Martin Rowley wrote:
... I've been meaning to post this for some time, but given the
situation over the next few days, now is as good as time as any.

Forecasting centres nowadays use a vast array of ensemble products
covering the /short-range/ (say, up to 60 hours) to assess risk of
weather having adverse effect. We don't unfortunately have access to
these, but courtesy of the US Air Force, we can at least get a flavour
of what's going on.

At this link ....
http://ows.public.sembach.af.mil/ind...ction=afwameps

The European hub for the USAF Air Force Weather Agency (AFWA) makes
available ensemble products for 'winter PPN' types. It is known as the
MEPS (Mesoscale Ensemble Prediction Suite), and looking at the
background notes, there are many other parameters available - though not
to us. [ This model is now used _extensively_ throughout the USAF for
critical operational decisions: a search will throw up much more.]

I'm NOT an expert, but to save you looking it up, the model is based on:-
... 10 members of the WRF-ARW [WRF = Weather Research & Forecast model,
based on NCAR modelling]; the ARW bit simply means ... Advanced Research
version of the WRF, which code is effectively (IIUC) open-source.

... Initial (boundary) conditions are driven by (UK) Global, (US/NCEP)
GFS, (Canadian) GEM and (US/USN) NOGAPS.

... PLUS, there is input from two /ensemble/ outputs: SREF and GFES;
SREF is the NOAA/National Weather Service Storm Prediction Center, Short
Range Ensemble Forecast suite ... hence SREF.

So, initial conditions are based on a goodly range of deterministic
models and two ensemble models, which of course themselves are based on
other models: I believe the SREF is 'driven' by no less than /four/
high-resolution models (though not many members are used) when used in
North American practice) - it may be fewer models out-with that domain
of course, as some of the input models are NAM specific.

So the MEPS is not so much a "poor man's ensemble", but a "moderately
well-off person's ensemble"!

They are only output *once per day*, with a nominal initialisation time
of 06Z, and if you access the page early in the day, you'll find the
fields are missing - I usually find that by 09Z they are available.

The display implies that the resolution is 20 km which is considerably
(in modern-day practice) courser than the models used by the Met O ...
1.5 & 4 km currently; but it's all we've got!

This is a 'shorter' link to the 06Z on Monday morning - interesting as
we say in these parts :-)

http://tinyurl.com/al98w7k

When I looked last year, the products are replaced in spring/summer
etc., by such as TS risk, high gusts etc.

Martin.



--------------------------------
Thank you for that Martin,
Dave