P.A. Semi: Integration of Greenwich and USAF Sunspot area data with farside interpolation

Posted: February 2, 2014 by tallbloke in Analysis, Astronomy, Astrophysics, Celestial Mechanics, Dataset, Solar physics, solar system dynamics, Tides

Brilliant Czech researcher P.A. Semi has sent us the fruits of some considerable labour, which he has asked us to share with the Solar-planetary community. Since the venue at Pattern Recognition in Physics was axed by Copernicus (The Innovative Science un-PublisherTM), he says he is not sure where to get this published, so the Talkshop it is for now. I will also add it to the repository I am building at ‘Solar System Science’, a new venture I’m setting up in collaboration with other researchers. Tim Channon will be interested in working with the dataset which can be generated from the resources Semi has provided, and I’m sure others will be too. Here’s a sample of the output:


Semi Writes:

Hello Tallbloke and others.

I’ve produced the Synoptic map of Sunspots 1874-2012 and Interpolated Sunspot Area, that allows to investigate sunspot record without smoothing, while removing the 27-day false signal of single-face problem another way – by interpolating individual Sunspot groups: if they can be matched on the next rotation, they are linearly interpolated to the new position and size, if they are not matched, they are interpolated in 17 days linearly to zero. This way, the far-side Sunspots are interpolated and the record does not need the usual monthly smoothing, that wipes away precise timings. (There still exists some single-face problem – the Sunspots, appearing on the far side first, are delayed until they get to the front side, and the 17-day fade-out makes a typical fade-out curves in the chart, but still better than if the group disappeared abruptly on the limb…)

There is also included a chart, marking Earth,Venus and Jupiter opositions, conjunctions and quadratures, showing, how the individual events influence Sunspot growth… (it seems, that the rule is different in cycle rise and in cycle descend, some cycles the opositions and conjunctions have influence, some times the quadratures. Also often the first sunspots of the new cycle appear soon after the “touching” oposition (where the line between the two crosses Solar body) of Earth and Jupiter, but not always…).

On the synoptic map, if it is used in conjunction with my EphView program, which can paint planet positions over the images, it is clearly visible, that the Sunspots usually form below the place, that was mostly influenced by the oposition etc, soon within the next rotation…

The source of the data is Greenwich Group Sunspot numbers for period 1874-1982 and USAF records for period 1982-2012. The USAF record (in which there are more stations reporting simultaneously) was manually cleared from cca 150 errors in 220k records, the method is described in the readme file…

The map is published such, that the charts are included directly, but the Sunspot images are distributed in the form of source data-file and program to generate 3.5Gb of PNG images, one for each day (run-time cca 1-2 hours), with a download size of some 25Mb… For those, who cannot run win32 application to generate the map, I probably could send the DVD (but for some fair cost of packing and postage), ask me on email, and also do not hesitate to ask, if there is some problem/question with this or with the EphView program on semi at gurroa dot cz.

[UPDATE 2, by Tim, 9th Feb,
After a lot of work by both of us we conclude there are timing mistakes. This is going to take time to correct so I am marking the downloadable files as strikeout although they might be correct. The basic ideas remain sound, about details.]



(Could you please notify others, who may be interested in this record, as I’m not sure where to publish this work…? )

[UPDATE] Tim writes, I have produced an experimental video of the image files. For more information see comments.

https://tallbloke.files.wordpress.com/2014/02/sun-1880-1910-a.zip (2.1MB, zip archive containing AVI file containing MP4)
[/update][/update 2]

  1. […] P.A. Semi: Integration of Greenwich and USAF Sunspot area data … […]

  2. tchannon says:

    Update. I have created an experimental video from the image files.
    Runs 1880 through end 1910 at 15 frames/sec, one frame per day, a compromise.

    AVI file inside a ZIP, encoder MP$, heavy compressed, size is a massive problem. Looks acceptable.

    Before you download note that this is adult content, in general extremely boring.
    This is a lot of work, hammers by poor old computer hard particularly the disk so I don’t want to do this for no real purpose.

    *if* this is useful, lets sort out what needs altering but remember I have to do it, takes my time, wears my equipment.

    Producing a whole dataset video is not too hard other than hours of disk hammering. If 30 years takes 2.1M, the whole thing, with luck will be 10MB, which is tiny given the source.


  3. tchannon says:

    Here is something I wanted to see, what happened mid-1951, thank you PA Semi.

    This is a fading sunspot peak where there is a sudden strange burst of activity. The papers and reports from then are unclear.

    [edit: expletive, WordPress “got” me again, upload didn’t take, wrong link last night]

    Here is the reason. There are only two of these since the concept of solar flares became accepted.

  4. tallbloke says:

    Tim, nice plot. A bit bigger than the last W&P effect you mooted. 😉

  5. tchannon says:

    That is the largest on record, a mix problem, on record but in general too old for there to be easy to find detail.

    Something to note: events during the fading period of a sunspot cycle might be of considerable significant. Cycle 23 is another case but is without an associated barycentre event. OTOH here I am aware of a curious train of events, not going into that now. We ought to be focusing on the work of Semi.

  6. tchannon says:

    Update I just spotted the wrong video file was linked for 1951, now corrected,
    this comment

  7. Scute says:

    I’m surprised no one but the webmasters have commented on this. It seems highly significant that pin-pointing exactly when and where the spots pop out on the surface is key to understanding the generation process below. I find it odd that they get counted, smoothed and filed away.

    The time between syzygy and appearance on the surface gives a precise convection speed. If it’s, say, ten days, the speed would be about 350 metres per second if the disturbance begins at the tachocline. That doesn’t sound excessive to me given the pressures, temperatures and latitudinal tachocline speeds.

    I’ve often thought that little attention seems to be paid to the possible processes that might explain the sporadic appearance of large numbers of spots, months apart with relatively quiescent periods in between. Getting this timing nailed down (assuming the syzygy really does set the process in train) would be a very good start in getting to grips with the mechanistic process.

    All the zip downloads failed for me so I’m admittedly going only on the basic description in the post.

  8. tchannon says:

    Can you handle this?
    https://tallbloke.files.wordpress.com/2014/02/sun-1951-parta.pdf (5M)

    Video as I have done is much more compact than straight images because it operates on change frame to frame, not much changes and so…

    Anyone not able to download, admittedly via HTTP or HTTPS instead of FTP is worrying.

    Uploads posing problems for der management is a completely different matter from downloads.

  9. Scute says:

    Thanks Tim

    Yes, that worked.

  10. oldbrew says:

    General interest solar story here (Dec. 2013).

    ‘Giant Plasma Spirals Found on the Sun’

    ‘These giant cells are expected to each last months, helping transport incredible amounts of heat generated in the sun’s core to its surface.’

    ‘ “They ultimately help drive the sun’s 11-year sunspot cycle, and should also help active regions — highly magnetic areas — form” ‘

  11. tchannon says:

    Scute, doesn’t make sense but life is life.

  12. tchannon says:

    https://tallbloke.files.wordpress.com/2014/02/sun-1990-a.pdf (6M)
    [mod: see article update, Plot below is completely independent work by me]

    This is the only other “major” barycentre “situation” during the timespan of the Greenwich/NASA dataset. Previous is 1851 presuming I have computed b/c data correctly.

    The 1990 case is what first brought this to my attention, wasn’t looking at this at all, just happened to notice an unusual spike in sunspot data, which is not all clear from casual inspection, was with what I was doing.

    This is unfortunately right at the top of a sunspot cycle, a split hump one, cycle 22.

    More remarkable is that there is the coincidence of _three_ events, won’t name number three yet. There is no third dataset during 1951.

  13. Brian H says:

    The first graph: pale yellow on white b/g? Really? A less visible combo could hardly be arranged. Deep purple on blue-black, perhaps.

  14. […] winter I’ve sent you one my works, and said, there is another work pending, which I’m attaching now… It is related to the […]