RealClimate logo

Curve-fitting and natural cycles: The best part

It is not every day that I come across a scientific publication that so totally goes against my perception of what science is all about. Humlum et al., 2011 present a study in the journal Global and Planetary Change, claiming that most of the temperature changes that we have seen so far are due to natural cycles.

They claim to present a new technique to identify the character of natural climate variations, and from this, to produce a testable forecast of future climate. They project that

the observed late 20th century warming in Svalbard is not going to continue for the next 20–25 years. Instead the period of warming may be followed by variable, but generally not higher temperatures for at least the next 20–25 years.

However, their claims of novelty are overblown, and their projection is demonstrably unsound.

First, the claim of presenting “a new technique to identify the character of natural climate variations” is odd, as the techniques Humlum et al. use — Fourier transforms and wavelet analysis — have have been around for a long time. It is commonplace to apply them to climate data.

Longyearbyen, at Svalbard

Using these methods, the authors conclude that “the investigated Svalbard and Greenland temperature records show high natural variability and exhibit long-term persistence, although on different time scales”. No kidding! Again, it is not really a surprise that local records have high levels of variability, and the “long-term persistent” character of climate records has been reported before and is even seen in climate models.

The most problematic aspect of the paper concerns the Greenland temperature from GISP2 and their claim that they can “produce testable forecasts of future climate” from extending their statistical fit.

Of course, these forecasts are testable – we just have to wait for the data to come in. But why should extending these fits produce a good forecast? It is well known that one can fit a series of observations to arbitrary accuracy without having any predictability at all. One technique to demonstrate credibility is by assessing how well the statistical model does on data that was not used in the calibration. In this case, the authors have produced a testable forecast of the past climate by leaving out the period between the end of the last ice age and up to 4000 years before present. This becomes apparent if you extend their fit to the part of data that they left out (figure below).

The Greenland temperature from GISP2 and a repeat of Fig. 8 in Humlum et al. (2011) (solid red) with two extended prediction (red dashed) with and without accounting for the trend fit.

I extended their analysis back to the end of the last ice age. The figure here shows my replication of part of their results, and I’ve posted the R-script for making the plot. The full red line shows their fit (“model results”) and the dashed red lines show two different attempts to extend their model to older data.

For the initial attempt, keeping their trend obviously caused a divergence. So in the second attempt, I removed the trend to give them a better chance of making a good hind cast. Again, the fit is no longer quite as remarkable as presented in their paper.

Clearly, their hypothesis of 3 dominant periodicities no longer works when extending the data period. So why did they not show the part of the data that break with the pattern established for the 4000 last years? According to the paper, they

chose to focus on the most recent 4000 years of the GISP2 series, as the main thrust of our investigation is on climatic variations in the recent past and their potential for forecasting the near future.

One could of course attempt to rescue the fit by proposing that some other missing factor is responsible for the earlier divergence. But this would be entirely arbitrary. Choosing to ignore the well known (anthropogenic) factors affecting current climate, on the other hand, is not arbitrary at all.

Humlum et al. also suggest, on the basis of a coincidence between one of their cycles (8.7 years) and a periodicity in the Earth-Moon orbital distance (8.85 years), that the Moon plays a role for climate change (seriously!):

We hypothesise that this may bring about the emergence of relatively warm or cold water masses from time to time in certain parts of oceans, in concert with these cyclic orbit variations of the Moon, or that these variations may cause small changes in ocean currents transporting heat towards high latitudes, e.g. in the North Atlantic.

How wonderfully definitive. They, however, admit that their

main focus is the identification of natural cyclic variations, and only secondary the attribution of physical reasons for these.

So, if the curve-fitting points to periodicities that are anywhere near any of the frequencies that can be associated with a celestial object, then that’s apparently sufficient. You can get quite a range of periodicities if you consider all the planets in our solar system, their resonances and harmonics, see any of Scafetta’s recent papers for more examples. And of course, if there are parts of the data that do not match the periodicity you believe to be there, you can just throw it away to make the cycles fit. Quite easy really.

In short, Humlum et al’s results are similar to those I discussed concerning meaningless numerical exercises, and their efforts really bring out the points I made in my previous post: arbitrarily splitting a time series up into parts generally does not allow one to learn anything.


  1. O. Humlum, J. Solheim, and K. Stordahl, "Identifying natural contributions to late Holocene climate change", Global and Planetary Change, vol. 79, pp. 145-156, 2011.

144 Responses to “Curve-fitting and natural cycles: The best part”

  1. 51
    Ray Ladbury says:

    Isotopious, I think it was Joan Baez who said, “If you have a choice between a hypothetical situation and a real one, take the real one.” Nearly 40 years of warming is pretty damned strong evidence that we are doing something–and then there’s the physics. Denialists seem to always want to ignor the physics. They also hate it when you point out that the warming was predicted over 115 years ago. A confirmed prediction counts as very strong evidence in science. Prediction is far more important than explanatory power. You say we should wait.

    We already have over 20% increase in the proportion of Earth’s land area in drought. We already have km wide plumes of methane bubbling up in the Arctic. We are already on track to have an ice-free Arctic before 2050. Rising temperatures are already decreasing crop yields. The health of the oceans is already being adversely affected by rising temperatures and acidification.

    Meanwhile, the human population is on track to crest (we hope) at around 10 billion sometime around midcentury. We will have to meet the needs of all of those people in an environment where petroldum has become scarce and expensive at a time when our agricultural methods have become critically dependent on it.

    Given all those facts and the strength of evidence that the current warming epoch is anthropogenic, what possible argument could a sane person construct for waiting? If we are not responsible, our progeny may not refer to the current epoch as the anthropocene, but rather as the anthropo-obs-cene.

  2. 52
    Hank Roberts says:

    “If you have the choice between a hypothetical situation and a real one, choose the real one.”
    – Joan Baez (to Michael Krasny, Forum, KQED radio, Feb. 4, 2003)

    See also her cousin John Baez:

    “… 3 points for every statement that is logically inconsistent.
    5 points for each such statement that is adhered to despite careful correction.
    5 points for using a thought experiment that contradicts the results of a widely accepted real experiment….
    … 10 points for arguing that while a current well-established theory predicts phenomena correctly, it doesn’t explain “why” they occur, or fails to provide a “mechanism”….”

  3. 53

    #52–Very amusing, and apposite, too. I wonder if this has ever actually been used? If so, what scores have been racked up?

  4. 54
    John E. Pearson says:

    # 52 Hank linked to:

    “10 points for each favorable comparison of yourself to Einstein”

    Not only Einstein, but lots of others. Recently someone one on RC compared themselves to Poincare.

  5. 55
    John E. Pearson says:

    51 Ray wrote: “We already have km wide plumes of methane bubbling up in the Arctic.”

    and the NY Times wrote too:

  6. 56
    Septic Matthew says:

    52, Hank Roberts: … 10 points for arguing that while a current well-established theory predicts phenomena correctly, it doesn’t explain “why” they occur, or fails to provide a “mechanism”….”

    Very droll. But why only “well-established”? Some “well-established” theories were critiqued in this fashion long before they became well-established. to my knowledge, gravity has never been explained (“why” or “how”) since Newton declined to “feign” hypotheses. Einstein turned “attraction” into curvature of space without explaining “how” or “why”, and that took a while to become “well-established” in its turn. Does any one know why the “universe” is expanding — i.e. what caused the explosion, or what mechanism held it all together just “prior” to the explosion?

  7. 57
    Brian Dodge says:

    @ Isotopious — 16 Dec 2011 @ 11:11 PM

    Your “Internal Combustion Engine Analogy” is internally inconsistent. If the engine changes temperature by 1deg/50RPM, and you change the RPM from 2350 to 2550 at t=1min, then the temperature would equilibrate one minute later(t=2) at +4 degrees((2350-2550)/50 = 200/50 = 4), or 111 deg(107+4), not the 108 deg you show.

  8. 58

    #49 Donald Oats

    The truth never loses. The truth is always the truth. It is us humans that can lose out when we ignore the truth.

  9. 59
    Chris Korda says:

    #51 sounds like a Church of Euthanasia sermon. How about Anthrobscene? It would make a good bumper sticker. My finely-honed spider sense for slogan marketing says: surf’s up!

  10. 60
    Brian Dodge says:

    Since Humlum et. al. is paywalled, I can’t see their references – is the prior work in this area by KLYASHTORIN and LYUBUSHIN,“CYCLIC CLIMATE CHANGES AND FISH PRODUCTIVITY”, which found cycles with periods of 16.8, 25.6, 26, 32, 32.5, 38.6, 39, 53.9, 54, 55.3, 60, 60.2, 50-70, 72, 75.8, 76, 99, and 108 years in the climate record consistent with their results?

  11. 61
    Paul Vincelli says:

    Folks, we scientists outside the realm of climate science depend on your expertise to sort the “wheat from the chaff”. (I guess it won’t surprise you to learn I am an agricultural scientist.) Frankly, some papers on climate change are too deep for me, but as a scientist, I understand the value of the peer-reviewed literature. I am helping to lead an effort to educate Extension agents (and ultimately farming communities) on climate change, and the most powerful thing I can point to is the peer-reviewed literature. So while blog exchanges are helpful, the most powerful rebuttal will be a thoughtful peer-reviewed response.

  12. 62
    David Horton says:

    #47 Thanks Eric, anyone contemplating the idea that the temperature rise of the last 4 decades can be duplicated by tssing some coins should read the Thomson PNAS paper. Would be useful if the exercise was repeated now with an additional 20 years data or so. The odds that the link between CO2 and temp are due to chance must now be infinitely small. And in turn the link between both and melting ice on cap and glacier, sea level and acidity, changes in ecology of plants and animals, increases in record and extreme weather events. The idea that what is happening to our planet is just a random natural event we happen to be witnessing right now is so absurd it falls over as soon as voiced for any intelligent person.

  13. 63
    mike says:

    *Septic Matthew* – living as we do in a rural area, our bodily waste products from the lavatory etc drain into a septic tank. Septic has always meant to me at any rate, poisonous, infected, or the accumulation of potentially disease inducing matter in a concrete lined underground tank that must be cleaned out periodically. I associate septic with pus, turds, and other such putrid material. Your very cogent posts are well written with correct spelling. It is puzzling then to see you describe yourself as “septic” rather than a sceptic/skeptic? Surely you do not want us to think of you as a turd? Or is this an attempt at irony? If it is, I have to tell you that the precise meaning of your irony laden name evades me. Are you really a sceptic about global warming etc, and present yourself as the microbe that reveals and heals the diseased wound afflicting a healthy body? Or can’t you spell?

  14. 64
    Pete Dunkelberg says:

    SM is racking up points!
    Hey SM did you know that an infinite regress of “Why?” is only in your imagination? Newton didn’t explain why F = mA so lets not depend on it.

    Here’s a “Why?” question for you: Why isn’t this in the open thread?
    Here’s another: Why do some people resort to questioning the most basic science when they don’t like a particular finding?

  15. 65
    EFS_Junior says:

    You can find a copy at the climate denier website klimarealistene;

    See also;

    IE9 with Google Toolbar can translate the Norwegian for you, now if I could only find a PDF translator, see for example;

    I have some more “stuff” but I’ll leave you with an easy riddle;

    Replace the letters e, h, and l with d in someone’s name and what do you get?

  16. 66
    Ammonite says:

    Isotopious 45: I am reminded of a statistician and a pit crew mechanic watching a car race. The statistician has an outsider’s view and makes deductions based on relative position and lap time. The mechanic has an insider’s view and makes deductions based on race congestion, fuel load and tire wear. What is the chance that they are both wrong when they independently reach the same conclusion?

  17. 67
    Isotopious says:


    Since there is a better chance that one will reach the right conclusion if they both make different conclusions, then I would say they are more likely to be wrong if they reach the same conclusion.

  18. 68

    #55–Thanks for the NYT story; it’s a good one, despite a couple of simplifications that grated me just a bit. A tease:

    “Dr. Walter Anthony had already run chemical tests on the methane from one of the lakes, dating the carbon molecules within the gas to 30,000 years ago. She has found carbon that old emerging at numerous spots around Fairbanks, and carbon as old as 43,000 years emerging from lakes in Siberia.

    “These grasses were food for mammoths during the end of the last ice age,” Dr. Walter Anthony said. “It was in the freezer for 30,000 to 40,000 years, and now the freezer door is open.”

    Unfortunately, a lot of it was also a bit–alarming. Increased wildfire has been well-documented in several more southerly locales around the world, but apparently it’s (anecdotally) increasing in tundra country, with highly carbonaceous results.

  19. 69

    #29 Vukcevic

    Problem is that the program proposed does not constrain model mis-specification error, even if its terms are all physical. An alternative is an intensely data-based model using stochastic processes, where the calibrations and dependencies come from the data series themselves.

    An example of this is addressed in the recent AGU presentation,

    The idea is to subsume the entirety of behavior in a windowed sample of the system-under-study’s recent past, and base diagnostics and remedial measures entirely upon that window’s charcterization.

    These could very well be informed by efforts to determine, for instance, marginal change in climate depending upon incremental adds of greenhouse gasses, from, for example, paleoclimate studies.

    Ultimately, a physics-based model would be expected to have greater explanatory and confirmatory power. Still, as a preliminary informant of further experiment and policy, it seems the empirical model is pretty good.

  20. 70
    Barb says:

    If I understand correctly, this article was peer reviewed. I find this really unsettling. Having only achieved a grade 12 science level over 20+ years ago, there is no way that I could ever determine that results were attained by “shoddy” practice. I depend on peer review as a guarantee that results & conclusions have been achieved using “best practice” available. Considering the paper’s claim that “most of the temperature changes that we have seen so far are due to natural cycles” goes against almost all credible findings this far, the review panel should have used even greater scrutiny. A quick google search of this Dr. Humlum should have alarmed the review panal. He acts as a science adviser to this Heartland Institute funded ICSC

    I am angry that this article was even published. Whoever reviewed this paper and gave it the green light did so with the intention of muddying the waters. It doesn’t matter if the paper gets trashed by the scientific community, it’s already out there, and not many people will check back to see what the reviews are. People will accept it as fact, or if by chance they hear about bad reviews, they’ll see proof that there really is debate about what causes global warming.

    I’m worried.

  21. 71
    vukcevic says:

    #69 Jan Galkowski says:
    Still, as a preliminary informant of further experiment and policy, it seems the empirical model is pretty good.

    Mr. Galkowski
    If your comment refer to this link:
    it is appreciated. Your suggestion although appropriate, is beyond my capacity and competence to withstand scrutiny of a professional statistician, however if you or anyone with appropriate skills is interested in pursuing the matter to the degree required for publishing, my email is on the graph in the above link.
    Mr. Galkowski thanks again.

  22. 72
    JGarland says:

    @41…That statement I do agree with.

    But it is not automatic and/or done in a short time.

    It took ~100 million deaths over a period of 6 decades post SG’s report to convince even a portion of the planet that tobacco kills.

  23. 73
    Ray Ladbury says:

    Chris Korda, As a general rule, I try to avoid introducing any term that contains in order the letters t, h, r, o and b. However, I would note that humans are showing themselves to be no more adept at avoiding the consequences of population biology than a colony of yeast in a bottle of beer.

  24. 74
    Ray Ladbury says:

    Ah, I see that Isotopious subscribes to the theory that a watch that is broken is more accurate than a watch that is miscalibrated by a microsecond, as the former will be right twice a day. Dude, did you ever even take a science class?

  25. 75

    Unfortunately I don’t have access to the original paper but from the description here it appears to be rubbish. A purely statistical analysis that doesn’t take into account physics can at best predict short term phenomena or situations that are exact analogues to past variations with similar causality to situation they purport to be modelling. To “disprove” anthropogenic change they are more or less claiming that the underlying causality has no effect. You have to be a brave scientist to claim that curve fitting trumps physics. Or stupid. Funny how denialists cling to arguments like “it’s just a model” or “it’s just statistics” when the numbers don’t agree with their prejudice but they are very happy to accept this sort of garbage when it suits their case.

    The difference between denial and skepticism: a skeptic rejects any bogus argument or evidence whether they like the conclusion or not; a denialist only rejects findings that don’t support their prejudices.

    On the positive side, check out these pix of animals near my home on my new blog. I hope they will give anyone tiring of the fight against destruction of the biosphere a lift.

  26. 76
    Hank Roberts says:

    Thank you Barb at 18 Dec 2011 at 2:21 AM for pointing out that press release.

    This is what Dan O’Neill long ago called the “over-the-counter-culture” — today it’s the “over-the-counter-consensus” — it’s about the money with these people, always the money, always short term.

    The PR stuff quotes:

    “… ICSC chief science advisor, Professor Bob Carter of James Cook University in Queensland, Australia, and author of … “ Climate: the Counter Consensus ” says, “Science has yet to provide unambiguous evidence …. New Zealand-based Terry Dunleavy , ICSC founding chairman and strategic advisor … ICSC energy issues advisor, Bryan Leyland of Auckland, New Zealand. … ICSC science advisor Professor Ole Humlum, of the Institute of Geosciences, University of Oslo …”

    I wonder what the editor of that journal was thinking?

  27. 77
    Hank Roberts says:

    Elsevier’s journal info says it requires authors to disclose the funding sources for their paper; was anything about funding on this disclosed in the published work somewhere? I’m curious.

    While the paper is paywalled, the figures in miniature are viewable with thye text of the captions– click the figures tab on the page:

  28. 78
    Hank Roberts says:

    “Climate4you” is Humlum’s climate blog, by the way. It’s read-only.
    Some of the claims and charts might be worth scrutiny — there’s a cloud-cover-vs-temperature chart for example, where he takes data from different published sources and draws a line showing change over time. I wonder if he published.

  29. 79
    Ray Tomes says:

    There is plenty of evidence for significant cycles in all ranges of periods. The Milankovitch cycles of around 400Ka, 100ka, 40ka and 23ka are widely accepted and are attributable to astronomical configurations. Cycles of 2300 and 208 years are found in many climate series as well as solar proxies. There are other longer and shorter period that are also well established.

    I do agree with you that finding a cycle of 8.7 years and attributing it to the 8.85 year lunar cycle is unjustified. The agreement over 4000 years data should 0.01 year or better unless the lunar period has changed over that time. Theer is too much of this sort or error made.

    Russian researchers (Afanasiev and others) have found evidence for the 9.3 year lunar cycle (half 18.6 year cycle) in very long salt deposit series. Using this data Afanasiev has devised a method that he calls “Nanocycles Method” which uses the change in the period and its interaction with the annual seasonal cycle to accurately determine the age of geological deposits for the last 600+ma. This work is well established in Russia and it is a pity that more Western geologists and climatologists are not aware of it.

    [Response: You need to be careful here. The cause of the Milanovitch cycles is very well established: predictable variations in the earth’s orbit, and hence the amount of sunlight received as a function of season. The seasonal cycle likewise of course. Furthermore, these cycles cause huge (tens of watts/m^2) changes in insolation, which means there is a clear physical mechanism for their influence on climate. They *should* have an influence on climate on purely physical grounds, and that’s why their influence is easy to detect on statistical grounds. Everything else you are referring to has either no known physical mechanism underling it or involves a known but very very small effect (e.g. .in the case of the the 11-year sunspot cycle, which is of order 0.1 W/m^2). If it exists, the supposed link between the moon and salt deposition could, I suppose, be related to the tides, but this would still have nothing to do with climate.–eric]

  30. 80
    flxible says:

    “The Milankovitch cycles of around 400Ka, 100ka, 40ka and 23ka are widely accepted and are attributable to astronomical configurations.”
    Sounds a lot like the sun revolving around the earth?

    Ray Tomes shtick wrt climatological cycles appears to be based here.

    CAPTCHA: head Thatvice

  31. 81
    Dan H. says:

    Interesting read, but I am not sure how relevant it is. Meanwhile, as Eric stated, Milankovitch cycles are fairly well known with predictable effects. Solar cycles are less well known, except for the 22-year sunpsot cycle, as are their effects.

  32. 82
    Ray Tomes says:

    To Flxible: Raymond Wheeler did much to bring attention to cycles in climate, but I would not consider his work definitive because the proxy data available at that time were so much less well known than at present. Here are my analysis of BE10 fluctuations and C14 variations over periods of about 10,000 years each.

    Response to response above: Agreed that Milankovitch cycles are stronger and so more easily established. However the nature of cycles is that as more data becomes available it is possible to establish the significance of weaker cycles also. See the above links please. I didn’t mention the cycle of 350-355 years in my first comment, because I wanted to limit that to cycles proven beyond reasonable doubt. But a cycle of ~355 years was reported by Chizhevsky more than 50 years ago, and shows up in both of these series also.

    There is a lot of well established cycles information that is not taught or understood well in Universities, simply because it is so interdisciplinary. I would highly recommend the paper by Edward R Dewey of the Foundation for the Study of Cycles (which went defunct about 1998 and should not be confused with the present FSC):

    “… insofar as cycles are meaningful, all science that has been developed in the absence of cycle knowledge is inadequate and partial. Thus, if cyclic forces are real, any theory of economics, or sociology, or history, or medicine, or climatology that ignores non-chance rhythms is manifestly incomplete, as medicine was before the discovery of germs.”

    – Edward R. Dewey (1967)

  33. 83
    Cugel says:

    Ray Tomes : “[I]nsofar as cycles are meaningful” makes Dewey’s statement a truism. The scientific (or economic, or whatever else) job is to determine whether or not they are indeed meaningful, a rather more difficult task than simply finding apparent cycles in time-series. It is only achieved by determining actual causes. While cycles may point to physical causes and thus be worth investigating, many won’t be.

    The last 30 years of climate data is not going to have much influence on, for instance, a 355 year cycle, nor will such a cycle say anything about the physics behind AGW. It is actual current events which are pertinent to the very different atmospheric conditions which mankind has created and continues to change. AGW explains much of those events, don’t you think?

  34. 84
    David B. Benson says:

    Ray Tomes @82 — To have cycles, even pseudocycles, there has to be a form of ‘elasticity’ in the physics. While there are a few (but of very much lesser importance) such configurations, more important is a strong human tendency to preceive patterns where actually none exist. Looking a short time time series of
    suggests ‘cycles’ where in fact none exists.

    I even have strong reason to doubt the apparant ~2 ky ‘period’ in Variability of El Niño/Southern Oscillation activity at millennial timescales during the Holocene epoch

  35. 85
    Dan Lufkin says:

    Analysis of cycles goes way back in meteorology and things just got worse with the invention of the fast Fourier transform. The only requirement on data for a successful Fourier analysis is that the series be bounded and (Lebesgue) integrable.

    The USAF once published a Fourier study of sunspot numbers that was so bad I tried to make fun of it with a study that related flying accident rate to unit number via a cosine function. It explained 80% of the variance.

    The lab that published the sunspot junk withdrew the report but got their revenge by submitting my study to the flying safety office. It took months to convince them that I was kidding.

  36. 86
    Ray Tomes says:

    David B Benson @84: Yes, there needs to be some form of restoring force for displacements (or what you called elasticity) for cycles to exist. You are right that people look for patterns – it is what science is about. In cycles research we use Bartel’s significance test to see whether we are justified in believing in a cycle. The p~0.05 significance of the 207.7 year cycle in C14 might be dismissed on its own, but the p<0.0002 significance of a 206.9 year cycle in Be10 confirms this as a strongly significant cycle.

    Cugel @83: Yes, as stated a truism. But if you look at Dewey's paper (which is just a tiny fraction of the research done) then the existence of cycles in every field of scientific research is well established. Also, agreed that finding a cycle does not tell us the cause, but if several matching cycles periods are found in conceivably related fields then it gives a strong hint for further study.

    Rather than use Fourier analysis which limits us to an integral number of cycles in our data, we allow the cycle frequency to vary continuously and determine the frequency to 100 times that precision and typically 10 times that accuracy.

  37. 87
    Ray Tomes says:

    In my view the question is not whether humans or natural cycles (or natural randomness) cause climate change. Very clearly both do. The question is how to break down the fluctuations into the various parts so that the human component can be isolated and the natural cycles can be extended.

    The levelling off in temperatures since about 1998 was entirely expected on the basis of the dominant cycles. The 2300 year cycle is rising for centuries to come. The 208 year cycle reached a maximum around 1998, and so was rising for the entire 20th century and will fall for the entire 21st century. The quasicycle of 50 – 60 years was at a high in 1941 and 1998. Lows therefore occur around 1970 and ~2025.

    If one adopts a view that only one of these factors is THE cause of fluctuations then one must be proved wrong in the future, whether or not humans act differently. Only by recognizing all factors can we hope to get near to the right answers.

    [Response: If it was all ‘entirely’ expected on the basis of your cycles then a) they should have been detectable prior to 1950 and all subsequent changes should have been predictable on the basis of what was known then. In which case, the ‘cyclists’ should have been warning people about increasing temperatures in the 1950s (perhaps you have a cite for that?). Otherwise, you are left with a post-hoc fit to the temperatures that doesn’t even come close to ‘recognizing all the factors’ and which has no predictive power. – gavin]

  38. 88
    Ray Ladbury says:

    David Benson, and then of course, there is my favorite example of periodicity in the first 10 digits of the base of Napieriand logarithsm, e-2.718281828…, which being transcendental,of course, cannot be periodic. The human brain is programmed to look for order, and it will find it whether it is there or not. It is risky to posit periodicity or even quasi-periodicity without understanding some sort of mechanism that could be driving it.

  39. 89
    CM says:

    EFS_Junior #65, re: the Norwegian denier group Klimarealistene,

    Count yourself lucky you don’t have a PDF translator. The booklet you refer to is slightly more coherent bunk than their first booklet, due to the involvement of Humlum and others with academic backgrounds, but bunk it is. The bit of it that is relevant to the present discussion: There’s no fun-with-Fourier stuff, but Humlum pulls a trick to hide the incline that uses the central Greenland reconstruction. This issue has been addressed at SkepticalScience (compare this for a slightly different version of local post-1855 warming).

  40. 90

    #87 Ray Tomes

    Interestingly, random fitting is what many denialists are accusing modern scientists of these days, when in fact the are not simply just looking at the patterns and saying ‘hey, these things go up and down’.

    This is because they are identifying the mechanisms. Raymond H Wheeler did not identify mechanisms, he merely said ‘hey, I think I see patterns’. Well, many that tried LSD also saw patterns… Many that look at the vastness of interactions in the universe see patterns. The trick is identifying the mechanisms that drive the changes…

    Funny how those that are most confused about climate change often claim change often claim the kettle is black and while simultaneously… such as it’s cooling and we are heading back to an ice age, or it’s warming, but it’s natural cycle…

    Can you answer this for me: How can it be cooling and warming at the same time?

  41. 91
    Ray Ladbury says:

    Ray Tomes: “Only by recognizing all factors can we hope to get near to the right answers.”

    NO! NO! NO! NO! NO!

    It is by no means necessary to understand everything to make reliable predictions. What is needed is to understand those factors and forcings contributing signficantly to the system. That is the biggest problem with the “Fun with Fourier” types. They make no assessment of physical significance or mechanism. Let them find enough cycles and they will fit any curve just by estimating Fourier coeffficients for the series.

  42. 92
    Jerry Steffens says:

    Imagine that a 29th-degree polynomial (for example) were found to be an excellent fit for a drunkard’s path home from his neighborhood pub. Now, suppose that this curve fit were used to predict the drunkard’s precise path home on subsequent trips. Clearly, it wouldn’t work since the staggering movements of the drunk are random; there is no particular physical reason that he lurches to the right at one spot and to the left at another. However, we CAN predict with a reasonable amount of confidence that our inebriated friend will end up at a particular address — his home. (If he’s sober enough to walk, he’s likely to be sober enough to remember where he lives.)

    Similarly, curve fits of fluctuations in past weather patterns are unlikely to have any predictive power. However, based on known physical laws, we can predict with a great deal of confidence that if greenhouse gases continue to increase we will arrive at a specific address — a warmer world.

  43. 93
    Dan Lufkin says:

    The Foundation for the Study of Cycles is still chugging along, apparently fueled by stock-market techies. The amazing power of cycles is best appreciated by looking at and clicking on the index of articles published in the foundation’s quarterly journal. It has everything from sunspots to caterpillars.

    The history of cycles in meteorology is largely that of Charles Greeley Abbott, who did pioneer work at Harvard in infrared astronomy and measuring the solar constant. He became enamored of cycles in the 1920s and published a couple of books and dozens of papers on the subject. His prestige drew a lot of effort into cycle study, but by the 1940s nobody had yet come up with a useful cyclic forecast. In the 1950 Compendium of Meteorology, an authoritative mid-century review of the state of the science, the cyclic approach is entirely dismissed for its lack of physical foundation.

  44. 94
    David B. Benson says:

    Since several of the comments address what I would otherwise have attempted to write (probably less eloquently), I will instead simply attempt to cite a poem [from memeory so I might not have it quite right]. This should be recited by every cyclist before deciding to (not) offer up yap, yet another pattern.

    I met a bear upon the stair.
    I met a bear who wasn’t there.
    He wasn’t there again today.
    Oh how I wish he’d go away.

    [Response: Beautiful. Is that A.A. Milne or…? -eric]

  45. 95
  46. 96
    David B. Benson says:

    Eric response @94 — The ghost of
    thanks you.

  47. 97
    sidd says:

    “The little man upon the stair,
    The little man who isn’t there
    He isn’t there again today,
    I think he works for the CIA.”

  48. 98
    Ray Tomes says:

    I find it quite amusing that so many people consider the use of cycles as some sort of hocus pocus. Are you aware that tides are predicted accurately this way? And many other things. You examples of curve fitting are all silly. It is essential to look at the statistical significance.

    Gavin @87 reply: The data available today are much better than in 1950. As are the computers. Why not start in 2000? It was already clear then that the 50-60 year cycle would decline until 2025 and the 208 year cycle would decline until 2100, although the 2300 year cycle will continue upwards. Do some searches for yourself on these cycles and you will find them well established. In the two links I provided for the Be10 and C14 proxies, the phase of the 208 year cycle is very clear and a peak around 2000 AD was not at all hard to predict.

    [Response: We’re very well aware of the presence of near-periodic variability in sunspots. And yes, we’re also well aware of the tides. The objection here is to the misuse of these facts to make unsubstantiated claims about the predictability of climate.–eric]

  49. 99
    Ray Tomes says:

    For links to the mentioned cycles, please see Seuss (or de Vries) cycle of 208 years and Hallstatt cycle of 2300 years (other spellings also)

  50. 100
    Ray Ladbury says:

    Ray Tomes,
    Since you are concerning yourself only with the frequency and not with the physics of how much a particular “cycle” contributes in terms of forcing, would you care to explain how your approach differs from simply approximating any time series with a Fourier series? How is that even interesting?