Guest Post by Willis Eschenbach
Well, I wasn’t going to mention this paper, but it seems to be getting some play in the blogosphere. Our friend Nicola Scafetta is back again, this time with a paper called “Solar and planetary oscillation control on climate change: hind-cast, forecast and a comparison with the CMIP5 GCMs”. He’s posted it up over at Tallbloke’s Talkshop. Since I’m banned over at Tallbloke’s, I thought I’d discuss it here. The paper itself is here, take your Dramamine before jumping on board. Dr. Scafetta has posted here on WUWT several times before, each time with his latest, greatest, new improved model. Here’s how well Scafetta’s even more latester, greatester new model hindcasts, as well as what it predicts, compared with HadCRUT4:
Figure 1. Figure 16A from Scafetta 2013. This shows his harmonic model alone (black), plus his model added to the average of the CMIP5 models following three different future “Representative Concentration Pathways”, or RCPs. The RCPs give various specified future concentrations of greenhouse gases. HadCRUT4 global surface temperature (GST) is in gray.
So far, in each of his previous three posts on WUWT, Dr. Scafetta has said that the Earth’s surface temperature is ruled by a different combination of cycles depending on the post:
First Post: 20 and 60 year cycles. These were supposed to be related to some astronomical cycles which were never made clear, albeit there was much mumbling about Jupiter and Saturn.
Second Post: 9.1, 10-11, 20 and 60 year cycles. Here are the claims made for these cycles:
9.1 years : this was justified as being sort of near to a calculation of (2X+Y)/4, where X and Y are lunar precession cycles,
“10-11″ years: he never said where he got this one, or why it’s so vague.
20 years: supposedly close to an average of the sun’s barycentric velocity period.
60 years: kinda like three times the synodic period of Jupiter/Saturn. Why three times? Why not?
Third Post: 9.98, 10.9, and 11.86 year cycles. These are claimed to be
9.98 years: slightly different from a long-term average of the spring tidal period of Jupiter and Saturn.
10.9 years: may be related to a quasi 11-year solar cycle … or not.
11.86 years: Jupiter’s sidereal period.
The latest post, however, is simply unbeatable. It has no less than six different cycles, with periods of 9.1, 10.2, 21, 61, 115, and 983 years. I haven’t dared inquire too closely as to the antecedents of those choices, although I do love the “3” in the 983 year cycle. Plus there’s a mystery ingredient, of course.
Seriously, he’s adding together six different cycles. Órale, that’s a lot! Now, each of those cycles has three different parameters that totally define the cycle. These are the period (wavelength), the amplitude (size), and the phase (starting point in time) of the cycle.
This means that not only is Scafetta exercising free choice in the number of cycles that he includes (in this case six). He also has free choice over the three parameters for each cycle (period, amplitude, and phase). That gives him no less than 18 separate tunable parameters.
Just roll that around in your mouth and taste it, “eighteen tunable parameters”. Is there anything that you couldn’t hindcast given 18 different tunable parameters?
Anyhow, if I were handing out awards, I’d certainly give him the first award for having eighteen arbitrary parameters. But then, I’d have to give him another award for his mystery ingredient.
Because of all things, the mystery ingredient in Scafetta’s equation is the average hindcast (and forecast) modeled temperature of the CMIP5 climate models. Plus the mystery ingredient comes with its own amplitude parameter (0.45), along with a hidden parameter for the zero point of the average model temperatures before being multiplied by the amplitude parameter. So that makes twenty different adjustable parameters.
Now, I don’t even know what to say about this method. I’m dumbfounded. He’s starting with the average of the CMIP5 climate models, adjusted by an amplitude parameter and a zeroing parameter. Then he’s figuring the deviations from that adjusted average model result based on his separate 6-cycle, 18-parameter model. The sum of the two is his prediction. I truly lack words to describe that, it’s such an awesome logical jump I can only shake my head in awe at the daring trapeze leaps of faith …
I suppose at this point I need to quote the story again of Freeman Dyson, Enrico Fermi, “Johnny” Von Neumann, and the elephant. Here is Freeman Dyson, with the tale of tragedy:
By the spring of 1953, after heroic efforts, we had plotted theoretical graphs of meson–proton scattering.We joyfully observed that our calculated numbers agreed pretty well with Fermi’s measured numbers. So I made an appointment to meet with Fermi and show him our results. Proudly, I rode the Greyhound bus from Ithaca to Chicago with a package of our theoretical graphs to show to Fermi.
When I arrived in Fermi’s office, I handed the graphs to Fermi, but he hardly glanced at them. He invited me to sit down, and asked me in a friendly way about the health of my wife and our newborn baby son, now fifty years old. Then he delivered his verdict in a quiet, even voice.
“There are two ways of doing calculations in theoretical physics”, he said. “One way, and this is the way I prefer, is to have a clear physical picture of the process that you are calculating. The other way is to have a precise and self-consistent mathematical formalism. You have neither.”
I was slightly stunned, but ventured to ask him why he did not consider the pseudoscalar meson theory to be a self-consistent mathematical formalism. He replied, “Quantum electrodynamics is a good theory because the forces are weak, and when the formalism is ambiguous we have a clear physical picture to guide us.With the pseudoscalar meson theory there is no physical picture, and the forces are so strong that nothing converges. To reach your calculated results, you had to introduce arbitrary cut-off procedures that are not based either on solid physics or on solid mathematics.”
In desperation I asked Fermi whether he was not impressed by the agreement between our calculated numbers and his measured numbers. He replied, “How many arbitrary parameters did you use for your calculations?” I thought for a moment about our cut-off procedures and said, “Four.” He said, “I remember my friend Johnny von Neumann used to say, with four parameters I can fit an elephant, and with five I can make him wiggle his trunk.”
With that, the conversation was over. I thanked Fermi for his time and trouble, and sadly took the next bus back to Ithaca to tell the bad news to the students.
Given that lesson from Dyson, and bearing in mind that Scafetta is using a total of 20 arbitrary parameters … are we supposed to be surprised that Nicola can make an elephant wiggle his trunk? Heck, with that many parameters, he should be able to make that sucker tap dance and spit pickle juice …
Now, you can expect that if Nicola Scafetta shows up, he will argue that somehow the 20 different parameters are not arbitrary, oh, no, they are fixed by the celestial processes. They will likely put forward the same kind of half-ast-ronomical explanation they’ve used before—that this one represents (2X+Y)/4, where X and Y are lunar precession cycles, or that another one’s 60 year cycle is kind of near three times the synodic period of Jupiter and Saturn (59.5766 years) and close is good enough, that kind of thing. Or perhaps they’ll make the argument that Fourier analysis shows peaks that are sort of near to their chosen numbers, and that’s all that’s needed.
The reality is, if you give me a period in years, I can soon come up with several astronomical cycles that can be added, subtracted, and divided to give you something very near the period you’ve given me … which proves nothing.
Scafetta has free choice of how many cycles to include, and free choice as to the length, amplitude, and phase of each those cycles. And even if he can show that the length of one of his cycles is EXACTLY equal to some astronomical constant, not just kind of near it, he still has totally free choice of phase and amplitude for that cycle. So to date, he’s the leading contender for the 2013 Johnny Von Neumann award, which is given for the most tunable parameters in any scientific study.
The other award I’d give this paper would be for Scafetta’s magical Figure 11, which I reproduce below in all its original glory.
Figure 2. Scafetta’s Figure 11 (click to enlarge) ORIGINAL CAPTION: (Left) Schematic representation of the rise and fall of several civilizations since Neolithic times that well correlates with the 14C radio- nucleotide records used for estimating solar activity (adapted from Eddy’s figures in Refs. [90, 91]). Correlated solar-climate multisecular and millennial patterns are recently confirmed [43, 44, 47]. (Right) Kepler’s Trigon diagram of the great Jupiter and Saturn conjunctions between 1583 to 1763 [89], highlighting 20 year and 60 year astronomical cycles, and a slow millennial rotation.
First off, does that graphic, Figure 11 in Scafetta’s opus, make you feel better or worse about Dr. Scafetta’s claims? Does it give you that warm fuzzy feeling about his science? And why are Kepler’s features smooched out sideways and his fingers so long? At least let me give the poor fellow back his original physiognomy.
There, that’s better. Next, you need to consider the stepwise changes he shows in “carbon 14”, and the square-wave nature of the advance and retreat of alpine glaciers at the lower left. That in itself was good, I hadn’t realized that the glaciers advanced and retreated in that regular a fashion, or that carbon 14 was unchanged for years before and after each shift in concentration. And I did appreciate that there were no units for any of the four separate graphs on the page, that counted heavily in his favor. But what I awarded him full style points for was the seamless segue from alpine glaciers to the “winter severity index” in the year 1000 … that was a breathtaking leap.
And as you might expect from a man citing Kepler, Scafetta treats scientific information like fine wine—he doesn’t want anything of recent vintage. Apparently on his planet you have to let science mellow for some decades before you bring it out to breathe … and in that regard, I direct your attention to the citation in the bottom center of his Figure 11, “Source: Geophysical Data, J. Biddy J. B. Eddy (USA) 1978″. (Thanks to Nicola for the correction, the print was too small to read.)
Where he stepped up to the big leagues, though, is in the top line in the chart. Click on the chart to enlarge it if you haven’t done so yet, so you can see all the amazing details. The “Sumeric Maximum”, the collapse of Machu Pichu, the “Greek Minimum”, the end of the Maya civilization, the “Pyramid Maximum” … talk about being “Homeric in scope”, he’s even got the “Homeric Minimum”.
Finally, he highlights the “20 year and 60 year astronomical cycles” in Kepler’s chart at the right. In fact, what he calls the “20 year” cycles shown in Kepler’s dates at the right vary from 10 to 30 years according to Kepler’s own figures shown inside the circle, and what he calls the “60 year astronomical cycles” include cycles from 50 to 70 years …
In any case, I’m posting all of this because I just thought folks might like to know of Nicola Scafetta’s latest stunning success. Using a mere six cycles and only twenty tunable parameters plus the average of a bunch of climate models, he has emulated the historical record with pretty darn good accuracy.
…
And now that he has explained just exactly how to predict the climate into the future, I guess the only mystery left is what he’ll do for an encore performance. Because this most recent paper of his, this one will be very hard to top.
In all seriousness, however, let me make my position clear.
Are there cycles in the climate? Yes, there are cycles. However, they are not regular, clockwork cycles like those of Jupiter and Saturn. Instead, one cycle will appear, and will be around for a while, and then disappear to be replaced by some longer or shorter cycle. It is maddening, frustrating, but that’s the chaotic nature of the beast. The Pacific Decadal Oscillation doesn’t beat like a clock, nor does the El Nino or the Madden-Julian oscillation or any other climate phenomena.
What is the longest cycle that can be detected in a hundred year dataset? My rule of thumb is that even if I have two full cycles, my results are too uncertain to lean on. I want three cycles so I can at least get a sense about the variation. So for a hundred year dataset, any cycle over fifty years in length is a non-starter, and thirty-three years and shorter is what I will start to trust.
Can you successfully hindcast temperatures using other cycles than the ones Scafetta uses? Certainly. He has demonstrated that himself, as this is the fourth combination of arbitrarily chosen cycles that he has used. Note that in each case he has claimed the model was successful. This by no means exhausts the possible cycle combinations that can successfully emulate the historical temperature.
Does Scafetta’s accomplishment mean anything? Sure. It means that with six cycles and no less than twenty tunable parameters, you can do just about anything. Other than that, no. It is meaningless.
Could he actually test his findings? Sure, and I’ve suggested it to him. What you need to do is run the analysis again, but this time using the data from say 1910 to 1959 only. Derive your 20 fitted variables using this data alone.
Then test your 20 fitted variables against the data from 1960 to 2009, and see how the variables pan out.
Then do it the other way around. Train the model on the later data, and see how well it does on the early data. It’s not hard to do. He knows how to do it. But if he has ever done it, I have not seen anywhere that he has reported the results.
How do I know all this? Folks, I can’t tell you how many late nights I’ve spent trying to fit any number and combination of cycles to the historical climate data. I’ve used Fourier analysis and periodicity analysis and machine-learning algorithms and wavelets and stuff I’ve invented myself. Whenever I’ve thought I have something, as soon as it leaves the training data and starts on the out-of-sample data, it starts to diverge from reality. And of course, the divergence increases over time.
But that’s simply the same truth we all know about computer weather forecasting programs—out-of-sample, they don’t do all that well, and quickly become little better than a coin flip.
Finally, even if the cycles fit the data and we ignore the ridiculous number of arbitrary parameters, where is the physical mechanism connecting some (2*X+T)/4 combination of two astronomical cycles, and the climate? As Enrico Fermi pointed out, you need to have either “a clear physical picture of the process that you are calculating” or “a precise and self-consistent mathematical formalism”.
w.
PS—Please don’t write in to say that although Nicola is wrong, you have the proper combination of cycles, based on your special calculations. Also, please don’t try to explain how a cycle of 21 years is really, really similar to the Jupiter-Saturn synodic cycle of 19+ years. I’m not buying cycles of any kind, motorcycles, epicycles, solar cycles, bicycles, circadian cycles, nothing. Sorry. Save them for some other post, they won’t go bad, but please don’t post them here.
‘It seems to me you are treating Scafetta like Hansen or Mann.?
Well Scafetta acts like Mann
When Steve McIntyre asked for Scafettas code, Scafgetta refused and turned the whole matter into a game.
When I asked for the code. Flat refusal.
Even Scafetta co author http://heartland.org/craig-loehle says that he wishes Scafetta would share code, but Scafetta continues to play the “mann” game and the “Jones” game.
Anyone with any ounce of integrity who fought with us to get data and code released by hansen, jones, mann and others should tell scaffetta to stop his childish nonsense and post
A) the code he used
B) the data he used
Not links, not descriptions, not arm waving attacks on the people who want to audit or build on his work. he needs to post the data and the code.
Until then folks like ferd should stop giving lectures on how science works. And in my opinion, WUWT should stop promoting Scafettas stuff.
Willis and Anthony and I disagree about a lot. But access to code and data is fundamental. You have to show your work, the ACTUAL WORK, and not merely descriptions of your work
Willis writes:
“Are there cycles in the climate? Yes, there are cycles. However, they are not regular, clockwork cycles like those of Jupiter and Saturn. Instead, one cycle will appear, and will be around for a while, and then disappear to be replaced by some longer or shorter cycle.”
Precisely, and that is not cycles as such, but an event series. The only fairly regular pulse is the sunspot cycle.
Anthony, I really do not understand why you have deleted my innocuous submissions from being posted here. If you cannot be open in the regulation of these comments then I shall refrain from wasting my time on composing comments for its admin to block. You must have a super-sensitive comment guideline? I’ll look over it, or I’ll just spark some discussion on Joanne Nova’s blog instead, who openly welcomes these sorts of off-topic discussions regarding AGW.
GabrielHBay says:
July 23, 2013 at 3:53 am
Gabriel, I can tell you that you may be the only one who didn’t notice the “Please” at the start of that request. Here’s a tip. When a man ASKS, not orders but asks you to do something, and says “Please”, that’s called a “polite request”. If you call it censorship, people will scratch their heads and wonder.
All the best,
w.
On cyclomania: http://www.leif.org/research/AGU%20Fall%202011%20SH34B-08.pdf
Ulric Lyons says:
July 23, 2013 at 8:20 am
The only fairly regular pulse is the sunspot cycle.
No. The frequency of the sun spot number is not constant. The frequency shift is in correlation to the global temperature (Yamal data)
http://www.volker-doormann.org/images/sunspot_tides.
V.
There is no Earthly reason for the major shifts in climate over the long term from ice age to interglacial over and over again on a somewhat regular basis according to palioclimatic data. I have concluded that there must be solar or other cosmic (Milky Way or larger scale astronomical) influences behind these major shifts. It also seems to me that there probably a is solar basis for the moderate scale shifts that lead to “little ice age” events. It is my hope that the astronomers and climatologists will work together in the decades ahead to understand these cause and effects and the means of energy transfer to make create these climatic influences by astronomical events. Many climatologists have dropped there eyes in dismisal when I have suggested all of this. None the less, I hold on to the idea that the answer is in the stars….
well; i do agree, he should focuse on finding out if there are periods…
too many parameters to do do forcecast
“One way, and this is the way I prefer, is to have a clear physical picture of the process that you are calculating. The other way is to have a precise and self-consistent mathematical formalism. You have neither.”
That is an either/or situation. When fitting observed data to a Fourier type harmonic, the statistical significance of each factor depends on the ratio of factors to degrees of freedom. So we may have a lot more confidence with lots more data and fewer factors than the other way. With regard to observed cycles, the measured precision of a cycle depends on the number of self-consistent cycles that are observed within a time frame. A statistically significant cycle that is longer than the time frame is not precise and is only slightly better as a predictor than a linear trend. I suggest that since radiation is essentially a “line of sight, speed of light” process, we should be using hourly met and CO2 data at different sites when analyzing changing mass and energy balances. There are too many assumptions “fudge factors” involved in the global models.
sorry, correct URL:
http://www.volker-doormann.org/images/sunspot_tides.gif
V.
http://www.webofstories.com/play/freeman.dyson/94;jsessionid=824C94DD263980B74AFB76282A33B896
This discusses how many free parameters are OK for a match, but more important what is the physical basis for the simulation. I think this discussion pretty much summarizes the discussion here
Leif at 08:42.
That I have to ingest. Logarigmic time onthe x-axis. Cool.
Monckton of Brenchley says:
July 23, 2013 at 3:49 am
Well, I wasn’t sure if I should write about this, but I guess I’m glad I did … not sure where to start here.
Christopher, you know that I’m a very big fan of yours. However, in this case you are supporting pseudoscience.
First, one cannot “test it in the simplest way” as you suggest. I spoke in the head post in a bit of detail about how you have to test these kinds of claims, by withholding half the dataset during the fitting so that you can use this later for testing. Scafetta has not done this to my knowledge. (Neither have the climate models, but that’s a different question)
Next, his cycle lengths are neither astronomical nor astrological. They are arbitrary parameters, fitted to the data. They are picked solely because they work. This should be evident from the fact that in some of his papers he uses 20 and 60 year cycles, and in others he uses 21 and 61 year cycles.
Next, I’ve never seen anyone other than Herschel find a correlation between grain prices and sunspots. Someone might have done it, but I’ve not seen it. In part that’s why citing work from 1801 is a bit sketchy. Many people, myself included, have looked at sunspot cycles to see what they correlate with. The answer is … not much. Occasionally someone finds some correlation, but often when you example it, the correlation is with some arbitrary cycle length like 11 or 21 years, and not with the actual sunspot cycle lengths.
But you delude yourself badly if you think that Scafetta is using sunspot cycles or anything like them. Scafetta is using sine waves. Sunspot cycles do not resemble sine waves at all. Their shape is different, and more to the point, the cycle lengths are greatly irregular. They vary from nine years to 50% more than that, 13.7 years.
Here’s the thing. If you want to compare to sunspot cycle lengths, you have to use the actual sunspot cycle lengths. You can’t use their average length, the growing grain isn’t responding to the average length.
As a result, if you talk about sunspot cycles, you are not talking about Scafetta’s work, and vice versa. Not only that, but “10.6 to 11 years” is nowhere near the values of the sunspot cycle, it goes from 9.0 to 13.7 years …
Mmm … as you are the acknowledged expert in these matters, let me draw your attention to the “tu quoque” nature of your argument about tunable parameters. Yes, the climate models have too many. But at least they have the advantage of being in some sense a representation of a known physical reality.
Scafetta has no less than twenty totally arbitrary tunable parameters, and no connection to any known physical mechanism. So I fear that justifying his parameters by saying “but they do it too” is a known logical fallacy.
But sure, put his predictions in your index. To date, he’s given us four different predictions, based on different combinations of the following cycle lengths in years: 9.1, 9.98, “10-11”, 10.2, 10.9, 11.86, 20, 21, 60, 61, 115, and 983 years.
So … which prediction do you plan to use? Will you use 20 years or 21 years as your cycle? Heck, why not make a “meta-prediction”, and include them all?
I am truly sorry to see you give this kind of nonsense any credence, my friend. It is not astronomy, the cycles don’t match. And I’m not saying it’s astrology, it’s not even that.
It’s pure numerology, as meaningful as giving the letters in your name each a number, and then adding up the numbers to predict the future.
My very best to you, and dang it, now I have to go see if I can replicate Herschel’s work. My thanks for your contribution to this post, and to the climate debate in general. Keep fighting the good fight, it’s always marvelous to watch you in action.
w.
vukcevic says:
July 23, 2013 at 4:29 am
Right. You have the actual real cycle numbers, and so the problem with Scafetta is just that he used the wrong numbers … vuk, I’m going to leave you to fight that out with Scafetta. He says 60 (or 61) years, and you say “64-65” years … go at it, guys.
This is why I’d said “Please don’t write in to say that although Nicola is wrong, you have the proper combination of cycles, based on your special calculations.” You’re debating with the wrong guy. Go over to tallblokes and put your case to him directly, I’m sure Nicola can give you chapter and verse about why “64-65 years” is absolutely and undeniably wrong.
w.
There is a much more scientific way that we express what Willis is saying and we say it over and over again “Correlation does not imply causation”.
You can take it to the absurd and say for example “the lack of pirates is causing global warming”
http://www.forbes.com/sites/erikaandersen/2012/03/23/true-fact-the-lack-of-pirates-is-causing-global-warming/
That’s the stupidity that Enrico Fermi was describing to Freeman Dyson in the story.
Many of the comments don’t get the message that even if Dr. Scafetti could match global warming perfectly it means absolutely nothing as does the correlation between lack of pirates and global warming, how we would we know that does not come about by pure chance like the match between pirates and global warming.
What Scafetti is doing is not even science and Willis has correctly identified that fact … don’t shoot the messenger … science demands that you link correlations thru a solid mechanism.
The problem for Scafetti is there is no mechanism to link astral cycles especially of ridiculously small planetary bodies versus massive earth local effects.
@Willis
“Christopher, you know that I’m a very big fan of yours. However, in this case you are supporting pseudoscience.”
You are absolutely correct Willis he is and it doesn’t seem to have dawned on him along with many of the others comments.
Sad really.
Willis,
One the things that you has to be learned in life is to get to know and understand the idea that you are criticizing. Your post shows me that you have not made that attempt. It is full of paper-tiger arguments which are used to make you look as though you have done your home work. You have not.
You are correct in pointing out that a plethora of adjustable constants can make any model fit the data. Point taken. However, there are many good reasons [some given by Lord Monckton above] as to why research like Nicola’s should be encouraged.
Basically, what Nicola’s heuristic model assumes that the Earth’s climate is influenced by external factors that are related to the overall level of solar activity. Given that there is some peer-reviewed research that indicates the possibility that the overall level of solar activity may be influenced by planetary torques and tides, Nicola (and others) are taking the logical next step, and that is to look for a connection between the long-term cycles in the Earth’s climate system and long-term cycles in the planetary toques and tides.
[You can read about the Tidal-Toquing model which is explained here:
http://astroclimateconnection.blogspot.com.au/2013/03/why-vej-tidal-torquing-model.html
http://astroclimateconnection.blogspot.com.au/2013/04/further-confirmation-of-vej-tidal.html
or you can read some per-reviewed science at:
J. A. Abreu1, J. Beer, A. Ferriz-Mas, K. G. McCracken, and F. Steinhilber
Is there a planetary influence on solar activity?
A&A 548, A88 (2012)]
Nicola has identified what he believes are the most important cycles that may [and I emphasis the word “may”] play a role in the interaction between the Sun and Earth’s climate.
Nicola is not the first person to point out these potential external forcing cycles. He is just he current lightening rod for those who want to show their unease with new and/or alternative ideas.
Here is a link to a post that I believe gives support to Nicola’s work:
http://astroclimateconnection.blogspot.com.au/2013/07/is-this-planetary-signature-in-our.html
Ian Wilson says:
July 23, 2013 at 9:45 am
He is just he current lightening rod for those who want to show their unease with new and/or alternative ideas.
This is not a new idea. It goes back to 1852 and has not gained acceptance because it has not worked and not advanced science.
pokerguy says:
July 23, 2013 at 5:42 am
You find my work “extremely distasteful”, and yet here you are reading my work … why am I so popular with people who want to tell me that I’m doing it all wrong?
pokerguy, perhaps you’re not aware of the backstory. Nicola and I have a history. I started out by asking him simple questions, like what a cycle of “10-11” years means. Here was my first question to him, plus a bonus question, and they are important questions:
He curtly and rudely refused to answer those and other questions. Many of them, including those simple questions I just quoted above, he has never answered. Instead, he kept telling me I should read his entire opus of work to find the answers … sorry, not going to happen. I’m not wading through miles of mire looking for a snipe. A simple question deserves a simple answer, and he has consistently refused to answer questions about his work, other than softball pitches from his followers. Instead, he rapidly descended into personal attacks. Here’s my comment from 2012, when someone asked why I wasn’t more polite to Nicola. I replied:
So yeah, we have some history.
Finally, anyone using twenty tuned parameters to do some kind of curve fitting, without the elementary precaution of only using half the data for fitting so you have the other half for verification, is practicing pseudoscience. It is a meaningless curve-fitting exercise.
What he has done is a kind of manual Fourier reconstruction of the climate signal. But as Fourier showed, any arbitrary signal can be decomposed into a set of sine waves. And that’s all Scafetta has done, decomposed some signal into sine waves. Are you impressed by that kind of pseudoscience? Because I’m not.
The problem, and the reason I decided to write this post, is that it is plausible enough pseudoscience to take in folks like Monckton of Brenchley, who should know better. So yes, I’m as scathing about this kind of garbage as I can be. I don’t wish for this site or the skeptical movement in general to be associated with Scafetta’s curve fitting exercise in any shape or form. I want to disavow it in the strongest possible terms.
All of the best,
w.
Willis Eschenbach says:
July 23, 2013 at 9:20 am
Monckton of Brenchley says:
July 23, 2013 at 3:49 am
Re: Herschel, sunspots & grain prices:
http://www.cxoadvisory.com/3642/calendar-effects/does-the-sunspot-cycle-predict-energy-and-grain-prices/
However, 20th & 21st century grain production & pricing differ in significant ways from the early 19th century British situation. Weather still matters of course, but inputs such as chemical fertilizers, petroleum-based cultivation practices, widespread irrigation, possibility of alternative crops, fallowing, government intervention & a global grain market, among other factors, could muddy the signal, if any.
GabrielHBay says:
July 23, 2013 at 5:54 am
Gabriel, I do like the description of “a tad too warmist”, although I’m not sure how it applies to me. However, if you think that Dr. Scafetta’s work is astronomically based, please disabuse yourself of that belief immediately. His cycles have nothing to do with astronomy, they are picked because they fit. As Vukcevic commented on an earlier thread:
Three synodic cycles of Jupiter and Saturn is 59.5766 years. Scafetta sometimes uses 60 years, and sometimes 61. His claim of an astronomical origin for those numbers is falsified by the data.
He’s just picking numbers, Gabriel.
w.
richard verney says:
July 23, 2013 at 5:58 am
What is it about the story of the elephant that you don’t understand, Richard? This is a meaningless curve-fitting exercise, and the number of free parameters turns it into a joke. If you don’t see the humor, read the story of the elephant again.
w.
John West says:
July 23, 2013 at 6:00 am
As you are the first person to mention Milankovitch cycles in this thread, I have no idea what you are referring to. If it was something I said, please quote my words.
Thanks,
w.
Willis Eschenbach says:
July 23, 2013 at 10:01 am
I don’t wish for this site or the skeptical movement in general to be associated with Scafetta’s curve fitting exercise in any shape or form. I want to disavow it in the strongest possible terms.
There are several other cyclomaniacs on this site pushing similar [or worse] nonsense. And I agree that this association harms the skeptical movement. Once you have been stung by the cyclomania bug, there is no way back and no salvation for you. That bug rears its stinger in every generation. A good example is the recent claims by Abreu et al. You can judge for yourself http://www.leif/research/Comment-Planetary-Peaks.pdf
http://www.leif.org/research/Comment-Planetary-Peaks.pdf