Nicola Scafetta Comments on “Solar Trends And Global Warming” by Benestad and Schmidt
From Climate Science — Roger Pielke Sr.
On July 22 2009 I posted on the new paper on solar forcing by Lean and Rind 2009 (see). In that post, I also referred to the Benestad and Schmidt 2009 paper on solar forcing which has a conclusion at variance to that in the Lean and Rind paper.
After the publication of my post, Nicole Scafetta asked if he could present a comment (as a guest weblog) on the Benestad and Schmidt paper on my website, since it will take several months for his comment to make it through the review process. In the interests of presenting the perspectives on the issue of solar climate forcing, Nicola’s post appears below. I also invite Benestad and Schmidt to write responses to the Scaftta contribution which I would be glad to post on my website.
GUEST WEBLOG BY NICOLA SCAFETTA
Benestad and Schmidt have recently published a paper in JGR. (Benestad, R. E., and G. A. Schmidt (2009), Solar trends and global warming, J. Geophys. Res., 114, D14101, doi:10.1029/2008JD011639).
This paper criticizes the mathematical algorithms of several papers that claim that the temperature data show a significant solar signature. They conclude that such algorithms are “nonrobust” and conclude that
“the most likely contribution from solar forcing a global warming is 7 ± 1% for the 20th century and is negligible for warming since 1980.”
By using the word “robust” and its derivates for 18 times, Benestad and Schmidt claim to disprove two categories of papers:
those that use the multilinear regression analysis [Lean and Rind, 2008; Camp and Tung, 2007; Ingram, 2006] and those that present an alternative approach [Scafetta and West, 2005, 2006a, 2006b, 2007, 2008]. (See the references in their paper.)
Herein, I will not discuss the limitation of the multilinear regression analysis nor the limits of Benestad and Schmidt’s critique to those papers. I will briefly focus on Benestad and Schmidt’s criticism to the papers that I coauthored with Dr. West. I found Benestad and Schmidt’s claims to be extremely misleading and full of gratuitous criticism due to poor reading and understanding of the data analysis that was accomplished in our works.
Let us see some of these misleading statements and errors starting with the less serious one and ending with the most serious one:
1. Since the abstract Benestad and Schmidt claim that they are rebutting several our papers [Scafetta and West, 2005, 2006a, 2006b, 2007, 2008]. Already the abstract is misleading. Indeed, their criticism focuses only on Scafetta and West [2005, 2006a]. The other papers used different data and mathematical methodologies.
2. Benestad and Schmidt claim that we have not disclosed nor detailed the mathematical methodology and some parameters that we use. For example:
a) In paragraph 39 Benestad and Schmidt criticize and dismiss my paper with Willson [2009] by claiming that we “did not provide any detailed description of the method used to derive their results, and while they derived a positive minima trend for their composite, it is not clear how a positive minima trend could arise from a combination of the reconstruction of Krivova et al. [2007] and PMOD, when none of these by themselves contained such a trend).” However, the arguments are quite clear in that paper and in the additional figures that we published as supporting material. Moreover, it is not clear to me how Benestad and Schmidt could conclude that our work is wrong if Benestad and Schmidt acknowledge that they have not understood it. Perhaps, they just needed to study it better.
b) In paragraph 41 Benestad and Schmidt claim that: “It is not clear how the lagged values were estimated by Scafetta and West [2006a]“. However, in paragraph 9 of SW06a it is written “we adopt the same time-lags as predicted by Wigley’s [1988, Table 1] model.” So, again, Benestad and Schmidt just needed to study better the paper that they wanted to criticize.
c) In paragraph 48 Benestad and Schmidt claim that: “over the much shorter 1980-2002 period and used a global surface temperature from the Climate Research Unit, 2005 (they did not provide any reference to the data nor did they specify whether they used the combined land-sea data (HadCRUT) or land-only temperatures (CRUTEM).” However, it is evident from our work SW05 that we were referring to the combined land-sea data which is properly referred to as “global surface temperature” without any additional specification (Land or Ocean, North or South). We also indicate the webpage where the data could be downloaded.
d) In paragraph 57 Benestad and Schmidt claim that: “The analysis using Lean [2000] rather than Scafetta and West’s own solar proxy as input is shown as thick black lines.” However, in our paper SW06a it is crystal clear that we too use Lean’s TSI proxy reconstruction. In particular we were using Lean 1995 which is not very different from Lean 2000. Benestad and Schmidt apparently do not know that since 1978 Lean 1995 as well as Lean 2000 do not differ significantly from PMOD because PMOD was build (by altering the published TSI satellite data) by using Lean 1995 and Lean 2000 as guides. Moreover, we also merge the Lean data with ACRIM since 1978 to obtain an alternative scenario, as it is evident in all our papers. The discontinuity problem addressed by Benestad and Schmidt in merging two independent sequences (Lean’s proxy model and the ACRIM) is not an issue because it is not possible to avoid it given the fact that there are no TSI satellite data before 1978.
3. In Paragraphs 48-50 Benestad and Schmidt try to explain one of our presumed major mathematical mistakes. Benestad and Schmidt’s states: “A change of 2*0.92 W/m2 between solar minimum and maximum implies a change in S of 1.84 W/m2 which amounts to 0.13% of S, and is greater than the 0.08% difference between the peak and minimum of solar cycle 21 reported by Willson [1997] and the differences between TSI levels of the solar maxima and minima seen in this study (~1.2 W/m2; Figure 6).” Benestad and Schmidt’s are referring to our estimate of the amplitude of the solar cycle referring to the 11-year modulation that we called A7,sun = 0.92 W/m2 in SW05. Benestad and Schmidt are claiming that our estimate is nor reasonable because in their opinion according to our calculations the change of TSI between solar maximum and solar minimum had to be twice our value A7,sun , so they write 2*0.92=1.84 W/m2, and this would be far too large. However, as it is evident from our paper and in figure 4a in SW05 the value A7,sun refers to the peak-to-trough amplitude of the cycle, so it should not be multiplied by 2, as Benestad and Schmidt misunderstood. This is crystal clear in the factor ½ before the equation f(t)= ½ A sin(2pt) that we are referring to and that Benestad and Schmidt also report in their paragraph 48. It is hard to believe that two prominent scientists such as Benestad and Schmidt do not understand the meaning of a factor ½! So, again, Benestad and Schmidt just needed to think more before writing a study that criticizes ours.
4) Finally, Benestad and Schmidt’s paper is full of misleading claims that they are reproducing our analysis. Indeed, Benestad and Schmidt’s paper is self-contradictory on this crucial issue. In paragraph 85 Benestad and Schmidt claim that they “have repeated the analyses of Scafetta and West, together with a series of sensitivity tests to some of their arbitrary choices.” However, in their paragraph 76 Benestad and Schmidt acknowledge: “In our emulation, we were not able to get exactly the same ratio of amplitudes, due to lack of robustness of the SW06a method and insufficient methods description.” It is quite singular that Benestad and Schmidt claim to have repeated our calculation, at the same time they acknowledge that, indeed, they did not succeed in repeating our calculation and, ironically, they blame us for their failure. It is not easy to find in the scientific literature such kind of tortuous reasoning!
In fact, the reason why Benestad and Schmidt did not succeed in repeating our calculation is because they have misapplied the wavelet decomposition algorithm known as the maximum overlap discrete wavelet transforms (MODWT). This is crystal clear in their figures 4 where it is evident that they applied the MODWT decomposition in a cyclical periodic mode. In other words they are implicitly imposing that the temperature in 2001 is equal to the temperature in 1900, the temperature in 2002 is equal to the temperature in 1901 and so on. This is evident in their figure 4 where the decomposed blue and pink component curves in 2000 just continue in 1900 in an uninterrupted cyclical periodic mode as shown in the figure below which is obtained by plotting their figure 4 side by side with itself:
Any person expert in time series processing can teach Benestad and Schmidt that it is not appropriate to impose a cyclical periodic mode to a non stationary time series such as the temperature or TSI records that present clear upward trends from 1900 to 2000. By applying a cyclical periodic mode Benestad and Schmidt are artificially introducing two large and opposite discontinuities in the records in 1900 and 2000, as the above figure shows in 2000. These large and artificial discontinuities at the two extremes of the time sequence disrupt completely the decomposition and force the algorithm to produce very large cycles in proximity of the two borders, as it is clear in their figure 4. This severe error is responsible for the fact that Benestad and Schmidt find unrealistic values for Z22y and Z11y that significantly differ from ours by a factor of three. In their paragraph 50 they found Z22y = 0.58 K/Wm-2, which is not realistic as they also realize later, while we found Z22y = 0.17 K/Wm-2, which is more realistic.
This same error in data processing also causes the reconstructed solar signature in their figures 5 and 7 to present a descending trend minimum in 2000 while the Sun was approaching one of its largest maxima. Compare their figures 4a (reported above), 5 and 7 with their figure 6 and compare them also with our figure 3 in SW06a and in SW08! See figure below where I compare Benestad and Schmidt’s figures 6 and 7 and show that the results depicted in their Figure 7 are non-physical.
Because of the severe and naïve error in applying the wavelet decomposition, Benestad and Schmidt’s calculations are “robustly” flawed. I cannot but encourage Benestad and Schmidt to carefully study some book about wavelet decomposition such as the excellent work by Percival and Walden [2000] before attempting to use a complex and powerful algorithm such as the Maximum Overlap Discrete Wavelet Transform (MODWT) by just loading a pre-compiled computer R package.
There are several other gratuitous claims and errors in Benestad and Schmidt’s paper. However, the above is sufficient for this fast reply. I just wonder why the referees of that paper did not check Benestad and Schmidt’s numerous misleading statements and errors. It would be sad if the reason is because somebody is mistaking a scientific theory such as the “anthropogenic global warming theory” for an ideology that should be defended at all costs.
Nicola Scafetta, Physics Department, Duke University
tallbloke (06:49:57) :
Any of these three possibilities will affect the value of the ‘crossing’ between cycles.
And how is that different from the ‘underlying trend’ [which would be measured how?]. The cross-over is just another [slightly different] marker of minimum. Page four of http://www.leif.org/research/Most%20Recent%20IMF,%20SW,%20and%20Solar%20Data.pdf shows the last three cross-overs. BTW, the cross-over [as you can see] does not happen at the statistical minimum [gray curve].
Jeff Id (05:48:46) :
The variometer does indeed seem easy to calibrate and is likely trustworthy.
They are very robust devices. You can even use one to open a can of worms. :0)
Leif Svalgaard (08:33:56) :
The HMF is the same now as 108 years ago at the minimum between cycles 13 and 14. Solar activity now is as low as it was then. TSI should then be that same then as now under the assumption that the Sun’s magnetic field is the underlying cause for variations in TSI. This is independent of ACRIM contortions.
The debate over the splicing of the data from different platforms concerns periods of time prior to now. Maybe TSI was the same at solar minimum in 1901, but I don’t think that is particularly relevant to the issue of the ‘adjustment’ of ACRIM data from the 1990’s.
The average B for cycles 17-18-19 is 6.892 nT, and for 20-21-22-23 it is 6.518. There is no upward trend. The central fact is still that HMF and TSI right now are back to 1901-1902 values. Cycle 23 was very much like cycle 13, and cycle 24 looks [so far] to be much like cycle 14. If solar activity and TSI is the main driver of our climate we would expect 1890-1910 temperatures to be similar to 2000-2020 temps. I don’t think this is [or will be] the case. You could argue that all the heat [or cold] is hidden somewhere and will materialize shortly, but the same argument could then be made for cycle 13/14 as SC10-11 were on par with SC21-22 http://www.leif.org/research/Heliospheric-Magnetic-Field-Since-1835.png
And what was the average from 1900-1930 from your graph? The upward trend was from then to now, not from cycle 17.
A great deal of extra heat has been stored in the oceans as the satellite altimetry tells us. Why you continue to pretend this is mysterious is beyond me. My calcs show that this has been building up since 1935 and so of course temperatures won’t be the same now as in 1911 just because the sun is suddenly less active after a long period when it was at it’s most active.
I’m no fan of Mother Earth hypothesis. There is nothing that says that Nature ’strives’ to achieve ‘equitable’ balance. Rather it has to live with the hand it is dealt, and it apparently does that by NOT being sensitive to small external changes.
If you bothered to read and digest the book you’d understand that the system is VERY sensitive to small external changes, but because there are a very large number of feedback mechanisms in the biosphere, these are dealt with in a myriad of ways. Try it, you might learn something worthwhile. Lovelock is not in the least sentimental about ‘mother earth’ in his book. It’s mostly factual stuff about atmosphere and ocean, and the biological processes which make our planet’s atmosphere and ocean very different from what they would be if there was no life on earth.
Why and how has earth’s near surface temperature remained almost constant while the sun has increased it’s output by 30% ?
Read the book and find out.
British soft power and its foreign policy are being recast with a diplomatic strategy to market Britain across the world as a climate change leader.
I believe Zimbabwe is actually the leader. They have done more to curtail their CO2 output than any other nation on earth.
AlanG (12:22:09)
A question for you sir: Have the Brits all gone daft?
tallbloke (09:53:49) :
TI don’t think that is particularly relevant to the issue of the ‘adjustment’ of ACRIM data from the 1990’s.
And the ACRIM data problem is not particularly relevant to the issue of the long-term variation of HMF and TSI. So, to keep introducing this red herring is a waste of bandwidth [peoples’ and networks’]
And what was the average from 1900-1930 from your graph? The upward trend was from then to now, not from cycle 17.
The trend stopped in the 1950s, as we are now back to where we started.
My calcs show that this has been building up since 1935
And why did the oceans not store heat before 1935?
Why and how has earth’s near surface temperature remained almost constant while the sun has increased it’s output by 30% ?
Try CO2 from outgassing. There was hardly any biosphere for the most part of the Earth’s age.
tallbloke (09:53:49) :
My calcs show that this has been building up since 1935
And why did the oceans not store heat before 1935? for example during the very active years 1840-1880, when average HMF B was 6.5 nT.
Angel (22:55:50)
Angel, there’s an involved discussion above between Leif Svalgaard and tallbloke: who are you dismissing, or both?
tallbloke (01:00:38) “[…] perhaps […] solarcyle24.com?”
No, but we can swap notes around here occasionally. Be sure to check out Figure 2 in the Trenberth & Stepaniak article I referenced upthread.
Leif: “Try CO2 from outgassing. There was hardly any biosphere for the most part of the Earth’s age.”
you might need to clarify or correct that.
The problem with peer reviews is that they are typically restricted to within the specialty of the writer. Many papers would be improved if the reviewers included experts in the methodologies used to reach the conclusions like mathematicians and statisticians.
tallbloke (01:00:38) :
Thanks for that Nasif. I’m a graph junkie at the moment. Got any good ones which illustrate the effect?
You’re welcome. Sorry for the delayed response. Regarding your petition, unfortunatelly I’ve not graphed the effect yet, though I think there are some papers on this issue.
Paul Vaughan (15:07:56) :
Be sure to check out Figure 2 in the Trenberth & Stepaniak article I referenced upthread.
I downloaded the pdf and fig 2 is the cross correlation plot of the +- lags of nino 3.4 vs the trans nino index. If I understand correctly, the magnitudes are the celcius difference of SST.
An area relevant to our previous discussion is the 1905-1930 period, a strong positive PDO and according to the graph TNI leading nino 3.4 by a year or so. The period end coincides with the anomaly in the Chandler period you pointed up with your wavelet analysis a while ago.
I’m struggling to ‘get’ what it is you are driving at beyond that so I hope you’ll give us your analysis.
Paul, I note the shifts also tie in with reversals of long term trends in length of day. Good result, thanks.
cba (15:20:29) :
Leif: “Try CO2 from outgassing. There was hardly any biosphere for the most part of the Earth’s age.”
you might need to clarify or correct that.
What he actually needs to do is overcome his prejudice and read Lovelock’s original book to better inform himself. Lovelock’s description of the evolution of the atmospere and oceans raises important questions about the temperature stability of earth in comparison to planets with no biosphere. Lovelock understood the vast power of the biosphere as a shaper and governor of the atmospheric and oceanic dynamic equilibrium.
Leif Svalgaard (11:40:55) :
And the ACRIM data problem is not particularly relevant to the issue of the long-term variation of HMF and TSI. So, to keep introducing this red herring is a waste of bandwidth [peoples’ and networks’]
It’s very relevant to the understanding of the fact that the sun is not as invariate as your thesis suggested, though the recent big drop of TSI has changed that analysis somewhat. The fact that despite our best efforts at calibration TIM reads 4.4W/m^ lower than the previous satellite, which saw 50% of the change between solar max and solar min that ACRIM did hardly seems irrelevant to me. It tells me there is a large enough amount of uncertainty in TSI measurement that nothing categorical can be deduced about it’s levels a hundred years ago for a start.
And what was the average from 1900-1930 from your graph? The upward trend was from then to now, not from cycle 17.
The trend stopped in the 1950s, as we are now back to where we started.
There is little point in getting into debates about when treds should be measured from and to, but I note the densely packed short minimum cycles of the late C20th produced almost the same average as the period containing the highest Rmax cycle despite their waning amplitudes.
I can see why you don’t want to answer the question, because I have calculated the percentage change in sunspot numbers between the various periods. The bottom line is that the linear trend was up over the C20th as a whole according to your graph.
My calcs show that this has been building up since 1935
And why did the oceans not store heat before 1935?
They did, and gained ocean heat content quite stronly during the high cycles of the mid C19th. But the gradual slight fall in ocean heat content according to my model from 1870 to 1935 due to the sunspot average being slightly below the equilibrium value is in big contrast to the sharp increase indicated by the much more active sunspot cycles between 1935-2003.
Your revisions to sunspot numbers will alter the curve and the equilibrium value if they become generally accepted. I sincerely hope Jeff Id is successful in his quest to look at the data himself, to corroborate or contradict your thesis. Are the original logbook records of the magnetic readings accessible by the public? Where are they kept?
My opinion/prediction is that the next few solar cycles will show us all just how variable the sun can be in it’s activity, and that will be the real test of the various hypotheses and data estimations. It’s going to be a fascinating story to watch unfold.
Your estimation based on the polar fields for solar cycle 24 is for an Rmax of ~75 around 2013.
My estimate based on barycentric nonsense(tm) is much less. ~45-55 a year later, if it really gets going at all.
But the sun is very unpredictable and difficult to read from previous behaviour as we are finding out, and I won’t take the result as proof of the correctness or falsity of either your theory or mine. So we’ll be able to carry on arguing about it for a long time to come I hope. 🙂
tallbloke (01:52:11) :
The fact that despite our best efforts at calibration TIM reads 4.4W/m^ lower than the previous satellite
The absolute value of TSI has always varied between satellites [and more than the 4.4 W/m2]. This is not relevant. What matters is the relative precision which is much better [something like 0.007 W/m2].
Are the original logbook records of the magnetic readings accessible by the public? Where are they kept?
yes, here: http://swdcwww.kugi.kyoto-u.ac.jp/hyplt/index.html and other places.
Thanks Leif.
By the way, using my data combination techniques and a bit of barycentricnonsense(tm) I’ve got a prediction for the Ap index to 2015 here with a hindcast to 1934 to show it’s strengths and weaknesses.
http://s630.photobucket.com/albums/uu21/stroller-2009/?action=view¤t=ap-prediction-1.gif
Lets see how it goes. If it’s on the mark I think solar cycle 24 will be long and low, with an Rmax 12 month average spot count of around 45-55 in 2015.
tallbloke (07:31:39) :
I’ve got a prediction for the Ap index to 2015 here with a hindcast to 1934 to show it’s strengths and weaknesses.
The hindcast to 1934 doesn’t look too good. Here is Ap back to 1844: http://www.leif.org/research/Ap-1844-2008.png or here http://www.leif.org/research/Ap%201844-2009.pdf
Hmmm, thanks. The correlation breaks down where the numbers of the two series get too far apart. I need to learn something about detrending algorithms I think.
Which of the long list of geomag stations have the furthest back records Leif?
tallbloke (09:42:52) :
Which of the long list of geomag stations have the furthest back records Leif?
At Kyoto, that would be Niemegk [NGK] back to 1890. There are others, like Greenwich and Prague and Helsinki that go back to the 1840s. Most of those have not been digitized and placed in public archives. I’m digitizing the old records from published ‘yearbook’ and have a lot of that old data and am working hard to get it all done and making it all available, but ain’t there yet [perhaps another ten years of work]. Section 7 [and Table 6] of http://www.leif.org/research/2007JA012437.pdf give more details on old stations.
tallbloke (07:31:39) :
I’ve got a prediction for the Ap index to 2015 here with a hindcast to 1934 to show its strengths and weaknesses.
when making such predictions [especially when using Ap at minimum in the process] one must be aware of the effect shown on page 4 of
http://www.leif.org/research/Ap%201844-2009.pdf namely that Ap at Odd-Even minima is about 3 units lower than at Even-Odd minima. The reason for this is understood [has been for 30 years] and has nothing to do with the Sun [to first order], but is a consequence of the geometry of the interaction of the solar wind with the Earth. You see, there are MANY such pitfalls lurking on the way. That Ap at this Odd-Even transition is so low is thus an artifact of the data, as far as the Sun is concerned, as the effect is terrestrial and not solar.
Leif Svalgaard (10:30:42) :
Maybe you aren’t the person to ask this question of, but how much energy is used by plants from TSI? Does changes in plant cover change the amount of TSI that is actually re-emitted by Earth? It would seem logical to me that ‘unused’ TSI is an important factor in climate studies, not TSI as a total figure.
I did not mean re-emitted, on second thought, I mean used to provide heat.