We (East Cambs) currently have a yellow warning for rain from the Met Office
(with a forecast last updated at 1400 (BST?) though I don't know about the
warning) but 2 columns to the right on our MO 3-hourly forecast page, the
probability of rain is 5% for the rest of the day. Is it that there's a
tiny chance of an inundation of truly biblical proportions? Or is the
warning system totally out of phase with the forecast updates?

I'm sure that there's a good and logical reason - and not wishing to indulge
in any UKMO-bashing - but it's beyond my waning powers of deduction.

JGD

On Monday, July 22, 2013 5:16:29 PM UTC+1, General wrote:

Is it that there's a tiny chance of an inundation of truly biblical proportions?


=================

More or less.

It's for a low probability high impact event. A typical warning for this sort of situation with scattered or isolated but potentially very heavy showers / thunderstorms. A high probability moderate impact event would also generate a yellow warning.

Look for the matrix on either of these pages:

http://www.metoffice.gov.uk/guide/weather/warnings

http://www.metoffice.gov.uk/weather/.../warnings.html

Stephen.

"Stephen Davenport" wrote in message
...
It's for a low probability high impact event.
===================================

OK, thanks. But in that case, I might be tempted to quibble with the way
that the probabilities are being interpreted. I tend to think of 95% as
beyond all reasonable doubt (to use a legal analogy). Therefore, 'beyond all
reasonable doubt' we're simply not going to see any significant rain before
midnight (p0.05). So I'd argue that no warning is justified. A low
probability event might be in the 10-15% band, which is not being forecast.

But that's just how I think.

JGD

On Monday, July 22, 2013 6:17:11 PM UTC+1, General wrote:
"Stephen Davenport" wrote in message

...

It's for a low probability high impact event.

===================================



OK, thanks. But in that case, I might be tempted to quibble with the way

that the probabilities are being interpreted. I tend to think of 95% as

beyond all reasonable doubt (to use a legal analogy). Therefore, 'beyond all

reasonable doubt' we're simply not going to see any significant rain before

midnight (p0.05). So I'd argue that no warning is justified. A low

probability event might be in the 10-15% band, which is not being forecast.

"Stephen Davenport" wrote in message
...

But they are not saying that there will be significant rain with 95%
probability. They are saying that there is low probability (5% if you like)
of a heavy shower or thunderstorm at any location but that if that
probability should manifest the impact from heavy rain could be significant.
==========================

That's not how I read it. The rain column is headed 'Precip. (%)' which I
take to mean the probability of anything more than a few drops of rain and
is a figure that stands on its own irrespective of whether any warning is in
force. Yesterday evening this column was saying 5%, which seems plain
enough to me - no rain was expected. It's not saying that there's 5% chance
of the warned circumstance happening.

(This is just a quibble about a detail of course - the storms did
eventually move through, though nothing too noteworthy for the time of
year - but it seems to me just the sort of blatant logical inconsistency
that brings the warning system into disrepute.)

I suspect that your other point, ie that the warning and the forecast are
generated by two different processes is the explanation. The progress of the
rain and storms was, in the event, perhaps somewhat slower than looked
possible when the warnings were generated?

JGD

General wrote:

"Stephen Davenport" wrote in message
...

But they are not saying that there will be significant rain with 95%
probability. They are saying that there is low probability (5% if you like)
of a heavy shower or thunderstorm at any location but that if that
probability should manifest the impact from heavy rain could be significant.
==========================

That's not how I read it. The rain column is headed 'Precip. (%)' which I
take to mean the probability of anything more than a few drops of rain and is
a figure that stands on its own irrespective of whether any warning is in
force. Yesterday evening this column was saying 5%, which seems plain enough
to me - no rain was expected. It's not saying that there's 5% chance of the
warned circumstance happening.

(This is just a quibble about a detail of course - the storms did eventually
move through, though nothing too noteworthy for the time of year - but it
seems to me just the sort of blatant logical inconsistency that brings the
warning system into disrepute.)

I suspect that your other point, ie that the warning and the forecast are
generated by two different processes is the explanation. The progress of the
rain and storms was, in the event, perhaps somewhat slower than looked
possible when the warnings were generated?

JGD

The biggest gripe I have is that the warning system is usually very static. It
gives a broad-brush catch-all warning that is left "current" without any notice
taken about what is actually happening. This morning is a perfect example where
there should be hourly, or even half-hourly, warning updates spelling out which
areas look like being hit in the next hour or two. It's not rocket science but
I suppose it does need manpower which costs money and so............

--
Norman Lynagh
Tideswell, Derbyshire
303m a.s.l.