"Bernard Burton" wrote in message
...
snip

I think that the Met Office would be well served if it did a thorough
investigation into why the GFS model got this one essentially correct,
and
the Bracknell model was so far off track.
.... When I produced a Shipping Forecast (up to August last year), in
critical situations I used to cross-refer to both GFS and Arpege models.
Arpege I found useful, but suffered from the Meso to some extent that it
would 'wind-up' a developmental situation rather too much. The GFS,
whilst not perfect, would maintain consistency, and although I can't
quote you facts and figures, I know from my own experience that on at
least 3 occasions, the GFS correctly identified the 'top-end' of wind
strengths when set against our models (in two of these some *four* days
ahead of the event). I would go along with you Bernard and ask why the
GFS managed to do quite well.

snip
Perhaps if less money was spent on ensembles, which
have very limited application in the real world, and more on improving
the
observational network, we may see some improvement.

.... I can't comment with sound knowledge regarding the ensembles - they
are a fact of life and some good work is being done there; however, I
wholeheartedly concur with your other comment - I believe that the area
where this disturbance originated from and transfered through was
particularly poorly-served with observations. I've said in another
thread: the models are *superb*, even more so when you realise just how
little data they often have to go on, but we are in danger of forgetting
that observing the weather is the key to a successful forecast - if we
haven't got the data (or the data are poor), then we start off with a
limp.

Martin.