Thread: Spring coming?
View Single Post
  #9   Report Post  
Old April 7th 08, 04:25 PM posted to uk.sci.weather
Norman[_2_] Norman[_2_] is offline
external usenet poster
 
First recorded activity by Weather-Banter: Sep 2007
Posts: 1,001
Default Spring coming?

Dawlish wrote:

On Apr 7, 1:58*pm, "Norman" normanthis...@thisbitweather-
consultancy.com wrote:
Dave R. wrote:

Another run with hints of a warming, but a subtly different
set-up and the outside chance of a plume heading our way.


http://www.wetterzentrale.de/pics/Rtavn1802.png


Still not enough confidence for any kind of forecast at T240. This
mildness has come in under my radar. Some warmth would be very
welcome, though.


Paul


Wow that looks fantastic I hadn't seen that one at the
wetterzentrale site and only have to wait until the 12th! fingers
crossed and thks Dave R.


The 06Z GFS ensemble mean show 850 mb temps rising only to around
average in the south of the British Isles, implying daytime max
temps probably around 10-12c. The 850mb ensemble for London can be
found at

* * * * * * *http://85.214.49.20/wz/pics/MT8_London_ens.png

The 06z GFS operational run is on the warm side of the ensemble
mean.

Norman
--
Norman Lynagh
Chalfont St Giles, Buckinghamshire
85m a.s.l.
(remove "thisbit" twice to e-mail)- Hide quoted text -

- Show quoted text -


Hi Norman,

At 10 days it could be on the backside, underside, or the inside out
of the ensemble mean for all it means in terms of forecasting! (insert
smiley!) Spaghetti is impossible to untangle at that distance.
Patterns becoming established in a significant number of sucessive
operational runs makes for a better forecasting tool, in my
experience. No patterns: no forecast, but if a pattern stays for over
a day on the gfs at T240+, you've got over a 75% of giving a
reasonable forecast of the outcome conditions at 10 days distant. I
don't see anyone using the position relative to the ensemble mean to
get that kind of accuracy at 10 days, though I do find people still
telling me it's the thing I should do!

The ensembles provide a good tool to 7 days, but why use ensembles to
give you a forecasting tool when each ensemble has so little chance of
being correct at 10 days? Accuracy falls away like rotten skin off a
chicken at 7days. The operational run is the one that the computer
has selected as the most likely to achieve outcome and even that is
often hopelessly wrong at 10 days and each single operational run has
a very low probability of achieving outcome. It is only when you
compare successive runs and see the same pattern emerging, from
different data inputs for over a day, that anything at that distance
cane be relied upon. Even then, I've found it very hard to maintain
80% accuracy with forecasting using this method. Forecasting at 10
days, with any kind of decent accuracy is just about impossible for
any organisation, including the Met Office - who effectively don't.
All they do is give a 6-15 day idea and they produce no accuracy
statistics at 10 days.

Paul



We'll just have to agree to differ over this :-)


Norman
--
Norman Lynagh
Chalfont St Giles, Buckinghamshire
85m a.s.l.
(remove "thisbit" twice to e-mail)