| || |
April 25, 2012 - Harry Eagar
We are often told that the science behind global warming is settled and that computer simulations demonstate the bad stuff that will shortly befall us.
As soon as anyone cites "computer simulations," I start doubting. For many reasons. One, and this is merely stating the obvious, though I bet you have never heard it before, is that to simulate something on a computer, you have to thoroughly understand it.
Some things are probably beyond thorough understanding, and climate and economics are two of them. Both have proven highly resistant to successful computer simulation.
An interesting example of betting the farm on a computer simulation and failing spectacularly is the Defense Department's hypersonic airplane, a test bed of which crashed in the ocean not long ago. Flying magazine has a report on DARPA's analysis of what went wrong:
"According to DARPA’s report, ground testing and thermal modeling that dictated the vehicle’s skin structure needs was completed based on extrapolations from known flight characteristics, and ultimately proved unable to 'successfully predict the harsh realities of Mach 20 atmospheric flight.'
'Those harsh realities took their toll on HTV-2’s aeroshell as the vehicle traveled around the 13,000 mph mark, creating gaps in the vehicle’s surface that triggered impulsive shock waves around the 13,000 mph mark, approximately 9 minutes after the vehicle successfully separated from a rocket in the upper atmosphere.' "
The really harsh reality is that computer modeling responds to GIGO, and worse than that, some systems may be inherently impossible to emulate.
I was talking with son about that this morning, and that will be the subject of the next post.
No comments posted for this article.
Post a Comment