On 26/02/10 20:38, Will Hand wrote:
Alex, the Great Storm of 1987 was over 22 years ago and NWP has moved a
long way. It too IIRC was a Shapiro-Keyser system with a seclusion (warm
core), but Shapiro-Keyser had not written their paper in 1987 so the
understanding was less. What that storm did was to focus forecaster's
minds on making better use of satellite imagery to spot conceptual
features. The modelling of this storm looks to be superb capturing the
west-east orientation and warm air seclusion process typical of this
kind of system. Because the model is capturing the dynamics it should
get the wind forecast right too, the only error is likely to be track
and possibly depth. Alex, models have moved a vast amount since the days
of 1987, those models could be run on a modern day PC, nowadays a fast
supercomputer is required.

That would have been in the days of the Cyber 205, I think. I see that
supercomputer had a theoretical peak performance of 200 MFLOPS. The
model installed in 1985 at Florida State had a core memory of 32 MiB and
7.2 GiB of on-line disk space. I've no idea how many MFLOPS PCs would be
rated at these days but memory and disk space is certainly larger.

I think the computer that ran the three-parameter model back when I was
new to the Office, would have been about the same specs as a Sinclair
Spectrum.

--
Graham P Davis, Bracknell, Berks., UK. E-mail: newsman not newsboy
"I wear the cheese. It does not wear me."