An on-line article states:
(http://www.skeptics.ca/articles/yack-computers.html)
>"Computers can accurately forecast the weather"
>This myth is easily dismissed, and not just by personal experience. >In the last half-century we've learned that the accurate prediction of >the future state of a "chaotic" system like the world's weather is not >a time-consuming computation problem; it's theoretically >impossible. Dynamical systems (the technical term) are inherently >unpredictable, because they exhibit what is called "sensitive >dependence on initial conditions". So computers forecasting next >year's weather are no more believable than computers squaring the >circle, or flying faster than the speed of light.
I believe the above to be rubbish - given a large enough amount
1) Time
2) Computational power
it must be possible.
I would appreciate any one else's opinion.
Originally posted by AlcraThe author cites 'squaring the circle or flying faster than light' as similarities, but this is plainly rubbish.These examples are both impossible to achieve as they would contradict the laws of physics. Predicting the weather acuurately does not contradict the laws of physics, it is just incredibly complex. I too would tend to believe that with enough time and computational power the results would be accurate.
An on-line article states:
(http://www.skeptics.ca/articles/yack-computers.html)
>"Computers can accurately forecast the weather"
>This myth is easily dismissed, and not just by personal experience. >In the last half-century we've learned that the accurate prediction of >the future state of a "chaotic" system like the world's weather is not >a ti ...[text shortened]... ime
2) Computational power
it must be possible.
I would appreciate any one else's opinion.
Originally posted by StarrmanExtending this idea, would it be theoretically possible to compute the state of the universe backward to a few milliseconds after its birth?
The author cites 'squaring the circle or flying faster than light' as similarities, but this is plainly rubbish.These examples are both impossible to achieve as they would contradict the laws of physics. Predicting the weather acuurately does not contradict the laws of physics, it is just incredibly complex. I too would tend to believe that with enough time and computational power the results would be accurate.
It should be, but at what cost in computational power and time (makes me think of the HHGTTG's planet Earth computer)
Originally posted by AlcraHmm, with the weather we can recalibrate methods as each point is reached, because we can see if we were right or not. Going back in time removes this ability and consequently I would presume the results would be increasingly inaccurate, especially when you factor in things like the ice age, K-T boundary mass event etc. for which we have no atmospheric data.
Extending this idea, would it be theoretically possible to compute the state of the universe backward to a few milliseconds after its birth?
It should be, but at what cost in computational power and time (makes me think of the HHGTTG's planet Earth computer)
It is impossible because even one slight disturbance in the atmosphere can lead to wildly different weather outcome in a weeks time.
There is that saying about a butterfly flapping its wings in China today can lead to the consequences of a hurricane on the otherside of the world next month.
Unless you can map the state of every particle in atmosphere, and model every possible disturbing factor (inc people), you haven't got a chance of predicting over long periods of time. Obviously noone is ever going to be able to get the state of every particle in the atmosphere, let alone predict how people are going to act, so the task is branded as impossible.
Grab a monday edition of a daily newspaper that prints 5 day weather predictions. Take a look at what it predicts for Friday. Then grab the tuesday edition and look again what it predicts for Friday. I did this for USA Today and the predictions change dramatically which clearly shows current weather prediction is not really accurate over more than 1 or 2 days.
Originally posted by PotatoErrorBut those things you talk of are not impossible. Improbable yes, but not impossible. They are within the realms of physical possibility and as such must be possible, even if it took 100,000 years and all the computers in the world, it can happen.
It is impossible because even one slight disturbance in the atmosphere can lead to wildly different weather outcome in a weeks time.
There is that saying about a butterfly flapping its wings in China today can lead to the consequences of a hurricane on the otherside of the world next month.
Unless you can map the state of every particle in atmosphe ...[text shortened]... hich clearly shows current weather prediction is not really accurate over more than 1 or 2 days.
According to Heisenberg's Uncertaincy Principle, it would be impossible as you cannot determain the momentum and position of a particle exactly at the same time. So the position of all the particles in the world would always be uncertain. Therefore if you can manage to track every single butterfly on Earth (or me even moving my fingers to type on this keyboard). You would not be able to track the momentum and position of all particles in the world with microscopic accuracy. These uncertaincies would multiply billions of times on macroscopic weather patterns.
Originally posted by lauseyTrue, so what's the solution, should we continue to pursue research into weather to get as accurate as possible within the operating parameters, or stop at the level of which human use for forecast beyond would be unnecessary?
According to Heisenberg's Uncertaincy Principle, it would be impossible as you cannot determain the momentum and position of a particle exactly at the same time. So the position of all the particles in the world would always be uncertain. ...[text shortened]... ould multiply billions of times on macroscopic weather patterns.
Originally posted by Starrman
True, so what's the solution, should we continue to persue research into weather to get as accurate as possible within the operating parameters, or stop at the level of which human use for forecast beyond would be unnecessary?
At the moment, I think the former.
I think there would come to a point where we reach a dead end though, hence leading to the latter. 🙂
Originally posted by StarrmanThe former. Theoretical science produces practical results indirectly.
True, so what's the solution, should we continue to pursue research into weather to get as accurate as possible within the operating parameters, or stop at the level of which human use for forecast beyond would be unnecessary?
Sometimes complex patterns are understood by the simplest of form. Birds can map the magnetic fields that cirlce the earth. A june bug is a fairly good predicter of the weather. Perhaps we are'nt collecting the right data.
By observing not the weather itself but how different species react before a change we can learn through observation. How does a tree know to produce bigger acorns for longer winters?
Sure we can make long complex formulae, but why not try finding the simple explanation first.
Nyxie
Originally posted by AlcraSquaring a circle is a matter of subtracting the slope of all the tangents using calculus, I believe, but that is better left to RC and Acolyte.
An on-line article states:
(http://www.skeptics.ca/articles/yack-computers.html)
>"Computers can accurately forecast the weather"
>This myth is easily dismissed, and not just by personal experience. >In the last half-century we've learned that the accurate prediction of >the future state of a "chaotic" system like the world's weather is not >a ti ...[text shortened]... ime
2) Computational power
it must be possible.
I would appreciate any one else's opinion.
The thing that makes computers able to "more and more accurately" predict weather and chaotic systems is that they ignore the quantum.
The data points are all of the macro-world. To try and implement any sort of prediction based on quantum physics would be absurd.
In a "very large heat trough" though -- such as the Sun-Earth system of measure... many data can be gathered and modeled.
It requires a lot of generalization and computers must use more and more "fuzzy logic", but we are getting better.
I would be surprised if the best super computer ever gets better than about 90% accuracy over 14 days time. That would be really amazing.
Originally posted by Alcra
An on-line article states:
(http://www.skeptics.ca/articles/yack-computers.html)
>"Computers can accurately forecast the weather"
>This myth is easily dismissed, and not just by personal experience. >In the last half-century we've learned that the accurate prediction of >the future state of a "chaotic" system like the world's weather is not >a time-consuming computation problem; it's theoretically >impossible. Dynamical systems (the technical term) are inherently >unpredictable, because they exhibit what is called "sensitive >dependence on initial conditions". So computers forecasting next >year's weather are no more believable than computers squaring the >circle, or flying faster than the speed of light.
I believe the above to be rubbish - given a large enough amount
1) Time
2) Computational power
it must be possible.
I would appreciate any one else's opinion.
To accurately predict the weather, not only would you need to take into accounts every atom, its mass and velocity on Earth, but also forces external to Earth which affect weater. It's impossible..
Originally posted by pcaspian[/b]You make a rudimentary logical mistake.
To accurately predict the weather, not only would you need to take into accounts every atom, its mass and velocity on Earth, but also forces external to Earth which affect weater. It's impossible..
A large "system" can be modeled based on probabilities.
The equivelent of what you are saying is that we would have to know exactly what every molecule in the universe could do in order to know if we should go to supper.
No. And furthermore, I recommend a good meal for all. We can predict what kind of meal awaits us without knowing all the atoms in the universe. We model that based on experience.
Originally posted by AlcraAccording to James Gleick, the author of 'Chaos' (the lay introductory
I believe the above to be rubbish - given a large enough amount
1) Time
2) Computational power
it must be possible.
I would appreciate any one else's opinion.
manual for that branch of mathematics/physics), if we had censors
on every square yard of the earth giving whatever amount of
meteorological data a scientist wanted, one would only be able to
predict the weather accurately about two weeks in advance.
As other people have stated, this is because one would have to know
the precise initial conditions of every single cubic millimeter of space
on the earth in order to know precisely how things are going to
behave. And, as pcaspian rightly observed, we would also need to
know precisely about those extra-terrestial influences (most notably
the moon and sun, which have a specific impact on our weather).
A single atom's being out of place (or mis-measured) can have
drastic outcomes on the system in time, hence the cliched 'If a
butterfly flaps it's wings in Bangladesh, it will snow in Cologne' (or
whatever and wherever the saying originally stated 😉).
I hope this commentary helps.
Nemesio