Guest Post by Willis Eschenbach
Normal carbon has six neutrons and six protons, for an atomic weight of twelve. However, there is a slightly different form of carbon which has two extra neutrons. That form of carbon, called carbon-14 or “14C”, has an atomic weight of fourteen. It is known to be formed by the interaction of high-energy cosmic rays with the atmosphere.
Therefore, the production of the carbon isotope 14C goes up and down with the number of cosmic rays.
Thus, other things being equal, the production of 14C could be a proxy for how many cosmic rays are passing through the atmosphere. And the number of cosmic rays striking the earth is regulated by a combination of the magnetic fields of the earth and the sun. When the combined magnetic field is strong, it deflects the cosmic rays away from the earth. When it is weak, more cosmic rays strike the earth.
So let me start, as I prefer to do, with the largest, longest view of the underlying raw data. In this case it is something called the “INTCAL13 Calibration Curve”. It is a record of historical variations in the levels of the carbon isotope 14C.
Figure 1. INTCAL13 calibration curve. The interval between values is five years in the recent part of the record (since 11950 BC). In the middle part of the record, from 23050 BC to 11950 BC values are ten years apart. And in the earliest part, from the beginning to 23050 BC, values are 20 years apart.
The large variations in the curve are said to be from slow changes in the earth’s own geomagnetic field over the millennia. However, our knowledge of geomagnetism millennia ago is not of the finest … given that, it does seem like a possible explanation.
Keep that INTCAL13 calibration curve in mind for a moment, and let me move on to discuss a guest post over at Judith Curry’s often-excellent website, by someone named “Javier”. The post is all about solar cycles. And there’s a new post on WUWT discussing Javier’s solar cycles. These are solar cycles that are two thousand four hundred years long, to be exact. How do they know that? Well, here’s Javier’s money graph.
Figure 2. Javier’s graph showing a claimed 2400-year cycle in the 14C record, which in turn is claimed to be a solar cycle.
So the obvious question is … how on earth did they get from the curve shown in Figure 1 to the curve shown in Figure 2? Javier says it is done by “removal of the long-term trend” … but how was that done? I went to the cited work of Clilverd et al. to find out the answer.
First, because the sun and the cosmic rays are negatively correlated, they flip the 14C record over as shown in Figure 3. In this orientation, warmer is at the top of the chart and cooler is at the bottom. That’s just a graphic convenience, no problem.
Figure 3. Inverted INTCAL13 calibration curve.
Then they throw away more than three-quarters of the data, leaving only the chunk since 9600 BC as shown in Figure 4.
Figure 4. Inverted INTCAL13 calibration curve since 9,600 BC.
Following that, they fit a linear trend to the data, and detrend it. Then they subtract out a purported 7,000 year signal of unknown origin. Figure 5 below from Clilverd et al. illustrates the procedure. Note that the upper panel of Figure 5 matches my Figure 4.
Figure 5. Figure 1 of Clilverd et al.
I note in passing that although Javier asserts a “correct” cycle length of 2400 years, Clilverd shows a 2300 year cycle. I guess that’s why Javier’s version is “adapted from” … but I digress.
Let me recapitulate the bidding. To get from the inverted 14C record shown in Figure 3 to the record used by Clilverd et al, they have
- thrown away three-quarters of the data,
- removed a purported linear trend of unknown origin from the remainder,
- subtracted a 7000-year cycle of unknown origin , and
- ASSERTED that the remainder represents solar variations with an underlying 2,300 year period …
I suspect that y’all can see the problems in each and every step of that process. First and foremost, why throw away three-quarters of the data? That alone disqualifies the study in my mind. But let us continue listing the difficulties:
Where did the claimed linear trend come from? What justifies removing it? Why use an exactly 7,000 year cycle, and where did it come from? How does one diagnose a 7.000 year cycle when you only have about 12,000 years of data, not even two full cycles? How do they know that the 7,000 year cycle is NOT solar-related and the 2,400 year cycle IS solar-related?
And finally, what evidence do we have that the remainder has anything to do with the sun?
But wait, as they say on the TV ads, there’s more. Let’s set the work of Javier aside entirely and return to the question of cosmic rays, which Javier does not discuss. Remember that the relationship between cosmic rays and the climate is supposed to work as follows:
In times when there are more cosmic rays, the rays cause more cloud nuclei to form. As a result more clouds form (and in addition, more 14C forms) and the world is colder. But in times when there are less cosmic rays (indicated by less 14C), there are less clouds, and thus the world ends up warmer.
And according to that theory, people claim that the final dip in the 2400-year cycle seen in Figures 2 & 5 is the cause of the cold times around the Little Ice Age. Back then it was a couple of degrees cooler than at present. If that theory is correct, this means that a change in ∆14C of about 10 per mil reflects a change in cosmic rays that is enough to cause a global temperature change of 2°C.
Now that all sounds good until you take another look at Figure 3. Let me replot it, and this time I’ll include the 10 per mil change in ∆14C, and hence in cosmic rays, rumored to be responsible for the Little Ice Age.
Figure 6. Inverted INTCAL13 calibration curve. Gray lines show the variation in ∆14C of 10 per mil claimed to be from cosmic rays and said to be responsible for the 2° cooling during Little Ice Age. The large swings are said to be due to changes in the strength of the geomagnetic field.
I reckon you folks can see the difficulty … according to this, about twenty thousand years ago it should have been about 100°C colder than today …
Now, about the only way out of this dilemma is to say that the peak-to-peak swing of about 500 per mil in ∆14C is from some kind of non-cosmic ray variations. You know, like the claimed 7,000 year cycle that was removed in Figure 5 that was ascribed to … hang on, I want to get this right … OK, they said it was from “changes within the carbon system itself”.
(Let me say that I like that particular bit of bafflegab a lot, “changes within the system itself”. Seems like that would cover a host of unpleasant variations in any dataset you might find … but again I digress.)
So to recap: IF the claims are true that the changes in ∆14C shown in Figure 6 reflect changes in cosmic rays and that the changes in cosmic rays result in the claimed changes in temperature, then twenty thousand years ago the earth should have been 100°C cooler. Even if I’m wrong by 100%, it is still saying that it was 50°C cooler back then … didn’t happen.
Since that is not possible, then it seems we must assume that “changes within the system itself” are causing the huge swings in ∆14C.
But if that is the case, then it is more than possible that these unknown changes within the system are also responsible for the smaller swings currently ascribed to variations in cosmic rays.
Anyhow, that’s my cosmic problem with rays. Here I have no problems. It’s two AM, I’m a night owl. There’s been rain for three days, wonderful rain. And there’s still rain in the area, a small cell passing north of us. But the wind has shifted. It was blowing strongly from the south or southwest for the last three days. Now the wind is just a mild breeze, and from the west. There are big gaps in the clouds, and the moon, aah, for the first time in a while the moon is finally showing its face. I can hear the distant hungry grumbling of the surf as it nibbles on the ribs of the coastline some six miles (ten km) away … a good night to be alive here in the redwood forest, with the giant trees standing stark and clear in the pale wash of moonlight, and silvered cumulus drifting across the sky.
Best to everyone,
w.
My Usual Request: Misunderstandings are the bane of the web. Please further understanding by quoting the exact words that you disagree with. That is the only way that we can all be clear about the exact nature of what you object to.
My Other Request: Bald statements that someone doesn’t know what they are doing, even if true, are of little use to anyone. If you think someone is using a wrong method or a wrong dataset, please further everyone’s understanding by demonstrating the right method or by linking to the right dataset.
Javier October 17, 2016 at 5:16 am
Javier, first, thanks for showing up to defend your work. Many do not do so, I commend you.
But I said nothing about your unwillingness to sign your own words, neither by word, tone, nor intent. Any interpretation of that nature is all you. I was uncertain how to write that sentence and gave it some thought. I couldn’t really write “a paper by Javier saying …” because that seems to assume that the reader knows who Javier is and it sounds like a last name. So I settled on “someone named “Javier”. I’m sorry you took that as me casting shade on you, it was not my intention.
I find exactly three mentions of the Hallstatt cycle in Google Scholar prior to the 21st century … and Bray claimed the cycle was 2600 years, viz:
You and Clilverd and Bray all disagree about the length of the cycle … which is what I was pointing at. When that happens, I have to wonder if we’re talking about a real cycle, or just another of the many appearing and disappearing cycles we see in the sun at all the time scales.
For example, there is a claimed “Gleissberg Cycle” in the sunspot data. But when it is examined closely, it’s seen to exist only in the first half of the data, and to die out in the second half. Here is a Complete Ensemble Empirical Mode Decomposition (CEEMD) of the sunspot data (source and CEEMD description here):
ORIGINAL CAPTION Periodograms of each of the intrinsic modes C1 through C7 of the CEEMD analysis of the annual mean sunspot number, 1700-2014. These show the strength of waves of the various periods in each on the intrinsic modes.
The so-called “Gleissberg Cycle”, upon which much ink has been expended, is visible in Mode C6, with a frequency of about 90 years. Now, based on that you’d say that a) the Gleissberg Cycle is real, and b) it’s the second-largest signal in the sunspot data. But here’s the problem.
ORIGINAL CAPTION CEEMD analysis of the mean annual sunspot numbers. Top panel shows the sunspot data, standardized to a mean of zero and a standard deviation of one. Panels marked C1 – C7 show the intrinsic modes of the signal. The bottom line shows the residual, meaning what remains after the removal of modes C1 – C7 from the signal. Note that all intrinsic modes are displayed at their true size, with all scales being the same.
As you can see, the so-called “Gleissberg Cycle”, upon which much ink has been expended, is visible in Mode C6 … but it only exists in the first half of the data, and even then it only exists for a couple of cycles. After that it dies away totally.
I’ve seen this pattern over and over in natural datasets—an apparent “cycle” comes into existence, lasts for a few cycles, I’ve seen as many as five cycles … and then it dies away altogether. All of which has made me highly skeptical of arguments about some purported 2600 or 2400 or 7000 or 2300 or 88 year solar cycle … they come and go.
Next, please note that the so-called “Gleissberg Cycle” persisted for a century and a half and then died out entirely, so a hundred years of data is far from sufficient for deciding if a real cycle exists … but let me point out something even more amazing. Here is the CEEMD analysis of your exact data that you say contains a 2300-year cycle:
This is actually a lovely analysis, as it highlights several problems with the standard analysis. First, take a look at mode C8. It contains a strong cycle with a length of just under a thousand years, a cycle which shows up in a Fourier or periodicity analysis … but over the period of the record, it dies out to zero. And not quickly either. It gradually goes from a strong, clear cycle 9000 years ago to no cycle at all.
So let me ask you this, Javier. Is there a ~1000 year cycle in the C14 data? Well … yes and no. You see the problem? Apparent cycles appear, exist for a while, even for millennia, and then disappear …
And the situation with the cycle of around 2400 years is similarly complex. It’s shown in Mode C9. It shows a ~ 2100-year cycle for a few cycles, then a long cycle of about 2700 years, then a final cycle of about 2100 years. All of these combine together to give a best Fourier fit for a 2400-year cycle … but that is not what is actually happening. In fact it is a string of shorter cycles with a long cycle in the middle.
So again, let me ask you: is there 2400-year cycle in the ∆14c data? Well, not as we generally understand a 2400 year cycle. However, the CEEMD analysis does make it clear why y’all get different values for the putative cycle … because it is a mix of various cycles coming and going. And remember, this is just a short cherry-picked section of the data. If we add more years, things don’t get better …
And of course, just like the very solid 1000-year cycle in Mode C8 of the same data, over the next twelve millennia the 2400-year cycle may indeed disappear like the Cheshire Cat, leaving only a faint ripple in the ether.
The underlying problem is that the sun, like many natural systems, appears to be chaotic in nature at all time scales.
Finally, Usoskin did an analysis of the∆14C data itself to claim significance … but like you he threw away most of the data first, with the same weak excuse that before 6700BC it was too uncertain. Unlike you, he only used the data from 6700 BC … this is what I meant about cherry picking. Everyone just uses the part of the 14C data that supports their thesis, and waves their hands at the rest and utters the magic words “too uncertain” …
But that’s not my main point. All tha Usoskin may or may not have established is that there is a 2400-year cycle in the 14C data … but you seem to think that he’s established that there is a corresponding cycle in the sun.
Here is the INTCAL data with the error estimate from the folks who put it together:
You don’t like the data older than 9600 years because it differs from what your theory might claim. So you wave your hands at changes in the “carbon mass budget” … we don’t even understand the current carbon mass budget, much less that of ten millennia ago.
And surely if there is a 7000 year cycle in the recent data, a cycle of that size would be clearly detectable in the data out to at least 25000 BC, despite the greater uncertainty.
While it is true that uncertainty increases with age, this is true of all natural datasets. Since the variations in the signal are much, much greater than the uncertainty of the signal, there is absolutely no justification for throwing the earlier years out.
In particular, the idea that you can cherry-pick your 9600 BC starting date based on uncertainty and that Usoskin can make the same claim about 6700 BC is totally unjustified—the uncertainty is quite small out to at least 25000 BC, and the larger swings prior to that are assuredly as significant as the 7000-year swing that you have removed from the data.
Not true in any sense. First, we only have good sunspot data back to about 1700, so there is not a “400 year period of overlapping”. Instead, we have the 250-year period from 1700 to 1950, with samples every five years … in other words, we have only fifty-one samples in the period of overlap.
It gets worse from there. Both the sunspot and the ∆14C data are highly autocorrelated, with high Hurst exponents (∆14C data, Hurst exponent = 0.88; sunspots, Hurst exponent = 0.76). As a result of the small dataset and the large Hurst exponents, the correlation of the two datasets (sunspots and ∆14C) have a p-value of 0.3 … in other words, we do not have anywhere near enough actual observational data to back up your claim that the relationship is significant.
Javier, my problem is that folks seem to take claims like “400 years of overlap” or “Gleissberg Cycle” for granted, without ever actually doing the math including the adjustment for autocorrelation. I don’t shy from such scut work, because unlike far too many PhD scholars, I take “nullius in verba” quite seriously.
I’d be cautious about claiming to speak for Leif … speaking for myself, I do think that the ∆14C data does contain solar information. The questions are how much information is encoded there, and what the signal to noise ratio is, and how we can reliably distinguish signal from noise … questions I don’t think have been asked a lot, much less answered.
My profound apologies, you are quite correct. I have changed the head post to reflect your very valid objection, and said specifically that you do not discuss cosmic rays.
See above. I’ve changed the head post.
I’ve now left you out entirely, again my apologies for the offense … but that is indeed the result of the popular claim that cosmic rays rule the temperature. That claim is not my strawman, it is the strongly held belief of many.
Many people claim that cosmic rays are the connection between the sun and the climate. It appears you don’t think that is true, and neither do I… but that’s just you and I, lots of folks disagree. I was addressing them.
However, I’d be curious if and how you think the tiny variations in solar output (none of which exceed about a third of a watt per square metre on a global 24/7 basis) affect the climate.
Oooh! Vague insults based on personal opinions! ! Can I play too? Please? OK, here’s my contribution:
You see how pointless that kind of nonsense is? Did that personal attack convince you that you were wrong? Of course not, and your attempt at mudslinging is no different. Could we stick to the science, please?
In that regard, Javier, you are the one defending throwing away three-quarters of a perfectly fine dataset with an established error estimate as being good solid science, not me. You are the one defending removing a 12,000-year-long linear trend from the data based on a handwaving claim about a perfectly-straight-line-for-twelve-millennia change in “internal carbon system dynamics”, not me. You are the one wielding a magic cycle of exactly 7000 years that removes unwanted variations in the actual data … and please don’t claim that the 7000 years is observationally determined, nature is never that neat. You are the one claiming Bray was wrong about the length of the Bray Cycle.
You can see why I might be less than convinced by your arguments … but I do thank you for raising them. As I said, not many people do so. I tip my hat to you.
Regards,
w.
Javier seems to have detrended to data because that’s what everyone in climatology does with everything.
I appoligise to J if he did explain why he detrended the but his C. Etc. post was massively long and meandering and did not seem to have much substance so I ended up scanning and then skipping.
I too am sceptical of all ephemeral cycles that people manage to see in solar activity.
However, that slope just reflects the basic nature of exponential decay. IntCat13 data is a “calibration” curve if you compare col 1 and col 2 , as presented here it is the transfer fn from d14C to sample age.
It is the deviations that ( may ) reflect changes in 14C creation due to cosmic rays. So his detrend seems to be approximately right. I think there is probably a more accurate way to do this mathematically since the decay half-like is believed to be very constant in our understanding of the process.
I would criticise the lack of scientific reason for detrending but I think it’s about right.
The fact that the data remains fairly straight out to 20ka suggests that, despite increased uncertainty it may be usable. The circa 7ka ripple seems to be visible in that period.
Here is a quick look at the spectrum of this data. Not I used first diff rather than detrend which attenuates as 1/f . We also see the data appears to be low-pass fitlered, this I explained above is a result of the exponential decay damping out fast changes in the early record.
The “Gleissberg” period is equally well defined in both the 14ka and 6ka subsets. The 148, 208 less so.
The circa 2500 year periodicity is obviously less well defined in the shorter data but there seems to general agreement.
Then I apologize too. I have been insulted recently here at WUWT for not revealing my name.
https://wattsupwiththat.com/2016/10/07/evidence-that-multidecadal-arctic-sea-ice-has-turned-the-corner/#comment-2315034
Well, there are many more mentions of the ~ 2200-2400 periodicity in radiocarbon data. Roger Bray described the periodicity without using radiocarbon data that was not available in 1968, so the initial relationship between solar variability and climate was not established based on cosmic rays.
The classical citation for the radiocarbon data is Damon and Sonnet 1991, although they recognize Houtermans 1971 as the first report of the periodicity in 14C:
III. THE ~ 2300 YR PERIOD
Aside from the aforementioned long (secular?) variation, the strongest feature in the ∆14C record is the long period of ~ 2300 yr. This component of the ∆14C spectrum, in addition to the 208 yr period, was first reported by Houtermans (1971). Its source is enigmatic but probably not attributable to the geomagnetic dipole field, for no periodic geomagnetic dipole field change of the required amplitude has been detected.
Damon and Sonnet 1991, pg. 366
My bold. Damon, P. E., & Sonett, C. P. (1991). Solar and terrestrial components of the atmospheric C-14 variation spectrum. In The Sun in Time (Vol. 1, pp. 360-388).
The differences in periodicity are due to the different methods to establish it, and on the irregularity of the cycle. The 11 year cycle is actually 8-15 years, a ±33% variability that in the case of a 2300 yr cycle would give a ±700 years. We accept the 11 year cycle because we have 23 oscillations. For the ~ 2400 yr cycle we only have 5 oscillations in the last 11,000 years, and with very old and imprecise data. Solar cycles are also famous for their differences in amplitude. It is in their nature.
The situation that you describe is exactly the same for the 11-year sunspot solar cycle. It also disappears between 1650 and 1700. It has irregular length so if you measure it between 1850 and 1900 it will be a 12 yr cycle and if you measure it between 1900 and 1950 it will be a 10 yr cycle. So let me ask you this, Willis. Is there a ~11 year cycle in the sunspot data? Well … yes and no. You see the problem? Whether we think it is due to a chaotic nature or to a poorly understood cause for solar variability, we must accept that this is the nature of solar cycles or abandon its study. Most people seem to believe in the 11 year cycle by conveniently ignoring the same reasons they use to not believe in the ~ 2400 yr cycle. Not very consistent.
That there is a 2400-year cycle in the 14C data is not discussed. It was found from day one. That 14C is a proxy for solar variability is generally accepted. That the 2400-year cycle represents solar variability is a reasonable assumption that is extraordinarily common in the scientific community. You are correct that we have no proof that the Sun varied with a 2400-year cycle but extending that unreasonable demand means we cannot use any proxy for anything.
You are making a crucial mistake. You are assuming that the error in carbon dating determination (That’s what IntCal is) can be assumed to be the error in solar activity estimation. Solar activity estimation involves several steps that each one has its own error source, since it requires two box models, one for the biosphere and one for the oceans. The final error I have already stated that becomes unacceptable when data is older than 12,000 yr BP. It is panel b in this figure:
http://www.clim-past.net/9/1879/2013/cp-9-1879-2013.pdf
Don’t be ridiculous. If the data doesn’t support my hypothesis I abandon it. I have already abandoned two hypothesis for lack of data support being the first that the world is warming almost exclusively because of the increase in atmospheric CO2, and the second that the solar variability is just too small to produce significative climate changes. I have nothing against older data except that the experts refuse to use it because they consider it unreliable and I agree with them.
Ok. I count 300 years and it is good enough of a match for me.
http://i1039.photobucket.com/albums/a475/Knownuthing/Goslar14CSSN_zpsskwzxv52.png
Considering that for the ~ 2400 yr cycle we are talking exclusively about grand solar minima as the lows in the cycle are the highest 14C production periods of all, I don’t think the signal to noise issue is that relevant. If 14C data represents solar activity the 2400 yr cycle is real. As I said the robustness of this feature in the data is what it made it to be discovered from day one.
Fair enough. I am glad we have cleared that and thank you for correcting the article.
I agree on sticking to the science. I was angry because your criticism was naming me but was directed to the work of Clilverd et al., and to the beliefs of “lots of folks” as if I was responsible for that or shared it.
It is not actually me. It is the published science on the issue. The solar variability estimate has to be extracted from the ∆14C data using two different box models and a reference. The models do not work for out of the Holocene conditions and thus the estimate becomes too unreliable to be useful. It is very easy to defend that if all of the authors of the estimates (there are over 10 articles on the solar variability reconstruction issue) do not consider data older than 12,000 reliable, it would be unwise to use it to extract conclusions. I can imagine the critics for doing so.
Again it is not me. There are half a dozen solar reconstructions in the literature, and although made by different authors with different methodology, they all perform a detrending. I have to say however, that it does not matter how you do that, because the Bray cycle is based on the position of grand solar minima, and that doesn’t change independently on the methodology employed because it is in the raw data before models and detrendeing are applied. This is what it is usually called robust feature as Damon and Sonnet noticed in the cite above.
On the contrary. I am open to the existence of longer oscillations in solar activity. Some authors propose a 6000 yr solar cycle that could be related to the 6000 yr Heinrich oscillations in ice rafted depositions in the North Atlantic and in Polar Circulation Index and δ18O variations in ice cores. I find that the evidence for such cycle is still insufficient to defend it. However again, everybody finds the ~ 2400 year cycle in the data, whether they subtract the longer periodicity or not.
He might still be right. Determining the length of the cycle is difficult with so few instances. I have preferred to take the very well dated low at 10.3 Kyr BP (figure 4 in my article) and the very well dated Spører minimum to calculate the average duration over 10,000 years. The uncertainty in the cycle length has no relationship with whether it is real or not.
It is not my goal to convince but to provide information. My article contains 54 references:
https://curryja.files.wordpress.com/2016/09/bibliography.pdf
Even if you find one of them to be incorrect that doesn’t affect the conclusions. If Clilverd et al., 2005 are wrong, I will bring forward McCracken et al., 2004; Solanki et al., 2004; Vonmoos et al., 2006; Muscheler et al., 2007; Steinhilber et al., 2012; Inceoglu et al., 2015; Usoskin, 2013; Usoskin et al., 2014.
Willis says that Javier has:
1) thrown away three-quarters of the data,
2) removed a purported linear trend of unknown origin from the remainder,
3) subtracted a 7000-year cycle of unknown origin , and
4) ASSERTED that the remainder represents solar variations with an underlying 2,300 year period …
However, Javier has used the only reliable part of the record: The first 11,000 years when we can count the years backwards in time via tree rings AND measure the amount of C14 in the same tree ring. Before that, any error in dating samples for calibration produces an error in the calibration curve. Your interesting demonstration that these cycles and their effect on temperature must be bogus (-100 degC) depends on the accuracy of the much less reliable data from before 11,000 BP.
Removing a linear trend may be justified if the wrong half-life is used to calculate the difference between observed and C14 predicted for a constant rate of C14 production. That can be tested.
There are techniques for judging the statistical significant of empirical methods of abstracting a sinusoidal signal. Properly characterizing a 7000-year period is 11,000 years of data is impossible. Will the 2,300 year cycle remain statistically significant if you don’t remove the 7,000 year cycle?
What is missing from your and Javier’s analysis is a comparison of the B10 and C14 proxy records. Both are alleged to be measures of solar activity. But one rarely sees them together. And despite our best efforts (and no effort by some), it is very easy to cherry pick among many datasets and have your attention drawn to those that appear to show a correlation. Chaotic systems often show periods of sinusoidal behavior interspersed with non-sinusoidal behavior. However, if “solar activity” is controlled by the planets, there should be sinusoidal behavior all of the time.
Frank October 17, 2016 at 6:57 pm
What is missing from your and Javier’s analysis is a comparison of the B10 and C14 proxy records.
Should be ‘Be10’.
Phil: Absolutely right.
“However, if “solar activity” is controlled by the planets, there should be sinusoidal behavior all of the time.”
When there are many planets the result could well have sections with reasonably stable periodicity and then others where this breaksdown and seems absent.
I’m making a case for the planetary model, but what you say here is incorrect.
Greg: Willis discussed finding some data sets with periods of stable oscillations behavior followed by periods with no oscillations. This sort of behavior can be observed in some chaotic systems, but my point is that this wouldn’t happen if the planets caused the solar cycle.
If the planets modulate the sun’s output, you are correct in pointing out that the phrase “sinusoidal behavior” doesn’t accurately describe the vector sum of the gravitational force from multiple planets. The orbit of Jupiter alone could produce sinusoidal behavior, modulated by the other planets.
The tides on Earth are caused by the rotation of the Earth under the influence of a slowly orbiting Moon (sinusoidal behavior) modulated by the influence of the Sun. Would you call the tides “sinusoidal?”
“but my point is that this wouldn’t happen if the planets caused the solar cycle.”
They do, and the ‘breakdown’ is how solar grand minima happen.
Guys, if you look back at the article on planetary influence on solar cycles, it is quite clear that planetary influence is a small influence that is contributing to and synchronizing an already cyclical phenomenon; thus the whole language of “resonance”. There may be, and surely are, other contributing factors, perhaps even more powerful ones.
+10, but I think you mean Be10.
Javier,
Post your data and code and then you can begin to engage in a discussion of your work.
I’m surprised this is not a condition of you being allowed to post here.
Anthony?
Javier’s article was on Climate Etc. , he was not “allowed” to post here. It was Wlllis who chose to discuss it here. Perhaps you would prefer that Willis did not discuss anything that does not show it’s works.
Willis seems to have dropped his excellent habit of providing his R code in recent posts too. Maybe you wish to talk to AW about not “allowing” W to post either.
The calibration curve (as I understand it) is simply to allow C-14 measurements to be used to date items based on known reference points from known items (historical or from dendrochronology).
The curve CANNOT be used to measure cosmic ray production. Why? as the C-14 is diluted by ordinary CO2.
CO2 (with C-12 and C-14) is incorporated into plants (or carbonate deposits, marine shells etc).
We know that CO2 goes up and down with temperature, but we do not know true values.
So, C-14 production from N-14 is based on cosmic radiation level, which varies.
The dllution of C-14 by normal CO2 also varies based in temperature and other factors.
The calibration curve can be used to date samples, but in itself cannot be separated into the two factors: C14 production, and CO2 dilution.
End of Lesson!
I’d like to add this to the Lesson …
Here’s the conclusion ..
“7. Conclusions
Meteorological effects in the form of cloud cover
changes, atmospheric temperature changes, surface pressure
changes, and strengthening or weakening of winter
cyclones show correlations with the measured or inferred
changes in the ionosphere-earth current density Jz in the
global electric circuit. These responses are consistent in
onset time, duration and sign of the response with the Jz
changes associated with both solar activity and internal
atmospheric forcing. The responses are consistent with
the production of electric space charge in conductivity gradients
at the boundaries of cloud and aerosol layers by the
flow of Jz through them, and consistent with the theory of
electrical effects on scavenging of cloud condensation
nuclei and ice-forming nuclei by droplets. However, the
overall effect for climate of the consequent variations of
cloud cover in latitude, altitude and cloud type remain to
be worked out.
The large changes in galactic cosmic ray flux on time
scales from decades through millennia produce changes
in Jz that have the potential to account for records of
long-term climate variations that correlate with cosmic
ray flux changes. Also, changes in temperature and humidity
in the thunderstorm-generating regions of land masses
at low latitudes modulate the current density flowing everywhere
in the global circuit. This may affect cloud cover
everywhere, especially with surface temperature changes
on the longer time scales.”
http://gacc.nifc.gov/sacc/predictive/SOLAR_WEATHER-CLIMATE_STUDIES/GEC-Solar%20Effects%20on%20Global%20Electric%20Circuit%20on%20clouds%20and%20climate%20Tinsley%202007.pdf
And the Geomagnetic Field is created by the “flow of liquid iron generates electric currents, which in turn produce magnetic fields. Charged metals passing through these fields go on to create electric currents of their own, and so the cycle continues. This self-sustaining loop is known as the geodynamo.”
http://www.physics.org/article-questions.asp?id=64
It truly is a electric universe.
As I’ve commented above. It is the deviations which ( may ) reflect 14C production. Temperature dependence of update may explain why the curve deviates into a different, messy and uncertain from during the last glaciation.
It’s interesting that he roughly linear section goes all the way back to 20-22ka BP . Maybe there is something to be inferred about climate or temp. dependance of d14C in that observation.
Sorry , a bit garbled.
Temperature dependence of uptake may explain why the curve deviates into a different, messy and uncertain form during the last glaciation.
Greg, you pointed out in your 5:11am comment above , d14C measurement of 400 per mil you can choose between 14ka and 43ka !
I asked the question, “how good are the dendochronologies, “ that are critical to the calibration curve. If the calibration curve is remotely correct for periods prior to 14ka, How is it physically possible for a 43ka wood sample to show the same d14C as a 14ka wood sample? The difference in years is 30,000, roughly five half-lives, or 2^5 = 32. It that holds, we must conclude that somehow the atmospheric [14C]/[12C] concentration ratio 43ka was 32 times richer in 14C than it was 14ka.
I am at a complete loss how such an 32x enriched 14C environment could be present 43ka. Even if we suppose that 14C production is eight times higher that present day (or 14ka day), you’d be forced to quadruple the amount of fossil 12C (depleted in 14C) into the atmosphere.
Personally, I think the easier explanation is the dendochronologies and other references for 14C calibration at least prior to 23ka are horribly wrong.
An interesting graphic would be the calibration curve with the currently accepted [CO2] ppm, partial pressure, curves, the sea-level curve, and the paleo mag curves.
Further on paleomag. From Wikipedia ( Leschamp event 41.4 +/- 2 ka.
Ok, how much greater 14C production?
Introduction
Earth is permanently bombarded by high-energy nucleonic particles – cosmic rays, which
produce nucleonic-muon-electromagnetic cascades in the Earth’s atmosphere. As a subproduct
of the cascade, radioactive isotopes can be produced, called cosmogenic nuclides.
Measurements of the abundance of long-living cosmogenic radionuclides in the atmosphere
and terrestrial archives (ice cores, tree trunks, sediments, etc.) form a very important
tool to study atmospheric processes and interaction between different reservoirs (see, e.g.,
books by Dorman [2004] and Beer et al. [2012]). This also offers a reliable quantitative
method to study solar activity on the long time scale [McCracken et al., 2004; Solanki
et al., 2004; Vonmoos et al., 2006; Muscheler et al., 2007; Steinhilber et al., 2012; Inceoglu
et al., 2015; Usoskin, 2013; Usoskin et al., 2014]. Most important for these purposes
are cosmogenic isotopes 7Be (half-life ≈53 days), 22Na (2.6 years), 14C (5730 years), 36Cl
(3 · 10^5 years) and 10Be (1.4 · 10^6 years), and many studies are based on these data.
4.1. Solar cycle in 10Be
As an example of an application of the approach presented here, we have computed the
deposition flux of 10Be in the northern polar region and compared it with the measurements
in the NGRIP ice core [Berggren et al., 2009]. The deposition flux was calculated
in two steps. First, a 3D time varying pattern of the isotope atmospheric production was
calculated using the yield functions presented here and applying the reconstruction of
the modulation potential φ based on data from the global neutron monitor network since
1951 [Usoskin et al., 2011]. For the atmospheric transport we used a parameterization by
Heikkil¨a et al. [2009, see Table 3 there], applying the mean latitudinal height profile of the
tropopause. Finally, we calculated the deposition flux of 10Be in the Northern polar region.
A 1-year delay due to the transport was applied. The calculated 10Be flux is in very good
agreement with the real data, especially for the period 1951–1970 (see Figure 8). A minor
(about 5%) discrepancy after 1970 is most likely related to the post-deposition effects in
firn and/or to the regional climate variability on annual-decadal time scale [Pedro et al.,
2006, 2012]. The good agreement with data validates the yield-function for 10Be.
https://arxiv.org/pdf/1606.05899.pdf
Summary
We have performed a new consistent and precise computation of the production of five
cosmogenic isotopes, 7Be, 10Be, 14C, 22Na and 36Cl, in the Earth’s atmosphere by cosmic rays. Computations were made by means of a detailed Monte-Carlo simulation by the CRAC model using a recent version of the GEANT-4 tool. The results are presented in the Supporting information in the form of tabulated yield (production) functions for a wide set of atmospheric depths. We provide, for the first time, a full detailed set of the altitude profiles of the production functions which makes it possible to apply the results directly as input for atmospheric transport models. Our results are in good agreement with most of the earlier published works for columnar and global isotopic production rates. Comparison of the computations with measured data of 10Be for the last decades and also for a period around 780 AD validates the approach also in quantitative terms.
https://arxiv.org/pdf/1606.05899.pdf
The question is whether ionizing radiation can grow? Research shows that it can. Will greater ionization in the past hastened evolution? Probably so.
In 1965 Paul D. Jose published his discovery that both the motion of the Sun about the center of
mass of the solar system and periods comprised of eight Hale magnetic sunspot cycles with a
mean period of ~22.375 years have a matching periodicity of ~179 years. We have
investigated the implied link between solar barycentric torque cycles and sunspot cycles and
have found that the unsigned solar torque values from 1610 to 2058 are consistently phase and
magnitude coherent in ~179 year Jose Cycles. We are able to show that there is also a
surprisingly high degree of sunspot cycle phase coherence for times of minima in addition to
magnitude correlation of peaks between the nine Schwabe sunspot cycles of 1878 through 1976
(SC12 through SC20) and those of 1699 through 1797 (SC[-5] through SC4). We further identify
subsequent subcycles of predominantly non-coherent sunspot cycle phase. In addition we have
analyzed the empirical solar motion triggers of both sunspot cycle phase coherence and phase
coherence disruption, from which we boldly predict a future return to sunspot cycle phase
coherence at times of minima with SC12 to SC20 for SC28 to SC36. The resulting predicted start
times ± 1 year, 1 sigma, of future sunspot cycles SC28 to SC36 are tabulated.
https://arxiv.org/ftp/arxiv/papers/1610/1610.03553.pdf
Presented at this week’s Living Planet Symposium, new results from the constellation of Swarm satellites show where our protective field is weakening and strengthening, and importantly how fast these changes are taking place.
The animation above shows the strength of Earth’s magnetic field and how it changed between 1999 and May 2016.
Blue depicts where the field is weak and red shows regions where it is strong. As well as recent data from the Swarm constellation, information from the CHAMP and Ørsted satellites were also used to create the map.
It shows clearly that the field has weakened by about 3.5% at high latitudes over North America, while it has strengthened about 2% over Asia. The region where the field is at its weakest – the South Atlantic Anomaly – has moved steadily westward and weakened further by about 2%.
In addition, the magnetic north pole is wandering east, towards Asia.
http://www.esa.int/Our_Activities/Observing_the_Earth/Swarm/Earth_s_magnetic_heartbeat
GCR flux is not constant, so magnetic field variation isn’t the only cause of fluctuation in C14 production.
GCR flux also depends on where the solar system is in its orbit of the galactic barycenter.
http://www.sciencebits.com/ice-ages
Whew, Dr. S. cycle on top of cycle on top of cycle.
Here’s a new cycle for you..lol..the “OUT OF CHAOS,” cycle.
Please, no comment needed.
Without a thorough understanding of of Earth’s geomagnetic field in time this all seems frugal.
However, a period of about 1,000 years fluctuations GCR are dependent more on solar activity.
t is more likely that the Earth’s magnetic field depends somehow from the Sun’s magnetic field (as evidenced by the long periods of changes in Earth’s magnetic field).
Wow! What an interesting discussion.
Willis,
Here are my comments, some have been stated by others above and I apologize for any repetition.
There are problems and traps with 14C data, we all know that. But, the ~2300 (or 2400, or 2450 or 2500) year cycle is well known and well documented without 14C data. Let’s not lose the forest for the trees here.
“Then they throw away more than three-quarters of the data, leaving only the chunk since 9600 BC as shown in Figure 4.”
The data they throw away (prior to 9600BC) is so corrupted by the geomagnetic oscillations, the Younger Dryas warming and radical sea level changes it shouldn’t be used. The main geomagnetic oscillation has a 7,000 year cycle (John Southon, 2002, GRL), this had to be removed from the 11,000 year dataset they did use. There is no problem with the procedure recommended by Clilverd, et al. that I can see. He was conservative as some recent work by John Southon (2002 GRL) has tried to extend the useful 14C data back 14.5Kyrs, but given the Younger Dryas good luck with that (see Roth and Joos, 2013). 14C data is very difficult to work with and it is easy to make major errors with it. But, used carefully, it is helpful. Regarding the ~2300 year cycle, we would not want to depend only on 14C data, but in combination with paleoclimate data and 10Be data it does help.
“subtracted a 7000-year cycle of unknown origin , and”
The 7,000 year sinusoidal cycle removed is a well-known and well documented geomagnetic oscillation (John Southon, 2002 GRL)
“ASSERTED that the remainder represents solar variations with an underlying 2,300 year period …”
This “assertion” is also well known and supported by multiple lines of evidence, including 10Be data. Paleoclimate data including geophysical, biological, ice raft data and glaciological data support a strong worldwide ~2300-year climate cycle. This cycle has been recognized by numerous studies going back to the original Bray article in 1968. That it exists is not in dispute that I know of. The 14C study by Clilverd is evidence, but it is not necessary to show the cycle exists. Neither is it necessary to show that it is a solar cycle. The 10Be data does that along with the worldwide glacier, paleontological and ice raft data. Other evidence suggesting a solar origin is found in an orbital oscillation of 2318 years identified by Nicola Scaffetta and others (Earth-Science Reviews, Nov. 2016). Historical records support the 1300 Wolf minimum, the 1500 Sporer minimum and the 1700 Maunder minimum. Clilverd’s comparison of these historical periods with 3 Bray (or Hallstatt) cycles ago in his Figure 2 is very impressive. So, using the term “assertion” is a bit over the top in light of all of the evidence gathered and published over the last 50 years on the ~2300 year cycle. I don’t think we can say the 2300 year cycle is proven or completely understood, but it is very well established in the peer-reviewed literature.
Your question regarding the removal of the linear trend (secular trend) in the data is a good one. No one knows about the very long solar trends, only that they are apparent in the data. It was removed to enhance the 2300 year cycle and study it. It is simply “detrending” the data, a useful practice. It was done by Wyatt and Curry to get their “Stadium Wave.”
How cosmic rays affect clouds and cloudiness is a subject of much debate right now. I don’t think we have figured this out yet, but this hypothesis is not relevant to the discussion of the Bray cycle.
This thread appears to be mildly alive, and I’ve completed some analysis using frequency domain analysis. (I really should be doing a wavelet decomposition, but I don’t have that library running yet in octave). So here goes.
?dl=0
?dl=0
Executive Summary: there are clear signals at ~1000 years and ~2400 years, and probably ~7000 years. I’m pretty picky about declaring valid signals in data but I’m satisfied it’s there.
I noticed in the data there’s an estimate of the error sigma so I included a measurement noise floor in the mix.
I did not investigate thoroughly the idea that the data is itself an autocorrelated process with 1/f noise spectrum. However if you mentally draw that line it’s pretty clear that the signals mentioned above are still above that kind of noise floor.
There’s a pile of other data (e.g. southern hemisphere, marine calibration curves) in the same directory as this data, I have not yet investigated to see if similar results will be obtained.
Also, I have no idea whether the signals are due to the method of creating this data. Not my area of expertise. I just do signal processing.
Source code and copy of the intcal13 data is here: https://www.dropbox.com/sh/8hchg7kw1ah4rkc/AACMdM0Ny9y3yYvJBCLUOBzHa?dl=0
uses code from: https://www.dropbox.com/sh/ej9eozfhyqgd0es/AACGnhc3yLzUDSXKBfZbLEEfa?dl=0
Ugh, horizontal units on 2nd graph are 5 years. Forgot to fix that.
Thanks Peter. I also noticed that a google search for the “Hallstatt cycle” turned up 154,000 items. Many of these were about the 2300-2500 year solar/climate cycle. One of them is this gem from the European geophysical society that does a power spectrum analysis of the data and has a very prominent peak at 2300 years, see figure 3. http://www.ann-geophys.net/20/115/2002/angeo-20-115-2002.pdf
They conclude that the 2400 year period is the dominant period and provide support that the cause is a solar cycle.
Abstract
We derive two principal components (PCs) of temporal magnetic field variations over the solar cycles 21-24 from full disk magnetograms covering about 39% of data variance, with λ=-0.67. These PCs are attributed to two main magnetic waves travelling from the opposite hemispheres with close frequencies and increasing phase shift. Using symbolic regeression analysis we also derive mathematical formulae for these waves and calculate their summary curve which we show is linked to solar activity index. Extrapolation of the PCs backward for 800 years reveals the two 350-year grand cycles superimposed on 22 year-cycles with the features showing a remarkable resemblance to sunspot activity reported in the past including the Maunder and Dalton minimum. The summary curve calculated for the next millennium predicts further three grand cycles with the closest grand minimum occurring in the forthcoming cycles 26-27 with the two magnetic field waves separating into the opposite hemispheres leading to strongly reduced solar activity. These grand cycle variations are probed by α-ω dynamo model with meridional circulation. Dynamo waves are found generated with close frequencies whose interaction leads to beating effects responsible for the grand cycles (350-400 years) superimposed on a standard 22 year cycle. This approach opens a new era in investigation and confident prediction of solar activity on a millenium timescale.
https://www.researchgate.net/publication/283862631_Heartbeat_of_the_Sun_from_Principal_Component_Analysis_and_prediction_of_solar_activity_on_a_millenium_timescale