Sandy Harris wrote:
> Would sheer numbers help much? There's a lot of connectivity
> options these days. If someone came up with a really cheap
> basic instrument pack and deployed thousands of them (maybe
> one on every farm?) with communications to some central
> computer, would prediction get any better?
>
As someone else has pointed out, numbers efforts are well underway.
However, I think the increase in predictive accuracy will be
incremental, at best.
Computer analysis of weather data in an effort to improve prediction
accuracy is famously one of the roots of Chaos Theory. You know the
story. A butterfly turns right or left where you live and a blizzard
does or doesn't kill people and cause havoc in the US Midwest.
The old ideas, based in Newtonian physics (and philosophical ideas of
determinism) believed that with enough computer power and enough input
data, any physical system could be accurately predicted in detail. Chaos
Theory showed that in non-linear systems that is simply not true. And,
Catch 22, all real world systems, at least at the level of scale where
we exist and interact with the world, are non-linear. Boundary
conditions can be determined, but the detailed state of the system at
any moment cannot be predicted.
And the unpredictability isn't the kind we would like, with our
deterministic thinking, a little fuzziness around a central value.
Complex, non-linear systems are subject to major breaks in behavior at
unpredictable times. And weather is a highly complex system or systems
with innumerable non-linear components.
Complexity Theory is starting to make important headway in explaining
systems behaviors in the area between orderly and chaotic systems, but I
doubt it will help with large scale, long term weather predictions in
the near future, if ever. Very interesting stuff, though.
Moose
==============================================
List usage info: http://www.zuikoholic.com
List nannies: olympusadmin@xxxxxxxxxx
==============================================
|