Planetary effects on long-term solar activity

[Chapette]

I’d like to bring my research to your valuable attention : ) Peer-review as well as publication and presentations (including at IAU meeting) have already happened without adverse effects. But the topic is certainly considered off-mainstream.

In brief, this is about a model based on the records of solar flares in the last four decades, proposing that solar activity is based on the synergy between an “internal” solar component (of unknown origin, but probably related to good old magnetic fields) and the relative position of the planets Jupiter and Saturn.

Starting from minimal data and assumptions, the model succeeds in reconstructing cycles 22-24 and makes predictions for 25. The “successful reconstruction” includes features that are considered weird by the mainstream, such as the long minimum between cycles 23 and 24, and the “sudden” activity in 2017.

Please advice me about the best way to start this discussion. Would it be better to write a summary of the study here? Personally I’d recommend that you take a look at any of two summaries that I’ve already written, one at my site and one as a guest blog post:
http://www.chapette.net/solar.html
https://www.science20.com/tommaso_do...r_cycle-233408
Of course there is the publication itself:
"A deterministic model for forecasting long-term solar activity", https://www.sciencedirect.com/scienc...64682618303869
https://arxiv.org/abs/1702.00641

I’m very eager to listen to others’ viewpoints. Finally, let me say that there are obvious similarities to pseudoscience, but I think that they are only skin-deep.

[PetersCreek]

Welcome to the CosmoQuest forums, Chapette. If you haven't already done so, please read our rules linked in my signature line below. Please note rule 13 in particular, which applies to the Against The Mainstread (ATM) forum. I also recommend that you read our ATM Forum Advice, also linked below.

While we do allow and welcome offsite references, our rules do require that your presentation take place here in the forum. Any links offered as support should be specific (page, paragraph, figure, etc.) and be pertinent to the point of discussion.

Again, welcome and good luck.

[profloater]

I’d like to bring my research to your valuable attention : ) Peer-review as well as publication and presentations (including at IAU meeting) have already happened without adverse effects. But the topic is certainly considered off-mainstream.

In brief, this is about a model based on the records of solar flares in the last four decades, proposing that solar activity is based on the synergy between an “internal” solar component (of unknown origin, but probably related to good old magnetic fields) and the relative position of the planets Jupiter and Saturn.

Starting from minimal data and assumptions, the model succeeds in reconstructing cycles 22-24 and makes predictions for 25. The “successful reconstruction” includes features that are considered weird by the mainstream, such as the long minimum between cycles 23 and 24, and the “sudden” activity in 2017.

Please advice me about the best way to start this discussion. Would it be better to write a summary of the study here? Personally I’d recommend that you take a look at any of two summaries that I’ve already written, one at my site and one as a guest blog post:
http://www.chapette.net/solar.html
https://www.science20.com/tommaso_do...r_cycle-233408
Of course there is the publication itself:
"A deterministic model for forecasting long-term solar activity", https://www.sciencedirect.com/scienc...64682618303869
https://arxiv.org/abs/1702.00641

I’m very eager to listen to others’ viewpoints. Finally, let me say that there are obvious similarities to pseudoscience, but I think that they are only skin-deep.

Welcome. This was exactly the subject of my first post a few years ago but i think the thread is lost, was lost during the amalgamation.
Will look forward to your, presumably, ATM presentation. My angle was messed up coriolis forces In the radial heat transport cells due mainly to eccentricity of Jupiter’s orbit, with Saturn in there too. I found the longer term records of sun activity did not stay in phase though. Short term predictability works rather well, so it is frustrating.

[Chapette]

Thanks to both, - I'll write more here then in a couple of days.

Profloater, I look forward to hearing more about your idea; you can tell I'm quite interested in proposed underlying mechanisms.

[Reality Check]

Welcome Chapette. This looks close to mainstream to me. You find a statistical correlation between M and X class solar flares and the alignment of Jupiter and Saturn and add an internal solar component to make the peaks diverge over time as in the data. I suspect that flares occurring in the solar atmosphere would be sensitive to gravitational effects.

The main issue I can see is the Maunder Minimum and other minimums and maximums. These imply that the internal solar component can dominate for many decades, suppressing any planetary effects. How do we know that we are not in such a period?
We were in the Modern Maximum until at least the peak of Cycle 23. Could using the earlier Cycle 21 during that maximum in your analysis affected the analysis?

[Chapette]

Hi Reality Check, sorry for the late reply and thanks for the comments and for reading the links! I will answer in more length after the posts below where I'll describe everything. But I can already reply that invoking planetary effects for anything solar is the not mainstream part (although it's not considered too off).

[Chapette]

[I am back with my detailed explanation. For clarity’s sake let me break it over a few posts.]

I will describe the development of a model for describing and predicting solar cycles in terms of solar flares activity. One of its main elements will be the relative motion of the planets Jupiter and Saturn. Let’s just start by saying that the usual solar observable is sunspots and no similar work has been done for flares, even though they have more impact on terrestrial life. Of course, there are longer records for sunspots; but it’s possible that after recording solar flares for forty years the time is right for us to also see what they have to say!

The data used are all M and X-class solar flares in years 1977-2019 incl., from the SMS and GOES satellites X-ray measurements. These cover the whole of the four latest solar cycles, 21-24, with the current cycle 24 nearing its end. Weaker solar flares are not considered.

The story starts with the empirical observation that the number of solar flares during these four cycles tends to increase as Jupiter and Saturn approach alignment, and decreases as they move towards quadrature (=normal angle). This applies equally to alignment on the same side of the sun and to alignment with the sun between them. This plot shows the count of flares as a function of the relative heliocentric ecliptic longitude between the two gas giants:

However, as we progress from cycle 21 to 24 the peak of activity is "dragged" further away from the alignment:

With naked eye, this looks compatible with the staggering between the different lengths of the sunspot cycle (~11 years) and half of the Jupiter-Saturn synodic period (~9.9 years).

So, the question arises whether the evolution of solar activity is a coupled effect between two temporal components: an internal cycle and the approach and retreat of the two giant planets.
In order to test this proposition we'll take three steps:
- Extract from the data two distributions, each corresponding to one of the two proposed components.
- Time the two distributions appropriately.
- Quantify their coupled effect.

[Chapette]

Let’s look at how the two distributions can be extracted from the data.

We will make the assumption that the "internal component" of solar activity is centered on each cycle's temporal middle and spans somewhat less than 11 years, and that the "Jupiter-Saturn component" is centered on the dates of the two planets' alignment and spans mostly -/+45o around it (as seen from the plots above). In addition, we'll assume that each component can be described by a Gaussian distribution. Then the constraints just described give the two Gaussians' means and deviations.

Now we’ll go on to use cycle 21 for the extraction. The reason is that in cycle 21 the dates of the temporal middle and of the alignment happened to lie close (237 days away); so, we'll assume that in that cycle the full deployment of the two effects can be observed, and in any case it is seen better than in the other available cycles.

We can extract each distribution independently by fitting the envelope of the data of cycle 21 with the constraints described above for the Gaussians' means and deviations. (In the plots, the solid red line shows the mean of each Gaussian, and the dotted lines show the distance of two standard deviations.) The two resulting Gaussians' constants are close, and they were refitted with equal values.

[Chapette]

Now we can use the two distributions extracted from cycle 21, for the rest of the cycles. We do this by repeatedly centering the "internal Gaussian" on the dates of the cycles' temporal middles, and the "Jupiter-Saturn Gaussian" on the dates of the two planets' alignments.

The middle of cycle 24 is still unknown, so it was estimated. On average, the distance between consecutive Jupiter-Saturn alignments and the cycles' temporal middle increases by 396 days between consecutive cycles. This number was used for estimating the temporal middle for the ongoing cycle 24, based on 23. (The number comes from the synodic half period average of 3,634 and the sunspot cycle average of 4,030 days).

The model was developed initially using data up to the end of year 2016, so the time range 1977-2016 is plotted, but see below for more.

Finally, let’s assume that the coupling of the two components is expressed by their common area, plotted here as a binned histogram:

This purple distribution is proposed to describe the long-term solar activity in terms of M and X-flares.

And here it is again, overlayed with the observational data:

The agreement is notable, as seen in the general features of each cycle, such as start, time span, intensity and evolution. There are short-scale deviations that still need to be understood, but the predicted distribution also contains short-scale features that are considered puzzling by the mainstream. Such are the deep minimum before cycle 24, and the abrupt decrease at the end of 2015 (around day 14,500 in this plot).

Again, the only assumptions are that there are two Gaussians, that they are centered repeatedly on specific dates, and that their common area quantifies their coupled effect. The only input data were the flares of cycle 21.

[Chapette]

Finally, let’s extend the model beyond 2016 by repeating the two kinds of Gaussians, and let’s include the new observations.

Note that during cycle 24, a second "Jupiter-Saturn Gaussian" overlaps with the "internal Gaussian". At the time of first publication it was predicted that this will lead to a new, smaller, surge of activity. This happened indeed and was the “inexplicable” solar activity in late summer 2017.

As about the next cycle, it is seen to begin around the summer of 2020 and be also characterized by a "double overlap". So, the model points to a spread-out weak cycle, with two distinct ranges of activity. (Note that this is a falsifying prediction.)

That was the model for the long-term modeling of solar flares activity, based on both internal solar processes (most probably good old magnetism) and the relative motion of Jupiter and Saturn. I will add that there are a few other similar models for sunspots, but afaik they are all based on frequency analysis, not on statistical analysis of data. I will finish by saying that so far I don’t have a proposal for the underlying mechanism (although I have a few guesses).

Looking forward to your comments and constructive attack!

[Reality Check]

Let’s look at how the two distributions can be extracted from the data....

The problems start with using cycle 21. As you note the two Gaussian distributions are close. The result is that cycle 21 could be fitted with a single Gaussian distribution, i.e. no separate contributions. There is nothing about the relative contributions of the internal component and planetary component. Maybe there is a hidden assumption that they are equal?
The next problem is applying this to the next cycles. This is the invalid assumption that the contributions do not change. The internal component contribution certainly changes (see my post above). The planetary contribution will also change (Jupiter and Saturn are not in circular orbits). That may be negligible but you need to show this.

[Shaula]

In your assessment above you have hand crufted in the Gaussian parameters. Have you done any sensitivity modelling on this or tried any approaches which make no such assumptions and seen what comes out of them? Are there combinations of Gaussians that fit the model better? And why Gaussians? Do other distributions or combinations of distributions fit the data better? If we assume that flares are probabilistic with the probability tied to the planetary positions wouldn't something like a Poisson curve make more sense?

Essentially my feedback would be that just eyeballing some graphs and looking at one model doesn't really prove much. I'm not convinced, looking at the graphs, that your curves are particularly good fits and would like to see some stats to show that they are at least the most plausible of a range of checks. You might run into the issue of low sample sizes here. Have you looked at flare proxy variables like beryllium-10 to try to go back over more cycles (recognising that the signal may be too noisy to do much with)?

[profloater]

Thanks to both, - I'll write more here then in a couple of days.

Profloater, I look forward to hearing more about your idea; you can tell I'm quite interested in proposed underlying mechanisms.

You are looking at correlations, which are interesting of course, I was looking for a cause. We know that the dominant forces inside the Sun are electromagnetic fields operating on the plasma. However My point was that there are still mechanical forces which must be present even if they are second or third order. The heat from the centre of the Sun travels outwards in zones in which the final zone is a kind of columnar flow. Because the Sun is rotating there will be Coriolis forces changing the radial flow into spirals. Jupiters gravity will exert tidal forces onto the Sun and Jupiters orbit is eccentric, so the amplitude of these forces must vary with Jupiters period and of course is modulated by all the other planets To a lesser degree. Every column of heat transfer includes a inward radial flow of the cooler plasma and this is also subjected to Coriolis forces, creating a complex picture. Seeing recent evidence that many stars are more volatile than our sun, I am also wondering whether this stirring by the massive planets has overall a stabilising effect on the Sun.

[profloater]

Are you aware of Landscheit’s work on solar flares? He correlates both with planets and direct effects on Earth. Same idea that you are working on.
https://link.springer.com/chapter/10...-015-7692-5_47

[Chapette]

In your assessment above you have hand crufted in the Gaussian parameters. Have you done any sensitivity modelling on this or tried any approaches which make no such assumptions and seen what comes out of them? Are there combinations of Gaussians that fit the model better? And why Gaussians? Do other distributions or combinations of distributions fit the data better? If we assume that flares are probabilistic with the probability tied to the planetary positions wouldn't something like a Poisson curve make more sense?

Essentially my feedback would be that just eyeballing some graphs and looking at one model doesn't really prove much. I'm not convinced, looking at the graphs, that your curves are particularly good fits and would like to see some stats to show that they are at least the most plausible of a range of checks. You might run into the issue of low sample sizes here. Have you looked at flare proxy variables like beryllium-10 to try to go back over more cycles (recognising that the signal may be too noisy to do much with)?

Thank you - there are four points made here:

1. > I'm not convinced, looking at the graphs, that your curves are particularly good fits

I think that the point missed here is that the criterion is not the best-fitting function; the criterion is the recosntruction of the three following cycles.

2. > In your assessment above you have hand crufted in the Gaussian parameters. (...) And why Gaussians?

Gaussians is the first choice for modelling physical systems with unknown behaviour, due to them being the most ubiquitous distributions in nature, in a word this is the simplest choice.
Please note the slight difference: Putting by hand some of the parameters helps making a fit more reasonable - not biasing it. (Actually any good fit should start with an initial range of accepted parameters.) The subsequent results serve as a test for the original assumption of the parameters.
In other words it's not a bug, it's a feature of the model which is being tested.

These two answers also apply to the question of whether there are "other distributions or combinations of distributions fit the data better". The idea is not to fit the data best, but to model the solar cycle. If there are other distributions or combinations, I haven't thought of a physical meaning to attribute to each of their components, so I don't know which of the endless number of combinations to choose.

3. > Have you looked at flare proxy variables like beryllium-10 to try to go back over more cycles

No, this is on the to-do list.

4. > If we assume that flares are probabilistic with the probability tied to the planetary positions wouldn't something like a Poisson curve make more sense?

Can you say more on this?
Actually, I didn't write about this above, but the Poissonian distribution was taken into account, in the statistical uncertainty of the model (if you look at the final plots there is both statistical and systematic uncertainty thrown together, but I avoided going into details). I think that this is the only way to apply it in such a case: there is Poissonian uncertainty in the count, because of it being probabilistic, but it doesn't affect the choice of other distributions in a model.

[Chapette]

The problems start with using cycle 21. As you note the two Gaussian distributions are close. The result is that cycle 21 could be fitted with a single Gaussian distribution, i.e. no separate contributions. There is nothing about the relative contributions of the internal component and planetary component. Maybe there is a hidden assumption that they are equal?
The next problem is applying this to the next cycles. This is the invalid assumption that the contributions do not change. The internal component contribution certainly changes (see my post above). The planetary contribution will also change (Jupiter and Saturn are not in circular orbits). That may be negligible but you need to show this.

Thanks again - three points here:

1. > As you note the two Gaussian distributions are close. The result is that cycle 21 could be fitted with a single Gaussian distribution

As written also in the reply above, the point is not to get the best fit but to reproduce the cycles, which seems to happen when using the assumption of two Gaussians.

2. > There is nothing about the relative contributions of the internal component and planetary component. Maybe there is a hidden assumption that they are equal?

If you look again at the description of the fitting there is a quick mention about this, "The two resulting Gaussians' constants are close, and they were refitted with equal values". To make it more clear, there was no initial assumption about the constants of the Gaussians (i.e. the height), but the fit on cycle 21 gave very close results for both of them, although they were fitted independently. Then, they were assumed to be equal for simplicity, and the fit was redone.

3. > This is the invalid assumption that the contributions do not change. The internal component contribution certainly changes (see my post above). The planetary contribution will also change (Jupiter and Saturn are not in circular orbits). That may be negligible but you need to show this.

There is a valid point here, although there are indications that it won't change much. Let me break it down:

> The internal component contribution certainly changes

There is no certainty about this - in the previous post you said that it might be an explanation for the Maunder minimum, which might be correct but it is a conjecture. (Even like this, there is probably no reason to expect that it will change dramatically over four contiguous cycles.)

> The planetary contribution will also change (Jupiter and Saturn are not in circular orbits)

This is a good point. I suspect that indeed it will be negligible, but it's something to consider.

Overall, I said that there are indications about not much changing if the "height" of the contributions is different. This comes from the plot with the two Gaussians over the extended time period (it's the plot after the words "binned histogram").
I have tried changing the height of either or both components. But in the end the model comes from their common area. This means that changes in the height don't really translate to significant differences in the common area, and thus in the final result.

[Chapette]

The main issue I can see is the Maunder Minimum and other minimums and maximums. These imply that the internal solar component can dominate for many decades, suppressing any planetary effects. How do we know that we are not in such a period?
We were in the Modern Maximum until at least the peak of Cycle 23. Could using the earlier Cycle 21 during that maximum in your analysis affected the analysis?

Every attempt at modeling the solar cycle hits the Maunder wall sooner or later : )

Although I don't have the data to quantify it (at least until I look at solar activity proxies), I have formed a possible explanation about the Maunder minimum and the modern maximum.
Let's assume that there is an effect here (probably gravitational) which is at its strongest when planets align and at its weakest when they are at quadrature. Let's also assume that planets Uranus and Neptune contribute to this effect, although only at very large timescales - they might be distant, but their presence is constantly there and it changes very slowly.
Now, in the recorded sunspot activity, there have been only three conjunctions of Uranus and Neptune, in 1650, 1821 and 1993. These coincide with the beginning of Maunder, the Dalton minimum and the modern maximum.
However, there is a crucial difference between the two minima and the maximum. At the maximum, when the conjunction of Jupiter+Saturn took place it was also close to alignment with Uranus+Neptune. At the minima, the two conjuncted pairs were at quadrature with each other. This could probably affect the total strength of those cycles.
Especially for Maunder, this quadrature happened to occur during two consecutive alignments of Jupiter+Saturn, i.e. possibly disrupting the development of two consecutive solar cycles. I think that this might have something to do with the following "unwinding" of the cycle process.

> Could using the earlier Cycle 21 during that maximum in your analysis affected the analysis?

This is a good point, and after all if an effect from the outermost planets exists then it will have some continuous nature as well. And even if it doesn't exist this is still a valid point. I'll put some more thinking into it.

[Shaula]

I think that the point missed here is that the criterion is not the best-fitting function; the criterion is the recosntruction of the three following cycles.

My point was I don't see any stats showing the you have reconstructed the cycle very well. All of the analysis seems to be that the fit 'looks' good (yes, you've included uncertainty bars but the model and observations frequently disagree at a level way beyond these ranges). So I don't think you have reconstructed the cycle - you've got some plots which sort of match up but not that well and seem to frequently disagree with what we are seeing.

Gaussians is the first choice for modelling physical systems with unknown behaviour, due to them being the most ubiquitous distributions in nature, in a word this is the simplest choice. Please note the slight difference: Putting by hand some of the parameters helps making a fit more reasonable - not biasing it. (Actually any good fit should start with an initial range of accepted parameters.) The subsequent results serve as a test for the original assumption of the parameters.
In other words it's not a bug, it's a feature of the model which is being tested.

I know why people use Gaussians - I wanted to see if you'd thought about the underlying processes or just gone with them as the easy option or first approximation.

And I disagree with your logic about putting in parameters by hand. It totally biases the assessment. Especially if you haven't looked at any other alternatives. Your hypothesis is that the effects are linked to planetary influences. If the only fit you have considered is one with parameters linked to planetary influence then you are not testing your hypothesis particularly strongly. Effectively you are only looking for what you want to see and ignoring anything that doesn't fit the narrative. Consistent with is about as strong as you could possibly get from this.

These two answers also apply to the question of whether there are "other distributions or combinations of distributions fit the data better". The idea is not to fit the data best, but to model the solar cycle. If there are other distributions or combinations, I haven't thought of a physical meaning to attribute to each of their components, so I don't know which of the endless number of combinations to choose.

As I said above - that means you are only really looking for what you want to see. I'm not talking about you doing a full breakdown of every possible combination but some testing would be nice. You are taking a very model driven approach, I guess that what I would regard as stronger evidence would be seeing your parameters coming out from a data driven approach. Even if they are not the 'best' fit coming back with evidence that any old pair of random distributions isn't better gives some confidence.

Can you say more on this?
Actually, I didn't write about this above, but the Poissonian distribution was taken into account, in the statistical uncertainty of the model (if you look at the final plots there is both statistical and systematic uncertainty thrown together, but I avoided going into details). I think that this is the only way to apply it in such a case: there is Poissonian uncertainty in the count, because of it being probabilistic, but it doesn't affect the choice of other distributions in a model.

I was more talking about replacing your Gaussians with a Poissonian. Make the assumption that flares are independent of each other but have a base probability driven by the planetary gravitational influence. Then you can model your variation in probability continuously based on planetary positions (range to Sun or something similar) and fit on a single scaling factor for each planet considered. That would remove the rather arbitrary assumption that the effect is only there for the 45 degrees around the peak influence. If you looked at the problem like this a Poissonian would make more sense than a Gaussian, which was where I was going.

So I guess what I take away from what you have done is that you have a basic model for flare count that is weakly consistent with observations (which is a good start) but that you are a long way from anything resembling proof that this is a valid model for what is going on. This isn't dismissal, by the way, it is just me trying to work out where you are in the analytical process. You need either considerably stronger statistical proof of what is being seen or a well constructed phenomenological model that naturally leads to your model of the observations (and your explanations are the starting point for that but not yet rigorous physics).

Good luck developing your work, its at a early stage and hasn't ruled itself out yet which is always a good start. But its not yet compelling.

[profloater]

I can only offer that sunspot activity appears to go out of phase over many cycles and that might explain the Maunder minimum if the internal frequency of the sun is close but not equal to the gravity forcing frequency, the minima and maxima then represent a beat frequency, too long for ancient observations to help. If planet gravity is a parameter why not use sinusoidal forcing? Also the internal process may not be as rhythmical as the planets, or may have non independent variables making the planet forcing important but unpredictable unless a complex model is forthcoming. I would like to come back in a few hundred years of good records to revisit the model

[Reality Check]

.Although I don't have the data to quantify it (at least until I look at solar activity proxies), I have formed a possible explanation about the Maunder minimum and the modern maximum.....

The Maunder minimum and the modern maximum are examples of the many minima and maxima which you will have to explain. Grand solar minima and maxima lists 8 such events. A guess that you have to add Uranus and Neptune to explain these events invalidates your idea. In your idea, Jupiter and Saturn are the major influences on solar activity. Adding the less influential Uranus and Neptune, says that they can sometimes dominate.

1650, 1821 and 1993 do not coincide with the beginning of grand solar minima and maxima. Close with the Wikipedia 1645 start of the Maunder minimum but the Dalton minimum is 1790 and Modern maximum 1914.

[slang]

30 days gone, looong gone... let's call it a few days of COVID-19 grace period. Sadly, not used. Anyway, thread closed, per rule 13 this topic may not be raised again without express prior moderator permission.

[Swift]

After a request from the OP and a discussion among the Moderation Team, we have granted a two week extension to this thread.

[Chapette]

Apologies for the late reply!

The Maunder minimum and the modern maximum are examples of the many minima and maxima which you will have to explain. Grand solar minima and maxima lists 8 such events. A guess that you have to add Uranus and Neptune to explain these events invalidates your idea. In your idea, Jupiter and Saturn are the major influences on solar activity. Adding the less influential Uranus and Neptune, says that they can sometimes dominate.

1650, 1821 and 1993 do not coincide with the beginning of grand solar minima and maxima. Close with the Wikipedia 1645 start of the Maunder minimum but the Dalton minimum is 1790 and Modern maximum 1914.

I see that you refer on extrema at very large timescales. This is definitely outside the scope of my current analysis. (I've seen "modern maximum" usually referred to as around cycle 20 and that's what I meant above.)

I think that if the presence of Uranus and Neptune plays a role then it has to be taken into account accordingly, but that this doesn't invalidate the overall idea: after all, if there are planetary effects then they apply (to different degrees) to all planets.

[Chapette]

I'll catch up on the comments here over the next days.

By the way, some of the time that I let slip was taken up by a relevant work - an attempt to involve the five innermost planets in the timing of solar flares. In case you were looking for something to read: https://arxiv.org/abs/2006.10694

[Reality Check]

I see that you refer on extrema at very large timescales. This is definitely outside the scope of my current analysis.....

Which is not correct. Neglecting that there are solar minima and maxima that do not match your idea invalidates your idea unless you can explain them. Consider someone who says that they have a theory that the solar output is constant in the long term because they restrict themselves to a scope of analysis where the TSI will be constant. They might use proxies that are only valid for thousand years and average over centuries to remove the solar cycle. Their theory is obviously wrong because textbook astrophysics and wider proxy data says that the solar output will and does vary.

"if there are planetary effects then they apply (to different degrees) to all planets" would make your idea invalid overall. There will be planetary effects from all planets (and even dwarf planets, asteroids, comets, etc. ). If you do not know what the relative effects from the planets are then Neptune and Uranus are guesses. You have to explain why you can omit Mercury but include Uranus.
Even with just Jupiter and Saturn, there is the question of why Saturn is included when people tend to treat Jupiter as the dominant influence on the Sun. An example from this thread is Landscheidt's paper Cycles of Solar Flares and Weather (a cyclic pattern in flares claimed to be related to "the tidal forces of the planets Venus, Earth, and Jupiter").

Speaking of the tidal forces: Resonance between Alfven Waves and Planetary Tides on the Sun (PDF) has a table of theoretical tidal forces on the Sun relative to that of Earth. Saturn is about a 20th of Jupiter and a tenth of Earth. More importantly look at Neptune and Uranus.

[PetersCreek]

I'll catch up on the comments here over the next days.

Chapette,

You're almost a week into your 2-week extension and in spite of promising to catch up on comments, you haven't posted anything during the last six days. This thread is once again closed.