Sunday, December 13, 2015

Simulation, Calibration, And Chaotic Systems

     With the recent conclusion of the Paris Conference has arisen a fresh set of cries from the warmistas: this time, to the effect that “there’s an international agreement now, so it’s time to cease your blasphemous dissents.” As usual, they point to their simulations, chant that “the science is settled,” and expect everyone to fall in line. And as usual, they’re furious that we’re not doing as they ask. One fellow with whom I recently crossed swords gave me this for a valediction:

     You dipsticks are terminal failures to mankind. Let's hope your wilful ignorance of science extends to medicine and you die a long, slow, painful death of dick bleeding disease.

     Charming, eh? But that’s what you get from persons who claim to march under the banner of “science” without having the faintest idea what science is.

     Among other things, simulation is not science – and as I’m an expert on real-time simulation, I believe I can speak on this subject with some authority.


     Simulation techniques can be useful to scientists, but in and of themselves, they are not science. The essence of science, without which the word means nothing, is scientific method:

  1. Collect a substantial amount of real-world data centered on a phenomenon of interest.
  2. Note patterns in the data, specifically with regard to sequences of events in time.
  3. Formulate a hypothesis about what causal mechanism might lie behind those patterns. (Steps 2 and 3 constitute the inductive phase of the scientific method.)
  4. Use the hypothesis to predict specific outcomes from related contexts and stimuli applied to them.
  5. Design experiments that can test the above predictions. (Steps 4 and 5 constitute the deductive phase of the scientific method.)
  6. Perform the experiments:
    • If the predictions come true, the hypothesis has been confirmed, and testing should continue. (Note: Confirmation is not proof; no scientific hypothesis is ever regarded as “proven.”
    • If the predictions don’t come true, the hypothesis has been falsified and must be discarded. From here, one must return to steps 2 and 3 to try again.

     The applicability of simulation to the scientific method is at step 3. Once a causal mechanism has been hypothesized, a simulation of that mechanism can be used to explore it further: in particular, to make the predictions which subsequent experiments will test. However, the simulation cannot and does not confirm the hypothesis. How could it? It’s an artificial environment under the complete control of the simulation writer. He has dictated its starting conditions and its rules, which don’t necessarily match reality...and reality is notably indifferent to anyone’s opinions about it.

     Only an experiment in the real world, with strictly specified and enforced starting conditions and a strictly specified and reproducible stimulus, can confirm or falsify a scientific hypothesis. The warmistas cannot produce such an experiment – no one can – because the Earth is an open system. So they fall back on simulations which:

  • Omit consideration of admittedly mysterious atmospheric, oceanic, and solar phenomena;
  • Employ disputed, heterogeneous data sets that they refuse to disclose;
  • Require repeated adjustments to that data, hewing to no widely agreed-upon standard;
  • Have yet to produce a result markedly similar to real-world conditions;
  • Cannot even reproduce previous eras’ climatic developments.

     As one of my physics professors once said of a bit of theorizing, this is so bad it’s not even wrong.


     A few words about calibration and chaotic systems and I’ll close for today.

     The atmosphere and oceans are empirically regarded as chaotic. Most laymen don’t understand the term, which is an arcane one that pertains to the equally arcane realm of perturbation theory. To be as brief as possible while remaining clear, in a chaotic system, there is no way to predict the extent of the system’s response to a proposed stimulus, regardless of the nature, magnitude, and duration of the stimulus. The response’s magnitude, or its scope, or its duration, or some combination thereof, can (and sometimes will) exceed any proposed upper bound.

     This is of central importance to the problem of calibrated measurement. In a chaotic system, if one’s measurements are not absolutely exact, then there is a possibility that either the starting conditions or the magnitude and duration of the stimulus will differ sufficiently from the experimental proposal to produce completely irrelevant results.

     There has never been a measurement technique for any measurable item or event that provides absolute accuracy. In the nature of things, there never will be one. Thus, experiments proposed to probe the laws that govern a chaotic system are inherently problematic. We can observe the behavior of such systems at a macroscopic order of magnitude. We can hope that the patterns we see in them provide some basis for prediction. But no matter how often we’ve used the pattern successfully, those predictions will sometimes be wrong, and there is nothing we can do about it.

     Ask your local weatherman.


     The idea of a law against bad weather was once a commonplace joke about the limits of governments and legislation. (The Soviets would have loved to pass such a law, but then, they’d have loved to make a lot of counterfactual things true.) Weather – the instantaneous condition of the atmosphere in a particular region – is merely one of the aspects of a chaotic system. Climate – our large-scale, long-term expectations about weather patterns in a given region – can never be anything but chaotic. No simulation can capture all its oddities. It’s not merely a matter of “not knowing enough yet.” Chaos is inherent in an open system of fluids whose dynamics are influenced by Brownian motion and turbulence.

     Max Born once said that should the opportunity be afforded him in the next life, he would ask God two questions: “Why relativity?” and “Why turbulence?” Frankly, I’d like to know the answers myself. But I think I already know what He would say:

“They make the game interesting.

     Hard to argue with that, isn’t it?

7 comments:

  1. Fran,

    Buy a copy of the book "The Chilling Stars" by Nigel Calder and Henrik Svensmark. The book describes the excellent work done by Svensmark in detail concerning the effect of the interplay between the solar magnetic field and cosmic rays from the rest of the galaxy on Earthly cloud formation, and hence on our weather and climate. Humanity have no control whatsoever over either the solar magnetic field or galactic cosmic rays, the best we can do is pray; which does not sit well with many people.

    A weak solar magnetic field lets more galactic cosmic rays reach the Earth, which nucleate more clouds, which reflect more sunlight back into space, cooling the earth. The opposite happens when the solar magnetic field is strong (assuming a constant flux of galactic cosmic rays). The Earth's magnetic field is a minor player in this interaction.

    The solar magnetic field, galactic cosmic rays, and the Earth's temperature in the past all leave various markers in the Earth and in meteorites (usually isotope variations) showing how intense they were at a given time in the past. Carbon dioxide levels in the atmosphere has only a minor effect on the Earth's climate; the solar magnetic field and the galactic cosmic ray flux being the two major drivers of climate change by far, over millions of years.

    Many people do not want to hear this, and do not want anybody else to hear this, either. The book appeared in bookstores for only one week when it first came out. I bought a copy immediately, which was a good thing, since I have never seen it in a bookstore since then. You will probably have to order your copy from Amazon or a similar web site; apparently it has been declared contraband by the far left.

    I hope this helps in your search for the truth about climate change.

    Karl Henriksson

    ReplyDelete
  2. I've just ordered a copy, Karl. It looks interesting, but as I said in the essay, as experiments are impossible, we'll never have anything but hypotheses.

    Myself, I severely doubt that human activity is a significant driver of the world climate, but I have to admit that the matter will remain forever open to dispute. One thing that's not in dispute, though, is that the Earth hasn't warmed appreciably for nearly twenty years...and the warmistas are unable to match that to any of their theses or models.

    ReplyDelete
  3. What is really annoying, and dangerous, is the conflation of "normal" and "average." If the high temperature of the day, for instance, is 56 degrees fahrenheit, and it happens to be 4 degrees above average, it is reported as 4 degrees above normal. There is no normal, only mathematical averages. There always will be above and below averages.
    Get ready for a windmill in your back yard.

    ReplyDelete
  4. You should submit that to Watts Up With That, Fran. I think Anthony would post it.

    A good argument for responding to the Warmists is to go retrieve the Holocene temperature record graphic from Wikipedia. And simply ask, "show me the problem."

    Global warming was sold as something both dramatic and unprecedented. It is neither. There is nothing happening that is inconsistent with our geological climate period (the Holocene), which began with the last really serious warming -- the emergence from the last Ice Age.

    ReplyDelete
  5. Recently, I spent a very long evening discussing climate change with a bunch of young, recently-graduated liberal arts majors (in the area of English and Literature). As a point of reference, I am a degreed meteorologist with >30 years experience in the following areas:

    - Operational weather forecasting

    - Weather satellite data processing for numerical weather prediction

    - Software development and implementation for optical, IR, and microwave radiative transfer analysis

    - Computer modeling and simulation

    - Solar/terrestrial interaction and effects on terrestrial weather

    - Climate modeling under a NASA program I worked on for a few years

    Yet, these liberal arts folks refused to accept any of my points and assertions (complete with published journal references), instead relying only on the information with which they were indoctrinated (very little of which they understood in the first place). Needless to say, the discussion proved fruitless....

    ReplyDelete
  6. I sympathize, Mark. My God, how I sympathize! Without going into too much detail, allow me to say that my experience in the modeling of large fluid systems is extensive -- and there is not one single thing that's predictable about them over the long term. It's a master's course in humility for the simulation expert.

    Smooth flow, laminar flow, differential flow, flow under uniform forces, flow under shear forces, flow under pulsed forces, flow under single-ended and double-ended thermodynamic influences, flow under longitudinal and transverse perturbations...I left it behind more than thirty years ago, and I still have nights I wake up in a cold sweat over it. And these...persons think they can predict what the average global temperature will be a hundred years from today? Pour me a tot of what they're drinking -- a large one. I could use it.

    ReplyDelete
  7. Well, you said what I've said at least 10 times in various discussions about this topic over the years, though you've covered the detail more thoroughly than I could. All I can say is I know a little about chaos theory, and what that implies. The only reason I ever bothered with this topic is I have a degree in computer science, and I studied atmospheric science a bit while I was in college. I had an interest in meteorology from the time I was a child, and thought I would go into the field. Then I discovered computers... While getting my degree, my CS professors discussed what was known about simulations that tried to model chaotic, non-linear systems at the time (this was around 1990). They said it was extremely hard to achieve anything close to accuracy, and didn't think the pursuit had a promising future. When I took economics, I heard the same thing. Economists used computers to do some game theory on their ideas, but never for prediction, because it was known they were not accurate. All they could do is look for patterns that would give them new ideas they hadn't considered.

    In the last 15 years, I've heard of climatologists using supercomputers to try to predict the climate, a coupled, chaotic, non-linear system, to no avail (though they swear by it), and I've heard of financial experts using such models to try to predict economic trends, also to no avail (recall the '08 financial crash). I keep thinking they could have listened to the experts in the field of computer simulation to avoid being misled by this shiny new technology, but somehow I don't think they wanted to.

    ReplyDelete

Comments are moderated. I am entirely arbitrary about what I allow to appear here. Toss me a bomb and I might just toss it back with interest. You have been warned.