Re: Societal collapse caused by Climate Changes


Posted by Uncle Al on Jul 14, 2003 at 10:39
(68.5.243.16)

Re: Societal collapse caused by Climate Changes (Mike Kremer)

Why don't you begin with basic undergrad physics of a gyroscope like precession and nutation?

Then you can add magnetic field pole wandering and reversal. If you collapse the magnetosphere the climate walks all over the place. Ionizing radiation triggers cloud formation (Wilson cloud chamber). The Earth's manetic field is only roughly approximated by a dipole - and that only externally. It is generated by a conductive fluid slosh heir to all sorts of perturbations.

The Earth orbits the Milky Way galaxy with detectable short term periodicities. The sun cycles with a major 14-year sunspot rhythm and others. Etc.

Given long term climate data, one does a Fourier transform or power spectrum to tease out peridicities and their amplitudes. Given *good* data... Given measured and inferred peridicities, one goes looking for climate data.

All this stuff has been extensively covered in refereed literature. There is a huge and expensive entire industry of "better models." At present, nobody can model a cloud. It is a good thing clouds almost enever happen anywhere (or so the models say).

Long term prediction of climate is as meaningful as economics or psychology extrapolating a model fit to known trends. That is to say, it is meaningless. How many 14-year solar cycles would you need to predict the Maunder Minimum, 1645-1715? Both economics and psychology have a special word to describe when a perfect mathematical model is cruelly japed by physical reality: Heteroskedasticity. How often do these disciplines furiously backpedal?

Google
Heteroskedasticity 25,500 hits
Heteroscedasticity 20,500 hits

Making excuses for empirical failure, prediction vs. observation, is pretty much the sum and substance of both.

--
Uncle Al
http://www.mazepath.com/uncleal/eotvos.htm
(Do something naughty to physics)

Follow Ups:


Post a Followup

Name:
E-Mail:

Subject:
Comments:

[ Forum ] [ New Message ]