Bradley published “Changes in time of the temperature of the earth” pdf(Earthquest 1991, Vol 5 No 1), a publication of the US Global Climate Research Information Office. Although Bradley, regarded as notoriously vain in the Climategate emails, has a comprehensive bibliography of his publications, this 1991 publication isn’t listed.
Eddy and Bradley 1991 presented six figures representing global temperature over scales ranging from 140 years to 180 million years. One of their figures was a slight modification of the IPCC 1990 Figure 7.1 bottom panel (the “Lamb” graphic); two other figures either matched or were only slight modifications of the top two IPCC 1990 panels.
Eddy and Bradley provided the following overall description of these reconstructions:
“reconstructions of the longer history of past variations of surface temperature are obtained through the chemical analyses of dated sediments of various kinds are as such are themselves regionally or even site specific.. A compilation representative of what is known today is shown in this set of graphs, which portray a sampling of estimated surface temperatures through the last 180 million years of earth history”.
Their most birds-eye graphic (see below) showed 180 million years (attributed to L.A> Frakes, Climates through Geologic Time, 1979). This showed “a global cooling trend since the time of the Cretaceous, when global surface temperatures were 8-10 deg C warmer than today” and shows that even the interglacials of the Pleistocene are cool in geological time.
Eddy and Bradley, 1991 Figure 1.
Their second panel (see below) showed temperatures for the past 850,000 years and closely matches the IPCC 1990 Figure 7.1 top panel (also shown below for comparison). Bradley and Eddy attribute it to an equatorial Pacific O18 proxy series from Shackleton and Opdyke (Quat Res, 1973), referring also to the iconographic version of this series in W.C. Clark, Carbon Dioxide review, 1982.
|
|
Top- Eddy and Bradley, 1991 Figure 2; bottom – IPCC 1990 Figure 7.1 top panel
Their third panel (not shown) shows a 150,000 year series, derived from Vostok dD isotopes (Jouzel et al 1987), a series shown in other contemporary multiscale panels. Their fourth panel showed 18,000 years and is said to be a modification of the IPCC 1990 Figure 7.1 middle panel. They say that this graphic is ‘estimated from a variety of sources, principally isotope ratios from Greenland ice cores”, but also mention the use of “European lakes”, stating that, during the Holocene maximum “summer temperatures may have been 1-2 deg C warmer than the present era”. Again, the corresponding IPCC 1990 panel is shown for comparison.
|
|
Top – Eddy and Bradley, 1991 Figure 4; bottom- IPCC 1990 Figure 7.1 middle panel.
Their fifth panel was a 1000-year graphic, clearly in the Lamb lineage. They describe their graphic as “modified from J.T. Houghton et al, 1990”. As above, the corresponding IPCC graphic is shown. [Update note: as Tom Curtis observed below, Bradley bodged (to borrow the Briffa phrase) the Lamb graphic by reducing the MWP estimate by about 0.25 deg C; Curtis asserted that the dashed lines in the Bradley graphic “clearly” show the bodge. In my opinion, the existence of the Bradley bodge would have been more evident if it had been mentioned in the caption or running text.] Bradley also modified the schematic in the period after 1600, varying the squiggles somewhat.
|
|
top – Eddy and Bradley, 1991 Figure 4; bottom – IPCC 1990 Figure 7.1 bottom panel.
They described its construction as follows:
Variations in surface air temperature estimated from a variety of sources including temperature-sensitive tree growth, indices and written records and accounts of various kinds, largely from western Europe and eastern North America. Of note is a possible protracted global warming through the Medieval period, when surface temperatures may have averaged about 0.3 deg C warmer than the AD1900 reference. It was followed by a longer period of much colder conditions, loosely termed the Little Ice Age, when the estimated global mean temperature may have fallen about 0.5 deg C below the reference norm, reflecting global temperatures almost 1 deg C lower than the values attained during the middle of the current century, modified from J.T. Houghton et al, 1990.
Their final panel (not shown) was the CRU temperature history (then covering about 140 years).
Bradley’s Confusion
Given the recent blog criticism of my incorrect attribution of the Lamb graphic in a 2005 blog post and contemporary presentation (ignoring that I had correctly attributed the graphic by June 2005 and subsequently), it is ironic to read Raymond Bradley’s thorough confusion on this point in Climategate emails.
Phil Jones had been trying to figure out the provenance of the IPCC 1990 Figure 7.1 bottom panel. In very early January 2007, Bradley (CG1 – 760. 1167752455.txt) wrote to Jones that Bradley himself had originated the Lamb-style graphic and that Chris Folland (“Dork” Folland in Bradley’s contemporary terminology in CG2-4621) had used it in IPCC 1995 (elsewhere complaining at the time that Dork Folland had been insufficiently deferential to Bradley’s pearls of wisdom):
I believe this graph originated in a (literally) grey piece of literature that Jack Eddy used to publish called “Earth Quest”. It was designed for, and distributed to, high school teachers. In one issue, he had a fold-out that showed different timelines, Cenozoic, Quaternary, last 100ka, Holocene, last millennium, last century etc.
The idea was to give non-specialists a perspective on the earth’s climate history. I think this idea evolved from the old NRC publication edited by L. Gates, then further elaborated on by Tom Webb [see Webb 1991 here] in the book I edited for UCAR, Global Changes of the Past. (This was an outcome of the wonderful Snowmass meeting Jack master-minded around 1990).
I may have inadvertently had a hand in this millennium graph! I recall getting a fax from Jack with a hand-drawn graph, that he asked me to review. Where he got his version from, I don’t know. I think I scribbled out part of the line and amended it in some way, but have no recollection of exactly what I did to it. And whether he edited it further, I don’t know. But as it was purely schematic (& appears to go through ~1950) perhaps it’s not so bad….
They also seem to have fabricated a scale for the purported temperature changes. In any case, the graph has no objective basis whatsoever; it is purely a “visual guess” at what happened, like something we might sketch on a napkin at a party for some overly persistent inquisitor….. (so make sure you don’t leave such things on the
table…).What made the last millennium graph famous (notorious!) was that Chris Folland must have seen it and reproduced it in the 1995 IPCC chapter he was editing. I don’t think he gave a citation and it thus appeared to have the imprimatur of the IPCC. Having submitted a great deal of text for that chapter, I remember being really pissed off that Chris essentially ignored all the input, and wrote his own version of the paleoclimate record in that volume.
There are other examples of how Jack Eddy’s grey literature publication was misused. In a paper in Science by Zielinski et al. (1994) [v.264, p.448-452]–attached– they reproduced [in Figure 1c] a similarly schematic version of Holocene temperatures giving the following citation, “Taken from J. A. Eddy and R. S. Bradley, Earth-quest 5 (insert) (1991), as modified from J. T. Houghton, G. J. Jenkins, J. J. Ephraums, Climate Change, The IPCC Scientific Assessment (Cambridge Univ. Press, Cambridge, 1990).”
But I had nothing to do with that one!
So, that’s how a crude fax from Jack Eddy became the definitive IPCC record on the last millennium!
Crowley quickly disputed Bradley’s recollection, pointing out that he did not believe that Eddy (and Bradley) had ‘”cooked up” that figure … purely out of thin air’. Crowley reported that he had used a version of the Lamb graphic in Crowley and Lowery 2000. The Climategaters quickly recognized that the graphic originated with Lamb, rather than Bradley, though the Climategate correspondence shown thus far did not trace it as far back as the Lamb 1965 later identified at Climate Audit.
Crowley 1996
At Lewandowsky’s blog, conspiracy theorist John Mashey asserted that use of the IPCC-Bradley graphic “after 1991” would only be done by someone adhering to the “Aniximander” plan:
Anyone citing this schematic as credible any time after 1992 might be likened to a flat-earther clinging to Anaximander. Think of the schematic as a flat-earth temperature sketch, which by 2005 got retroactively elevated (by a few people) and promoted as Absolute and Unchanging Truth being hidden by a cabal of climate scientists, just as the way NASA is hiding the absolute truths known by the Flat Earth Society. There we find The Conspiracy:, the active faking of the whole space program, not just Apollo! It has a good FAQ. They show a newer map than Anaximander, still flat.
Mashey didn’t explain exactly what happened in 1992 to exclude Eddy and Bradley’s 1991 publication from membership in the Aniximander plot. The IPCC 1992 report is almost totally silent on the MWP: its only mention of the topic was a passim comment about Briffa’s Tornetrask reconstruction as follows:
in this region [Scandinavia] for a “Medieval Climatic Optimum” (S7, p 202) around 870-1110, another warm period around 1360-1570, and a “Littie Ice Age” (S7, p202) period around 1570-1750.
The Briffa reconstruction cited by IPCC 1992 was, of course, his Tornetrask reconstruction, which Briffa had “bodged” to reduce the MWP – a topic later of interest at CA. (This data recurs in the recent Esper et al 2012 article, discussion of which is long overdue here.) But I am not aware of anything in 1992 (either in the IPCC report or elsewhere) evidencing that climate scientists had resiled from the Eddy and Bradley 1991 summary.
Bradley and Jones (Holocene 1993) was a sort of opening shot by the Hockey Team. It did not discuss the MWP, but did question the concept of the Little Ice Age, citing the inconsistency of a number of NH proxies. (From a 2012 AR5 perspective, it seems to me that the concept of a global Little Ice Age is now relatively widely accepted in the paleoclimate community and, to that extent, Bradley and Jones 1993 took an incorrect fork in the road, though I recognize that this interpretation might not be widely shared within the climate “community”.)
Hughes and Diaz (1994) extended the Bradley and Jones 1993 project to the MWP. It was cited by AR4 as authority against the MWP concept. However, Hughes and Diaz 1994 primarily relied on tree ring chronologies that had been homogenized in a way that removed all centennial variability. Although it has been widely cited by climate scientists, its flawed methodology meant that its conclusions did not follow from its data – a point that ought to be evident even to climate scientists. In the review of AR4, both Jan Esper and myself, using identical arguments and in unusual agreement, argued against AR4 relying on Hughes and Diaz 1994, but our criticism was ignored. Nonetheless, the criticisms are valid. BTW draft versions of AR5 admit the concept of a “Medieval Climate Anomaly” and are far more accepting of the idea of a global reorganization of climate than AR2 and AR3, but that’s a different story.
While Hughes and Diaz criticized the concept of a MWP, through 1995, neither they nor anyone else had presented an alternative to the Lamb graphic and variants of the Lamb graphic continued to be used in academic literature. As Bradley noted in his 2007 Climategate email, the Eddy and Bradley 1991 version was used in Zielinski et al (Science 1994.)
[Update: Oct 10] Jean S observes in comments below that a version of the Lamb graphic was used in Mayewki et al (JGR) 1993, a presentation of GISP2 ice core results. The bottom two panels of Mayewski et al 1993 are shown below.
[Update: Oct 10] Bradley’s recollection of Zielinski et al 1994 was only partly correct. Zielinski et al re-drew from IPCC 1990 bottom and middle panels as shown below:
Zielinski et al 1994 Figure 1. Caption stated: Composite of the postulated variation in Northern Hemisphere surface temperatures from the present to 7000 B.C. (35) Temperature change was determined by a deviation from the mean value at the turn of the last century. (35 – Taken from J. A. Eddy and R. S. Bradley, Earthquest 5 (insert) (1991), as modified from J. Houghton, G. J. Jenkins, J. J. Ephraums, The IPCC Scientific Assessment (Cambridge Univ. Press, Cambridge, 1990).
Here is a parsing of the Zielinski bottom panel, showing the scale as a grid over the figure, together with the MWP and Holocene Optimum maxima.
Update- Oct 11.] Another Jean S find: Hass and Kamp 1995 here and Hass 1996 here.
Additional adherents to the Aniximander plan revealed by these articles are: G. A. Zielinski, P. A. Mayewski, L. D. Meeker, S. Whitlow, M. S. Twickler, M. Morrison, D. A. Meese, A. J. Gow, R. B. Alley, K. K. Ferland, M. R. Legrand, and J. P. Steffensen, twelve of the most prominent ice core specialists of the time, and Christian Hass and Michael Kass.
Crowley 1996
The last authors meeting for the IPCC Second Assessment Report was held in Madrid in Novermber 1995, but the report itself would not be published until June 1996.
In winter 1996, during this hiatus, Tom Crowley, a very prominent climate scientist, re-visited the multi-scale temperature history (covered in 1991 by Eddy and Bradley) for Consequences, a peer-reviewed publication by the US Global Climate Change Research Office. Crowley’s 1996 article had three named peer reviewers: including Anthony Broccoli, the present editor of Journal of Climate; William Ruddiman, another prominent and prolific climate scientist; and Lisa Sloan.
Like Eddy and Bradley 1991, Crowley 1996 presented temperature histories over scales varying from the instrumental period to geological time. It drew directly on the Eddy and Bradley 1991 graphics for the three panels in the IPCC 1990 report.
Crowley’s longest-scale graphic (his Figure 6 – not shown here – derived from Crowley et al 1990) covered 120 million years and was even more schematic than the corresponding graphic in Bradley and Eddy.
Crowley’s next scale (his Figure 5 shown below) covered the 850,000 years of the Pleistocene. Crowley attributed this graphic to Imbrie et al 1984, a book section in Milankovitch and Climate (1984), edited by Berger and others. However, the data itself appears unmistakeably to be the same series as in the corresponding graphic in Eddy and Bradley 1991 (see above) and thus ultimately derived from the same data as the top panel of the IPCC 1990 graphic.
Crowley 1996 Figure 5
Crowley’s next scale (his Figure 3) was 125,000 years; it was attributed to Bradley and Eddy 1991, based on Jouzel et al 1987 (Vostok dD).
Crowley’s next scale (his Figure 2) was the 15,000 year period covered by the IPCC 1990 middle panel. Again, Crowley attributed his figure to Eddy and Bradley 1991.
Crowley 1996 Figure 2
Crowley’s final scale (his Figure 1) was the 1000-year period of the IPCC 1990 bottom panel. Crowley’s version was a re-drawing of the corresponding Eddy and Bradley 1991 figure (which Crowley acknowledged). Both figures are shown below for comparison.
Top – Crowley 1996 Figure 1; bottom – corresponding figure from Eddy and Bradley 1991
Crowley’s caption was as follows (drawing on the corresponding caption in Eddy and Bradley 1991):
Figure 1 Example of regional variations in surface air temperature for the last 1000 years, estimated from a variety of sources, including temperature-sensitive tree growth indices and written records of various kinds, largely from western Europe and eastern North America. Shown are changes in regional temperature in° C, from the baseline value for 1900. Compiled by R. S. Bradley and J. A. Eddy based on J. T. Houghton et al., Climate Change: The IPCC Assessment, Cambridge UniversityPress, Cambridge, 1990 and published in EarthQuest, vol 5, no 1, 1991.
Thus, if Mashey were correct in his allegation that users of the Lamb graphic after 1992 were adherents to the Aniximander plan, Crowley, his reviewers (Ruddiman, Broccoli and Sloan) and his editor (Eddy) would have to be counted among the adherents to the Aniximander plan.
In June 1996, a few months after Crowley 1996, the IPCC Second Assessment Report was finally published. Its Chapter 3 combined assessment of recent instrumental data and assessment of paleoclimate data. Coordinating lead authors include names familiar from Climategate: Neville Nicholls, Tom Karl, David Parker; Phil Jones was listed as a “Key Contributor”. It cited both Bradley and Jones 1993 and Hughes and Diaz 1994 and took the very disbelieving views of those articles, as, for example, the following:
it is not yet possible to say whether, on a hemispheric scale, temperatures declined from the 11-12th to the 16-17th century. Nor, therefore, is it possible to conclude that global temperatures in the Medieval Warm Period were comparable to the warm decades of the late 20th century.
Crowley’s Recollection
Crowley’s reflections in 2007 on the prior events are instructive. He explained in 2007 (CG1-760) that he used this graphic in 1996, because “even though I didn’t like it – there was not an alternative”:
I might also point out that in a 1996 Consequences article I wrote – and that Fred Singer loves to cite — Jack [Eddy] (who was the editor of the journal) basically shoehorned me into re-reproducing that figure even though I didn’t like it – there was not an alternative. in the figure caption it has a similar one to Zielinski except that it states “compiled by R.S. Bradley and J.A. Eddy based on J.T. Houghton….so that puts a further twist on this because it point to Houghton not Bradley/Eddy as the source. Jack must have written that part of the figure caption because I don’t think I knew those details.
Deming in Context
In March 2005, shortly after the publication of our GRL article criticizing the MBH hockey stick, David Deming recounted the following story containing the now notorious phrase “get rid of the MWP”:
With the publication of the article in Science [in 1995], I gained significant credibility in the community of scientists working on climate change. They thought I was one of them, someone who would pervert science in the service of social and political causes. So one of them let his guard down. A major person working in the area of climate change and global warming sent me an astonishing email that said – We have to get rid of the Medieval Warm Period.
I first learned of Deming’s recollection from an article by Deming posted at the SEPP website here. Deming later published a version of this article in a lower-tier journal (Journal of Scientific Exploration); this version was subsequently cited by Huang et al in a 2008 GRL article here. In 2006, Deming gave similar testimony to a Senate Committee (see here.) I first referred to Deming’s story in a March 2005 blog post. Ross and I also referred to it in contemporary presentations in which we sought to explain the remarkable prominence of the MBH hockey stick in the IPCC (2001) report, where it appears on no fewer than 6 occasions.
Conclusion
In previous reflections on this topic (most recently in 2007 here), I had made the (to-me-obvious) point that it would be difficult to motivate policy makers or the public with any sense of urgency using the Lamb graphic.
Crowley’s Climategate recollection clearly shows the frustration of Hockey Team scientists in 1996: they felt (incorrectly IMO) that Bradley and Jones 1993 and Hughes and Diaz 1994 had refuted concepts of the Little Ice Age and MWP, but had nothing to substitute in its place. Thus, Bradley in 1991 and Crowley in 1996 had both ended up using the Lamb graphic – reluctantly but used nonetheless. It seems entirely plausible to me that Hockey Team scientists in 1995-96 really did want to “get rid of the MWP”. Even as late as 2007, Overpeck, as an IPCC Coordinating Lead Author, still wanted to “deal a mortal blow to the MWP” (see CA post here).
Mashey contested the validity of Deming’s story on the basis that “Nobody in 1995 would have been worried about “getting rid of the MWP,” although they might have wanted to dispel the idea the Lamb schematic was Truth. They had years before abandoned the schematic.” However, it seems to me that the closer re-examination of the documents carried out here, if anything, adds circumstantial support for Deming’s recollection (I emphasize circumstantial). Mashey’s claim that no one used the Lamb graphic after 1992 is refuted on its face by Crowley 1996. This doesn’t prove that Deming’s recollection is valid, only that Mashey’s argument that “nobody in 1995 would have been worried about “getting rid of the MWP” is completely imaginary.
Postscript: Ross and my criticism of the Mann graphic was not on the basis that the Lamb graphic was “right”, but only that the Mann reconstruction had used questionable methods, did not live up to its warranties of statistical skill and robustness and therefore did not permit Mann or the IPCC to assert with statistical confidence that 1998 was the warmest year of the millennium. (Our actual argument has been misconstrued by both supporters and critics alike.) We did not preclude the possibility that someone could show that 1998 was the warmest year of the millennium using better data and statistics, only that Mann et al hadn’t done so. Subsequently, mainly at Climate Audit, I’ve argued that other Hockey Stick reconstructions are also defective, but that’s a long and different story.
Climate Audit 
159 Comments
How strange to present data, charts, graphics, whatever, that you don’t think are correct but feel compelled to because “there is nothing else available”. Maybe a little scientific agnosticism and/or humility in this field could do everyone a lot of good.
Fred Pearce in the Guardian gave credence to David Deming
Pearce is an honest journalist who remembers these conversations with climate scientists when it matters. We only need a few more like him.
Re: Richard Drake (Oct 9 14:40),
In fact, we need a LOT more like him.
Having been conversant with the older (pre-2000) literature, such as above, it has been astonishing to me the vehemence with which my and other’s statements to the effect that the MWP and Holocene maximum might have been warmer have been met. I recall official Global Change Program reports from around 2000 or so mentioning summer N. Hemisphere temperatures being up to 2-4 deg C higher. And yet to mention this was to be a denier…go figure.
Steve, many thanks for your work! I think from the “history of science” perspective your insight is immensely valuable.
By the time I had grown skeptical of the “consensus”, I was wondering how exactly the consensus had formed in the first place. The progression from “purely schematic” (“We suspect it is like this”) to “robust replication” (“We had always known it is like this”) seems to have missed the “Is it really like this?” step – and all the time it seems what was “known” has been a bit of a moving target.
I for one find all the name calling and breathless hyperventilating from the hockey team members (real and wannabe), like now from Mashey, counterproductive, to say the least. You on the other hand have a wonderful “stick to the facts” way. From that point I find the quote about “Bradley, regarded as notoriously vain in the Climategate emails” a bit unnecessary, at least without a link to the actual Climategate mails – surely someone will criticize you for that sentence, instead of the original author in the Climategate mails.
Steve,
It is ridiculous to treat these two uses as though they are the same as that in the IPCC 1990. Most importantly, the two graphs are significantly modified from that of IPCC 1990, which shows a MWP temperature 0.6 C above the baseline. In contrast, that by Bradley and Eddy 1991 shows a warming above the baseline of about 0.3 C and shows the line as dashed over the period of the MWP to clearly indicate the relative temperature during the MWP is uncertain. Crowley 1996 reduces the MWP temperature above baseline still further to about 0.24 C and very explicitly describes it as a schematic of a regional, not a global temperature. (Note that in the IPCC graph the baseline equates to 1900 temperatures on the graph.)
It takes a special talent to see the use of a graph showing a regional MWP of 0.24 C above the baseline as endorsing a prior graph showing a global 0.6 C excursion.
This is not to deny the (obvious) lineage of the graph. The lineage, however, drives the point home more strongly, for it means the alterations were conscious and reflect a change in understanding.
Further, given that Crowley expressed reservations about using a modified IPCC 1990 graph with the MWP reduced by 60%, described purely as a regional graph – it is a long bow to suggest this refutes John Mashey’s claim that by that period climate scientists would not have accepted the use of the original graph as representing global temperatures. In fact, it rather supports it.
Re: Tom Curtis (Oct 9 15:27),
If “there was not an alternative,” where did the “change in understanding” come from?
Tom, I’ve noticed that you haven’t commented on Skepticalscience for nearly a month now, since the Lewandowsky post. Have you been banned from posting there?
If you just redraw the same figure and shift the baseline and redescribe it as regional vs global, this is all still conceptual, not data-driven. That is, all versions of the graph in question are the opinions of the authors. Nowhere is the rescaling described or justified, and several alterations of figures are described which have no basis except the intuition of those doing the drawings. This is science?
Re Tom Curtis
Exactly where do you get the value of 0.6°C from. The IPCC graph has only an ARBITRARY scale, and nowhere in Ch7 is it referenced back to an absolute scale. As far as I can tell, the scale divisions could be 0.5, 0.2 or Kilngon Spaceships for that matter.
John Mashey? ROFLMAO
Tom Curtis – I have the same question as sleeper. Where did the supposed information come from that brought it from 0.6 to 0.24 or whatever? It does not really matter what the scale is, the salient point is the shape and height of MWP vs. LIA and Current.
I’ve edited the post a little, primarily by including the IPCC 1990 versions for more direct comparison.
Tom C, the ratios of the peak warmth shown for the MWP relative to the peak warmth shown in the modern era relative to a “1900” baseline are (approx time of peak modern warmth shown in brackets):
IPCC 1990 – 3.7 (1950)
Eddy and Bradley 91 – 1.3 (1945)
Crowley 96 – 1.7 (1945)
The increased relative ratio in Crowley 96 relative to Eddy and Bradley 91 is due to a reduced modern warming. By your chosen measure, the graphs show a clear change in understanding, and are not evidence that climate scientists post 1990 would have been happy to use the IPCC 1990 graph as is.
RE Tom Curtis again
Where exactly in Ch7 does the IPCC refer to a scaling of Figure 7.1, or alternatively where in Ch7 can I find cross refs to do so. If you can find I would be appreciative of you sharing it. Scaling it using later works has nothing to do with the fact that it is an arbitrary scale in ch7. Your claim that Fig 7.1 shows an MWP of 0.6°C is nonsense.
Terry, if you cast your eye over Figure 7.1(c) reproduced by Steve above, you will notice on the left the words, “Temperature Change ([degrees]C)”
Next to those words you will find a scale. The claim that Fig 7.1 (c) of IPCC 1990 WG 1 has an arbitrary scale is a fiction.
WEll then I guess my eyesight must be decieving me more than I think. Fig 7.1 has units in °C but for the life of me, and I expect most other normal folk, I see no scale. Can you tell me with 100% certainty that the divisions are 1.0 or 0.5 or 0.2 OTOH The scale might be labeled using Special “Tom” ink that no-one else can see. If I shine a UV light at it will it magically appear ?
Craig Loehle, every version of the graph right back to Lamb 1965 was the (informed) “opinion of the author”. If you have a problem with that, your correct stance should be that any talk of MWP, LIA, RWP, Miocene WP, and HCO prior to, at earliest MBH 98, and at latest within the last couple of years is not science. What is more, your hero should be Mann who, whether successfully or not, at least tried to move paleoclimatology of the recent past beyond (mere) informed opinion onto a genuine scientific basis.
Somehow, however, I suspect you will consider the mere informed opinion of Lamb to be science, while considering the merely better informed opinion of Crowley to not be science.
I take it you’re being facetious. The genuine science was in a folder called “CENSORED”.
Ross McKittrick, I know the fiction you tell about the “Censored” file, and it is just that, a fiction.
Tom Curtis: you can suspect whatever you want, but I do not accept the Central England record of Lamb as anything more than indicative. Thank you for agreeing with me that these graphs represent opinions (not the ones directly from ice cores but the recent one at issue). And if that was what they thought in the 1990s, then claims by Mann later that no one ever believed in the MWP are false.
What they believed in the 1990s was that the Medieval Warm period “may not have been global”. How often must that be quoted before auditors realize that fig 7.1(c) is not the whole story of the opinion of 1990s paleoclimatologists?
Tom Curtis –
Saying merely that the MWP “may not have been global” understates the FAR’s view. The Executive Summary, in discussing climatic fluctuations, mentions as an aside that the MWP “may not have been global.”
A fuller expression is in section 7.2.1 (next to figure 7.1): “The late tenth to early thirteenth centuries (about AD 950-1250) appear to have been exceptionally warm in western Europe, Iceland and Greenland (Alexandre 1987, Lamb, 1988). This period is known as the Medieval Climatic Optimum. China was, however, cold at this time (mainly in winter) but South Japan was warm (Yoshino, 1978) This period of widespread warmth is notable in that there is no evidence that it was accompanied by an increase of greenhouse gases.”
Section 10.0 adds: “For example, the boreal forests of Canada extend [sic] well north of the current timber line during the Medieval Warm Epoch (800 to 1200 AD), a time when temperatures in that region were about 1°C warmer than today’s. At the same time, farmers in Scandinavia grew cereal crops as far north as 65° latitude (Lamb 1977).”
Where my previous post (5:16 am) has “Executive Summary”, read “Chapter 7 Executive Summary.”
This is unusually specious, even for Mashey. If, as he claims, no-one in CliSci wanted to ‘get rid of the MWP’ after 1992 because its existence was no longer credible to climate scientists, then that means that we have a clear and distinct example (by Mashey’s reasoning) of climate scientists being wrong, and the ‘settled’ consensus changing.
sleeper, the change in understanding came from a greater number of regional proxies from a greater diversity of geographical locations coupled with a concerted attempt to determine the extent to which regional warm periods actually coincided.
That, however, is beside the point. At issue is whether or not the later used of similar, but crucially adjusted graphs represents an ongoing endorsement of the conjectured relative (possibly) global warmths shown in IPCC 1990. Clearly they do not. McIntyre may be inclined to swallow that camel, but I don’t see any reason to forget the ability to reason and follow him.
Tom, you don’ seem upset by isotope series from (say) Vostok being used as a proxy for global climate, even though it is from a single site.
One important contrast between Lamb and the Bradley-Jones-Mann-Crowley school was that Lamb believed that global circulation caused organized changes in world climate i.e. that the Little Ice Age and Medieval Climate Anomaly (if you prefer) were global phenomena though, in the tropics, the changes were mainly manifested in precipitation.
Bradley-Jones-Mann-Crowley argued that regional climate could move more or less independently. AR5 drafts clearly resile from this position, recognizing the Little Ice Age and Medieval Climate Anomaly as real epochs. IN my opinion, the case for disjointed regional climate is not articulated in the literature, as much as arm waved. It is clear that so-called “proxies” frequently are inconsistent, but that may merely mean that the inconsistency is with the proxies rather than with the climate. If you can identify any references that establish the point through more than arm waving, I’d appreciate it. Please note that I’m obviously very familiar with the literature and the point is not trivial.
Steve, I have already answered precisely this question from you from another thread:
“As to the use of a single site, there is a clear difference between using a single site to represent global temperatures in the holocene, during which variations in regional temperatures have been of the same magnitude as centenial variation in temperature within individual regions; and using a single site to show the timing of onset and end of glacial and interglacials over the last million years. Despite that, the IPCC FAR clearly labels the first panel as a “schematic” thereby showing they make no claim that minor fluctuations are reproduced globally. Further, I personally far prefer the use of multiple site proxies such as that by Lisiecki and Raymo even for glacial/interglacial intervals.
http://lorraine-lisiecki.com/stack.html”
As you responded to that comment, you are certainly aware of my opinion; so your raising the question again must be regarded as a deliberate red herring.
Speaking of which, the issue raised by you in the OP was whether the IPCC 1990 graphic was used by climate scientists in the period 1992-1995 in a way that would rebut John Mashey’s claim that by 1992 it had been rejected as misrepresenting what really occurred. You suggest two putative counter-examples, and as I have clearly demonstrated, they are not actually counter-examples. Both were ammended significantly in ways that reduced the presumed MWP warmth, and the later (the only actually relevant one, as being post 1991) clearly only treats it as a regional proxy, not a global proxy (however qualified). As such you have no counter-example.
I have noted since then that you and your friends have been piling on with points to debate, none of which are relevant to the main point. I am not here to debate all of climate science. Therefore I have no interest in being distracted by your various red herrings. Either defend your initial claim that IPCC 1990 Fig 7.1(c) was still considered an adequate estimate of global MWP warmth in the mid 1990s, or admit your error. Continued red herrings will only show me that you neither have the means to do the former, nor the integrity to do the later.
Tom Curtis wrote:
“Both were ammended significantly in ways that reduced the presumed MWP warmth…”
Lamb made the most accurate chain saw cut he could at the time.
Bradley and Eddy redrew it with a burnt stick.
Crowley redrew it with an unsharpened pencil.
Tom Curtis insists that we need to measure it with a micrometer to understand it.
Tom Curtis – Steve’s comment just above also points to a fatal contradiction of Mann et. al.’s work. The argument that Bristlecone pine dominated PCs fairly represent world GMT becuase of “teleconnections” is at odds with the claim that the MWP was just a regional event. When climate around the globe has to be explained to fit our “genuine belief” we invoke “teleconnections”; when 3 solid centures of unusually high temperatures show up at multiple spots around the world everything is “regional”.
I’ve often wondered about how people characterize your work Steve.
” We did not preclude the possibility that someone could show that 1998 was the warmest year of the millennium using better data and statistics”
You are widely labeled “skeptic” across the Internet, yet you have made little statement about warming in general. This sentence is very carefully worded as are most in this post. Why people miss the nuance is beyond me but the recognition of both the nuance and the reaction by the main stream climate community are what makes your work so influential.
Better “data” and “statistics” is well worded because both would be required to solve the problem. CA readers know that Journal published interpretations of unusual data are often less than objective. It is heartening to me that some of the more recent work has recognized the pitfalls of automated data sorting and begun to mitigate its damage to their field. Still, it is disappointing that more don’t label the work for what it was.
From my experience, when I look at the plots, all I see are data and math problems. I don’t always know where they are by paper but do know that they are definitely there. It sounds horribly dismissive, but that is the reality of the field.
If paleoclimatolgy fixes their entire field’s work at this point, I would propose a new name paleo-paleoclimatolgoy.
Dave, a scientific finding is only supported by a settled consensus if it is accepted by the vast majority of scientist within the relevant field (ie, it is a consensus) and it has been so supported for a long period of time (ie, it is settled). Scientific consensuses, no matter how settled, can always be over turned which is the reason mavericks should not be denied funding or a place in universities. However, there has never been a consensus among paleoclimatologists that the Earth’s global mean surface temperature was warmer in the MWP than in the mid-twentieth century. Even the IPCC 1990, the much miss cited source for the claim of such a consensus can only say the “Medieval Warm Period around 1000 AD (which may not have been global)”. And if no consensus, still less has it been a settled consensus.
You may find it easy to simply make up “facts” to suite your case, and expect such fictions to be simply swallowed wholesale at various “skeptic” blogs. However, that cavalier attitude to truth leaves me very cold.
snip
Steve: the issue of “consensus” is an interesting one but O/T here.
Tom Curtis – Fair enough explanation. However, I really don’t believe for a minute that enough data exists to truly calculate whether these ratios are 3.7 or 1.3 etc. It would likely be embarassing for Crowley were he asked to show his calculations. I think the MWP was much warmer than now but my opinion is based much more on historical evidence than proxy science. When something can’t be calculated it is best to admit that it can’t be calculated rather than make something up because “there was not an alternative”.
I say this constantly, even within the context of my own research. Sometimes the answer we want can not be obtained with the data we have. And that is the correct answer. Saying we have a result with R-sq=0.1 is simply wishful thinking. You have not explained anything and probably captured spurious relations.
Steve: The verification r2 of the Mann AD1400 reconstruction was 0.02 and of the AD1600 step something like 0.00004. But of course, Mann didn’t report these horrific results. I guess that’s what Tom C regards as realscience.
I agree with Tom C that ‘when something can’t be calculated it’s best to admit it can’t be calculated rather than make something up “because there was not an alternative”.
There is sufficient evidence from Asia and the Pacific regions to suggest the MWP was at least as warm as now – we may have to leave it at that until we have more definitive real evidence.
Tom C, talking about the actual estimates (rather than the graphs themselves), of course they were not accurate enough in 1990 or 1996 to be capable of exact numerical resolution. That does not mean they ought not to be represented graphically, so long as it is made clear that the graphic representation is a schematic, not a reconstruction. That is no more problematic in paleoclimatology than diagrams of the internal structure of atoms are in quantum mechanics. In the former case, it is known that exact numerical values cannot be given, but a diagram can give a reasonable idea of the best estimates. In the later case, (fairly) exact numerical descriptions can be formulated, but it is known that any two dimensional representation will be significantly inaccurate. They are still useful as aids to the imagination in grasping complex concepts.
However, because the schematic representations of global mean surface temperature over time have changed, we are able to trace the trajectory of “informed estimates” over time. Clearly both the estimated peak of temperature in the MWP and the confidence that it was global declined from 1990 to 1996.
As to McIntyre’s main contention, the mere fact that firs Eddy and Bradley, then Cowper felt it necessary to modify the schematic shows that they at least would not have accepted using the Lamb schematic unaltered. McIntyre is here presenting evidence in favour of Mashey’s claim as though it some how refutes it.
Steve: Tom, again your zeal blinds you. Crowley’s Climategate recollections clearly show that he and others were very frustrated by having to use Lamb-based graphics and that they wanted to get rid of its implication of an elevated MWP. Overpeck still wanted to deal a “mortal blow” in 2007. Because the Mann reconstruction gave them the answer that they were looking for, it was inhaled far too quickly. Its many defects were overlooked.
The canonical HS shape, despite claims that it is “NH”, is imparted by Graybill bristlecone chronologies, which are no more representative of world temperature (perhaps less representative) than CET. But of course, the climate community never discusses this.
Tom Curtis,
You have some points about the scale being changed and the Warm period not as pronounced and that it is labeled “regional”.
However, all of the graphs have “Medieval Warm Period” written clearly across the major warm period and all show it as similar to or higher than today’s temperatures.
Thus, I think that this fits just fine with the idea that people were not satisfied with these graphs and would want to “get rid of the MWP”. The fact that (as you pointed out) they modified the graphs also suggests they wanted to get rid of it, as they did not think it was global, or as warm as depicted.
Bill, all have MWP written across the graph, and all show temperatures higher than in the 1930s to 1940s. Unless you are living in a time warp, the 1930s and 40s are not today, and you should not misrepresent them as being so.
Steve, the CRU hack emails clearly show Crowley to be frustrated with the lack of a good alternative because he genuinely believed the Lamb graphic overstated MWP temperatures. They show Overpeck to likewise genuinely believe MWP to be much less than shown by Lamb and IPCC 1990. In both cases, nothing shows these beliefs to have been formed in anyway other than by examination of the evidence.
The emails also clearly show that Overpeck is keen to deal a mortal blow to myths about the MWP, not the notion of a MWP itself. That you conflate the two in order to stitch Overpeck up speaks little of your integrity.
Re: Tom Curtis (Oct 9 22:46),
Having a “genuine” belief is fair enough but asking how the “genuine” belief arose is also fair I think. I think it is the history of how the elevation of one flavour of “genuine” belief above all others occured which is the major question here. Chicken and egg? 😉
For where would ClimateAudit be without it’s endless imputations of motives? A scientist cannot honestly believe the evidence demonstrates something. There must always be a more devious motive hidden somewhere.
If you want to get upset about endless imputations of motives, read the climategate emails. Or any interview, speech, or letter to the editor authored by Michael Mann.
Yes, Tom Curtis’ reading of these emails is correct,and should not be controversial.
Overpeck repeatedly questions warmer MWP proxies (Jasper) based on sample depth of 5 or less, but its cricket sounds when the single tree YAD061 validates the stick. How does this dichotomy demonstrate unbiased reading of the evidence?
Example: http://www.ecowho.com/foia.php?file=1534.txt
“5) Talking specifically about Jasper, it is interesting that the 20th
century is as warm or warmer than everything in the last 1000 years EXCEPT
before ca. 1110 AD. Since the sample depth before this time is 5 or less,
how much faith should we put in those warmer than modern temps??”
Steve McIntyre:
Indeed all of the good “long-period” proxies are geographically limited.
There are plausibility arguments (shorn up partially by climate models) that long-period variability (e.g., 100-year scales) track globally (i.e., “telecommunication”). If you don’t make that argument, I don’t see how you can even try and perform a reconstruction.
I guess Tom Curtis is saying he doesn’t believe any of the paleo-reconstructions. >.<
Carrick, there is a huge difference in “catch” area of different proxies. Some indeed are very limited in area, but e.g. the Ireland and South England (Lamb) temperatures largely reflect the Atlantic Ocean temperatures in the main (SW) wind direction. The Vostok dD and d18O isotope ratio changes are from evaporation and precipitation of most of the whole SH oceans. Thus they largely reflect the SH temperatures over time. The coastal Antarctic ice cores reflect a more nearby catch area: the oceans near the Antarctic coast which show an opposite variability between the Peninsula and the rest of the coast (with an ENSO related modulation). The Greenland ice cores reflect the nearby North Atlantic Ocean temperatures, in general with a variability opposite to the NW Europe temperatures (NAO?), etc…
Thanks FerdiEgb. That was very helpful.
Carrick, do not lie about what I believe. You obviously are not interested in what I actually believe or you would quote me.
Yeah Carrick, don’t intentionally say things you know are untrue when you say stuff like:
You evil liar, you!
Yep, evil liar. For a statement that starts with “I guess…” That’s the descriptor I’d use too.
kek
Carrick, both you and Brandon know that the “I guess” was entirely rhetorical. The statement was intended to attribute to me a ridiculous position for rhetorical purposes, based on a comment by McIntyre that I had already answered, albeit on another thread. I admit that when I made that comment, I had not seen that Steve had, for rhetorical purposes, repeated a comment from another thread while ignoring my direct response to that comment which showed his comment to be inaccurate. Therefore I had assumed that you had quoted Steve from another thread while ignoring my response on that thread which rebutted the opinion you attributed to me as well. Therefore, the primary fault lies with Steve who deliberately, and knowingly attributed to me an opinion in this thread which he knew I did not hold because of my response on the other thread. Never-the-less, your attribution of an opinion to me based solely on Steve’s attribution of an opinion to me, all without any reference to what I have actually said shows a contempt for truth.
Good grief.
Tom Curtis — calm down! I was kind of admiring your persistence up to here….
—
I live in a country which had a failed coup d’etat over a guy getting a blow-job. It’s really hard to stop laughing at this. — Bruce Sterling
Tom Curtis, just remember when you see the >.< emoticon, which I also used in that statement that you took such affront with, that signals that everything I say is to be taken literally. Me thinks there is a bulb out somewhere. It's a bit dim in here.
😛
Carrick, never seen that emoticon before in my life. Neither, apparently has google (which returns zero hits for a search of those symbols). Nor html, which does not convert it.
Further, that has no bearing why you thought your idle speculations about my beliefs should be shared with the world. If people want to know my beliefs, they should ask me, not present their idle speculations as a source of confusion for others.
You won’t find any emoticon Tom, as Google doesn’t search ANY punctuation. Not even if you put it in quote marks. It can be very annoying, as it happens, when you actually want to specify punctuation.
The emoticon in question is however in the top ten list of emoticons in Wikipedia’s list, and took me 30 seconds to find.
Mooloo, thank you for the heads up. Unfortunately, wikipedia also indicates that the emoticon does not indicate that everything before hand should be taken literally.
tom now you missed the sarcasm. sheesh. >.<
Tom Curtis:
No it doesn’t. Does it? /sarc
I think you need to find a different sport than being a rhetorical soccer ball. I mean it’s your choice,if you like it keep going. 😕
Tom Curtis:
Oh good grief. It’s a friggin’ facebook emoticon.
Here, let me help you with that google search.
(Search for >.< after that point. I'm assuming you know how to execute a Find in your browser.)
More “flat-earthers” (GISP2):
Mayewski, P.A., L.D. Meeker, M.C. Morrison, M.S. Twickler, S.I. Whitlow, K.K. Ferland, D.A. Meese, M.R. Legrand, and J.P. Steffensen. 1993. Greenland ice core “signal” characteristics: An expanded view of climate change. Journal of Geophysical Research 98:12839-12847.
Steve:
The bottom two panels of Mayewski et al 1993 are shown below.
Mayewski et al (1993) use the diagram from Eddy and Bradley (1991), complete with a 0.3 C MWP warming relative to 1900, ie, just half that in IPCC 1990.
Are you saying now that Mayewski et al do not count as “flat-earthers” (per Mashey) as they used the “bodged” diagram instead of the original? Does a similar argument apply to Zielinski et al (1994), who used the E&B version of the middle panel? Did I just see goal posts move?
Jean S, Mayewski, and later Zielinski, had the choice of using either the IPCC 1990 diagram, or the Eddy and Bradley 1991 diagram. The fact that they chose the later indicates that they thought it more accurate. The substantive difference is that it halves the peak indicated warmth of the MWP relative to 1900. Their choice of the later, therefore, clearly indicates that they though the MWP was significantly cooler than was though by Lamb. That is not an endorsement of the earlier view, but a significant amendment of it. Pointing that out does not represent a goal post shift. On the other hand, pretending that a schematic with a plus 0.24 C MWP on a regional basis somehow represents an endorsement of the IPCC diagram as a global representation, as McIntyre does is beyond belief.
Tom Curtis wrote:
Unfortunately, Tom is just making stuff up here. Zielinski et al 1994 did NOT pick a graphic that “halved” the warmth of the MWP relative to 1900. On the contrary, as Jean S pointed out, they picked a graphic that enhanced MWP warmth relative to modern. They showed the MWP as 0.7 deg C warmer than the 20th century, as shown below. Zielinski et al presented this graphic as representing “Northern Hemisphere”, not global temperatures. This suggests to me that, in their article the previous year Mayewski et al 1993), they had not turned their minds to whether the Bradley bodge was or was not an improvement on the IPCC schematic (as opposed to simply using a figure at hand.) Nor on the same principles do I regard their use of the version shown below as an “endorsement” of it, as much as recognition that it was a published view of recognized scholars in the field.
Bradley was obviously aware of the Zielinski et al 1994 graphic as he referred to it with some annoyance in a 2007 Climategate email. That he remembered it at all shows that this sort of thing grated on him. I’m sure that one can include Bradley among those who wanted to get rid of figures like the bottom panel of Zielinski et al 1994.
Again, I re-iterate that I’m not “endorsing” any of these figures. My issue was that people inhaled the replacement figures too readily because they “liked” the answer without fully evaluating the flimsiness and pseudo-statistics in the Mann-style replacements. As I repeatedly say, problems with the Mann HS do not mean that Lamb graphics are “right”.
Tom, I see, they are moving fast. Zielinski et al caption (my bold):
Steve: Zielinski et al re-drew from IPCC 1990 bottom and middle panels as shown below:
Zielinski et al 1994 Figure 1. Caption stated: Composite of the postulated variation in Northern Hemisphere surface temperatures from the present to 7000 B.C. (35) Temperature change was determined by a deviation from the mean value at the turn of the last century. (35 – Taken from J. A. Eddy and R. S. Bradley, Earthquest 5 (insert) (1991), as modified from J. Houghton, G. J. Jenkins, J. J. Ephraums, The IPCC Scientific Assessment (Cambridge Univ. Press, Cambridge, 1990).
Co-authors of flat-earth Zielinski et al (Science 1994) were G. A. Zielinski, P. A. Mayewski, L. D. Meeker, S. Whitlow, M. S. Twickler, M. Morrison, D. A. Meese, A. J. Gow, R. B. Alley. Richard Alley has been mentioned from time to time here.
Jean S, I note that Zielinski et al 1994 is not as you describe it:
http://digitalcommons.library.umaine.edu/cgi/viewcontent.cgi?article=1191&context=ers_facpub
If you are referring to a different Zielinski et al 1994, perhaps you could cite it correctly.
In the mean time, it appears that all you have to add to the discussion is a slogan. Chant it loud enough and you may be able to avoid all troublesome thoughts.
Tom, perhaps it would helpful if you actually read these posts. From Steve’s post:
Jean S:
No, that is not helpful. That gets in the way of him making stuff up. 🙄
kek
1) In my previous comment I mistook Jean S as saying that Zielinski 94 used the analogue of IPCC 1990 fig 7.1(c) as a second panel, rather than as saying they used an analogue of 7.1(b). Consequently I made a mistaken inference for which I apologize.
2) I note that Steve has taken the trouble to mark temperature differences relative to the baseline for Zielinski 94, but is inconsistent in doing so. Specifically, if the temperature difference matters, then it ought also to be marked on the two analogues of 7.1(c). If it does not, then there was no need to mark it for Zielinski 94. As it stands, by marking those differences where it appears to support his case, but not where it fails to support his case he gives the impression that he is more interested in persuasion than in analysis.
3) On a side note, and having read one of Steve’s updates, he purports that I said the dashed line in the Eddy analogue of 7.1(c) “‘clearly’ show[ed]” that the graphic had been, in his words, “bodged”. What I actually wrote was that the dashed lines “… clearly indicate the relative temperature during the MWP is uncertain”. That is an entirely different thing.
Steve also suggests that a dashed line is not an appropriate method of indicating a “bodge” (again, not its purpose), neglecting the fact that Eddy and Bradley clearly label the graph as “Modified from J.T.Houghton et al, 1990”
Indeed, Eddy and Bradley write of the analogues of Figures 7.1(b) and 7.1(c) that, “the temperature excursions depicted in this and the subsequent graph are the most conjectural of the set”.
John Mashey’s view, that Steve is determined to rebut, was that the IPCC 1990 figure 7.1(c) was not accepted as “the Truth” during the period 1992-1995 (or indeed at any time by serious climate scientists), and that after 1992 it was not accepted as an accurate presentation what was in fact known about global temperatures.
Curiously, Eddy and Bradley explicitly discuss the temperature differences, giving the MWP as being 0.3 C above reference, LIA as being 0.6 C below reference, and the middle twentieth century as being “almost” 1 C above LIA temperatues. That makes the middle twentieth century at least equal to the MWP peak, and possibly just above it, according to Eddy and Bradley. If the middle twentieth century is equal to MWP temperatures,then clearly the 1980s and (later) the 1990s were above them.
4) That brings me back to Zielinski. To rebut Mashey, McIntyre needs to show a use of fig 7.1(c) or analogue which endorses the temperature differences shown therein. It is not sufficient to show a use which represents just an historical quotation, ie, an indication of what had been accepted in the past for comparison purposes.
Instead, what he shows us is so crude a shematic that it represents the last 230 years of temperature terminating in 2000 AD as a flat line. If we are to take the fractional temperature differences from the baseline in this graphic at all seriously, which we must if it is to stand as a counterexample to Mashey, then we must say that the IPCC 1990 directly contradicted themselves in that figure 7.1(b) and 7.1(c) have inconsistent curves.
Nor can Steve bring himself to say the authors endorsed the graphic, saying, “Nor on the same principles do I regard their use of the version shown below as an “endorsement” of it, as much as recognition that it was a published view of recognized scholars in the field.” That being the case, this no more represents a counter example to Mashey’s claims than Steve’s own use of IPCC 1990 Fig 7.1(c) in 2005 to show what experts thought at the time (although he continues to ignore said experts qualifications).
Tom, I’m sure you are innovative enough to shift the goal posts once again (Figure 11):
Hass, H.C. 1996. Northern Europe climate variations during late Holocene: evidence from marine Skagerrak. Palaeogeography, Palaeoclimatology, Palaeoecology 123: 121-145.
Addition: Does anyone have access to the book (in German):
Schönwiese, C.-D., 1994. Klimatologie. Ulmer, Stuttgart, 436 pp.
It would be interesting to see, how (if) the FAR account on climate history was dealt there.
Desperation must be setting in. Now Jean S cites an article with no analogues of any panel from IPCC 1990 WG1 Fig 7 as a refutation of a thesis about the use of one panel from Fig 7. She further neglects the fact that the articles deals with just one site, and for that site finds MWP comparable to, but not warmer than modern times, defined as 1950-1980. In temperature terms, that makes it equivalent to Eddy and Bradley, not IPCC 1990.
The Hass 1996 figure caption says “Global temperature variations (after Folland et al 1990)” – which is of course the IPCC report. The paper also has a discussion of the ‘worldwide’ nature of the MWP and LIA from p 136-138. (And I think Jean S = he).
C’mon, Tom, I thought you could do better. The second bottom panel in Fig 11 (“Global temperature variations (after Folland et al. 1990)) in Hass (1996) is a redrawn version of IPCC 1990 Fig 7.1(c). In case you still have problems visualizing it, compare it to Figure 3/2nd panel in a proceedings paper from the same author:
Hass, H.C. and Kaminski, M.A. (1995) Change in atmospheric and oceanic circulation reflected in North Sea sediments during the late Holocene. In: Langbein, R. and Niedermeyer, R.-O., (eds.) Sediment ’94: 9. Sedimentologen-Treffen, Greifswald 1994 / Sediment ’94: Proceedings of the 9th Sedimentology Meeting, Greifswald, Germany, 1994. (pp. pp. 51-65). E. Schweizerbart’sche Verlagsbuchhandlung: Stuttgart, Germany.
BTW, it’s “he” not “she”.
Steve: Here is the graphic referred to by Jean S.
Jean S, the second bottom panel of figure 11 is a grey scale representation of the duration and sequence of warm and cold periods based on, as you note, IPCC fig 7.1(c). However, it contains no temperature scale of any sort and so cannot show an acceptance of MWP warmth significantly warmer than the modern period. As such it is even less analogous to IPCC fig 7.1(C) than the examples provided by McIntyre. Given the several statements in the paper that MWP warmth at at least this location was equivalent to that in the modern era (mid 20th century), this is an even worse counter example than Eddy and Bradley.
The proceedings paper, in contrast directly uses fig 7.1(c), which therefore makes it the best possible counterexample proposed. I would argue that, given the close logical relationship between the two papers, and the comments on the scope of MWP warmth, and comparison with the modern period in the former paper, Hass and Kaminski do not endorse fig 7.1(c) as representing the global relationship between modern and MWP temperatures.
Against that, Hass explicitly indicates European MWP temperatures up to two degrees above the mid-twentieth century average; and that the warming was wide spread. That data for the one region he analyses show approximately equal warmth does not indicate automatically that global temperatures where not greater in the MWP than in the modern era. Consequently it is certainly possible on available evidence that Hass and Kaminski thought Fig 7.1(c) fairly represented global temperatures.
Consequently I must congratulate you on finding a possible counter-example to Mashey’s claims. If anything hung on the issue I would contact Hass, if possible to see which of the two possible positions better represents his view. As nothing of significance does hang on the issue, I won’t bother.
I will note that the mere possibility of a counter example shows that Mashey’s rhetorical flourishes (flat earth theory) are unjustified. On the other hand, his basic thesis could easily be couched in less florid terms, in which case the existence of one obscure counter-example would not rebut the basic facts – the evidence gathered in the period after 1990 showed that Lamb’s graphic had significantly overstated MWP warmth. Paleoclimatologists of the time knew that, and would have felt no need to destroy a theory of the MWP as warmer than mid to late 20th century when they had every reason to think that theory was already well on the way to the scientific graveyard.
Tom, when you say “the evidence gathered in the period after 1990 showed that Lamb’s graphic had significantly overstated MWP warmth”, what evidence are you referring to? I agree that Bradley’s version of the graphic showed a lower estimate of MWP warmth than the underlying Lamb graphic, but it shows Bradley’s judgement and is not itself evidence, let alone evidence gathered in the period after 1990. Do you have anything else in mind?
Tom Curtis:
Actually, relative to the LIA minimum, Lamb gets the MWP warming about right. What he clearly gets wrong, and they would have known this in 1990, is it grossly understates the modern warming period. Indeed if you look at the figure, the temperature very bizarrely is “flatlined” from about 1900 and extends to 2000, 30+ years after the end of any available data for the original graphic.
If you tack on the instrument temperature record since about 1850 to 1970, 1970 (where his graph should have ended) would be about as warm as 1970. If you update it to 2000, the instrumental record would be at about 1.25°C on that temperature scale, well above the MWP maximum.
Here’s an ensemble of proxy reconstructions.
Contrary to the claim that Lamb has “significantly overstated” the MWP, if you compare the temperature from sequence from the the RWP, the DACP, the MWP to the LIA with modern proxy reconstructions, Lamb’s delta-T is reasonably good agreement with recent reconstructions.
This figure also illustrates just how starkly different MBH is from later reconstructions, including Mann’s own 2008 EIV series (which we recognize to have problems, but anyway).
I think it also illustrates the talking point that the modern warming period is warmer than the MWP hadn’t been as widely adopted yet. In post 1990 publications that included renditions of Lamb’s figure, they were still more focused on temperatures of the past, it seems, and less on how that related to recent global warming.
Regarding Mashey’s commentary, if you red ink anybody’s enough, certainly you can make what it says “right.”
Steve, knock yourself out.
http://scholar.google.com.au/scholar?as_q=medieval+warm+period&as_epq=&as_oq=&as_eq=&as_occt=any&as_sauthors=&as_publication=&as_ylo=1991&as_yhi=1995&btnG=&hl=en&as_sdt=0%2C5
Steve: what a ridiculous response. I had asked:
Googling “medieval warm period” is not answer.
In other words, when you said: “the evidence gathered in the period after 1990 showed that Lamb’s graphic had significantly overstated MWP warmth”, you had no knowledge of whether that statement was true. You just made that up. Shame on you.
You can find the article reprinted here.
Carrick, the Lamb graphic in IPCC 1990 shows a 1.3 C temperature difference between MWP maximum and LIA minimum. Your graphic shows 0.8 C for the same interval. The Lamb graphic shows mid twentieth century warmth as 0.86 C warmer than the LIA minimum. Your graphic shows the difference as being 0.74 C. Given the finer temporal resolution of your graphic, we would expect the Lamb graph to show smaller differences not larger. Clearly it is absurd to claim:
“Actually, relative to the LIA minimum, Lamb gets the MWP warming about right. What he clearly gets wrong, and they would have known this in 1990, is it grossly understates the modern warming period.”
Did you even bother to determine the differences?
In the event, Lamb and the IPCC 1990 only slightly over stated the LIA/middle 20th century difference, but greatly over stated the relative warmth of the MWP,whether compared to LIA or to middle twentieth century. The difference between LIA and middle twentieth century is, I believe mostly due to an understatement of twentieth century warming relative to 1900.
For the record, the graph as adjusted by Crowley 96, with a minimum to peak difference of 0.8 C most accurately reflects the LIA/MWP difference as shown by your graphic. He also shows a LIA minimum to mid twentieth century maximum of 0.74 C. Both values are, within measurement error, exactly the same as the equivalent values in your graphic. Never-the-less, he was wise to treat it as a schematic regional proxy only. First, your schematic averages only reconstructions with a high value for the MWP, and includes two (of four) hemispheric only reconstructions. NH only reconstructions are likely to exaggerate temperature differences if treated as global both because the greater the area included, the more varying regional trends will average out, and because of the far greater sea surface to land surface in the SH. Second, your graphic has a finer temporal resolution, meaning short term peaks are better retained.
Consequently an accurate schematic based on your graphic would show a reduced MWP and LIA relative to that shown by Crowley, let alone that shown by Lamb.
Tom, if you look at my graph, the vertical axis is “pseudo-temperature”. There’s a reason for this nomenclature—it’s related to the “loss of variance” (descaling) of the low-frequency content of the reconstructions.
When you compare different methods against each, you have to perform a variance adjustment or you end up with different curves having temperature scales. The extent to which different algorithms “descale” during the reconstruction period is a hotly debated one, with I would say no firm agreement by any parties as to the proper way to treat it. It’s been discussed in an intelligent way on this blog too (see the Gergis posts), and IMO there is no question that any reconstruction method proposed to date descales low-frequency information.
So to compare a “real temperature” to proxy data, you would have to know what descaling factor you need to use to covert proxy reconstructed temperature into a “real” temperature axis. If we’re going to say that Lamb has overstated the warming of the MWP, that tacitly assumes we’re talking “real” temperature, not the descaled proxy temperature.
[And by the way this whole discussion of descaling is a concept that hadn’t even been recognized during the period you claim that scientists were persuaded by unstated on your part “evidence gathered in the period after 1990”.]
I’ve put on the same graph a digitized version of Zielinski (in haste so it’s noisy) together with the ensemble I showed previously , with an “eye-ball” baseline shift (yes the vertical offset is arbitrary until you pin it to real data, I matched the mean of the ensemble to Zielinski by eye, sue me) and with descaling factors of 0, 20 and 50%. No descaling is certainly overly optimistic (it would require nearly noiseless proxies for starts), 50% is probably pessimistic.
If you looked at this graph objectively, would you say the biggest problem with it is too large of a MWP or an unrealistic portrayal of post 19th century warming?
Lamb may have overstated MWP, but we really don’t know that even in 2012, let alone with the “evidenced gathered in the [unspecified] period after 1990”. What we do know is, even if Lamb “overstated it”, that it is warmer now than it was during the MWP even according to Lamb’s account, once you properly take into account 20th century warming in his figure.
I think it’s conceivable that Lamb’s estimate of actual temperature may be closer to true temperature (in terms of low frequency content) than the current status of proxy reconstructions. All of this without overturning any important dogma. The Earth is still round, we really did land on the Moon and so forth.
I don’t think anybody particularly remarked about the botched up 20th century in 1990 because there wasn’t all of the skeptical talk about it being warmer then than now, so as I said previously:
“I think [this] also illustrates the talking point that the modern warming period is warmer than the MWP hadn’t been as widely adopted yet. In post 1990 publications that included renditions of Lamb’s figure, they were still more focused on temperatures of the past, it seems, and less on how that related to recent global warming.”
And while we’re on the subject of getting things wrong, which is more truth centered, Lamb or the poster-child of the warming movement, MBH’s hockey stick?
I mean, it’s plausible that the MWP was less than portrayed by Lamb, possibly even likely..
But it’s an absolute certainty that MBH flubbed things up badly in their reconstruction.
Steve: the MBH reconstruction is really nothing more than Graybill bristlecone chronologies plus a fuzz of white/low order red noise. Bristlecones have a very unusual growth situation (both high altitude and VERY dry.) Graybill said that they weren’t a temperature proxy. It is bizarre that the climate community was so distracted by the bells and whistles that they failed to squarely address whether these chronologies could plausibly be a thermometer for the entire world.
Carrick, your observation about the MWP-LIA delta looks right to me. There has been considerable movement since IPCC 1995 to AR5 in restoring concepts of LIA and what-they-now-call Medieval Climate Anomaly, the latter term also used in Soon and Baliunas 2003. IPCC 1995 (chapter lead author Neville Nicholls, a Climategate correspondent very active in the persecution of Willie Soon) made the following statement on MWP-LIA differential based on Bradley and Jones 1993 and Hughes and Diaz 1994, both, in my opinion, very weak articles:
This position seems to me to be a retrogression from Lamb. I think that it was possible to comment on this problem with the information available at the time. However, Bradley and Jones 1993 and Hughes and Diaz 1994, and thus IPCC 1995 on this point, were wrong turns.
Carrick,
1) Taking what you say about rescaling at face value, your original claim that the Lamb graphic understates modern warming is simply nonsense. If you are allowed to rescale and rebaseline at will, essentially what you are claiming, then you can make any two points on the graphs being paired to align, and doing so gives you absolutely no information as to which is an over estimate or an underestimate.
2) Lamb and the IPCC 1990 did not use “Pseudo-temperature”. If you did not think your “pseudo-temperatures” in your graph were a reasonable approximation of actual temperatures, your making any comparison between the graphs was simply deceptive. On the other hand, if you thought they were reasonably comparable, then the comparison does not bear you out.
3) Further, contrary to your practise, not just any rescaling or rebaselining is acceptable. The early part of the Lamb graph is explicitly speculative, while the later part, including most of the LIA and the modern interval is based on an instrumental recored, albeit of just one location. Prima facie, this means the relationship in temperature between those two is much better justified than is the relationship between the MWP and the LIA. Rescaling and rebaselining to force the temperature difference between LIA and MWP to match therefore represents a determination that a speculative relationship should be given greater weight than one based on observation.
4) Your graph of Zielinski is irrelevant as it is based on IPCC 7.1(b) which is inconsistent with IPCC 7.1(c) in large part because it does have a too low modern temperature. Showing your thesis holds for a different, and inconsistent graph in no way supports your original claims. Given that you could have with equal ease digitized the Lamb/7.1(c) graphic, I have to wonder why you switched to the irrelevant Zielinski graph.
😀
Cool.
More of Tom and his endless cerebrally-deprived prattle. Football practice.
1) The discussion hinges on the fact that Lamb’s temperature series is not a weighted sum of temperature proxies and hence does not suffer from loss of scale. In order to place the digitized version of Zielinski onto the proxy ensemble, you do have to adjust for scaling. Remember we’re shifting Zielinski onto the proxy pseudo-temperature scale, and a scale factor of one is equivalent to assuming no bias in the modern reconstructions (which only happens, for among other reasons, if the signal to noise ratio is very high, an assumption we know not to be true here).
The range that I showed effectively gives an estimate of the confidence bounds of Zielinski’s curve when plotted onto the ensemble of proxy reconstruction. Seen that way, there are two real differences, one is the onset of the LIA, which probably has to do with the geographical locality of Lamb’s original series and secondly of course the flatlined post 1900 warming that you continue to pretend isn’t there. The MWP period difference? Not so important. One other point—it would take one hell of a rescaling to get the MWP hotter than say the temperatures post 2000. And of course there’s no physical argument to increase the scale on Zielinski’s curve before superimposing it on the proxy temperature ensemble. None.
2) “Lamb and the IPCC 1990 did not use “Pseudo-temperature””. Sorry this is just silly prattle. If Tom doesn’t understand how pseudo-temperature relates to (the modern reconstructions), then he clearly has no understanding of what he’s even looking at. “Pseudo-temperature” is a term of art that describes a quantity that relates to temperature, but lacks the correct scaling factor, and (as in this case) contains uncorrelated coherent variations that aren’t temperature related (if you want, it’s a multivariate index which we think relates to temperature for long-period fluctuations). Short-period variability in climate reconstructions in some of the reconstructions (likely Mann 08 EIV, which uses a lot of precipitation proxies) may be more related to precipitation or solar exposure changes rather than temperature changes, for example. Of course they didn’t understand this in 1990, but in making a comparison with a 1990 figure (or a later rendition of this) to proxies that didn’t occur until 15+ year later… frankly it’s just a silly thing to complain about. I expect more football practice on this issue.
3) Not any baselining is acceptable… I agree. I gave one version which I think has a stronger theoretical justification. The other reasonable one (matching the temperature at the end of Lamb to the temperature record at that point is noisy, but it is a reasonable alternative) only shifts Zielinski’s curve up by about 0.1°C. Neither choice changes the conclusion, which is that the MWP according to Lamb (or Zielinski’s rendition) is not warmer than current temperatures.
4) And the graph of Zielinski is not irrelevant in the slightest. Tom knowing his other three points are weak, adds a fourth even weaker one. Writing without thinking? The sum of series of arguments is no stronger than the arguments, so tacking more sloppy logic to the end of a series of poor arguments doesn’t somehow bolster his case one iota.
As a “flat earther”, Zielinski’s curve it’s as good as any other post 1990 curve. But if Tom wants to use a different curve, I suggest he digitize it, generate his own ensemble and show us his results. I can easily most digitize the figure that is sitting on my screen front of me, when I get ready to digitize it. I’ll counter claim that using the other figure won’t change the conclusions to be drawn and I am completely certain Tom know’s this which “makes me wonder” why you bring this fake issue up. 😮
Counterfactual to Tom’s bloviations, the main conclusions I gave above still hold:
A) The main defect of Lamb’s curve (and others based on it) is the post 1800 flat-lined temperature, which gives the erroneous impression that the MWP was more prominent compared to the modern warming period than it really is.
B) Inclusion of the modern warming period via e.g. instrumental records clearly shows that modern temperatures have exceeded those of Lamb using any reasonable criterion for matching up the two temperature series. (I think the proxy-to-proxy baselining is more appropriate, but you can’t account for other people’s tastes.)
C) MBH is clearly much more substantially flawed than Lamb’s treatment of the MWP, which makes the degree of infatuation of warmers like Tom even more interesting. (They are very selective about which “bad science” gets decried.)
D) It’s possible that Lamb may have overstated the amount of temperature variability (the geographic location of his data source and unadjusted for polar amplification factor is one explanation for that), but even if so, the difference is hardly something to warrant a hysterical treatise from John Mashey or the credulous defense of it by Tom Curtis that we’ve seen on this blog.
The point being is even know we really don’t know for a fact that Lamb has overstated the amount of historical variability in temperature. And even if true, nor does it upset any established dogma, namely the prominence of the modern warming period, approximate ellipsoidal shape of the Earth, the lunar landings, and so forth.
Carrick, you may recall some of the discussions we had about this at tAV when we were comparing the Loehle and Ljungqvist reconstructions. Besides the loss of variance issue with individual reconstructions, one of the problems with forced pegging of proxy reconstructions to the overlapping instrumental period is that when you start comparing reconstructions or assembling spaghetti graphs you get these ridiculous comparisons of reconstructions with obviously differing scales and/or offsets. AFAIC all comparisons between reconstructions should at least explore centering and scaling based on common overlapping relationships between the proxies or reconstructions in addition to instrumental calibration.
Layman Lurker, yes that was a very fruitful discussion on JeffId’s blog. Here’s my take on the issues brought up there and as always I’m glad for insights that improve my understanding of the issues:
One of the problems that people who are acting as advocates (tamino in this case), who just spew party line and move one, is they never come to grips with the fact that the vertical axis of the different reconstructions can vary by as much as a factor of two in scale, and that there can be a substantial offset bias during the reconstruction period relative to the supposed baseline used in the calibration period.
Briefly, we have biases that gets introduced when people select by correlation or equivalent methods. That’s by now well documented and IMO an unassailable point. And this problem is made worse when you have more noise and you are forced to make a tighter restriction on correlation.
Secondly different geographical distribution and proxy selection in general affects the overall scale (temporal response of differing proxies is an issue too that doesn’t get remedied by a weighted sum over proxies using a single weighting value for each proxy).
One major issue with geographical scaling is of course polar amplification. Distributions of proxies that don’t include tropical proxies will have a different scaling factor … by as much as a factor of four… than say those that are proxies for SST (which don’t exhibit any polar amplification… at least until you get to the Arctic Ocean.. but they will also underestimate global temperature trend, and will be more closely aligned with tropical temperature trends than non-tropic latitudes).
The issues with scale and offset bias gets even more complicated when you take into account that the actual number of proxies vary over time. Changing the number of proxies over time will in general influence the net offset bias, and you will get a low-frequency signal that tracks with number (and quality aka SNR) of proxies.
On top of all that, all of these algorithms at some level are inverse problems, and one of the qualities of inverse problems is they have a feature called noise amplification associated with them. Robustness is another issue that rears it’s head when incorrect proxies are feed into the inverse problem. Sheep Mountain and Tiljander are examples that come to mind there.
And the point is that since the available proxies changes over time, the amount of problems that individual algorithms exhibit is also time dependent. It’s a dizzying mess.
Really the best you can hope for is that the long-period portion of the reconstructions reflect temperature, and when you correlate the reconstructions against each other, that the coherent portion is related to a common temperature signal.
That’s really the philosophy behind using ensembles created from methods that use very different inversion techniques, we can cross our fingers and hope that systematic effects associated with individual reconstructions will look like noise when compared to disparate methods for reconstruction.
[By the way, we don’t have to cross our fingers if we have reasonable climate field histories that we can use as input and decent Monte Carlo methods. We can show the effect on offset and scale using different methods, and possibly remove this from a multi-reconstruction ensemble by modeling its effects.]
Anyway, that’s the basis for why I compare proxies the way I do and I don’t use words like “temperature” unqualified when I discuss the proxy index value that gets returned from various reconstruction methods….
I may get around to a blogpost in response to Carrick later this month (I have higher priorities, and other things may come up). I will make the following points:
1) Carrick says that:
“As a “flat earther”, Zielinski’s curve it’s as good as any other post 1990 curve.”
However, the issue is whether, in Carrick’s words, “… relative to the LIA minimum, Lamb gets the MWP warming about right”, and “What he clearly gets wrong, and they would have known this in 1990, is it grossly understates the modern warming period”. One of the most notable differences between the Zielinski graph used by Carrick and the Lamb based IPCC fig 7.1(c) is that the later shows warmer temperatures than the former in the twentieth century. To suggest that “any graph will do” when they differ exactly on the point at issue is absurd.
2) When baselining Zielinski against his multi-reconstruction, chooses to place the baseline of Zielinski (purportedly approximating 1900 temperatures) 0.2 C below the 1900 temperature in his ensemble mean. This convenient baseline shift halves the difference between the Zielinski MWP peak and his ensemble mean MWP peak, while changing the mid 20th century temperatures in Zielinski from above those of 1900 AD. That is, the baseline shift biases heavily in favour of his opinion. It is, however, certainly not justified. The Zielinski graph notionaly shows differences in temperature from 1900 AD, so the comparison should make align that zero point with 1900 AD in the ensemble mean.
3) Carrick assumes that the historical range of temperatures over the MWP, LIA and modern era must be greater than that shown in reconstructions. There are several reasons this may not be so, of which the most basic is that proxy reconstructions are based on a limited number of proxies, which means there variability will be greater than actual global temperatures.
To compound the problem, the resonstructions in his ensemble consist of two NH only reconstructions, one NH extratropics, and one reconstruction that weights NH relative to SH 15:3; NH extratropics to Tropics, to SH extratropics as 12:6:0; and NH Atlantic to the rest of the globe as 10:8. The large regional bias introduced by this heavy NH dominance will also bias the ensemble mean to overstating rather than understating actual change in temperatures. This is particularly the case as the SH extratropics (entirely unsampled in his ensemble) is predominantly ocean, and hence will show a much reduced change in temperature.
There are also factors which will bias the ensemble towards reduced indicated change, but there is no guarantee that that bias will be equal to, let alone greater than the biases in the other direction.
4) As Carrick notes, the temperature scale in Fig 7.1(c)is just that, ie, intended to represent real departures in temperature from the 1900 value. In contrast, those in the ensemble are pseudo-temperatures that need to be rescaled in order to correctly represent temperatures. Because it is the pseudo-temperatures that need rescaling, the rescaling must be done using the baseline of the ensemble as the zero point. If that is not done, relative differences in temperature will not be preserved after rescaling. This necessity is a simple function of the fact that (x + a) * b does not automatically equal (x * b) + a. In fact, and incorrectly, Carrick rescales around the baseline of the Zielecki graph, which results in a far greater relative reduction of MWP warmth. This error is compounded by the previous relative baselining error (2).
As it stands, Carrick’s effort would make a fine example for a post on “how to lie with graphs”.
Tom Curtis, I could care less if you blog on this. Don’t expect my participation on your advocacy = by definition dishonest SkS blog.
1) Zielinski works just fine because it is a counter to John Mashey’s ravings, and truthfully to your own claims I I invited you to make a plot of IPCC 7.2c, I noticed you weren’t up to the task. If you had done so (i have now) you’ll find that no, actually, IPCC 7.2c does not show that the MWP was warmer than the modern warming period. I of course never said that “any graph will do” because obviously different graphs say different things. I note that you accuse me of dishonesty in your post, but you can’t even honestly represent my own comments on this matter. What a complete dishonest hypocrite you are. Anyway you yourself said:
I guess it was more accurate until it wasn’t.
2) Zielinski actually doesn’t go to 1900, more like 1850. The “flatlining” starts about then. As I mentioned above, 0.1°C shift is possible, but I think not as defendable as bringing the two datasets (ensemble and Zielinski) onto a common baseline. If you don’t understand the issue about offset biases, I’d suggest you learn something about it before you embarrass yourself with a blog post. And in any case, you do is going to “fix the problem” so that the MWP from this graph is warmer than the current temperatures nor fix this graph, which you defended earlier as “more accurate”, so that it doesn’t flatline after 1850.
It’s funny seeing people like you who are so bent on being right that you won’t even admit when you and the dogma you’ve embraced have gotten it wrong. Oddly you pay more attention to skeptics comments on this (you’ve fixed belief that these graphs show the MWP to be warmer based on comments by skeptics) than you do to comments from non-skeptics pointing out that this premise is false.
3) “Carrick assumes that the historical range of temperatures over the MWP, LIA and modern era must be greater than that shown in reconstructions” This is again a inaccurate representation of my views on your part. I never made any such comment, nor as it turns out, do I assume that. I don’t, and I have no idea how you could have arrived at that conclusion based on anything I’ve said.
In fact, I said rather the opposite, my comment were to the effect is that it is likely that Zielinski overstated the warming, but that we can’t state with certainly even in 2012 that he did in fact overstate the warming.
(That is bold-faced so you won’t miss this comment this time.)
I certainly understand the issues about polar amplification as I brought it up above. What I understand that you apparently don’t is that there really isn’t an expectation that polar amplification will remain constant over different climate periods. You can’t just invent data when data don’t exist. If it were possible to extend the extratropicals to the tropicals and SH, they would have done this already.
That observation really hadn’t occurred to you before? Or are you just assuming that you’re the most clever person in the room again, and you can WAG it and nobody will notice?
Also I notice that Ferdinand’s comment went over your head. Perhaps you should read that one again, as the issue of how regional the sampling of a proxy is important (simply because you’re measuring a proxy in Greenland doesn’t mean that this represents the mean temperature in Greenland).
There aren’t guarantees that geographical bias will have either sign, though I think we would both agree that land proxy reconstructions from northern climates will tend to overestimate global temperatures. But again, what we have to compare against is the proxy reconstructions against each other, not Tom’s made up hand-waving version. Regardless, none of them (including IPCC 7.2c) show the MWP to be warmer than current temperatures, the claim made by skeptics on this is false.
I think you need to check your police work there. There are two sources of error, offset bias and scaling bias. So you first pick an algorithm for correcting the offset. Matching temps to 1850 is one, and the other is matching the means over the reconstruction period (which is likely to be more accurate). You then apply the scale bias correction.
You certainly don’t do them simultaneously. What a goofy idea.
I’m going to start by apologizing for being a bit snarly in my last post. I detest people who spend their entire post attacking the honesty of another person, as Tom Curtis is doing (typical of the tactics used on SkS), but I will keep my own responses more civilized. If he can only engage at the level of constantly accusing others of dishonesty, as he has consistently done on this thread, that speaks louder and more effectively than any comment I can make in response.
Here’s a bit more on offset versus scaling bias and defendability of different methods for matching offsets (during reconstruction periods) between different reconstructions. First a link to the relevant post on this blog by Roman. Roman mostly discussed the scaling bias introduced by screening, Nick on this blog. Nick also did some nice work see here. Lucia also has some good posts.
An offset bias is a vertical shift in scale between instrumentation and reconstruction period created by the reconstruction technique. It is something I became aware of in comparing reconstructions against each other. Nick did some Monte Carlo’ing that exposed this was a property of at least screened reconstructions.
Different SNRs, different numbers of proxies, and different reconstructions appear to produce different offset biases. I’m not sure if we’ve established whether there is a definitive sign for offset bias or not. My impression is it can have either sign (which if true is good because it means it will cancel out in an ensemble average across reconstructions). The issue of offset is a point that Tom Curtis seemed to understand at some level as he posted on it on this blog.
If you want to compare different reconstructions against each other, it is a mistake to try and align them against each other using the instrumentation period. This is effectively what Tom is advocating and on this point he is dead wrong.
A scaling bias is a change in scale between the calibration and reconstruction periods, and again different SNRs, different numbers of proxies, and different reconstructions will produce different scale biases. Unlike offset biases, scaling bias will always produce a smaller scale. There is no expectation that averaging over different reconstructions will yield a bias-free
I’ll cover the issue of geographical distributions and their effects in a separate comment.
Carrick wrote:
Carrick, this is the same flaw as Mann’s “short centered” principal components. By standardizing proxies in the instrumental period, Mann ended up mining series for HS shapes.
I noticed at the time that screening proxies by temperature correlation in a short instrumental period ended up having a very similar effect as Mannian principal components – a phenomenon independently reported (as you observe) on numerous “skeptic” blogs (CA, Jeff Id, Lucia, Lubos Motl, David Stockwell) but still unreported in academic literature (other than, perhaps, a short note by David Stockwell in an Australian mining news journal).
The Graybill bristlecone chronologies have a real HS shape, which is familiar to readers as the Mann hockey stick shape. Whether they are valid temperature proxies is a substantive issue (their originator, Donald Graybill, thought not.) This is a different issue than the statistical bias from screening or short centering. Climate scientists, for the most part, seem baffled by being confronted with two problems at tone time.
This comment focuses on the effect on scale relating to changing geographical distributions.
The effect is very complex and not as nearly straightforward to correct for as Tom Curtis suggest it is. As a simple example, let’s take tree ring proxies. In Northern climates, these are known to be seasonal proxies (their growth patterns register summer time temperature much more than winter-time ones, otherwise you wouldn’t have tree rings ;-)).
On the relationship between tree-ring proxies, polar amplification and changing geographical distributions. I believe this is not an issue that is well understood, but it is instructive to break down polar amplification by season. This is GISTEMP data. You can produce your own version of this curve using the GISTEMP online tools.
What you see from this figure is summertime there is almost no geographic effect associated with polar amplification. So not only do different proxies sample in non-regionally (with different geographic weightings) as pointed out by Ferdinand above, there is a seasonality to polar amplification and that influences the effective scaling associated with different proxies.
Tom’s seems to think that you can assume a constant polar amplification profile over time and all proxies only sample local temperature. Neither of these points are correct. If it were that easy to correct for, people would do it, instead of trying to work on improving the geographical distribution of proxies.
I also think it’s fair to say that the NH proxies don’t necessarily reflect global temperatures at least on short-enough time periods (over periods of 60 years or so, there seems to be an anti-correlated component for example). On longer time scales, we can get some information by correlating Northern Hemisphere reconstructions against ones that include SH proxies, like Loehle (I’ve been meaning to amend his reconstruction with more SH proxies to see if it makes a difference), and in general by correlating the reconstructions, all of which have their own geographic distributions, against each other.
In this plot, I show the Pearson correlation coefficient of different reconstructions against Ljungqvist, including that of Loehle against Ljungqvist, produced using a 500-year sliding window.
Notice that Loehle, which includes both Northern and Southern hemispheric contributions, correlates very well with Ljungqvist, which is strictly and extratropical NH reconstruction. This of course argues against the idea that the MWP is entirely a NH phenomenon, and further suggests that there is a high degree of coherency between NH and SH temperature variability at these very low frequencies.
What this does not suggest is that the temperature scale of any of these will be the same due to different geographical samplings, even assuming otherwise “prefect” reconstructions. Nor would one expect the offset of the temperature scale of the reconstruction period to be the same (let alone zero) between different reconstructions, when compared to global mean temperature.
In fact an offset bias (relative to global mean temperature) will be introduced by the non-uniform nature of the geographical sampling. This is a statement that is true regardless of SNR of the proxies, although SNR of the proxies will definitely affect the bias, and arises from the nature of regional climate variability itself.
I’ll address my views on what I think the “best” way to compare proxies in a third comment.
This is the third of three comments on proxy reconstructions and bias. The other two are currently in moderation–too many links–but I’ve assembled the entire set of comments here. A certain amount of editorializing is left out, and I will probably revised the document over time. (If so, I will archive older versions.)
This comment relates to what iI think the “best” way to compare proxies with each other is.
I’ve discussed there are two main biases introduced by the reconstruction methodology, one of these is the oft discussed scaling bias, the other is the less well understood offset bias. As Tom correctly points out what this means in practice is we can assume the relationship
Tgmt = alpha * Tproxy + beta
alpha is a descaling factor that depends on proxy reconstruction algorithm, SNR of the proxies and geographical sampling (including seasonal effects) on proxy response relative to global mean temperature.
beta is an offset between the proxy “pseudo-tempeature” and global mean temperature. All of the effects discussed above affect beta, but also regional scale variability in climate (the degree to which this affects the reconstruction is also influenced by what temperature series the reconstruction algorithm used, for example, did it correlate against global mean temperature, or regional scale temperature?).
If you want to compare proxies against each other, the wrong way is to assume alpha = 1 or that beta = 0 (which is effectively what Tom is doing by advocating for aligning reconstructions using global mean temperature).
Using an ensemble of reconstructions (where you “scale” correct them) is expected to reduce the influence of the temperature offset, since as I’ve discussed this generally may have either sign. If you want to see if a “new” (or old) reconstruction is in agreement, you regress against both alpha and beta between “new” proxy series and proxy ensemble, and you make the comparison on that basis.
If I get a chance, I will show versions of Zielinski and IPCC 7.2c. Actually I would have done this to start with, but I knew that Tom would not understand the point of such a transformation.
Note that none of this addresses whether the scale of the ensemble reconstruction is the same as that of the global mean temperature.
Addressing that fully would amount to a full-fledged research paper. Nonetheless, with the uncertainties in comparing reconstructed to true temperature, while it seems unlikely, it is at least possible that MWP temperatures could be similar to current global mean temperatures. I would predict they are not warmer, but we are also missing a lot of frequency content for that period, so I do not think you can categorically state even in 2012 that current temperatures are warmer than the MWP.
Unless you are an advocate. Who writes posts for an warmer advocacy site. In which case you’re engaged in rhetorical arguments. Which means you aren’t really interested in truth, merely in coming up with a plausibility argument that furthers the cause you are advocating.
Excellent comment Carrick (I believe Ed Cook expressed vaguely similar insight in one or more of the climategate emails). While spaghetti comparisons of reconstructions force fit an alignment with instrumental data with no consideration of offset uncertainties, standardized comparisons of reconstructions can give some perspective in framing the +/- uncertainties of “beta”. I think simulated data experiments showing the effect of differing SNR’s, algorithms, proxy selection methods, etc, would make a great paper.
Steve McIntyre:
I wouldn’t say climate scientists in general, but the “Team” certainly affects bafflement.
I don’t think it’s a hard concept, and most scientists grok it, that you can have more than one error in a paper, and that until you clean up all of the problems, the answer you’ll get is wrong.
It’s claimed correcting MBH to use a centered PCA “doesn’t matter”, of course it matters or you’d get an identical answer. You get a different, equally wrong answer. Until you’ve corrected all of the mistakes (or enough of them for the the answer to start converging), your answer will always be nonsense.
It isn’t a test of a method that it returns noise. That should be a point of concern, not a point of praise.
Layman Lurker:
I wouldn’t be surprised if most people in the field understand most of this. I’m not sure they get the relationship between the offset bias and proxy network or I think there’d be a paper on that already.
Interesting paper from Christensen and Ljungqvist..
It may relate to this comment I made:
Steve: it looks like another screen by correlation paper a la Gergis. I have corresponded with Ljungqvist from time to time and have data for this.
The talking point crowd(s) will have fun with one.
For what it’s worth, I discussed that Christiansen and Ljungqvist (2012) paper a few months ago over at The Blackboard. I even wound up exchanging a few e-mails with the authors. I notified them of an error where they used a series (#25) with insufficient resolution for their 1500 year reconstruction. The series has sufficient resolution for a few hundred years, but the earlier data is at a lower resolution, and thus it shouldn’t have been used. Interestingly, that series has the highest correlation of all their series.
I also tried to get clarification on their decision to use the Dulan series, a precipitation series, as a temperature proxy. The response was basically that one author had discussed the matter with Chinese scholars other than the ones who created the series, and he was told it could be interpreted either way and the region has a strong correlation between temperature and precipitation. I tried to pursue the matter further, but after not getting a response to an e-mail, I lost interest and dropped it.
Long story short, the paper has issues though I don’t know how much impact those issues have. What I do know is at least five of the 16 series they used in their millennial reconstruction are questionable, at best. That includes one series which doesn’t meet their criteria for inclusion, as well as the Avam-Taimyr, Yang, Dulan and Yamal series.
I have a comment in moderation discussing the paper Carrick linked to above. Long story short, the paper has issues.
Steve McIntyre:
As I said, we’ll have to look at the paper and judge it on its merits. I know Christiansen knows about the variance loss, he’s debated this on different blogs, including Lucia’s blog, Jeff Id’s blog where his correspondence with Jeff was reproduced, including this:
See Eduardo’s post here too.
Steve: Variance loss and bias from correlation-screening are different issues.
WUWT has a post up this morning regarding a new Esper et al paper. Here is a money quote from the abstract that falls into the category of “speak of the devil” for this discussion:
Steve: This is a companion paper to the Esper et al 2012 paper in Nature last summer. I found the Esper papers interesting for different reasons than WUWT. There were a couple of points in it that I hadn’t thought about it. The Nature paper made a nice sensitivity discussion of Kaufman et al 2009 that was very much in CA style – but it was something that I hadnt noticed and rather smiled at them pointing it out.
I imagine this is why John Mashey is important enough to have a wiki entry..
yet William Connolley has just deleted another Marcel Leroux from Wiki.
William, has this to say about Marcel Leroux’s work..
http://scienceblogs.com/stoat/2012/10/04/death-of-a-salesman/
(comment by Jan Morten, defening Leroux, a little Oct 9th, 5.02, WC responds
[I haven’t read his books, or any of his work, but what you say fits roughly with what I’ve seen: he had some ideas, mostly in a few limited areas, that might be good or not. But his “contributions” to GW were not useful -W]
–
fairly typical I suppose, delete all other voices, especailly one you haven’t even bothered to read
At 0433hrs your time, I posted a comment that has not yet appeared.
This may be because, to get through moderation on this website, one needs to provide links to peer-reviewed work that refutes the argument being made. In this instance, this would require me to cite peer-reviewed work that suggests that the MWP and LIA have not been made to disappear, were just regional events, and/or that it is now warmer than the MWP…
However, whilst I am certain such work exists, I suspect that links to them posted here would be even more short-lived than a Higgs-Boson particle in the Hadron Collider at CERN. So, then, is this comment not going to appear either?
Steve: this post is about what people thought in the immediate aftermath of IPCC 1990. Assessing the LIA and MWP is a much larger topic and requires discussion of other materials not discussed in this thread. Editorial policy here is to avoid trying to discuss everything all at once on every thread.
In view of 329 consecutive months of above 20th Century average temperatures, I must confess I find this entire debate about the MWP and LIA to be completely anachronistic. As a statistician, Steve, surely you must concede that this could not happen unless the Earth was warming-up? If so, since no Sun-related factor can explain the multi-decadal warming trend of the last 100-200 years, what is your alternative explanation?
Steve: this post is about what people thought in the immediate aftermath of IPCC 1990. Editorial policy here is to avoid trying to discuss everything all at once on every thread.
I don’t much care what people thought 22 years ago. I think it is much more important what they think now. e.g. http://climatecrocks.com/2012/10/10/extremes-driving-voter-shift-on-climate/
Steve: I’m interested in archaeology as well. It’s too bad that you’re so uninterested in past history – it’s really very interesting.
Re: Martin Lack (Oct 10 10:56),
Martin, your comment/statement is OT because there’s not a lot of argument over whether there is warming in general today. The question is whether the warming is unusual. You must understand that to understand that MWP/LIA are important topics.
I would refer you to my earlier comment (0934hrs) in which I mentioned the Higgs-Boson and the Hadron Collider.
I would agree, Steve, that learning from the past is important. However, given all that we have already learned from the Earth’s 4.6 billion-year history, I think it would be far more useful if people would focus on solving the problem (i.e. that we have far more fossil fuel on this planet than it would now be safe to burn).
http://lackofenvironment.wordpress.com/2012/10/10/saved-by-the-sun/
Ooh! Yet another OT comment by our most recent arrogant troll demonstrating his deep knowledge of statistics (or lack thereof)!
So you think that the monthly temperatures form a sequence of independent observations and a simple-minded evaluation of the probability of an occurrence such as the one you describe provides such overwhelming corroboration of equal simple-mindedly understood beliefs. The behavior of the temperature is closer to that of a random walk where the next temperature will result as a relatively smaller increase or decrease from the current position.
Perhaps you did not run across the Arcsine Law for random walks. It states that the least likely a random walk will do is to spend half of its time above and half of its time below the starting point of the walk. The most likely is that it will always stay above the starting point or always stay below the starting point with equal probability. In this case, observing long seuences of consecutive observations on the same side of the start is not unlikely at all, but is more the norm.
However, you already knew that , didn’t you? 😉
You use many long words; but say very little. If you hoped to impress; please embrace disappointment. You failed to answer my (OT) questions; so don’t expect me to answer yours.
I take it then that you haven’t a clue about what I wrote and that you still don’t understand what I indicated to be scientifically nonsensical with your original claim of “unusual” for the sequence of monthly temperatures.
Spoken like a true ClimateScientistTM.
I am just complying with Steve’s request not to turn every post into a discussion of all issues. Given this stipulation, you (or at least he) cannot berate me for supposedly dodging issues (and/or not answering questions).
RomanM,
Just reading a peer reviewed climate paper, despite not being “a true ClimateScientistTM,” means you wear a black hat in Martin’s world. On a different thread, Martin quoted some spin from SkS on Dessler 2010. I asked him why he did not simply read the paper, since the conclusion did not really support the spin. He responded:
“I am not interested in reading Dessler’s paper because I am not a climate scientist. I am content that climate scientists who support the consensus do so because they are genuinely concerned about the decades we have wasted arguing about whether or not anthropogenic global warming is a problem (i.e. about 5 decades so far). I am equally convinced that those few climate scientists (like Lindzen) who dispute the consesnus understanding of the problem, do so for reasons that have nothing whatsoever to do with science.”
The simple fact that you and Steve are better statisticians than the Hockey Team is enough to make his head explode. Stand clear so it does not get on your shoes.
I ran into a problem related to the Arcsine Law while working with a telecom company. Our system reconciled the total monthly usage as recorded by the billing system and the technical logs. If the difference was above a certain threshold, this would trigger an alert. The billing system rounded off to the nearest second, whereas the technical logs had a few decimal places.
With increasing usage, the threshold would have to be raised from time to time to allow for the increased expected error. In hindsight, we probably should have included a sqrt(n) term in determining the threshold. Oh the joys of red noise.
“Steve: this post is about what people thought in the immediate aftermath of IPCC 1990. Editorial policy here is to avoid trying to discuss everything all at once on every thread.”
Which would be why Steve mentions Bristlecone Pines every time he feels he is doing poorly on the actual subject under discussion.
The Team substituted Mann’s bristlecones for Lamb’s CET.
Simple as that.
The argument that the MWP may not have been global was a red herring introduced by the IPCC. It was designed to offset the significance of the MWP that contradicted their claim that the late 20th century was the warmest period in history. David Deming’s claim about receiving an email from a significant person that starkly stated, “We must get rid of the MWP” is believable in light of the mood and thinking about “the cause” displayed in the leaked emails. Besides he reported the information in his presentation to the US Senate.
http://epw.senate.gov/hearing_statements.cfm?id=266543
I realize many have given false information before such hearings, however, it was a very serious charge and one that would elicit much scrutiny. (There is extensive discussion elsewhere on this site about the originator of the email.)
I spoke at length with Lamb personally at the CRU when working on my doctoral thesis about the MWP and the LIA. This was because my work involved two records of instrumental and weather diaries that spanned from the nadir of the LIA up to the modern instrumental era. Among many other issues the discussion involved determining the approximate beginning and end of both periods.
Lamb’s view was that the MWP was global, but the start and finish times varied considerably based upon local conditions and the sensitivity of the proxy being used. It was also part of the discussion we had about relative homogeneity, that is the need to produce long records for two stations far enough apart to be able to separate local changes from those that were regional, hemispheric or even global. Mine were about 200 km apart.
The MWP not being “global” is indeed a red herring. In past epochs when the arctic was ice free (and thus many degrees warmer) the tropics were only a few degrees warmer. Climate models indeed predict that warming will be mainly at higher latitudes. At the MWP the slight warming in the north has been observed across Canada, Greenland, Europe, Japan, and Siberia. Pretty much covers it. Until very recently there has been very little data from the S. hemisphere, so that can’t be evaluated, and short-term the arctic and antarctic exhibit a see-saw. If there was less warming in the tropics at the MWP it would probably not even be detectable over noise.
One study using Indo-Pacific Warm Pool cores stated the following:
“The Little Ice Age, which occurred around A.D. 1700, was a cool period, but its magnitude was only about 0.5 to 1˚C cooler than modern winter temperatures. Water temperature during the late Medieval Warm Period, between about A.D. 1000 to 1250, was within error of modern annual sea surface temperatures.”
http://www.whoi.edu/page.do?pid=7545&tid=441&cid=92918&ct=61&article=59106
I realize that this is not about what the circa 1995 concensus was, so no worry’s if snipped.
Craig Loehle, the facts of Arctic amplification are as true today as in the MWP. Therefore if you are going to take the northern latitude proxies as being indicative of MWP warmth, you must take the instrumental record of similarly northerly latitude for comparison. Good luck with that.
Tom Curtis wrote:
“Therefore if you are going to take the northern latitude proxies as being indicative of MWP warmth, you must take the instrumental record of similarly northerly latitude for comparison.”
Nonsense. All you have to do is standardize by age and then compare ring width and latewood density between the two eras. What difference does it make whether you can calibrate to temperature when you are comparing ring widths from different eras?
Tom Curtis:
“[I]f you are going to take the northern latitude proxies as being indicative of MWP warmth, you must take the instrumental record of similarly northerly latitude for comparison.”
It seems that the IPCC did precisely that when writing in FAR WG1 Section 10.0: “For example, the boreal forests of Canada extend [sic] well north of the current timber line during the Medieval Warm Epoch (800 to 1200 AD), a time when temperatures in that region were about 1°C warmer than today’s.”
Tim Ball, the correct attitude to Demings comment if he does not present the email is to regard it as what it is – hearsay. It is, IMO, very informative that virtually all comentors on this site, including Steve, are happy to condemn people on the basis of hearsay.
If he received the e-mail, it is hardly hearsay. You may not believe him, but he has stated what he was direct witness too.
By the way, you can be convicted on hearsay, provided it is reliable enough. I think that if it is good enough for a conviction, we can at least weigh the evidence (and I note that people at CA are not taking it for granted IMO).
Steve: I have a post in moderation in reply Tim Ball’s post. Yea or nay, please.
Tom Curtis:
The correct attitude according to whom? You can call it “unsubstantiated”, but it’s not hearsay, regardless of whether or not he produces the email.
I find it very interesting that somebody who has such basic problems with simple English words would invest such an enormous amount of energy in defense of bizarre and implausible propositions, like that of john Mashey’s.
Hearsay means, “Unverified information heard or received from another; rumor.” The information that Deming received an email, as he purports, is certainly unverified for all but Deming, and “received from another” for all but Deming. In fact, it is difficult to distinguish, “I received an email from x, that said y” from paradigm examplar of hearsay, “I heard x say y”.
The simple fact is that until Deming either produces the email, or his account is corroborated by the person who sent the email, or somebody to whom it was copied, his account is hearsay.
If you want to take it more legally,
“The hearsay rule does not apply to representations in an e-mail, fax, lettergram or telex but only so far as the representation is as to the identity of the sender, the date the message was sent, and the identity of the addressee”
http://www.lawlink.nsw.gov.au/lawlink/pdo/ll_pdo.nsf/pages/pdo_ruleagainsthearsay
That is, Deming can state the name of the person who sent him the email, the date he received it, and the adressees of the email, and all of that is not hearsay evidence. But if he wishes to establish the contents of the email, he must present the email. Beyond that his allegation is hearsay.
He claimed he got an email. That’s not hearsay, that amounts to a “personal account”.
The contents could contain hearsay, but him producing the email won’t remedy that.
Re: Carrick (Oct 10 22:26),
In the U.S. (where I think Deming made his statement), hearsay is an out of court statement offered for the truth of the matter asserted therein. It’s an evidentiary concept that doesn’t really fit or have any application here. Hearsay can be admissible in court, so the idea that hearsay is shorthand for “worthless information” is not accurate. There are several reasons why Deming’s statement would not be hearsay anyway (to show bias, motive, intent, knowledge).
A better objection for Tom Curtis to raise in the courtroom would be under the “best evidence rule,” which generally states that one should not be allowed to describe the content of a document when the document can speak for itself; the document is the best evidence, not someone’s description of it.
Think of all the instances in every day life where people make decisions (some mundane, some important) based on someone else’s description of information, e.g., advice from your stock broker, information in a scientific journal, etc. To say such information is unreliable and should not be considered is quite remarkable.
I interpreted the intended substance of Tom Curtis’s comment as meaning he personally will believe the content of Deming’s email when he sees it because Deming’s description just isn’t good enough.
Tom.
I suggest you remember the standard of proof you just established.
I suggest you remember it when people claim they get death threats vial email. I suggest you remember it when we question some of the “investigations” that have cleared mann and others.
Lack of documentary evidence is a problem for the Deming claim. True. But if you use that same rule on other cases I’m sure you see some resist it
Unscientific Lawyer:
Thank you, that was very precisely put!
Steve Mosher, I do:
http://www.abc.net.au/mediawatch/transcripts/s3507732.htm
In contrast, “skeptics” seem to be happy to ridicule the notion that death threats existed because a location of which it was claimed (in the original report, not the distortions) that threats of violence were made in person was found to have no death threats by email in a period one year after the threats were made in person.
Meanwhile they ignore the published excerpts of emails as, apparently, not being evidence of death threats at all.
Click to access 1216_canberra2.pdf
But if somebody should say something derogatory about climate scientists ten years after the supposed event and with no documentary evidence, that has to be believed without question.
…. and then there are people who just hand-wave at Google and call it evidence:
Mooloo, get with the program. I never said death threats were made in person, only that threats of violence were made (in person, and only at ANU that I am aware of). Death threats were made by email to other people at other locations, one of which was reproduced by the Canberra Times. The other emails reproduced by the Canberra Times are abusive and indicate that the senders wished the climate scientists were dead. That is, they are hate mail rather than death threats.
I hope Tom is not trying to resurrect the old meme:
http://joannenova.com.au/2012/05/pathological-exaggerators-caught-on-death-threats-how-11-rude-emails-became-a-media-blitz/
I wonder how so many people can claim anything based on virtually no data for the Southern hemisphere. The MWP cannot be stated to be global or local based just on the data. It’s all hearsay!
Incidentally I understand Deming says it was Overpeck and Overpeck when asked doesn’t deny the incident. [Steve: there has been blog speculation that it was Overpeck, but Deming has not said so. ] It may not be his exact words but the climategate emails happily tell all who read them that there was indeed an active attempt to get rid of the MWP in the IPCC reports. It doesn’t seem to have been taken seriously by the rest of the climate science world though, as they continued to write papers assuming there still was a widespread MWP. And now even Mann does the same, principally to attempt to link the MWP with hurricanes; so it is all truly academic.
In summary; the MWP was prominent in the Northern hemisphere but beyond that everyone is guessing. Anyone who disputes that is not following any consensus but giving his/her opinion.
snip – this is about the history
Here’s a link to the latest study of both the MWP (MCA in newspeak) and the LIA in the Andes:
http://www.clim-past-discuss.net/8/4295/2012/cpd-8-4295-2012.html
So, no data from the Southern hemisphere? Is the AGW-camp trying to redefine the meaning of “data”, in addition to redefining the meaning of “peer review”?
Interesting results Pethefin. Looks like a case of global warming causes more extreme weather until it doesn’t (note their comment about the South American Monsoon Index being the lowest in a millennium), every 30-year trend is a long-term secular trend and so forth.
For some reason a previous comment of mine that contained a link to a visual overview of research on the MWP all over the world got stuck in moderation, so here’s something that should put an end to the “no MWP outside the Northern hemisprere” meme:
http://www.co2science.org/data/timemap/mwpmap.html
The map contains links to short descriptions of the actual studies…
@Tom Curtis Posted Oct 11, 2012 at 5:01 PM
Risking adding to an OT here, but I remember following that issue being genuinely interested at the time about just what was going on – and with the broader Australian climate psyche – there, and I have to say your woolly response here reminds me how that issue allowed the uncertainty principle to reign by an impressive technique of switching away the emphasis on which part of evidence is supposedly crucial at any point – the time, medium or location – whenever clarification was risked in any of those realms 😉
Very instructive. You are Australian aren’t you?
Tom.
I think you may have missed my point.
You rightly request some sort of documentary evidence for deming’s claim. Absent that you are justifiably skeptical of it.
I am fine with that and would join you in your skepticism.
I make the following points however.
1. When people have made claims of death threats and I have requested evidence of that, documentary evidence, some people complain about my requests. I assume you will join me in my skepticism. As with demings claim, if somebody has a claim about death threats they can be expected to show some evidence.
2. I would expect investigations to also collect documentary evidence where it is available and not to take a witness’s word on a matter that could be supported with the right documents. I would also expect them to keep transcripts of their questions.
When an investiation reports that they asked professor so and so a question, I am skeptical, and rightly so, if there is no transcript to back up what was said.
So lets just call this Tom Curtis’ rule. The Deming case gives us a clear example. We are justified in asking for documentary evidence ( like CRU confidentiality agreements, like investigation transcripts, etc) and we are justifiably skeptical if these documents are not produced.
Steven Mosher,
I’d like to propose a corollary to that which is, if the evidence is produced and it isn’t confirmatory, we are allowed to conclude that the original claim was wrong.
We are also allowed, like with the emails that were sent to the researchers, to react in approbation to the contents of the emails as being otherwise abusive, while simultaneously having to say “they were death threats.”
For example Death threats debunked
And UEA death threats published
You can read the comments on this later thread, and see that many people who didn’t think the emails qualified as death threats, still thought they were awful. (Yes you can find people with Asperger’s Syndrome among the skeptics community that think the emails are OK, just as you can find warmers who think skeptics should be rounded up and killed. /slightexaggeration)
Get my verbiage right here, I meant to say:
Carrick, had your read my original link, you would have seen that the Australian got the facts wrong because it was in pursuit of an agenda.
Mosher,
you are perfectly right to expect documentation of a claim unless there is additional supporting evidence. For instance, it seems probable that the ANU scientists were threatened on site, because the University took the trouble of shifting their location, and providing security guards for their new location (which they did not publicize at the time). In the case of the CRU non-disclosure agreements, if the original source of the data sells the data commercially, then the existence of a non-disclosure agreement with the CRU requires no further evidence. If, on the other hand, they provide it to all comers, the claim of a non-disclosure agreement would be highly suspect and I would require sight of the documentation to assuage my skepticism.
Of course, in the later case the willingness or unwillingness of CRU to disclose the data, is irrelevant as any person attempting replication can simply obtain it from the original source.
@Tom Curtis Posted Oct 13, 2012 at 7:05 PM
Moving of sites and hiring of security guards in Universities are also the sort of thing that can happen when money is available or the legislation requires it or numerous other reasons. Must happen all the time in campuses everywhere. I may read the wrong papers but never hear of it either. Putting on the purchase request at the time “Reason: Death threats from deniers” may “seem probable” evidence for you, but you must know others would still be in their rights not to be persuaded even if they saw that? Don’t you?
For instance, it seems probable that the ANU scientists were threatened on site,
You were claiming a short while back that we can’t take anything on hearsay. Yet you are prepared to speculate! So do we accept your speculation, or do you have evidence?
I don’t believe any person ever made a serious death threat in person. Tracking people down like that is very uncommon because it is difficult to do it anonymously. A death threat in person is extremely serious. The police would have been called at once.
Tom you also got it wrong because you too are clearly pursuing an agenda. Many skeptics clearly condemned the hate speech as do I, even if we don’t agree that the majority meets the legal standard for death threats.
This is an important distinction, because saying hateful things is vile, but it doesn’t land you in jail. Making death threats can lead to you losing your freedom.
So yes, it is clearly a distinction with a difference.
Carrick, I did not get it wrong. I never said the threats made in person were death threats. Quite the opposite, I clearly indicated that they were only threats of violence.
I even carefully distinguished between the actual death threat published by the Canberra Times and the other emails that were merely hate mail.
You on the other hand are happy to quote the long debunked rubbish from The Australian and get the facts wrong.
Tom
“Mosher,
you are perfectly right to expect documentation of a claim unless there is additional supporting evidence. For instance, it seems probable that the ANU scientists were threatened on site, because the University took the trouble of shifting their location, and providing security guards for their new location (which they did not publicize at the time).
##############
I don’t see the move as additional evidence of anything except that a claim was made. It is not evidence about the claim, it is evidence that a claim was made.
“In the case of the CRU non-disclosure agreements, if the original source of the data sells the data commercially, then the existence of a non-disclosure agreement with the CRU requires no further evidence. If, on the other hand, they provide it to all comers, the claim of a non-disclosure agreement would be highly suspect and I would require sight of the documentation to assuage my skepticism.”
Well then you need to actually review the case. Since the data in question was supplied to peter webster and we had evidence of that the claim of non disclosure was suspect. Second, when we asked for a list of those countries which had non disclosures so we could approach them, we were rebuffed. Third, when I consulted the CRU guidelines about entering into non disclosures and found that CRU was not allowed to unless the data was NECESSARY TO THE MISSION, I considered that as additional evidence that there was no agreement.
Fourth, when they denied a request on a false basis ( they argued a use restriction that made no legal sense given my years of experience with licencing ) I considered that as additional evidence.
The final argument of course is that we were right to be skeptical and the ICO agreed. game set match.
“Of course, in the later case the willingness or unwillingness of CRU to disclose the data, is irrelevant as any person attempting replication can simply obtain it from the original source.”
Wrong. The goal was not to replicate the study. My goal was to AUDIT the work. To audit the work I would need both the copy provided by the source to ME and CRU’s copy of the “same” data. You don’t understand this work. The first order of business is to get a list from CRU of all the stations they used. They fought that. They next stage would be to get their copies of the data AS USED. The last step would be to get the data from the original source and compare. basic quality checks. You can stop loosing arguments by simply agreeing with the ICO.
Steve, broken link above in post for the Thompson Webb 1991 (there seem to be a couple of extraneous spaces in the URL) in the hyperlink for the Thompson Webb 1991. I got to it with this URL after I removed a couple of spaces (thanks for posting this item!):
http://www.climateaudit.info/pdf/history/Thompson Webb 1991.pdf
I note that Webb’s exhortation in the final paragraph seems particularly important and also ironic in light of the fact that Mann & co. have produced such dubious millenium reconstructions. I don’t doubt the Obstacles involved in taking up Webb’s 1991 challenge, but can paleoclimatologists really claim that Mann & co. have attained “precise SST estimates for the past 1000 years”??
[emphasis added]:
still trouble with that Thompson Webb link, trying again:
Thompson Webb 1991
Steve,
Love your work, but would it be possible to write a summary at the beginning? This post seems hugely important to me, but starts off being about some conspiracy theorist commenting on a website that I have never read. 4 lines of summary would be fantastic
Al: Steve encourages people to search the archives if they need to get up to speed on an issue. In other words, it’s the readers’ responsibility. The first post on Lewandowsky is below and there were probably a half a dozen more after that.
Steve, I am re-posting the following part of my earlier comment (which has still inexplicably not appeared) because it is not off-topic. However, just in case you have filter on incoming comments, a certain word has been redacted:
Does it not strike you as even the tiniest bit ironic for you to describe John Mashey as a [C-word] theorist? What is the continuing disputation of the reality, reliability and reasonableness of the scientific consensus; if it is not [C-word] theory?
Steve: the basis of disputing the Mann-style proxy reconstructions is because of poor statistical methodology and poor data. Things like upside-down Tiljander, unreported verification r2 of ~0. Precisely why the community doesn’t object to these practices is a sociological puzzle, but I leave the interpretation of that to others.
Thanks for embedding your response into my comment, Steve. However, this raises more questions than it answers: If Mann et al have made basic errors and deduced unwarranted conclusions (as indeed Anthony Watts would appear to have done recently), can you please explain to me why Mann’s work has been replicated by others using all kinds of alternative proxies; and/or why Mann has been investigated and found innocent of any charge of scientific dishonesty, fraud, incompetence, or malpractice?
Assuming you do not think they are all deceitful or just perpetuating their research funding, do you consider that every climate scientist that says recent warming is unprecedented is just seeing what they want to see in the data
-snip – let’s stick to one thought, please
Steve: first of all, Anthony’s paper has nothing to do with Mann’s paper. Second, I’ve never precluded the possibility that someone might show that the 20th century s warmer than the MWP using data and statistics fit for the purpose. If someone did that, that doesn’t prove that Mann did it. (A flawed mathematical proof doesn’t become “right’ because someone else establishes the theorem without flaws.) Now I happen to think that there are major defects in the “independent” reconstructions e.g. continued dependence on Graybill bristlecones; data mining and other problems. But those are different issues.
Whether Mann’s results are statistically flawed is an entirely different issue than whether he committed academic misconduct. The former is possible without the latter. I am not aware of any actual investigations into whether Mann committed academic misconduct. Penn State elected not to hand the principal issues over to the investigation committee without interviewing critics. On the one issue considered by the investigation committee (bizarrely whether we requested an excel spreadsheet), Mann gave false evidence, but this was accepted by the Investigation Committee without taking rebuttal evidence.
Hi Guys,
Really enjoyed the thread.
Perhaps tangential, but don’t the Climategate emails on the MWP demonstrate the integrity of the climate science community rather than the reverse ?
From “Dealing A Mortal Blow To The MWP” on this blog (https://climateaudit.org/2010/04/08/dealing-a-mortal-blow-to-the-mwp/) the documentary record of Climategate MWP emails between Overpeck and Briffa demonstrates that climate scientists reacted negatively to any hint that they should conceal or manipulate data and that this commitment to impartiality was forcefully re-stated by Overpeck.
The exchange reads like this:
Overpeck: I wish to deal a mortal blow to misuse of the MWP.
Briffa: Are you asking us to cheat ? I find that disturbing.
Overpeck: Absolutely not.
But surely if Climategate proves deceit the exchange should have said:
Overpeck: I wish to deal a mortal blow to the MWP.
Briffa: Great. Let’s cook the data.
Overpeck: Yeah. Go for it. Let’s all get rich.
The Climategate trove is meant to be an open window on the venality of the climat eresearch culture. But on the MWP they demonstrate integrity.
2 Trackbacks
[…] https://climateaudit.org/2012/10/09/the-afterlife-of-ipcc-1990-figure-7-1/#more-16892 […]
[…] includes a intermediate version of the graph in Tickell 1986—and further discussion here and here. I am grateful for the feedback that lead to these changes as posted below and on Bishop Hill. […]