Earth Science Supercomputers race to predict storms

Discussion in 'Earth Science' started by Cinderloft, Sep 19, 2004.

  1. Cinderloft

    Cinderloft Premium Member

    http://edition.cnn.com/2004/TECH/internet/09/16/hurricane.supercomputers.ap/index.html


    This would be a boon to forecasters. While even if it comes through, it would not be 100% accurate (nothing ever is) but it would be the most accurate forecast available, and with updates constantly feeding the simulation, you would always have the most probable scenerio available.

    Mods: I wasnt sure whether to put this in the Tech forum or here. Do with it as you feel necessary.
     
  2. pineappleupsidedown

    pineappleupsidedown Premium Member

    I wonder how the competition would be with those. If the flap of a butterflies wing in south america can affect the waves on the other side of the ocean, how would the computer be able to factor that in?

    ---pineapple
     
  3. crayon

    crayon New Member

    I used to work for THE supercomputer people. Up until 2 months ago I was a computer programmer for Cray Inc. (formerly Cray Research Inc.).

    I have often thought about what would be required to give a 100% accurate prediction for something further out than a few hours.

    The number of variables is astonishing. Water temp, air temp, plate movement, currents, winds, ice caps, the butterfly effect, animal and plant life, air pollution, beached whales, and the list goes on and on. And not only would the list of things to measure be a mile long, the number of measurements needed would be beyond anything we could possibly imagine. I doubt there are enough harddrives in the world to store all the data. Not to mention a computer that can crunch numbers fast enough to keep up with all the data comming in.

    I am all for it though. If we can predict the path of a storm even a few hours earlier than we can now, lives would be saved. And that is the name of the game I think.