Subject: | |
From: | |
Reply To: | |
Date: | Fri, 15 Jan 2016 11:21:01 -0500 |
Content-Type: | text/plain |
Parts/Attachments: |
|
|
On Sat, 09 Jan 2016 19:52:46 -0500, Peter Loring Borst
<[log in to unmask]> wrote:
>> In the results section the authors talk about the model predictions.
>> What are the thoughts about the quality of the developed model?
>
> * The answer to this is: models may be state of the art but they are
> notoriously unreliable. Garbage in=garbage out. They are barely adequate
> to predict the weather 3 days in advance.
I did want to follow up on this because that's not entirely accurate and
depends on which models are we talking about. Computer modeling of real
events is as old as computer science itself (the first computers were used
to calculate bullet trajectories for artillery - a model if there ever was
one!), but some things can be modeled quite perfectly. (Full disclosure: I
work for a company that sells such simulators). Specifically, modeling
physical world, especially mechanical behavior is able to achieve 1-1
match with real world experiments (car & airframe crash tests, for
example). A lot of car commercials showing disintegrating car parts
originate from those simulation models - no one crashes a new car for a TV
commercial, much less an airplane. It was in 1995 that Dassault Systemes
and Boeing teamed up to develop the first digital airplane mockup and the
first 777 off the assembly line flew successfully without ever building a
physical mockup for flight tests. The rest, as the say, was aviation
history.
Weather models, as much as we might complain about them have also
progressed by leaps and bounds and generally are dependent primarily on
the computational power available and the size of the sample data. After
all, we have hundreds years of data in certain cases to use as base for
those models. That's why the current European weather prediction model is
superior to the US because European Union has given itself a huge facility
where the weathermen can compute more cycles in their models. US is
playing catchup and weather computers are less powerful, but with our
disfunctional Congress I wasn't able to find out if the money was
allocated in this year's budget as promised.
On the other hand, I agree, modeling life events for which we do not have
bullet-proof theories is hard and such models need to be approached with a
very skeptical eye.
Przemek
***********************************************
The BEE-L mailing list is powered by L-Soft's renowned
LISTSERV(R) list management software. For more information, go to:
http://www.lsoft.com/LISTSERV-powered.html
|
|
|