Difference ISN vs 10.7 Solar Flux vs High Resolution SN

[Patrick P.A. Geryl]

17 uren geleden, Philalethes zei:

To me it doesn't look like the formula was a very good proxy to begin with. The values are all over the place from what it looks like to me, looks like someone has tried to force linear regression on a relationship which isn't necessarily linear at all, or which has more variance than you know due to the low sample size of just a few cycles; when you already have some errors as large as 10%, finding subsequent errors of 20% doesn't seem very surprising.

As I've mentioned previously there is also some literature on both inter- and intracycle variance in relative sunspot sizes, such as e.g. this paper that reports how there were fewer small spots during the beginning of SC23, compensated by more such spots later in the cycle.

Your argument falls away due to the linearity of the high resolution 1K and 2K data series. They proof there is a linearity between the amount of sunspots and the 10.7 Solar flux. Therefore the ISN should be linear. It isn’t … and I know why… They don’t count sunspots that are only a few hours on the Sun! They don’t count grey sunspots… 

Edited May 3, 2023 by Patrick P.A. Geryl

[Philalethes]

16 hours ago, Patrick P.A. Geryl said:

Your argument falls away due to the linearity of the high resolution 1K and 2K data series. They proof there is a linearity between the amount of sunspots and the 10.7 Solar flux. Therefore the ISN should be linear. It isn’t … and I know why… They don’t count sunspots that are only a few hours on the Sun! They don’t count grey sunspots… 

But your point is that the relationship between high-resolution and flux still holds, is it not? Even if they were linear before and linear now, that wouldn't be relevant for the other measure of sunspots being used if that one was neither linear before nor is linear now, or if the variance is too high to establish the actual relationship. That could simply be variability in relative sunspot sizes, as has been described in the literature before (such as the paper I mentioned). By all means, it could be interesting to look at and see how the relative sunspot sizes develop over time, but to proclaim based on this that SC25 will be the the weakest since the Dalton minimum and that we might be headed for another Maunder minimum seems premature, to put it mildly.

[David Silver]

This bizarre logic just doesn’t add up. A minor peak in activity recently may indicate that we will see a double-peak cycle as has happened many times before. But there is no reason to believe the cycle will be two years shorter for the first time in history. Why does the fundamental knowledge of the 11 year cycle not factor into your extremely peer reviewed calculation?  If you are right, I’ll be happy to try and learn something new from you about space weather. But this is a stretch even for the most open-minded of us. Wwhhhyyyyyyyy

[Jesterface23]

@Patrick P.A. Geryl

I want to come back to this from your document. When and how are the different 10.7cm radio flux sources used?

[Patrick P.A. Geryl]

5 uren geleden, Jesterface23 zei:

@Patrick P.A. Geryl

I want to come back to this from your document. When and how are the different 10.7cm radio flux sources used?

This is the best specific and most important question I got since a very long time…

Jan Alvestad found that Canada didn’t use flare compensation!

Then I researched the older data and found DOZENS OF FAULTS! Some more then 10 percent…

MONTHLY DATA!

The data from Moscow did use flare compensation… but not as good as Jan Alvestad…

Even to this day, peer reviewed articles use the wrong data!

it is extensively described in my article…

Edited May 4, 2023 by Patrick P.A. Geryl

[Jesterface23]

12 hours ago, Patrick P.A. Geryl said:

This is the best specific and most important question I got since a very long time…

Jan Alvestad found that Canada didn’t use flare compensation!

Then I researched the older data and found DOZENS OF FAULTS! Some more then 10 percent…

MONTHLY DATA!

The data from Moscow did use flare compensation… but not as good as Jan Alvestad…

Even to this day, peer reviewed articles use the wrong data!

it is extensively described in my article…

Within your data, what year/month/day did you stop using the wrong data? At what points were you using the wrong data?

[Patrick P.A. Geryl]

2 uren geleden, Jesterface23 zei:

Within your data, what year/month/day did you stop using the wrong data? At what points were you using the wrong data?

Jan Alvestad since 2012, We started using the right data from the beginning of measurements 5 years ago. But! Till this day peer reviewed articles are published with the wrong data! Only after reading my article they see what is wrong.

The wrong data set is one of the reasons they never were able to find a formula like I did….

[Christopher S.]

4 hours ago, Patrick P.A. Geryl said:

Jan Alvestad since 2012, We started using the right data from the beginning of measurements 5 years ago. But! Till this day peer reviewed articles are published with the wrong data! Only after reading my article they see what is wrong.

The wrong data set is one of the reasons they never were able to find a formula like I did….

I remember reading your older articles and discovering this. I'm pretty sure the issue is continuously gaslit moving forward and the lack of acknowledgement towards observational capabilities given higher resolution imagery is severely hampering any further developments in this specific field of study. My hope is that we wipe the slate clean and start over with a more uniform and viable method of determining the cyclical behavior of the Sun, which is used to deride shorter-term activity. Given the erroneous and over-eager conclusion-jumping you have successfully uncovered, there ought to be doubts in the current methods and narration which are used to facilitate warnings of satellite damage, over-reaching to conclusions about periodic phenomenon.

[3gMike]

5 hours ago, Patrick P.A. Geryl said:

Jan Alvestad since 2012, We started using the right data from the beginning of measurements 5 years ago. But! Till this day peer reviewed articles are published with the wrong data! Only after reading my article they see what is wrong.

The wrong data set is one of the reasons they never were able to find a formula like I did….

Therein lays the problem with your claims. You claim to have tracked ascending phase of the last seven phases yet your supposedly 'right data' only became available in 2012. So your formula for high resolution spots was based only on the last cycle.

This paper https://agupubs.onlinelibrary.wiley.com/doi/pdf/10.1002/swe.20064 indicates that those responsible for the long-term Penticton record of 10.7cm flux became aware of changes relative to the ISN as early as 2013. So it is fair to say that there has been a change, but I believe that your claim that it relates to small spots only visible in 1K / 2K data is tenuous, and certainly does not support your claim that we may be entering a new Maunder Minimum. Given smoothed sunspot number of 98.7 for October 2022 it would require monthly mean sunspot numbers to remain static at around 100 for the next 6 months in order for peak value to fall below your target of 118. That seems very unlikely given current data.

It has also been noted that flare activity in 2022 was particularly high relative to sunspot number. The first four months of 2023 also show significant flaring. This is not consistent with your claim for reducing activity.

[Philalethes]

1 hour ago, 3gMike said:

Given smoothed sunspot number of 98.7 for October 2022 it would require monthly mean sunspot numbers to remain static at around 100 for the next 6 months in order for peak value to fall below your target of 118. That seems very unlikely given current data.

I think it could be somewhat higher and still fall short of that, actually; at least if I'm understanding it right. If you're talking about how low the mean SN has to be per month for the next 6 months for the SSN to not reach 118, the value is closer to 130 per month if I'm not mistaken. Here's a projection of what the SSN will be from 2022-11 to 2023-04 if the mean SN is set to various values for the next six months, including this one (i.e. 2023-05 to 2023-10):

If the mean SN were to remain at only 100, it'd fall quite short of a 118 SSN; it'd still have a good lead on SC24, though, and be well set to keep that lead as SC24 tumbles.

[Jesterface23]

7 hours ago, Patrick P.A. Geryl said:

Jan Alvestad since 2012, We started using the right data from the beginning of measurements 5 years ago. But! Till this day peer reviewed articles are published with the wrong data! Only after reading my article they see what is wrong.

The wrong data set is one of the reasons they never were able to find a formula like I did….

You have a formula that seems to work and that is fine.

I come back to this,

On 4/7/2023 at 4:53 PM, Jesterface23 said:

On 4/7/2023 at 12:03 PM, Patrick P.A. Geryl said:

So I repeat my question: what is going on? Why the never before seen difference?

There's a first time for everything. With there being data from 6 prior complete solar cycles, maybe the chance for error is possibly higher.

Your dataset for this no longer goes back to solar cycle 19, now it only goes part way though solar cycle 24. You still need to discover and learn the limits of your new dataset along with what new solar cycles will bring, and it will take a while.

[Patrick P.A. Geryl]

Op 5/5/2023 om 00:49, Philalethes zei:

I think it could be somewhat higher and still fall short of that, actually; at least if I'm understanding it right. If you're talking about how low the mean SN has to be per month for the next 6 months for the SSN to not reach 118, the value is closer to 130 per month if I'm not mistaken. Here's a projection of what the SSN will be from 2022-11 to 2023-04 if the mean SN is set to various values for the next six months, including this one (i.e. 2023-05 to 2023-10):

If the mean SN were to remain at only 100, it'd fall quite short of a 118 SSN; it'd still have a good lead on SC24, though, and be well set to keep that lead as SC24 tumbles.

YOU overlook some basic math.

You should compare the 10.7 flux graphs.

Then the worsening relationship with the SSN.

and keeps worsening🤨🧐

Look surprised🤔

 

Edited May 6, 2023 by Patrick P.A. Geryl

[Philalethes]

@Patrick P.A. Geryl:

Quote

 

So yes, it is all related.

By the way, you focus on the SSN, while you should focus on the 10.7 flux. Will fail.

 

If you want to discuss the flux, let's do it here; your discussion in the other thread is only marginally related to the terminator framework (it's only about your criticism of it and your own ideas).

So, first of all: when you state "you will fail" like that, I hope you realize that this is a deeply unscientific attitude; your models are not the actual future, and you don't know what the future will bring. Until you come up with a model that successfully predicts the future, you should refrain from telling others that what they're doing will fail, especially when it's based on reasonable evidence and seems to fit the current cycle well so far (talking about Scott's model here, which I'm not necessarily saying is entirely right either). So far you only have a model that you've fitted to the past (and not always so well at that either, apparently), not one that predicts the future.

Also, you're talking about focusing on the flux instead, so let's do that. Could you tell us exactly what your prediction is with regards to the flux? You talked about how you believe this cycle will peak in May, are you talking about that representing peak flux, or peak SSN, or both? And in those cases, using which data sources and what smoothing? You need to at least make some clear predictions so that you and others can judge your own hypotheses objectively, otherwise you can just keep hand-waving and talking about how everything supposedly fits great with the latest ad-hoc correction to your model, which is perhaps the oldest trap in the book in a scientific context (confirmation bias).

And if we're talking about the fluxes, I really don't see any signs of what you're claiming either. You say SC25 will be the weakest since the Dalton minimum, but since we don't have any measurements of flux from then you have to use proxies like sunspot numbers to measure that, which apparently you now try to hand-wave away. And even if we just look at recent developments, what exactly are the indications of flux peaking? Using 30-day smoothing there's been a drop recently from ~200 down to ~140, a ~30% drop, but even SC24 had a drop just like that before surging back up again. Similar has been seen in many cycles. And that's just using the F10.7; if we look at the F30, which I've read in some papers is typically an even more consistent measure, the drop has been even smaller, from ~130 to ~105, only ~20%. And using 365-day smoothing there's still no drop at all, and the peak is higher than SC24; and given how using 30-day smoothing also shows a peak that's already higher than SC24 (as well as SC20) in terms of flux, then your claim would already technically be invalidated if that's your criterion.

So really, I still struggle to see most of your points, but feel free to keep clarifying; maybe you're right and everyone else just doesn't get it.

Edited May 6, 2023 by Philalethes
typo

[Philalethes]

10 hours ago, Patrick P.A. Geryl said:

You should compare the 10.7 flux graphs.

Well, I outlined some data about the fluxes above, but let's look at some. First of all, let's look at a pitfall that would be easy to fall into. Let's look at the minima of the 10.7 fluxes of the recorded cycles using a 365-day smoothing:

Notice something here? The roughness at the troughs is clearly interfering with an accurate placement of the minima, so if we were to use this to compare the cycles in terms of flux we'd end up with mismatches. Note particularly how the last minimum is placed a small bump that's clearly much further left than most of the other minima (and conversely how the minimum between 1980 and 1990 is placed much further to the right in the same manner). If we were to compare the cycles like this, we would expect to see SC25 getting displaced to the right and SC22 getting displaced to the left. Let's see:

Lo and behold, exactly as expected: there seems to be a clear relationship between rise time and amplitude, but SC25 here instead starts out with a long flat stretch (due to the displaced minimum) and then suddenly surges through SC24; and conversely, SC22 somehow ends up slightly to the left of the mighty SC19, rising unexpectedly fast due the minimum displaced in the other direction.

To fix this, I would suggest still using a 365-day smoothing for the values we're looking at, so we can see the detail better, but to use a much longer smoothing period to get a better approximation of the actual large-scale troughs. Let's see what the first plot looks like with 5-year smoothing instead:

Much better. There are still hints of the same irregularities, but they're not nearly as pronounced now. Using these dates for the minima we should expect to get a much more sensible plot of the different cycles. It looks like this (note how SC19 falls away due to the data not going far enough back to include its preceding minimum when smoothing it this way):

That looks much more sensible to me. Here the relative rise times and amplitudes seem to be much more coherent. What that holds for the future is of course uncertain, because it could flat out like SC20, or it could keep rising like some of the other cycles (whether that violates the general relationship or not depends on whether there's still some displacement of the minimum or not, but this is a small sample size anyway, for all we know cycles can full well violate this relationship). What is at least clear is that we're already above SC24, so it's not possible for this to be the weakest cycle since the Dalton minimum by virtue of that alone.

We can also look at the 30-day smoothed flux for more detail (only including SC23 and SC24 here, since more quickly get messy due to the rougher nature of the plots):

Looking at this and all of the above I still have a hard time understanding why you make some of the statements you do. To me they sound very sensationalist, almost as if you want there to be an amazing discovery here more than there really is one.

[3gMike]

14 hours ago, Patrick P.A. Geryl said:

YOU overlook some basic math.

You should compare the 10.7 flux graphs.

Then the worsening relationship with the SSN.

and keeps worsening🤨🧐

Look surprised🤔

 

Patrick, I think you might be well advised to re-examine your formulae. You have been vague about the precise source of your data, but since you use the STAR 1K and 2K data, and claim that these are better than ISN, I have used the graph from Jan Alvestad's Solen website as a starting point. It is reproduced below with numbers added to indicate measurement points.

Measurement point 1 is June 2012 and represents the first point where this dataset became available, and possibly where formulae were calculated

ISN = 85 x 0.635 + 64 = 118 )

1K = 104 x 0.5 + 64 = 116 )       Flux = 119 so all looks good for your formulae

2K = 160 x 0.326 + 64 = 116.2 )

Measurement point 2 represents the maxima of Cycle 24.

ISN = 118 x 0.635 + 64 = 139 )

1K = 170 x 0.5 + 64 = 149 )     Flux = 142 so we see some spread, with ISN actually being closest

2K = 250 x 0.326 + 64 = 145.5 )

Measurement points 3, 4 and 5 let us examine what happens in ascending phase of Cycle 25

             ISN calc flux        1K calc flux      2Kcalc flux      measured flux

Point 3      109.7                 117.5                 118.8                 115

Point 4      116.7                 124.5                 126.6               123

Point 5      127.5                 138.5                 140                  140

We see that the 1K and 2K figures are closely related to each other, and to measured flux while the ISN numbers seem to be low.

Comparing measurements 1 and 7 we find something interesting

                                         Point 1             Point 7       Difference (7 - 1)

ISN sunspot number         85                 76               -9

1K sunspot number         104                118              +14

2K sunspot number        160                178              +18

So it seems that the original figure used to set the multiplier for ISN sunspot number may have been non-representative.

In order to test that theory I altered the ISN multiplier to 0.75. Note that this is by no means a refined number, but it produces these results :-

             ISN calc flux        1K calc flux      2Kcalc flux      measured flux

Point 1       128                    116                    116.2               119

Point 2      152.5                 149                   145.5               142  

Point 3      118                     117.5                 118.8                115

Point 4      127                    124.5                 126.6               123

Point 5      139                    138.5                 140                 140

Point 6       88                      85                      82.3                84

Point 7      121                    123                    122                  124

Apart from Point 1, where I thought ISN spot number might be anomalous, the calculated fluxes are now much better matched. It might be possible to tune it a bit more but I hope that you can understand my point.

[Philalethes]

4 hours ago, 3gMike said:

Patrick, I think you might be well advised to re-examine your formulae. You have been vague about the precise source of your data, but since you use the STAR 1K and 2K data, and claim that these are better than ISN, I have used the graph from Jan Alvestad's Solen website as a starting point. It is reproduced below with numbers added to indicate measurement points.

Measurement point 1 is June 2012 and represents the first point where this dataset became available, and possibly where formulae were calculated

ISN = 85 x 0.635 + 64 = 118 )

1K = 104 x 0.5 + 64 = 116 )       Flux = 119 so all looks good for your formulae

2K = 160 x 0.326 + 64 = 116.2 )

Measurement point 2 represents the maxima of Cycle 24.

ISN = 118 x 0.635 + 64 = 139 )

1K = 170 x 0.5 + 64 = 149 )     Flux = 142 so we see some spread, with ISN actually being closest

2K = 250 x 0.326 + 64 = 145.5 )

Measurement points 3, 4 and 5 let us examine what happens in ascending phase of Cycle 25

             ISN calc flux        1K calc flux      2Kcalc flux      measured flux

Point 3      109.7                 117.5                 118.8                 115

Point 4      116.7                 124.5                 126.6               123

Point 5      127.5                 138.5                 140                  140

We see that the 1K and 2K figures are closely related to each other, and to measured flux while the ISN numbers seem to be low.

Comparing measurements 1 and 7 we find something interesting

                                         Point 1             Point 7       Difference (7 - 1)

ISN sunspot number         85                 76               -9

1K sunspot number         104                118              +14

2K sunspot number        160                178              +18

So it seems that the original figure used to set the multiplier for ISN sunspot number may have been non-representative.

In order to test that theory I altered the ISN multiplier to 0.75. Note that this is by no means a refined number, but it produces these results :-

             ISN calc flux        1K calc flux      2Kcalc flux      measured flux

Point 1       128                    116                    116.2               119

Point 2      152.5                 149                   145.5               142  

Point 3      118                     117.5                 118.8                115

Point 4      127                    124.5                 126.6               123

Point 5      139                    138.5                 140                 140

Point 6       88                      85                      82.3                84

Point 7      121                    123                    122                  124

Apart from Point 1, where I thought ISN spot number might be anomalous, the calculated fluxes are now much better matched. It might be possible to tune it a bit more but I hope that you can understand my point.

The problem does indeed seem to be that there have been some somewhat unrepresentative values lately, which I believe Patrick is misinterpreting as some radical change occurring, whereas in reality it's likely just some expected variability. Patrick's formula is in fact fitting the overall relationship in the data going as far back as I have flux data at least, so I assume he's used linear regression on long-term data to come up with it. In other words, altering the coefficient as you are doing here results in a better fit more or less due to what Patrick is saying, namely that lower sunspot numbers now yield higher fluxes than they've done previously; that does indeed appear to be correct, as we shall see below, so altering the coefficient is essential overfitting the formula to match the recent values. What I don't think is correct is to conclude that this represents a huge change that heralds historical lows.

First off, let's look at a scatterplot of all the data going back to ~1952, using a 365-day smoothing for both the daily SN and the F10.7, including the line of best fit using linear regression:

First thing to note is that the fit is very close to the formula that was already presented, ~0.65x + ~63 in this case. Presumably the difference is because the formula was fitted earlier, and that the recent values have also deviated from the norm, pushing the coefficient upward (closer to the 0.75 you just briefly fitted).

Second thing to note is how there's actually quite a bit of variation on both sides of the line; sometimes the flux is lower than you'd expect from it, sometimes higher (as it is now).

Now, to quickly confirm that many of the recent values are indeed among the anomalous ones, let's remove the last ~11 years from the data, essentially removing the entire interval in the plot you posted (from Solen):

As is readily seen, all those anomalous values on the lower left have disappeared (note that I haven't fitted this again, that's the same as previously). So it is correct that for the last ~11 years or so, the flux values have been higher than predicted from the overall relationship so far.

To see more clearly where the anomalies lie, let's plot the actual flux against the predicted one:

Here we see clearly what this is all about, i.e. the higher fluxes relative to the predicted ones, especially near the bottom (hence the values near the bottom left in the scatterplot); some higher values for higher SNs too, but that is clearly nothing new as we can see here (and from the upper right part of the scatterplot, where there's a higher density of anomalous values on that side of the fit).

Going through the same for the F30, we get this fit instead:

Overall it's still clearly a good fit and only has a slightly lower score, but here we can see noticeably more anomalous values on either side. I believe the reason for this is that the 30 cm flux is more sensitive to various types of Solar emissions that the 10.7 cm flux is less sensitive to, such as those from plages/faculae, and if Patrick is correct that we're currently seeing more smaller sunspots and fewer larger ones, that could definitely be part of the explanation. Plotting the actual against the predicted here looks like this:

What's interesting here is that we can see more of it in the later cycles than above, and conversely also more below the earlier cycles. This could possibly mean that it's rather some longer-scale variation in relative sunspot size, as it almost seems to cross over around the second half of SC21 (between ~1980 and ~1990); but that's speculative and something we'll get a better picture of in the future with more and better data, especially the high-resolution data, which if it actually continues to follow the flux closely (as I believe Patrick claims) would indicate to me that there's not actually anything radical going on.

All in all I still see no reason why this would be sensational. Interesting, sure, but nothing revolutionary as far as I can tell.

Also, I still really struggle to see how what we're seeing should be construed as lower activity; to me it almost seems like the opposite is the case, we're essentially getting a higher flux than we'd expect, while the SN is developing more or less as expected. Sounds great if you ask me.

Edited May 7, 2023 by Philalethes
axis label correction

[Patrick P.A. Geryl]

7 uren geleden, Philalethes zei:

Well, I outlined some data about the fluxes above, but let's look at some. First of all, let's look at a pitfall that would be easy to fall into. Let's look at the minima of the 10.7 fluxes of the recorded cycles using a 365-day smoothing:

Notice something here? The roughness at the troughs is clearly interfering with an accurate placement of the minima, so if we were to use this to compare the cycles in terms of flux we'd end up with mismatches. Note particularly how the last minimum is placed a small bump that's clearly much further left than most of the other minima (and conversely how the minimum between 1980 and 1990 is placed much further to the right in the same manner). If we were to compare the cycles like this, we would expect to see SC25 getting displaced to the right and SC22 getting displaced to the left. Let's see:

Lo and behold, exactly as expected: there seems to be a clear relationship between rise time and amplitude, but SC25 here instead starts out with a long flat stretch (due to the displaced minimum) and then suddenly surges through SC24; and conversely, SC22 somehow ends up slightly to the left of the mighty SC19, rising unexpectedly fast due the minimum displaced in the other direction.

To fix this, I would suggest still using a 365-day smoothing for the values we're looking at, so we can see the detail better, but to use a much longer smoothing period to get a better approximation of the actual large-scale troughs. Let's see what the first plot looks like with 5-year smoothing instead:

Much better. There are still hints of the same irregularities, but they're not nearly as pronounced now. Using these dates for the minima we should expect to get a much more sensible plot of the different cycles. It looks like this (note how SC19 falls away due to the data not going far enough back to include its preceding minimum when smoothing it this way):

That looks much more sensible to me. Here the relative rise times and amplitudes seem to be much more coherent. What that holds for the future is of course uncertain, because it could flat out like SC20, or it could keep rising like some of the other cycles (whether that violates the general relationship or not depends on whether there's still some displacement of the minimum or not, but this is a small sample size anyway, for all we know cycles can full well violate this relationship). What is at least clear is that we're already above SC24, so it's not possible for this to be the weakest cycle since the Dalton minimum by virtue of that alone.

.

You write we are already above cycle 24….

Nope…

as endlessly explained you use wrong flux data. Look at the graphs of Solen…😳🤨🤔

[Philalethes]

56 minutes ago, Patrick P.A. Geryl said:

as endlessly explained you use wrong flux data. Look at the graphs of Solen…

As was explained to you previously, the only data you're claiming to be the "right" data does not go back long enough to establish any significant relationships. As you can clearly tell from the above, variability in the relationship between SN and flux is nothing new. I don't even see why you'd call it the "right" data at all when it dates back that shortly, because for all you know a few more cycles will come and completely mess your newfound relationship up, leaving you to declare that it is now the "wrong" data, and that your most recent fit is the "right" data instead.

As for the data I'm using, it's the raw CLS data, not adjusted for anything, and I don't really see any reason to adjust for something like flares either given how that's an inherent part of the Solar activity you're measuring; and additionally, the linear fit is virtually the same as your formula for the ISN anyway.

[Patrick P.A. Geryl]

1 uur geleden, Philalethes zei:

 

As for the data I'm using, it's the raw CLS data, not adjusted for anything, and I don't really see any reason to adjust for something like flares either given how that's an inherent part of the Solar activity you're measuring; and additionally, the linear fit is virtually the same as your formula for the ISN anyway.

Warning for all astronomers

As explained in my article you need to use flare compensation.

Jan and I spend a lot of time to find the right values…

As an example there is a more than 4 percent fault in the flux data from cycle 25! See the graph above… 

By not using flare compensation… the above graphs and conclusions are null and void.

So mine are still standing🧐🤩

 

 

 

Edited May 7, 2023 by Patrick P.A. Geryl

[Jesterface23]

1 hour ago, Patrick P.A. Geryl said:

So mine are still standing🧐🤩

At least a little less than 1/6 of your graph, the most recent, would be by your current standards to be fair.

[dave]

The visible sunspots seem close to the Equator. Is 'the butterfly' closing in rather early? Or is it just a chance thing? 

 

[Philalethes]

7 hours ago, Patrick P.A. Geryl said:

By not using flare compensation… the above graphs and conclusions are null and void.

I'm not sure that's actually reasonable at all, because flaring is an inherent part of Solar activity that it makes little sense to me to remove. Adjusting for Earth-Sol distance makes more sense, but this annual variability would be unlikely to affect the above relationships much, since we're talking about anomalies occurring over longer periods of time.

Anyway, let's have it your way for the sake of discussion and use the flare-corrected data; but let's just get this clear: what exactly is the procedure you're applying to "correct" the flux? Are you using the adjusted data you were using in this article uploaded in 2020, i.e. this data, or are you using something else? When you say you started using the "right" data in 2012, do you mean that you switched over to that data source, or that you started making your own "adjustments" to the data? Because if you're still using historical data, then at least we can still compare. Also, if you're using that above data, are you using the data that has been corrected for distance too, or only for flaring and other factors?

I'm just asking to make sure we're using the exact same data, because I already have all the corrected and adjusted data downloaded too, I just need to know which ones you are using.

Edit:

Actually, let's just assume you're using historical distance-adjusted and flare-corrected data for the sake of the discussion. That data source uses the data from Penticton, which is what I already have downloaded (the F10.7 of CLS is taken from there), so let's use that instead and see what we find:

As you can see, this is virtually identical, and the line of best fit remains practically unchanged. In fact, plotting it we find pretty much the exact same plot as above:

So no, the problem here has nothing to do with correction for flares, because (as I find more reasonable) only significant flaring activity is corrected for, whereas on the vast majority of days there's no correction at all, and as I suspected the adjustment for distance doesn't really affect the long-term trends over many years. Take a look at this list of all dates where there's a difference between the flare-corrected value and the uncorrected one; you can clearly see that there are almost no dates where this is happening overall, fewer than 100 days out of over 26,000 days in total, they simply don't affect the overall development noticeably at all, especially not when smoothing it over 365 days.

The reason why the plot from Solen still doesn't show that SC25 has surpassed SC24 I believe is because it has still yet to incorporate the last couple of weeks of fluxes, which are all very high; in fact, the plot above is cut off at original data limited to the end of April, so even there the very last week of fluxes is not included, and that week is even higher, with an average of 158.9. So once Solen gets around to plotting the latest fluxes, you will see the SC25 flux above the peak of SC24, invalidating your Dalton minimum claim.

Also, all of this shows that all the data I've provided above is far from "null and void", but perfectly in line with the adjusted and corrected data, since the differences are not discernible at all.

Edited May 7, 2023 by Philalethes

[Patrick P.A. Geryl]

1 uur geleden, dave zei:

The visible sunspots seem close to the Equator. Is 'the butterfly' closing in rather early? Or is it just a chance thing? 

 

You make my day🤩🤔

As I found… May 2023 is max smoothed month for SSN.

[3gMike]

4 hours ago, Philalethes said:

I'm not sure that's actually reasonable at all, because flaring is an inherent part of Solar activity that it makes little sense to me to remove. Adjusting for Earth-Sol distance makes more sense, but this annual variability would be unlikely to affect the above relationships much, since we're talking about anomalies occurring over longer periods of time.

Anyway, let's have it your way for the sake of discussion and use the flare-corrected data; but let's just get this clear: what exactly is the procedure you're applying to "correct" the flux? Are you using the adjusted data you were using in this article uploaded in 2020, i.e. this data, or are you using something else? When you say you started using the "right" data in 2012, do you mean that you switched over to that data source, or that you started making your own "adjustments" to the data? Because if you're still using historical data, then at least we can still compare. Also, if you're using that above data, are you using the data that has been corrected for distance too, or only for flaring and other factors?

I'm just asking to make sure we're using the exact same data, because I already have all the corrected and adjusted data downloaded too, I just need to know which ones you are using.

Edit:

Actually, let's just assume you're using historical distance-adjusted and flare-corrected data for the sake of the discussion. That data source uses the data from Penticton, which is what I already have downloaded (the F10.7 of CLS is taken from there), so let's use that instead and see what we find:

As you can see, this is virtually identical, and the line of best fit remains practically unchanged. In fact, plotting it we find pretty much the exact same plot as above:

So no, the problem here has nothing to do with correction for flares, because (as I find more reasonable) only significant flaring activity is corrected for, whereas on the vast majority of days there's no correction at all, and as I suspected the adjustment for distance doesn't really affect the long-term trends over many years. Take a look at this list of all dates where there's a difference between the flare-corrected value and the uncorrected one; you can clearly see that there are almost no dates where this is happening overall, fewer than 100 days out of over 26,000 days in total, they simply don't affect the overall development noticeably at all, especially not when smoothing it over 365 days.

The reason why the plot from Solen still doesn't show that SC25 has surpassed SC24 I believe is because it has still yet to incorporate the last couple of weeks of fluxes, which are all very high; in fact, the plot above is cut off at original data limited to the end of April, so even there the very last week of fluxes is not included, and that week is even higher, with an average of 158.9. So once Solen gets around to plotting the latest fluxes, you will see the SC25 flux above the peak of SC24, invalidating your Dalton minimum claim.

Also, all of this shows that all the data I've provided above is far from "null and void", but perfectly in line with the adjusted and corrected data, since the differences are not discernible at all.

Thanks for all the effort you have put into making this clear.

Given the data presented in Solen daily reports I would also question the value of referring to the 1K and 2K data, let alone consider them to represent a 'gold standard'. These additional spots are generally reported as being quiet and stable, or as having tiny spots. Whilst they make some (probably small) contribution to 10.7cm flux they are rarely, if ever, reported as a source of flares.

4 hours ago, Philalethes said:

So once Solen gets around to plotting the latest fluxes, you will see the SC25 flux above the peak of SC24, invalidating your Dalton minimum claim.

Their plot will not be updated for a little while yet, because they use a centred 81 day average for flux. Even so, looking back to 2014 they recorded a peak for Cycle 24 of 154, and have already recorded a peak of 168 for January this year.

[Philalethes]

12 minutes ago, 3gMike said:

Their plot will not be updated for a little while yet, because they use a centred 81 day average for flux. Even so, looking back to 2014 they recorded a peak for Cycle 24 of 154, and have already recorded a peak of 168 for January this year.

Hmm, I'm not quite sure I understand. As far as I can see they only use a 81-day average for the first plot, the one on top of their page, whereas they're using a 365-day average for the one in question, the one at the bottom of the page; but in either case I don't see why they wouldn't update it daily as long as flux data keeps coming in each day. At least with a 81-day average I'm arriving at the same final day plotted as they are (27/03/2023); however, the peak I'm registering is higher than the one of ~168 on their page:

Not sure what the difference stems from; at least I've double-checked that I'm indeed using the distance-adjusted and flare-corrected data, so I can only assume we still might be using different datasets, but if that's the case it beats me what exactly they're using for the data.