"TudorHgh" wrote in message
...
Add to that the possibility of human intervention in some model
output, and
you increase the uncertainty further.

You didn't really mean that, did you, Bernard? I always thought
the
purpose of human intervention was rather like a parent trying to guide
their
brilliant but erratic child, i.e. to stop it doing something really
stupid.



.... there is actually a long-running debate about the advisability of
'intervention' in NWP simulation of the atmosphere. The modellers almost
certainly (though I may be doing individuals a dis-service & for
Tropical features, 'support'/positive intervention is definitely an
advantage) would rather that the models were allowed to run without
outside help; and in any case, the purpose of intervention is to aid the
*analysis*, rather than nudge the model into one solution rather than
another - though those that perform this 'deft art' are all too aware
that too much tinkering (particularly with humidity), can have rather
dramatic effects! I'm not sure how many of the main models we all look
at now actually have such intervention; the US models and the EC don't;
not sure about the other European models though.

The following is quite an interesting read:...

http://www.metoffice.com/research/nw...00/Storms.html

Martin.