Rahmstorf’s Third Trick

Rahmstorf et al 2015 Figure 5 shows a coral d15N series from offshore Nova Scotia (see left panel below). The corresponding plot from the source is shown on the right.  Original captions for both follow.  There’s enough information in the figures and captions to figure out Rahmstorf’s next trick. See if you can figure it out before looking at my explanation below the fold.

mann-rahmstorf-temp-proxies-fig5

N15

Figure 1. Left – Rahmstorf et al Figure 5. Original caption: Figure 5 A compilation of different indicators for Atlantic ocean circulation. The blue curve shows our temperature-based AMOC index also shown in Fig. 3b. The dark red curve shows the same index based on NASA GISS temperature data-48 (scale on left). The green curve with uncertainty range shows coral proxy data – 25 (scale on right). The data are decadally smoothed. Orange dots show the analyses of data from hydrographic sections across the Atlantic at 25 N, where a 1 K change in the AMOC index corresponds to a 2.3 Sv change in AMOC transport, as in Fig. 2 based on the model simulation. Other estimates from oceanographic data similarly suggest relatively strong AMOC in the 1950s and 1960s, weak AMOC in the 1970s and 1980s and stronger again in the 1990s (refs 41,51). Right – Sherwood et al 2011 Figure 3 excerpt. Original caption: time series … annual mean bulk d15N from six colonies of the deep-sea gorgonian P. resedaeformis. Shaded areas represent 95% confidence intervals around annual means. Dashed lines indicate long-term trends, where significant. Note the cold periods (blue bars) of the 1930s/1940s and 1960s and sustained warm period (red bar) since 1970. Bulk d15N is most strongly correlated with NAO at a lag of 4 years (r= -0.19) and with temperature at a lag of 3 years (r=-0.27, p<0.05). … Squares in bulk d15N plot show values of the eight individual samples used for d15N-AA analysis.

The Trick

Rahmstorf’s Figure 5 shows a positive correlation between temperature and coral d15N (both decline together), while Sherwood et al reported a negative correlation (r=-0.27) between temperature and coral d15N.  How’d they do that?

Sherwood et al 2011
Sherwood et al 2011 stated that coral d15N was negatively correlated to gridcell temperature. Their figure 3 (a larger excerpt is shown on left) shows gridcell temperature increasing over the 20th century, while coral dN15 is going down.  Sherwood et al 2011 attributed the negative correlation to higher d15N values in the cold Labrador Slope Water (LSW) relative to the warm Western Slope Water (WSW):

Differences in nutrient processing histories underlie the distinct end-member d15N_NO3 signatures between WSW and LSW. The d15N_NO3 of WSW ( 5.02+-0.27% ) is identical to the nominal mean ocean value of 5% (26), reflecting its origin in the nutrient-rich waters beneath the Gulf Stream (27,28). The d15N_NO3 of LSW (6.00+-0.28%) reflects the low nutrient surface currents from which it is formed in the northern Labrador Sea. Because LSW is depleted in nutrients and not replenished by upwelling, the residual nitrate is enriched over that of WSW by 1%. These end-member d15N_NO3 signatures reflect basin-scale circulation and likely did not vary significantly over the late Holocene

Sherwood et al also observed d15N values similar to the early 20th century in corals dated several centuries older.

On the right, I’ve digitized the data in the Sherwood figure and re-scaled and re-oriented the d15N data to fit to the temperature data: see inverted d15N scale on the right axis. This is opposite to the orientation shown in the Rahmstorf figure.

sherwood_2011_figure3_plus

sherwood_fig3_annotated

Figure 2. Left – excerpt from Sherwood et al 2011 Figure 3, showing gridcell temperature (as per Sherwood et al reference); right panel: digitized version of Sherwood 2011 Figure 3, with d15N shown in inverted scale.

Rahmstorf et al 2015

Rahmstorf et al orient d15N to temperature in an opposite orientation to Sherwood et al, justifying their opposite orientation as follows:

Figure 5 illustrates corroborating evidence in support of a twentieth-century AMOC weakening…The green curve denotes oceanic nitrogen-15 proxy data from corals of the US north-east coast from ref. 25. These annually resolved d15N data represent a tracer for water mass changes in the region, where high values are characteristic of the presence of Labrador Slope Water. The time evolution of the d15N tracer agrees well with that of our AMOC index (Fig. 5). Ref. 25 reports four more data points from ancient corals preceding the twentieth century, the oldest one from AD500. These lie all above 10.5, providing (albeit limited) evidence that the downward excursion to values below 10 between 1975 and 1995 and the corresponding watermass change may be unprecedented in several centuries.

Discussion
In their abstract, Rahmstorf says that offshore Nova Scotia coral d15N values are one of the “consistent” lines of evidence supporting a cooler subpolar gyre and, in turn, reduced Atlantic Meridional Overturning Circulation:

Here we present multiple lines of evidence suggesting that this cooling may be due to a reduction in the AMOC over the twentieth century and particularly after 1970. Since 1990 the AMOC seems to have partly recovered. This time evolution is consistently suggested by an AMOC index based on sea surface temperatures, by the hemispheric temperature difference, by coral-based proxies and by oceanic measurements.

[Apr 7 – The Nova Scotia coldwater coral series are, like the alkenone series offshore east coast North America discussed at CA last month, located near the front between the Labrador Current and Gulf Stream, as shown in the location map from Sachs et al 2007, updated below to show the location of the Nova Scotia coldwater corals.

sachs_2007_map_annotated_2

In that earlier post, I reported that very large SST decreases had been estimated off the east coast during the Holocene, considerably larger than the North Atlantic as a whle. I noted the following explanation from Julian Sachs in terms of changing ocean currents:

Sachs observed that a relatively small coastward displacement of the Gulf Stream could account for the difference and plausibly speculated that the Gulf Stream hugged the East Coast much more closely in the mid-Holocene.

-Apr 7]

But all that the coral d15N series show is (at most) that there has been increased [Apr 7 – mixing proportion of lower d15N Gulf Stream waters, which could account for] increased temperatures offshore Nova Scotia) – but the significance of this relationship is very tenuous.  On its face, the d15N series does not show that subpolar gyre temperatures have decreased [the definition of Rahmstorf’s AMOC index].

Rahmstorf chose the right-axis d15N scale and orientation so that smoothed versions of each trending series will have similar scales.  But this could be done with any smoothed and trending series.   This is tautological mathematics, not science of deep insight.   One could equally use coral-14 values from the Red Sea or the Dow Jones Index.    Rahmstorf’s third trick.

The idea that coldwater corals offshore Nova Scotia can be thermometers for ocean temperature in the subpolar gyre has little more plausibility than the belief that stripbark bristlecones in the distant Sierra Nevadas or contaminated Finnish sediments can be thermometers for the subpolar gyre.

It’s not even well established that coral d15N is a proxy for local ocean temperature [Apr 7 – or mixing of water masses]. Coral d15N is not a well-studied proxy, to say the least. Only a few examples are reported with results from only one article being archived at NOAA. In one of the original articles on the proxy, co-authored by the lead author of Rahmstorf’s citation, d15N values were thought to depend on distance from sewage source:

While coral reefs decline, scientists argue, and effective strategies to manage land-based pollution lag behind the extent of the problem. There is need for objective, cost-effective, assessment methods. The measurement of stable nitrogen isotope ratios, d15N, in tissues of reef organisms shows promise as an indicator of sewage stress.

Alkenone Proxies

To the extent that climate scientists believe that offshore East Coast d15N is relevant to the interpretation of the subpolar gyre, you’d think that they would have canvassed information from alkenones – an important SST proxy with a number of samples offshore East Coast.  Curiously, on January 21, 2015, a week before acceptance of Rahmstorf et al, I’d looked at East Coast alkenone proxies, including some high-resolution data, in a CA post here. These comparisons seem incomparably more informative than Rahmstorf’s meager coral d15N data and I’ll discuss it in a forthcoming post.

Update: Coral d15N values offshore Hawaii also show a decline over the 19th and 20th centuries. On Rahmstorf’s reasoning, this is further evidence of cooling in the Atlantic subpolar gyre.  Or maybe warming. Or maybe both.

maka_sherwood-2015

sherwood_2014_np_annotated

217 Comments

  1. Posted Apr 2, 2015 at 12:08 AM | Permalink

    Quality science, as proxy
    Suffers much when the PI’s a doxy
    Not insight, just subsistence
    When there’s too little distance
    From sewage/grant funds orthodoxy

    ===|==============/ Keith DeHavelle

    • Beta Blocker
      Posted Apr 2, 2015 at 9:58 AM | Permalink

      There once was a proxy from Nantucket
      Rumor has the source a slop bucket
      As a proxy for hire
      It fits every gyre
      And from slop we surely can pluck it

    • Follow the Money
      Posted Apr 2, 2015 at 4:53 PM | Permalink

      Dendrooceanology

      Telesedimentology

      Curvefittingabation

      Anthozoicdataflagellation

      Climatescientology

  2. ThinkingScientist
    Posted Apr 2, 2015 at 2:18 AM | Permalink

    I can’t help it. Distance from sewage source seems such an apt metaphor for the work of some ClyScyentists

  3. Posted Apr 2, 2015 at 2:28 AM | Permalink

    Great post Steve..!

    The money quote being..

    “This is tautological mathematics, not science of deep insight.”

    What a surprise..

  4. Frank
    Posted Apr 2, 2015 at 3:08 AM | Permalink

    It’s very interesting that SR seems to be inconsistent with himself: See http://www.realclimate.org/index.php/archives/2015/03/a-hypothesis-about-the-cold-winter-in-eastern-north-america/ , update 1st of april (not a april fool’s?). He shows a figure from Zhang et.al 2011 http://www.realclimate.org/images//Zhang-2011.jpg where one can see clearly the impact of AMOC-change ( in this case from cold to warm) on the SST of the region. Nova scotia is not influenced by the AMOC as everybody can see. In the text below the figure:” This experiment is quite neat in that there is no change to the surface forcing; it is pure isolated effect of the AMOC, triggered by a change way below the surface in the deep overflow.” . So I understand that the SST-pattern itself is the effect of AMOC, what about the claim from “Nature” AMOC=SSTspg – Tnh??

  5. Ursus Augustus
    Posted Apr 2, 2015 at 7:28 AM | Permalink

    Is three tricks a record or something?

    It seems a lot of tricks to me.

    The paper must really be utter smelly stuff if it required three tricks.

    • S. Geiger
      Posted Apr 2, 2015 at 7:44 AM | Permalink

      Inverted slope via trickanometry?

      • Posted Apr 2, 2015 at 8:01 AM | Permalink

        +1 LOL

        • S. Geiger
          Posted Apr 2, 2015 at 9:20 AM | Permalink

          Still don’t know how our host, as a Candian no less, missed his opportunity for using the title: Rahmstorf’s Hat-Trick

          Steve: prettt bad miss. Especially with the Hockey Team.

  6. Craig Loehle
    Posted Apr 2, 2015 at 8:38 AM | Permalink

    I am going to venture a guess that there is a fourth trick. In the top figure the uptick after 1990 on left seems steeper than the green curve. Sherwood has data to 2002. A 10 yr moving average should stop 5 yrs before the end of the data but on left they go all the way out to 2002 (as far as I can tell from graph). Thus I suspect the usual endpoint padding by mirroring that makes the line go up more steeply than it should.

    • Steve McIntyre
      Posted Apr 2, 2015 at 9:50 AM | Permalink

      I am going to venture a guess that there is a fourth trick.

      I’d give very short odds on that:)

    • Gerald Machnee
      Posted Apr 2, 2015 at 11:06 AM | Permalink

      And i bet that Steve has long sleeves!

    • Spence_UK
      Posted Apr 5, 2015 at 8:08 AM | Permalink

      Sounds like someone needs to invent a trick to hide the tricks…

      • Green Sand
        Posted Apr 5, 2015 at 6:33 PM | Permalink

        Complacency

  7. Posted Apr 2, 2015 at 8:47 AM | Permalink

    The meaning of 15N enrichment is not clear to me. A quick goggle search for images relating to “oceanic nitrogen cycle” doesn’t show the relative importance of terrestrial vs. atmospheric inputs. (Presumably this depends on proximity to the coast.)

    It seems that d15N is high in effluent sources. It also seems that 15N is enriched by N-transformations so that in water masses with low water exchange d15N is high (because it has undergone more transformations). Temperature presumably has an effect in the same direction. Do nutrient-poor oceanic waters, effluent-rich waters, and warm waters all have elevated d15N? Confused.

    • Steve McIntyre
      Posted Apr 2, 2015 at 10:15 AM | Permalink

      Coral d15N values offshore Hawaii also show a decline over the 19th and 20th centuries. On Rahmstorf’s reasoning, this is further evidence of cooling in the Atlantic subpolar gyre.  Or maybe warming. Or maybe both.

      maka_sherwood-2015

      • MikeN
        Posted Apr 2, 2015 at 12:29 PM | Permalink

        I think we will see Rahmstorf et al 2016 that uses this proxy as well.

      • Posted Apr 3, 2015 at 12:48 PM | Permalink

        Well, maybe it is proof of the action-at-a-distance theory that the ideologies underlying the climate alarmist mentality often peddle.
        /sarc

        Or maybe the worldwide phenomenon of fewer whales defecating in the water, because of over-harvesting. 

      • John A
        Posted Apr 5, 2015 at 3:40 PM | Permalink

        Or maybe, just maybe, d15N values mean nothing climatologically being an index of biologic chaotic variation. But how would anyone get further funding for such a mundane explanation?

  8. Jeff Westcott
    Posted Apr 2, 2015 at 9:21 AM | Permalink

    Thank you Keith, I have missed your poetic contributions. They help soften the edge with no loss of impact.

  9. PhilH
    Posted Apr 2, 2015 at 9:53 AM | Permalink

    Are they paying any attention to these posts over at RC? I don’t go there anymore.

    • MikeN
      Posted Apr 2, 2015 at 12:29 PM | Permalink

      Oh they are paying attention, they just pretend not to.

    • Steve McIntyre
      Posted Apr 2, 2015 at 12:47 PM | Permalink

      Phil, when we did detailed commentary on Mann et al 2008, Mann quickly made changes at his Supplementary Information, always without acknowledgement or credit, although such conduct would appear to constitute plagiarism under most codes.

      • PhilH
        Posted Apr 2, 2015 at 1:56 PM | Permalink

        But not Mannian code!

  10. Posted Apr 2, 2015 at 10:30 AM | Permalink

    There once was a proxy fantastic
    Whose math was hyper-gymnastic
    Round and round we would go
    All the errors to show
    While Mann’s pals all remained quite bombastic

    This hockey stick concept is plastic
    Twisted, misguided, elastic!
    No critique can compete
    With Mann’s magical feat
    Making recent years’ temperatures drastic

    Nonetheless, my friends, do not despair
    McIntyre and McKitrick are there
    To research and check
    Mann’s great pile of dreck
    And revive climate science with fresh air

    • Craig Loehle
      Posted Apr 2, 2015 at 10:58 AM | Permalink

      A new competitor to Keith and Kim appears! hahahaha

      • Steve McIntyre
        Posted Apr 2, 2015 at 12:01 PM | Permalink

        People have to be wary of having too heavy a touch in such doggerel. We did some doggeral a number of years ago. https://climateaudit.org/2008/07/08/bull-dogs-have-little-dogs/ Mann is a tempting target but too much slagging comes across as simply angry.

        Here were a couple of my entries, the first on Hansen’s bull dogs, Tamino and Gavin, who in turn inspired even more mean-spirited lesser bull dogs.

        Bull dogs have little dogs
        That feed on their ferocity
        And little dogs have lesser dogs
        And thus to animosity

        Cherry trees have tasty fruit;
        And pickers need dexterity;
        But not as much as paleos,
        Who claim unprecedentity.

        • Posted Apr 2, 2015 at 1:50 PM | Permalink

          Our multi-tricker (Upside down?
          Again? And thought we’d miss it?)
          Is quite the mathematic clown
          As you make quite explicit

          A pity that these accidents
          (If such they are) in science
          Corrupt some useful evidence:
          Insight becomes defiance

          ===|==============/ Keith DeHavelle

      • kim
        Posted Apr 2, 2015 at 12:18 PM | Permalink

        Often when I see Keith’s stuff in comparison to mine I think, like McIntyre/Mann, there is a man who just wants to show you how to do it right.
        ==================

        • Posted Apr 2, 2015 at 1:46 PM | Permalink

          I much respect your cunning tell
          For mine would be much worse
          I do not do “no meter” well
          Though surely not a verse

          In single lines, in sly haiku
          I’ve seen you deftly skewer
          With literary thrust, right through
          The heart. You’re quite the doer!

          ===|==============/ Keith DeHavelle

  11. Posted Apr 2, 2015 at 12:14 PM | Permalink

    Do such people think through what they’ve drafted?

    While a binary error is easy to make, checking and thinking through should catch errors.

    A benefit of real peer review is that someone other than the writer is more likely to spot mistakes than someone steeped in the report.

    Not to absolve co-authors from checking each others’ work.

    (Yeah, I know, many of thing the errors are deliberately made.)

  12. Hoi Polloi
    Posted Apr 2, 2015 at 2:48 PM | Permalink

    So, already been some counter attack from Real Climate on this demolision.

  13. Hoi Polloi
    Posted Apr 2, 2015 at 3:02 PM | Permalink

  14. Posted Apr 2, 2015 at 3:25 PM | Permalink

    “Rahmstorf chose the right-axis d15N scale and orientation so that ”

    Something similar to what I mentioned in

    https://climateaudit.org/2015/01/08/more-mann-grafting/#comment-748452 ?

    Scaling data includes the option to switch the sign, double-y plots with LS-fit must be in their toolbox.

    Steve: as usual, you’re onto these tricks:) It’s a technique made for arm-waving.

    • Posted Apr 8, 2015 at 9:32 AM | Permalink

      I’ll start selling the toolbox once my version is ready.

      (calibration can be done otherway as well, as we know)

    • Posted Apr 14, 2015 at 12:05 PM | Permalink

      ..and when you zoom out you see that scale factor errors need to be taken seriously in calibration

      Figure S5 is interesting plotyy as well. Why Monte Carlo?

    • Posted Apr 17, 2015 at 7:14 AM | Permalink

      My guess at the moment to reproduce S5:

      1) AMOC index reconstruction from 900-1850 is used to obtain ML estimates for ARMA(1,1) model
      2) With resulting ARMA(1,1) model, 10 000 Monte Carlo series were generated, and then smoothed with 21-year running average
      3) For each series, minimum value of that series is calculated. These order statistics are collected to estimate cdf (shown in green).
      4) cdf and proxy recon uncertainty pdf @1975-95 mean value are multiplied and the result integrated to get the reported 0.0045 value

      S5

      I’m quite puzzled with this (but, at the same time, not so familiar with standard techniques developed during the past two decades in the paleoclimate community). The calibration methods they use tend to push the results toward calibration mean if the proxies are non-informative (example from https://climateaudit.org/2009/03/18/mann-2008-replication-ii/ ) :

      ICE

      Now, if you use that 900-1850 data to build a statistical model for the ‘true’ series, where does the noise (data uncertainty of the proxy reconstruction) go?? In Rahmstorf15 it appears in the observation whose unprecedentedness is about to be tested (1975-1995 mean). But disappears again if one reads carefully,

      “However, the 20th Century decline in the proxy-based AMOC index closely follows that found also in the instrumental data with their much smaller uncertainty, so that the chances that the proxy data are that far off are in reality much smaller than their formal uncertainty range would suggest”

  15. seanbrady
    Posted Apr 2, 2015 at 4:24 PM | Permalink

    “But this could be done with any smoothed and trending series. This is tautological mathematics, not science of deep insight. ”

    That great line remided me of Wolfgang Pauli’s famous putdown … and when I looked it up on Wikipedia I found some excellent context that makes it even more apt:

    “Regarding physics, Pauli was famously a perfectionist. This extended not just to his own work, but also to the work of his colleagues. As a result, he became known in the physics community as the “conscience of physics,” the critic to whom his colleagues were accountable. He could be scathing in his dismissal of any theory he found lacking, often labelling it ganz falsch, utterly wrong.”

    “However, this was not his most severe criticism, which he reserved for theories or theses so unclearly presented as to be untestable or unevaluatable and, thus, not properly belonging within the realm of science, even though posing as such. They were worse than wrong because they could not be proven wrong. Famously, he once said of such an unclear paper: It is not even wrong!””

    http://en.wikipedia.org/wiki/Wolfgang_Pauli

  16. Chuck L
    Posted Apr 2, 2015 at 4:38 PM | Permalink

    Tricksy scientists
    approxymate gyring sea
    currents quite smoothly

  17. Geoff Sherrington
    Posted Apr 2, 2015 at 6:45 PM | Permalink

    Here I offer a small prize to the first person to document a horizontal flip -one on the X axis – of a graph said to be relevant to climate analysis.
    Climate authors have provided the vertical or Y axis flip several times now, as Steve has noted, in this tposy-truvy (sic) world of unreal science.
    Why should one axis be more flip-preferred?
    An Ignobel Award awaits the first example of the XYZ all 3 axes example.
    But God only knows the tricks of the multi-dimensional hyperspace method, by which I forecast that an example will be found because of the abundance of those authors who equate their abilities.

    Steve: what about Mannian end=point smoothing?

    • M Happold
      Posted Apr 3, 2015 at 9:21 AM | Permalink

      Must the graph be of a time series such that they would be flipping the arrow of time?

    • ztabc
      Posted Apr 6, 2015 at 10:34 AM | Permalink

      Dessler already has an award for the correlation coefficient most indistinguishable from zero published in a ‘scientific’ journal (Science): https://climateaudit.org/2011/09/08/more-on-dessler-2010/

  18. David Brewer
    Posted Apr 2, 2015 at 7:37 PM | Permalink

    Assuming Rahmstorf et al actually believe in their results, the question is why? How are they misleading themselves? Two ideas:

    1) Fooled by pictures. They trawl for correlations and once they find them are convinced there is a relationship. Notice how the first figure here shows two versions of their “AMOC index”: “our[s]” and one “based on NASA GISS temperature data-48”. The lines are virtually indistinguishable and much be drawing on essentially the same data. But showing both lines gives the impression of mutual reinforcement, of verification. The question for them then is then whether the Sherwood proxy fits their two “verified” “instrumental” series. They fit it over with generous uncertainty shading and, hey presto, it works. They just forget that they inverted the temperatures to get their AMOC index, whereas Sherwood is still showing warm as up and cool as down.

    2) Fooled by words. Notice how the language they use assumes that their data series have succeeded in measuring the phenomenon in question. They describe Figure 5 as “A compilation of different INDICATORS for Atlantic ocean circulation” and their own proxy as an INDEX of that circulation. But are they, and is it? They skip that step. The very title of their article assumes what has to be proved: “Exceptional twentieth-Century slowdown in Atlantic Ocean overturning circulation.” Compare also Mann’s original article titled “Global SIGNATURES and Dynamical Origins of the Little Ice Age and Medieval Climate Anomaly”, similar to the frequent use of the word FINGERPRINT when attributing temperature change to CO2. Such terms imply unambiguous identification, as if it would be churlish or a waste of time to inquire further.

    Acting together, these two delusions help elide the step of proving mechanism, repressing elementary questions concerning physical dependence and even proximity in space to the phenomenon of interest.

    • David Jay
      Posted Apr 2, 2015 at 8:40 PM | Permalink

      +1

      Correlation is a serious flaw in human perception. We latch onto it and assume causation.

      Mechanism? We don’ need no stinkin’ mechanism!

    • Posted Apr 3, 2015 at 4:58 AM | Permalink

      David Brewer:

      Your insights reinforce the need for legitimate peer review of any and all scientific papers.

      • M Happold
        Posted Apr 3, 2015 at 9:42 AM | Permalink

        opluso: You are right, but it is completely impossible within the current system where perverse incentives to publish as much as possible mean that there will be a flood of papers that swamp the set of qualified peer reviewers. Ioannidis has made
        some excellent suggestions on how to remedy this, but they have no chance of being adopted.
        Much of peer review is done by graduate students because the professors are too busy writing grant applications and generating more papers. Good luck changing that.

        • Craig Loehle
          Posted Apr 3, 2015 at 11:29 AM | Permalink

          M Happold: You are correct about perverse incentives. Evidence that past work is flawed is of no interest to journals, even though that means flawed work keeps being cited and used. Likewise, refinements of past results are of less interest than completely new stuff. The incentive to publish lots of work vs quality means that many studies have inadequate sample sizes to be reliable.

        • Clark
          Posted Apr 5, 2015 at 11:25 AM | Permalink

          Peer review is mad worse by:

          1. Reviewing is entirely a volunteer activity. In a finite world, it’s easy to how the choice would be made between paying close attention to my own research, which benefits me directly, or checking someone else’s research, which benefits me not at all.

          2. It’s often hard to find reviewers, so that journals typically ask for “suggested reviewers”, amplifying the prevalence of pal review.

          3. Self-interest is also apparent in the one situation guaranteed to get a close review – when the reviewer is a competitor whose work is being superseded or refuted by the paper under review. As has been reported for papers attempting to question the climate consensus or refute Mann, the peer review process turns into a gatekeeping effort to keep ideas and questions out of the literature.

          Peer review needs a major overhaul. I think an open on-line review system might work much better than the current system

      • Duster
        Posted Apr 3, 2015 at 6:53 PM | Permalink

        The dismaying aspect is determining what “legimate peer review” would be. Presently, climate science claims that the only “peers” must be other climate scientists. They discard criticisms from statisticians and mathematicians, geologists and geographers as being not “peers.”

    • Posted Apr 3, 2015 at 7:56 AM | Permalink

      Only a true ‘peer’ would accept this paper. Like so many in politics, this represents a doubling down on the nonsense. No thinking person would ever accept this as a scientific contribution. The fact that it was published should concern a lot of people.

      • TimTheToolMan
        Posted Apr 6, 2015 at 8:40 AM | Permalink

        There seem to be more and more embarrassing features of the earth’s climate that are emerging …in the blogosphere at any rate. I didn’t even know this particular region of the ocean was defying the long term warming trend and so needed to be “explained”

    • Beta Blocker
      Posted Apr 3, 2015 at 9:49 AM | Permalink

      David Brewer, in my personal opinion, the pattern we have seen since 1998 with this ever-growing series of Mannian Methodology papers suggests either of two possibilities: (1) Climate scientists who subscribe to Mannian Methodology are both hopelessly deluded and scientifically incompetent; or (2) climate scientists who subscribe to Mannian Methodology are canny scientist-businessmen who are creating a series of contrived research products which are consciously tailored to service a strong market niche within the Global Warming Industrial Complex.

      • Don Keiller
        Posted Apr 3, 2015 at 12:21 PM | Permalink

        Beta Blocker, I strongly believe the latter. They have realised that their peers will not call them out on these “constructs” and are milking them for all it is worth. For them quite a lot.

      • mpainter
        Posted Apr 3, 2015 at 2:31 PM | Permalink

        More realistically, these are scientists who eschew any critical faculty; that is, insofar as the science supports the particularist/cultist point of view, they embrace it uncritically. It will get worse as the “hiatus” continues.
        Rahmstorf has opened a new vein of fool’s gold and it will be mined assiduously. The net message is that global warming has caused cooling in the NH via slowing of the AMOC.

        • kim
          Posted Apr 3, 2015 at 2:56 PM | Permalink

          Yes, this is well and truly through the looking glass. There is no end in sight for this tangled web.
          ===============

  19. Geoff Sherrington
    Posted Apr 3, 2015 at 1:09 AM | Permalink

    “Steve: what about Mannian end=point smoothing?”
    Only when the reflections meet in the middle on their way back.
    No prize, but almost.

  20. Posted Apr 3, 2015 at 1:51 AM | Permalink

    Reblogged this on I Didn't Ask To Be a Blog.

  21. PhilH
    Posted Apr 3, 2015 at 9:21 AM | Permalink

    This whole thing reminds me of something that used to happen in the legal community here in Charlotte some years ago. There was a woman attorney who specialized in land title work and she was incredibly sharp. When another lawyer who had certified title and closed a transaction on a piece of property, say, a year before was told by his secretary that Miss Leilia was on the phone and wanted to talk to him about that title, his drawers would start to get a little tight because the chances were nine out of ten that she had found a defect in his title work.

    You did not mess with Miss Leilia!

  22. Don Keiller
    Posted Apr 3, 2015 at 10:43 AM | Permalink

    I’m struggling to tell the difference between Rahmstorf et al’s “tricks” and bare-faced lies.

    Steve a letter to Nature pointing out these egregious practices is in order.

  23. Posted Apr 3, 2015 at 6:03 PM | Permalink

    Beta Blocker…instead of an either 1 or 2 choice..a combination of both is probably more valid.

  24. jorgekafkazar
    Posted Apr 3, 2015 at 10:56 PM | Permalink

    Windungszusammenbringen

  25. Ron Graf
    Posted Apr 4, 2015 at 2:23 PM | Permalink

    Michael Crichton on climate:
    (To says the least.)

  26. gallopingcamel
    Posted Apr 4, 2015 at 10:49 PM | Permalink

    Thank God for poets.

  27. Bill H
    Posted Apr 6, 2015 at 12:31 PM | Permalink

    Steve, Having read the rahmstorf paper it is clear to me that nitrogen 15 content in coral is a proxy for water mass changes NOT temperature. Your entire post would consequently seem to be based on a false premise.

    • Carrick
      Posted Apr 6, 2015 at 1:32 PM | Permalink

      Bill H: What am I missing.

      Figure 5 of Rammstorf 2015 clearly shows the nitrogen-15 proxy data converted to a temperature scale.

      The errant premise seems to be in Rammstorf et al, and not in this post, which does not endorse that premise in any case.

      • Bill H
        Posted Apr 6, 2015 at 1:51 PM | Permalink

        No, Carrick, the graph has two vertical scales: one on the left for temperature and one on the right for Nitrogen 15 concentration. No “conversion” is indicated anywhere in the paper or required for the graph.

        Such graphical representation to demonstrate a correlation between two variables is commonplace in the scientific literature. From what I know of your background I would be very surprised if you had not come across such graphs before.

        • Carrick
          Posted Apr 6, 2015 at 3:00 PM | Permalink

          Thanks Bill. I just missed the axis on the right: When I do these sorts of graphs I usually put arrows on curves to indicate when vertical axis to read from. As it was, I didn’t even notice there was a right axis, because of the microscopic font size for that axis label (when viewed in full panel mode).

        • Steve McIntyre
          Posted Apr 6, 2015 at 3:56 PM | Permalink

          Bill H, I do not understand what exactly you are objecting to. I do not claim to be infallible and try to correct errors when they are pointed out to me. Can you provide a direct quotation from statements that you believe to be erroneous. Simply saying that the entire post is “based on a wrong premise” is arm-waving. It is my understanding that Rahmstorf’s Figure 5 purports to show a positive correlation between Nova Scotia coral d15N and gyre temperature and I have seen nothing in your statements to change this understanding.

        • Bill H
          Posted Apr 6, 2015 at 4:33 PM | Permalink

          Steve, Thanks for calling by. I would suggest the following quote from you makes plain your assumption about Ramstorf et al attempting to use nitrogen-15 as a temperature proxy:

          “The idea that coldwater corals offshore Nova Scotia can be thermometers for ocean temperature in the subpolar gyre has little more plausibility than the belief that stripbark bristlecones in the distant Sierra Nevadas or contaminated Finnish sediments can be thermometers for the subpolar gyre.
          It’s not even well established that coral d15N is a proxy for local ocean temperature.”

          If this not an attack on the paper what on Earth is it?

        • Bill H
          Posted Apr 6, 2015 at 4:38 PM | Permalink

          Also, Steve, I detect maybe a bit of “revisionism” on your part. You are now saying that the paper “purports to show a correlation between N-15 and gyre temperature.” If you are referring to the specific gyre in question then I would agree with you and I think they have a pretty strong case. If you are claiming a correlation with gyres in general then I would suggest you re-read the paper.

        • Don Monfort
          Posted Apr 6, 2015 at 4:50 PM | Permalink

          You are meandering, billy. Steve said: “It is my understanding that Rahmstorf’s Figure 5 purports to show a positive correlation between Nova Scotia coral d15N and gyre temperature and I have seen nothing in your statements to change this understanding.”

          You need to address that, billy. If you want to be taken seriously.

        • Bill H
          Posted Apr 6, 2015 at 5:32 PM | Permalink

          Don, I appreciate you’ve arrived rather late to this debate, but Steve has asked me to address where in his original post he has made an untenable assumption, not anything he has written in the ensuing correspondence. What he says about there being merely a correlation is uncontroversial.

        • Don Monfort
          Posted Apr 6, 2015 at 5:46 PM | Permalink

          billy:Steve, Having read the rahmstorf paper it is clear to me that nitrogen 15 content in coral is a proxy for water mass changes NOT temperature. Your entire post would consequently seem to be based on a false premise.

          Steve: It is my understanding that Rahmstorf’s Figure 5 purports to show a positive correlation between Nova Scotia coral d15N and gyre temperature and I have seen nothing in your statements to change this understanding.

          billy:You are now saying that the paper “purports to show a correlation between N-15 and gyre temperature.” If you are referring to the specific gyre in question then I would agree with you and I think they have a pretty strong case. If you are claiming a correlation with gyres in general then I would suggest you re-read the paper.

          Are we done now, billy? Is that what you got?

      • Posted Apr 6, 2015 at 2:39 PM | Permalink

        Indeed, there are graphs of two different scales
        One one each side, different coding called out
        And clearly the Rahmstorf Figure 5 fails
        To make such distinctions. He leaves you no doubt.

        Since there are no different kinds of chart trace
        Just collapse down the chart so the two axes touch
        There’s the conversion, right in your face
        To say “a false premise” is saying too much.

        ===|==============/ Keith DeHavelle

        • Bill H
          Posted Apr 6, 2015 at 3:31 PM | Permalink

          Keith,Instead of producing doggerel how about reading Rahmstorf et al. You’ll find it’s pretty clear from that.

    • Posted Apr 6, 2015 at 2:30 PM | Permalink

      Steve, Having read the rahmstorf paper it is clear to me that nitrogen 15 content in coral is a proxy for water mass changes NOT temperature.

      Could be, but co-plotting them (with the dual y-scales) shows they *are* comparing it with temp. Otherwise, why would you display some time-series plotted vs. temp, and others plotted w/N15?

      As UC pointed out, independent scaling of two (as you say, unrelated) y’s gives the graph maker a large degree of freedom in squiggle matching.

      • Carrick
        Posted Apr 6, 2015 at 3:04 PM | Permalink

        Terry the figure caption does say “A compilation of different indicators for Atlantic ocean circulation”. You’re supposed to infer that both proxies point to a change in the rate of circulation.

        • Bill H
          Posted Apr 6, 2015 at 3:27 PM | Permalink

          Carrick,
          If you read Rahmstorf et al.’s paper it’s pretty clear that the two variables point to a change in rate of overturning, as Carrick describes. It’s interesting that Steve is rather less helpful on these matters, preferring, as is his wont, to leave such things as “exercises for the student”. Rather like when he passed his Monte Carlo software on to Wegman, leaving it to Wegman to make sure he didn’t inadvertantly print out copies of Steve’s data from earlier Monte Carlo runs. So much for independent verification of the “hockey sticks from red noise” claim

          Or maybe Steve just made a mistake.

          Steve: I do not claim to be infallible, but do not believe that I’ve made a mistake here. I try to write clearly and to fully document things. In our 2005 articles, I placed code online to enable people to help clarify what we did. In doing so, I was attempting to be more helpful than those authors who provided no such details. I don’t believe that such attempts deserve sneering. At the time, I had not developed the concept of “turnkey” code and later code has moved in that direction. I didnt “pass” this code to Wegman – I placed it online, where it was accessed by, among others, Huybers, Wahl and Ammann, and Wegman and/or one of his assistants.

        • Carrick
          Posted Apr 6, 2015 at 3:38 PM | Permalink

          Wegman is a different issue—He was paid a nice sum to independently vet that work, and in my opinion should have written his own code. Not even understanding the code isn’t an exercise for the reader in his case, Wegman was supposed to have been operating in the capacity of a professional.

          Steve: WEgman wasn’t paid anything. On the other hand, the NAS panel had a substantial budget. As I wrote at the time, I was surprised at the very limited actual due diligence done by the panels. At the time, I was surprised that the NAS panel recommended that stripbark be avoided in temperature reconstructions, but then used reconstructions with stripbark in their spaghetti graph. I had a chance to ask North about this in an online colloquoium and North had no coherent answer.

        • Bill H
          Posted Apr 6, 2015 at 3:55 PM | Permalink

          Well, Steve could have told him about the bug, and should surely have drawn attention to it promptly when that appalling piece of work was issued. As it was it was hushed up for years, till the pesky DeepClimate unearthed the matter, also falsifying Ross McKitrick’s claim about 99% of the Monte Carol runs leading to Hockey Sticks. Funny how 99% DIDN’T give ’em.

          By the way, you seem a very knowledgeable person, Carrick. You wouldn’t by any chance know if Ross ever publicly withdrew that 99% claim? Maybe Steve could help if he’s still following this thread?

        • Posted Apr 6, 2015 at 4:11 PM | Permalink

          Bill H – not saying that you keep changing the subject, but can you maybe prioritize the questions you’ve posed? Thanks!

        • Bill H
          Posted Apr 6, 2015 at 5:04 PM | Permalink

          Steve. Wow, so you had no knowledge until Deepclimate revealed it in 2012(??), that Wegman was using your code and then produced a report containing your and Ross’s data and passing it off as his own work, whether deliberately or through incompetence.

        • Carrick
          Posted Apr 6, 2015 at 6:37 PM | Permalink

          Steve McIntyre:

          WEgman wasn’t paid anything

          I know he originally claimed it was pro bono. I can’t find a direct link right now, but I’m pretty sure he amended that later to say he charged time from of his grant(s) to work on this project.

      • Bill H
        Posted Apr 6, 2015 at 3:15 PM | Permalink

        Here are some examples of 2 – vertical scale graphs of the sort that Rahmstorf et al. use. Now is anyone going to suggest that this automatically implies a proxy relation of the sort that Steve is claiming?

        https://www.creditwritedowns.com/2010/10/shifting-correlations.html

        • clays
          Posted Apr 9, 2015 at 5:44 PM | Permalink

          Bill H: your writing has hardly been the model of clarity (you sadly seem to be well and truly out of your depth in this thread), but providing this link is just meaningless. You previously state that such graphical representation to demonstrate a correlation between two variables is commonplace in the “scientific literature”. Now you link to, well, what exactly? It seems to be nothing more than a quick observation written on a random finance blog. It is certainly not the “scientific literature”. There are many perfectly fine, prestigious peer-reviewed finance journals. This just isn’t one of them. It doesn’t pretend to be one of them. It is certainly possible you are correct that that these type of graphs are common in the scientific literature. But providing worthless links of this type isn’t going to help make your case.

          (And enough with the tangents on Wegman etc. It is a silly debating tactic).

    • Beta Blocker
      Posted Apr 7, 2015 at 1:19 PM | Permalink

      Re: Bill H (Apr 6 12:31),

      Bill H: Steve, Having read the rahmstorf paper it is clear to me that nitrogen 15 content in coral is a proxy for water mass changes NOT temperature. Your entire post would consequently seem to be based on a false premise.

      Mr. Bill, from a science argumentation perspective, it is not obvious where you are coming from with your commentary regarding Steve’s critical analysis of the Rahmstorf paper.

      I have read both the paper and your commentary about it several times, and there appears to be some disagreement between you and Steve concerning what material the Rahmstorf paper actually contains in terms of its stated objectives, its stated methodology, and its stated key conclusions.

      If we are to make any sense of your commentary, what is needed from you is an understanding of what you yourself believe the paper contains in terms of its stated objectives, its stated methodology, and its stated key conclusions.

      Could you be so kind as to give us a short synopsis describing your own personal interpretation of each of these three facets of the Rahmstorf paper: (1) its objectives; (2) its methodology; (3) its key conclusions?

  28. Bill H
    Posted Apr 6, 2015 at 4:09 PM | Permalink

    Ooh the moderators are after me. Just as well I’ve made screen shots of all this.

    Steve: please do not think that your observations are particularly challenging. One of your comments went into automatic moderation because you used a blacklisted word. Surely you should be able to discuss coral d15N without using the word “nazi” – a blacklisted word. Pathetic that you resort to such language.

    • davideisenstadt
      Posted Apr 6, 2015 at 4:34 PM | Permalink

      Bill H:
      whats up with your snark?
      and who is “bill H”?

      • Bill H
        Posted Apr 6, 2015 at 4:51 PM | Permalink

        David,

        Considering the amount of abuse that’s been heaped on Rahmstorf et al. in this thread, much of it in verse I find it strange that you single me out for “snark”.

        • Bill H
          Posted Apr 6, 2015 at 5:07 PM | Permalink

          Steve, Thanks for calling by. I would suggest the following quote from you makes plain your assumption about Ramstorf et al attempting to use nitrogen-15 as a temperature proxy:

          “The idea that coldwater corals offshore Nova Scotia can be thermometers for ocean temperature in the subpolar gyre has little more plausibility than the belief that stripbark bristlecones in the distant Sierra Nevadas or contaminated Finnish sediments can be thermometers for the subpolar gyre.
          It’s not even well established that coral d15N is a proxy for local ocean temperature.”

          If this not an attack on the paper what on Earth is it?

        • Bill H
          Posted Apr 6, 2015 at 5:13 PM | Permalink

          Hi, Don, Did you see my reply to your comment at Prof Curry’s place where I invited you to view my efforts on this blog. You had been suggesting to a “warmist” that s/he come to Climate Audit to debate with Steve, then going on to say, a tad unkindly, he/she wouldn’t be able to take the intellectual heat. If your comments are anything to go by the intellectual level would not seem to be quite so fearsome.

        • mpainter
          Posted Apr 6, 2015 at 5:30 PM | Permalink

          Correct, Bill H, Rahmstorf et al explicitly state that that the corals are a temperature proxy via 15 N. You seem to reject the idea that 15 N can be useful as a temperature proxy, as we all do. Wherefore your complaint?

    • Bill H
      Posted Apr 6, 2015 at 5:14 PM | Permalink

      Hi, Don, Did you see my reply to your comment at Prof Curry’s place where I invited you to view my efforts on this blog. You had been suggesting to a “warmist” that s/he come to Climate Audit to debate with Steve, then going on to say, a tad unkindly, he/she wouldn’t be able to take the intellectual heat. If your comments are anything to go by the intellectual level would not seem to be quite so fearsome.

    • Bill H
      Posted Apr 6, 2015 at 5:15 PM | Permalink

      Oh, bless me: moderated for quoted the Noble Viscount.

      • Don Monfort
        Posted Apr 6, 2015 at 5:22 PM | Permalink

        You are hysterical in both commonly used meanings of the word, billy. Try to calm down and get on with whatever it is you are trying to do. Show us what Steve Mc. has got wrong, if you got the goods. This is your big chance. Make sure you save screen shots of the whole spectacle.

        • kim
          Posted Apr 6, 2015 at 5:35 PM | Permalink

          Don’t anybody tell Bill that the moderating screen doesn’t kowtow to Viscounts, either; wouldn’t want to halt his snit.
          ===========================

      • Bill H
        Posted Apr 6, 2015 at 5:51 PM | Permalink

        ONce again, Don, you’re rather late to the party. I have already stated where Steve has made an untenable assumption. See my post above, way back at at 4:33 Toronto(?) time. Still waiting for an answer from Steve. Incidentally you might also want to check out Carrick’s contribution (3:04 pm) to my original post: he agrees with me that Steve is in error here.

        • Stephen McIntyre
          Posted Apr 6, 2015 at 6:07 PM | Permalink

          Bill H, thus far the only statement in the above article that you’ve directly quoted is the following:

          The idea that coldwater corals offshore Nova Scotia can be thermometers for ocean temperature in the subpolar gyre has little more plausibility than the belief that stripbark bristlecones in the distant Sierra Nevadas or contaminated Finnish sediments can be thermometers for the subpolar gyre. It’s not even well established that coral d15N is a proxy for local ocean temperature.

          I stand by my claim that contaminated Finnish sediments and strip bark bristlecones are not valid proxies for Atlantic ocean currents. Do you dispute this? I stand by my claim that it is not well established that coral d15N is a proxy for local ocean temperature. It is a novel proxy and still poorly documented. Sherwood et al reported a negative correlation to local ocean temperature (as opposed to the positive correlation to gyre temperature illustrated in Rahmstort Figure 5), but without a coherent understanding of the properties of the d15N proxy, I do not believe that any statistical meaning can be attached to the sort of squiggle provided in Rahmstorf’s Figure 5.

          Please do not assume that I am online all the time or able to provide room service responses.

        • Don Monfort
          Posted Apr 6, 2015 at 6:17 PM | Permalink

          billy:Steve, Having read the rahmstorf paper it is clear to me that nitrogen 15 content in coral is a proxy for water mass changes NOT temperature. Your entire post would consequently seem to be based on a false premise.

          How is nitrogen 15 content in coral offshore Novia Scotia related to water mass changes in the gyre somewhere else? Got anything to do with temperature?

          Steve:Rahmstorf’s Figure 5 shows a positive correlation between temperature and coral d15N (both decline together)

          That’s true, ain’t it?

          Where is the Figure in Rahmstorf that shows a positive correlation between Nova Scotia coral nitrogen 15 content and the gyre water mass crap?

      • Bill H
        Posted Apr 6, 2015 at 5:53 PM | Permalink

        Oh, and Don, since this is a science blog, do please adduce some evidence for my hysteria.

        • Jeff Norman
          Posted Apr 6, 2015 at 6:50 PM | Permalink

          Hysteria – a psychological disorder (not now regarded as a single definite condition) whose symptoms include conversion of psychological stress into physical symptoms (somatization), selective amnesia, shallow volatile emotions, and overdramatic or attention-seeking behavior.

          So far Bill H posted this afternoon at 1:51, 3:15. 3:27, 3:31, 3:55, 4:09, 4:33, 4:38, 4:51, 5:04, 5:07, 5:13, 5:14, 5:15, 5:32, 5:51 and 5:53 (blog time), not including the moderated posts. Several times in response to themselves, and occasionally repeating the same message again.

          The visible posts seem dramatic and apparently attention-seeking, but this is a subjective opinion.

        • Michael Jankowski
          Posted Apr 6, 2015 at 7:53 PM | Permalink

          As a starting point, Billy, can tell us how d15N goes down as T goes up in Sherwood et al whereas d15 does up as T goes up in Rahmstorf? How do you reconcile this 180 degree difference in correlation?

        • Bill H
          Posted Apr 8, 2015 at 9:49 AM | Permalink

          I might comment if you apologise for the patronising “billy” address.

      • Bill H
        Posted Apr 6, 2015 at 7:36 PM | Permalink

        @Steve Mc:

        “I stand by my claim that contaminated Finnish sediments and strip bark bristlecones are not valid proxies for Atlantic ocean currents. Do you dispute this? I stand by my claim that it is not well established that coral d15N is a proxy for local ocean temperature. It is a novel proxy and still poorly documented. Sherwood et al reported a negative correlation to local ocean temperature (as opposed to the positive correlation to gyre temperature illustrated in Rahmstort Figure 5), but without a coherent understanding of the properties of the d15N proxy, I do not believe that any statistical meaning can be attached to the sort of squiggle provided in Rahmstorf’s Figure 5.”

        The straw men are coming thick and fast. Can you please not digress onto subjects such as whether Finnish sediments are valid proxies for ocean current (whether “local” or otherwise): the authors are not claiming this, as I think you are aware. As for there being “no coherent understanding of the properties of the d15N proxy”, well that’s for you to research. For all you know it could be a well-understood proxy. Basically your attack on them is centred on their using 15-N concentrations as a “thermometer”, and, judging by this comments section: this is the message that your readers, overwhelmingly, have received . I cite the specific examples of: Carrick (before I put him right), Monfort, ,Terry MN, MPainter, JIT Indeed when JIt specifically asked “do warm waters all have elevated d15N?” you replied, not with a caution that Rahmstorf et al. were not claiming a general correlation between water temperature and N-15 concentration, but instead you reinforced his/her misunderstanding with the following:

        “On Rahmstorf’s reasoning, this is further evidence of cooling in the Atlantic subpolar gyre. Or maybe warming. Or maybe both.”
        Face it, Stephen, you have comprehensively misled your readers on what exactly the authors’ trick is in this case. So effectively have they been misled that they have , with the honourable exception of Carrick, personally and sometimes rudely attacked me when I have attempted to put them right on the matter. I would suggest an apology, a retraction and re-write of your post, stating exactly what trick 3 is. Also maybe a rebuke for the tone some of your followers have displayed, especially after you described my language as “pathetic”: if you actually read my “pathetic” post you will observe that I was not accusing anyone of being a n*z*, eco- or otherwise.

        • HAS
          Posted Apr 6, 2015 at 7:55 PM | Permalink

          The thing I’m confused about here is why Rahmstorph et al didn’t orient the δ15N tracer the other way up and lag the temp dependent indices relative to it in their fig. 5 given the known relationship with temp from Sherwood et al (p <0.001).

        • Steve McIntyre
          Posted Apr 7, 2015 at 12:37 AM | Permalink

          You say:

          Can you please not digress onto subjects such as whether Finnish sediments are valid proxies for ocean current (whether “local” or otherwise): the authors are not claiming this, as I think you are aware.

          Puh-leeze. The starting point of this series of posts was Rahmstorf and Mann’s use of contaminated Finnish sediments and stripbark bristlecones as proxies for Atlantic ocean currents. This is hardly a digression, it’s one of the central points. Your assertion that the authors “are not claiming this” is contradicted by the article itself, which purports to be a reconstruction of an Atlantic ocean current using contaminated Finnish sediments and stripbark bristlecones as proxies. Both are integral part of the Mann et al 2009 network.

          Their use of d15N is secondary. As stated in the above post, Sherwood et al 2011 reported a negative correlation between coral d15N and temperature. They attributed this negative correlation to relatively higher d15N values in (cold) Labrador Slope Water relative to (warmer) Western Slope Waters. On the other hand, as shown above, Rahmstorf et al showed a positive correlation between d15N values and gyre SST. My issue with these squiggles is that one could scale and orient virtually any trending series and achieve a similar rhetorical impression and thus no valid statistical statistical conclusions can be drawn from the comparison in Rahmstorf Figure 5: it is, as I observed, a trick.

          You haven’t acknowledged the ludicrousness of Rahmstorf and Mann’s use of contaminated Finnish sediments in the calculation of their reconstruction. This is not a “digression” but a central issue. I find it hard to believe that you endorse the use of contaminated sediments in proxy reconstructions, but, for good order’s sake, I’d appreciate confirmation on whether or not you endorse their use of contaminated sediments. If you are unwilling to criticize such absurdities, it will clarify your position.

        • Frank
          Posted Apr 7, 2015 at 3:16 AM | Permalink

          Bill, I don’t full understand your point. You say, that the d15N from the island of Nova Scotia is a proxy for water mass, not for temperature, wright? If not for temperature… for salinity? The SPG is 1500 km away from this island and IF some physical quantity from Nova Scotia show something with a significant correltion – negative or postive- to the SSTspg ( for this quantity it’s used as a proxy as it’s shown in fig. 5 of the paper) it should be shown, shouldn’t it? What is your point exactly?

        • Posted Apr 7, 2015 at 9:30 AM | Permalink

          That is the point. The source of water flowing off Nova Scotia is a balance between two currents, the cold Labrador current and the Gulf Stream. By looking at d15N in Sherwood’s proxy, one can conclude which of these currents dominate at any time. Knowing which current dominates off of Nova Scotia is a proxy for what is happening in the sub-polar gyre.

          There is considerable confusion here because what Rahmstorf et al discuss is a teleconnection, paraphrasing the Wikipedia, applied to sea currents not the atmosphere where climate anomalies are related to each other at large distances (typically thousands of kilometers),

          The teleconnection is an observation. Rahmstorf, et al, describe a mechanism for this teleconnection, that both are controlled by the relative strengths of specific ocean currents.

          Steve: The relationship of the Labrador Current and Gulf Stream offshore East Coast was discussed at CA last month here. In that post, I observed:

          Sachs observed that a relatively small coastward displacement of the Gulf Stream could account for the difference and plausibly speculated that the Gulf Stream hugged the East Coast much more closely in the mid-Holocene.

          In my post on d15N, I pointed to recent discussion of alkenones and commented adversely on Rahmstorf’s failure to consider these actual proxies. You say:

          By looking at d15N in Sherwood’s proxy, one can conclude which of these currents dominate at any time.

          At present, there are only a few studies of coral d15N. One can conjecture that the information sheds light on the relative contributions of the Labrador Current and Gulf Stream offshore Nova Scotia, but the information is very limited. That’s one of the reasons why I observed that the alkenone information should have been considered by the authors.

        • Posted Apr 7, 2015 at 10:41 AM | Permalink

          Eli Rabett said

          …where climate anomalies are related to each other at large distances (typically thousands of kilometers)…

          I assume this assertion traces back to the early work of Hansen & Lebedeff? If so, the “thousands” should probably be reduced to “hundreds” and, even then, presented only with certain exclusions and caveats.

        • Posted Apr 7, 2015 at 1:48 PM | Permalink

          While Hansen and Lebedeff described a correlation between weather stations some distance apart (~1200 km or so) and there has been some interesting more recent work on that showing the distances vary with season (don’t ask Eli to find it, but he has read several papers on the subject) teleconnetions are much longer and more general. See this for a start

          http://www.cpc.noaa.gov/data/teledoc/telecontents.shtml

          Steve: Spatial autocorrelation with exponential decay (1200 km) of station data appears supported by data. I recall discussing this in connection with our commentary on Steig et al 2009 and in connection with Chladni patterns. This isn’t what is meant by “teleconnection”. Nor does it support very pronounced negative correlations.

        • Posted Apr 7, 2015 at 3:53 PM | Permalink

          Of course teleconnections (AMO, PDO, etc.) exist and influence weather patterns around the globe. But since you mentioned “anomalies” I assumed you were specifically referring to the long-distance in-filling/adjustment technique promoted by Hansen, et al. The 1200 km distance was selected in Hansen, et al., because that was the point at which correlation coefficients fell below 0.5, on average. Of course, this means that many of the in-filled anomalies do not achieve “strong” correlations although the focus has always been on those that do. This radius also enabled them to assert approximately 80 percent global “coverage” with existing weather station records. However, regardless of whether Hansen’s 1200 km distance was sufficiently robust to support the strength of his claim, to stretch it to “thousands of kilometers” would seem to be unwarranted.

        • Steve McIntyre
          Posted Apr 7, 2015 at 4:11 PM | Permalink

          In my opinion, there is considerable evidence for spatial autocorrelation of temperatures up to 1200 km or so. Indeed, this motivated the discussion of Chladni patterns in connection with our commentary on Steig et al 2009, as principal components applied to spatially autocorrelated data in a finite geometric region yields Chladni patterns.

          The average distance from the Nova Scotia corals to the subpolar gyre gridcells used in R15 is 2500 km, double the Hansen radius. In addition, Rahmstorf is not arguing for a positive correlation, but for a NEGATIVE correlation. While NH wave patterns do exist, it’s a large jump to claiming that these are stable enough to yield a permanent negative correlation of temperatures at a distance of 2500 km or so.

          This (relatively) uncontroversial idea of spatial autocorrelation does not mean that stripbark bristlecones can act as thermometers for Southern Hemisphere.

        • Posted Apr 8, 2015 at 10:21 AM | Permalink

          Steve, your comment continues to remind Eli of Stephan Leacock’s

          It was a wild and stormy night on the West Coast of Scotland. This, however, is immaterial to the present story, as the scene is not laid in the West of Scotland. For the matter of that the weather was just as bad on the East Coast of Ireland.

          But the scene of this narrative is laid in the South of England and takes place in and around Knotacentinum Towers (pronounced as if written Nosham Taws), the seat of Lord Knotacent (pronounced as if written Nosh).

          But it is not necessary to pronounce either of these names in reading them.

          As the coral d15N measurements characterize the flow of nutrients into the area, not the temperature. As currents which carry different concentrations of nutrients shift and strengthen or weaken this is reflected in the d15N record. The temperature is a secondary issue for Sherwood et al.

          Steve: Thank you for reminding me of that excellent Leacock quote, Leacock being part of the youth of Canadians of a certain age. Leacock’s phrase seems excellently apt for Rahmstorf’s attempt to reconstruct North Atlantic ocean currents using stripbark bristlecone pine chronologies and contaminated Finnish sediments. Please keep in mind that these are what Rahmstorf used to reconstruct the temperature of the subpolar gyre, which Rahmstorf equated to an index of ocean circulation. To paraphrase:

          It was a wild and stormy night on the North Atlantic just south of Greenland. This, however, is immaterial to the present story, as the scene is not laid in the North Atlantic just south of Greenland. For the matter of that the weather was just as bad on the west Coast of Iceland.

          But the scene of this narrative is laid in California and takes place in and around Knotacentinum Towers (pronounced as if written Sheep Mountain), where, during the short summer, bristlecone pines were quietly converting carbon dioxide and water to stripbark cambium.

        • Matt Skaggs
          Posted Apr 9, 2015 at 9:26 AM | Permalink

          Perfect! And now I can (thankfully) stop reading this thread.

        • Matt Skaggs
          Posted Apr 9, 2015 at 9:28 AM | Permalink

          Hmmm, my comment was supposed to appear under Steve’s paraphrase of Leacock. Comment nesting seems to have failed again.

        • DB
          Posted Apr 9, 2015 at 9:35 AM | Permalink

          Although I enjoy reading through comments here immensely, it does seem the commenting system is a bit flawed, and I am using the CA. It does make it exceptionally difficult to follow conversations at times, although I tend to search for dates in conjunction with names to follow back and forths.

        • Sven
          Posted Apr 9, 2015 at 10:14 AM | Permalink

          DB, the commenting system is just screwed up at the moment. Seems to be a wider issue

        • MikeN
          Posted Apr 9, 2015 at 11:04 AM | Permalink

          Now if the numbers would come back for each comment, it would be excellent.

        • medge
          Posted Apr 11, 2015 at 12:21 AM | Permalink

          I’m trying to let Steve know that I dove into the water off Peggy’s Cove (The Swiss Air Memorial) and floated effortlessly. That was in August last year..The salt was in the water..Getting back to shore was fun having to get a grip on seaweed covered rocks which is why I kept my shoes on..The flow or drift was right to left,or south to north..Cold at first but not so bad after 30 seconds or so..But oh the salt..
          Don’t know how you can put up with these idiots like “h”..Amazing..

        • tty
          Posted Apr 8, 2015 at 2:36 PM | Permalink

          If you go back to Hansen & Lebedeff’s original 1987bpaper and look at the actual data you will find that the claimed 0.5 correlation at 1200 km only applies to the extratropical northern hemisphere (30 % of the Earth’s surface).
          In the tropics the correlation is <0.5 even for nearby stations and in the southern extratropics the 0.5 correlation only extends to about 600-700 km.

        • Geoff Sherrington
          Posted Apr 9, 2015 at 6:19 AM | Permalink

          Steve,
          I’d be delighted if one day you made a separate thread for graphs of this type as discussed

          I have worries about this graph, but lack the concentration to discern my unease.

          Not wanting to divert this thread, so I’ll just point to some work I did on Melbourne Regional BOM 86071 from 3 years ago.
          Simply, I lagged temperatures and calculated correlation coefficients for lags of 1, 2, … 12 days and other data choices as explained.
          Rationale was that a lag of 1 day was rather like a weather system moving some km in a day , to simulate weather stations separated by those km.
          http://www.geoffstuff.com/chasing_R.pdf
          (Note: There are occasional minor errors in the text, but they make little difference).

          Geoff.

        • Posted Apr 7, 2015 at 8:07 PM | Permalink

          Steve said

          Steve: Spatial autocorrelation with exponential decay (1200 km) of station data appears supported by data. I recall discussing this in connection with our commentary on Steig et al 2009 and in connection with Chladni patterns. This isn’t what is meant by “teleconnection”. Nor does it support very pronounced negative correlations.

          Thank you for agreeing on the teleconnections point.

          As to the range of spatial autocorrelation, it sort of, depends on where and when. See New, Hulme and Jones for example. Eli had a post on this way back when.

          The interesting point is that the correlations in Hansen and Lebedeff were much higher for high latitudes than they were in the tropics. At mid and high latitudes the correlation was attributed to large scale eddy mixing. They picked a 1200 km radius as the distance at which correlation was at least 0.5 at middle and high latitudes and 0.33 at low ones and used this correlation to construct their first global temperature record.

          Inherent to this is the thought that the correlation can vary from season to season, and indeed as New, Hulme and Jones show it does. This offers a possible improvement to surface temperature methods.

        • Posted Apr 7, 2015 at 9:37 PM | Permalink

          Eli Rabett said:

          As to the range of spatial autocorrelation, it sort of, depends on where and when. See New, Hulme and Jones for example.

          In New, et al., it also depends on the “what.” For mean temperature they utilized a Hansen-esque 1200 km radius but for other climate elements the correlation decay distance (CDD) was much smaller. In addition, the “where” and the “when” distinctions were, to some degree, ignored by their process (shortcut?) of creating a global average CDD.
          As for your original parenthetical comment (“typically thousands of kilometers”), you have not provided any additional justification for correlations at such distances.

          Steve: nor did Rabbitt support large negative correlations. A complete red herring.

        • Steven Mosher
          Posted Apr 8, 2015 at 12:55 PM | Permalink

          “The interesting point is that the correlations in Hansen and Lebedeff were much higher for high latitudes than they were in the tropics. At mid and high latitudes the correlation was attributed to large scale eddy mixing. They picked a 1200 km radius as the distance at which correlation was at least 0.5 at middle and high latitudes and 0.33 at low ones and used this correlation to construct their first global temperature record.

          Inherent to this is the thought that the correlation can vary from season to season, and indeed as New, Hulme and Jones show it does. This offers a possible improvement to surface temperature methods.”

          Well the ‘higher correlations for higher latitudes” holds, until you get to the arctic. And then all hell breaks loose.

          “The statistics of surface air temperature observations obtained from buoys, manned drifting stations, and meteorological land stations in the Arctic during 1979–1997 are analyzed. Although the basic statistics agree with what has been published in various climatologies, the seasonal correlation length scales between the observations are shorter than the annual correlation length scales, especially during summer when the inhomogeneity between the ice-covered ocean and the land is most apparent. During autumn, winter, and spring, the monthly mean correlation length scales are approximately constant at about 1000 km; during summer, the length scales are much shorter, i.e. as low as 300 km. These revised scales are particularly important in the optimal interpolation of data on surface air temperature (SAT) and are used in the analysis of an improved SAT dataset called IABP/POLES. Compared to observations from land stations and the Russian North Pole drift stations, the IABP/POLES dataset has higher correlations and lower rms errors than previous SAT fields and provides better temperature estimates, especially during summer in the marginal ice zones. In addition, the revised correlation length scales allow data taken at interior land stations to be included in the optimal interpretation analysis without introducing land biases to grid points over the ocean.”

        • Steven Mosher
          Posted Apr 8, 2015 at 1:02 PM | Permalink

          Eli.

          you wrote:

          However, this correlation IS encouraging for polar regions, where we see that the 1200 km range is a good solid estimate all year long between latitudes 60 and 90 N. One may assume that the same could, maybe even should, hold true for the southern polar regions.”

          This is Wrong.

          The polar region as I cited above has correlation lengths that drop to 300km in the summer. Please stop talking about temperature when you have no publication history in the field.

        • Posted Apr 8, 2015 at 3:48 PM | Permalink

          Is that a rabbit punch? Thought that was illegal.

        • Posted Apr 8, 2015 at 3:49 PM | Permalink

          Dang, that was supposed to be a reply to Steven Mosher at 1:02 PM.

        • Beta Blocker
          Posted Apr 8, 2015 at 5:38 PM | Permalink

          HaroldW: Dang, that was supposed to be a reply to Steven Mosher at 1:02 PM.

          Here on Climate Audit, and also on Climate Etc., WordPress seems to have a one track mind today.

        • Posted Apr 8, 2015 at 8:23 PM | Permalink

          There is also a problem with WordPress plugins being exploited by hackers. Probably unrelated to the above, but moderators should be aware.
          http://arstechnica.com/security/2015/04/beware-of-pro-isis-script-kiddies-exploiting-wordpress-sites-fbi-warns/

        • Don Monfort
          Posted Apr 7, 2015 at 11:39 AM | Permalink

          prof. halpern says:

          “The source of water flowing off Nova Scotia is a balance between two currents, the cold Labrador current and the Gulf Stream. By looking at d15N in Sherwood’s proxy, one can conclude which of these currents dominate at any time.”

          So, we are interested in finding a proxy for the water temperature. Right, prof. halpern?

          prof. halpern says:

          “By looking at d15N in Sherwood’s proxy, one can conclude which of these currents dominate at any time”

          How do we know that? Because the d15N is a proxy for the nutrients in the water, which is determined by what prof. halpern?

          It’s the variations in temperature of the water that causes the variations in nutrients, which causes the variation in the d15N in the corals.

          Are alkenones better temperature proxies than corals, prof. halpern?

          Steve: Don, please calm down. In addition, I do not agree with your assertion that it is known that variation in water temperature that causes the variation in coral d15N – a claim that I certainly haven’t made. My point was that little is known about coral d15N and that the variation in coral d15N offered negligible support for Rahmstorf’s attempted history of subpolar gyre temperatures. I see no basis for the stronger assertions that you’re making. In the specific location, ocean temperatures depend in part on the interplay of the Labrador Current and the Gulf Stream, a point already made at CA, though ignored by Bill H

        • Don Monfort
          Posted Apr 7, 2015 at 12:38 PM | Permalink

          I’ll calm down. I can appreciate why you deleted some of my comments. And I suggest that you calm down, Steve. I haven’t put any words in your mouth. And I don’t see why I can’t make assertions that you don’t necessarily agree with. Is there a blog rule I don’t know about?

          Doesn’t the d15N in the coral vary because of the availability of nutrients in the water? Isn’t the availability of nutrients largely influenced by water temperature?

          Steve:”In the specific location, ocean temperatures depend in part on the interplay of the Labrador Current and the Gulf Stream, a point already made at CA, though ignored by Bill H’

          Yeah, that’s why Rahmstorf et al was interested in the coral d15N. They used it as a proxy for temperature.

          halpern:“The source of water flowing off Nova Scotia is a balance between two currents, the cold Labrador current and the Gulf Stream. By looking at d15N in Sherwood’s proxy, one can conclude which of these currents dominate at any time.”

          There’s the confirmation from the warmist camp. The d15N tells them whether the warm current or the cool current dominates. Warm and cool are temperature variants. Period. I am done here.

        • Carrick
          Posted Apr 7, 2015 at 12:56 PM | Permalink

          Don Monfort:

          Doesn’t the d15N in the coral vary because of the availability of nutrients in the water? Isn’t the availability of nutrients largely influenced by water temperature?

          Probably more importantly by where the current is coming from.

          For example,if it’s upwelling, it’s going to be nutrient poor.

          If it’s from tropical oceans in the plankton rich zone, it’ll be nutrient rich.

          This explains why you’d expect a correlation between temperature and N-15 (that is, it justifies the comparison of SST and N-15 concentration in Rahmstorf’s Figure 5), without predicting that N-15 makes a proxy that can be used to construct a temperature scale.

        • Don Monfort
          Posted Apr 7, 2015 at 1:33 PM | Permalink

          Thanks Carrick, We are talking about d15N in coral, in a specific location, where the cold Labrador current and the warm Gulfstream intermix.

          Rahmstorf et al are interested in whether the warm GS current dominates at any particular time, or the cold Lab. That’s why they are interested in d15N. Is it warm, or is it cold? I don’t think they are interested in nutrients, or as far as I know upwelling and downwelling at that particular place. They are interested in the temperature to determine which current is dominating. Is it warm, or is it cold? I don’t know how they demarcated warm from cold, because I ain’t going to pay to read a bogus paper. But warm and cold to me is about temperature. They might be interested in mass water movement, but they get at it by looking at temperature. The corals as used in Rahmstorf et all 2015 are a temperature proxy. JMVHO.

        • stevefitzpatrick
          Posted Apr 7, 2015 at 2:03 PM | Permalink

          Hi Carrick,
          I think you have this backwards. Cold upwelling water, or water in regions of deep seasonal convection, most often at high latitudes, is usually rich in nutrients, while tropical surface water is usually low in nutrients. Which is why warm tropical surface water tends to be extremely clear… not much microscopic life to scatter light. (eg, http://en.wikipedia.org/wiki/Upwelling http://earthobservatory.nasa.gov/IOTD/view.php?id=4097) When ENSO is in the El Nino phase (weakened eastern Pacific upwelling) fishing catch off the Pacific Coast of South America falls drastically relative to La Nina conditions.

        • Carrick
          Posted Apr 7, 2015 at 2:06 PM | Permalink

          Thanks for the correction Steve. I had a 50% chance of remembering it correctly, as it turns out. It’ll stick this time, pretty sure.

        • Steve McIntyre
          Posted Apr 7, 2015 at 2:40 PM | Permalink

          In deference to some misunderstandings, I’ve added some text to better clarify the perspective (arising from Sachs’ alkenone series) that I had in mind in this post. I had noted the connection at the close of the post but had not articulated it. I had been very struck by the similarity of the location of the Nova Scotia coldwater corals to Sachs’ alkenone series and show this in the diagram added today. I re-examined the post carefully to see if there were any errors and did not see any, though undoubtedly some points could have been made more clearly.

          I added the following:

          The Nova Scotia coldwater coral series are, like the alkenone series offshore east coast North America discussed at CA last month, located near the front between the Labrador Current and Gulf Stream, as shown in the location map from Sachs et al 2007, updated below to show the location of the Nova Scotia coldwater corals.

          sachs_2007_map_annotated_2

          In that earlier post, I reported that very large SST decreases had been estimated off the east coast during the Holocene, considerably larger than the North Atlantic as a whle. I noted the following explanation from Julian Sachs in terms of changing ocean currents:

          Sachs observed that a relatively small coastward displacement of the Gulf Stream could account for the difference and plausibly speculated that the Gulf Stream hugged the East Coast much more closely in the mid-Holocene.

          -Apr 7]

        • Steve McIntyre
          Posted Apr 7, 2015 at 2:49 PM | Permalink

          I think that both of you are mixing up nutrient content with d15N isotopes. The two are undoubtedly related but are not the same thing.

          Also, as I understand it, the Labrador Current is not “upwelling” water, but primarily return flow of cold water on more or less the same horizon from poleward flow of warm waters. In some recent articles, it is estimated that the volumes in the lateral gyre are much larger than volumes in the overturning circulation.

        • stevefitzpatrick
          Posted Apr 7, 2015 at 3:45 PM | Permalink

          Steve,
          I was not mixing these things up; I don’t know anything about dN15, and I make no comment about it. I was commenting only about how nutrients in ocean near-surface waters tend to vary with location due to upwelling and/or convective overturning.

        • little polyp
          Posted Apr 7, 2015 at 4:08 PM | Permalink

          not on topic but for those of us on little islands in the western pacific….the fish are truly jumping into the boat

        • mpainter
          Posted Apr 7, 2015 at 11:50 AM | Permalink

          Bill H:
          I can see for myself the spurious correlation between temperature and coral d15N
          posited in Rahmstorf et al and displayed in figure 5 of their study.

        • Michael Jankowski
          Posted Apr 7, 2015 at 7:29 PM | Permalink

          Exactly!

          “…Bulk d15N is most strongly correlated with NAO at a lag of 4 years (r= -0.19) and with temperature at a lag of 3 years (r=-0.27, p<0.05)…"

          Why call attention to this correlation and (incorrectly) plot it overlapping temperature anomaly if the intent is not to portray it as a proxy?

      • Bill H
        Posted Apr 7, 2015 at 3:43 AM | Permalink

        Steve, Thanks for your reply. The reason I regard your comments about Finnish sediments as a digression is that “Trick 3” is, as far as I , and indeed most others on this board who have actually addressed the point at issue can tell, is about using d15N, in your words, “as a thermometer”. An additional reason is that there is no evidence in the paper that they are using contaminated Finnish sediments as “a proxy for ocean currents”, though they are referencing global temperature reconstructions which use, inter alia, data from Finnish sediments, towards which I know you claim insufficient circumspection has been shown. I disagree with you on the lack of circumspection, and the agreement with subsequent PAGES reconstruction does seem to bear this out.

        You do seem to be moving away from your accusation that they are using d15N as “a thermometer”. Your suggestion that they should have used alkenone proxies – standard temperature proxies – instead of d15N is pretty convincing evidence for your earlier version of “trick 3”. However, your new version of Trick 3 seems rather nebulous: that they have produced a graph showing “squiggles” with no “statistical significance”, whatever that might mean, and therefore of no value.

        In an earlier post on this board I included a link to various correlation graphs using two separate vertical axes. These included graphs showing both negative and positive correlations between variables such as equity yields and US treasury yields. Are these “squiggles” equally “of no value”, merely tricks of scaling, and if not, why? Here they are again:

        https://www.creditwritedowns.com/2010/10/shifting-correlations.html

        • kim
          Posted Apr 7, 2015 at 4:14 AM | Permalink

          To correlate or not to correlate, that is approximately the question.
          =============

        • kim
          Posted Apr 7, 2015 at 4:42 AM | Permalink

          I really gotta thank Bill H for helping to doubleunderline the sketchiness of some of this stuff that goes wow to Nature and bang to the WaPo, galloping around the world while the poor damn coral is trying to get out of bed.
          ==============

        • HAS
          Posted Apr 7, 2015 at 5:44 AM | Permalink

          “Are these “squiggles” equally “of no value”, merely tricks of scaling”

          Quite possibly.

          As I noted below we (including hopefully Rahmstorf et al) know the linear relationship between temp and the δ15N tracer as calculated by Sherwood.

          In light of this knowledge if you were writing this paper would you flip the index upside down relative to temp and move them along a bit to give a better apparent fit?

          Or would you think that was being a bit too tricky?

        • Steve McIntyre
          Posted Apr 7, 2015 at 9:14 AM | Permalink

          Bill, once again, Rahmstorf’s and Mann’s use of contaminated data is not a “digression” but was the opening point in my commentary on Rahmstorf’s et al, which started here. I stated:

          Rahmstorf and Mann’s results are not based on proxies for Atlantic current velocity, but on a network consisting of contaminated Tiljander sediments (upside-down or not), Graybill’s stripbark bristlecone chronologies, Briffa MXD series truncated to hide-the-decline and hundreds of nondescript tree ring series statistically indistinguishable from white noise. In other words, they used the same much-criticized proxy network as Mann et al 2008-9. It’s hard to understand why anyone would seriously believe (let alone publish in peer reviewed literature) that Atlantic ocean currents could be reconstructed by such dreck, but Rahmstorf et al 2015 stands as evidence to the contrary. After so much controversy about Mann’s prior use of contaminated data, it defies credulity that he and Rahmstorf have done so once again.

          You now purport to justify their use of contaminated data as follows:

          An additional reason is that there is no evidence in the paper that they are using contaminated Finnish sediments as “a proxy for ocean currents”, though they are referencing global temperature reconstructions which use, inter alia, data from Finnish sediments, towards which I know you claim insufficient circumspection has been shown. I disagree with you on the lack of circumspection, and the agreement with subsequent PAGES reconstruction does seem to bear this out.

          First, I do not claim that there has been a “lack of circumspection” on the contaminated data. It is well known that the modern portion of the Tiljander data is contaminated by agricultural runoff. As I said, it defies credulity that Mann and Rahmstorf continue to use this data. Finnish paleolimnologist Atte Korhola supported my criticism of Mann’s use of the contaminated data as long ago as OCtober 2009 (see here). Kaufman et al 2009 conceded that they had used contaminated data.

          I did not raise Rahmstorf’s use of contaminated data to “move away” from another dispute. It was the first issue that I raised in respect of this proxy network. I am astonished that you believe that it is valid science to use the contaminated data as a proxy for Atlantic ocean currents.

          My point about Figure 5 is unaffected by your comparison. Rahmstorf’s figure 5 does not contain any statistical analysis of the supposed relationship between coral d15N offshore Nova Scotia and the subpolar gyre temperature, the key ingredient in their pseudo-AMOC index. Nor is my point about d15N being a little-known proxy something for me to “research”. I have detailed knowledge of the proxy literature and I am unaware of any prior use of this data by third parties. I have a longstanding criticism of ad hoc use of proxies, the properties of which have not been explored in multiple populations and re-iterated this criticism here. I stand by this criticism.

        • davideisenstadt
          Posted Apr 7, 2015 at 10:58 AM | Permalink

          but Steve, you miss the point: you believe that one must first establish a plausible relationship between a proxy and the phenomenon it is purported to represent before conducting any further analysis. This crew sees things differently. These guys go fishing for time series that have differing patterns of variance…these are simply colors in their palette…used to simulate some other time series.
          Physical relationships? geographical proximity?
          Feh. Youre hopelessly old school.
          Inversion of the time series?
          Contaminated time series?
          Misused time series?
          Cherry picked time series?
          These are totally immaterial to the analysis at hand.

        • Craig Loehle
          Posted Apr 7, 2015 at 9:18 AM | Permalink

          It seems that Bill H will not disavow the use of bristlecones in California and contaminated sediments in the proxy to represent something unclear about the circulation of the Atlantic. Bill, it is not about being “sufficiently circumspect” it is about an implausible and in part corrupted set of data relative to the phenomenon (circulation) being used with no attempt to prove they are relevant except that they sort of look the same as d15N data. You don’t keep using corrupted data just because it looks similar to the Pages graph–that doesn’t prove anything.
          So glad you clarified your position. Fail.

        • Bill H
          Posted Apr 8, 2015 at 3:22 AM | Permalink

          Craig,

          I keep being asked to “disavow” different things. First of all it’s stripbark bristlecones, now it’s all bristlecones. Am I to disavow entire sediment records in Finland on the grounds of partial contamination? Incidentally, the supposed problem with Stripbarks has been traced to to Graybill and Idso’s (1993) data processing rather than any problem with the Stripbarks themselves. As Salzer et al (2009) describe when Graybill and Idso’s raw data (as opposed to their “data after standardisation”) are plotted the problem of divergence in the 1850-1990 that their processed data shows disappears.

          Maybe Steve should write a post about Graybill and Idso’s “torturing and molesting” of Bristlecone data?

        • Carrick
          Posted Apr 8, 2015 at 4:48 AM | Permalink

          Bill H:

          m I to disavow entire sediment records in Finland on the grounds of partial contamination?

          How about just in places the original researches annotated as being contaminated specimens?

          I would consider that established practice, rather than allowing software code to override the judgment of the original researchers based on possibly specious correlations with temperature—especially when that code inverts the relationship between temperature and proxy noted by the original researchers.

          Incidentally, the supposed problem with Stripbarks has been traced to to Graybill and Idso’s (1993) data processing rather than any problem with the Stripbarks themselves. As Salzer et al (2009) describe when Graybill and Idso’s raw data (as opposed to their “data after standardisation”) are plotted the problem of divergence in the 1850-1990 that their processed data shows disappears.

          I think this is a more general issue than Graybill and Idso. Most of Mann 2008’s individual proxies don’t show a divergence for example.

          In general, I believe the divergence in tree-ring based proxy reconstructs is generated in the processing that attempts to create regional scale or global temperature reconstructions from the tree-ring proxies and the divergences are not inherent in the proxies themselves.

          [This does point to underlying issues with the proxies, in my opinion. I don’t think it’s as simple as the reconstruction algorithms are “torturing the data.”]

          Considering what strip-barking represents to tree health, I think it would be a very improbable claim that strip-barked tree-ring samples remain as valid temperature proxies, regardless of whether one can find literature to support it.

          The problems with supposedly temeprature-limited bristlecone growth also remains near legendary. From Salzer’s abstract:

          In the White Mountains of California, eight bristlecone pine (Pinus longaeva) tree-ring width chronologies were developed from trees at upper treeline and just below upper treeline along North- and South-facing elevational transects from treeline to ~90 m below. There is evidence for a climate-response threshold between approximately 60–80 vertical m below treeline, above which trees have shown a positive growth-response to temperature and below which they do not. Chronologies from 80 m or more below treeline show a change in climate response and do not correlate strongly with temperature-sensitive chronologies developed from trees growing at upper treeline. Rather, they more closely resemble lower elevation precipitation-sensitive chronologies.

          This is exactly what anybody who understands what affects tree growth would have predicted.

        • Posted Apr 8, 2015 at 5:14 AM | Permalink

          Try reading some of Steve posts on this. See

          https://climateaudit.org/2009/11/17/salzer-et-al-2009-a-first-look/

        • Steve McIntyre
          Posted Apr 8, 2015 at 9:11 AM | Permalink

          Maybe Steve should write a post about Graybill and Idso’s “torturing and molesting” of Bristlecone data?

          There have been many posts at Climate Audit about strip bark bristlecones. It appears that we have found common ground that Graybill’s bristlecone chronologies should not be used in temperature reconstructions, though for somewhat different reasons. I am quite content to stipulate that Graybill’s stripbark chronologies not be used in temperature reconstructions and welcome your agreement on this point.

        • Beta Blocker
          Posted Apr 8, 2015 at 9:43 AM | Permalink

          Bill H: As Salzer et al (2009) describe when Graybill and Idso’s raw data (as opposed to their “data after standardisation”) are plotted the problem of divergence in the 1850-1990 that their processed data shows disappears.

          Which implies that the indispensable power of appropriately-applied statistical mathematics can overcome the methodological issue that no one has yet published credible research which demonstrates that tree rings can reliably record atmospheric temperature information in ways which are separable from other physical effects such as moisture availability, nutrient availability, and location.

        • Craig Loehle
          Posted Apr 8, 2015 at 12:14 PM | Permalink

          Stripbark trees are damaged, where one side of the tree bark has died. They exhibit compensatory growth on the remaining side for many years. They should not be used for anything.
          The particular sediments in question are altered by agricultural runoff after a certain data. should not be used ever.
          Both are still used in this paper. Does it “matter”? If you keep using contaminated data, you ruin your credibility and possibly produce completely wrong results. The details matter in science. It has been found that many cancer cultures used for testing chemo drugs are not the cell types assumed but were taken over by aggressive melanoma cells due to lab contamination–it matters because drugs got to market that are irrelevant to the patient being treated, who likely died as a result. A single mistaken outlier in a regression can change the slope and therefore the conclusions. If you think the details don’t matter in your own work then simply don’t do it.

        • Carrick
          Posted Apr 8, 2015 at 3:51 PM | Permalink

          Craig Loehle:

          Stripbark trees are damaged, where one side of the tree bark has died. They exhibit compensatory growth on the remaining side for many years. They should not be used for anything.

          Absolutely agree on this point, except I would amend this to say “They should not be used for anything…except perhaps comic relief.”

          I don’t know any legitimate field of science where the use of such badly flawed data would be even tolerated, let alone vigorously defended in the manner that Bill H is apparently willing to do here.

        • Craig Loehle
          Posted Apr 8, 2015 at 4:23 PM | Permalink

          Carrick: Thanks. It seems some people don’t even understand that data can be screwed up/contaminated. You would not put a telescope in mexico city and expect to get good star spectrum data. You wouldn’t attempt to study electro-magnetism right next to a transformer. If you tried to study effects of cold medicine on people on life-support, you might not get useful info. Yet they don’t blink when told how goofy strip-bark trees are.

        • M Happold
          Posted Apr 9, 2015 at 8:38 AM | Permalink

          There is a reason for this: there is no measurable effect for what they are doing. Their work is purely observational and there is no way to check their results for, say, 1710 AD. Everything you mentioned in your list is experimental with measurable and comparable outcomes. If you give garbage data to a system controlling a robot, for example, it drives off a cliff. Give garbage data to one of their reconstruction algorithms, it simply produces a slightly different squiggle with no hope of any validation and no catastrophic consequences. So the squiggles that please them are the ones they publish. There is no danger of being shown to be wrong.

      • Bill H
        Posted Apr 8, 2015 at 10:49 AM | Permalink

        REPLY TO STEPHEN MCINTYRE

        Steve, What gave you the idea that I was agreeing with you re Stripbark pines being no good as proxies? If you read my post you will see that I am discussing how Salzer et al (2009) dispose of the objection of NAS to the use of Stripbark pines for paleo reconstruction, which was the 150 year-long divergence problem reported by Graybill and Idso in 1993. I note that you claim in a 2009 post that all Salzer et al have done is to “re-center” G and I’s divergence graph, but you don’t explain what this means.

        I still don’t see why Rahmstorf et al.’s use of Mann et al’s (2008) global temperature reconstruction is part of a “trick”. WHether or not the reconstruction is seriously flawed they make it clear that’s what they are using: where’s the deception. So, let’s get back to the original question of what exactly the “trick” is. You seem to be saying that what they have produced is of negligible value because they haven’t carried out statistical analysis. But how does that constitute a “trick”?

        And is it of no value? they have demonstrated a pretty strong correlation between quite disparate variables: something of interest in itself. You claim that you can demonstrate a correlation between any data sets with enough scaling of graphs strikes me as ludicrous. How would you for instance show correlation between an oscillatory function and a monotonic function (excluding the trivial case of making one or other axis scale so small that they both appear to be horizontal/vertical)?

        Lack of statistical analysis doth not a worthless paper make.

        • miker613
          Posted Apr 8, 2015 at 11:01 AM | Permalink

          Bill H, I’m having a lot of trouble following what you are doing here. You claimed that McIntyre made several serious mistakes, and your claim has been trumpeted on various pro-AGW websites. McIntyre has responded that you were mistaken that he misunderstood these points, he has modified his original post to clear up the misunderstandings.
          Now it would seem to be your turn. You should either (a) admit that you misunderstood him/there was a miscommunication and McIntyre’s post did not make that mistake (at least), or (b) write a much clearer comment explaining exactly what you think his mistake was, quoting his relevant quotes, quoting from Rahmstorf, etc., taking into account all McIntyre’s updates.
          Instead, all I’m seeing in the last day or so are complaints and comments on other people’s comments, and complaints about various other far-flung issues. Unless you do more I’m going to see that as a concession.

        • miker613
          Posted Apr 8, 2015 at 11:46 AM | Permalink

          This particular comment seems to be completely wrong. You may want to check with a statistician, or anyone who knows these things. Just because you found an example where there is no correlation doesn’t mean that there aren’t many many cases with spurious correlation. The GDP growth of Latvia (http://upload.wikimedia.org/wikipedia/commons/d/d9/Lv_real_gdp_growth.svg) and Standard Oil assets from 1885 to 1895 (http://en.wikipedia.org/wiki/Standard_Oil#/media/File:Standard_oil.gif) are very strongly correlated, because both of them are going up. Any two time series that change monotonically with time are strongly correlated. It means nothing.

        • miker613
          Posted Apr 8, 2015 at 11:47 AM | Permalink

          This comment was supposed to be in reply to https://climateaudit.org/2015/04/01/rahmstorfs-third-trick/#comment-756978

        • Steve McIntyre
          Posted Apr 8, 2015 at 12:44 PM | Permalink

          Bill, may I request that you take the simple courtesy of directly quoting remarks to which you take exception. You tend to race off without necessarily understanding the texts. Among such incidents in your present comment (not all of which I’m responding to right now), you say:

          What gave you the idea that I was agreeing with you re Stripbark pines being no good as proxies?

          But I did not make any such assumption. I only relied on your condemnation of the Graybill bristlecone chronologies for their supposed “data processing” errors. You said – and I make a practice of directly quoting people –

          the supposed problem with Stripbarks has been traced to to Graybill and Idso’s (1993) data processing rather than any problem with the Stripbarks themselves. As Salzer et al (2009) describe when Graybill and Idso’s raw data (as opposed to their “data after standardisation”) are plotted the problem of divergence in the 1850-1990 that their processed data shows disappears. Maybe Steve should write a post about Graybill and Idso’s “torturing and molesting” of Bristlecone data?

          From your criticism of Graybill’s “data processing”, I presumed that you rejected the Graybill chronologies as temperature proxies, while reserving the possibility that some other series computed from bristlecones might be a valid temperature proxy. I wrote back:

          It appears that we have found common ground that Graybill’s bristlecone chronologies should not be used in temperature reconstructions, though for somewhat different reasons.

          I did not attribute to you the belief that stripbark bristlecones were “no good as proxies”, only that the data as “molested and tortured” by Graybill as no good as a proxy – though for “different reasons”. While our reasons for rejecting Graybill stripbark chronologies were very different, sometimes people reach agreement for different reasons.

          While I do not agree with your reason for rejecting the Graybill stripbark chronologies, you seemed to strongly hold the view that Graybill had “tortured and molested” the data and that his chronologies should be set aside, a policy that, as I said, I am prepared to stipulate to, though for different reasons.

        • Sven
          Posted Apr 9, 2015 at 6:13 AM | Permalink

          Et, comme un éclair, M. Bill est disparu…

        • Hoi Polloi
          Posted Apr 9, 2015 at 8:11 AM | Permalink

          “Enjoy The Silence”

          This is how post normal climate science works.

        • Jeff Norman
          Posted Apr 9, 2015 at 8:59 AM | Permalink

          Perhaps taking his screen caps back to his bridge.

        • tomdesabla
          Posted Apr 10, 2015 at 9:03 AM | Permalink

          So, to clarify this in my mind – this is another moving pea situation right? We are not supposed to be talking about stripbark pines in general; we are talking about the specific stripbark chronologies used by Mann in 2008, and now by Rahmsdorf.

          So we are talking about Graybill and Idso 1993 period, and Bill H. admits that those chronologies are flawed. So glad we’ve cleared that up.

          Steve: Graybill had other stripbark chronologies other than those discussed in Graybill and Idso 1993, but they were done the same way. It’s always been a bizarre irony that the distinctive shape of the Mann hockeystick came from the Graybill stripbark chronologies. However, Bill H seems to have gone silent on the Graybill chronologies.

      • Bill H
        Posted Apr 8, 2015 at 11:06 AM | Permalink

        Steve, sorry, one more thing. You mentioned that you do not need to do any research to know that d15N is little-known proxy. I just did a search on Google Scholar and found 1100 literature references for it. Maybe not so little-known after all.

        Steve: coral d15N is not used as a proxy in any of the multiproxy reconstructions cited in AR4 or AR5. prior to late 2014, there wasn’t a single dataset archived at NOAA. Nor do the properties of coral d15N appear to be well understood. Further, many/most of the references to d15N in coral are connected to studies of sewage, fertlizer runoff and other non-climatic impacts. It’s not at all clear to me that a decrease/increase in coral d15N offshore Nova Scotia, let alone without accounting for nonclimatic effects, has very direct implications for the Atlantic Meridional Overturning Circulation. However, it obviously has a much better chance than the stripbark chronologies or Finnish lake sediments contaminated by agricultural runoff actually used by Rahmstorf and MAnn.

        • mpainter
          Posted Apr 8, 2015 at 4:33 PM | Permalink

          Bill H
          My understanding, gained from perusal of internet sources, is that d 15 N is much used as a proxy for inorganic vs organic sources of nitrogen and has been so utilized for decades. Naturally the internet will have an abundance of references to such.Nowhere could I find an example where the d15N was utilized as a temperature proxy.
          As I commented earlier, all seem to agree that d15N does not serve as a temperature proxy. Also it seems that any plot which attempts to show a correlation between d15N and temperature is highly suspect, particularly when no support is given for such a correlation.
          Also, SMc showed that Rahmstorf et al reversed the actual temperature trend to cooling (from warming), which reversal is spurious. There is the trick that you have been looking for.

  29. Craig Loehle
    Posted Apr 6, 2015 at 7:34 PM | Permalink

    Rahmstorf is using a temperature proxy to say something about the gyre. I don’t see where it is anything more than speculation. Throwing in a correlation with d15N does not help when no one knows what d15N means or if it has any relationship to the gyre. But in correlation-land, they don’t seem to be bothering with mechanisms…

    • Bill H
      Posted Apr 6, 2015 at 7:45 PM | Permalink

      Steve, Oh you can include Craig as being another of the people you misled into supposing that Rahmstorf et al. are claiming d15N is a “temperature proxy”.

      • HAS
        Posted Apr 6, 2015 at 8:03 PM | Permalink

        Good tricks never explicitly claim things, they are always done with suggestion.

        • tomdesabla
          Posted Apr 9, 2015 at 9:26 AM | Permalink

          Just so you know Bill, as a lurker who is admittedly better equipped to assess methods of argument than the details of math and science, you are not impressing me. It looks like two series are plotted together on the same graph for the purpose of showing that they are correlated. Clearly the implication is that the correlation has meaning.

          It looks to me like you are nitpicking, not making a reasonable criticism.

      • Craig Loehle
        Posted Apr 6, 2015 at 8:54 PM | Permalink

        Bill H: what are you talking about? Their proxy is based on tree rings which are an index of temperature. They show this correlating with d15N over time. I did not say d15N is a temperature proxy. I said no one knows what it is, least of all Rahmstorf.

    • Bill H
      Posted Apr 6, 2015 at 8:18 PM | Permalink

      Michael Jankowski, What are you getting at? There’s nothing to be “reconciled” if N-15 concentration isn’t a proxy for temperature. Are you saying that Rahmstorf et al. claim that it is? If so, on what grounds?

      • Michael Jankowski
        Posted Apr 7, 2015 at 12:10 PM | Permalink

        What I am getting at is quite clear. Sherwood et al show temperature and d15N moving in opposite directions…Rahmstorf et al shows temperature and d15N moving in the same direction. You can claim it’s not a proxy until you’re blue in the face, but that’s how Rahmstorf presented it…but backwards.

        • Carrick
          Posted Apr 7, 2015 at 12:50 PM | Permalink

          Michael Jankowski:

          You can claim it’s not a proxy until you’re blue in the face, but that’s how Rahmstorf presented it…but backwards.

          Again, Bill H isn’t claiming that N-15 concentration isn’t a proxy.

          Bill H is saying N-15 concentration is a proxy for water mass change.

          Bill H is saying -15 concentration is not a proxyfor sea temperature.

          I reiterated your point about the inconsistency issues with N-15 concentration below.

      • Carrick
        Posted Apr 7, 2015 at 12:46 PM | Permalink

        Bill H: I agree with you that Steve’s post certainly conflates N-15 concentration with sea temperature. For example here:

        But all that the coral d15N series show is (at most) that there has been increased temperature trends up in one region (offshore Nova Scotia) – and the significance of this relationship is very tenuous. On its face, the d15N series does not show that subpolar gyre temperatures have decreased.

        As you pointed out, N-15 concentration is an indicator of water mass change, not temperature, and Rahmstorf is correctly using it in this manner.

        I even agree that this conflation leads to confusion in the underlying point, because the underlying issue is the rate of Atlantic Ocean circulation, and not SST (as an indicator of rate of circulation).

        But noting the conflation of temperature with rate of circulation happens several time within his post just a quibble: It is not the case that Steve consistently conflated N-15 concentration with SST.

        So this error affects the presentation of Steve’s post, but not the substance, which I would summarize by this observation:

        Rahmstorf’s Figure 5 shows a positive correlation between temperature and coral d15N (both decline together), while Sherwood et al reported a negative correlation (r=-0.27) between temperature and coral d15N. How’d they do that?

        […]

        Sherwood et al 2011 stated that coral d15N was negatively correlated to gridcell temperature. Their figure 3 (a larger excerpt is shown on left) shows gridcell temperature increasing over the 20th century, while coral dN15 is going down. Sherwood et al 2011 attributed the negative correlation to higher d15N values in the cold Labrador Slope Water (LSW) relative to the warm Western Slope Water (WSW)

        This seems like a really big deal to me.

        On the other hand, when you say:

        There’s nothing to be “reconciled” if N-15 concentration isn’t a proxy for temperature.

        That’s not actually true.

        The proxies for temperature and water mass change are being used in Rahmstorf to argue for a change in the rate of Atlantic Ocean circulation.

        If N-15 concentration isn’t a consistent indicator of Atlantic water mass change, which I think is actually the main point in Steve’s post, then there is something that needs to be reconciled.

        This is not a quibble, not even close. Rather it strikes to the heart of the Rahmstorf paper.

        This is, in my opinion, a substantive issue that needs to be addressed.

        Steve: Carrick, in a very recent post (which was in my mind when writing the above post), I showed a diagram of the Labrador Current and Gulf Stream offshore east coast North America. While I may not have spelled this out in complete detail in the present post, I am obviously aware of the differences between the Labrador Current and Gulf Stream. Because the two currents have very different temperatures, Julian Sachs (See discussion here) took the position that changes in SST through the Holocene reflected varying proportions of these two currents. In my earlier post, I noted up Sachs’ comment that slight displacements of the Gulf Stream could cause disproportionate changes in SSTs offshore east coast North America. While I may not have spelled this point out sufficiently for someone who like Bill H contests whether the Finnish sediments are contaminated, I do not see any incorrect understanding of this in my post.

        • Tom Gray
          Posted Apr 7, 2015 at 1:59 PM | Permalink

          Somehow I must have become brilliant because I understood all of this water mass/temperature stuff from Steve McIntyre’s postings and previously I was not in anyway brilliant. I couldn’t quite understand what BillH was writing about not being a temperature because it couldn’t have been that obvious. I see Michael Mann describing SMc’s posting as some kind of error. I couldn’t really understand his comments since the temperature proxy for water mass idea was quite a straightforward idea of the type that I have seen in my engineering career. Where did SMc make a mistake?

        • Carrick
          Posted Apr 8, 2015 at 4:14 AM | Permalink

          Tom Gray, I think it’s just an issue with language, but there are multiple places where N-15 concentration get conflated with temperature. I gave one example above in the comment you just replied to: “But all that the coral d15N series show is (at most) that there has been increased temperature trends”.

          According to Rahmstorf’s paper, which Steve is clearly critiquing in this blog post, “increased temperature trends” are not what the coral d15N is showing.

          Since you can obviously read, I’ll let you COMMAND-F (or CONTROL-F) to find the others.

          I didn’t count the number but in most places where N-15 concentration are discussed, Steve is referring properly to water mass changes rather than to temperature change.

          Steve McIntyre:

          Because the two currents have very different temperatures, Julian Sachs (See discussion here) took the position that changes in SST through the Holocene reflected varying proportions of these two currents.

          Technically what you have in this case is a model for why you’d expect N-15 concentration to track SST in this case.

          But that still doesn’t make N-15 concentration a temperature proxy. It’s still just a proxy (in the best case scenario) forchanges in water mass, that happens to track here with temperature.

          While I may not have spelled this point out sufficiently for someone who like Bill H contests whether the Finnish sediments are contaminated, I do not see any incorrect understanding of this in my post.

          I don’ view this as an issue with understanding, but rather one of clarity in language. HotWhopper has just told another one, apparently. It’s about all Sou is good for.

          […] whether the Finnish sediments are contaminated

          Or whether d15N is itself a contaminated proxy in this case.

        • Bill H
          Posted Apr 8, 2015 at 6:19 AM | Permalink

          Carrick, The disparity between correlations reported by Sherwood et al and by rahmstorf et al is hardly surprising when one considers that they were investigating locations about 2,000 km apart with quite different patterns of nutirent flow. Another mistake Steve made inhis original post was to claim that Rahmstorf et al were investigating an area off the coast of Nova Scotia. He’s confusing the two papers.

          Steve: Bill, you continue to fabricate stuff. I did not “claim” that Rahmstorf et al were investigating an area offcoast Nova Scotia. Conspicuously, you did not provide a single quotation to support your false claim that I “confused” the two papers. Whatever my defects may be, this sort of elementary confusion isn’t one of them. Your continued fabrications are becoming tiresome.

          I definitely challenge whether coral d15N data offshore Nova Scotia is a useful proxy for ocean temperatures over 2000 km away. But I challenge even more strongly whether contaminated Finnish sediments and stripbark bristlecones are useful proxies for Atlantic ocean currents – the ludicrous claim by RAhmstorf and MAnn that you endorse.

        • Bill H
          Posted Apr 8, 2015 at 11:46 AM | Permalink

          Steve, I apologise for saying that you thought Rahmstorf et al. were discussing temperatures in near Nova Scotia. That was a misreading of the opening of my post, and not a deliberate fabrication as you have assumed.

          I would be interested to know what my other “fabrications” are. I know you specialise in the exposure of fraud, but do please remember the rule: “Don’t ascribe to malice what can be adequately explained by incompetence”

          Steve: you say that “you know that [I] specialize in the exposure of fraud”. Do you have any basis for this allegation or did you make it up? Obviously I’ve criticized people for making false statements and take considerable care in being accurate in such criticism. I have very seldom made accusations of “fraud”, have blog policies against commenters making such accusations and have publicly discouraged the mentality of making fraud accusations. Even Mann acknowledged this in his pleadings in Mann v Steyn.

  30. Pat Frank
    Posted Apr 6, 2015 at 8:30 PM | Permalink

    Figure 5, in Sherwood, et al., reference 25 in Mann & Rahmstorf, provides M&R the, “four more data points from ancient corals preceding the twentieth century, the oldest one from AD ~500. These lie all above 10.5‰, providing (albeit limited) evidence that the downward excursion to values below 10‰ between 1975 and 1995 and the corresponding water mass change may be unprecedented in several centuries.

    The four data points are at CE 600(+/-)200, 1770(+/-)150, 1810(+/-)130, and 1860(+/-)240. The vertical N-15 values are all about 10.9(+/-)0.06%%.

    The earliest point is after the end of the Roman Warm Period, and the other three are during the final fourth of the LIA. None of them represent periods of climate warmth comparable to the present.

    Therefore, these data cannot support the very definitive-sounding conclusion made by Sherwood, et al., that, “the persistence of the warm, nutrient-rich regime since the early 1970s is largely unique in the context of the last approximately 1,800 yr.” They can’t know that uniqueness in the absence of Roman and Medieval N-15 proxy data.

    Mann & Rahmstorf’s derivative claim of “unprecedented in several centuries” is hardly more supportable. The data begin in 1925 during recovery from the LIA. The N-15 proxy declines over this whole period, from 1925 on through to the present.

    The lack of proxy data from the Roman and Medieval times leaves the modern data disconnected from any comparative time. One is left with a proxy decline that has no larger relevance, because no one knows whether there were comparable proxy declines in earlier warm climate regimes.

    The whole Figure deserves no more than a ‘so, what’? It’s not telling us anything new.

  31. David Springer
    Posted Apr 7, 2015 at 2:36 AM | Permalink

    Looks like Bill H lost this from word go by refusing to accept the fact that figure 5 is squiggle matching to show that coral squiggle has high correlation with temperature squiggles.

  32. Salamano
    Posted Apr 7, 2015 at 7:21 AM | Permalink

    Apparently there’s some preening over at HotWhopper that I saw Michael Mann tweet out…

    http://blog.hotwhopper.com/2015/04/steve-mcintyres-big-blooper-mistaking.html

    Some mistakes being made..?

  33. RoyFOMR
    Posted Apr 7, 2015 at 9:36 AM | Permalink

    Michael Mann has already got his FB account out with ‘Steve-McIntyres big blooper – mistaking water mass movement for water temperature’ via HotWhopper.
    Looks like Bill H accepted and achieved the mission.

  34. Steve McIntyre
    Posted Apr 7, 2015 at 11:52 AM | Permalink

    Another of the very few articles on coral d15N is Sherwood et al 2014, which discusses d15N values in the North Pacific Subpolar Gyre (near Hawaii). These values also decline in the 20th century without the help of the Atlantic subpolar gyre.

    sherwood_2014_np_annotated

    On this occasion, Sherwood et al posit a variety of hypotheses, including (1) an expansion of the North Pacific Subpolar Gyre since the Little Ice Age; (2) an increase in N2 fixation linked to the supply of iron-bearing dust aerosols.

    They also note that terrestrial runoff of fertilizers (values
    near 22%) can depress local seawater d15N, noting that this is not an issue in the Pacific subtropical gyre, though it’s something that would have to be checked offshore Nova Scotia. There is no evidence that Rahmstorf carried out such a check.

    • Frank
      Posted Apr 7, 2015 at 12:59 PM | Permalink

      Steve, because the fig.5 is very small and smoothed in the paper I have re-done the operation of Rahmstorf et.al and made a high resolution comarison with a high resoltion figure from the supps of the Sherwood-paper ( http://www.pnas.org/content/suppl/2011/01/03/1004904108.DCSupplemental/pnas.1004904108_SI.pdf#STXT ). The result:
      http://fs2.directupload.net/images/150407/6z754jp6.jpg . It looks similiar, okay. Anyway, there are very big deviations between the records: look at the sharp dips in 1943 and 1947 obs ( bottom) , not replicated in the proxys( upper part of the figure). The max. in decline was in 1994 (obs), in the proxys it was in 1980. The sharp increase after 1994: not replicated in the proxys. In my opinion there is no statistical significant “teleconnection”. The basis of the paper?

  35. mpainter
    Posted Apr 7, 2015 at 2:10 PM | Permalink

    Most definitely, Carrick, you need to read up on upwelling of deep oceanic water and learn about the rich nutrient supply thereof. Such upwelling provides the richest fishing because of the nutrient base circulated to the photic zone.

  36. human1ty1st
    Posted Apr 7, 2015 at 9:31 PM | Permalink

    OT … but did you see the new paper from a whole bunch of European Paleoclimatologists ( I recognise Esper and Ljungqvist) that looking at the issue of the limitations of trees as thermometers

    http://journals.ametsoc.org/doi/abs/10.1175/JCLI-D-14-00673.1

  37. human1ty1st
    Posted Apr 7, 2015 at 9:32 PM | Permalink

    http://journals.ametsoc.org/doi/abs/10.1175/JCLI-D-14-00673.1

    Limitations of trees as thermometers. Some interesting authors.

  38. Bill H
    Posted Apr 8, 2015 at 3:40 AM | Permalink

    Steve (McIntyre), despite our differences I’d like to thank you for deleting some of the more unkind ad hominems against me. I notice there’s still one accusing me of paranoia and one about my being a “baby daddy” to Rahmstorf, but in view of the amount of bile to start with you’ve cleaned things up very well.

    With Thanks, Bill

    Steve: I moderate after the fact and have not been online all that much the last few days. I ask readers to comply with blog policies and unfortunately not all readers have done so in the past few days. I apologize for this. I sometimes miss comments that do not comply with blog policy, but will go back and deal with others.

    • David Young
      Posted Apr 11, 2015 at 8:16 PM | Permalink

      If only Sou exercised the same restraint and civility as Steve does here. She is unusually personal and nasty and allows all sorts of insults.

  39. Posted Apr 8, 2015 at 10:24 PM | Permalink

    This post is nearly incomprehensible. Links to the referenced papers would help – as would using their actual titles.

    For instance, Sherwood et al 2011 is “Nutrient regime shift in the western North Atlantic indicated by compound-specific δ15N of deep-sea gorgonian corals

    The paper says: “The δ15N of organic material has long been used as a tracer of both trophic level of organisms and of the source of nitrogen (N) at the base of the food web.”

    Google Scholar indeed yields 26,900 results for ‘δ15N’; 13,500 results for ‘δ15N trophic’; and 1,890 results for ‘δ15N trophic coral’. The vast majority of all these search results uses or addresses δ15N value as a proxy.

    Contrast this to what this blog post asserts: “Coral d15N is not a well-studied proxy, to say the least.” Ahem. A more accurate statement by our host would have been. “I am not that familiar with this area of study and the use of specific amino acids in stable isotope analysis to untangle the Nitrogen source versus subsequent trophic transfers is a relatively new technique.”

    Of course if our host *had* said that – it would have made clear why he botched the rest of the analysis. Our host has conflated temperature with mass, conflated different geographic areas, and seems obsessed with temperature – though neither Sherwood et al nor the putative subject of this post (“Exceptional twentieth-century slowdown in Atlantic Ocean overturning circulation” Rahmstorf et al 2015, doi:10.1038/nclimate2554) are really ‘temperature’ papers. They are both about water movement. Sherwood using δ15N to determine the source of Nitrogen in an area off of Nova Scotia (subtropical versus subpolar slope waters) and Rahmstorf looking at the AMOC. Thermohaline circulation is a subject that Rahmstorf has been writing about for 20 years.

    There is nothing controversial in the Sherwood paper. The Sherwood results support Rahmstorf’s view that the recent weakening of the AMOC is unique in the past couple thousand years. Believing otherwise is a WUWT level mistake.

  40. Ed Snack
    Posted Apr 8, 2015 at 11:01 PM | Permalink

    On Bristlecones, perhaps Bill H would care to comment on why Mann (and others) continue to use the Idso & Gray records when a later collection is available courtesy of Linah Ababneh. Mann and co should be aware of it, her PhD supervisor was one of the team, and she was awarded her doctorate on the basis of the study.

    Why then, are the results ignored and in fact are under some sort of embargo ? Surely it couldn’t be that the results don’t fit the narrative could it ? BTW I have seen the results in that brief period when they were available, and there certainly didn’t appear to be a “stick” to see. Same coverage as I&G but apparently ignored strip-bark samples mostly.

  41. DB
    Posted Apr 9, 2015 at 7:04 AM | Permalink

    It would appear that Bill H got what he wanted for Mann’s Hot Whopper and left, unable to engage in a level of civil debate that involves actually quoting authors instead of inserting strawmen into what others say to make his argument. From the way things developed it seems he just was a parrot for Mann, to try and throw out arguments not based on anything SM said, but to provoke a desired response on items entirely disconnected.

  42. donna summer
    Posted Apr 9, 2015 at 9:04 AM | Permalink

    hi Steve,

    i think you mad some very wrong conclusions…see here:

    http://blog.hotwhopper.com/2015/04/steve-mcintyres-big-blooper-mistaking.html?spref=tw

    • DB
      Posted Apr 9, 2015 at 9:23 AM | Permalink

      I think you needed to continue reading responses instead of cherry picks from a longer conversation by a vested party interest. That seems to be difficult, but try, despite the difficulty you may have following Bill H’s train of thought.

    • HAS
      Posted Apr 9, 2015 at 4:41 PM | Permalink

      donna, you’ll see from the thread that I’m trying to understand how the AMOC index got created and the potential for the apparent relationship between it and the d15N tracer being simply an artifact of that. Unfortunately no one over there seems to be able to help and the paper itself is opaque on this point.

      Can you help?

    • HAS
      Posted Apr 10, 2015 at 1:45 AM | Permalink

      Got myself banned at Hotwhopper. The Aussies have got sensitive now our dollar is almost at parity (the Canadians seem more relaxed about being relegated).

      However in the process I did work out/realise a number of things about Rahmstorf and related issues:

      – They don’t report any statistics on the relationship between the d15N tracer and their AMOC index
      – It isn’t clear how the index is constructed (this is more an issue with the next point)
      – The d15N tracer being correlated with NOA clearly then will likely have relationships with AMOC and the components of their index (if I knew how it was done)
      – GCMs don’t do SST well in the `subpolar gyre’ but somehow they are used in the reconstruction of the index based on SST.

      I was trolling and off topic. Said I thought Rahmstorf was lightweight so probably was the latter.

      • Bill H
        Posted Apr 10, 2015 at 5:29 PM | Permalink

        HAS, Maybe the reason you were “banned” might have had something to do with the fact that you never once expressed a word of gratitude for the huge amount of assistance you received particularly from Sou but also from Rattus Norvegicus and even a little from me. Instead you were insulting and disdainful . You sneered at Rattus simply because he hadn’t read the original paper, even though he had read commentaries thereon whic gave him a rather better understanding than Steve seems to have. You then sneered at me on the evidence-free grounds that I was at the limits of my understanding simply because I had rebuked you for your earlier rudeness and exposed the logical absurdity of your position.

        Now having shown all this rudeness, disdain and total ingratitude when offered copious, seriously copious assistance you now have the brass neck to take the moral high ground because Sou finally decided to block your trolling.

        Oh, and HAS, I suppose it’s pure coincidence that when Steve found himself in very hot water over the maths after criticising Lewandowsky’s paper there was a character called HAS producing a very large number of very repetitive “criticisms” of Lewandowsky. Funny how when Steve’s in trouble a guy called HAS goes on a troll offensive against whoever has had the temerity to challenge Steve.

        Steve: Bill H, once again you are fabricating stuff. I was never in “hot water over the maths” in criticizing Lewandowsky. Lewandowsky’s results were highly flawed.

        • HAS
          Posted Apr 10, 2015 at 6:35 PM | Permalink

          I suspect that my personal attributes are well off topic but for any reader that feels I might exhibit these traits please read the thread and make your own assessmenthttp://blog.hotwhopper.com/2015/04/steve-mcintyres-big-blooper-mistaking.html?showComment=1428489066533#c1430637422105294318.

          And as I recall on the Lewandowsky issue my concerns about the paper were separate from and prior in the experimental design from our host’s concerns and don’t recall any hot water being present. What SM was saying was just based on mainstream statistical analysis and experimental design (as were mine).

          Coming back on topic, my point here simply extends on SM’s. I ignored the various physical interpretations being placed on the data and time series, and tried to work out whether the tracer/index relationship could be inherent in the structure of the index.

          To repeat what I’ve said above: the dN15 tracer was correlated with the NOA that in turn was correlated the components of the AMOC index and the AMOC itself, so it was likely the relationship could derive from this source; I couldn’t recreate the index from the paper (and when I started looking at the millennium reconstruction found it appeared to rely on GCMs that apparently couldn’t reproduce the SST in the area); and the paper reported no statistics on the tracer index/relationship, when this seemed an obvious thing to do under the circumstances. I expressed a view on the quality of the paper and transparency of the authors in this regard, but in the end suggested we just agree to differ on this point.

          Putting aside the physical interpretation of the data didn’t seem to be something commentators at HotWhopper were either familiar or comfortable with. We never really got past go.

        • Gerald Machnee
          Posted Apr 10, 2015 at 11:55 PM | Permalink

          Bill H, you come to this site and make comments without supporting evidence (of course you assume you are correct), then you go to blogs like Unrealclimate and brag that you found errors at Climateaudit. And of course the sheep there go and repeat your statements such as #113,#114, and #115. You ASSUME reasons that HAS was banned and you again fabricate statements about HAS critiquing Lewandowsky. Everyone has the right to do so and Lewandowsky has made enough goofs to be criticized. I will not repeat them as they have been cover here and in WUWT. You can look them up. HAS does not need to thank SOU or anyone else for “helping”. You complained at the other site that you were not thanked for “pointing out hie error”.
          Well, you did not find an error, but you have the freedom to say anything on those sites as any of us who attempt to correct it are deleted no matter how polite we are. Yes, maybe you are a hero on that site, but then there is no opposition allowed. You should be thankful you have been given the opportunity to comment here even though is has been much like race horsing.

        • HAS
          Posted Apr 11, 2015 at 4:57 AM | Permalink

          Just to note that Sou at Hotwhopper has checked with the authors and the index is simply average SPG temp – average NH temp. My eyesight is not what is was and the typography leaves something to be desired (but it is a bit of an object lesson in relying on only publishing graphs to document your work – not wishing to mention the tracer/index relationships).

          Quite what such an index represents I’m not sure. It isn’t immediately clear how accurately it portrays the deviation of the SPG area temps from the rest of the NH temps, which is what the authors claim for it.

          Putting that aside, given the negative lagged correlation between dN15 and NAO, and NAO as I understand it not being correlated with SPG at high frequencies, but being positively correlated with NH temp, a positive relationship between the index and dN15 would be indicated.

          A more detailed analysis of the index/tracer relationship seems to be called for before attributing particular physical phenomena to it.

        • Frank
          Posted Apr 11, 2015 at 10:21 AM | Permalink

          HAS, a more detailed analysis between (SSTspg – Tnh) and the Proxies just like this one? https://climateaudit.org/2015/04/01/rahmstorfs-third-trick/#comment-756896

        • HAS
          Posted Apr 11, 2015 at 2:26 PM | Permalink

          Frnak, yes and perhaps some of the statistics on the relationships I’m sure they did.

        • Carrick
          Posted Apr 11, 2015 at 7:47 PM | Permalink

          Steve:

          Steve: Bill H, once again you are fabricating stuff. I was never in “hot water over the maths” in criticizing Lewandowsky. Lewandowsky’s results were highly flawed.

          Sorry it’s very hard for me to picture how somebody with Sou’s primitive math skills is going to get Steve in “hot water over the maths”.

        • HAS
          Posted Apr 13, 2015 at 5:43 PM | Permalink

          I did have some time and digitalised a couple of the series (d15N and AMOC) using my now demonstrated fallible eyesight and calculated SPG temp from GISS NH temp (and the Hurrell December to March station-based NAO series).

          On first blush d15N and AMOC look pretty good and the relationships with SPG temp and NH temp (-ve) all look good too, but not quite so impressive.

          But here’s the thing, the relationship with time was pretty good in all cases too (as is obvious in retrospect from the graphs). Putting aside the debate about exactly which ARIMA form these series take and living with the limitations of using Excel to do any sensible analysis, I did look at first differences of the series (no consideration of lags or autocorrelations).

          No discernible relationship between first differences of d15N and AMOC or NI temp or SPG temp (the latter series are pretty highly intercorrelated). Putting aside philosophical arguments about what constitutes evidence of causality, perhaps an explanation for why no analysis of the relationship was reported. Leading or lagging first differences of d15N and AMOC by a couple of years makes no difference.

          As an aside first differences of d15N and the NAO index did appear significant, which would be consistent with what Sherwood reports.

          So I was wrong about the thought that it might be the relationship between d15N and ANO that was driving the apparent relationship with AMOC, it was the common ARIMA structure that seems to cause it.

        • Sven
          Posted Apr 14, 2015 at 5:36 AM | Permalink

          You are quite a hero in your circles, Bill, aren’t you, doing your strange and inaccurate insinuations/shoot-bys over here and then going around in your camp’s grateful audiences and bragging about them? Hey, look what I told them! Childish!

          “You sneered at Rattus simply because he hadn’t read the original paper…”
          You somehow failed to include that the “sneering” at the Rattus’ “teaching moment” (without having read the paper), started with Rattus (who was the one being wrong) calling HAS a dimbulb. Now why is that, Bill?

        • Sven
          Posted Apr 14, 2015 at 5:53 AM | Permalink

          Just in from Sou, on HAS’s highly relevant comment (showing extremely convincingly that Sou was wrong from the beginning):

          “HAS’s comment has been moved to the the HotWhoppery.”

          The End

  43. mpainter
    Posted Apr 9, 2015 at 10:24 AM | Permalink

    Matt Skaggs:
    Comment nesting has flown the coop.

  44. MikeN
    Posted Apr 11, 2015 at 10:56 AM | Permalink

    Has Mann 2008 or 2009 been corrected? I ask because I notice that Pages 2K EIV shows modern warming as not unprecedented.

    • Beta Blocker
      Posted Apr 11, 2015 at 12:43 PM | Permalink

      MikeN: Has Mann 2008 or 2009 been corrected? I ask because I notice that Pages 2K EIV shows modern warming as not unprecedented.

      MikeN, it would be a useful exercise in comparing various interpretations of hockey stick iconography for someone with access to the baseline data to plot Mann 2009 on top of Pages 2k EIV, and then also to plot Mann 2015 on top of Pages 2k EIV, using a common datum for both comparisons.

      If this were to be done, would one picture be worth a thousand Hotwhopper/Bill H words?

  45. MikeN
    Posted Apr 15, 2015 at 1:11 PM | Permalink

    d15N and teleconnections, but has nothing to do with Mann!

    “That’s one thing this kind of research is really helpful for — showing the teleconnections in the climate system,” Gibson says. “So you see something in this one 4,000-square-kilometer basin off the northeast coast of Venezuela, but you see similar changes in the Arabian Sea and in the tropical Pacific, and you can link it all back to changes seen in an ice sheet in Greenland.

    http://wattsupwiththat.com/2015/04/15/strong-evidence-for-rapid-climate-change-found-in-past-millenia/

  46. RCB
    Posted May 23, 2015 at 1:31 PM | Permalink

    Hello Mr. McIntyre – greetings from the States! I hope that you and your family are all well.

    I thoroughly enjoy the topics presented in your weblog, and have followed your presentations for some years now. (I share this site with all of my friends.)

    I am very much looking forward to your next topics.

    Best regards,
    RCB

5 Trackbacks

  1. […] Rahmstorf did not actually measure velocity of the Gulf Stream; instead he used a “multi-proxy temperature reconstruction” to postulate that cooling of the Atlantic meridional overturning circulation (AMOC), possibly from increased ice melt, may be slowing Gulf Stream velocity. Rahmstorf also called this an “unprecedented event.” Such characterization seems to be a favorite among climate alarmists. Rahmstorf ignored natural cycles in Gulf Stream velocity such as the Atlantic Multidecadal Oscillation. (Mathematician Steve McIntyre takes Rahmstorf’s reconstruction methods apart in three very technical posts here  , here. and here) […]

  2. […] Rahmstorf’s third trick […]

  3. […] Rahmstorf’s third trick […]

  4. […] https://climateaudit.org/2015/04/01/rahmstorfs-third-trick/ […]

  5. […] fagfolk, deretter fulgte flere kommentarer fra Steve McIntyre da Rahmstorf hadde flere triks. Hele tre til nå. I korthet kan det slås fast at artikkelen og medias fremstilling ikke hadde tatt hensyn til en […]

Follow

Get every new post delivered to your Inbox.

Join 3,633 other followers

%d bloggers like this: