I see the flash warnings for most of central England have now shrunk to the
blob on the radar around and just to the North of London. Certainly appears
to have been where the height of the activity was today. My daughter
reported violent storms with torrential rain and pea sized hail from around
the Finsbury Park area.
Dave

On 27 June, 22:11, "Dave Cornwell"
wrote:
I see the flash warnings for most of central England have now shrunk to the
blob on the radar around and just to the North of London. Certainly appears
to have been where the height of the activity was today. My daughter
reported violent storms with torrential rain and pea sized hail from around
the Finsbury Park area.

The warnings today were interesting - the absence of a warning for
Kent was soon rectified when the sea breeze convergence-induced storms
over Kent kicked in. Ditto the gap over London as the storms rumbled
around there this afternoon. Hopefully examples of this show a) how
horribly difficult convective warnings are to put out and b) the MO do
"look out of the window" during times of potential severe weather.

Great afternoon of weather though - from dozing on Blackheath watching
clouds bubble to trying to film some very odd cloud behaviour outside
my flat !

Richard

In message
,
Richard Dixon writes
On 27 June, 22:11, "Dave Cornwell"
wrote:
I see the flash warnings for most of central England have now shrunk to the
blob on the radar around and just to the North of London. Certainly appears
to have been where the height of the activity was today. My daughter
reported violent storms with torrential rain and pea sized hail from around
the Finsbury Park area.

The warnings today were interesting - the absence of a warning for
Kent was soon rectified when the sea breeze convergence-induced storms
over Kent kicked in. Ditto the gap over London as the storms rumbled
around there this afternoon. Hopefully examples of this show a) how
horribly difficult convective warnings are to put out and b) the MO do
"look out of the window" during times of potential severe weather.

Great afternoon of weather though - from dozing on Blackheath watching
clouds bubble to trying to film some very odd cloud behaviour outside
my flat !

Richard

Looking at the Meteosat IR loop there did appear to be an almost
enclosed low around London looking at the circulation. Possible a heat
low?

James
--
James Brown

There was an upper low over southern UK at 1200z, and the location of the
storms in the south were probably associated with this.
Watch out for the line wrap
http://weather.uwyo.edu/cgi-bin/uama...&TYPE=obs&TYPE
=an&LEVEL=400&TIME=2009062712

--
Bernard Burton
Wokingham, Berkshire, UK.

Satellite images at:
www.woksat.info/wwp.html

"James Brown" wrote in message
...
In message
,
Richard Dixon writes
On 27 June, 22:11, "Dave Cornwell"
wrote:
I see the flash warnings for most of central England have now shrunk to
the
blob on the radar around and just to the North of London. Certainly
appears
to have been where the height of the activity was today. My daughter
reported violent storms with torrential rain and pea sized hail from
around
the Finsbury Park area.

The warnings today were interesting - the absence of a warning for
Kent was soon rectified when the sea breeze convergence-induced storms
over Kent kicked in. Ditto the gap over London as the storms rumbled
around there this afternoon. Hopefully examples of this show a) how
horribly difficult convective warnings are to put out and b) the MO do
"look out of the window" during times of potential severe weather.

Great afternoon of weather though - from dozing on Blackheath watching
clouds bubble to trying to film some very odd cloud behaviour outside
my flat !

Richard

Looking at the Meteosat IR loop there did appear to be an almost
enclosed low around London looking at the circulation. Possible a heat
low?

James
--
James Brown

On Jun 27, 10:28*pm, Richard Dixon wrote:
On 27 June, 22:11, "Dave Cornwell"

wrote:
I see the flash warnings for most of central England have now shrunk to the
blob on the radar around and just to the North of London. Certainly appears
to have been where the height of the activity was today. My daughter
reported violent storms with torrential rain and pea sized hail from around
the Finsbury Park area.

Hopefully examples of this show
a) how horribly difficult convective warnings are to put out .

They are not difficult to put out; that's the easy bit, all that has
to happen is the trigger of a threshold i.e. the forecasters have to
believe, through their experience, that something will have x% chance
of occurring. The difficult bit is actually getting the warnings
correct by the outcome of an event happening in a particular area -
that's the difficult bit, especially in convective situations.

I would think the MetO have more chance of getting their heatwave
warnings correct, as it is far easier to forecast temperatures above a
threshold, that it is to forecast rainfall above a threshold. I would
have far more belief in a temperature forecast than I would have in a
convective forecast for my area (and my area is NOT the whole of
Devon!!). The MetO is not yet capable of forecasting convective events
accurately, at any distance greater than a very short time (for which
you could better use a freely available weather radar and watch the
sky as to when to move your BBQ indoors, or abandon the beach), either
in terms of 1. location, 2. intensity of rainfall, or 3. duration of
rainfall. Again, that is not a criticism, the tools are simply not
there to do that with good accuracy, it's just a fact.

MetO forecasters are some of the best in the world, but accurate
forecsating of those 3 perameters in these events would be beyond
almost anyone (forecasting of convective events in the mid-west of
America may provide the exception to "anyone". Maybe the threat to
life from tornados forces more resources and a higher level of skill -
just a thought). Yesterday's too wide severe weather warning, that
you've referred to, shows that yet again. There have been a number of
other examples already this year.

I must say, I do think the post mortem explanations are a good idea,
though.

On 28 June, 08:43, Dawlish wrote:

They are not difficult to put out; that's the easy bit, all that has
to happen is the trigger of a threshold i.e. the forecasters have to
believe, through their experience, that something will have x% chance
of occurring. The difficult bit is actually getting the warnings
correct by the outcome of an event happening in a particular area -
that's the difficult bit, especially in convective situations.

When convection on days such as this can pop up quite randomly, I'd
love to see your methodology for assigning a probability to the chance
of convective precipitation, I really would.

I would think the MetO have more chance of getting their heatwave
warnings correct, as it is far easier to forecast temperatures above a
threshold, that it is to forecast rainfall above a threshold. I would
have far more belief in a temperature forecast than I would have in a
convective forecast for my area (and my area is NOT the whole of
Devon!!).

You expect them to forecast convective precipitation for your town?
Goodness me - you're a hard man to please. Yes, they'll have more
chance of forecasting temperatures for this week correctly than
convective events.

The MetO is not yet capable of forecasting convective events
accurately, at any distance greater than a very short time (for which
you could better use a freely available weather radar and watch the
sky as to when to move your BBQ indoors, or abandon the beach), either
in terms of 1. location, 2. intensity of rainfall, or 3. duration of
rainfall. Again, that is not a criticism, the tools are simply not
there to do that with good accuracy, it's just a fact.

The Met Office is capable of forecasting convective events accurately,
just not all the time. Have you ever looked at any diagnostics for
convection like CAPE? The key letter here is the P. Potential. The
whole of England and Wales was no doubt under a region of high CAPE
yesterday. In some areas, the surface air parcels were able to break
through the usual summertime capping lid on instability and form
thunderstorms. Did you not see the other thunderstorms on the radar in
the Reading, Birmingham, Manchester etc. areas? I would have though
the people in these regions would have been grateful for the severe
weather warning. And yet you see it as a chance to point out the
failings, which probably says more about you than the severe weather
warnings themselves, if I may be rude.

MetO forecasters are some of the best in the world, but accurate
forecsating of those 3 perameters in these events would be beyond
almost anyone (forecasting of convective events in the mid-west of
America may provide the exception to "anyone". Maybe the threat to
life from tornados forces more resources and a higher level of skill -
just a thought). Yesterday's too wide severe weather warning, that
you've referred to, shows that yet again. There have been a number of
other examples already this year.

Forecasting the location, intensity and duration of individual
convection cells would probably need cloud-resolving (1km
resolution?) models. The best mesoscale high-resolution models will
struggle to get them. Until then, you'll have to accept that blanket
warnings will be the way of things. Expecting these perameters (sic)
to be bang on - well, you're in cloud cuckoo land my friend.

I must say, I do think the post mortem explanations are a good idea,
though.

Or, in the case of the heavy rain mid-June, an explanation for their
forecast and the reasoning for their (correct) severe weather warning.
Funny how I never see you on here praising the forecasters for a
correct forecast. Always up for a gripe though, aren't we.

Richard

On Jun 28, 11:18*am, Richard Dixon wrote:
On 28 June, 08:43, Dawlish wrote:

They are not difficult to put out; that's the easy bit, all that has
to happen is the trigger of a threshold i.e. the forecasters have to
believe, through their experience, that something will have x% chance
of occurring. The difficult bit is actually getting the warnings
correct by the outcome of an event happening in a particular area -
that's the difficult bit, especially in convective situations.

When convection on days such as this can pop up quite randomly, I'd
love to see your methodology for assigning a probability to the chance
of convective precipitation, I really would.

I would think the MetO have more chance of getting their heatwave
warnings correct, as it is far easier to forecast temperatures above a
threshold, that it is to forecast rainfall above a threshold. I would
have far more belief in a temperature forecast than I would have in a
convective forecast for my area (and my area is NOT the whole of
Devon!!).

You expect them to forecast convective precipitation for your town?
Goodness me - you're a hard man to please. Yes, they'll have more
chance of forecasting temperatures for this week correctly than
convective events.

The MetO is not yet capable of forecasting convective events
accurately, at any distance greater than a very short time (for which
you could better use a freely available weather radar and watch the
sky as to when to move your BBQ indoors, or abandon the beach), either
in terms of 1. location, 2. intensity of rainfall, or 3. duration of
rainfall. Again, that is not a criticism, the tools are simply not
there to do that with good accuracy, it's just a fact.

The Met Office is capable of forecasting convective events accurately,
just not all the time. Have you ever looked at any diagnostics for
convection like CAPE? The key letter here is the P. Potential. The
whole of England and Wales was no doubt under a region of high CAPE
yesterday. In some areas, the surface air parcels were able to break
through the usual summertime capping lid on instability and form
thunderstorms. Did you not see the other thunderstorms on the radar in
the Reading, Birmingham, Manchester etc. areas? I would have though
the people in these regions would have been grateful for the severe
weather warning. And yet you see it as a chance to point out the
failings, which probably says more about you than the severe weather
warnings themselves, if I may be rude.

MetO forecasters are some of the best in the world, but accurate
forecsating of those 3 perameters in these events would be beyond
almost anyone (forecasting of convective events in the mid-west of
America may provide the exception to "anyone". Maybe the threat to
life from tornados forces more resources and a higher level of skill -
just a thought). Yesterday's too wide severe weather warning, that
you've referred to, shows that yet again. There have been a number of
other examples already this year.

Forecasting the location, intensity and duration of individual
convection cells would probably need cloud-resolving (1km
resolution?) models. The best mesoscale high-resolution models will
struggle to get them. Until then, you'll have to accept that blanket
warnings will be the way of things. Expecting these perameters (sic)
to be bang on - well, you're in cloud cuckoo land my friend.

I must say, I do think the post mortem explanations are a good idea,
though.

Or, in the case of the heavy rain mid-June, an explanation for their
forecast and the reasoning for their (correct) severe weather warning.
Funny how I never see you on here praising the forecasters for a
correct forecast. Always up for a gripe though, aren't we.

Richard

Outcomes Richard. Outcomes. Funny how you never read the posts where I
defend the MetO and prasie the forecasters - as I have in this
"gripe".

Funny also how you don't believe in outcomes as a method of judging
the accuracy of forecasts and would rather attack the questioner, than
agree with the comments about the lack of general accuracy in many of
these "severe" weather warnings. And I mean many; pointing to one
correct one - which I actually praised, but you obviously failed to
notice - does not address the many that cause people to think that the
MetO are crying wolf and covering their backs whenever a sniff of
severe weather is in the offing. Yesterday's wide area "be prepared"
warning was another example of many people being warned.......then
nothing happens for them, but only happens for a few. When a real
warning is needed how many of the vulnerable groups, that really need
to heed the warnings, will really "take action"?

FWIW and I'll repeat for your benefit, I don't think it is the fault
of the MetO at all. I think the difficulties lie in exaggerated public
expectation, combined with forecasting at the limits of what is
possible, in terms of location, intensity and duration of the rainfall
(or snowfall, in the very poorly forecast and very localised devon
event in February; or the very local and again poorly forecast St Just
event; or the very local and again poorly forecast Ottery St. Mary
event; or the less local, but very poorly forecast and expensive for
many local communities, non-event over the Spring Bank holiday - and
they are just recent SW events this year which were poorly forecast at
12-36 hours notice).

Now Richard; I understand that conundrum. Falling back on defensive
posts like the last one of yours really doesn't help. An open
assessment of the state of forecasting of convective events would be a
better standpoint.

Like I said, the issuing of the warnings is easy. However, the
achievement of good outcomes at 12-36 hoursand not just for the areas
on which the rain fell - not difficult to get a result on that one
when the majority of England was highlighted as a risk zone, whereas
heavy rain fell in a very small part of that area yesterday, judging
by reports - is the difficult part. I have every sympathy with the
forecasters this week, as those kind of situations are likely to be
repeated.

On 28 June, 14:03, Dawlish wrote:

Outcomes Richard. Outcomes. Funny how you never read the posts where I
defend the MetO and prasie the forecasters

No - I didn't see them.

Funny also how you don't believe in outcomes as a method of judging
the accuracy of forecasts and would rather attack the questioner, than
agree with the comments about the lack of general accuracy in many of
these "severe" weather warnings.

Because I realise your view of "outcomes" is far, far different from
mine in weather forecasting. Yesterday there were scattered severe
thunderstorms. Apart from the Kent convergence line cock-up (that I
pointed to in another thread) that led to an amended warning, they did
a pretty good job. In your "outcome" book, they didn't as it was just
a wishy-washy broad warning. We have two entirely different points of
view.

Yesterday's wide area "be prepared"
warning was another example of many people being warned.......then
nothing happens for them, but only happens for a few.

Which is down to the basic difficult of forecasting the situation
which I get the impression you fail to grasp. Convective storms were
forecast to be isolated events yesterday, the MO didn't have a full
grasp on where they were to form so what do you do? Issue nothing? No,
you issue warnings for where thunderstorms could happen
(unfortunately, the Kent example evaded their model).

Maybe you'd like to note today's severe weather warning which states
"Slow moving thunderstorms will develop this afternoon giving
torrential downpours and hail, although some places will miss them and
stay dry. However, in areas affected, 20-40mm of rain is likely in 3
hours or less". That's pretty clear in saying we've got some showers
around, we're not quite sure where exactly they're going to fall, but
you could catch one, and it could be pretty nasty. It's called
"dumbing down" if you like and trying to get the uncertainty across to
the general public.

FWIW and I'll repeat for your benefit,
I don't think it is the fault
of the MetO at all.
I think the difficulties lie in exaggerated public
expectation, combined with forecasting at the limits of what is
possible, in terms of location, intensity and duration of the rainfall
(or snowfall, in the very poorly forecast and very localised devon
event in February; or the very local and again poorly forecast St Just
event; or the very local and again poorly forecast Ottery St. Mary
event; or the less local, but very poorly forecast and expensive for
many local communities, non-event over the Spring Bank holiday - and
they are just recent SW events this year which were poorly forecast at
12-36 hours notice).

Take a breath, man! Semi-colon overload! I think "exaggerated public
expectation" is probably where you're right here, and are probably the
champion of the movement.

Now Richard; I understand that conundrum. Falling back on defensive
posts like the last one of yours really doesn't help. An open
assessment of the state of forecasting of convective events would be a
better standpoint.

In your humble opinion, of course. Do you really think you local
butcher wants to know about the state of forecasting convective
events? I would say that probably at most 1% of the population is
interested in this. Given your keenness, maybe you could start a
career in research - having come from that area, it's quite rewarding
- and helps you to understand the demands, time and effort of other
involved in forecasting research. Just ask Will Hand..

However, the
achievement of good outcomes at 12-36 hoursand not just for the areas
on which the rain fell - not difficult to get a result on that one
when the majority of England was highlighted as a risk zone, whereas
heavy rain fell in a very small part of that area yesterday, judging
by reports - is the difficult part.

This is where I give up with you - you're expecting them to forecast
accurately everywhere that's going to get wet. As before, I refer you
to the dictionary definition of warning: "An intimation, threat, or
sign of impending danger". There was a *threat* of storms throughout
England and Wales yesterday that the forecasting models were not able
to suggest locations for as it can be such a random process. Given
this threat, but no guidance of where they could be, would you want
them to stay quiet?

Richard

On Jun 28, 2:46*pm, Richard Dixon wrote:
On 28 June, 14:03, Dawlish wrote:

Outcomes Richard. Outcomes. Funny how you never read the posts where I
defend the MetO and prasie the forecasters

No - I didn't see them.

Funny also how you don't believe in outcomes as a method of judging
the accuracy of forecasts and would rather attack the questioner, than
agree with the comments about the lack of general accuracy in many of
these "severe" weather warnings.

Because I realise your view of "outcomes" is far, far different from
mine in weather forecasting. Yesterday there were scattered severe
thunderstorms. Apart from the Kent convergence line cock-up (that I
pointed to in another thread) that led to an amended warning, they did
a pretty good job. In your "outcome" book, they didn't as it was just
a wishy-washy broad warning. We have two entirely different points of
view.

Yesterday's wide area "be prepared"
warning was another example of many people being warned.......then
nothing happens for them, but only happens for a few.

Which is down to the basic difficult of forecasting the situation
which I get the impression you fail to grasp. Convective storms were
forecast to be isolated events yesterday, the MO didn't have a full
grasp on where they were to form so what do you do? Issue nothing? No,
you issue warnings for where thunderstorms could happen
(unfortunately, the Kent example evaded their model).

Maybe you'd like to note today's severe weather warning which states
"Slow moving thunderstorms will develop this afternoon giving
torrential downpours and hail, although some places will miss them and
stay dry. However, in areas affected, 20-40mm of rain is likely in 3
hours or less". That's pretty clear in saying we've got some showers
around, we're not quite sure where exactly they're going to fall, but
you could catch one, and it could be pretty nasty. It's called
"dumbing down" if you like and trying to get the uncertainty across to
the general public.

FWIW and I'll repeat for your benefit,
I don't think it is the fault
of the MetO at all.
I think the difficulties lie in exaggerated public
expectation, combined with forecasting at the limits of what is
possible, in terms of location, intensity and duration of the rainfall
(or snowfall, in the very poorly forecast and very localised devon
event in February; or the very local and again poorly forecast St Just
event; or the very local and again poorly forecast Ottery St. Mary
event; or the less local, but very poorly forecast and expensive for
many local communities, non-event over the Spring Bank holiday - and
they are just recent SW events this year which were poorly forecast at
12-36 hours notice).

Take a breath, man! Semi-colon overload! I think "exaggerated public
expectation" is probably where you're right here, and are probably the
champion of the movement.

Now Richard; I understand that conundrum. Falling back on defensive
posts like the last one of yours really doesn't help. An open
assessment of the state of forecasting of convective events would be a
better standpoint.

In your humble opinion, of course. Do you really think you local
butcher wants to know about the state of forecasting convective
events? I would say that probably at most 1% of the population is
interested in this. Given your keenness, maybe you could start a
career in research - having come from that area, it's quite rewarding
- and helps you to understand the demands, time and effort of other
involved in forecasting research. Just ask Will Hand..

However, the
achievement of good outcomes at 12-36 hoursand not just for the areas
on which the rain fell - not difficult to get a result on that one
when the majority of England was highlighted as a risk zone, whereas
heavy rain fell in a very small part of that area yesterday, judging
by reports - is the difficult part.

This is where I give up with you - you're expecting them to forecast
accurately everywhere that's going to get wet. As before, I refer you
to the dictionary definition of warning: "An intimation, threat, or
sign of impending danger". There was a *threat* of storms throughout
England and Wales yesterday that the forecasting models were not able
to suggest locations for as it can be such a random process. Given
this threat, but no guidance of where they could be, would you want
them to stay quiet?

Richard

It's the same pitch from you Richard. There were so many areas in
which storms were forecast yesterday and didn't happen. You admit, in
one sentence "Which is down to the basic difficult of forecasting the
situation which I get the impression you fail to grasp. Convective
storms were forecast to be isolated events yesterday, the MO didn't
have a full
grasp on where they were to form so what do you do? Issue nothing? No,
you issue warnings for where thunderstorms could happen
(unfortunately, the Kent example evaded their model)."

After saying in the previous; "Because I realise your view of
"outcomes" is far, far different from mine in weather forecasting.
Yesterday there were scattered severe thunderstorms. Apart from the
Kent convergence line cock-up (that I pointed to in another thread)
that led to an amended warning, they did a pretty good job. In your
"outcome" book, they didn't as it was just a wishy-washy broad
warning. We have two entirely different points of view.

I understand well the difficulties of forecasting in this situation.
At 12-36 hours it is impossible to forecast accurately on the 3
perameters I've mentioned and you have studiously ignored. Why not
admit to that instead of trying to defend such poor outcomes? It would
be so good if the MetO did not "dumb down" the situation and treated
it's public, on this particular (note, particular) part of its site to
a lot more honesty. Why not say, outright "We know the difficulties of
forecasting this difficult situation, but we've produced the best
forecast we can. Some of you will get showers, some won't"? Then
explain in some detail why forecasting this kind of situation is so
difficult.

My semi colons separated (accurately - you can't have read Keats, or
Virginia Wolfe *)) ) 4 different, recent examples from just the SW
where 12-36 hour "flash" warnings were not correct (could any of those
4, well known and well discussed examples have been good forecasts, by
any outcome means that you could conjure from your "different"
viewpoint?) From the point of view of the general public living in
those areas, or stuck in the snow on the A380, or A30 on that February
night, or were flooded out, or died in flash floods - and that's the
viewpoint that actually counts - they were not correct and the pretty
extreme outcomes were not as predicted in any of the "flash" warnings,
only 12 hours in advance.

On 28 June, 15:36, Dawlish wrote:

It's the same pitch from you Richard. There were so many areas in
which storms were forecast yesterday and didn't happen.

I refer you to my earlier comment. The forecast was only ever for
isolated storms.

You admit, in
one sentence "Which is down to the basic difficult of forecasting the
situation which I get the impression you fail to grasp. Convective
storms were forecast to be isolated events yesterday, the MO didn't
have a full
grasp on where they were to form so what do you do? Issue nothing? No,
you issue warnings for where thunderstorms could happen
(unfortunately, the Kent example evaded their model)."

After saying in the previous; "Because I realise your view of
"outcomes" is far, far different from mine in weather forecasting.
Yesterday there were scattered severe *thunderstorms. Apart from the
Kent convergence line cock-up (that I pointed to in another thread)
that led to an amended warning, they did a pretty good job. In your
"outcome" book, they didn't as it was just a wishy-washy broad
warning. We have two entirely different points of view.

And?

I understand well the difficulties of forecasting in this situation.
At 12-36 hours it is impossible to forecast accurately on the 3
perameters I've mentioned and you have studiously ignored. Why not
admit to that instead of trying to defend such poor outcomes?

Yesterday wasn't a poor outcome, in my book. It two different points
of view again.

It would
be so good if the MetO did not "dumb down" the situation and treated
it's public, on this particular (note, particular) part of its site to
a lot more honesty. Why not say, outright "We know the difficulties of
forecasting this difficult situation, but we've produced the best
forecast we can. Some of you will get showers, some won't"? Then
explain in some detail why forecasting this kind of situation is so
difficult.

My semi colons separated (accurately - you can't have read Keats, or
Virginia Wolfe *)) *)

It's more from a "readability" point of view.

4 different, recent examples from just the SW
where 12-36 hour "flash" warnings were not correct (could any of those
4, well known and well discussed examples have been good forecasts, by
any outcome means that you could conjure from your "different"
viewpoint?)

I'd like to see the before, during and after before I take your word
for their incorrectness given you see yesterday's warnings as poor.
I'd like an impartial view rather than bee-in-bonnet view. Flash
warnings will cover an area. Same story - a severe thunderstorm cell
will not cover an entire county. People are always going to miss out.

From the point of view of the general public living in
those areas, or stuck in the snow on the A380, or A30 on that February
night, or were flooded out, or died in flash floods - and that's the
viewpoint that actually counts - they were not correct and the pretty
extreme outcomes were not as predicted in any of the "flash" warnings,
only 12 hours in advance.

Fair enough - if the forecast failed to miss a severe weather event
then slapped wrists all round but I will happily defend the MO for
issuing warnings when there is a broad-scale threat of severe weather.

I can't help thinking though that you'll always be right in your own
mind and discussions with you go round and round in circles, as a
couple of people have mentioned to me off-line. As before - on severe
weather warnings, we'll have to agree to disagree on their purpose and
extent - not that you'd let it lie of course, judging from your
discussions with Lawrence and Mr Weatherlawyer.

Richard