Leif Svalgaard writes in comments:
We plan to submit tomorrow to JGR the following…
http://www.leif.org/research/IDV09.pdf (preprint)
…showing the run of the heliospheric magnetic field since 1835 [not a typo]. I plan to discuss the whole peer-review process here on WUWT, complete with nasty comments by the reviewers and our responses. This will be an illustration of the peer-review process as it unfolds. Should be interesting.
I’ll say. I’ve taken some of the most interesting graphics and put them up for WUWT readers, along with the abstract.
IDV09 and Heliospheric Magnetic field 1835-2009
Leif Svalgaard1 and Edward W. Cliver2
Stanford University, HEPL, Cedar Hall, Via Ortega, Stanford, CA 94305-4085
Space Vehicles Directorate, Air Force Research Laboratory, Hanscom AFB, MA 01731-3010
Abstract.
We use recently acquired archival data to substantiate and extend the IDV index of long-term geomagnetic activity, particularly for years from 1872-1902 for which the initial version of the index (IDV05) was based on observations from very few stations. The new IDV series (IDV09) now includes the interval from 1835-2009, vs. 1872-2004 for IDV05. The HMF strength derived from IDV09 agrees closely with that based on IDV05 over the period of overlap. Comparison of the IDV09-based HMF strength with other recent reconstructions of solar wind B yields a strong consensus between the series based on geomagnetic data, but significant lack of support for a series based on the 10Be cosmic ray radionuclide.
The reconstructed data in the graphic below, from the paper, is quite interesting. Currently, we appear to be at the lowest point in the record.
Click for larger images.
Here’s the comparison with the Be10 isotope record:
Leif Svalgaard (17:59:25) :
Your all over the place Leif, SC20 has all the data you care to look at, almost all the solar metrics are reduced compared with the cycles around it. Of course the dynamo is still running, the issue is at what level. A grand minimum is going to be substantially lower in those metrics than SC20.
We seem to be talking past each other on the wave graph issue…so lets just drop it. Anyone who cares to look can see the wave in multiple records over the last 400 years. Sure it would be nice to go back further….but we cant. If you think we can, have a go at it yourself and show me.
Carla (18:26:44) :
Welcome to the discussion Carla…you might find it safer here.
A function of nothing other than the relative positions of Jupiter, Uranus, & Neptune:
http://www.sfu.ca/~plv/f_UN_J.png
A function of Sunspot Number:
http://www.sfu.ca/~plv/f_R.png
Suggested: Open in separate tabs & blink between them.
To be clear: The first plot is not a sine wave fit. It is based on nothing other than NASA Horizons output for J, U, & N.
[If anyone wonders “why start at 1840?” the answer is simple: I investigate terrestrial polar motion and the records start in the 1840s.]
Paul Vaughan (23:22:10)
In theory, due to Titus-Bode Law, any function relating to the orbits, can be shown to have a major component with an oscillating period in the region of 10-11 years, as vukcevic and Sharp have demonstrated, hence no surprise that the sun cycles are of a similar period.
Leif..
On pages 17&18 of IDV09 where comparison of Observed is used, I started noticing fluxuations between IDV and Obs, bound to be some, no big deal, negligible.
Exceptions 1978, 1989, 2001 seems to have issues?
Was checking out the high ends for the entire period when I noticed.
Geoff Sharp (20:50:05) :
Welcome to the discussion Carla…you might find it safer here.
Thanks Geoff. Did I tell you that my flywheel is a primary and necessary component of my system? None of the lesser systems on the drive train move without it. But the linkage… oh man…now that’s a sticky wicket. lol
fabron (01:35:30) :
major component with an oscillating period in the region of 10-11 years, as vukcevic and Sharp have demonstrated, hence no surprise that the sun cycles are of a similar period.
The surprise is the coincidence that the Sun cycles with a similar period. The ~10-yr period is also found in sun-like stars, so seems to be set by properties of the star [Sun] rather than the planets [unless you postulate the Sun-like stars also have Sun-like planetary systems, which is possible].
Carla (05:45:30) :
Exceptions 1978, 1989, 2001 seems to have issues?
Near solar maximum when the HMF is higher than normal, the deviations will also be larger. Crudely, we expect the error to be a proportion, say 5%, of the total, so 5% of 9 nT = 0.45 is naturally higher than 5% of 4 nT = 0.20. This is one part of the answer. Another part is that we expect missing data [spacecraft coverage] to increase the error in the observed data. If one plots the difference between inferred HMF and HMF observed, one finds that the difference increases as the proportion of missing data [which can be as high as 60-70% in some years] go up. And 1978 and 1989 has much missing data. 2001 is an exception and may be a clue [that needs more investigation] as to why there is this difference. If you look at the lower plot on page 3 of http://www.leif.org/research/Most%20Recent%20IMF,%20SW,%20and%20Solar%20Data.pdf you’ll note that the standard deviation of observed B [the orange curve] seems to have the same ‘hump’ in 2001 as B from IDV, so perhaps there is a clue in the variability of B. But all this is nitty-gritty stuff, that eventually will be resolved, that does not change the bigger picture.
Paul Vaughan (23:22:10) :
A function of nothing other than the relative positions of Jupiter, Uranus, & Neptune: http://www.sfu.ca/~plv/f_UN_J.png
Trivial sleight of hand, of course. Jupiter’s period is not quite right, but inclusion of other ones fixes that. To make it a bit [but only a bit] more meaningful, Saturn should have been in the mix.
Paul Vaughan (23:22:10) :
A function of Sunspot Number: http://www.sfu.ca/~plv/f_R.png
Not to speak of the semi-fictitious ‘function’.
If one considers the set {J,S,U,N,JS,JU,JN,SU,SN,UN} the lowest & highest frequency elements are UN & J.
Paul Vaughan (17:39:50) :
If one considers the set {J,S,U,N,JS,JU,JN,SU,SN,UN} the lowest & highest frequency elements are UN & J.
Interesting correlation Paul, have you tried graphing them together and going back to 1700 to see if the match continues?
Paul Vaughan (17:39:50) :
If one considers the set {J,S,U,N,JS,JU,JN,SU,SN,UN} the lowest & highest frequency elements are UN & J. And?
What are the two functions plotted?
Geoff Sharp (20:50:05) :
Sure it would be nice to go back further….but we cant. If you think we can, have a go at it yourself and show me.
Yes we can. For this purpose the 10-yr values from the 14C series are sufficient, and they do not show neither a 107-yr cycle no a 172- yr cycle, but their power spectrum has a broad peak around 88 years, so that settles that, and, as you say, perhaps it is time to drop that settled matter.
Paul Vaughan (17:39:50) :
Cant quite understand the sunspot plot….how was it derived?
I’m still working on this, but here’s something to consider:
http://www.sfu.ca/~plv/DRAFT_VaughanPL2009CO_TPM_SSD_LNC.htm
Geoff Sharp (18:29:24) “Interesting correlation Paul, have you tried graphing them together and going back to 1700 to see if the match continues?”
This was just something that suddenly dawned upon me late at night after a long day – so I just quickly used the data I had on-hand, generating the graphs within a few minutes. I’ve been swamped with my other research, but eventually I’ll get around to obtaining a 3000BC-3000AD series of coordinates (which I need for other work) from Horizons.
It may be interesting to use the approach to explore an alternate to Desmoulins’ – and compare all 3 series using cross-wavelet phase-differencing, to see the phase-lag patterns. If there’s just phase-drift, it will be evident quite quickly; on the other hand, if there is sustained loose-resonance, that will also be apparent quite quickly.
So many interesting things to investigate. What I really want to do is keep working on precipitation, as I see it as the weak link in human understanding of climate.
Geoff Sharp (21:25:16) “Cant quite understand the sunspot plot….how was it derived?”
Perhaps I’ll check my notes sometime (I grabbed the plot quickly from a file full of columns), but my intuition / memory is:
(Log_2(R+1))'[11.1a/2] where the ‘ indicates rate-of-change (i.e. differencing), the square brackets indicate integration over a sliding-window, “_2” indicates base 2 (but any other base you like for interpretive purposes will do), & “a” (annum) indicates years.
This is just a standard transform based on the “ladder of powers” covered early in Stat 101, combined with some lessons I’ve learned, about the effects of integrating over harmonics of cycles, from programming & running time-integrated cross-correlation analyses during my M.Sc. education/research.
It doesn’t actually matter what function one uses, as all the transform accomplishes is a scatter (heteroscedasticity) adjustment (via 1:1 transform) & a phase-shift; it doesn’t change the period. Thus, if I used any of the other columns from the file from which I grabbed the plot, I would just have found myself needing to use a different column from the other (UN,J) file – in which I had sines, cosines, -sines, & -cosines of the angles with which I was working …which cover phase-shifts for all multiples of quarter-cycles …which is exactly what you get when you difference & integrate strategically over harmonics. In the absolute worst-case scenario I’d find myself using a double- or half-angle theorem (from basic trigonometry) – and the curves would still phase-match loosely.
If this doesn’t make sense, just use plain sunspot numbers and slide your curve in steps of 11.1a/4 over the red graph in an image editor (after scaling 1840-2010 to unit length via stretching/squeezing).
Paul Vaughan (22:38:40) :
Your in a different league to me Paul, but interested to see what your research produces, The Desmoulins test is in place this cycle, will the lag behind catch up? I see he is still updating his website so it will be good to watch. But whoever finally works out the 11 yr cycle will have bronze statues erected in their honor.
I had trouble matching up the sunspot records on your graph, the peaks and troughs didnt seem to line up with the actual count….is there a reason?
something i will learn about. wow, is an understatement. luv the blog
Leif Svalgaard
How does the next installment work Leif, what is the normal time period for the reviewers to respond. Looking forward to seeing their nasty comments 😉
Geoff Sharp (08:19:06) :
what is the normal time period for the reviewers to respond. Looking forward to seeing their nasty comments 😉
This is the status of the paper as I write:
Waiting for Reviewer Assignment 2009-10-08 13:38:40
Initial Quality Control Complete 2009-10-08 13:38:40
Initial Quality Control Started 2009-10-07 23:04:41
Author Approved Converted Files 2009-10-07 23:04:40
Preliminary Manuscript Data Submitted 2009-10-07 21:59:05
What is happening is that a request has gone out to a number of reviewers [at the editor’s discretion] to see if they are willing and able to review the paper. This stage can be either short [a week perhaps] or long [a month or more] depending on the reviewer. If it takes too long, the editor will assigned another reviewer.
The author is encouraged to submit a list [minimum 5 items] of reviewers that are knowledgeable about the subject. The editor is not bound to select from that list and often does not. Another list can be gotten from the list of references in the paper. It is my experience that the selection usually is ‘fair and balanced’ for reputable journals [and JGR is].
Leif Svalgaard (08:55:43) :
Is there a submission fee?….and do the reviewers receive remuneration?
Geoff Sharp (08:19:06) :
Looking forward to seeing their nasty comments 😉
After a reviewer has accepted, he usually has a month to respond. In some cases he drags it out [for perhaps personal reasons]. The worst case I personally have endured took six months.
The innocently sounding items:
Initial Quality Control Complete 2009-10-08 13:38:40
Initial Quality Control Started 2009-10-07 23:04:41
cover a nasty reality. In the distant past, the paper was often handwritten or typed and was typeset from the submitted text. Then about 30 years ago, the journals thought they could save money by requiring the authors to submit ‘camera-ready’ manuscripts that did not require typesetting and could be reproduced in the journal at no cost to the publisher. This brought about an unevenness in style and ‘print’ quality of the papers [scientists may be lousy typesetters, an example: http://www.leif.org/research/Using%20Dynamo%20Theory%20to%20Predict%20Solar%20Cycle%2021.pdf ].
But, the publishers discovered another [business] problem. They want to assert copyright over the paper, but if it is ‘camera-ready’ there is no ‘creative’ element involved in their publication, so the publisher cannot claim copyright. Mainly for this reason, they went back to the old way of doing it. To make it clear that the publisher is doing ‘creative’ work [necessary for copyright], the submitted version has to be in a [virtually] unreadable style [see below], and the publisher then turns the paper into a readable ‘work’, by arranging the pieces such that the paper can be read. The ‘initial quality control’ is a check that the paper meets the standard of unreadability required by the journal.
Here is our submitted version:
http://www.leif.org/research/IDV09-Submitted.pdf
It is made hard to read, by the requirement that the text and the Figures, and [most distractingly] the Figure captions be separated into different sections of the paper. This means that the reader does not see where the Figure goes in the text, and cannot interpret the Figure without having to hunt down its caption somewhere else. All in all, it is an enormous pain and extra burden on the reviewer, but the publisher can now claim copyright because of the ‘creative’ work of making the paper readable by putting things together where they belong.
Geoff Sharp (09:13:46) :
Is there a submission fee?….and do the reviewers receive remuneration?
No to both.
There might be a publication fee [‘page charge’] when the paper is finally published. for my last paper in JGR, it came to US$ 11,000. Admittedly, that was a large paper with many [expensive] color Figures, and a more typical charge would be $3,000.
The issue with unpaid reviewers is an interesting one. If there is money involved, there might be issues with impartiality [who pays for what?]. Most reviewers do it because there is a certain ‘honor’ in being selected as a reviewer, or at least recognition that your peers regard you as an ‘expert’ on the subject. Here is a note from the editor from our IDV05 paper: “[26] Arthur Richmond thanks Joseph King, Kalevi Mursula, and Ian G. Richardson for their assistance in evaluating this paper.” Interestingly there were three reviewers [normally only two]. This can happen for several reason, mostly because there was dissent, but also [and we choose to believe that this was the case here 🙂 ] in case the editor felt that this was an important paper, and therefore should be submitted to extra scrutiny.
Leif Svalgaard (09:49:38) :
This is a real learning curve, I am blown away. I cant believe the costs incurred by your last paper. This is not how science should work.