Guest Post by Willis Eschenbach
In a recent post here on WUWT, someone yclept “Javier” has written about the Bond Rafted Ice data. In a comment, he said:
This is a frequency spectrum from Bond data. It shows the 980-year Eddy cycle, and the 2400-year Bray cycle I have written so much about.
Well, that looks like it establishes the existence of the 960-year cycle in the Bond data beyond any doubt …
Intrigued by that, and knowing next to nothing about the Bond data or the “980-year Eddy cycle” referred to by Javier, of course, I had to go take a look at it. Javier provided a link to the paper by Debret et al. he was discussing, entitled “The origin of the 1500-year climate cycles in Holocene North-Atlantic records“
The abstract of the Debret et al. paper says:
Since the first suggestion of 1500-year cycles in the advance and retreat of glaciers (Denton and Karlen, 1973), many studies have uncovered evidence of repeated climate oscillations of 2500, 1500, and 1000 years. During last glacial period, natural climate cycles of 1500 years appear to be persistent (Bond and Lotti, 1995) and remarkably regular (Mayewski et al., 1997; Rahmstorf, 2003).
Hmmm … however, I was interested in the data itself, so I dug around and found out that it’s available here.
I found out that what the paper analyzed was called the “stacked ocean record”. What does that mean? Well, they combined a record of hematite grains with two records of Icelandic glass from different drill cores and one record of detrital carbonate … add them together, divide by four, and presto! A “stacked” record.
So I did a Complete Ensemble Empirical Mode Decomposition (CEEMD) analysis of the stacked record. I used the “CEEMD” function in the R package “hht” for the analyses. To begin with, I found what the authors found about the “Eddy cycle”. Their comment was
… the application of a 1000-year filter to the composite series of IRD does not provide very conclusive correlation during 0–5000 years, …
Now, here is the periodogram of the CEEMD analysis:
This shows the strengths of the cycles in each of the empirical modes. It sure looks like there is a 960-year cycle in there in empirical mode C3, along with strong cycles at about 2300 and 7000 years, and a weaker cycle at around 1,300 years in length. And this is what we saw in the periodogram provided by Javier above.
But a look at the individual empirical modes gives a deeper understanding of the situation with the 960-year cycle. These empirical modes are the actual signals that when added together recreate the original raw data signal shown in the top panel below.
As the authors found, the putative 960-year cycle in empirical mode C3 only has significant strength in the earliest 6,000 years of data. On the other hand, it weakens and nearly disappears in the most recent 5,000 years. As I’ve said many times, this kind of appearance and subsequent disappearance of “cycles” is quite common in natural datasets.
More to the point, however, upon learning that it was a “stacked” record, my further thought was “Wait a minute, whenever you add different records you can get all kinds of artifacts from constructive and destructive interference”. So I did a CEEMD analysis of the four individual underlying Bond datasets. Here are the pairs of CEEMD graphs for each of the four individual datasets that make up the “stacked” dataset. In each case, as in the “stacked” data, the putative 960-year Eddy cycle is in empirical mode C3. First, the hematite stained grains dataset. Click on the graphic if you’d like a larger image.
Curious. In this one, there is no clear peak in C3. There’s still a peak at ~2,300 years, in empirical mode C5, but the 7,000-year cycle is much weaker. And there’s a wide peak around 1,400 years in empirical mode C4. This is very different from the stacked data.
Next, here’s the first of the two Icelandic Glass datasets.
In this one, the 7,000-year cycle is back … but there’s even less sign of the putative 960-year cycle in empirical mode C3. However, the previous cycle in C4 has moved down to a weak peak near 1000 years.
Then we have the second Icelandic Glass dataset, from a different drill core.
Once again, there is no 960-year cycle. It’s smeared out from 500 to 1,500 years, basically non-existent, but the 7,000-year cycle is strong. Overall, the two Icelandic Glass datasets are nearly identical, which increases the confidence in these results.
Finally, here’s the detrital carbonate data.
Most interesting. Almost no sign of the 7,000-year cycle, and once again there’s only a weak 960-year cycle. However, in empirical mode C4, there is a small peak around 1,500 years in length.
So … what have we learned?
Well, the first thing I learned is that the putative 960-year “Eddy Cycle” only exists in the “stacked” dataset. There is little sign of it in the four underlying datasets. It is an artifact of the averaging process was used to stack the four datasets.
Next, Debret et al. say that there is a 1,500-year cycle in the Bond data … however, although it appears to show up in the “stacked” data, it is only found in one of the four individual datasets. Again, this appears to be an artifact.
Next, the 7,000-year cycle is strong in the Icelandic Glass datasets, weak in the hematite data, and basically non-existent in the detrital carbonate data. Go figure.
Next, there is a cycle in all four of the datasets at somewhere between 2,200 and 2,600 years. This is the “Bray Cycle” referred to by Javier. Since it appears in all four datasets, can we believe that it is real?
Well, not so fast. Here’s empirical mode C5 for all four of the datasets:
At about 5,500 years BP, they all line up. And moving towards the present, the adjacent cycle is almost exactly 2,500 years long.
But the next cycle nearer to the present is quite different. For the first three datasets, it’s only about 2,000 years long … and for the detrital carbonate, a bit longer than that. Also, all of the cycles are disappearing as we get nearer to the present.
Going back in time from the 5,500-year peak, however, things get worse rapidly. The correlation of the data falls apart, and by the time we’re back to 10,500 years before present, it’s all over the map. The datasets are completely different. However, instead of decreasing in size as they did near to the present, they maintain their amplitude back to the earliest part of the record.
Note, however, that the two Icelandic Glass datasets (blue and red) stay in very tight lockstep over the whole dataset. This is not good since they are both given equal weight in the “stacked” dataset … meaning that the Icelandic Glass data gets counted twice and thus is given twice the weight of the other two datasets. This is bad practice because it distorts the end result.
So … is there actually a 2,500-year cycle as claimed by Debret et al.?
Well, it’s possible, but the evidence is far from clear. There’s only one 2,500-year complete cycle in the data, from ~ 3,000 to 5,500 years BP … but before and after that, the cycles vary in both length and amplitude in the four datasets.
I see this as yet another cautionary tale of expert analysis gone wrong. Here are the cautions, in no particular order:
• As Richard Feynman observed, “Science is the belief in the ignorance of experts”. I can’t tell you how many times I’ve looked at some paper like this one, written by experts, only to have it fall apart under closer examination.
• Periodograms and Fourier Analyses are limited in that they can be fooled by a signal that only appears in part of the record, and then disappears.
• What I call “pseudocycles” are quite common in nature. These appear to be real cycles, but over time they get larger, or get smaller, or disappear altogether, only to be replaced by some other pseudo cycle.
• The fact that four datasets all show say a ~ 2,500-year cycle does not mean that the cycles are in phase.
• Using “stacked” datasets can easily create artifacts through both constructive and destructive interference.
• Using two basically identical datasets in a “stack” of four datasets will overweight that data, leading to incorrect conclusions.
• Humans are very good at detecting patterns and cycles, even where none exist. For example, almost all cultures see “constellations” in random groupings of stars. I hold that this is a result of using our eyes to find predators—there is no penalty for seeing a pattern of stripes that is not there, but there is a huge penalty for not noticing the pattern of stripes that is a tiger. And as a result, we tend to see patterns everywhere, including cycles in natural climate datasets, even though they may not exist at all.
• This implies that all such claims of cycles in natural climate datasets need to be investigated very, very carefully. For example, if you are using a periodogram or doing a Fourier analysis, at a minimum it is imperative to divide the dataset in two and see if the claimed cycle exists in both halves. Or, as I have done, you can use a CEEMD analysis to investigate the nature and changes in the cycle visually. However, I see experts all the time do one Fourier analysis on a full dataset and declare victory …
My best wishes to you for all the good things—laughing with your family, walking in the rain, sunlight far-reaching on the sea, gentle breezes …
w.
MY USUAL REQUEST: Please, when you comment, QUOTE THE EXACT WORDS YOU ARE DISCUSSING so that we can all understand who and what you are talking about. In addition, it’s not possible to refute someone’s claim unless you quote it first. Note that while this request is polite, I am likely to get grumpy and say inappropriate things about your ancestry and personal habits if you repeatedly refuse to identify what you’re talking about.
In support of that, I’ve posted this graphic before, and I’ll post it again. Please structure your comments to keep them up near the top of the pyramid.
The graphic is based on How to Disagree by Paul Graham, which is well worth reading.
One thing I’d like to highlight is that in the linked article the author says (emphasis mine):
DH5. Refutation.
The most convincing form of disagreement is refutation. It’s also the rarest, because it’s the most work. Indeed, the disagreement hierarchy forms a kind of pyramid, in the sense that the higher you go the fewer instances you find.
To refute someone you probably have to quote them. You have to find a “smoking gun,” a passage in whatever you disagree with that you feel is mistaken, and then explain why it’s mistaken. If you can’t find an actual quote to disagree with, you may be arguing with a straw man.
Words to live by …
“yclept” in the first line?
Yep.
w.
And a delight to encounter! Thanks Willis, your erudition coruscates 😉
It’s good to see Middle English is not yet dead…
You also had me going on artefact v. artifact.
It appears that modern English has diverged towards a preference for the former, whilst the USA and Canada prefer the latter.
Apparently clepe is denigrated.
Willis, Do you have a link for the pyramid graphic? I like that!
billw1984 March 15, 2018 at 7:42 am
Sorry, no link. I got it so long ago … wait, hang on … OK, I found it here on Wikimedia Commons.
w.
OK, A new word for me so I’ll bite.
Yclept – called, past participle of clepe to call; to name
Old English: Clipian to speak, call
Still in use in modern Scots: Clype to tattle, to tell tales.
“yclept”
A great word for Scrabble!
Especially if the “y” falls on a double/triple letter square.
Oh yclept is the one reference that I DID understand but then I did read Chaucer as a younger man!
As a younger man, I used to enjoy Dickens…
…but I haven’t been invited to one for years (:-((
I just hope these predictive spell check algorithms do not find their way into driverless car software.
Willis, many of us have plenty of reasons to not out ourselves. Like, my goofy opinions could hurt future employment. In some cases, I fear people could start behaving badly as there are not many republicans in the Republic of Finland. I don’t know about Javier, but scientists who take part on contrarian blogs, can be seriously attacked. Please bear us.
Hugs, I understand the desire to be anonymous. And I made no attempt to out Javier. I just pointed out that it was a screen name because it is important to note, for reasons I explain below.
My own feeling on the matter is that if you can’t reveal your own views in public, you’re in the wrong job or in the wrong situation. I also think that people not revealing how they feel is destructive, because it gives the appearance that there is perfect agreement where no such agreement exists. It also prevents people in the minority from realizing that there are others out there who feel as they do.
This is more than theoretical to me, as I’m a cartoonist who has published my cartoons of Mohammed on the web. You can bet before I did it that I gave a lot of thought to doing it under an alias. But at the end of the day, I decided that I’d be damned if I was going to run scared and hide my identity.
So I published them under my own name, knowing that there is a chance that I could get killed for my cartoons … after all, I wouldn’t be the first cartoonist killed by Muslims for drawing the Prophet.
But I’m not going to live as a coward in fear and hiding. Unless we stand up to the oppressors, they’ve won, and no way I’m going to let that happen while I’m hiding behind an alias.
Next, if you post anonymously you can never claim your own ideas. My ideas are my own, and I sign every one with my sigil, “w.”. Good or bad, they’re mine, I own them. You, on the other hand, can never claim your own ideas, because you can’t show that you wrote them.
This gives me power and also keeps me in check, because unlike you, I know I can be called to account for my words even years after the fact … and it happens more often than you’d think. So I have to reflect, I have to take positions that I can defend.
In addition, it makes me responsible to my wife, my daughter, and any of my friends who read what I wrote. They can call me on any of it … so I pick my words cautiously.
You, on the other hand, can walk away from some inappropriate post, change your screen name, and never have to take responsibility for what you say. This leads to all kinds of inappropriate claims and actions.
And the same is true of “Javier”, whoever he may be. He can get on my case, make any outrageous claim he wish, and then disappear entirely.
This is the part that people posting anonymously don’t seem to understand—their words are inherently less believable because they take absolutely no responsibility for them.
People think this means I am opposed to people posting anonymously. I’m not. I’m just pointing out that there are big costs to posting anonymously, and those costs have nothing to do with me or my opinion.
w.
“Willis, many of us have plenty of reasons to not out ourselves. Like, my goofy opinions could hurt future employment. In some cases, I fear people could start behaving badly as there are not many republicans in the Republic of Finland. I don’t know about Javier, but scientists who take part on contrarian blogs, can be seriously attacked. Please bear us.”
As long as your show your work and do good work you are in no “danger”
sure there will always be people who slime you, look at how willis and I get slimed ( by opposing sides)
for being ‘self taught’ and not having the proper university pedigree.
Chances are if Javier used his real name he would take more care to double check at least ONE
of the charts he copies from other people.
Willis has clearly shown that the claim made about the cycle is bogus.
Will WUWT or Javier do a retraction? nope. But if it was Willis making a mistake using his own
name you better be damn certain he would correct the error or admit he was wrong.
I don’t have any problem with using my real name or providing general details about my background.
I don’t identify my current employer because I am not acting as a representative of my company.
In the case of Javier, I don’t think it’s really the anonymity that is the core issue, it’s the anonymity plus the personal attack. Personal attacks by people who hide behind anonymity is just cowardly. When it’s over an issue that has been rehashed so many times I’ve lost count, it’s also boring and a waste of time.
It’s fine for people to jump in with a comment they think relevant, despite it being discussed and understood for years, if they do it respectfully. We can’t all read every single post and comment.
However, if that jumping in is grounded in a personal attack that’s poor behavior and deserves to be put down hard in my opinion. That kind of comment isn’t science, it’s a lame attempt to discount and silence people without using science.
I took a somewhat middle ground, as I didn’t want my employer to google my name, and then get fired for having an unpopular opinion, since I am a consultant, and it could “insult” my potential customers.
Micro is my name, It’s just not how it’s spelled, but it is a fair point. My twitter account uses my full name.
Mosher says: “As long as your show your work and do good work you are in no “danger”” but that is not true. Several state climatologists and professors have lost their jobs because of their climate views. It is great if your employer is tolerant, but many people are not in that position. One can have grants turned down, papers not published due to the “wrong” views.
Is this the same Javier who wrote the long post at Climate, Etc. on “Modern Global Warming”, and has contributed there several times?
scraft1 March 15, 2018 at 6:55 am
Probably, but you’ve put your finger on one of the problems with posting anonymously … there’s no way to know.
w.
In practice I agree with Hugs. I use my first initial and not my full name so as web searches find my LinkedIn profile and work persona rather than my comments on controversial subjects.
This is a compromise. My name in both places is true. But linking the two is deliberately difficult. It’s just being cautious.
Who would employ a rabid Green or a staunch Sceptic? They might be disruptive to team spirit. Must I choose between vocation and freedom of speech?
And further more, if I were fully anonymous my behaviour would be the same, most assuredly.
You either have integrity. Or not.
Willis
Yes if like you one is an independently wealthy and kind of retired Pacific island-hopping gentleman of leisure, then sure one has the socio-economic status to let you say and write/draw what you want. But for us lesser mortals lower down the food chain with jobs to hold down and kids to support, anonymity retains its attraction. The freedom to say what you want is very dependent on sociobiology. Not everyone is free to indulge it. All are free but some are freeër than others.
ptolemy2 March 16, 2018 at 7:02 am
“Independently wealthy”? I like the sound of that. I’m actually retired and living on savings and Social Security, and my latest trip to the South Pacific was for work, as have been all but one of my trips there … and if you hear of any short-term jobs in the S. Pac. let me know.
Next, I published my cartoons of Mohammed on the web. I had to consider that action very carefully, because my socio-economic status doesn’t somehow magically protect me from the kind of Islamists who have killed people for doing that very thing. And believe me, I thought hard about doing it anonymously.
I’m not insulated from the risks of the world—instead, I’ve made a conscious decision to take my chances.
I understand that and have no problem with it. I’m just pointing out that when you post anonymously, there are costs as well as benefits. It costs you credibility because there is no need to stand behind your words. And this lack of accountability often leads to bad behavior, or at a minimum it allows bad behavior to go undiscovered. So, reduced credibility and ample opportunity for wrong-doing. Not a good position, but that has nothing to do with me and my socio-economic status. It’s the cost of posting from behind an alias.
Regards,
w.
I have learned so much over the many years that I have followed Anthony’s invaluable website, much if it from the gifted writer, analyst and world traveller that is Willis. But I can honestly say that this is the first post that has gone completely over my head. I have no idea what Willis and Javier are talking about. Not a clue. Can’t even get near it. And that after reading the whole thing from top to bottom. Could one of you gifted guys out there sum it up in a paragraph using words that half-decently educated folk (I promise, I am half-decently educated) can understand? Thanks in advance!
In very simplified form, you have five (5) sets of data as follows: 3, 5, 11, 11, 13. The first set of data shows a strong signal at the number 3 and so on. (Incorrectly) Trying to find a pattern in this data by adding 3+5+11+11+13 = 43. Then divide by 5 to get 8.6. Wow a signal at 8.6! Actually it would be wrong to say there is a strong correlation of these numbers at 8.6. All you did was the average five completely different numbers.
Further, if I told you the two number 11 data sets came from the exact same location, of course there is a good chance they will match. Including both “identical” data sets, just because you took more of them in the average just skews the average closer to the 11. As these are essentially the same data point one of the data sets should have been discarded to keep even weighting.
Lastly, you can really only add similar numbers. If your samples are made from different numbers you have to be very careful adding then averaging dissimilar numbers and what you intend to show with it. It would be ok to average 4 apples & 1 oranges, then say someone eats 5 fruit servings per day. It would be very wrong to average number of apples and number of cars, to then say people eat more apples if they have more cars.
Sorry, the last paragraph should have been 5 fruit servings per day.
Fixed … I hate typos.
w.
TY
The purpose of this kind of study is looking for long term periodicity where, if it exists, retreating glaciers aren’t that newsworthy nor entirely mankind’s “fault”. On the other hand, where no periodicity can be found then some non-periodic cause must exist which necessarily includes athropogenic greenhouse gases.
So there’s a bit of excitement when a new periodic function is discovered or believed to exist and amateur scientists attempt to replicate the outcome with varying degrees of success.
It’s physics, but not basic physics. If it was basic then everyone with even a high school education could arrive at the same conclusions based on the same observations.
John V. Wright, Many researchers over the past 50 years have looked into evidence of “cycles” or “oscillations” in the energy output of the Sun. These are investigated by using tools like Willis used in this post against available data, generally isotopes of beryllium and carbon that have a Solar or cosmic ray origin. Using frequency analysis, strong periodicities show up at about 980 years and 2450 years. Seeing this, they look for climate cycles that match to confirm the solar cycles. Evidence exists in glacial advances, biological records, human history (the “Roman” warm period, the Medieval Warm period and the Modern Warm period are a 1000 years apart for example), and in other proxies like Bond’s analysis of North Atlantic sediment cores. This and other data convinces Javier and I, but not Willis or Leif. Definitive proof eludes us, but the investigation continues. I actually think much of the data and analysis Willis provides supports our view, he thinks it disproves it, only time will tell.
Thanks, Andy. I’ll make you the same offer I’ve made to many others. Send me two links, one link to the very best paper, the one you think establishes most strongly the connection between the sun and some surface climate phenomenon, and the other link to the dataset(s) as used in the paper.
I’m more than happy to take a look and see if it is valid, or if like the dozens of other papers and data that people have submitted to me, it falls apart on close examination.
Your move …
w.
Sorry Willis, if one or two links would do it, it would be easy and no fun. Besides, you look at each line of evidence in isolation, that is with a microscope. Hardly scientific. Geological research requires the big view. Your microscopic view is valuable, brings us back to Earth, but in the early stages of discovery the larger view is most important. Eventually, when working on mechanisms the details will rule. Key point, it takes both.
Andy May March 15, 2018 at 6:27 pm
Andy, I’ve made this offer to a lot of people, some of whom have sent me two links and some of whom haven’t, and I have to give you big props. That is the most hilarious excuse for not putting your money where your mouth is that I’ve heard to date.
First, science “looks at things in isolation” all the time. It’s one of the main ways we do science—we divide and conquer. We pull the universe apart and we look at each part in isolation. You may not realize it, but we have things called “specialties” and “specialists” in science, who take just one small part and look at that particular line of evidence until their eyes glaze over. Of course we look at things in isolation.
Second, looking at things with a microscope is “hardly scientific”? I must have missed the memo ruling microscopes non-scientific .. man, that’s real knee-slapper.
But if you think only the “big view” is important, then surely you could provide a scientific study giving us the “big view” of the purported solar/surface climate correlation … but noooo. Your excuse is that that would be “easy and no fun” … say what? You refuse to provide the slightest scrap of actual evidence to back up your point of view because that would be too easy???
Finally, it appears you have no problem with Javier picking a line of evidence in isolation, say the evidence from ice rafting in a small segment of the North Atlantic, hardly a “big view”. But when I look at the same study, I’m the bad guy? You are missing your calling, you truly should be a comedian.
Like I said, I’ve never heard a funnier excuse. You’ve moved in one jump to the top of the list.
But sadly, like far too many other solar advocates you’re unwilling to back up your claims with even the tiniest bit of evidence. Got it. I certainly won’t bother asking you for facts again, my stomach can’t take laughing this hard more than once in a great while …
w.
Willis, you even quoted me as writing “it takes both.” Science is based on disagreement and argument, it’s a process, but you need to learn to disagree without being disagreeable. Javier and I have presented a mountain of evidence for our ideas, we have not proven anything, nor has anyone else, but we have presented the evidence. You can decide on your own whether the solar cycles match the climate cycles and whether or not it is likely the Sun dominates our climate, we have presented the evidence and the data that we could find. Your request for one or two papers that explains it all for you is spurious and insulting, our posts and bibliographies should be sufficient. The magic paper or data doesn’t exist because the debate continues and it will until the data and analysis is presented that shows us what controls the climate. I think, at least among those familiar with the data, it is pretty clear man does not control the climate with CO2 emissions. After 50 years of trying the CO2 effect still can’t be measured.
Do solar variations control the climate? Logic and proxies suggest this is a real possibility. Is the data we have presented in numerous posts definitive? Of course not. Have you or Leif disproved the idea? Hardly. Let the debate continue, but try and be more polite when you disagree.
Andy May March 16, 2018 at 4:34 am
Look, Andy, I get that you are unwilling to put your ideas to the test by providing a link to even one single study plus data that supports your claims. What I don’t understand is why. But the “why” pales in the face of the fact that you don’t want me to look at whatever you think underpins your certitude.
And as a result, that whole comment of yours is just another attempt to justify you running for the door when I made a simple request for a single study and its associated data. However, the fact that you are unwilling to have someone actually examine the basis of your beliefs has nothing to do with whether I’m polite or not. I didn’t point and laugh until you’d refused to put your money where your mouth is, at which point you richly deserved it.
I truly don’t understand this. Surely you must have one paper that you firmly believe shows the sun/surface climate connection, no? So why this sudden coyness? You’re willing to go on and on about your beliefs, but you are totally unwilling to have them examined closely. What’s up with that?
Instead, you make excuses that microscopes are not scientific and I’m not looking at the “big picture” … Andy, I’m not looking at any picture, because you refuse to reveal just which picture you think will stand up to being looked at.
And no … you can’t blame your reluctance to face the music on whether or not I’m being polite …
w.
PS—You say:
“Do solar variations control the climate? Logic and proxies suggest this is a real possibility.”
Actually, since the variations in TSI are less than 1%, logic suggests that this is possible but doubtful. And the fact that despite literally centuries of looking there is still no bulletproof smoking gun showing the putative connection also logically suggests that the possibility is low …
You also say:
As a scientist, you surely know that one cannot prove a negative like “The minor sunspot-related solar variations don’t have a visible effect on the climate at the surface”. Can’t be done. So your question holds no meaning, it’s just an attempt to denigrate Leif’s and my contributions to understanding the situation.
And since one cannot prove a negative, since we can’t show it’s false, then all Leif and I can do is to examine the claims that it is true … except you’ve refused to provide what you see is the strongest claim for us to examine.
Funny how that works …
Than you Andy, that has helped my understanding a lot. About those periodicities, would they not have to be correlated with Milankovitch cycles to get the full solar picture?
John Voight, Javier has correlated the Milankovich cycles to the periodicities as best he can. The post https://judithcurry.com/2017/07/11/nature-unbound-iv-the-2400-year-bray-cycle-part-a/
Figure 55, compares obliquity to the transition from the Holocene Climatic Optimum to the Neoglacial.
Willis, I’ve already said there is no one study plus data that definitively supports our claims, the data does not exist. The best description of our hypothesis is in our posts, which presumably you have read.
That statement is absurd in the extreme. We have put up numerous posts for others to examine, supplied complete bibliographies and mountains of data. We have definitely “put our money where our mouth is.” You are flat wrong.
I realize you have never worked on a scientific project, but most of the projects I’ve worked on are just like this one. First you study the literature to understand the previous work and analyze it. You try and put up a synthesis of it for discussion, not insults or spurious demands for one simple paper, they don’t exist at this stage. Later you try and come up with a mechanism or idea to test your best hypothesis. If it is successful the idea survives, if not, back to the drawing board.
The simple “one” paper is years away from now, if it ever exists. Certainly, you have not disproven the idea, nor have you proposed an idea to explain the data Javier and I have presented, this is the key weakness of your argument, Leif’s as well. As I said your comments are spurious, trivial and mostly irrelevant. No one cares about the temperature trends in central England in a global context – that is a dot in the ocean. Your criticisms are many years too early, the solar/climate connection investigation is just getting started.
Another point, I agree that TSI changes are probably too small to have an effect. TSI is not the only relevant solar property that is changing. The solar magnetic field changes a lot and so does UV. The Sun is a variable star and it is not the only variable star in the universe that never changes.
The solar magnetic field changes a lot and so does UV
The solar magnetic field as such [10,000 times smaller than the Earth’s magnetic field at the surface] is extremely unlikely to have anything to do with our climate. Variations of the magnetic field show up as variations in TSI [and in UV which is a tiny part of TSI] which is where the energy is. Indirectly, the field also controls the modulation of cosmic rays. All of these solar indicators vary with the sunspot number. The variation of the sunspot number and of the climate for the time where we have good data are very different.
Andy May March 17, 2018 at 4:59 am
Thanks, Andy. I had no idea that you had said somewhere that there is no study that definitively supports your claims. That claim, that no other scientist has published a good study that supports your work, changes everything. So let me ask you this, since I’m quite unfamiliar with the work of whatever group you are referring to as “your”, as in “your work”:
Who are the other scientists associated with “your work”? Is it just Javier, or are there others?
Which of your posts do you think is the strongest, the most bulletproof, the best of the lot, the pick of the litter? A link to that post, and a link to the data used if it is not in the post, would be wonderful.
As I said, your admission that there are no scientific studies that definitively support your claims is a game-changer. My apologies, I see that you weren’t naming the best study that supports your views, not because you were unwilling to do so, but because no such study exists.
Let me start by saying that claiming that “no one cares about the temperature trends in Central England” is simply not true. I can guarantee that if there were a clear sunspot effect on the CET, people would be trumpeting it around the world as evidence that the sunspot/climate theory is correct, and indeed it would be evidence of that … but since it doesn’t support your claims, you diss it. It’s just one more in a host of data points showing no solar effect on surface weather, but it happens to be one of the longest datasets we have.
Next, the idea that “the solar/climate connection investigation is just getting started” is laughable. This whole area has been closely looked into since at least the early 1800s, when William Herschel wrongly claimed that sunspots affected wheat prices. As evidence of the effect of Hershel’s claims on the field, a search on Google Scholar for “sunspots and wheat prices” yields over 5,000 results, and a search for “sunspots and temperature” yields over 33,000 results … so no, the investigation is not “just getting started”, it’s been going on for two centuries.
More to the current point, I’m quite unclear what you are calling “the data that Javier and I have presented”, as I have never seen you both listed as authors on a post. And I’m also not clear what you are calling “the idea” when you say I have “not disproven the idea”. (And in any case, I’m not under any obligation to “explain” that or any other data, that’s a totally separate question from whether your explanation is valid … but I digress.)
Of course, this is likely just my lack of familiarity with your work. So as I requested above, if you’d please provide a link to whatever post you and Javier wrote that you think is the most solid and bulletproof, plus a link to the data used if it doesn’t exist in the post, that would at least give us a starting point for discussion.
I can do no better than quote Leif on this subject:
lsvalgaard March 17, 2018 at 7:09 am
The key idea in there is that all of these solar phenomena vary in lockstep with the sunspot number … so if we don’t see the imprint of that sunspot variation in a given surface dataset, it means that it is highly unlikely that ANY solar phenomenon is affecting the dataset, be it solar wind, UV, TSI, or anything else.
And the good news is that the sunspot variation is quite unique, it has a very distinct shape and signature due to its changes in both amplitude and period. Which is why I’ve looked for that signature in everything from sea levels to river flow to cloud cover to a whole host of temperature datasets to tree rings to lake levels to 10Be levels to Norwegian child mortality data, of all things … and so far I’ve found nothing.
Now, as you point out, this doesn’t “disprove your theory”. Absence of evidence is not evidence of absence. I’m just pointing out that to date, despite looking at dozens and dozens of datasets, many of them recommended by folks who believe in the solar/climate connection, I haven’t found it.
Best regards, Andy. I completely retract and sincerely apologize for the harsh comments I made. I had not realized that there are no papers that definitively support your claims, so you couldn’t point to one. My bad.
In lieu of that, I await your link to what you think is the best post from you and Javier.
My best to you,
w.
And then there is this:
https://www.iau.org/news/pressreleases/detail/iau1508/
Leif,
[Willis, I think this addresses your comment as well]
I’m aware that this is your opinion and respect it, but others who have studied the data as deeply as you disagree with your view. They believe the solar magnetic field variations can have a large effect on Earth’s climate and weather (Svensmark and Svensmark, in these pages recently ( https://wattsupwiththat.com/2018/03/16/an-interview-with-henrik-svensmark-cosmic-rays-clouds-and-climate/ ).
Solar UV is very energetic and can vary in strength rapidly (at the Earth) by 8% or more and it appears to affect the weather ( http://joewheatley.net/cold-winter-sun/ and Ineson, et al., Nature Geoscience, 9 Oct. 2011). The solar magnetic field changes affect the number of cosmic rays that strike the Earth, weather is clearly affected in some fashion (Prof. Haigh, Feb. 2011, see link below), climate??
Further, UV and cosmic rays can penetrate the ocean surface to great depths. I’m not sure changes in the atmosphere, whether from CO2, TSI, UV or cosmic rays have much of an effect on climate, the atmosphere’s heat capacity is only 0.07% of the heat capacity of the surface. 99.9% is in the oceans, so for long term effects it always seemed to me we should be concentrating on the oceans. How quickly is solar energy going into the oceans? How deep is it penetrating? These factors are not getting as much attention as they deserve IMHO.
I also do not think estimating TSI or counting sunspots will get us very far from the standpoint of climate, these are very general measurements, not very accurate, and if they correlated with climate with any degree of accuracy we would not be having this debate. Further, the satellite measured TSI data is woefully inconsistent and the differences between satellites is many times the estimated solar variation ( https://www.imperial.ac.uk/media/imperial-college/grantham-institute/public/publications/briefing-papers/Solar-Influences-on-Climate—Grantham-BP-5.pdf ). We need long term accurate granular data on solar activity, by wavelength and magnetic field strength and details on the effect the Sun has on Earth’s magnetic field.
In summary, our focus on the atmosphere is OK for weather predictions of a couple of weeks, but not very important for climate. The oceans pick up most of the solar energy that reaches the surface, the shorter the wavelength the deeper the penetration, how is this thermal energy dispersed? How long before it reaches the surface and is passed to the atmosphere? What affects the timing? Solar activity correlates with cloudiness, why? How?
In short, I find your opinion interesting, as well as other opinions of Sun’s influence on climate, but I consider the jury still out on the subject – and data is needed to resolve the debate. That said, solar variability clearly affects our weather, we can see that in our short lifetimes. TSI? Who knows, I don’t think it has even been measured accurately, frankly. And what is the true relationship between sunspots and solar variability?
I’m aware that this is your opinion and respect it, but others who have studied the data as deeply as you disagree with your view
This is not my view, but the official sunspot/group numbers.
there are a few die-hards who cling to the old numbers because they are wedded to a sun-climate connection.
TSI? Who knows, I don’t think it has even been measured accurately, frankly. And what is the true relationship between sunspots and solar variability?
TSI is measured with great accuracy. the various disagreements between instruments have been resolved and the record [to the accuracy needed for sun-climate studies] is firm back to the 1970s.
The [revised] sunspot number is a very good representation for solar variability. The agent responsible for all observed solar variation is the sun’s magnetic field and we find that the relationship between sunspots and the magnetic flux is very strong and not in doubt:
http://www.leif.org/research/EUV-F107-and-TSI-CDR-HAO.pdf
[the calibration problem with the SORCE TIM record is too small to affect the conclusion]
Leif, Thanks for the terrific powerpoint. Lots of good data analysis in there. I believe your EUV to sunspot correlation and I appreciate that the Sun’s magnetic field drives solar variability. I don’t have your confidence in the accuracy of the TSI record however. It could be that neither sunspots nor TSI are the appropriate proxies to compare to climate, they are just what we can see and try to measure today. You know how it goes, when you have a hammer, everything looks like a nail.
I don’t have your confidence in the accuracy of the TSI record however. It could be that neither sunspots nor TSI are the appropriate proxies to compare to climate, they are just what we can see and try to measure today [..] when you have a hammer, everything looks like a nail.
The TSI record is very good. What do you build your lack of confidence on? If you knew how the instruments work and how the data is reduced, you would have very high confidence. The disagreements are so small that they don’t affect the use of TSI as a climate driver. But perhaps you don’t know.
And there is a ‘hammer’ here: the magnetic field.
If the sun’s magnetic field [and things derived from that] is not an appropriate measure, then what do you build your confidence on sun-climate relations on? There is no other.
Andy May March 17, 2018 at 1:35 pm
Andy, I fear I don’t understand this. You claim that the solar variability has an effect on surface weather/climate. Now, we know that the various variable solar phenomena (e.g. TSI, EUV, UV, heliomagnetic field, solar wind) all vary in close agreement to the sunspots.
So if, as you say, sunspots are not the proper measure of solar variation, and all of the other variable solar phenomena are closely related to sunspot variations … just what are you proposing to use to determine if your claim is true?
w.
This image doesn’t really need an explanation:
It is from Haigh, Feb. 2011
Another try:
you are confusing [deliberately or involuntarily] accuracy and precision [or absolute and relative calibration]. The absolute calibration is VERY hard [but also not important because TSI is so large]. What matters is the relative variation [i.e. the variation over time – not between instruments] and that is much better than the absolute calibration. We finally found out that the early instruments had a construction defect that allowed extra light to scatter into the instrument. This is easy to correct for and has been done in the bottom panel of the graph. The issue is ‘degradation of the sensors’ due to the harsh conditions in space. This is usual done by having several sensors where one [or more] is only exposed to sunlight a few minutes every week [say] to minimize its degradation. The latest instrument recently launched to be deployed on the Space Station will allow instrument to be returned to Earth for End-to-End calibration. bottom line: The TSI data is VERY good.
Sorry, I can’t get the image to appear, but the link should work OK.
Leif and Willis, I think it is likely that solar variability (including the Earth’s orbit, obliquity, etc.) account for much of climate change. I do not think we have any understanding of how that works. All we really have are a bunch of proxies that appear to correlate. This includes our TSI “measurements,” which are not measurements at all, but a “PMOD” composite, stitched together with guesswork. It also includes sunspots which relate to some portion of solar variability. It’s all correlation, no real mechanism and it is all qualitative and not quantitative. It will be a while before we get there. This argument is useful, but the longer it goes, the less we seem to know. The real problem is “climate change” is supposedly caused by a 2 W/m2 change. Look at the plot above, the satellite measurements vary almost 15 W/m2! How will we ever be able to measure that accurately?
This includes our TSI “measurements,” which are not measurements at all, but a “PMOD” composite, stitched together with guesswork.
Not at all. PMOD is a composite of two real measurements by two different instruments on the same spacecraft.And PMO is just one of the several series used in making the latest composite. Your comment betrays profound ignorance of the situation.
It also includes sunspots which relate to some portion of solar variability.
As pointed out, sunspots are a very good indicator of all the parts of solar variability.
It’s all correlation, no real mechanism and it is all qualitative and not quantitative. It will be a while before we get there.
Not so. TSI is a direct measurement of the total incoming solar radiation. EUV reconstruction is completely based on well-understood physics, completely quantitative. Perhaps what you have is all correlation with no understanding. What we have is then much better than that.
[Formatting fixed. -w.]
Andy, I fixed your image above. The URL for the image needs to be alone on its own line for it to appear.
You say:
Andy May March 17, 2018 at 2:41 pm
Handwaving. Please provide links, bearing in mind that you’ve already dismissed these studies by saying that you know of no one study that supports your claims.
Guesswork? Your own image says it is “obtained by inter-calibration of the data from the individual instruments”. The scientists didn’t just stand back and “guess” what the values should be, that’s an insult to their work.
This is not clear. WHAT “also includes sunspots”? And given your claim that sunspots only “relate to some portion of solar variability”, could you please be so kind as to point out which solar variations of interest in this discussion do NOT relate to sunspots?
Again, that’s not clear. WHAT is “all correlation, no real mechanism”?
Sorry, not buying that for one moment. The amount of knowledge about the purported sun/climate connection is not going down just because people disagree with you …
Huh? Look at the graph. NONE of the satellite measurements “vary almost 15 W/m2”. Despite different baselines, they all agree that the variation is on the order of 1 W/m2. This agrees with the CERES data and the data from the SORCE satellite … and there’s more bad news. You’re not using the same units for solar and CO2. The 2 W/m2 change from CO2 is the 24/7 average. The change in the TSI on a 24/7 average is on the order of 0.25 W/m2, less than a tenth of the CO2 figure you’ve given.
Regards,
w.
Leif, precision in this case would apply to the variation seen in one record. Accuracy refers to the long term (baseline) change. The PMOD composite has to assume a baseline and in tis case a flat baseline is assumed. Thus, the reconstruction by moving the records, with a spread of 15 W/m2 has no long term significance. Perhaps the new instrument will tell us what the actual accurate solar output is. Right now we don’t have a clue.
Willis, I meant sunspots are a proxy, which they are. Given the TSI measurements in the graph, TSI could be varying over 15 W/m2. Assuming a flat baseline and sticking them together doesn’t change that.
Leif, precision in this case would apply to the variation seen in one record. Accuracy refers to the long term (baseline) change. The PMOD composite has to assume a baseline and in this case a flat baseline is assumed.
Betrays deep ignorance.
First of all, PMOD is not the only composite. Second, there is no assumption of a constant baseline. Two overlapping records are compared and the difference between them is used to calibrate one to the other.
Here is more about the PMOD reconstruction:
http://www.leif.org/EOS/TSI-Uncertainties-Froehlich.pdf
One thing he discovers is that TSI depends on the sunspot number to the power of 0.7.
Willis, this is incorrect, think about it.
Andy May March 17, 2018 at 6:40 pm
Here’s what I think you are answering …
I assume you were responding to this interchange, although it’s not clear, since you’ve neglected to quote whatever it was you’re talking about … in any case, I’m still interested in the answer to my question.
Seriously? You actually believe that TSI could be varying over 15 W/m2?? Here’s your graph again:
The only difference between the various TSI measurements is the baseline. As Leif pointed out, we know why one of them is ~ 15 W/m2 higher than the rest, because of light leakage that falsely increased the data.
However, nowhere in there is there any single instrument that shows a ~15 W/m2 variation … nor have I EVER heard anyone claim that TSI has ever varied by 15 W/m2 in the modern era. That claim is absolutely unsupported by anything I’ve ever seen.
Andy May March 17, 2018 at 6:43 pm
You’re right. 0.25 is not “less than a tenth” of the 2 W/m2 change, it’s slightly more than a tenth of the value. Not sure why you think that is important … as to the calculation, I stand by what I wrote. The 2 W/m2 is indeed a 24/7 average. The ~ 1W/m2 change is an instantaneous TOA value. You can’t compare the two.
If you meant something else, you’ll have to point it out.
w.
Something people forget is that TSI is an exceeding precise measurement of the distance to the Sun [corrected for tiny effects of relativity]. Here is the variation through the year for 2004-2014 (one cycle):
http://www.leif.org/research/TSI-through-a-year.png
There are 11 curves plotted . They all fall on top of each other, except for tiny wiggles when very large sunspots cross the disk. The annual variation of TSI is some 70 times larger than the variation due to the solar cycle.
This variation can also be used to calibrate the various instruments against each other.
For amusement, here is my investigation of the question: “is TSI different on Fridays?”
http://www.leif.org/research/TSI-SORCE%20Friday%20Effect.pdf
We can now account for variation of TSI [not due to the Sun] of the order of one per million.
Leif and Willis,
OK, lots to read and look into. Too much for comments, I think. I’ll try and digest all of this and do some more reading and put together a post specifically on solar variability, pro and con, and we can continue the discussion there. There are a variety of opinions, very little definitive data, if any, so it should be interesting.
Andy May March 18, 2018 at 4:04 am
Thanks, Andy, stay well. Although I may often disagree with you, your contributions are always appreciated.
w.
“MY USUAL REQUEST: Please, when you comment, QUOTE THE EXACT WORDS YOU ARE DISCUSSING so that we can all understand who and what you are talking about.”
‘someone yclept “Javier” has written’
Yeah…why this compulsion to insinuate Javier is not his real name…not the first time you have done this either. I like reading both your posts’, so why denigrate him by calling him out on his name. Obviously he is in academia (probably recently graduated) that will likely persecute him if they ever find out his real name. We are all skeptics here and sometimes some people need anonymity.
Lay off on the theatric’s, eh? Except for your original witty stuff…
Thanks, Earthling. I said he was “yclept Javier” because at the top of the post it said it was written by Andy May. Plus see above for a fuller discussion.
w.
Willis,
Andy May is a regular contributor with editor privileges. He frequently “guest authors” posts by Javier and Renee Hannon.
Javier and Renee are the actual authors of the posts. English is not Javier’s primary language, so Andy provides some assistance.
Andy is a very real petrophysicist. Renee is a very real geologist. Javier clearly has a geological background, probably of an academic nature… Which is probably why he opts for anonymity.
Thanks, David. My apologies for the insult to Andy May, I’d hate to confuse anyone with Javier. Regarding why I said “yclept ‘Javier'”, I merely wanted to point out that as I mentioned above, in addition to inherent benefits there are inherent costs to posting anonymously.
Sorry for my misunderstanding, thanks for clearing up my confusion,
w.
From the South Pacific to climate topics I have so enjoyed your comments the last few years that I read everything you write. I have always thought ad homenin comments were negative but is it possible that they could be flattering. You Mr. Eschenbach are no ass hat but rather Top Hat.
Thanks for the kind words, Ben. If you’re not reading it you might also enjoy my own personal blog, Skating Under The Ice, as well as my comments on Twitter @Willis Eschenbach.
w.
I much appreciate your post, Willis, as always. Here, in particular, you draw attention to cycles than come and go – part of nature. Please take a look at the paper below – the link gives open access, and all the data is available in Supplementary Materials. I had meant to write a post to wuwt and commentary, but have been traveling these past two months and Jackson Davis also. We would appreciate a thorough criticism! Jackson Davis has examined every data point in the time series chosen (226 kyr) and mainly used cross correlation analysis, supplemented by spectral analysis – in the latter case, we being aware that dealing with a non-stationary time series, Fourier analysis has limitations. Reviewers had insisted we did our own spectral analysis even though we had referred to previous such analysis on the same data – and we confirmed their results. What we find is that the period of the centennial cycle is contracting and amplitude increasing toward the end of the data set (especially during the Holocene). We were asked late-on to do a CEEMD as well as wavelet analysis, but this would have required more resources than we had for this study, so any such that you might have time to do, would be of great interest.
with much appreciation for your posts.
http://www.mdpi.com/2225-1154/6/1/3/
The Antarctic Centennial Oscillation: A Natural
Paleoclimate Cycle in the Southern Hemisphere
That Influences Global Temperature
W. Jackson Davis 1,2,*, Peter J. Taylor 1 and W. Barton Davis 1
1 Environmental Studies Institute, Santa Cruz, CA 95062, USA; ethos_uk@onetel.com (P.J.T.);
wbartdavis@yahoo.com (W.B.D.)
2 Division of Physical and Biological Sciences, University of California, Santa Cruz, CA 95064, USA
* Correspondence: JacksonDavis@EnvironmentalStudiesInstitute.org or jacksondavis@earthlink.net
Received: 5 October 2017; Accepted: 3 January 2018; Published: 8 January 2018
Abstract: We report a previously-unexplored natural temperature cycle recorded in ice cores from Antarctica—the Antarctic Centennial Oscillation (ACO)—that has oscillated for at least the last 226 millennia. Here we document the properties of the ACO and provide an initial assessment of its role in global climate. We analyzed open-source databases of stable isotopes of oxygen and hydrogen as proxies for paleo-temperatures. We find that centennial-scale spectral peaks from temperature-proxy records at Vostok over the last 10,000 years occur at the same frequencies (2.4%) in three other paleoclimate records from drill sites distributed widely across the East Antarctic Plateau (EAP), and >98% of individual ACOs evaluated at Vostok match 1:1 with homologous cycles at the other three EAP drill sites and conversely. Identified ACOs summate with millennial periodicity to form the Antarctic Isotope Maxima (AIMs) known to precede Dansgaard-Oeschger (D-O) oscillations recorded in Greenland ice cores. Homologous ACOs recorded at the four EAP drill sites during the last glacial maximum appeared first at lower elevations nearest the ocean and centuries later on the high EAP, with latencies that exceed dating uncertainty >30-fold. ACO homologs at different drill sites became synchronous, however, during the warmer Holocene. Comparative spectral analysis suggests that the millennial-scale AIM cycle declined in period from 1500 to 800 years over the last 70 millennia. Similarly, over the last 226 millennia ACO repetition period (mean 352 years) declined by half while amplitude (mean 0.67 C) approximately doubled. The period and amplitude of ACOs oscillate in phase with glacial cycles and related surface insolation associated with planetary orbital forces. We conclude that the ACO: encompasses at least the EAP; is the proximate source of D-O oscillations in the Northern Hemisphere; therefore affects global temperature; propagates with increased velocity as temperature increases; doubled in intensity over geologic time; is modulated by global temperature variations associated with planetary orbital cycles; and is the probable paleoclimate precursor of the contemporary Antarctic Oscillation (AAO). Properties of the ACO/AAO are capable of explaining the current global warming signal.
Peter, thanks most kindly for the invitation to look at your work. I’ve just spent two hours extracting your data from the PDF … grrrr, it’s most difficult to get from that PDF to a nice spreadsheet. However, I’ve finally gotten there, the spreadsheet is here in CSV format.
My first question has to do with the individual data lines with the TOC350V number containing the letter “a” (I’ve replaced the “a” with “.5” in the spreadsheet data so the column is entirely numeric for easier processing.
How do these differ from the regular data?
Next, the line labeled “24a” has a start date of 4690BP, a peak date of 4552BP, and an end date of 4690BP … one or more of those must be in error.
Answers to those questions will allow me to move forwards.
My best to you,
w.
PS—I reformatted your comment to remove the pesky line breaks in the Abstract.
A couple more errors. The data with the peak time of 195940 is listed as “2”, far too large. Also the peak data in line 219 is listed as “4.8.0”, presumably 4.80.
w.
Willis, Is your last graph experimental mode c3? or c5?
I don’t follow your logic on the last part of the piece. Isn’t the Bray cycle related to the c5 signal?
Thanks, sebmagee. You’re correct, it’s C5. I’ve fixed it.
Much appreciated,
w.
There are lots of climate related cycles that are quasiperiodic. Glaciations/interglacials are an example. Sunspots are another. It looks like an oscillation but the period varies.
This paper is clear that the 1500 year signal evolves with time.
It then goes on to use wavelet analysis. That’s probably valid (I’m not a mathematician) and probably does show that there’s ‘something there’. For sure it’s possible to use Fourier Analysis in a manner that disappears a signal. We should never forget that an FFT is basically a filter and is often used as such.
Having said the above, I think the authors of the linked paper have squeezed too much out of the data (but, again, I’m not a mathematician). It’s way too easy (using Matlab or R) to subject a data set to a number of different analyses until you get something that looks significant.
Thanks, Bob. I’m not talking about something that is quasiperiodic, like say sunspots. The ~11-year sunspot cycle never disappears. In addition, the period varies around some certain value, for sunspots about 11 years.
I’m talking about what I call “pseudocycles”, which often exist for a while, and then disappear entirely, or change to another cycle length. Here’s an example:
While it seems like there is a correlation, a closer examination shows at the start of the record, sunspots and sea level moved in opposite directions. Then for most of the time, they were in phase. In both those cases, sunspots were leading sea level, suggesting the possibility that sunspots might affect sea level … except in opposite directions at different times. And in addition, in about 20% of the data, the sea level moved first, followed by the sunspots, suggesting the possibility that at times, the sea level might affect the number of sunspots … you can see the problem.
And yes, you can use wavelet analysis. However, I find CEEMD to be more intuitive. In part because the CEEMD analysis divides the underlying cycles empirically, based on the shape and form of the data itself. I also see people cherry-picking wavelet analyses, circling some allegedly “significant” cycle in some section of the data and claiming that it is important. See the paper you linked to for examples …
Also, in general, I see no attempt to correct wavelet analysis significance regarding the “Bonferroni Correction”. This is the correction needed due to repeated trials, or the use of subsets. For example, if I throw two dice, the odds of “snake eyes” is one in thirty-six. So if I start by specifying I’m gonna roll snake eyes and I do throw it on the first roll, that’s unusual enough to be “statistically significant” at the p<0.05 level.
But if I throw the pair of dice ten times, I have one chance in four of throwing snake eyes, so that wouldn't be unusual.
But when you do a wavelet analysis you are looking at all frequencies over all the times … I wouldn't even know how to correct that, but it's clear that we should expect something in there somewhere to be statistically significant at the p<0.05 level. Remember, we expect something with a p-value of 0.05 to occur by chance one time in 20 …
Best regards,
w.
I said:
You said:
We’re both saying:
🙂
Willis,
This graph has bugged me for years. Before I make any comments I’ll let Steve McIntyre speak first:
“The maxima and minima of the solar cycles seem to match the fluctuations in sea level rise rather uncannily. While the resemblance is impressionistic (I don’t have a digital version of Holgate’s series), offhand, I can’t think of any two climate series with better decadal matching. I think that this resemblance is pretty obvious.”
https://climateaudit.org/2007/02/11/holgate-on-sea-level/
When you originally debunked this, I think your main argument was a very low R2 correlation. But R2 assumes a linear relationship. A low R2 can simply mean that, if there is a real correlation, it is more complex, and it cannot be captured by any linear R2 analysis. I found that the Internet is littered with warnings about the improper use of R2.
You also point out that, in the graph above, there are details that do not fit. Yes, in some cases the sea level starts to increase before the sunspots. But of course it’s not sunspots that would drive sea level rise. It would be more likely solar magnetic strength, which may well occur before an increase in sunspots.
But here’s the problem: despite some detailed mis-matches, the overall fit in the shape of the sunspot and sea level graphs is remarkable. If you don’t agree with that, take it up with Steve McIntyre!
Clearly, one cannot expect different data series to line up perfectly. This is climate, after all.
Several years ago I tried a practical test. I wrote a program to generate random data (drunkard’s walk). I ran it maybe a hundred times and compared it with the sea level data. Not a single run created an apparent correlation even remotely close to the graphs above. This suggests the probability that there is no connection is no more than 1 in 100. I suspect that if I ran it 1000 times the result would be similar.
I hope you agree that, at the very least, there is a strong *apparent* correlation. If so, are you saying that is purely random?
********************************************************************
One other thing. Very recently you posted a graphic showing solar activity, NH temperature, Mediterranean temperatures, and some ice and glacier data.
The first thing that struck me was this: if you exclude the first three, all the others don’t even correlate with each other! So you can’t choose one of those and draw any conclusions if they don’t correlate with solar activity.
But here’s the thing: the two series in your graphic that show by far the best correlation is solar activity and NH temperature. Again, the correlation isn’t perfect, but it’s good. I think the probability that they line up by chance is less than 1 in 100.
So, ironically, that graphic that you posted actually gave good support for the solar connection!
Best regards and thanks for all the great work you’ve done,
Chris
Willis,
I think that it is implicit in ‘signals’ that change in period or amplitude significantly that they may be the result of constructive interference of primary forces, and are not themselves primary. That is, the “stacked” data may well be an artifact of different components responding differently to different drivers. Neither the drivers nor the stacked resultant may be correlated with things we have greater interest in (such as temperature) than we do of things like abundance of carbonate or magnetite grains. Some sort of physicality has to be demonstrated between the things we measure (proxies) and the thing(s) we are actually interested in. The “stacked” data may be interesting, and may provide some insights, but we have to look more closely at what they really represent and why they may be showing “pseudo-cycles.” However, even destructive interference may contain information because it may mean negative correlation between the signals that we chose to composite.
Chris Wright ,
Rather than eyeballing two periodic data sets to establish an informal correlation, it is probably better to create a scatter plot of one variable against the other (leaving out time) and seeing if the data cloud appears linear. If it does, then fitting a least-squares regression will give you a number to hang your hat on.
Chris Wright March 15, 2018 at 6:50 am
My original post on the subject is here. If you don’t like R^2, what are you proposing we use?
Details? The two are totally out of phase at the end, at the beginning, and in the middle. This is hardly a “detail” …
In addition, part of the time what is claimed as the “cause” lags what is claimed as the “effect” and part of the time it leads the effect … sorry, nature doesn’t work like that.
And your claim that “solar magnetic strength … may may well occur before an increase in sunspots” … I dislike this kind of handwaving greatly. The data is out there, please grab it and determine what it is actually doing before providing this kind of vague “explanation” for obvious discrepancies in the fit.
First, I looked at Steve’s post. He makes no mention of the graphic that I showed above.
Next, as I’ve mentioned more than once, we invented statistics because the “overall fit in the shape” is a very bad gauge for whether two series are related.
This is called “Monte Carlo” testing, and you need to be very, very careful in the generation of the random data that you use. For example, you have assumed that the solar signal is adequately captured by a random walk. This is not the case at all. Your pseudodata will have little resemblance to actual solar data, so your test is absolutely not valid in the slightest.
To do a valid Monte Carlo test, your pseudodata needs to have statistics (mean, variance, heteroskedasticity, cycle lengths, Hurst exponent, autocorrelation, trend, etc.) that are very similar to the real data … and that’s not an easy ask.
In fact, in my experience making pseudodata that resembles solar data shows that it is quite difficult to do, much more so than making pseudodata for many other datasets. Inter alia, solar data is curious because it shows large variations in peak values, but the values in the troughs are nearly identical. Despite some extensive trials I still haven’t found any way to reliably generate what I call “lifelike” solar pseudodata … maybe I’ll give it another try.
Actually, it was Javier who posted up that graphic, I was just discussing it. I have no idea whether any of it is valid or of the provenance of any of the data displayed therein.
My best to you, keep up your excellent curiosity, ask lots of question, well done,
w.
Thank you Willis. I’m glad you went to the effort of what Javier should have done.
Grabbing peoples charts from their papers and merely trusting them because they fit your narritive is not science. It’s novel writing.
Thanks, Mosh. After much experience I don’t trust anything, particularly claims about solar/climate connections. I haven’t found one yet that has held up to close examination.
w.
Willis Eschenbach
March 15, 2018 at 11:28 am
Willis,
Thanks for your comments.
“First, I looked at Steve’s post. He makes no mention of the graphic that I showed above. ”
I’m puzzled by this. By Steve’s post, did you mean the one I linked to? This is from ten years ago. He is referring to precisely the same sea level data, Holgate 2007. And he is referring to precisely the same apparent correlation we’re discussing now. So his comment stands. Surely you can see that, at the very least, there is a strong *apparent* correlation. Whether the correlation is in fact true is what we’re discussing now.
“If you don’t like R^2, what are you proposing we use?”
A very good question. But as I’m not a statistician I don’t know the answer.
“Details? The two are totally out of phase at the end, at the beginning, and in the middle. This is hardly a “detail” …”
I agree my comment about sunspots and magnetic strength etc were vague. I was really making a more general point, that sunspots have their own causes and there may be additional, variable lags that could affect the relative timing of sunspots and sea level. Quite likely solar magnetic activity is a better metric than sunspots.
They are indeed out of phase right at the beginning.This is by far the biggest discrepancy. That it does occur right at the start might suggest a problem, perhaps with the data, particularly as it specifically refers to the rate of change and not absolute values.
But apart from this, the rest of the graphs do line up remartkably well, just as Steve noted. You can’t conceivably expect all genuine correlations to be perfect – this is climate we’re talking about, after all. I’m sure no one seriously suggests that the only thing effecting climate and sea levels is solar activity. There are many other things, such as El ninos and ocean cycles that will also affect sea levels.
So, I think it is very likely that data problems and the effect of other climatic processes being superimposed may well reduce the correlation. But, taking these problems into account, overall, the peaks and troughs line up remarkably well. Although it starts with a bad match, it’s instantly followed by a good hit – the cycle 21 peaks line up perfectly. Are you really saying that the peaks line up one after another simply from chance?
“This is called “Monte Carlo” testing…”
Yes, I nearly referred to that. I agree that a drunkard’s walk probably isn’t the best method. By the way, Monte Carlo testing was described in Richard Muller’s book Nemesis (about how the Alvarez group, with his help, discovered what probably killed the dinosaurs – if you or anyone else can find a copy, I thoroughly recommend it, it’s definitely the best science book I have read).
I’m quite sure R2 has serious problems. I assume there are lots of other statistical methods for assessing correlation that don’t assume a linear relationship.
But some kind of Monte Carlo method should be effective, as it should make no assumptions about the kind of relationship. If you could perfect such a method I’m sure your post analysing this problem would be fascinating!
Yes, I had assumed you had originated that graphic, but in fact it was Javier’s. I’ll check to see the source of the data.
Best regards,
Chris
Good stuff. Perhaps add that humans are very good at seeing trends -usually just before they stop being trends!
‘Science is the belief in the ignorance of experts ‘.
Pseudo-science is the assumption of the infallibility of experts?
Sad but true, Ed.
w.
On pseudo cycles:
Chaotic systems display what I prefer to call ‘quasi-periodic’ behaviour.That is there will be times when they appear to be periodic, before they switch to a different mode and a new ‘period’ appears.
We dont have a good set of terms to describe chaotic behaviour other than ‘chaotic’ which just says ‘can’t see the pattern really’
I do wish instead of always chasing cycles, people would get to grips with chaos as an emergent property of complex non-linear dynamic systems. I dont think anyone denies (sic!) That climate is a complex non linear dynamic system, but they still go looking for ‘cycles’…
When all we can solve is linear equations there is a strong temptation to treat every phenomena as if it were the result of (enough superimposed) linear equations.
This is a recipe for disaster, or the IPCC. (is there a difference?)
weather is chaotic. climate … i have only seen claims. no evidence.
If weather is chaotic wouldn’t it follow that climate is chaotic; by reason alone and without evidence? If I stand back and say that on average the earth’s surface is about 9 degrees centigrade aren’t we just saying that the underlying chaos has bounds? I see your point though. At some point the chaos seems to disappear. The act of taking a temperature by itself has a great deal of underlying randomness.
“At some point the chaos seems to disappear.”
Does the chaos disappear or is it simply being obscured by averaging the data over a long period of time?
HankHenry,
I agree with you. If weather is truly chaotic I don’t see how averaging it will eliminate the chaos. On the other hand, if what Mosher claims is chaos is actually just a trend with a large random component (noise) then averaging will reveal the long-term trend.
i have only seen claims. no evidence
==============
Lots of work on this has been done in other fields. Most notably economics. In general, if a dataset is chaotic then the average is chaotic, because you cannot predict the time it will switch from one attractor to the next.
The average will only appear to be predictable so long as the climate (local average) continues to orbit the same set of local attractors. When the orbit shifts to new attractor set there will be a new average (new climate) established. When this will happen and what the new average will be, this is unpredictable by current mathematics.
So while the climate will appear to be stable over short periods of time, over longer periods of time it will appear to be unstable. Even the IPCC agrees:
The climate system is a coupled non-linear chaotic system, and therefore the long-term prediction of future climate states is not possible.
https://www.ipcc.ch/ipccreports/tar/wg1/501.htm
chaos is actually just a trend with a large random component (noise)
======================
that describes a system with a single attractor. it is the 2-body problem in orbital mechanics and mathematically predictable. This explains the interest climate gurus have in adopting such a model of climate.
However, when you look at our long term climate it is clear that even a very simple model has 2 attractors, Ice-age and inter-glacial. This is the 3-body problem in mathematics and not predictable under current mathematics. This explains the interest climate gurus have in dismissing such a model of climate.
In other words, our climate fluctuations are not (solely) due to random noise. They are due to deterministic chaos. For example, imagine a solar system with two widely spaced stars and one planet. The stars are called ice-age and inter-glacial. The planet is called climate. The planet is not orbiting in the orbital plane of the stars. Rather, the planet was captured from space, not formed with the stars.
As the planet climate orbits these stars, the temperature of climate will vary. And for all intents and purposes, this variance will appear to be random. Depending upon the orbit, the climate may be orbiting ice-age, or it may be orbiting inter-glacial, or it may be orbiting both and rapidly changing between ice-age and inter-glacial.
However, from our point of view, the mathematics to predict the orbits does not exist. It may as well be random because try as we might, whenever we calculate what will happen in the future, what actually happens soon begins to diverge from what we calculated in a random, unpredictable fashion.
“The climate system is a coupled non-linear chaotic system, and therefore the long-term prediction of future climate states is not possible.”
What does the IPCC mean by “Climate System”? As I understand it, “Climate” is just the weather averaged over a fixed period. Weather is definitely a system, but isn’t “Climate” just a mathematical abstraction of the Weather System?
That is how some of it is defined (where the 30 year average comes in), But you can also use the physical location, ie Latitude, Altitude, distance to bodies of water, and GHG’s.
What’s wrong with this, is it only has the noncondensing GHG’s, and the condensing GHG plays a role in both how warm it get, and how cool it gets, and that’s not included, or it’s just averaged in a way the you do not see it’s regulatory response.
So that shows a great increasing signal from the increase of Co2, it doesn’t show how WV has negative feedback response to that increase.
“ie Latitude, Altitude, distance to bodies of water, and GHG’s. ”
Aren’t all of these things just parts of the total Weather System?
Steven Mosher March 15, 2018 at 3:38 am
Well, here’s what Mandelbrot said about it, and he should know … from Mandelbrot and Wallis, 1969. Global dependence in geophysical records, Water Resources Research 5, 321-340.
My best to you,
w.
Well that was damned good work Willis, myth ablated to nix.
There can still be a cause and effect relationship even where cycles last for a while and then shift to another cycle length.
Multiple other factors can cause adjustments in the relationship between the cause and the effect.
During the shift to another cycle length the cause and the effect will of course be out of phase for a while.
Thus, I don’t think Willis’s analysis tells us much whereas the persistence of the relationships Javier points to in the historical record heavily suggests that there is a cause and effect relationship between solar activity and a climate response even if there are shifts in and out of phase over time.
SW,
Those compound pendulums that display ‘classic’ chaotic, unpredictable patterns serve in my mind as a template. The drifting around of cyclicity groups happens in the mind when you oil them, or forget to oil them and you get irregular drifts with sudden breaks.
In the example given above from Debret et al, the analog of oiling might be a side effect change in relative humidity or long-term cloud fraction or as Mosh would have it, some unicorn. I would not expect an absence of unstudied side effects that disturb the regularity of sedimentation or isotope fractionation. Geoff.
Sure Stephen, possible, but Willis showed that what was claimed by Javier was not even there.
If there is something to it, it has to be a lot less touchy-feely than layering multiple datasets on each other and looking for regular cycles in the resulting cumulative error.
Stephen Wilde March 15, 2018 at 3:44 am
And you know this how? I fear that science by assertion doesn’t work all that well.
w.
I think that the given cycle did not necessarily change its length, but that current conditions of the previous state have carried over to prolong the original state in place prior to the shift.
Thank you Willis for taking the time to show the analysis for what it is. It’s worrying when people get their hands on tools they don’t know the limits of. Even scientists fall prey to it.
Recently WUWT seems to be posting more from amateur scientists. Great – if they’re careful. But it’s a problem when people learn a little about a tool that they don’t know the limits of. They can look like “experts” without having learnt the hard way of dead-ends and pitfalls. We live in an age where “arguments from authority” are immediately suspect. But sometimes a little experience goes a long way.
There is always the possibility that one uses cadillac methods on horse-and-buggy data. This may lead to misinterpretations, especially when the methods are meant to apply to normally distributed data, and data distribution has not been tested. My reading would question results when one “good” data set is analyzed relative to a “not so good” data set. I’m not sure how “stacked” data sets would play out.
Willis,
Your deductions here are compelling.
It is easier, I think, to agree with some of your comments like “As I’ve said many times, this kind of appearance and subsequent disappearance of “cycles” is quite common in natural datasets.” if you have worked in geological sciences with their long time spans in which constant concepts like uniformitarianism can vary. Like your cycles that come and go. Sometimes the variation is on the Y-axis, but it can be on the X-axis, often a time axis, where a constancy of events cannot be assumed and where measurement errors are often less scrutinised.
With my experience – and I have not been hands-on for a while now- I doubt that I would have even considered ‘stacking’ the 4 measures shown here. There is too much possibility that the time series can go out of phase, for a start. Then I would be initially sceptical of measures like glass 1 & 2 that devolve into such similar looking graphs. I might be missing some points, but it is rare in earth science work to obtain such good agreement. We drilled and analysed thousands of holes, some very close together, others a long way apart, for geochemical interpretation. Correlation coefficients below about 0.6 were often the norm when we were expecting high correlation, above 0.9 was almost unheard of. But then, there was no value in making adjustments to the raw numbers, that was only fooling ourselves, so we did not do it.
Your essay here should complement the work from Javier, hopefully tighten up some parts of his overall thesis, make it more solid. Thanks Geoff.
Well here is a contradiction…. Or worse. If the sun completely winked out like a 17th street lamp i’m pretty sure the climate will change. Nothing like the warmth ftom those first rays at sunrise. same gies for the relief at midday when a stray cloud blots it out. The math from Wilis is just playing around. Ha!
Whenever I have ventured to point out some problem with the underlying assumptions of some physical theory (for example, when I disputed the relativistic interpretation of “gravitational waves” because I do not believe in the literal existence of a space-time continuum), there have been people–often enough people on this very website–who were happy to get all Popperian on me and proudly aver that “science doesn’t have to explain the world, it simply has to describe it mathematically.” The name David Hume is often brought up in this connection. As another darling of the empiricists, he was all too willing to hang his entire system on the strength of correlations which he deemed it unnecessary to dissect. In order to have empirical value, it was only required that they repeat often enough to form an impression in the mind.
On what grounds, then, do the same people even dare to moot the concept of “pseudo-cycles”? By their own lights, what right have they to say that some correlations are significant while these other ones are not? Once they have done that, they have stepped outside the boundaries of a strict empirical order. They have made use of some foundational and probably entirely unconscious image of the world which they never examine in further detail. Yet if some discovered correlation allows us to predict anything about the behavior of the natural world, it is not necessary (again, on their own theory) to inquire even whether there this is a causal linkage involved, let alone what it might be; but this bit of Popperian piety is conveniently forgotten by these amateur defenders of science whenever they are in pursuit of their own ends, which may be as shallow and petty as simply ensuring that another does not encroach upon his limelight.
+10
Intelligent Dasein, you are not being intelligent when you refuse to QUOTE WHATEVER IT IS THAT YOU ARE BABBLING ABOUT. You throw out all kinds of ugly accusations involving unknown actions by unnamed people. How is anyone supposed to respond to that pile of reeking rubbish when we don’t know what you said, what they said, or even who they are?
Pathetic … get with the picture, my man. Your anonymous accusations of unknown crimes against unidentified people do nothing but ruin your reputation.
w.
Willis,
at first glance you are entirely right, however discarding them as pseudocycles is a bit of a short “easy answer”.
before i explain let me say this: it would be interesting to see if one of the “doubles is removed from the stack that every dataset as it’s unique value and weight. now i will place a conceptual thought from a sound engeneer’s point of view:
i don’t see artefacts, i see resonances. when you add up white noises together you will find cycles that come and go These are – if strong enough – very audible. As climate is a result of many white noises added together it’s change can be just the response on these resonances that come and go.
All we know from all these white noises is that some have cycles, others don’t. I do agree with your post that these cycles are results of the additions of the datasets, but then comes the interesting question: how do these resonances influence the climate system? How are the resonances when global temp is warming, how are the resonances when it is cooling? Have these resonances a periodicity (phasing) or are they just random? As some inputs in the climate system have some periodicity some resonances will have that come and go as well and others won’t.
i am aware that this may be beyond the scope of your article, as like you said: to bring them as real cycles is indeed incorrect they are a result of stacking the data.
Frederik:
“… All we know from all these white noises …’
White noise is not many varieties, it is just one thing, all frequencies being produced at equal unchanging amplitude.
No resonance, no cycles, it is uniform invariant noise at the same amplitude at all frequencies in the wavelenght spectrum.
yes i see i used a bit of a wrong term to explain the analogy. In fact i wanted to say a “varying noise with varying individual frequencies and amplitude” so that some can resonate. White noise is indeed not that type of “noise”
the idea was that in this varying noise that some frequencies can get that important that they create resonance or so called artifact cycles
to describe the concept a bit better: in the above data from your article all 4 datasets do have “some hidden ground tone” at 960 years, that goes as unimportant in each signal, but they suddenly show a “resonance” that is temporary in the “stacked dataset”
of course these resonances are pure “luck”. With longer datasets that may just show up as something that appears and then dissappears.
Frederik,
The question is whether the naturally occurring noise amplitudes are being modulated by some other force. That is, with noise envelopes varying over time, one can get constructive or destructive interference in the presence of other noise envelopes similarly modulated, but by different drivers.
Willis,
Stacking records is somewhat of an art form that basically comes down to wiggle matching of time series with a few fixed points that tie the records together chronologically. It can be done by eye or using a technique to constrain the matching. In the end it’s an approximation and sometimes the best that can be done. Valuable information can come from it, but always with caveats.
https://www.sciencedirect.com/science/article/pii/0025322784900094
The central thesis of this article is that
But the article fails to consider the alternative hypothesis presented by the authors of the study:
Bond, G., Kromer, B., Beer, J., Muscheler, R., Evans, M. N., Showers, W., … & Bonani, G. (2001). Persistent solar influence on North Atlantic climate during the Holocene. Science, 294(5549), 2130-2136.
Working with the numbers without consideration for what they represent can lead to wrong conclusions. Clearly the author of this post has failed to read the article he is critizicing, because otherwise he would have noticed its figure 1:
With its long legend:
“Fig. 1. Map of coring sites described in the text that provide the basis for inferring sources and transport routes of ice carrying the petrologic tracers. Dashed blue lines: subpolar cyclonic circulation. The main frontal boundaries are labeled in blue. Red dots are core-top measurements of all tracers. Areas enclosed by shading indicate core tops with >10% of tracers as keyed by colors [red: >10% hematite-stained grains (HSG); yellow: >10% Icelandic glass (IG); blue: >10% detrital carbonate (DC)]. Documentation for core-top percentages of HSG and IG are from (2); red numbers next to core-top locations are percentages of DC in core tops. Colored arrows indicate inferred direction of transport of tracer-bearing drift ice. Gray lines are mean (1900 to 1992) ocean-surface temperatures from LEVITUS94 (52) for spring when iceberg discharge into the North Atlantic reaches a maximum. EIC: East Iceland Current; EGC: East Greenland Current; LC: Labrador Current. VM28-14: 64°47’N, 29°34’W, 1855-m water depth; VM29-191: 54°16’N, 16°47’W, 2370-m water depth; VM23-81: 54°15’N, 16°50’W, 2393-m water depth; KN158-4 MC52: 55°28’N, 14°43’W, 2172-m water depth; KN158-4 MC21, KN158-4 GGC22: 44°18’N, 46°16’W, 3958-m water depth; and EW9303 JPC37: 43°58’N, 46°25’W, 3980-m water depth. Petrologic analyses of more than 120 core tops demonstrates that most tracer-bearing ice today circulates in the cooler waters north and west of the subpolar front. Lower tracer percentages to the south and east are consistent with observational evidence that icebergs there come mainly from south and west Greenland where tracer-bearing rock types are rare, if present at all. Increases in DC off Newfoundland, therefore, reflect southward shifts of the cooler Labrador Sea surface water and carbonate-bearing drift ice. Peak percentages of HSG and IG off Newfoundland rarely reach the correspond- ing peak values of those two tracers in the eastern North Atlantic (MC52- VM29-191) (Fig. 2). That rules out transport of HSG and IG through the East Greenland–Labrador Sea current system at times of peak drift-ice transport. The eastern North Atlantic drift-ice records, therefore, require that at times of peak tracer percentages, ice-bearing surface waters from north of Iceland were advected southeastward toward the coring site. That was accompanied by cooler ocean-surface temperatures (1) and, by analogy with transport mechanisms of modern drift ice (53), must have been aided by northerly or northeasterly surface winds in the Nordic Seas and eastern subpolar North Atlantic. The concentrations of IRD (lithic grains >150 µm), although small, covary with the petrologic tracers, and peak percentages reflect true increases in the tracer concentrations rather than dilution by other grain types.”
If the author would have read the article he would have noticed that the different tracers come predominantly from different areas. And as the original authors state:
“As demonstrated previously, the tracers are particularly sensitive to changes in the amounts and trajectories of glacial ice and/or sea ice circulating in the surface waters (1, 2). We interpret percentage increases in the tracers as reflecting advections of cooler, ice-bearing surface waters eastward from the Labrador Sea and southward from the Nordic Seas, probably accompanied by shifts to strong northerly winds north of Iceland, as explained in Fig. 1 and (1).”
So stacking the records results in sampling a larger area than analyzing the records one by one. And as the authors state:
“The rationale for stacking is that all five records are in % petrology and each petrologic tracer reflects the same parameter, i.e., change in drift ice.”
All the records are measuring the same phenomenon in the same way, so it is adequate to stack them.
Let’s now consider a different example. We have a large area (or a globe) on which we have temperature measurements from different stations. Do we stack those measurements to get a better representation? Yes we do. Is the final result reproduced in every station? No it isn’t. Do we say that the final result is an artifact arising from the stacking? No we don’t. Then why apply a different criteria here?
Sampling from a larger area gives a better representation of a regional phenomenon, the climatic effect of the 980-year Eddy cycle on North Atlantic iceberg rafting. It is not only irrelevant, but expected, that local effects are more variable.
This post fails to demonstrate that the hypothesis and the evidence on which the original article is based are not correct. It proposes but does not demonstrate that the signal is an artifact from stacking, versus the more correct interpretation in my opinion, that the fact that the signal intensifies with the stacking actually supports that it is a more general phenomenon even if it affects differently the different areas.
I completely agree with the conclusion of the article, one of the highest cited climatology articles, with over 2500 citations in Google Scholar:
As a referee I would have to recommend that the post is rejected, because it is fatally flawed by failing to consider the most obvious alternative explanation, that the signal intensification is not an artifact, but a consequence of the phenomenon being analyzed.
“Dashed blue lines: subpolar cyclonic circulation.” As of today, I presume. Are there indications that it was the same 12,000 years ago when the sea level was some 400 feet lower?
I completely agree with the conclusion of the article, one of the highest cited climatology articles, with over 2500 citations in Google Scholar:
Most of those [go have a look] cite a 1500-yr cycle, not a 1000-year cycle. ..
And ADS [http://adswww.harvard.edu/] records only 1232 citations…
Leif, Harvard has a much smaller database than google scholar. Google scholar samples virtually all periodicals in all languages. It is more accurate as a result. In any case, even 1200 citations is large for a geological article.
Leif, there is a paper that compares citation counts. It found that google scholar finds an average of 53% more citations that WoS and Scopus combined. Here is the link: https://arxiv.org/ftp/cs/papers/0612/0612132.pdf
It is interesting that combining GS with WoS and Scopus increases the number of citation by 93%! Thus, almost half of citations are missed by WoS and Scopus.
You probably know that Gerard Bond wanted to push the 1500 year cycle into the Holocene scene. He had already discovered and published about a 1500-year cycle in benthic cores for the last glacial period, related to the Heinrich cycle and to Dansgaard-Oeschger events. But one thing is his conclusions and another the evidence he uncovered which became very important for two reasons:
– The close correlation between solar activity and Holocene Bond events, that I have reported here:
https://wattsupwiththat.com/2018/03/13/do-it-yourself-the-solar-variability-effect-on-climate/
And more importantly:
– Bond events are recognizable in many climatological records (for example the famous Dongge Cave speleothem), from tree rings to lake sediments, pollen, other benthic cores… Many articles that cite the article is because they find evidence that agrees with one, or more frequently several, Bond events.
The 1500-year cycle in the Holocene is not what Bond expected. It is a lot harder to detect, and it is clearly not solar in nature.
https://judithcurry.com/2017/09/15/nature-unbound-v-the-elusive-1500-year-holocene-cycle/
Most professional [peer-reviewed] science papers today are published in English, so it matter not that Google covers many more languages. And ADS is pretty good for high-quality papers. To wit, it reports 4563 citations to my papers vs. Googles 4686.
And most of the citations of the Bond paper are about the climate [not the solar connection, although some lip-service is paid to that] and the 1500-yr cycle [which Javier claims are not solar to begin with].
“The results of this study demonstrate that Earth’s climate system is highly sensitive to extremely weak perturbations in the Sun’s energy output, not just on the decadal scales that have been investigated previously, but also on the centennial to millennial time scales documented here.”
You still have to with data, that shows the ~2300 yr cyclicity broke down in the last few thousand years.
That should have read… have to go with the data….etc
That seems correct to me Javier. Presumably there is quite a lot of other corroboration from other sources, but even without that it’s pretty clear you can’t just dismiss the result as an artifact of stacking.
Javier March 15, 2018 at 7:18 am Edit
Thanks, Javier. What is the “alternative hypothesis” you are referring to?
Yes, I know that.
Agreed, not sure what your issue is here.
Here’s where we diverge. IF all of the records were actually “measuring the same phenomenon in the same way”, their internal cycles as revealed by the CEEMD analysis would also agree … but they don’t. The two Icelanding glass records agree, so yes, they are obviously “measuring the same phenomenon in the same way” … but also obviously the others are not.
It’s well known that the temperature anomalies are very well correlated at distances out to ~ 500 km apart. That is obviously not the condition with this stack regarding the putative 960-year cycle … but it IS the situation with the ~2,500-year cycle, which appears in all four of the underlying records.
You still don’t seem to have grasped the nettle. In total contrast to the ~2,500-year cycle, the 960 year pseudocycle only appears in one place—in the earlier half of the stacked records. It does NOT appear in any of the individual records. As a result, it is clearly an artifact of the stacking process.
You seem fixated on the number of citations … Michael Mann’s Hockeystick paper has been cited well over 2000 times as well.
Does that mean that the hockeystick paper is right? … don’t make me laugh. The number of citations means NOTHING in climate science. Lots of total crap gets cited over and over again. Yes, lots of folks cited the Bond paper, because it agreed with their preconceptions … so what? Science is the belief in the ignorance of experts …
Regards,
w.
Good discussion on cycles. The problem is trying to determine what, if anything, is driving the purported periodicities. What leads me to conclude this could be an artifact is that the correlation goes away with the oldest data.
Willis, thank you for putting a lot of work and time into this analysis. My personal problem is that I don’t understand what exactly is being analyzed: hematite grains, Icelandic glass, and detrital carbonate. Why should they have anything in common, and why should that common thing – if any – be a proxy of solar activity or of climate?
Good question, George. This is all data from what is called “ice rafting”, where ice forms, drifts to another location, and then melts. When it melts, all the bits and pieces of stuff like hematite grains, tiny bits of volcanic glass, and detrital carbonate that were in the ice fall to the ocean floor.
In theory, this will give us information about the weather conditions that formed the ice.
w.
will give us information about the weather conditions that formed the ice
And of the [well-known] contamination of the cosmic ray record by the weather and climate [especially in and around Greenland].
lsvalgaard March 15, 2018 at 2:58 pm
Yes, I’ve deliberately stayed away from putative solar connections for exactly that reason, and just focused on the Bond data itself.
w.
The results of this study demonstrate that Earth’s climate system is highly sensitive to extremely weak perturbations in the Sun’s energy output, not just on the decadal scales that have been investigated previously, but also on the centennial to millennial time scales documented here.”
My commentary is the solar changes with earth’s climate is not an isolated system meaning the same changes in solar activity are not going to give the same result.
This is the flaw in trying to connect the dots when it comes to the climate. In addition to solar changes versus earth’s climate not being an isolated system the magnitude and duration of change have to taken into account.
Then the geo magnetic field. I am especially going to keep harping on the geo magnetic field which will enhance or diminish given solar effects.
In addition so much else has to be taken into consideration from the arrangement of continents to the given state of the climate to what is going on in near by space (within 60 light years to pick a number) just to name some that have to be incorporated into what the sun is doing or not doing at a given time to give the x result.
The thresholds also and I know they are there because the climate has changed abruptly at times.
This is why we go in circles and why everyone can argue their points and be convincing to a degree because there is so much wiggle room because the entire picture can not be put together at a given time.
The only conclusions I have reached is, if solar/geo magnetic fields weaken enough (what is enough?) they are going to cool the climate to one degree or another.
.
The results of this study demonstrate that Earth’s climate system is highly sensitive to extremely weak perturbations in the Sun’s energy output
The results of this study demonstrate that Earth’s climate system is highly effective in producing weak perturbations in our proxies for the sun’s output…
The only conclusions I have reached is, if solar/geomagnetic fields weaken enough (what is enough?) they are going to cool the climate to one degree or another.
I can tell you what ‘enough’ is. It is that weakening that will bring about a cooling of one degree or another.
Perhaps you can see how meaningless that is as an expression of your knowledge about this. It is like saying: “if it rains enough, I’ll get wet”.
I have observed countless times that members of the deer family take no notice of a stationary person. I, on the way other hand, notice all sorts of stationary animals. I notice that my dog also sees stationary deer. People also see patterns in human relations, weather, the effects of diet on health and pretty much everything we do. I am not buying predator recognition as the primary, or even an important aspect of patttern recognition.
This will drift a bit aside, but Willis himself gave this opening with his theory about pattern recognition, so I guess it is OK.
It helps (me) to define an extreme position, i.e. all phenomena in nature considered in their entirety show us total chaos. Not chaos as defined in mathematics or chaos theory, but chaos as experienced in say “random noise”, “incomprehensible”, “beyond recognition”, “can mean anything”, “completely beyond me”. The positive about this position is it encompasses the maximum degrees of freedom you give yourself to give any phenomenon you encounter a place on the continuum from familiarity and recognizing it on the one hand, and completely being chaotic on the other.
The closest I know of “total chaos” in the physical sciences is the phenomenon of “vacuum fluctuations”. As far as I know (I am not into quantum physics so bear me on this) these fluctuations are not limited to a vacuum but they are ubiquitous. I refer to this phenomenon here to show this extreme position of “total chaos” is not just nonsense but makes some sense in the real world.
If this is so, any pattern and regularity observed may have its origin in the predisposition of the observer. Predator or prey, doesn’t matter. They all see their relevant patterns. Science? Wouldn’t exist without this predisposition. Its origin? Life itself I would guess. If so, something very basic in any life form.
Now whether this predisposition gives you useful information all the time is another question to which the answer would be “no” in my opinion. The discussion then should center on not why we have this pesky predisposition (“of course we do this all the time, we are life forms after all, aren’t we!”), but how to tackle it to see the good ones.
BCBill March 15, 2018 at 9:26 am
Thanks, Bill, and if you have a better explanation for why we see patterns where none exist, now would be the time to break it out …
w.
Thanks for commenting Willis. The following is not my original idea but I have added some thoughts to things that I read in evolutionary biology. I can get you a reference if you like. I and others contend that pattern recognition is perhaps the most important human skill, though other animals have it too. There is a huge advantage to being able to leap ahead, to see a pattern with incomplete data. In fact a premise of life and science is that we never have absolute confirmation so the question is when is it time to act? Traditionally the decision of when to act was probable based largely on merit. For example, I have read that in the Lakota tradition there was no such thing as a chief in the western European context. If Willis was good at predicting where the bison would be for the fall hunt, then others would move camp with him to prepare for the hunt and he became a de facto hunting chief. There was a high cost to moving camp too soon and an even higher cost to moving camp too late. As anybody who has hunted knows, it is not trivial to predict where game will be at any time. Past and present weather, forage availability, predator behavior, phase of the moon and other factors all play a role in creating a behaviour that can look chaotic. The same is true for plant crops. I have read that hunter gatherers such as the Ju/’hoansi or the aboriginal people of Australia harvested up to hundreds of different types of plants in different areas, which they would access depending on the weather patterns. Hunter gathering probably required the greatest development of pattern recognition. However, that does not negate the likelihood that pattern recognition was useful in all aspects of human life (are those people conspiring against me, when I ate that plant I felt better, when I shape the stone like this it appears that I have better success at bringing down an animal (finite analysis not necessary), etc.
With the advent of agriculture and most importantly with the increasing control of factors affecting production, it has become increasingly less necessary to be successful at pattern recognition. Most people today can probably survive quite well with almost no pattern recognition skills and so it is perhaps not surprising that the ability to recognize patterns may have atrophied (any farmer who has had the misfortune of working with citiots can attest to this). However, another point of view is that there was always a high error rate in making predictions with inadequate data. Some people were good at it and some people were bad at it and some people were good at making it appear that they were good at something other than what they were really good at. This gets into my whole conman theory of social evolution. Briefly, the original con may have gone along the lines that a keen observer noticed that people with some types of injury or illness were more or less likely to survive. All that remained was for that person to figure out how to take credit for the outcome. The con is that the ability to predict an outcome is attributed to the ability to create an outcome. Both abilities could come from astute observation and intuitive thinking, however one ability is more palatable than the other. But I digress. The point about making wrong decisions it that you would hopefully learn from them. Perhaps the true evolutionary dead end would be the inability to make a decision without complete data. I read one time that the only difference between a good scientist and a bad scientist is that a good scientist tries more things and therefore succeeds at more things, even though the success rate is the same. I don’t entirely believe that, deciding what to study is a highly developed pattern recognition skill. However, I do think that there is an element of truth in that. As long as bad choices aren’t lethal, they probably make us better and so people at one time (before the tyranny of the manager) were predisposed to act, even if they risked being wrong. The global warming nonsense could benefit from this understanding. Many are focused on trying to prove what is the cause of climate change, or worse have accepted a false explanation for climate change. It would be so much more beneficial to recognize that climate and the world are in a constant state of flux and that trying to stop that is akin to trying to stop nature herself. We can only progress as we have always progressed, by having the flexibility to adapt to the situation, to intuit the best response.
I am not a surfer (can’t stand the cold water), but I have spent many hours from the beach watching sets of waves come in. First one or two in the set are weak, but then they build and usually 4,5,6 are the strongest then the wave set starts to subside.. 7,8,9 till the set disappears. Surfers out just beyond where the waves are building watch for the one they will pick to try and get a ride on the rising front edge. But they do come in sets. On a very high surf day, the time between sets may be so minimal it seems they never separate into individual sets, but they do if you watch closely enough.
This informs me in several ways about the nature of wave sets.
– First, as the name implies, they come in sets. There are variably large or small large time gaps between sets. Very active (high energy periods) the sets come in so quickly there is really never a break.
– Second, somewhere far out in the ocean, a physical disturbance process (a surface wind shear with a low pressure perhaps) starts the process in motion. Across a vast swath of ocean, many such small scale events can constructively or destructively interfere. Those that constructively interfere may move many hundreds of miles before arriving at the shallowing sea bed as it nears the shore to build the wave set. A chaotic process that defies deterministic analysis or prediction.
– Third, the state of the tide (height of the water at the beach) can affect the sets with apparently different amplitudes due to depth (to sea floor changes) from hour to hour/day-to-day. The state of the tide would be the “Residual” in the mode decomposition.
What does this mean for Willis’s analyses?
– First, don’t disregard what the value of the residual may be telling you. The decay of the residual may be related to physical reason the “the putative 960-year cycle in empirical mode C3 only has significant strength in the earliest 6,000 years of data. On the other hand, it weakens and nearly disappears in the most recent 5,000 years.”
– I can only sit on the beach so long to watch waves before I have to leave to go do something else, like eat, sleep, pee. So, the severe problem that cannot be overcome with the dataset at hand is that it only goes back to less than 12,000 years. Yet if these millennial scale cycles are present in the solar activity, they have been occurring for many millions of years. The trees falling in the forest continues to make noise even after everyone has left.
– Stacking the sets is similar to radar signal integration processing, in that you want random fluctuations (noise) to be cancelled with successively filtered radar returns, while enhancing the SNR on real targets to achieve signal detection above your threshold, this is in effect constructive and destructive interference happening by careful design of the signal processing architecture. False alarms (noise artifacts) are suppressed with more integrations (stacking). Thus this supports Javier’s assertion where he says, “…it [this post] is fatally flawed by failing to consider the most obvious alternative explanation, that the signal intensification is not an artifact, but a consequence of the phenomenon being analyzed.”
My random musings.
watching sets of waves come in
=============
if you watch long enough, you will eventually see a wave that is bigger than any wave before that. This is proof that we are experiencing more extreme waves as a result of climate change. not.
This is identical to what is happening with temperature, storms, floods, etc. They all appear to be getting bigger simply because we are watching them, and as expected if you watch long enough you will eventually see one that is bigger than all the others.
What is changing is not climate. It is the sample size that is changing. Climate science has made a fundamental statistical error. They have changed the sample size over time, and created a false trend.
I certainly agree that climate science is making a fundamental mistake (maybe intentionally) of insufficient sample size for linear trends that are mere cycles.
joelobryan March 15, 2018 at 10:35 am
Well, as a long-time surfer and seaman, I can assure you that that is absolutely not true. The ocean is nowhere near as orderly as that. See the marvelous text “Oceanography and Seamanship” for an excellent discussion of the random probabilistic nature of waves.
w.
Sorry, I am not going to buy a 3 lb book from Amazon.
The observation that waves come in sets is true. This is no different than when a peebble is thrown into a pond. The waves comes in ripples (sets). Keep throwing pebbles in slowly and you have sets of waves. Throw a large handful in all at once and the time between the sets casued by each individual pebble means that the sets are arriving continuously, asynchronously. Similarly, when the energy state of the ocean is high A distant storm is passing off-shore) then it’s “Surf’s Up!”, and the time between sets makes them appear continuous.
joelobryan March 15, 2018 at 7:46 pm
Your call … me, I don’t weigh knowledge by the pound.
Absolutely.
You missed my point, likely my lack of clarity. You’d said:
Waves do NOT come in any kind of regular sets. Nor are the fourth, fifth and sixth waves “usually … the strongest”. This is just more of our insistence that a pattern exists when there is no pattern.
When I started surfing, the legend was that the 7th wave is the biggest. And I can still find it all over the web. But a quick look at Google finds things like:
and
and
Best regards,
w.
The big sets on the north shore of Oahu tended to be third wave is a monster. If anything, longer sets (more waves) tended to also be milder; heavy storm surf tends to have short sets but peak considerably higher.
I have a photo, shows me starting to run to higher ground at Chun’s Reef. It’s hard to estimate the size of the wave I was fleeing. I estimate up to 50 feet on the face just before breaking. It was huge and went over the rocks like they weren’t even there, came all the way up to Kamehameha highway in places.
Willis,
My observation is that the 4th or 5th wave in set is usually the biggest. Also I don’t let the “good enough” become the enemy of the “perfect”. Life is too short.
Regards.
Actually, I wrote that backwards,
Correction: I don’t let “perfect” become the enemy of “good enough.”
Too many beers tonight. time to go to bed.
Cheers.
Thanks, Joel, stay well.
w.
Michael 2 March 16, 2018 at 1:47 pm
As someone who used to surf Pupukea, very occasionally even successfully, I can testify that a wave getting that far would indeed be huge.
w.
joelobryan March 15, 2018 at 10:35 am
Joel, you’ve missed my point, likely my lack of clarity. In general, the only difference between two successive radar sweeps is noise. So yes, you can stack those with great success. I’ve used such stacking in my own research. See e.g. Stacked Volcanoes Falsify Models.
But that does NOT mean that you can simply grab some random datasets from different locations measuring different things and stack them. Be clear that I am not objecting to stacking in general.
I’m objecting to stacking in this particular case, based on the fact that the stacked dataset contains results that are not present in any of the underlying datasets.
w.
I saw the graph titled “Empirical Mode C5 Four Bond Ice Rafting Datasets”, and so I did my own overlay of all of the empirical modes. One thing I saw is that the two glass ones are practically identical and that it could be useful to combine them into one dataset. Another thing I saw is that C3 lined up well in all four datasets for most of the past 5,000 years, impressively well for nearly 7.5 cycles, and C4 lined up impressively well for the past nearly 6,000 years or 5 cycles. For that matter, C4 lines up well through the past nearly 9,000 years or nearly 9 cycles in all four datasets, except for the hematite dataset missing one positive peak around 6500 years ago. I see this as some evidence that a periodic item in C4 with a period close to 1,000 years is for real.
As for 9,000 to 12,000 years ago, it appears to me that the ~1,000 year cycle is shifted from C4 to C3 for the glass and hematite datasets, so it mostly exists in all datasets for the full 12,000 years with the main exception being the hematite dataset missing one positive peak. I wonder if this is related to C5 having phase disagreement in the datasets around and before 7,000 years ago, with either C3 or C4 being temporarily or intermittently a harmonic of C5. If the fundamental waveform of the solar cycles is not a sinusoid but a distorted one with faster risetime and slower falltime as is the case with the 11-year cycle, then I think CEEMD being modified to consider this could make it more clearly separate and identify any actual longer period solar cycles. Another thing I noticed is that the carbonate dataset has an irregularity a little over 9,000 years ago that is mostly in C1 and C3 and the C3 component of this irregularity just happened to be pretty much in phase with the ~1,000 year cycle, so I wonder if the ~1,000 year cycle itself got temporarily misidentified as something else because of this nonperiodic irregularity.
Something else I noticed: The squiggle in C1 around 3,000 years ago shows up in all of the datasets. That was something that held up well as periodic for about 1.5 cycles or a little more with agreement in all of the datasets, but is obviously not an ongoing oscillation.
One more thing: The ~11-year solar cycle is an obvious ongoing oscillation that does not have steady frequency or amplitude, and even looked close to nonexistent in the sunspot record during the Maunder Minimum. Any longer period cycles modulates the ~11-year cycle, and I wonder if CEEMD has trouble distinguishing harmonics and/or modulation of one cycle by another.
I immediately thought of this:
Penguin In Bondage
https://youtu.be/FWqWI5diKPA
You could do the world a favor by studying the AMO cycles with modern data combined with USGS drill core data from coral reefs. We need some answers and not modeler simplifications with averages.
resource guy.
looking at data requires models.
typically statistical models.
We need to decide why we are doing so much of this expensive research.
We need to set objectives that are more important than satisfaction of intellectual curiosity.
How important are cycles to our present well being and quality of life?
I propose them as irrelevant. Geoff
ResourceGuy March 15, 2018 at 12:53 pm
You could do the world a favor by providing links to the “AMO cycles with modern data” and the “USGS drill core data” you’re babbling about. I’m happy to look at most anything, I’m a curious guy, but I’m not going to try to guess what the heck it is you are referring to. I don’t go on a snipe hunt for any man … or woman, if that’s what you are, don’t want to make sexist assumptions …
Finally … what do you expect to find? That temperatures in the AMO are somehow related to temperature proxies in the drill core data? And if they are … so what? What new insights do you think it will provide?
w.
I have 2 comments that are unrelated so will treat them separately.
Although I appreciate that this post is of scientific interest, is it relevant to ongoing CAGW debate? The cycles being discussed appear to be over periods long enough that we might not observe them over the 30 year period from 1989 to present.
Thanks, pmh. It is relevant to the ongoing debate because the underlying paper is a good example of the kind of sketchy science that goes on in the field. As I said, it is a cautionary tale to encourage people to be skeptical, not only of alarmist claims, but even more so of claims that agree with our preconceptions.
w.
I much appreciate the time, dedication, and sacrifice Mr. Watts has made with this blog. I have nothing original to contribute and have learned much. My problem is that if my comments/questions are not in the first 12 hrs (or so) from when the post appears, everyone has moved on the to the next post and subsequent comments/questions are either ignored or lost. Not having spent my life in this field, what is lost to me is the educational aspect. But, perhaps that is not one of the purposes of the blog. In any event Wattsupwiththat is still the go-to blog on climate related issues and I offer my thanks and appreciation.
Not true for comments on my posts. I subscribe to each of them so I get an email whenever someone comments. I reply to as many of them as seem relevant.
w.
Thanks for doing this analysis, Willis. Very educational. Especially some of the comments on anonymity.
Quasi-periodic data that breaks down over time may be due to earth dynamics interfering with more regular solar events. The Pleistocene Ice Age began when continents assumed their modern placement and caused major changes in ocean currents and atmospheric circulation.
Milankovic Cycles were highly likely to have been active before this time but to smaller effect, perhaps causing a globally moderate cooling. The much more disruptive Pleistocene Ice Age resulted from concentration of the cooling in the polar regions, particularly in the NH because of the concentration of land that inhibited distribution of heat by ocean currents.
Major events like collision of bolides with earth could also interrupt whatever “cycles” were in play. A giant one could probably end a glacial maximum and a Yellowstone grade volcanic blowup could bring one on early.
That quasi-cycles aren’t enduring shouldn’t discourage their investigation. They could still be quite useful for forecasting. An intelligent humanoid understanding Milankovic Cycles 2.5 mya would have had a good long run forecast going!
Today’s 60ish year cycle of warming and cooling seems to have arrived on schedule to give us the dreaded Pause and with it, given the great over estimation by CAGW proponents, support for the recent warming having had a substantial boost by the rising phase of this cycle. IIRC, someone in the early years of the new millennium predicted the cooling going forward. I myself cautioned about 5 years ago here on WUWT that instead of crowing about the hurricane drought, sceptics should pre-empt the certain exploitation of renewed major hurricane activity to come – a repeat of the busy 1950s hurricane cluster- by forecasting the return in a few years. By golly I went and predicted Harvey, Maria and friends on schedule, and, of course the doomsday crowd’s rejuvenation. IIRC, I threw in the ending of the Texas drought as a bonus. This cycle may end eventually but might be useful for a few centuries.
Another possibility I have never seen ruled out is a slowly decreasing sea level atmospheric pressure (SLAP). A slow secular decline in SLAP over millions of years would eventually make ice ages begin as the weakening warm attractor would allow the climate trajectory to veer toward Ice Age attractor under the Milakovitch INSOLATION cycling. Chaotic, but stochastic.
Thanks for the idea, Joel. I find this from Science magazine, which says the opposite … but who knows?
w.
Willis, A couple of points – Firstly Fourier transforms do not get confused – if you want to reconstruct a signal as a sum of complex exponential then the coefficients you need are those you get from a Fourier transform anything else would not give the correct signal.
On the other hand Fourier transforms do not tell you what frequency components are present at which moment in time. For that you need to calculate the instanteous frequency. This is not what the empirical mode decomposition does although it is a step on the way. To do that you would need to take the Hilbert transform then get the analytic signal, calculate the phase and then take the derivative of the phase. Simply eyeballing the amplitude of empically calculated mode and claiming that it seems to grow or decay is like finding patterns in stars. It is just not good enough.
I am also slightly skeptical about the modes you show. I downloaded the data from the website ran it through two different programs to calculate the empirical mode decomposition and got very different results. Firstly they both found 9 modes not 6 and also the 2nd mode is very different. The issue is that the empirical mode decomposition is not unique and so different methods for calculating them will find different modes and also find different Fourier coefficients for each mode.
Germonio March 16, 2018 at 12:19 am Edit
I didn’t say Fourier transforms get confused, that’s just some random fantasy of yours. QUOTE WHAT YOU ARE DISCUSSING!
I love folks like you who want to lecture people on things you don’t understand.
Fourier transforms are only one of a number of ways to completely decompose a signal. Periodicity transforms can do it. SubBandAnalysis can do it. PCA can do it. Tensor decomposition can do it. Low frequency PCA can do it. S-Transforms can do it. A Non-Uniform Filter Bank can do it. Wavelets can do it. Your idea that “anything else would not give the correct signal” is a joke.
And more to the current point, the “Complete” in Complete Ensemble Empirical Mode Analysis (CEEMD) means that unlike EEMD, it contains all of the information necessary to give what you call “the correct signal”. When you add up all the different empirical modes together (C1 through C6 plus the residual) you reconstruct the original signal perfectly.
So no, my friend, I fear you truly don’t grasp the underlying idea of decomposing signals into simpler signals. You’ll have to go lecture someone else. I’m not interested in being talked down to and patronized by you, that’s no fun at all.
Please address future comments to anyone but me, thanks.
w.
Willis,
You said that Fourier analyses get “fooled” I wrote confused. I am happy to go with fooled. But the same
point applies. Taking a Fourier transform of a sympony will tell you whether or not you need a piccolo player but not whether they can go home after the first five minutes. Fourier analysis tells you one particular thing – namely the amplitude of the complex exponential that you need to reconstruct the signal. Changing the amplitudes of those exponential will give you a different answer when you reconstruct the signal.
Of course you can write a signal as a sum of polynomials or any other set of orthogonal functions. Each
decomposition tells you different things about the signal. CEEMD does not tell you about what periods are
present – they do not give you a spectrogram. For that you still need to calculate the instanteous frequency
of each mode. Looking at it by eye is not good enough – the brain sees patterns everywhere as you have pointed out.
Germonio, was I truly that unclear in my last post? I’ll give it another shot.
Go away. Slope off. Not interested. You’re addressing the wrong man. Talk to the hand, the head ain’t listening. You’re too stupid to quote my words despite being asked to do so both here and in other threads. You tried to stuff words in my mouth. Disappear. Try your schtick on someone else. Go bother yer momma, she might not be able to see through you. You’ve burnt your bridges with me. I’ve been through this movie with you both here and elsewhere, not interested in re-runs.
You getting the picture?
w.
Hi Willis,
Which level of the pyramid of refutation does your response fit with? Name calling i.e. “too stupid”
or something lower?
Germonio writes “Which level of the pyramid of refutation does your response fit with?”
Communication happens best, IMO, when participants speak similar languages and use similar styles. It really doesn’t work for one person to use formal language while the other uses vernacular. Being able to wrestle in the mud is at times a useful skill if that is where everyone happens to be.
I wonder about your choice of handle or online name; it could be a clever reference to the book Guns, Germs and Steel, about forces that civilized such places as are civilized, or it could be a misspelling of “Geronimo”.
i love how geronimo tells you how to do it.
then claims to have done it.
then….
doesnt show his work.
geronimo.
you cannot refute willis with words about methods.
you have to show your work.
numbers
data
code.
or no cookie.
Germonio writes “Firstly Fourier transforms do not get confused – if you want to reconstruct a signal as a sum of complex exponential”
Fourier transforms do not use exponentials; but rather sines and cosines harmonically related and in varying amplitudes.
Confusion in this case is “aliasing” or violation of the Nyquist Theorem.
Michael you are wrong when you say: “Fourier transforms do not use exponentials”
..
Here is the definition of a Fourier transform: https://wikimedia.org/api/rest_v1/media/math/render/svg/97ad0938a279c4846d42a4bbd212f6a1f0ca4c0f
…
That little “e” in the formula is an exponential.
Thank you for showing me the existence of Fourier Transform as an integral; I now agree it contains an exponential (or it can do so as one representation). I have used Fourier Transform as shown as figure 1 of https://en.wikipedia.org/wiki/Discrete_Fourier_transform which is a way to actually implement DFT in a computer using sines and cosines.
While I’ve known a relationship exists between “e” and sine and cosine, until today I haven’t focused my attention on that interesting phenomenon.
From wikipedia:
Euler’s formula, named after Leonhard Euler, is a mathematical formula in complex analysis that establishes the fundamental relationship between the trigonometric functions and the complex exponential function. Euler’s formula states that for any real number x
e^( i x) = cos x + i sin x
https://en.wikipedia.org/wiki/Euler%27s_formula
Willis – love your work – just wondering, however:
Would you consider some day looking into the “sudden stratospheric warming events”? We just had one, as you’re undoubtedly aware, and I think I noticed something that might be interesting:
During this latest event it stood out to me: Where the stratosphere warmed pressure at the surface increased, and where it cooled the pressure was lowered. Now I know you are interested in the tropical management of temperatures, so here’s the hook:
The lower stratosphere cooled in the tropics as the warming in the Northern hemisphere progressed. The response was immediate and fairly radical: Tropical SST rose under the cooling umbrella while Arctic SST cooled under the warming umbrella.
Now I think, that these phenomena can be seen as signs of suppressed/enhanced convective regimes governed by the lower stratospheric density/temperature – acting as a radiative filter where convective processes end at the tropopause. Essentially: back-radiation seem to be highly sensitive to density witch would make physical sense.
Interestingly these stratospheric warming events seem to have absented themselves prior to the major warming event of the late nineties during a period of high solar activity.
Dispensing with a fruitless debate whether cycles can be identified or not, we might make some headway into the mechanisms of short term temperature variation. I am convinced that you are the guy who could do that – would you? please…
Per
Speaking for myself, I have enjoyed and learned from both Javier and Willis over the years. As far as I know, Javier has done some great work on major Holocene events, and beyond that the Pleistocene, especially glaciation and de-glaciation. My sense as a lay person is that it is difficult to identify any one “external” factor that “caused” massive (to us) climatic changes, but there are three things that change with the earth in relation to the sun: axial tilt or obliquity, orbital precession, and eccentricity. There is reason to think that while any of these alone might have little effect, two (or more) together could have been the main drivers of significant climate change.
Willis, as far as I know, has focussed more on short-term solar changes, and whether they can be linked to short-term temperature or climate change. He keeps saying no, as he does in this post and comments. He teaches us to be skeptical about leaps from small amounts of data, which only speak clearly when they are treated in a specific way, to big picture explanations. If four smaller data sets do not support a conclusion, but support for the conclusion suddenly appears when you stack the data sets, there is still a problem with your conclusion. I enjoyed Javier’s recent work on the northern hemisphere cryosphere (not “the globe”); noticeable or significant temperature increase, low humidity/water vapour; ice loss. Judith Curry weighed in with soot/albedo. I can’t resist adding: no known harm so far to any living species, including homo sapiens.
The warmists have made big careers out of saying they have identified the one critical man-made factor that causes blah blah blah. There is a temptation to jump in and say something like “no you idiots; here is the one critical factor”; water vapour or sun or whatever. Maybe every “symptom” that is being discussed is quite complex (I won’t annoy Mosher by saying “part of a chaotic system”); maybe sea level, for example, is affected by additions of water from deep underground–small changes from year to year, but with a significant long-term effect, including a surprising stability as opposed to dramatic or frightening changes (Jim Steele on Curry’s site). One thing I like in Willis’s work is the reminder that the climate system, assuming that phrase even conveys much meaning, is remarkably stable, and short-term events that are quite dramatic from a human perspective generally come to an end with a return to a status quo we are used to. There is an overall resiliency in nature, and of course we are the species that can adapt to change better than any other. We are supposed to prove we are good people by saying “I believe climate change is real” (today’s profession of faith, without which persecution may be in order); the consensus doesn’t want us to say “I believe climate stability and resiliency are real.”
Anyway, keep up the good work.
Willis, This is boring. What the heck are you trying to do? The rock crystals are from glacial advances in the middle of a interglacial period.
Did you forget about the ocean sediment analysis that you presented in this forum? (O16/O18 ratio in the shell of a marine animal that is deposited in the ocean sediment is analyzed to determine past surface temperatures.) The ocean surface temperature cools up to 10C, cyclically.
Do you know what a Bond event is? Have you heard of the super large Bond event that is called a Heinrich event?
Have you heard about the Younger Dryas (12,900 year abrupt cooling event) or the 8,200 year cooling event? The observations are real.
The glacial/interglacial cycle is real. It physically happened.
It is a fact, that there are periods of millions of years in the paleo record when atmospheric CO2 has been high and the planet is cold and vice versa. There is not even correlation of atmospheric CO2 levels and planetary temperature in the paleo record.
It is a fact that the planet’s climate changes cyclically both poles. It is a fact that after 30 years there is no mechanism, no physical explanation to explain the past cyclic climate change.
http://www.agu.org/pubs/crossref/2003/2003GL017115.shtml
Greenland ice temperature, last 11,000 years determined from ice core analysis, Richard Alley’s paper. William: As this graph indicates the Greenland Ice data shows that have been 9 warming and cooling periods in the last 11,000 years.
http://www.climate4you.com/images/GISP2%20TemperatureSince10700%20BP%20with%20CO2%20from%20EPICA%20DomeC.gif
http://wattsupwiththat.files.wordpress.com/2012/09/davis-and-taylor-wuwt-submission.pdf
then comes the question and the point of the article: why does the bond event sometimes just get supressed?
the Bond events are called Oescher Dansgaard events during glaciations and here you see how irregular they appear
http://1.bp.blogspot.com/_tG8JCC_Tnp0/TK3hw4fmpOI/AAAAAAAAAFo/oJUYikI5zuQ/s1600/Pleistocene_Holocene.JPG
that’s also why they called it “events” not a cycle as they don’t appear like clockwork but they seem to appear in bundles
Nothing has changed since Wally Breocker’s Model’s to the rescue paper.
Wally’s comment below comment is correct, that climate scientists do not have a first order idea (do not have a clue) what causes abrupt climate change such as the Younger Dryas or 8200 BP cooling event and almost no one can imagine the planet’s climate during the 100,000 year glacial phase.
Wally it is the sun.
Wally it is the sun.
You have no evidence for that. And what made the Sun vary. The sun is a million+ times bigger than the Earth.
Why the magnetic fields at the solar south(?) pole when the field switches polarity at the end of cycle!
That, and the big iron core, and large magnets that are waved around in the Suns magnetic field.
I would like to respond, but your comments make no sense. Try again.
I believe we’ve discussed that the next solar cycles was strongly influenced by the residual magnetic field, I think at the south pole, when that field switches polarity.
Do I remember that correctly?
I am not sure what you are hinted at. You may find your answer here:
http://www.leif.org/research/Super-Synoptic-Maps-and-Polar-Fields.pdf
Lief your comment concerning a lack of physical evidence to solve the problems is incorrect.
There are piles and piles of independent physical evidence, in peer reviewed papers, to solve the problems.
The trick to solving the problems was following the different fields, pulling out the anomalies and paradoxes, organizing them, and then looking for possible solutions.
But what is the point in arguing without looking at the observations in question?
Up above, Andy May claimed that “the solar/climate connection investigation is just getting started”. I replied:
Inspired by Andy’s comment, I thought I’d take a look at the wheat data. I went to the excellent UN Food and Agriculture Organization (FAO) dataset, and got the wheat yield for all countries for the years of record, 1961 – 2016. There are 85 countries which have complete records for all of those years. I looked to see the p-value of the correlations, using the method of Koutsoyiannis to adjust for autocorrelation.
The result? Out of the 85 countries with complete records of wheat yields, not one showed a p-value less than 0.05, the usual threshold in the climate field. Minimum p-value for the 85 countries is 0.16, far from significant. Nor do things get better when I look at the cross-correlation between sunspots and any country’s wheat yield.
So … once again I’ve looked for a connection between sunspots and a surface dataset, in this case wheat yield, and found nothing. And although as I’ve said many times this doesn’t prove anything, because you can’t prove a negative, it is one more in the increasingly long list of places that I’ve looked for such a connection and found nothing.
Best to you all,
w.
Irregular natural cycles come in a great variety of bandwidths, which dictate the degree to which they resemble the behavior of strictly periodic cycles. That’s why power spectra (F. transforms of autocovariance functions) are used in serious signal analysis of random geophysical time-series. There are trade-offs to be made in such analysis between the confidence limits and the frequency-resolution of the spectral estimates.
While raw periodograms, such as employed here, are useful in analyzing strictly periodic time-series, they provide notoriously unreliable, asymptotically inconsistent estimates of the spectral density that characterizes the underlying random process. It’s a grave mistake to conclude on that basis that the natural oscillations are “pseudocycles,” as if their unpredictable irregularity makes them physically less real.
This is for Andy:
You have in recent posts expressed several misconceptions about TSI [PMOD in particular]. I don’t think you inventted those yourself, so I am interested in where you got them from.