Nicola Scafetta and Richard C Willson: Planetary harmonics in the historical Hungarian aurora record (1523–1960)

Posted: February 14, 2013 by tallbloke in Analysis, Astrophysics, cosmic rays, Dataset, Geomagnetism, Natural Variation, Solar physics, solar system dynamics
Tags: , , ,

Congratulations! to Nicola Scafetta and Richard C Willson on the publication of their new paper: Planetary harmonics in the historical Hungarian aurora record (1523–1960). This is another excellent paper, published in Planetary and Space Science. Grabbitquick before I take it offline. Scafetta always makes papers available later if you miss this one. The Hungarian record goes back to a very early date and this makes the paper especially interesting to those of us eager to see more validation of the solar planetary theory, which is rapidly becoming the best show in town for matching paleo records. Geoff Sharp will be particularly pleased to see the strength of these Uranus-Neptune synodic correlations with solar activity levels.



The historical Hungarian auroral record extends from 1523 to 1960 and is longer than the sunspot record. Harmonic analysis reveals four major multidecadal secular cycles forming an approximate harmonic set at periods of 42.85, 57.13, 85.7 and 171.4 years. These four frequencies are very close to the four major heliospheric oscillations relative to the center of mass of the solar system caused by Jupiter, Saturn, Uranus and Neptune. Similar frequencies are found in solar radiation models based on long cosmogenic isotope records (Steinhilber et al., 2012) and in long records of naked-eye sunspot observations (Vaquero et al., 2002). Harmonic regression models are used to reconstruct and forecast aurora and solar activity for the period 1956–2050. The model predicts: (1) the multidecadal solar minimum in the 1970s that is also observed in the sunspot record; (2) a solar maximum in 2000–2002 that is observed in the ACRIM total solar irradiance satellite composite; (3) a prolonged solar minimum centered in the 2030s. These findings support a hypothesis that the Sun, the heliosphere and the terrestrial magnetosphere are partially modulated by planetary gravitational and magnetic forces synchronized to planetary oscillations, as also found in other recent publications (424345461 and 54).



  1. vukcevic says:

    Only skimmed through, it looks an interesting paper.
    However I would point that strength of aurora, and ability of it to be observed and recorded depends on the strength of the Earth’s magnetic field.
    NASA has found 85 year component in the Earth’s magnetic field:
    “Four of these oscillations were robust, occurring at periods of 85, 50, 35 and 28 years. Since the scientist’s data set goes back to 1840, the recurrence period of the longest oscillation (85 years) is less well determined than the other oscillations.”
    Limited amount of data is unlikely to resolve if it is 85 or 85.7 ( 2 x 85.7 = 171.4) or even 84 years (84 yr.= Uranus sidereal period).
    If NASA is correct then Uranus/Neptune are also causing change in the Earth’s field, which I personally have strong reservation about such possibility, since I have spent lot of time studying and analysing the Earth’s magnetic field.
    Alternative possibility that aurroral frequency is capable to penetrate to the outer core, the source of the Earth’s field, is equally unlikely.
    Nevertheless there is always Gleissberg cycle to fall back on to, in ‘The probable behaviour of sunspot Cycle 21’ Gleissberg states :” After an explanation of the method of forecasting based upon the 80-yr sunspot cycle, reasons are given for assuming that the maximum of the present 80-yr cycle now has passed. ”

  2. Geoff Sharp says:

    While it is good to see Uranus & Neptune synodic periods appearing in the literature there are 2 questions and a statement that I will make.

    1. Is there enough data in the Hungarian record to perform any sort of analysis?

    2. When is the scientific community going to have the courage to reference non peer reviewed papers that are perhaps more relevant than other peer reviewed references.

    While FFT analysis is useful it only skims the top layer. There will almost always be a spike in the 170-210 year region because every 172 years (avg) around 3 solar disturbances (around 40 years apart) occur. The disturbances because of astronomical precession are of different strength each time varying from nothing to extreme. When science drills down further this will all be revealed (perhaps this year).

    A brief article which may provide some understanding is available from:

    Some basics via Youtube:

  3. Ninderthana says:


    I noticed that Scafetta & Wilson 2012 used your idea about the Z-axis Barycentric motion of the Sun being a good indicator without actually acknowledging that you came up with the idea.

  4. Roger, thank you

    Just a short reply:

    1. Is there enough data in the Hungarian record to perform any sort of analysis?

    Evidently yes. There are 430 year of aurora data. In any case, in the paper I also study Steinhilber’s record in details that is 9000 year long.

    However, it is important to understand one point: the difference between actual solar related observations such as sunspot and Aurora, and solar proxy models such as Steinhilber’s record. Somebody may think that because proxy models are longer than actual observational records, they are “better.” This reasoning is twisted physics. The very reason why people believe that records such as
    Steinhilber’s record has something to do with solar activity is because of its approximate correlation with sunspot records, which are real solar data.

    So, what needs to be done is to analyze both the data and the proxy models, and this is what it is done in the paper.

    2. When is the scientific community going to have the courage to reference non peer reviewed papers that are perhaps more relevant than other peer reviewed references.

    This is an important issue. And I may understand some frustration of people that had good ideas around. And I have often encouraged people with good ideas to try to write short papers expressing those ideas perhaps with the help of somebody with some expertise.

    However, there is a technical difficulty in referencing non-peer reviewed literature in a paper that needs to be peer reviewed. Essentially, papers use references as extensions of the paper itself. Thus, because the paper is supposed to be peer reviewed also the references need to be peer reviewed if they are used to claim a specific physical property of a system. So, in the peer reviewed literature, non peer reviewed literature is allowed only in very specific cases that do NOT claim physical properties. If not, a referee or the editor is entitled to reject the paper that uses non-peer reviewed literature to claim specific physical properties. Essentially, scientific claims in the non-peer reviewed literature need to be re-demonstrated in a paper that can be accepted as peer reviewed literature.

    So, there is the need to understand that non-peer reviewed literature is simply a different kind of literature and as such needs to be treated.

    To Ninderthana says: I noticed that Scafetta & Wilson 2012 used your idea about the Z-axis Barycentric motion of the Sun being a good indicator without actually acknowledging that you came up with the idea.

    As explained in the paper, the z-axis was chosen not for a specific reason. The purpose of the paper is to get the planetary frequencies in a simple compact way. Any function of the planets may be used for the purpose. This is explained clearly in the paper and a reference to Bucha et al. 1985 paper is used, Also the referenced Jakubcova´ 1986 paper is important on this topic. The frequencies can also be calculated without any function, just from the periods of the orbits. But it is better to use a function.

    If Roger had used a complex function that I re-used as a necessary function of my paper then I needed to reference Roger. If I had used the “z” axis in the same exact way in which Roger used it by reproducing his same pictures in my paper, then I needed to reference Roger work.

    However, I do not think that Roger or somebody else can claim a copyright on the generic “x”, “y” or “z” axis coordinate of a system as if everybody in the world that uses the “z” axis coordinates of the sun from now on must reference Roger.

    So, I invite Ninderthana to be reasonable.

  5. tallbloke says:


    It was Ray Tomes who originally started studying the Z-Axis back in the 90’s. Taking his work a step further, I found that Z-axis motion of the solar system Barycentre relative to the solar equatorial plane, when smoothed over a couple of Jupiter orbital periods matches well to the Gross LOD record (with a 30 year lag).

    We all have our strengths. One of mine is spotting, developing and flagging up good ideas. One of Nicola’s is further developing them, performing professional grade stats analysis on them, and, most importantly, getting them into the literature. Science benefits from this kind of teamwork.

    Nicola did reference the talkshop by hyperlink in the non-peer reviewed report he produced for presentation to the EPA, and he has recognised other contributions from the talkshop here in comments before.

    Nicola himself should have got a mention by Abreu et al last year but didn’t, so we’re all in the same boat.

    The planetary-solar hypothesis is an idea whose time has come and we should work together to make it a big success.

    Now that Paul Charboneau has reviewed Abreu et al in Nature and recanted his 2003 opinion that the planetary-solar theory was dead anything can happen, and it will!!

    Did you both get the emails I sent you yesterday with my new draft paper attached? Ian, I will resend you the latest version I sent Nicola.

  6. Roger, thank you for your comments.

    I was thinking whether you may add also figure 3 to your article. It shows the aurora record analysis vs. the planetary record and Steinhilber’s record. In particular the analysis of Steinhilber’s record highlights a lot of planetary frequencies, as explained in the text.

  7. tallbloke says:

    Nicola, Done. Now, put me out of my misery and email me about my paper. 🙂

  8. Ulric Lyons says:

    I would tend to go with the 83yr peak in the power spectrum that Silverman noted. It makes far more sense when considering which bodies have the dominant effect on sunspot numbers.

  9. tchannon says:

    Is the Hungarian aurora data available in numeric form?

    Incidentally, I requested and was kindly supplied with the unpublished Eddy aurora data. This is a very mixed bag where I’ve looked but not felt able to do much with it.

  10. tchannon says:

    copy from the table in the paper. The data are published inside the paper

  11. tchannon says:

    As a graphic image, oh well, might copy by hand sometime.

  12. vukcevic says:

    Hi Tim

    [ list transferred to simple .xls and put on server here, done for shortness — Tim]

  13. Greg Goodman says:

    Congratulation Nicola. Another useful set of physical data to add to the list.

    (note that the implemented 11-year moving average algorithm dampens higher
    frequencies) (Horne and Baliunas, 1986; Press et al., 1997; Ghil et al., 2002).

    That ref is paywalled 😦 You call this an algorithm so it must be more than just a running mean. Is this something like a three stage running mean with periods of 1/1.3771 each time ?

  14. tchannon says:

    Thanks vuk

  15. Greg Goodman says:

    Anyone have more detail on the Horne and Baliunas method? Sounds very relevant to what I’m trying to get out of SST.

  16. Horne and Baliunas have just written a paper introducing the “periodogram” to the astrophysicist.

  17. Paul Vaughan says:

    Nicola has correctly cited Charvátová (formely known as Jakubcová & also Charvátová-Jakubcová). (No offense to Ray, but Ivanka was there a decade earlier — and I don’t doubt many others had already been there, probably countless many without ever publishing.)

    Remember Ivanka’s ~1930 trefoil? Here’s some follow-up on that:

    Solar Recurrence Plots

    Lots more to say when time/resources permit.

  18. Greg Goodman says:

    Just thought I’d put this preliminary result out , it looks interesting.

    Tim has kindly sent be his spectral analysis software , which I cannot claim to master, so emphasis on preliminary here.

    For some time I’ve been looking at ICOADS SST (since Hadley processing does some funny things to the frequency spectrum). In particular I am looking basin by basin for periodic structure.

    I am finding more and more indications that both solar and lunar effects are present and to a large degree are obscuring each other if care is not taken to see both.

    I think this _provisional_ plot show will look familiar to anyone who had kept up with Nicola Scafetta’s published work.

    The split of the “solar” signal in to two close peaks is very similar to figure 2 in his JS tides paper. Indeed it may be resolving the peaks more accurately than his sigma=0.5 line. I does seem totally consistent with his numbers.

    The 9.15 peak would clearly coincide with his 9.1+/-0.1 peak in variations of the motion of the Earth’s orbit that he demonstrated was due to the presence of the Moon.

    I had been somewhat distracted to interpreting the two close peaks seen here as one peak centred on 8.85. I was thus suggesting this seemed linked to lunar apse, though the two are surely related.

    The bottom line is the three dominant and almost equal peaks in the analysis of N. Atlantic SST show both lunar and solar influence to be equally important. Ignoring one will produce a phase crisis when the two go out of phase and will confound efforts to show long term correlation.

    Again, I have not mastered the software, so I do not know how much confidence to place on this result, but the tie-in with N. Scafetta’s results is intriguing.

  19. Greg Goodman says:

    Multi-scale harmonic model for solar and climate cyclical variation
    throughout the Holocene based on Jupiter–Saturn tidal frequencies plus the
    11-year solar dynamo cycle
    Nicola Scafetta

    The grouping of individual cycle lengths show in this plot match closely with the the peaks in my N. Atlantic SST periodogram. This would seem to firm up the idea sunspot cycle length being the key climate factor.

  20. tallbloke says:

    Excellent work Greg!

  21. Greg Goodman says:

    Just rescanning Scafetta’s JS tides paper noted above the peak of SSN eh shows in the following figures 3a, 3b show that same circa 10, 11 year and much smaller 12y periods.

    circa 9 year is clearly lunar.

    TB, would you like to integrate my plots here? Tinypic is convenient but getting messy.

    [Reply] Done. You might be interested in my new post in relation to the J&S Tidal periods:

  22. Greg Goodman says:

    Here is a extra tropical N. Atl as an example of why I prefer to work with ICOADS rather than Hadley regridded datasets for this sort of thing.

    The three main peaks are essentially the same but there are significant differences elsewhere.

    Until Hadley can demonstrate that doing distortions like doing running-means over adjacent cells inside an iterative loop until it “converges” actually improves the signal rather than corrupts it , I’ll stick with ICOADS version.

    [Yes, I saw the post, not had time to digest it yet, looks interesting]