The Sun Is Lopsided

[SDJ]

No, that would not explain why P > C for the whole day for days where there are only one group on the disk, like these days:

Well, the estimates were made by other people 5,000 miles away several years ago, so I can't really explain the cause of the discrepancy. There may have been "first readings" from the projections that were then corrected after further examination, estimated by other people, etc.

I don't think the differences are that great in relative terms.

[lsvalgaard]

Mar 25, 2010 21:31:34 GMT SDJ said:

No, that would not explain why P > C for the whole day for days where there are only one group on the disk, like these days:

Well, the estimates were made by other people 5,000 miles away several years ago, so I can't really explain the cause of the discrepancy. There may have been "first readings" from the projections that were then corrected after further examination, estimated by other people, etc.

I don't think the differences are that great in relative terms.

I asked the people at Debrechen. Here is their explanation:
"The projected area is measured in millionths of the solar disc but the corrected area is measured in millionths of the solar hemisphere. Thus, if the corrected area is computed from the projected area, then not only the correction for foreshortening is made but the change of the unit too. In this way the corrected area of a spot at the center of the disk is half of its projected area. At about 60 degree LCM the two kinds of area are equal, and closer to the limb the projected area is larger than the corrected one."

[SDJ]

OK, glad they were able to offer an explanation, but it only applies to the relative weights given to the determinations of umbra and penumbra areas depending on the angle of observation for a single spot.

If they've been consistent, the ratio of the P/C estimates of the areas of individual spots in a single AR is irrelevant. The main thing is the ratio of the relative areas assigned to different spots in different AR groups.

That's what I'm trying to look at, the ratios of the sizes of different AR area estimates in the SRS data, and of the umbra and penumbra areas in the Debrecen data, which should correlate a bit and eliminate the "smearing" issue in the latter case.

If AR###1 included a single spot with an area that the Debrecen folks measured at XXX millionths of the sphere/hemisphere and the individual spots in AR###2 were measured in a similar fashion, with the largest individual spot in AR###1 being 1000 times the size in area of the largest individual spot in AR###2, the ratio of the Projected vs. Corrected values for all spots or for individual spots in each AR doesn't really matter that much. The issue is the ratio of AR###1 to AR###2, by whatever measure, SRS or Debrecen.

It's like arguing that a 7.5 kg stone dropped on your toe is different from a 7.3 kg stone, when the alternative might have been a 7.5 g stone opposed to a 7.3 g one.

My interest is still in the unevenness of activity around the sphere in longitudinal terms, and both datasets are showing some interesting patterns.

[SDJ]

Here's an example of what I'm looking at. It is a simple plot of the number of sunspots in Active Regions that appeared during 2000 by their recorded Carrington longitudes.

There were a total of 55,206 spots logged in the Debrecen database for 2000, many of them multiple times if the AR persisted for more than one day or reappeared as a different AR number during more than one rotation.

I filtered them to remove any spots that had Projected Umbra Areas less than 1, which would include the more "inconsequential" spots. That left a total of 18,987 spot records that are plotted here.



Please remember that I am not doing this because of some grand theory other than an uneven distribution of solar activity around the sphere that might offer the possibility of predicting the return of regions that are historically more active than others as a result of the uneven distribution.

I have not contacted the people who prepared the data, nor have I verified the accuracy of every data element in it. I merely found a dataset and decided to have a look at what it shows. This also is not a job for me, just some individual curiosity that I fool around with when I have a few spare minutes.

I've also offered to provide the very substantial content of the dataset (22 MB of Excel files covering 21 years) that I've rendered into a form that can be examined with readily-available software if anyone else wishes to have a look.

[lsvalgaard]

Mar 27, 2010 19:07:38 GMT SDJ said:

Here's an example of what I'm looking at. It is a simple plot of the number of sunspots in Active Regions that appeared during 2000 by their recorded Carrington longitudes.

There were a total of 55,206 spots logged in the Debrecen database for 2000, many of them multiple times if the AR persisted for more than one day or reappeared as a different AR number during more than one rotation.

I filtered them to remove any spots that had Projected Umbra Areas less than 1, which would include the more "inconsequential" spots. That left a total of 18,987 spot records that are plotted here.

Please remember that I am not doing this because of some grand theory other than an uneven distribution of solar activity around the sphere that might offer the possibility of predicting the return of regions that are historically more active than others as a result of the uneven distribution.

I have not contacted the people who prepared the data, nor have I verified the accuracy of every data element in it. I merely found a dataset and decided to have a look at what it shows. This also is not a job for me, just some individual curiosity that I fool around with when I have a few spare minutes.

I've also offered to provide the very substantial content of the dataset (22 MB of Excel files covering 21 years) that I've rendered into a form that can be examined with readily-available software if anyone else wishes to have a look.

One problem with this is that the Carrington longitude is somewhat arbitrary. Its synodic period is 27.2753 days. There is good evidence that the 'real' active longitudes follow a shorter period, closer to 27.0 days. See: www.leif.org/research/Long-term%20Evolution%20of%20Solar%20Sector%20Structure.pdf

[SDJ]

That's a 1 percent difference.

[lsvalgaard]

Mar 27, 2010 20:38:23 GMT SDJ said:

That's a 1 percent difference.

Yes, but it accumulates, After 1 rotation it lags 1% behind, after 2 rotations it lags 2% behind, ... after 13 rotations [~one year] it lags 13% behind. Now 13% of 360 degrees is 47 degrees, so the peaks will be smeared out; after 2 years, more than 90 degrees and after 4 years on the other side of the sun...

[SDJ]

So I guess their meticulous recording of the longitude data was purely a waste of time and the activity really is uniform around the sphere?

Or would the supposed cumulative 1% difference per rotation just mean that the same pattern would just migrate slightly each time? Maybe it does end up on the opposite side after 4 years, but would that destroy the pattern during any single rotation?

I have the data for a period of 21 years, but unfortunately not for 2004. I'll see if it's 3/4 of the way around to the other side in 2003.

[SDJ]

There were 23,789 total spot records in 2003, 7,520 of which had Projected Umbra Areas greater than 0.

[lsvalgaard]

Mar 27, 2010 23:34:14 GMT SDJ said:

So I guess their meticulous recording of the longitude data was purely a waste of time and the activity really is uniform around the sphere?

Or would the supposed cumulative 1% difference per rotation just mean that the same pattern would just migrate slightly each time? Maybe it does end up on the opposite side after 4 years, but would that destroy the pattern during any single rotation?

I have the data for a period of 21 years, but unfortunately not for 2004. I'll see if it's 3/4 of the way around to the other side in 2003.

It is not a waste of time to carefully record the longitude and latitude. This is basic data and very valuable. You will find that if you order the data in rotations that are 27.0 days long that the clustering is much stronger. Or better, 26.85 days in one part of the cycle and 27.15 days in the other part. See my paper I referred to.

[SDJ]

You will find that if you order the data in rotations that are 27.0 days long that the clustering is much stronger.

Thank you!

You do see what I've been talking about, that there is "clustering" of activity in longitudinal terms with substantial degrees of variation in terms of relative activity levels around the sphere.

I can't do anything about the observations in Hungary that were recorded years ago to make corrections for the Carrington longitude determinations, but I hoped that someone else would see from the data that activity is unevenly distributed around the sphere, and I appreciate your having a look at the question.

[lsvalgaard]

Mar 28, 2010 3:42:31 GMT SDJ said:

You will find that if you order the data in rotations that are 27.0 days long that the clustering is much stronger.

Thank you!

You do see what I've been talking about, that there is "clustering" of activity in longitudinal terms with substantial degrees of variation in terms of relative activity levels around the sphere.

I can't do anything about the observations in Hungary that were recorded years ago to make corrections for the Carrington longitude determinations, but I hoped that someone else would see from the data that activity is unevenly distributed around the sphere, and I appreciate your having a look at the question.

The 'clustering' was first noted by Maunder more than a century ago. E.g. articles.adsabs.harvard.edu//full/1905MNRAS..65..556M/0000556.000.html
Here is modern study:
arxiv.org/ftp/arxiv/papers/0909/0909.2973.pdf

[SDJ]

Thanks for the links. It fits another pattern I've noticed in the past, in which I come up with an idea, people say I'm crazy ("astrologer", for instance), and then I find out someone else has already had the same idea, many times long ago, but i'm still crazy.

The relevance to now is that with all the data sources available at present, it should be possible to come up with some rough predictions of when and where significant activity has a higher probability of occurring, which would be of interest to people involved with radio and many other areas.

Back in my first period of more frequent activity on Kevin's board, around the end of 2008, there were some postings in regard to prediction of the "next sunspot" since there was very little going on then. I had a few successful predictions simply based on eyeballing the sphere as it had rotated. That's why I decided to spend all the time tracking down and going through these datasets.

Again, many thanks.

[grian]

Mar 28, 2010 13:47:59 GMT SDJ said:

people say I'm crazy ("astrologer", for instance),

Crazy? No. Simply a poor attempt at being a scientist. I recommend you read Carl Sagan's "The Demon Haunted World" especially the chapter called "The Fine Art of Baloney Detection" Three of your replies to perfectly valid questions I posed failed Sagan's baloney detection kit.

Also instead of entering into a rational debate you complained that "some people were getting upset" about your work every time you could find a sympathetic ear. Pretty much as you've done in your reply to Dr. Svalgaard.

So as far as I'm concerned you have no credibility whatsoever.

[SDJ]

This is AR9169, the largest region that appeared during Cycle 23. Is it still your contention that activity is "smeared" uniformly around the sphere?