Fine tuning Physics

Defects of computer models of solar system

Spread the love
Dr Sheldon
Rob Sheldon

Physicist Rob Sheldon writes to respond to the claim that Jupiter does not shield Earth:

Kevin Grazier, PhD, at the NASA Jet Propulsion Laboratory in Pasadena, describes the study in which he simulated the evolution of tens of thousands of particles in the gaps between the jovian planets for up to 100 million years. Based on the results, Dr. Grazier concludes that the widely reported shield role attributed to Jupiter is incorrect. More.

Sheldon notes,

While I have the greatest respect for people who run tens of Newtonian mechanics simulations of 10,000 bodies for 100 million years, the models only give results that were programmed into the simulation. If, for example, tomorrow someone were to find that the magnetized solar wind could impart a tangential drag to charged planetesimals, then all of those simulations would be worthless.

Accordingly, we should be very careful to qualify the results of a simulation which a brief exploration of a small corner of phase space. It is as if I should declare it impossible to drive to work because my keys were not on the hook by the door. Contrast this conclusion with the observation that my son had wrapped the car around telephone pole on the street. The first is an inference, the second a certainty.

We (and I speak for all scientists), have gotten so lazy that we confer upon our models the same certainty as observation. We call our computer runs “data”, even when our computer models are doing forward-modelling, Monte-Carlo integration of infinite-dimensional subspaces.

I am not simply arguing for modesty and good manners, but against the distortion of the education of a generation of young scientists who take these pronouncements as equally valid as astronomical statistics on binary stars. I taught high school astronomy from a leading college Astronomy textbook and was astonished how as we went from planetary to stellar to cosmology, the degree of speculation and wrong-headed models grew exponentially. I have nothing against speculation–I have been accused of practicing it too enthusiastically–but we should be very careful to label it as such.

See also: Paper: Jupiter doesn’t shield Earth, comets kickstarted life. Researcher: Jupiter teams with Saturn to kick a significant fraction of the particles into the inner Solar System, beneficial for the development of life.

Follow UD News at Twitter!

7 Replies to “Defects of computer models of solar system

  1. 1
    velikovskys says:


    Robb Sheldon:
    While I have the greatest respect for people who run tens of Newtonian mechanics simulations of 10,000 bodies for 100 million years, the models only give results that were programmed into the simulation. If, for example, tomorrow someone were to find that the magnetized solar wind could impart a tangential drag to charged planetesimals, then all of those simulations would be worthless.

    Accordingly, we should be very careful to qualify the results of a simulation which a brief exploration of a small corner of phase space.

    We (and I speak for all scientists), have gotten so lazy that we confer upon our models the same certainty as observation. We call our computer runs “data”, even when our computer models are doing forward-modelling, Monte-Carlo integration of infinite-dimensional subspaces.

    “Jupiter as a shield ” is also a model based on simulations not observations.

  2. 2
    anthropic says:

    Two words: Shoemaker-Levy.

  3. 3
    velikovskys says:

    One word: Chicxulub

  4. 4
    mahuna says:

    velilovoskys @ 3

    And since Chicxulub NOTHING. Except minor events like Tunguska. For serious cosmologists, the benefit of a Large Shield, like Jupiter, and a Last Ditch Shield, like Luna, is that they reduce the odds of Chicxulub to once every 100 million years and Tunguska to once every 10 million years or something. If Earth received a Chicxulub once ever million years and a randomly located Tunguska ever 10,000 years, advanced life would not be possible.

    Even with our Shields, the Toba Event right here on Earth severely reduced the number of humans. If such Earthly disasters were only slightly more common, humans may have survived as a species, but we would never have established cities or any technology much beyond the spear.

  5. 5
    velikovskys says:

    Muhuna:
    And since Chicxulub NOTHING. Except minor events like Tunguska

    For serious cosmologists, the benefit of a Large Shield, like Jupiter, and a Last Ditch Shield, like Luna, is that they reduce the odds of Chicxulub to once every 100 million years and Tunguska to once every 10 million years or something.

    You mean that is the model. The paper discusses that model and objections to the assumptions.

  6. 6
    kairosfocus says:

    News, a good clip of sobering modesty. S-L9 does demonstrate that the gas giants can shield, and a moonful of craters the same. We need to be very reserved about substituting simulation outputs for data, and we need to be very sensitive to how empirical support is very provisional on pain of affirming the consequent. KF

  7. 7
    frediklen says:

    Quite a bit is curious about our solar system and every day I’m finding the impossible to be possible and the other way around. Good for you for keeping an open mind. FRED from solarpanelsuit.com

Leave a Reply