Thread: GFS Reliability
View Single Post
  #4   Report Post  
Old July 28th 10, 11:17 PM posted to uk.sci.weather
Dawlish Dawlish is offline
external usenet poster
 
First recorded activity by Weather-Banter: Mar 2008
Posts: 10,601
Default GFS Reliability

On Jul 28, 9:19*pm, David Gartrell wrote:
I was wondering how far ahead the GFS pressure, temperature & rainfall
charts can be relied upon. *The reason for the question is that we're
off on our Holidays to Switzerland in August and the Gfs charts seem to
be suggesting a much improved outlook which quite frankly i'm
desperately hoping is correct. *I had read somewhere that the charts up
to +240 can be considered fairly accurate but become rather fanciful
after that.

If someone could advise me please that would be great.

Many thanks


That very much depends David. I maintain that there are occasions when
it is possible to forecast using the gfs at 10 days. I've proved that
with a 78% success rate, over 89 forecasts over 5 years, forecasting
at those particular times. I've only issued those forecasts in
circumstances, where consistency in the gfs output agrees with ECM
(and other output). Determining those times is not easy, I can assure
you.

The rest of the times, inconsistency and lack of agreement makes
forecasting impossible - and that's the majority of the time, IMO.

I couldn't comment on the Alps' weather over the next 10 days, sorry;
I haven't been watching the models for that area, only the UK.
Whatever it presently shows may, or may not, come to outcome, but
you'd only know to trust it if the output had been consistent and
agrees with other models over the last 24 hours - even then, you are
only loolking at being correct 3 times out of 4.

Forecasting at 10 days is a mug's game. It's impossible with regular,
daily forecasts and even the MetO hardly bothers. I always think that
their 6-15 days forecast (from which you may be able to divine the
forecast at 10 days, on occasions) is written in the tea break, in
between doing crossord clues. You'd be better tuning in closer to your
hols. 5/6 days is pretty accurate, as evidence by NASA's model success
monitoring . 7 days is still reasonable, but after that, outcome
success plummets.

Hope you enjoy your holiday. I remember a visit to the Bernese
Oberland in August and skiing on the Jungfrau glacier. Magnificent.
Snow-blind in the tunnels walking to the train, but still magnificent!
Have lunch in the Schilthorn revolving restaurant. Again, truly
magnificent! The view of a lifetime! You slowly turn to see the
Jungfrau, Monch and the Eiger pass, whilst eating lunch (which was
amazingly crap! - but who cares! Most of it ended up dribbled down
your chin as you concentrated on the panorama out of the window).

I'm sorry