It seems to me that every low coming up the channel since the great 1987
storm causes the same problems at the Met Office. They don't seem to have a
clue until the last minute, but go mad with the warnings anyway. Exactly
the same situation happens when there is a slight risk of snow and they
always get that wrong, especially in the SE.

I have no problems with gusts of 50-60mph. They should have listened to Joe
*******i after all!!!!!!

Shaun Pudwell.

Met office bashing is easy. Doing better than them not.


"Shaun Pudwell" wrote in message
...
It seems to me that every low coming up the channel since the great 1987
storm causes the same problems at the Met Office. They don't seem to have
a
clue until the last minute, but go mad with the warnings anyway. Exactly
the same situation happens when there is a slight risk of snow and they
always get that wrong, especially in the SE.

I have no problems with gusts of 50-60mph. They should have listened to
Joe
*******i after all!!!!!!

Shaun Pudwell.

In article ,
Shaun Pudwell writes:
It seems to me that every low coming up the channel since the great 1987
storm causes the same problems at the Met Office. They don't seem to have a
clue until the last minute, but go mad with the warnings anyway.

If a low comes up the Channel then a very small difference in its track
can make an enormous difference in the strength of the winds experienced
in southern England. The forecasters have to mention the possibility of
the worst case scenario, given the damage that it could cause, but they
might perhaps sometimes fail fully to get across the degree of
uncertainty in the forecast.

Exactly
the same situation happens when there is a slight risk of snow and they
always get that wrong, especially in the SE.

Again, a very small difference in the temperature or the humidity of the
air, not just at the surface but all the way up to the freezing level,
can have a major impact on what happens (and the temperature can change
quite quickly if the intensity of the precipitation changes). In
situations like this, we are still a long way from getting infallible
forecasts. The models seem now to handle the "broad brush" aspects
remarkable well, but the "devil is in the detail" still.
--
John Hall
"Sir, I have found you an argument;
but I am not obliged to find you an understanding."
Dr Samuel Johnson (1709-1784)

Being a computer programmer by trade, I can say with confidence "Crap In,
Crap Out". In otherwords, no matter how good the model, if you put in an
inadequate amount of information or the wrong type of information for that
matter, the resulting forecast will always be less than perfect.

From my point of view, tomorrows low pressure system looks like being far
more impressive than today's ( locally speaking of course ). I am sure the
French won't be seeing it that way though!

Shaun Pudwell.


"John Hall" wrote in message
...
In article ,
Shaun Pudwell writes:
It seems to me that every low coming up the channel since the great 1987
storm causes the same problems at the Met Office. They don't seem to
have a
clue until the last minute, but go mad with the warnings anyway.

If a low comes up the Channel then a very small difference in its track
can make an enormous difference in the strength of the winds experienced
in southern England. The forecasters have to mention the possibility of
the worst case scenario, given the damage that it could cause, but they
might perhaps sometimes fail fully to get across the degree of
uncertainty in the forecast.

Exactly
the same situation happens when there is a slight risk of snow and they
always get that wrong, especially in the SE.

Again, a very small difference in the temperature or the humidity of the
air, not just at the surface but all the way up to the freezing level,
can have a major impact on what happens (and the temperature can change
quite quickly if the intensity of the precipitation changes). In
situations like this, we are still a long way from getting infallible
forecasts. The models seem now to handle the "broad brush" aspects
remarkable well, but the "devil is in the detail" still.
--
John Hall
"Sir, I have found you an argument;
but I am not obliged to find you an understanding."
Dr Samuel Johnson (1709-1784)

"Shaun Pudwell" wrote in message
...
Being a computer programmer by trade, I can say with confidence "Crap
In,
Crap Out". In otherwords, no matter how good the model, if you put in
an
inadequate amount of information or the wrong type of information for
that
matter, the resulting forecast will always be less than perfect.

.... I think you summed it up quite well. The problem was always going to
be initialising the models correctly - get the starting point wrong and
the outcome will drift wildly in these particular situations,
particularly as the models become ever more sophisticated - a slight
nudge T+0 to T+2 could lead to major deviations at T+18. The best that
could be done in this situation (and I feel that, given the
uncertainties, the forecasters did do a creditable job), is to outline
the possibilities and let people judge.

Where I *do* have problems is the "up to xxx" terminology .. we've
talked about this before. I don't know where this crept in, but it
really isn't good enough to say 'up to 70 mph', or 'up to 40mm' or 'up
to 5 cm' etc. This is where we (professionals) are being let down by
dissemination system: I'm not sure where the 'blame' lies, but we really
have to tighten up on presentation.

Martin.

The other distinct problem is the fact that the GFS model was spot on, even
a full 24 hours out. During the 1987 storm, I am sure some of the other
models was predicting Hurricane force winds for the SE. Again the MO model
failed. They may have a new super computer, but have they really updated
their software? It looks like a variation of the same model with the same
weaknesses as before. Someone correct me if I'm wrong.

That's two major BAD forecasts so far from the MO this winter. First we
didn't get the xx cm of snow and blizzards a few weeks ago and now no wind,
and I mean ( NO WIND ). Its absolutely dead calm outside!

Shaun Pudwell.


"Martin Rowley" wrote in message
...

"Shaun Pudwell" wrote in message
...
Being a computer programmer by trade, I can say with confidence "Crap
In,
Crap Out". In otherwords, no matter how good the model, if you put in
an
inadequate amount of information or the wrong type of information for
that
matter, the resulting forecast will always be less than perfect.

... I think you summed it up quite well. The problem was always going to
be initialising the models correctly - get the starting point wrong and
the outcome will drift wildly in these particular situations,
particularly as the models become ever more sophisticated - a slight
nudge T+0 to T+2 could lead to major deviations at T+18. The best that
could be done in this situation (and I feel that, given the
uncertainties, the forecasters did do a creditable job), is to outline
the possibilities and let people judge.

Where I *do* have problems is the "up to xxx" terminology .. we've
talked about this before. I don't know where this crept in, but it
really isn't good enough to say 'up to 70 mph', or 'up to 40mm' or 'up
to 5 cm' etc. This is where we (professionals) are being let down by
dissemination system: I'm not sure where the 'blame' lies, but we really
have to tighten up on presentation.

Martin.

Just out of curiosity, is it at all possible that the situation could have
been the other way around. i.e, no warnings and a forecast of calm
conditions but when it comes to the event, damaging storm force winds? I
had thought that the MO had learned their lesson from the billions of pounds
worth of damage caused in 1987. From recent events, it looks like things
are no better now than they were back then.

I just hope someone from the MO has the guts to stand up and admit
catastrophic failure.

Shaun Pudwell.

"Shaun Pudwell" wrote in message
...
The other distinct problem is the fact that the GFS model was spot on,
even
a full 24 hours out. During the 1987 storm, I am sure some of the other
models was predicting Hurricane force winds for the SE. Again the MO
model
failed. They may have a new super computer, but have they really updated
their software? It looks like a variation of the same model with the same
weaknesses as before. Someone correct me if I'm wrong.

That's two major BAD forecasts so far from the MO this winter. First we
didn't get the xx cm of snow and blizzards a few weeks ago and now no
wind,
and I mean ( NO WIND ). Its absolutely dead calm outside!

Shaun Pudwell.


"Martin Rowley" wrote in message
...

"Shaun Pudwell" wrote in message
...
Being a computer programmer by trade, I can say with confidence "Crap
In,
Crap Out". In otherwords, no matter how good the model, if you put in
an
inadequate amount of information or the wrong type of information for
that
matter, the resulting forecast will always be less than perfect.

... I think you summed it up quite well. The problem was always going to
be initialising the models correctly - get the starting point wrong and
the outcome will drift wildly in these particular situations,
particularly as the models become ever more sophisticated - a slight
nudge T+0 to T+2 could lead to major deviations at T+18. The best that
could be done in this situation (and I feel that, given the
uncertainties, the forecasters did do a creditable job), is to outline
the possibilities and let people judge.

Where I *do* have problems is the "up to xxx" terminology .. we've
talked about this before. I don't know where this crept in, but it
really isn't good enough to say 'up to 70 mph', or 'up to 40mm' or 'up
to 5 cm' etc. This is where we (professionals) are being let down by
dissemination system: I'm not sure where the 'blame' lies, but we really
have to tighten up on presentation.

Martin.

On Mon, 12 Jan 2004 10:17:54 -0000, "Shaun Pudwell"
wrote:

Being a computer programmer by trade, I can say with confidence "Crap In,
Crap Out". In otherwords, no matter how good the model, if you put in an
inadequate amount of information or the wrong type of information for that
matter, the resulting forecast will always be less than perfect.

Despite the availability of satellite imagery and information, the amount of
data available over the Atlantic is pitifully small and this represents 3,000
miles of nearly bugger-all, datawise; and all of it upwind of us.

So it is more a case of very little in, in the circumstances pretty good out.

JPG

"JPG" wrote in message
...

Despite the availability of satellite imagery and information, the amount
of

QuikScat was supposed to be the NWP dream data input for situations like
this. The ability to calculate wind at 10 levels, temperature, humidity,
plus let's not forget the new Meteosat Satellite.

Perhaps Jon/Martin/Will know? Does the UKMO model use QuikScat as the GFS
does?

Regards

Andrew

"Andrew Bond" wrote in message
...


QuikScat was supposed to be the NWP dream data input for situations
like
this. The ability to calculate wind at 10 levels, temperature,
humidity,
plus let's not forget the new Meteosat Satellite.

Perhaps Jon/Martin/Will know? Does the UKMO model use QuikScat as the
GFS
does?


.... can't answer your question as put - i.e. does the *model* assimilate
the QuikScat data objectively (Jon or Will will be more up to date with
this aspect). These data *are* certainly used by duty staff to monitor
the NWP output. However, the problem lies in the PPN contamination - in
the areas that Q-S might be most useful, the data are regarded as less
than reliable: however, I have found in the past that this caveat was
over-stressed and that Q-S was always worth using across the full range
of output.

The problem with model assimiliation is often with humidity. There is
much data based on satellite irradiance retrieval that purports to
'load' the model atmosphere with the best-fit humidity profile - the
problem is that sometimes the humidity is assigned to the wrong level -
and thus spurious areas of precipitation etc., where none should occur.
Whether this humidity mis-match might feed through to other aspects of
the output (through latent heat exchanges etc.), is beyond my competence
to comment upon.

Martin.