the Air Vent

Because the world needs another opinion

DeWitt Payne – FARIMA Mann 08

Posted by Jeff Id on September 1, 2010

I’m sure I freaked out some of the science minded around here with my posts on Lomborg and Lee. It’s just a blog though and it’s about whatever catches my attention next.  You are all welcome to contribute and dilute my opinions at your whim.  DeWitt Payne, is very much science first, and he’s done something which is very interesting.  Unfortunately, getting the documentation/post from him is sometimes like pulling teeth.🙂 LL

I’m going back to science stuff for a while, unless someone drops another crazy link on a thread.  Basically, after some time the politics of climate become so sickeningly bad that it’s impossible to listen to.

What DeWitt discovered/revealed was that with a different red noise model matched to the Mann08 proxy data than I used in my posts, he was able to retain 31 percent of the series, even though they had zero signal.

DeWitt Payne said

September 1, 2010 at 8:00 pm e

After a few false starts with the loop and array indexes, I created 1209 synthetic random series with the same parameters as the Mann infilled data with an ARFIMA (1,d,1) model. There were 4 series that returned AR coefficients greater than 1 and one with an AR coefficient less than -1. I set those coefficients to 0.999 and -0.999 so I wouldn’t see warnings about non-stationary series. <b>I still get acceptance of 31% of the no signal series, i.e. a Hockey Stick. That’s after scaling all the series to a mean of zero and an sd of 1. </b>I’ll have to try not scaling, but if I don’t scale, the average over all series is not zero. But then maybe it shouldn’t be. Adding the signal 1:1 with the noise, the acceptance was 80%. At a ratio of 5x noise to signal, the acceptance was about 40% with the usual behavior of squashing the ‘reconstructed’ sine wave signal relative to the calibration period.

What he’s done apparently,  is run some of the R code from my previous posts with his own modifications.  I used a (1,0,0) model which matches Mann07,  whereas he has used a higher order better fitting methods.  What’s very interesting in DeWitt’s work is this quote from Mann08 SI:

Although 484 (40%) pass the temperature screening process
over the full (1850–1995) calibration interval, one would expect
that no more than 150 (13%) of the proxy series would pass
the screening procedure described above by chance alone

DeWitt has just demonstrated another confirmation of the fact that chance correlation in M08 was substantially underestimated.

This means that unless Mr. Payne has made a horrible error equal to my own independent and yet to be discovered horrible error, Mann has dramatically underestimated the number of proxies which would pass by random chance.  If this percentage were not the single basis for the claim that the proxies held temperature info in them, it would have little meaning.  Unfortunately for Paleo………

What I find additionally interesting is that Mann passed 40 percent of the proxies by correlation, whereas in M08 ,  Luterbacher’s work had 71 series of actual temperature data included (or 6% of the total series).  So according to DeWitt, 31% by  chance plus 6% by direct insertion of Luterbacher temperature series and we’re at 37%.  Mann found 40 percent from the quote above,  and claimed that was proof of non-random signal.  In engineering we test the instrument, in paleoclimate, they correlate and use obscure stats.   My recent AR1 post at CA and here estimated a conservative 8.5% signal to noise amplitude, so with DeWitt’s recent work, everything is coming into alignment.

There isn’t a statistically demonstrable signal in these Mann08 proxies.

27 Responses to “DeWitt Payne – FARIMA Mann 08”

  1. Brian H said

    I.e. Hockey Stick = Carefully selected noise.

  2. boballab said

    I can’t help myself, I think Mann has been moonlighting here:

    There are certainly many options available to players looking for the perfect hockey stick. With all of the different hockey stick brands available it would be almost impossible not to find one to your liking, but for the fussy or hard-to-please hockey player, it can also mean a lot of experimenting and research. Most companies manufacture a selection of wood hockey sticks and composite hockey sticks as well as stick blades and stick shafts. However, with the popularity of the composite hockey stick, some smaller manufacturers are concentrating on producing a few select composite models only.

    http://www.prohockeystuff.com/hockey-sticks/brands.aspx

    Now just change the words a little:

    There are certainly many options available to Climatologists looking for the perfect hockey stick. With all of the different hockey stick reconstructions available it would be almost impossible not to find one to your liking, but for the fussy or hard to please Climatologist, it can also mean a lot of experimenting and research. Most Paleo-Climatologists mannufacture a selection of wood hockey sticks and composite hockey sticks as well as stick blades and stick shafts. However, with the popularity of the composite hockey stick, some unknown deniers are concentrating on reproducing a few select composite models only.

    So what do ya think? About sums it up?

  3. MikeN said

    Perhaps this is the answer to a question I meant to ask. Why is Mann’s result so sensitive to Tiljander and bristlecones? Given the algorithm, and its performance against random noise, I would expect it to be able to produce a hockey stick given any large number of proxies. Are the proxies actively working against a hockey stick, to the point that they are worse than random data?

  4. Brian H said

    Mike;
    To the extent that the proxies reflect reality, then the answer to your question is, “Yes; necessarily so!”

  5. cementafriend said

    Off topic but I do not know how to otherwise to get DeWitt Payne’s attention.
    What does he think of the following http://claesjohnson.blogspot.com/2010/09/updated-computational-blackbody.html and download the paper.
    Maybe Jeff would like to make a separate post. There are others casting doubt on the photons concept. Miskolczi in the abstract of his new Aug 2010 paper has “Apparently major revision of the physics underlying the greenhouse effect is needed”
    Chasing that hypocrite Mann is waste of time if the whole AGW concept based on the Trenberth cartoon falls in a heap.

  6. Jeff Id said

    #5

    “it shows that the concept of “backradiation” as the driver of the “greenhouse effect” supposedly causing global warming, is fictitious non-physical and confusing,”

    Be careful what you read, studies have shown that goofy science facts can lower IQ.

  7. Kenneth Fritsch said

    This means that unless Mr. Payne has made a horrible error equal to my own independent horrible error, Mann has dramatically underestimated the number of proxies which would pass by random chance.

    Jeff, I must have missed something here – to what horrible error are you referring? I think you are attempting to say an error as your own possible, but yet to be discovered error that turned the Mann 08 conclusion upside down.

    Also when do we get to see the details of DeWitt’s analysis?

  8. Jeff Id said

    Kenneth,

    I fixed the wording — independent and yet to be discovered horrible error.

  9. Jeff Id said

    I was hoping DeWitt might share, but he can’t be at all the blogs all the time😉

  10. Brian H said

    #5;
    Einstein vindicated, Born walks the Planck, and takes probability QM with him? Oh, frabjous day!

  11. DeWitt Payne said

    Jeff,

    This is still a work in progress. For one thing, the ‘na.rm=TRUE’ in the rowMeans command doesn’t seem to work as expected. I noticed this when I hadn’t changed the array size to 1000,1209 instead of 1000,10000. Everything was squashed down compared to the calibration line. It’s summing the correct numbers, but it’s dividing by the full array dimension, not the actual number of objects using the cormask logical vector. The obvious work around is to create a new matrix with the correct dimensions and use rowMeans on that. But I shouldn’t need to do that.

    I did do an ARIMA(1,0,0) fit to the synthetic data and got a histogram of the rho values that looks somewhat like the histogram for the actual proxies, except for a large spike in the -0.1-0 bin.

  12. DeWitt Payne said

    One of the problems with trying to write this up is that there are no goodness of fit statistics for the coefficients when using the fractional difference model as implemented in the fArma package.

    There’s a lot of work I’d like to do as well. I really need to do replicates for each set of coefficients, or perhaps a designed experiment, to see if there is a specific volume in the three dimensional coordinate space where it is more likely that hockey sticks will be found. If there is, then I have to look at the proxies that passed the Mannian screen and see if they do, in fact occupy that volume.

    I don’t have time the rest of this week to do any more.

  13. cementafriend said

    Jeff (6)I was inclined to not comment but just having read the post http://motls.blogspot.com/2010/09/hawking-churches-hit-back-at-heretic.html I just thought that maybe the concept of photons is wrong and it you who does not understand. I suggest that you read Professor Claes Johnson article. He indicates he accepts string theory. Do you regard that Stephen Hawking has a low IQ? Is it ethical to regard Professor Claes Johnson mathematical treatise “goofy science”? It has been shown that radiation can be treated as waves (note the name micro-waves). It has been proven that waves can be canceled.
    My own measurements in furnaces (flame temperatures, flue gas composition, flue gas temperatures, heat balances etc) have shown that flame temperatures can not be higher than the theoretical based on energy input. One can not create energy by radiation within our earthly boundaries.
    Have you read the article by Chilinger et al 2008,”Cooling of the Atmosphere Due to CO2 Emission” Energy Sources 30 ? This relies on Thermodynamics which many engineers should understand but clearly most climate scientists do not.

  14. DeWitt Payne said

    Re: cementafriend (Sep 5 03:14),

    One can not create energy by radiation within our earthly boundaries.

    And one doesn’t. The energy comes from the sun. That means the upper limit on temperature for solar derived radiant energy is ~5,880K. That leaves plenty of room for an increase of up to a few degrees from an average surface temperature of 288K.

    Waves cancel, but not everywhere all the time unless the waves are exactly the same frequency and exactly 180 degrees out of phase with exactly the same polarity and originating from exactly the same point. Otherwise, they only cancel at some places and reinforce at other places.

    I’ve counted a lot of photons over the course of my career doing atomic spectroscopy.
    Don’t try to tell me they aren’t a useful concept.

    Do you accept that electrons are particles or at least have many of the characteristics of particles? If so, then how do account for electron diffraction? A wave only description of EM radiation does not account for the photo-electric effect, or do you think Einstein was wrong about that and didn’t deserve his Nobel prize?

  15. cementafriend said

    DeWitt Payne (14) I am not an expert in Einstein’s field so I can not answer if he was right or wrong. Some people, who are cleverer than I am, believe that he did not deserve his Nobel prize. However, on all matters, I have an open mind and am willing to learn something new(and judge if someone is trying con me and other people- like believers in AWG).
    I read Hawking’s early book “A brief history of time” a couple of times and found it interesting but could not agree with all his thoughts. I will be interested to read his new book. I have preferred concepts of multidimensions. I once wrote a spreadsheet program for my Apple II (I still have the red book) which poked information into multiple dimensions (eight I think)
    I recall in Chemistry learning about Linus Pauling’s theory for which he got a Nobel prize and then learning about his attempts to twist his theory to fit properties his theory could not explain. My (3)daughters who did chemistry at University have never heard of him. His Nobel peace prize was also probably not deserved. His attempts in the medical field (babies, vitamin C) did not go well in my country.
    Do you think Al Gore or Yasser Arafat deserved their Nobel Prizes?
    Anyway, back to the topic, I have never found, in my measurements of properties and heat balance calculations any back radiation from a cold receiver which increased the temperature of a hot source. If I was a reviewer of Kiehl and Trenberth’s 1997 paper I would have rejected it as contrary to the fourth postulate (second law)of Thermodynamics. This postulate (law) has not been falsified. If the hypothesis of particle like photons is in conflict with the laws of Thermodynamics then I would have to reject that no matter who put the hypothesis forward or who supported that. Einstein himself said it takes only one person to falsify an hypothesis.
    From my reading of peer reviewed articles and various text books there are many, many, others who have found radiation (long or short) can not flow from a cold source to a hot.
    Finally, I am of the considered opinion that the best way to defeat the pseudo science around AGW is to point out that the proponents do not understand,in total or even partially, the engineering subjects of Thermodynamics, Heat Transfer, Fluid and Particle Dynamics, Reaction Kinetics, Process Measurement and Control, Mathematics and Statistics, and Economics which all play a part in the complex assessment of earth’s climate. One could also add electrical and magnetic fields.
    regards and keep strong.

  16. Jeff Id said

    Radiation flows from all non-zero temperature objects. The net sum is from hot to cold but the net sum is affected by the coldness of the receiver.

  17. DeWitt Payne said

    Re: cementafriend (Sep 7 09:30),

    The Science of Doom has a lot of detailed articles on how the greenhouse effect works and why it doesn’t violate thermodynamics. An article that might be helpful as a starting point is: Do Trenberth and Kiehl understand the First Law of Thermodynamics?. There are lots more that should answer all your objections if you read them with an open mind.

    Comparing the Nobel Peace Prize to the Physics or any of the other hard science prizes is, to use a William F. Buckley word, nugatory.

  18. Brian H said

    #16;
    Jeff, DeWitt;
    Prof. Johnson’s take is not that radiation is not emitted by the cooler body and does not arrive at the hotter, nor is it that incoming waves are cancelled. It is that there is a resonance phenomenon such that the warmer material cannot be warmed by the lower frequency radiation from the cooler except very inefficiently. Only higher frequency EM than its own temperature-determined resonance can do so. The result would be that such low-f IR would reflect and scatter, I presume being re-absorbed (within the usual few meters) by CO2, which would heat up and either/both lose heat in thermal collisions or re-re-emit some of it (a swift logarithmic declining fraction of a fraction would be re-re-…-emitted, etc.) In any case, little or no surface heating would result, and what little did would black-body radiate away at frequencies higher than CO2’s fingerprint range.

    This seems to me to be a perfectly valid hypothesis (it is in principle quite falsifiable, I believe. I am also reminded, though I don’t have access to the original, of G&T’s citation of the thermodynamics engineer Schach, “… the radiative component of heat transfer of CO2, though relevant at the temperatures in combustion chambers, can be neglected at atmospheric temperatures. The influence of carbonic acid on the Earth’s climates is definitively unmeasurable.” (Because it is negligibly small.)

    It should be noted that even the implied slight air warming would not result in much or any surface warming, since the specific heat and conductivity of the gas is so low, and the actual energy involved at the widest edge of the swift log decline of re-radiation is a very small fraction of the incident EM.

  19. DeWitt Payne said

    Re: Brian H (Sep 7 16:53),

    Once a photon of a given frequency is emitted, how does it know it’s coming from a warm body or a cool body? Conversely, how does the absorbing body know that the photon comes from a body that’s warmer or cooler? Is there some new property of photons other than frequency/wavelength that physicists have previously ignored? I’d read Professor Johnson’s paper, but after G&T and Beenstock & Reingewertz (unit roots), my capacity for putting up with drivel has been exceeded.

  20. DeWitt Payne said

    In spite of myself, I scanned Johnson’s paper, extremely unconvincing. I’m reminded of Ptolemaic astronomy. You can describe the observed positions over time of all the planets by adding enough epicycles. In fact it’s easier to calculate eclipses by using an Earth centered coordinate system. But having the planets orbit the sun is easier for everything else. The concept of the photon and quantum physics, particularly quantum electrodynamics , has allowed extremely accurate predictions of quantities like the anomalous magnetic moment of the electron, and the Lamb shift of the energy levels of hydrogen.

  21. Brian H said

    Neither Claes nor I (though I’m not an advocate as such) are making any claim of photon knowingness. That’s a misrepresentation, I think. It’s the frequency, pure and simple. He postulates, I wot, that (e.g.) a body b-b emitting at x wavelength cannot (efficiently) absorb and be warmed by EM of x/n wavelength, where n is a constant or variable, somewhat >1, which determines a “cut-off”. This seems like a perfectly valid falsifiable hypothesis to me.

  22. DeWitt Payne said

    Re: Brian H (Sep 7 19:52),

    The problem is that Claes’ theory requires Kirchhoff’s Law, absorptivity = emissivity, to be invalid. Light impinging on an opaque surface can only be reflected or absorbed. That’s required by conservation of energy. The best approximation to a black body is a box with a light absorbing inner surface maintained at constant temperature with a small hole in it. I don’t see how that would give you a black body spectrum if Claes is correct about the implications of his theory. If he’s not correct about the implications and Kirchhoff’s Law is still valid, then the predictions of his theory are no different than for quantum theory. The main reason a colder object has only a small effect on a warmer object is that the radiative flux is proportional to T^4.

    We haven’t scratched the surface on molecular and atomic emission/absorption spectra either.

  23. Suibhne said

    DeWitt Payne said

    ….”The problem is that Claes’ theory requires Kirchhoff’s Law, absorptivity = emissivity, to be invalid.”…….

    Kirchhoff’s radiation law is far from being universally accepted.
    http://www.highbeam.com/doc/1G1-181856625.html

    Its interesting to note that Feynman in his three volume review of Physics did not include any mention of Kirchhoff’s Radiation Law.
    I think that the best that can be said for it is it is works in certain circumstances but it is certainly not a Law say in the category of say the Inverse Square Law of Gravitation

  24. RuhRoh said

    Mr. DeWitt Payne;

    Perhaps you will be so kind to give one further response to comment ~165 on the Radiative Physics Simplified II thread.
    I’m still stuck in the past, and unable to get closure on the nagging question of where to find the +5 K of ‘CO2 warming’ on Mars.

    Regretting off-thread comment, but work was consuming all free time.
    TIA
    RR

  25. Layman Lurker said

    #11 DeWitt Payne

    This is still a work in progress. For one thing, the ‘na.rm=TRUE’ in the rowMeans command doesn’t seem to work as expected. I noticed this when I hadn’t changed the array size to 1000,1209 instead of 1000,10000. Everything was squashed down compared to the calibration line. It’s summing the correct numbers, but it’s dividing by the full array dimension, not the actual number of objects using the cormask logical vector. The obvious work around is to create a new matrix with the correct dimensions and use rowMeans on that. But I shouldn’t need to do that.

    So would I be correct by saying your masking vector retains the values of the 1209 and replaced the other 8791 with NA’s? I can’t picture where your snag could be. I tried rowMeans (x, na.rm = TRUE, dims = 1) using a simple 3X2 dimensional array with NA’s and the missing values were properly omitted from the mean calculation.

  26. One of the longest continuous thermometer temperature records for a single location is that for Northern Ireland dating from 1796 – as shown on my Home page here http://climate-change-theory.com

    Being an island, this would be fairly representative of North Atlantic ocean temperatures which themselves affect the Arctic and perhaps even other oceans.

    Such records show absolutely no sign of a hockey stick.

  27. DeWitt: No. You are incorrect in stating “The problem is that Claes’ theory requires Kirchhoff’s Law, absorptivity = emissivity, to be invalid. ” You yourself have said (on SoD) that gases only start to absorb when the emitter starts to get warmer than the gas. This surely is one example of Claes conclusion in practice. Kirchhoff’s Law applies to a true blackbody which is surrounded by space – such as the whole Earth system including its atmosphere. Such a blackbody is thus insulated from its surroundings and can only transfer energy by radiation – never by conduction etc.

    The Earth’s surface is not a true blackbody because it is not insulated from its surroundings, namely the atmosphere and the rest of the crust and mantle below. Hence, it does not have to immediately re-radiate the SW radiation which it converts to thermal energy. We know it can store thermal energy from day to night and from summer to winter in a given hemisphere. But, yes, it does immediately re-radiate any IR backradiation with exactly the same intensity and spectrum, so no energy is left for any warming by such backradiation. It is as Claes says, acting like a high energy filter, converting SW to thermal energy, but not absorbing LW.

    Regard your question on SoD (where I cannot respond due to censorship) regarding lasers: these do not emit spontaneously with frequencies linked to their physical temperature. What they emit is induced radiation, not spontaneous radiation. There is also a blowing process which no doubt causes some cooling when cutting metal. Remember, it is really the frequency which determines whether absorption takes place. This frequency is only proportional to actual temperature (as per Wien’s Displacement Law) when the emission is spontaneous.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

 
%d bloggers like this: