In article ,
Dave Cornwell writes:
Whilst I agree with some of the comments as far as I can make out he is
only doing what lots of us do on here do for themselves. Look at the
major ensembles and come to a consensus as to what the weather will be
like based on those. This is part of what Will does, the Met office and
what all forecasters do. The major difference is that they use other
factors, especially when there is poor agreement, to arrive at a
forecast. Dawlish only does a forecast when there is good agreement.
This would inherently lead to a better accuracy. I think he is
basically saying that when the model ensembles are all over the place a
forecast is pointless based on them as it is no more than a guess. This
is a conclusion I have also come to and is borne out in practice.
Nevertheless real forecasters don't have the luxury of missing out days
that are "difficult" as there is an expectation upon them, hence the
occasional poor forecasts.

I agree with all that. If Dawlish wants to do it the way he does, then I
can't see any good reason why he shouldn't. However this does have to be
borne very much in mind when considering his percentage accuracy.

Meanwhile, perhaps by Sod's Law, the 00Z ECMWF now shows a cold easterly
setting in at 10 days, though it's very much an outlier and also has no
support from the GFS (at least not at that timescale). The odds are that
it will have disappeared when we get the 12Z run.
--
John Hall

"The covers of this book are too far apart."
Ambrose Bierce (1842-1914)