D-day: Loehle and Scafetta create the beachhead

Posted: July 26, 2011 by Rog Tallbloke in Astrophysics, climate, Ocean dynamics, Solar physics, solar system dynamics

A new paper by Nicola Scafetta and Craig Loehle has been published in the Open Atmospheric Science Journal 5:74-86.
Building on Scafetta’s previous recent papers it posits that 60 and 20 year cycles linked to the solar barycentric orbit affect the terrestrial climate and coupled with a trend possibly due to co2 corresponding to the non-feedback value and a low end climate sensitivity, explains the historical surface temperature record. They then provide a projection to 2100 which predicts a 0.66C/century warming due to co2. However, in comments on WUWT, Nicola says:

We clearly state in the paper that there are other cycles such a the millennial one explaining the MWP and LIA. However, in this paper we are dealing only with the data since 1850. The first approximation that can be done with this data is a linear one which yields to an upper estimate for our 21st forecast.

Have Loehle and Scafetta played a clever card to gain a foothold in the literature for barycentric cycles linked to climate change, or are they rebranding themselves as lukewarmers to curry favour with the mainstream climate science institutions? Judging by initial reaction on Judy Curry’s blog and WUWT, Steve Mosher may have a point when he says:

Since I know you both believe in radiative physics as your paper indicates, your estimate of sensitivity, puts you at the lower end of luke warmer. your bumper stickers are in the mail. prepare to be savaged by both sides.

Steve knows a thing or two about that situation, so we should tread lightly as we assess this interesting paper.

Climate Change Attribution Using Empirical Decomposition of Climatic Data
Craig Loehle and Nicola Scafetta

How do we detect the influence of humans on the climate system? Current methods based on climate models are unfortunately circular: their estimate of human effects is only valid if the models are correct, but the models make certain assumptions and also are fitted to the historical temperature record. A model-independent estimate of climate response is needed and is provided by this study.

The climate change attribution problem was addressed using empirical decomposition. Previously observed cycles in solar motion and activity of 60 and 20 years were used to develop an empirical model of Earth temperature variations. The model was fit to the Hadley global temperature data up to 1950 (the time period before anthropogenic (GHG+Aerosol) emissions became a significant forcing mechanism), and then extrapolated from 1951 to 2010 (Fig. 1A). The residuals (Fig.1B) showed an approximate linear upward trend after 1942. It is assumed that this residual upward warming has been mostly induced by anthropogenic emissions, urbanization and land use change. The warming observed before 1942 is relatively small and is assumed to have been mostly naturally induced because anthropogenic (warming + cooling) forcing would approximately compensate each other before 1950.

Figure 1.

The resulting full natural plus anthropogenic model (below, Fig. 2) fits the entire 160 year record very well. Residual analysis does not provide any evidence for a substantial cooling effect due to sulfate aerosols from 1940 to 1970. In fact, the cooling observed during that period is well predicted by a natural 60-year cycle, which from 1940 to 1970 was in its cooling phase and contributed about 0.3 oC cooling, plus an estimated +0.66 oC/century anthropogenic Note that a quasi 60-year cycle is visible in the global temperature since 1850 and has been observed also in numerous multisecular climatic records. New solar activity proxy models developed in the paper suggest a mechanism for both the 60-year climate cycle and a portion of the long-term warming trend. About 60% of the warming observed from 1970 to 2000 was very likely caused by this natural 60-year climatic cycle during its warming phase. Figure 2B shows the components of the signal in our model.

Figure 2.

A 21st Century forecast (below, Fig. 3) suggests that climate may remain approximately steady until 2030-2040, and may at most warm 0.5-1.0°C by 2100 at the estimated 0.66°C/century anthropogenic warming rate, which is about 3.5 times smaller than the average 2.3°C/century anthropogenic warming rate projected by the IPCC during the first decades of the 21st century.

Figure 3.

Conclusions

1) The estimated AGW component matches theory, since the log of an exponential rise in carbon dioxide should give an approximatelinear trend (as in fact the climate models do). The timing of AGW effects (beginning in 1942) also matches expectations.

2) The fitted components match solar model forcings within their uncertainty.

3) The estimated sensitivity matches a no-amplification (neutral) climate sensitivity, or even a slight negative feedback case.

4) Warming due to anthropogenic GHG+Aerosol of 0.66 oC/Century is not alarming, in comparison to the IPCC protected 2.3 oC/Century This 0.66 value is an upper bound in our estimation (due to possible poorly corrected UHI and LULC effects that may explain part of the observed warming trend since 1950).

5) Cooling/flat temperatures till 2030 are likely (as also predicted by others).

6) Our result matches the historical record better than any other attribution study and better than GCM output

The paper and the SI are available at WUWT

Loehle & Scafetta 2011 (full paper PDF)

Loehle & Scafetta Supplemental Info (PDF)

Comments
  1. adolfogiurfa says:

    Scafetta refers to LOD in previous papers.
    55 year cycle seen in LOD:
    ftp://ftp.fao.org/docrep/fao/005/y2787e/
    See graph on page 50, at:
    ftp://ftp.fao.org/docrep/fao/005/y2787e/y2787e08.pdf
    Then, there will be smaller variations within this curve: As we are in a general downward direction, this varies again with the evolution of the current solar cycle:Warmer when more active and cooler when less active.

  2. Doug Proctor says:

    By my understanding (far from perfect):
    1. Figure 1 shows the temperature influence by the barycentric model until about 1942, and then the combined barycentric + CO2 influence since 1942.
    2. In Figure 1 there is a linear trend to the barycentric model, such that between 1880 & 2000 the temperature attributable to the barycentric model increased 0.2C.
    3. Between 1942 and 1965 the barycentric model predicts a COOLING of 0.25C, while temperatures stayed the same.
    3. The1965 – present linear trend at 0.66C/century represents the COMBINED influences on temperatures since 1965.
    4. Removing the linear barycentric influence reveals the CO2 influence (assumption: only two significant processes acting on our climate!!!!!!).

    Thoughts:
    1. The influence of CO2-only won’t show as linear.
    2. The rate of temp increase of 0.66C/century is ONLY if the barycentric influence continues for the next 90 years as it did for the past 130+.
    3. If UHIE is even half as significant as some suggestions, then there is precious little left for CO2 OR
    4. A UHIE corrected graph may not match the barycentric model data.

    Conclusions:
    1. The radiative forcing of CO2 is way too small for concern, fixing by man and for the IPCC models (so far, so good).
    2. The increased injection of CO2 has not been taken into account.
    3. Who says the barycentric activity will continue as before?

    Steve McI is right. A lukewarmer to annoy both sides. They are playing music at a Garden Party (music reference for the ’70s crowd).

  3. Malaga View says:

    They are playing music at a Garden Party (music reference for the ’70s crowd).

    Thank you…
    I have wonderful memories of a garden party at Crystal Palace with Lou Reed…
    And I also remember that CO2 had nothing to do with the temperatures in the ’70s…
    And I am just as sure that CO2 did not suddenly start popping speed 1950…
    Just goes to show what you have to do to get published these days…
    /start of rant
    Any paper that relies upon the concept of an Average Global Temperature is not a scientific paper by definition…
    It is a travesty of science…
    Any paper that relies upon Hadley, GISS and the other Manufactured Anthropogenic Data sources is not a scientific paper by definition…
    It is a corruption of science…
    /end of rant

  4. adolfogiurfa says:

    @Malaga View
    and also those cases where there is a sincere and naive “groping in the dark” or “being surprised for”…
    As Galileo Galilei said: “E pur si muove”

  5. tallbloke says:

    Doug and MV: I think Loehle and Scafetta have done well to get the paper published. Yes, they have made a nod to co2, but hidden in the paper there are references to longer cycles which they acknowledge not including…

  6. Roger Andrews says:

    “Our result matches the historical record better than any other attribution study”. Actually the Scafetta and Loehle paper shows essentially the same results as I presented in my earlier solar cycle thread (http://tallbloke.wordpress.com/2011/05/23/global-warming-projections-using-solar-cycles/).

    However, there are three basic flaws in S&L’s methodology. The first is that they compare their model output against HadCRUT3, and HadCRUT3 is an apples-and-oranges average of different SAT and SST trends (mean global SAT has increased by about 0.3C since 1997 while mean global SST has increased by only about 0.1C). By using HadCRUT3 to define “surface temperatures” S&L implicitly assume that SAT and SST temperature trends are the same, which they aren’t, and that natural and anthropogenic forcings have the same impact on SAT and SST, which they almost certainly don’t.

    The second flaw, as Doug Proctor noted above, is that S&L allow for only two processes – a +/- 60 year barycentric period and greenhouse gas warming. If we include longer solar periods we can in fact explain all of the SST warming since 1900 in terms of solar forcing (see the earlier thread), although we do need an anthropogenic component to match SAT after about 1970.

    The third flaw – and one which is common to all analyses of this type, including mine – is that S&L’s barycentric temperature wiggles (which, incidentally, correlate quite closely with the AMO) are confined largely to high northern latitudes. They show up as distinct features on the global record only because their amplitudes are so large, particularly in the Arctic. But exactly how barycentric motions generate large temperature fluctuations in high northern latitudes and nowhere else remains unexplained, so far as I know.

  7. tallbloke says:

    Hi Roger,
    Good comment. On the third flaw, the southern hemisphere is mostly ocean, and the ocean is good at absorbing and shifting solar energy in order to maintain equilibrium. You can hardly blame L&S for plate tectonics!

    In fact, thinking about it, the fact of the stability of the temperature of the southern mid latitudes is a good argument for the existence of negative feedbacks in the climate system.

  8. Roger Andrews says:

    Hi TB:

    Your comment prompted me to go back and take yet another look at the North Hem and South Hem SST and SAT records.

    The SST records show the about same amount of warming since 1880 in both hemispheres (+/- 0.7C) and both show the rapid increase in temperature between 1900 and 1950 which I think is a solar effect, although it’s more marked in the NH. Overall the two records are not that different.

    The SAT records, however, are that different. The NH record shows the “barycentric bump” peaking around 1940 and about 1.2C of warming since 1880 (in fact it’s the only one of the four records that really matches the Loehle and Scafetta model). But the SH record shows no sign of the bump and only about 0.6C of warming since 1880, with effectively all of it occurring after 1965.

    One can argue that all we’re seeing here is the land warming faster than the sea, but I don’t think this is the case. Buried somewhere in my many gigabytes of disorganized spreadsheets is a graph of SAT change between 1970 and 2000 at 500 surface stations plotted against latitude. It shows a straight-line relationship between the two, with about 1.2C of warming in the Arctic, about 0.6C at the Equator and 0.0C in the Antarctic. And I get the same trend when I segregate the plot into “land” and “ocean” stations. The controlling influence is obviously latitude, not differential warming over the land and the sea.

    And one thing that increases linearly with latitude between the Arctic and the Antarctic is the area of open ocean.

    So the more ocean we have, the less SAT warming we get. Any ideas?

  9. Ray Tomes says:

    If the 208 year cycle is not included in such an analysis, then it must be flawed. This cycle rose for the whole of the 20th century and will fall for all of the 21st century. How can predictions be made without incorporating it? The 2300 year cycle is also important but for shorter periods can be ignored as it is still just trending on up.

    Several years ago I did a cycles analysis of Loehle’s two millenium non-tree-ring temperature construction. I got a cycle of 204.4 years. Then I had a computer worm and lost all my work and had to restart. With the newly obtained series the cycle became about 234 years. This is a bit of a mystery to me. I think there may be two versions of the data.

  10. Roger Andrews says:

    Hi Ray:

    This is one of the objections Doug Proctor and I raised in previous comments. You can’t just consider one solar cycle. You have to consider all of them.

    And the 208 (or 204, or 234-year cycle) may indeed be the most important. As you point out, it was on the upswing during the 20th century, and according to my phenomenological model, which used Tim Channon’s 238.9 year cycle, it accounts for about 75% of the SST warming since 1880 (link above).

  11. Paul Vaughan says:

    Did anyone notice Svalgaard trying to claim (over at WUWT) that irradiance & insolation are basically equal? And that spatial patterns of clouds have no effect on global averages? I no longer trust him.

  12. tallbloke says:

    Ray and Roger, L&S include a secular increase from 1850 in their model which they say is roughly equivalent to a long period sinusoid. So they do recognise that this could be due to a longer cycle. But they then project the additional anthropogenic component on top of that secular increase right out to 2100. They say this is so that they obtain an ‘upper bound’ result for potential warming.

    I think we need to take it all with a grain of salt. Nicola Scafetta knows as well as we do that there are longer term cycles. Indeed, they mention a MWP-LIA oscillation in the comments at Curry’s site. I think they made a ‘policy decision’ to keep it simple and include an anthropogenic term equal to the non-feedback value for co2 at a ‘reasonable’ sensitivity value in order to get the paper published and discussed. It’s a beachead in the modern literature for the barycentric planetary-solar-climate theory.

    I anticipate there will be followup papers which introduce longer cycles not included in this one…

    Paul: sounds like the typical Svalgaard agenda in operation. Flatten the TSI curve, then deny the terrestrial amplification and the changes in albedo. Looking a bit desperate IMO. copy the URL of the comment if you find it again, I’ll go have it out with him. :)

    Edit: Found it. here’s Leif’s comment:
    “Paul Vaughan says:
    July 26, 2011 at 7:30 pm
    INSOLATION (NOT irradiance).
    The irradiance is what the sun puts out, insolation is what the earth’s surface receives and over time scales under a thousand years, there is a good relationship between the two. Modulated a bit by the climate [climate, albedo, etc], but in essence just a constant fraction. But your comment is, as usual, too cryptic to be of any use.”

    I see you already made the appropriate response: “bit” indeed! Lol :)

  13. tallbloke says:

    Roger: re SAT increase: If the SAT dataset includes land there may be a UHI issue in the mid northern latitudes. The poles do seem to oscillate oppositely temperature-wise. My guess is that this may be for a reason linked to the asymmetry of sunspot production in the solar hemispheres. Both being caused by the heliocentric latitude of the barycentre as the big outer planets wimble up and down in the z axis. I think this affect the geomagnetic situation and links in to Vucevic’ work.

    http://tallbloke.wordpress.com/2011/02/12/sunspot-asymmetry-barycentre-to-solar-centre-link/

    http://www.vukcevic.talktalk.net/MF.htm

  14. vukcevic says:

    I was critical of the L & S paper on the two major climate blogs (WUWT & Climate etc).
    I’ve just posted this (a bit more moderate view) on the WUWT:

    All these cycles, as far as the climate is concerned, are irrelevant unless there is an extra energy input at a particular point along the cycles’ progress. All indications are that such input is insufficient to account for the extra thermal energy as reflected in the measured temperatures on decadal or century scale.
    So, why and how is the Earth warming or cooling?
    Earth receives certain amount of energy, up to ~ 40th parallel is greater than what is radiated back, but for the higher latitudes the ratio is reversed; it is obvious that the higher latitudes are supplied with the excess via oceanic currents. Currents are by no means constant, measured either by volume or velocity. More current volume (heat content) or higher velocity (less heat loss during the transport process) across middle latitudes will result in global warming. Reduction in these two parameters results in the excess of the equatorial heat being re-radiated back into space, instead of it being taken towards the poles; the result is global cooling.
    No solar variability need be present into the equation to explain simple process of global warming and cooling, just discover where and why the currents are affected, and that is precisely what I’ve researched. To my mild astonishment, or if you whish a great surprise, the process is very simple and results are more than satisfactory (considering simplicity of the approach) as seen here:

    http://www.vukcevic.talktalk.net/AP.htm

  15. tallbloke says:

    Hi Vuk,
    You got some good interest from Craig and Nicola on Climate etc. They are all telling you to finish your full write up which explains the ‘drivers’ on your mysterious graphs.

    I just posted on WUWT too:

    The same solar system oscillations which involve the timing of planetary orbits and solar activity also (IMO) lie behind the multidecadal changes in Earth’s length of day, and these are in phase with the oceanic oscillations seen in the SST records, PDO, AMO etc. Maybe the cause is the slowing down or speeding up of the Earth causes water to pile up against the sides of continents, producing cold upwelling. probably it’s more complex than that.

    These multidecadal trends in LOD are maybe amplifying the terrestrial response at that 60 year frequency.

    The first post on my blog posed the question:

    http://tallbloke.wordpress.com/2009/11/29/planetary-solar-climate-connection-found/

    Nicola Scafetta is a smart guy he knows this, we have conversed via email, but trying to introduce too many new ideas resting on as yet poorly quantified mechanisms in a single ground breaking paper is probably not a good plan if you want to get it published. I think you should see this as a ‘first foray’ to be followed up by many more studies. Hopefully these follow up studies will get published because they are referring to something already in the literature. This kind of bootstrapping has been long employed in many scientific fields. I see this paper as the Omaha beach head which will help the important study of cyclic phenomena to regain it’s rightful place in the body of peer reviewed science.

  16. Ray Tomes says:

    Rog, 2300 year cycle can be treated as “trend” for a couple of centuries. 208 year cycle is entirely different as we all know. Also, the ~20 year cycle that I find in 350 years of England temperatures is actually over 24 years. It is too long to be J-S cycle or sunspot cycle.

  17. Paul Vaughan says:

    Even from those overextending collegiality, the response is at best tepid. Tallbloke’s right: Sequence matters, particularly avoidance of cart before horse. It’s pure comedy watching oblivious overreactions at WUWT & Climate Etc. The intentions of the authors are not what many either assume or glean from the surface.

  18. Paul Vaughan says:

    @Roger Andrews (July 27, 2011 at 12:46 am)

    The fulcrum for this factor is not at the equator:

    http://wattsupwiththat.com/2011/07/25/loehle-and-scafetta-calculate-0-66%C2%B0ccentury-for-agw/#comment-706158

  19. tallbloke says:

    Paul, Scafetta’s response to my post there quoted above is illuminating in this regard

    http://wattsupwiththat.com/2011/07/25/loehle-and-scafetta-calculate-0-66%c2%b0ccentury-for-agw/#comment-706772

    Ray: Agreed about the 208 year cycle. In their defence, we haven’t determined a celestial origin for it, and it’s too long to be isolated in temperature series from 1850. Also, I think their purpose in including the term for co2 was as much political as scientific. Wait and see what happens in subsequent papers now they got this one past the reviewers. ;)

    Regarding CET and the 24 year pulse; I don’t think looking at one geographic location will give us the answer to this.

  20. Roger Andrews says:

    TB:

    “If the SAT dataset includes land there may be a UHI issue in the mid northern latitudes”.

    I know a lot of people won’t agree with me, but after having picked through the surface station records in very considerable detail I’ve satisfied myself that there is no significant UHI component in the SAT record in the mid-northern latitudes or anywhere else (specifics too complicated to go into here).

    “The poles do seem to oscillate oppositely temperature-wise.”

    The only support for this comes from a handful of surface station records in the Antarctic Peninsula that show net warming between 1940 and 1970, when the Arctic was cooling down. There’s no evidence for this antiphase relationship in the rest of the Southern Hemisphere records over this period, or at any other time for that matter. The Antarctic mainland record goes back only about 55 years, but so far it has shown no sign of oscillating in antiphase with the Arctic either (or the Antarctic would have been cooling rapidly over the last few decades).

  21. tallbloke says:

    Roger: what about maritime air temperatures? Is there a good dataset?

    The opposite oscillation of the poles (not hemispheres) could be happening even if the Antarctic were warming slightly and the arctic warming quickly, if we regard the oscillation as travelling on the back of a larger oscillation.

  22. Roger Andrews says:

    TB:

    Do we have a good maritime air temperature data set? There are five published versions to choose from, all based on air temperature measurements taken on board ships.

    The first is the Rayner 2006 NMAT (MOHMAT) data set, which is the one that usually gets used. It’s at http://www.metoffice.gov.uk/hadobs/indicators/data/MAT_mohmat.txt.

    The second and third are ICOADS Tair (which includes the DayMATs as well as the NMATs) and MOHMAT4.3, another NMAT series. Both are available at the KNMI site (http://climexp.knmi.nl/selectfield_obs.cgi?someone@somewhere)

    The fourth and fifth are Ishii Interpolated and Uninterpolated NMATs. These are (somewhere) at http://www.ncdc.noaa.gov/bams-state-of-the-climate/2009-time-series/mat?del%5B%5D=ishii_1

    Are any of them any good? Well, they all show similar wiggles, but Rayner shows 0.72C of warming since 1910, Tair shows 0.64C, Ishii uninterp. shows 0.45C, Ishii interp. shows 0.40C and MOHMAT4.3 shows 0.32C. So the answer is, probably none of them.

    (Yet the HadCRUT3 “surface temperature” record is largely derived from the Rayner NMATs. HadCRUT3 is about 70% based on HadSST2, and HadSST2 is adjusted, under the guise of “bias corrections”, to match the Rayner NMAT as closely as possible. And the Rayner NMATs are bias corrected too, and in some cases they’ve even been adjusted to match the SSTs. So if you want a really dodgy surface temperature record, HadCRUT3 is the one for you. But I digress.)

    The marine SAT series I use is the one I constructed myself using records from stations on islands in the ocean. It’s more robust than any of the MAT data sets, and it also tracks the SAT record I constructed using records from stations in landlocked areas (i.e. no difference between SAT trends over land and SAT trends over the oceans). So I think the GISS “met station only” record is about as good an indicator of marine air temperatures are you are going to get.

    Got to run, More on the South Pole later.

  23. vukcevic says:

    I noticed a short comment from Scafetta on the WUWT :
    About Vuk’s idea concerning the Jupiter Saturn conjunctions towards the forward moving part of the heliopause, as I said it is an interesting idea that may well fit another idea that I add to explain the phenomenon. But I cannot talk about it now. Hopefully, we will have another occasion to discuss it extensively.
    If any of you may wander what is all this above you can find comprehensive review, which I recorded for a reference, and is available here:

    http://www.vukcevic.talktalk.net/L&S+vukcevic.htm

  24. Paul Vaughan says:

    Vukcevic, I see the following at your link (presumably from Climate Etc.):

    tonyb | July 26, 2011 at 5:45 pm | “He has good ideas but they tend to be random one off thoughts rather than a closely written argument. Let’s hope he takes the hint and produces something before too long.”

    Being tied up at committee producing communications squanders precious limited time that can be used advancing exploration. It can literally take orders of magnitude more time to summarize findings than to find them.

    There are many roles to be played in these discussions. We are all volunteers. Each of us is free to decide independently what role we play. There will always be a few bright members of the audience who will understand and/or be sparked by a concise observation.

  25. Paul Vaughan says:

    @Roger Andrews (July 27, 2011 at 10:39 pm)

    You appear to be overlooking the significance of the fulcrum being not over the equator but rather the Antarctic Circumpolar Current of the Southern Ocean (a SHARP discontinuity caused by the current distribution of continents).

    See figure 10 here:

    Carvalho, L.M.V.; Tsonis, A.A.; Jones, C.; Rocha, H.R.; & Polito, P.S. (2007). Anti-persistence in the global temperature anomaly field. Nonlinear Processes in Geophysics 14, 723-733.

    http://www.icess.ucsb.edu/gem/papers/npg-14-723-2007.pdf

    Compare with the following to develop awareness of low-heat-capacity high-variance-leverage (that misleads those lacking adequate conceptualization of the role of aggregation criteria in spatiotemporal summaries into perceiving AMO &/or AMOC as “drivers”):

    Some of the dominant contrasts driving flow are asymmetric. I encourage you to spend some time becoming acquainted with EOP (Earth Orientation Parameters).

  26. Roger Andrews says:

    Hi TB:

    Back on the South Pole question.

    I was going to write a quick comment restating my earlier opinion that your polar antiphase relationship was pure speculation when I remembered my maxim – “engage data before waving arms.” So I engaged the data, and having done so I’m now going to have to eat my earlier opinion.

    What I did was correlate detrended and smoothed versions of the GISS latitude-band SAT series (http://data.giss.nasa.gov/gistemp/tabledata/ZonAnn.Ts.txt) against the Loehle and Scafetta 60-year cycle, which I analogued with a sine wave. Here’s what I got over the 1940-2010 period:

    64-90N, R= 0.70, peak-trough amplitude = 2.5C
    44-64N, R= 0.89, peak-trough amplitude = 0.7C
    24-44N, R= 0.87, peak-trough amplitude = 0.5C
    0-24N, R= 0.87, peak-trough amplitude = 0.4C
    0-24S, R= 0.93, peak-trough amplitude = 0.3C
    24-44S, R= 0.71, peak-trough amplitude = 0.2C
    44-64S, R= 0.36, peak-trough amplitude = 0.1C
    64-90S, R= minus 0.79, peak-trough amplitude = minus 0.8C

    Peak-trough amplitudes decrease rapidly as we go south, but the correlation is positive everywhere except in the Antarctic, where it’s strongly negative. So there’s your polar antiphase relationship. :-)

    One problem, however is that temperatures are effectively uncorrelated with the 60-year cycle before 1940, which is why I used only the post-1940 data. Can we have a planetary cycle that affects temperatures only some of the time? I guess I’m going to engage the data again before I do any arm-waving on that one.

  27. tallbloke says:

    Roger: classy my friend. That fits well with Erl Happ’s work on the arctic and antarctic oscillations. Now, the pre 1940 period has a big solar dip in it around 1900, plus a fairly wild oceanic oscillation. When I integrated the sunspot numbers to get an approx ocean heat content guide from the late C19th I found that the minimum in OHC would have been in the mid 1930s. Paul Vaughan can tell us something about the interesting 1934 ‘watershed’ in the data he has been looking at too.

    But we had that strong positive excursion of the PDO from 1910 or so, which indicates to me that lots of heat made its way up from further down in the ocean, warming the atmosphere and depleting OHC. That idea fits with my hypothesis that heat comes out of the ocean when solar activity is low, and continues coming out for a while after solar activity starts to recover, boosting surface temperature.

    If you look at the graph I made two years ago at the bottom of this comment, you can see that the upsurge in temp to the 1940 peak would nicely fit the 1900 hole if you turned it upside down and dropped it in. I think this is why the 60 year cycle ‘doesn’t work’ pre 1940. Notice though, that Scaffeta’s 20 year modulation of the 60 year cycle copes with the 1900 dip admirably well. As does my unphysical but interesting z-axis study:

  28. vukcevic says:

    I think your problem is in mainly in the PDO, pre 1940 it has 40 year period, post 1940 just under 60. Also the SOI (~ negative ENSO) pre 1940 is indeterminate, post 1940 is ~60 y.
    Only the AMO ‘behaves’ along the most of the time scale.
    The PDO and ENSO are the big beasts, one reason why I don’t go for the Scafetta’s theory.

    http://www.vukcevic.talktalk.net/AP.htm

  29. Roger Andrews says:

    HI TB & Vuk:

    A lot of stuff to think about here, so I’ll just hit on a few points before I go out for my steak dinner.

    TB:

    1. I’m going to have to go back and see how my results dovetail with Erl Happ’s work on the Arctic and Antarctic oscillations, because at first glance I can’t see any relation between them (R^2 = 0.04).

    2. The positive excursion in the PDO that began in 1920 was accompanied by a positive excursion in the AMO that also began in 1920. So a lot of warm water rising to the surface all at once.

    3. Not sure how turning the 1940 peak upside down and dropping it into the 1900 hole fits with the SAT and SST data. Have to think about this.

    Vuk:

    1. “Only the AMO ‘behaves’ along most of the time scale.” The AMO fits the North Hem SAT record quite closely, but with a +/-five-year lag relative to SAT. This makes it hard to attribute SAT changes to changes in the AMO.

    2. The PDO index shows troughs in 1922 and 1956 (a 34 year period) and pronounced peaks in 1941 and 1987 (a 45 year period). I don’t see a 60-year period.

    FWIW:

    The AMO (lagged 5 years) gives the best fit to detrended SAT in the North Hem and the PDO gives the best fit in the South, but neither of the fits are all that good. (I don’t compare the AMO and PDO with the SST record because the AMO and PDO are derived from the SST record, which means that I’m bound to get a match of some kind)

    Maybe more later.

  30. Paul Vaughan says:

    My tolerance for neverending PDO misinterpretations expired awhile back.

    A strong indicator that a climate discussion participant is misinterpreting PDO is mention of PDO as if it’s analogous to AMO.

    Proper interpretation of PDO requires conceptual foundations that MOST climate discussion participants (including academics) don’t have.

    Comments in this thread about PDO indicate substantial misconceptions about early 20th century terrestrial climate variations.

    My advice:
    If you don’t conceptually understand the calculations that generate PDO and the things that can go wrong with such calculations (including how to detect such potentially seriously misleading problems), then seriously consider not positioning yourself to look incredibly foolish by basing arguments on what is, from your perspective, an index spit out of a black box, the internal workings of which you have NO understanding.

    Regards.

    [Edit] My tolerance for your brand of condescencion expired a while back too.

    My advice: keep it civil and good natured or sod off.

  31. vukcevic says:

    Paul
    Anyone who understands what the PDO means for each season’s climate probably has a brain chip implant, not to mentioned 2 sets of indices. Risking being even more foolish, I opted for the minimum of knowledge contained in this illustration which I consult regularly (couldn’t guaranty that it is correct). http://www.vukcevic.talktalk.net/PDO.gif

  32. Stephen Wilde says:

    Bob Tisdale drummed it into my head that PDO is a statistical derivative of ENSO so if one is to talk about longer term events than ENSO then a different term should be used.

    He suggested Pacific Multidecadal Oscillation, if my memory is correct so I have been using that term since. I recommend it.

  33. Paul Vaughan says:

    Roger Andrews, you appear to be a quick study. Before you go mucking around with cross-correlations (& their seriously misleading lags), please have a look at the references I included here [ http://wattsupwiththat.com/2011/05/15/interannual-terrestrial-oscillations/ ] whenever you can spare some time.

    The Tsonis, Swanson, & Kravtsov (2007) article is the one that has gained the most popularity in online climate discussions so far, but Schwing, Jiang, & Mendelssohn (2003) is even more informative. Maraun & Kurths’ (2005) concise section 3 primer is consistent with astute remarks (reflecting lucid awareness) made by Climate Etc.’s “Cheif Hydrologist”, which I enthusiastically quoted here [ http://judithcurry.com/2011/07/14/time-varying-trend-in-global-mean-surface-temperature/#comment-87429 ]. More generally, see everything Tomas Milanovic has written at Climate Etc. (about spatiotemporal chaos) as an expeditious means of gaining a handle on the nature of the interannual variability conceptualization problem in online climate discussions.

    Reiteration of my (perhaps insufficiently stern) cautionary note to fellow data explorers:

    “[...] it needs to be pointed out assertively & clearly that the cross-correlation approach, while informative, is patently insufficient for determining the full nature of terrestrial spatiotemporal phase relations.”

    To be blunt:
    Don’t even try to interpret lags until you gain a handle on what the Cheif, Milanovic, Tsonis, Swanson, Kravtsov, Schwing, Jiang, Mendelssohn, & others are saying.

    For those of you wondering what to pair with ENSO to replace “PDO”, see Table 3 & Figure 3 of Trenberth, Stepaniak, & Smith (2005) for an idea of where Bob is driving (via a different route, so I’m not sure if he has considered the equivalence) with his references to North Pacific pressure variations.

    It will have to be another day that I address so-called “60 year cycles”. (People are missing something simple & important.)

    Best Regards.

    [Edit] To be blunt: It’ll have to be another day soon because I won’t allow vague assertions that “People are missing something simple & important” to stand unless they are backed up with clear explanation.

  34. Roger Andrews says:

    TB:

    OK. Here’s something simple and important, backed up by a brief explanation.

    As I noted earlier, when we compare the L&S 60-year cycle against detrended zonal surface air temperatures we get positive correlations everywhere except in the Antarctic. However, there’s no lag in the temperature response.

    When we compare the cycle against detrended zonal sea surface temperatures we also get positive correlations (up to R=0.93), but there’s a fairly consistent +/- 10 year lag in the temperature response.

    A forcing that warms and cools the air immediately but takes its time warming and cooling the ocean has to be external. An internal forcing generated by, say, ocean cycles would work in the opposite sense, with the oceans warming and cooling first and the air warming and cooling in response a few months later (we discussed this recently on the Ray Tomes thread.)