Sunspot Magnetism---Livingston & Penn

[clive]

Mar 21, 2010 15:29:43 GMT lsvalgaard said:

Mar 21, 2010 5:08:23 GMT clive said:

The method of data collection is dubious. 2 sunspot groups were measured over Jan/Feb when we have experienced quite high activity compared to last year. The major active regions have not been measured (1040,1045 etc). The last measurement was taken during speck activity.

The method is unbiased. It works like this:
Bill has fixed slots with telescope time decided months ago. If the weather permits, he observes all spots visible, specks or monsters alike during his slot. So the method is not dubious. With increasing activity and better weather [as we move towards summer] we should get a fuller measure.

The method is unbiased, but limited by the available telescope time. If the major regions of Jan/Feb were measured we may likely see an upswing which goes against their trend. The major regions are not included in the study.

The peer review process would not allow data collection on an ad hoc basis.

How this graph is ending is more important than the start.

[lsvalgaard]

Mar 21, 2010 23:26:23 GMT clive said:

Mar 21, 2010 15:29:43 GMT lsvalgaard said:

The method is unbiased. It works like this:
Bill has fixed slots with telescope time decided months ago. If the weather permits, he observes all spots visible, specks or monsters alike during his slot. So the method is not dubious. With increasing activity and better weather [as we move towards summer] we should get a fuller measure.

The method is unbiased, but limited by the available telescope time. If the major regions of Jan/Feb were measured we may likely see an upswing which goes against their trend. The major regions are not included in the study.

The peer review process would not allow data collection on an ad hoc basis.

How this graph is ending is more important than the start.

The unbiased method used is just perfect for this. It is as little ad-hoc as can be.
And we'll know in a few years time, so what's the rush?

[clive]

Mar 22, 2010 1:24:01 GMT lsvalgaard said:

Mar 21, 2010 23:26:23 GMT clive said:

The method is unbiased, but limited by the available telescope time. If the major regions of Jan/Feb were measured we may likely see an upswing which goes against their trend. The major regions are not included in the study.

The peer review process would not allow data collection on an ad hoc basis.

How this graph is ending is more important than the start.

The unbiased method used is just perfect for this. It is as little ad-hoc as can be.
And we'll know in a few years time, so what's the rush?

Disagree, a perfect method would be to measure every spot. They cant do that, so through no fault of the researchers their study is questionable. The results for this year could be very different (and probably are).

Waiting might not solve the problem, it is possible that most of the high future readings could be missed with the ad hoc problem that is part of this study, especially if this cycle has reduced activity as expected. The study in my opinion because of the inability to measure all of the data is not of a high scientific standard.

[lsvalgaard]

Mar 22, 2010 2:50:04 GMT clive said:

Mar 22, 2010 1:24:01 GMT lsvalgaard said:

The unbiased method used is just perfect for this. It is as little ad-hoc as can be.
And we'll know in a few years time, so what's the rush?

Disagree, a perfect method would be to measure every spot. They cant do that, so through no fault of the researchers their study is questionable. The results for this year could be very different (and probably are).

Waiting might not solve the problem, it is possible that most of the high future readings could be missed with the ad hoc problem that is part of this study, especially if this cycle has reduced activity as expected. The study in my opinion because of the inability to measure all of the data is not of a high scientific standard.

You misunderstand the nature of measurements. It would be nice to measure every single spot, but it is not possible. How about the spots that form and die on the backside of the Sun and we never see? The Sun already makes a selection for us by only presenting us with half the spots. Weather and telescope slot narrow that choice, but does not invalidate the method, as long as the decision about when to observe is not determined by the data [i.e. the spots]. So, the high standards are certainly met.

[clive]

Mar 22, 2010 4:14:23 GMT lsvalgaard said:

Mar 22, 2010 2:50:04 GMT clive said:

Disagree, a perfect method would be to measure every spot. They cant do that, so through no fault of the researchers their study is questionable. The results for this year could be very different (and probably are).

Waiting might not solve the problem, it is possible that most of the high future readings could be missed with the ad hoc problem that is part of this study, especially if this cycle has reduced activity as expected. The study in my opinion because of the inability to measure all of the data is not of a high scientific standard.

You misunderstand the nature of measurements. It would be nice to measure every single spot, but it is not possible. How about the spots that form and die on the backside of the Sun and we never see? The Sun already makes a selection for us by only presenting us with half the spots. Weather and telescope slot narrow that choice, but does not invalidate the method, as long as the decision about when to observe is not determined by the data [i.e. the spots]. So, the high standards are certainly met.

We will just have to agree to disagree. Its a bit like doing a retail sales report for December but leaving out the last 3 weeks.

[grian]

Mar 22, 2010 5:08:25 GMT clive said:

Mar 22, 2010 4:14:23 GMT lsvalgaard said:

You misunderstand the nature of measurements. It would be nice to measure every single spot, but it is not possible. How about the spots that form and die on the backside of the Sun and we never see? The Sun already makes a selection for us by only presenting us with half the spots. Weather and telescope slot narrow that choice, but does not invalidate the method, as long as the decision about when to observe is not determined by the data [i.e. the spots]. So, the high standards are certainly met.

We will just have to agree to disagree. Its a bit like doing a retail sales report for December but leaving out the last 3 weeks.

Then you misunderstand the nature of sampling and statistics if you believe that's what L&P are doing.

[clive]

Mar 22, 2010 10:07:51 GMT grian said:

Mar 22, 2010 5:08:25 GMT clive said:

We will just have to agree to disagree. Its a bit like doing a retail sales report for December but leaving out the last 3 weeks.

Then you misunderstand the nature of sampling and statistics if you believe that's what L&P are doing.

No misunderstanding. Its simple science, you cant report on a trend by taking 2 sunspot samples over 2 months. The first sample in Jan is 1039 which is a reasonable group but nothing near some of the northern hemisphere groups. Bypass all the heavy groups like 1040 and 1045 and a few others that were strong then move onto to the end of Feb where we have speck activity and take another sample.

If you think that is good sampling that reflects the trend accurately then perhaps I wont be looking at anything you might produce. L&P are hamstrung through conditions outside of their control, but that doesnt mean their results represent what's actually happening.

[jinki]

I agree with clive.

Good science is built on good data. I dont think we have good data in this case, it could be done better.

[csspider57]

One thing we all seem to agree on is this. We all would like to see Mr. Bill have more scope times.

[sranderson]

Since things such as the cancellation of observations due to weather, and the scope time schedule, are unrelated to what the sun is doing, the sampling is random and good. The only question then is if there are enough measurements/samples being performed. This is particularly important since the raw data is highly variable. By looking at the chart, it can be seen that the yearly averages lie quite close to the trend line. This is an indication that enough observations are being performed.

Over short periods of time (months), the observations are not statistically significant anyway (unless they happen to be way out of bounds). It doesn't matter if L&P catch large spots or small spots as long as the sampling is random. Just don't make conclusions based on a month or two of data.

By the end of this year, I would think that it should be obvious whether their observed trend is continuing despite the upturn we are seeing as SC24 takes off.

While it would seem to be a good idea to give them additional time during large spot outbreaks, this actually would corrupt the random sampling process that they have used for the previous data. Keep on going on is best, unless you want to devote a whole telescope to them (along with a larger staff).

[lsvalgaard]

Mar 22, 2010 14:45:03 GMT sranderson said:

Since things such as the cancellation of observations due to weather, and the scope time schedule, are unrelated to what the sun is doing, the sampling is random and good. The only question then is if there are enough measurements/samples being performed. This is particularly important since the raw data is highly variable. By looking at the chart, it can be seen that the yearly averages lie quite close to the trend line. This is an indication that enough observations are being performed.

Over short periods of time (months), the observations are not statistically significant anyway (unless they happen to be way out of bounds). It doesn't matter if L&P catch large spots or small spots as long as the sampling is random. Just don't make conclusions based on a month or two of data.

By the end of this year, I would think that it should be obvious whether their observed trend is continuing despite the upturn we are seeing as SC24 takes off.

While it would seem to be a good idea to give them additional time during large spot outbreaks, this actually would corrupt the random sampling process that they have used for the previous data. Keep on going on is best, unless you want to devote a whole telescope to them (along with a larger staff).

Yes, one should not change the sampling procedure. There will [hopefully] soon be more spots as we approach maximum to allow a good average to be computed.

[clive]

Looking at Leif's graph the sampling is all over the place, maybe if they solicited some help from other observers and performed a thorough analysis, their work might have a greater chance of passing peer review?

Until that happens I personally will not have confidence in the results.

[sranderson]

I guess I don't understand what you mean by "the sampling is all over the place". If you mean that there are significant gaps in time where no measurements were made, well, a lot of that is due to the fact that over the last 3 years there were significant amounts of time where there were no sunspots to measure. Add telescope scheduling and weather cancellations and you have the gaps (since 2001 or so). If you mean that the data appears "noisy" with samples spread out quite a ways out from the trend lines, well , L&P are just measuring the way the sun behaves -- and these particular parameters vary quite a bit. Regardless, there are thousands of data points that taken together have a clear trend -- so far.

It is a mystery why the trend exists however. I have not heard of any model, or even a general idea of why, except to say that the sun sometimes quiets down and sometimes is more active.

The interesting thing, if the trend continues, is that it provides insight into how a Grand Minimum may happen: Not just less sunspots occur, but rather a universal fading away of sunspots that may suggest a weakening of the entire underlying magnetic process within sunspots.

[waynelim]

Mar 23, 2010 5:18:09 GMT sranderson said:

I guess I don't understand what you mean by "the sampling is all over the place". If you mean that there are significant gaps in time where no measurements were made, well, a lot of that is due to the fact that over the last 3 years there were significant amounts of time where there were no sunspots to measure. Add telescope scheduling and weather cancellations and you have the gaps (since 2001 or so). If you mean that the data appears "noisy" with samples spread out quite a ways out from the trend lines, well , L&P are just measuring the way the sun behaves -- and these particular parameters vary quite a bit. Regardless, there are thousands of data points that taken together have a clear trend -- so far.

It is a mystery why the trend exists however. I have not heard of any model, or even a general idea of why, except to say that the sun sometimes quiets down and sometimes is more active.

The interesting thing, if the trend continues, is that it provides insight into how a Grand Minimum may happen: Not just less sunspots occur, but rather a universal fading away of sunspots that may suggest a weakening of the entire underlying magnetic process within sunspots.

What is the speculation as to correlation with the recent articles about the speed of the meridional circulation, like here:

sohowww.nascom.nasa.gov/hotshots/2010_03_15/

Is the speed of circulation the cause of the sunspot contrast decreasing, or are they both facets of some deeper, underlying mechanism?

[grian]

Mar 22, 2010 10:43:21 GMT clive said:

If you think that is good sampling that reflects the trend accurately then perhaps I wont be looking at anything you might produce. L&P are hamstrung through conditions outside of their control, but that doesnt mean their results represent what's actually happening.

It's random sampling, which is good - unlike the sales for December example you gave which was an excellent demonstration of cherry-picking data.

And while it would be nice to get data from 100% of the sunspot population it's just not practical.

To draw an analogy, how many industrial processes do you know of that use 100% sampling for quality assurance? Very few I suspect, most use random sampling methods and good statistical process control. In some ways this is no different.