Steve Simon wrote :
> There is an inherent distrust of SIMULATION models. Look at all the
> controversy about simulation models for global warming.
Quelle surprise !
Why shouldn't there be distrust of simulation modelling ? Almost all
the simulation models I have seen are published with no external
validation whatsoever. Often the data on which they are based is
limited, and questionable assumptions are made.
Global warming models are an excellent example, but one in which I
will admit many modellers have done remarkably well :-).. Most of the
models described over the last few years have been pretty poor. Yes
the researchers have been working with both hands tied behind their
backs because there is no way to externally validate the model - no
alternative earths handy; historical data has been limited; and they
have not envisaged the complexities of the system fully. Unlike many
simulation models the results (of the last factor) have been that
global warming models have almost universally under-predicted the
rate of change ! The modellers have done "remarkably well" in that
they have given us an inkling of likely problems, something many of
us have found worthy of attention.
So, of course we should examine simulation models critically,
especially in light of the methodology applied, before they are
published. They have their uses, but their limitations cannot be
ignored.
Mike.
--
The University of Edinburgh is a charitable body, registered in
Scotland, with registration number SC005336.
|