Chilly few days!
Graham
Max yesterday was only 10.7c and today 10.3c after a low of 5.0c this
morning.
Recorded my wettest September on record (1960-2008) with 154.3mm a
massive 222% of the average, beating the 147.1mm set way back in 1965!
The last 3 months have all recorded over 100mm of rainfall a total of
383mm, which is 190% of the average.
The same 3 months have produced only 332 hours sunshine, just 77% of the
average.
Plenty of rainfall for the first two days of October and plenty more for
the coming weekend. Just 20mm short of my total for the whole of last
Autumn
and only 58mm short of my 1971-2000 Autumn average with still a month
and 29 days to go!
Have the met office change their prediction of a drier than normal
autumn
yet, obviously they need too fast:)
Weston Coyney weather station (North Staffordshire) 220 metres asl
--
Graham
Keith(Southend)
> Have the met office change their prediction of a drier than normal autumn
> yet, obviously they need too fast:)
>
> Weston Coyney weather station (North Staffordshire) 220 metres asl
Not to mention changing the mild winter forcast to cold as well ;-)
--
Keith (Southend)
http://www.southendweather.net
e-mail: kreh at southendweather dot net
Will Hand
"Keith(Southend)" <keith_...@hotmail.com> wrote in message
news:gc3gr9$f38$1...@registered.motzarella.org...
> Graham wrote:
>
>> Have the met office change their prediction of a drier than normal autumn
>> yet, obviously they need too fast:)
>>
>> Weston Coyney weather station (North Staffordshire) 220 metres asl
>
> Not to mention changing the mild winter forcast to cold as well ;-)
You know what Keith, I have a feeling about this winter, just look at the
fantasy GFS charts, I have seen warmer looking charts in recent Januarys,
they have an old-fashioned look about them, or am I just seeing too much
into them - probably?
Will
--
Scott W
> "Keith(Southend)" <keith_harr...@hotmail.com> wrote in message
Don't you "have a feeling" about most winters, Will?
The autumn will probably promise so much and then, hey presto, come
December 26th the charts will forecast the vengeful return of the
south-westerlies...
We can only hope I suppose...
On saying that I've heard that there are many parallells between this
year and 1978 - at least events wise - and the winters then were superb
Graham Easterling
Just glanced, and they look warmer to me (and least they do at the
moment!)
http://91.121.93.17/pics/Rtavn1441.png
http://91.121.93.17/pics/Rtavn2161.png
http://91.121.93.17/pics/Rtavn2881.png
I rather hope so, seems a bit early to pack the board and wetsuit
away.
Graham
Penzance
John Hall
Will Hand <wi...@lyneside.demon.co.uk> writes:
>You know what Keith, I have a feeling about this winter, just look at the
>fantasy GFS charts, I have seen warmer looking charts in recent Januarys,
>they have an old-fashioned look about them, or am I just seeing too much
>into them - probably?
Doesn't GFS produce dramatic northerly plunges at 10-15 days range
almost every winter, though, which almost never come to pass? However
one thing seems likely to be different this winter from recent winters;
the North Sea and the Channel seem likely to be several degrees cooler,
given the weather over recent months. So that must tend to favour
slightly cooler conditions than those of recent winters. Of course if
the synoptic patterns give us predominantly roaring SWerlies, then it's
not going to make much difference.
--
John Hall
"If you haven't got anything nice to say about anybody, come
sit next to me."
Alice Roosevelt Longworth (1884-1980)
Norman
For what they are worth, the latest NCEP progs show cold NE'lies across
the country in the week leading up to Christmas but with strengthening
mild W-SW winds spreading in on Christmas Day and intensifying on
Boxing Day. The next run will probably tell a different story :-)
Norman
--
Norman Lynagh
Chalfont St Giles, Buckinghamshire
85m a.s.l.
(remove "thisbit" twice to e-mail)
Keith(Southend)
I think we're just keeping the dream alive.
<now wait for anti snow & cold grief>
Pete L
I've waded threough the NCEP site but cannot find any mention of their
'forecast' charts for Europe. Can you send a link please, Norman? Tks.
Norman
Not available for free, I'm afraid. I see them on Netweather Extra, the
subscription part of www.netweather.tv. They show surface forecast
charts for Europe and the North Atlantic as far west as the Canadian
Maritimes based on the Climate Forecast System, a model run by NCEP.
It's run once daily and produces global output at 12-hourly intervals
out to 9 months ahead. All good light entertainment :-) The computing
power involved must be mind-boggling.
lawre...@sky.com
> "Keith(Southend)" <keith_harr...@hotmail.com> wrote in message
Will, I feel exactly the same way. It almost feels like an Autumns of
our childhoods, Just witness MoTD earlier this evening. espescially
the Blackburn v United game,
Dawlish
I'll give you 5/2 the winter being colder than average!
It's been a chilly start to October, but I don't see it continuing. I
wrote this morning update on http://totallyweatherandclimate.co.uk/
this morning:
"I think the gfs got the collywobbles yesterday. Not a very
meteorologically exact term, I know and the operational runs on the
12z and 18z were certainly outliers, but for the model to decide that
high pressure over the Baltic was likely (12z), followed by high
pressure over us (18z), followed by this morning's T240 chart (left)
as the most likely outcomes over 3 runs, is bizarre. I described it as
"head-shakingly odd" on the "Talk models" thread last night and and so
it proved!!
The ECM was having none of it and generally continues to push the
returning polar maritime air option, certainly for the NW of the UK,
from the 12z yesterday. I wonder if it will have the same misgivings
about the proposed SW flow when the 00z is released? It provides an
awkward position for me in T240 forecasting as the gfs' uncertainty
leaves me with the same misgivings, as my 5-run criterion will remain
unfulfilled until this evening, at the earliest. A SW airstream is
likely, but yesterday evening's gfs "collywobbles" add enough
uncertainty to leave me below my 80% certainty forecasting threshold.
One subtle change from yesterday is that mT air is shown to have more
of an influence, at times, over the next 10 days, on the SE, whereas
the further north and west you go, the slightly cooler, returning
polar maritime air, is shown to be in charge.
All this leaves this morning's likelihoods looking like this:
70% a SW flow, with high pressure to our SE and low pressure to our NW
20% High pressure dominance for the SE of the UK as the European high
exerts more of an influence.
10% A warm, southerly feed in a plume.
After a transient ridge produces a frosty morning tomorrow - and
possibly the first one for some in the Midlands - I don't see a great
deal of frosts out to mid-month. Lots of cloud and mT air over central
areas should see higher night time minima and the CET into positive
territory by mid month, though it has been a cool start (Manley CET at
-2.5C to the 3rd. C/O www.climate-uk.com).
Finally; quite a split across the UK this morning between sunny and in
some parts frosty, north of the M62 and some dreadfully wet
conditions, overnight and continuing today, to the South of that."
Paul
John Hall
<5e14a351-ee74-4dd1...@c60g2000hsf.googlegroups.com>,
Before I'd consider taking that bet, I'd want to know what you mean by
average.
Dawlish
> In article
> <5e14a351-ee74-4dd1-9aaa-61cbafe2b...@c60g2000hsf.googlegroups.com>,
Have a look on totallyweatherand climate.co.uk John.
Martin Rowley
Have a look on totallyweatherand climate.co.uk John.
... he's getting desperate!
Martin.
--
Martin Rowley
West Moors, East Dorset (UK): 17m (56ft) amsl
Lat: 50.82N Long: 01.88W
NGR: SU 082 023
Pressure In Berne
A spam abuse report needed I think.
Richard Dixon
> Doesn't GFS produce dramatic northerly plunges at 10-15 days range
> almost every winter, though, which almost never come to pass?
Model availability does not a good model make ! Sadly so many of the
discussions / outputs on the internet these days revolve around the
GFS just because you can get a plethora of freely-available (and
useful, in diagnostic sense - shame about the forecast) diagnostics
every 6 hours. Come December people will be hanging on the every word
of 06Z GFS out to T+144.
Although it might be nice if some of the European agencies followed
the GFS lead and made available some of their model output at a higher
frequency - even they create the diagnostics themselves and just
update pictures. We've seen radar at 15 mins available on the web, and
the next-generation Meteosat also, so a little bit of model output
over and above PMSL charts wouldn't go amiss.
Richard
Richard Dixon
> Have a look on totallyweatherand climate.co.uk John.
Politely - do we *really* need yet another UK weather forecasting
site? Don't tell me you'll be issuing your own weather warnings too
like all the others? The word "overkill" springs to mind.
Regards,
Richard
John Hall
<71a01f18-8e01-4e3a...@d70g2000hsc.googlegroups.com>,
Richard Dixon <rdng...@yahoo.co.uk> writes:
>Model availability does not a good model make ! Sadly so many of the
>discussions / outputs on the internet these days revolve around the
>GFS just because you can get a plethora of freely-available (and
>useful, in diagnostic sense - shame about the forecast) diagnostics
>every 6 hours. Come December people will be hanging on the every word
>of 06Z GFS out to T+144.
I think that at the longer time-frames there ought to be much more
emphasise on what the ensemble is saying, rather than concentrating
almost entirely on the operational run as most discussions seem to do.
Dawlish
> Richard Dixon <rdngem...@yahoo.co.uk> writes:
>
> >Model availability does not a good model make ! Sadly so many of the
> >discussions / outputs on the internet these days revolve around the
> >GFS just because you can get a plethora of freely-available (and
> >useful, in diagnostic sense - shame about the forecast) diagnostics
> >every 6 hours. Come December people will be hanging on the every word
> >of 06Z GFS out to T+144.
>
> I think that at the longer time-frames there ought to be much more
> emphasise on what the ensemble is saying, rather than concentrating
> almost entirely on the operational run as most discussions seem to do.
> --
> John Hall
> "If you haven't got anything nice to say about anybody, come
> sit next to me."
> Alice Roosevelt Longworth (1884-1980)
At 10 days, the ensembles are simply spaghetti. I defy anyone to
forecast with any accuracy from them and no-one can, emphasising my
point perfectly. You are much better off following changes in the
operational run. After all, the computer model selects this as the
most likely option.
Paul
Dawlish
> In article
> <71a01f18-8e01-4e3a-81de-7e3892774...@d70g2000hsc.googlegroups.com>,
>
> >Model availability does not a good model make ! Sadly so many of the
> >discussions / outputs on the internet these days revolve around the
> >GFS just because you can get a plethora of freely-available (and
> >useful, in diagnostic sense - shame about the forecast) diagnostics
> >every 6 hours. Come December people will be hanging on the every word
> >of 06Z GFS out to T+144.
>
> I think that at the longer time-frames there ought to be much more
> emphasise on what the ensemble is saying, rather than concentrating
> almost entirely on the operational run as most discussions seem to do.
> --
> John Hall
> "If you haven't got anything nice to say about anybody, come
> sit next to me."
> Alice Roosevelt Longworth (1884-1980)
So you don't fancy the bet, John? If you feel that this UK warming
trend is not likely to continue, 5/2 has to be decent odds! <ignores
stalker>
John Hall
<8b532c0e-61fe-44ce...@i76g2000hsf.googlegroups.com>,
Dawlish <pjg...@hotmail.com> writes:
>So you don't fancy the bet, John? If you feel that this UK warming
>trend is not likely to continue, 5/2 has to be decent odds! <ignores
>stalker>
I do expect the warming trend to continue, but there is still likely to
be the occasional somewhat colder winter IMO. I find it hard to believe
that the amount of warming could have already swamped the effect of
year-to-year variability. So 5/2 looks tempting. If when you say "colder
than average", you are referring to the 1971-2000 average, it's much
more tempting than if you're using 1961-90, though. That was what lay
behind my question of what average you were using.
John Hall
<f78e3c79-6a79-4b6e...@k30g2000hse.googlegroups.com>,
Dawlish <pjg...@hotmail.com> writes:
>At 10 days, the ensembles are simply spaghetti. I defy anyone to
>forecast with any accuracy from them and no-one can, emphasising my
>point perfectly.
On the rare occasions that the majority are in fair agreement, you can
say with reasonable confidence that that is the likely outcome. Or they
may fall into two distinct camps, which can be informative. Of course if
almost all of them are showing different solutions then one isn't much
the wiser.
> You are much better off following changes in the
>operational run. After all, the computer model selects this as the
>most likely option.
I believe it's run at a finer resolution than the other models, because
of computer power constraints. So yes, other things being equal it gives
the most likely solution. But if it's on its own, or has only one or two
supporters amongst the other ensemble members, then at the least it
should be viewed with greater caution than usual.
Steve Willington
> operational run. After all, the computer model selects this as the
> most likely option.
>
> Paul
Now that statement could do with some explanation...
John Hall
John Hall <nospam...@jhall.co.uk> writes:
>In article
><f78e3c79-6a79-4b6e...@k30g2000hse.googlegroups.com>,
> Dawlish <pjg...@hotmail.com> writes:
>
>> You are much better off following changes in the
>>operational run. After all, the computer model selects this as the
>>most likely option.
>
>I believe it's run at a finer resolution than the other models, because
>of computer power constraints. So yes, other things being equal it gives
>the most likely solution. But if it's on its own, or has only one or two
>supporters amongst the other ensemble members, then at the least it
>should be viewed with greater caution than usual.
Sorry, that was supposed to read "it's run at a finer resolution than
the other ensemble members".
Dawlish
All points very true and well made. I'm not sure I've ever seen the
ensembles in any kind of agreement at 10 days. Has it ever happened?
It would be unbelievably unusual for runs set with different
perameters to come out in agreement at a distance of 10 days. The odds
against that must be astronomical. I do feel that there is a terrific
amount of percieved wisdom about the use of the ensembles:
professional weather forecasters, especially the MetO use them, so by
association, they must therefore be useful.
Not you John, as you write with understanding on this topic, but
others, sometimes professionals, then go on to say something like;
"they should be used together with other tools". I despair a little
when I hear that. The outcomes, in terms of accuracy, at 10 days plus,
shout to me that the "tools" being used, including the use of
ensembles, are very crude and of limited use. The Met Office do know
this. That's why they group their longer period forecasts under the
6-15 day banner and don't track the forecast accuracy well, in the
public domain.
Forecast accuracy has to be judged by outcomes alone and not by the
tools used, or the professional qualifications of the forecasters
concerned. If the outcome accuracy forecast percentages are low, what
does it say about all the inputs?
What it says to me is that forecast accuracy at 10 days is
frustratingly poor and we should celebrate any kind of accuracy. To
Richard's earlier question of "Do we need another forecasting site?",
I'd ask; "how accurate is anyone's forecasting at 10 days plus?" If
you are happy with forecasting at that distance, we don't need
anything more, I agree.
Totally! is certainly not simply a forecasting site, however. It's a
climate site, as well as a weather site and there's a lot more
besides, which I don't believe is offered on other similar sites. It
has been very enjoyable setting it up and I've always follow the maxim
"If you don't like what's available, don't just sit and complain, or
put up with what's there, but go and do something about it!". So I did.
Dawlish
Richard Dixon
> I think that at the longer time-frames there ought to be much more
> emphasise on what the ensemble is saying, rather than concentrating
> almost entirely on the operational run as most discussions seem to do.
There is also the added complication of the ensembles being run at a
lower resolution and the higher resolution main run going completely
out of kilter with the lower resolution ensembles !
Richard
Richard Dixon
> I do feel that there is a terrific
> amount of percieved wisdom about the use of the ensembles:
> professional weather forecasters, especially the MetO use them, so by
> association, they must therefore be useful.
Equally the blind faith put in a single model (especially the GFS) and
inter-run differences. The only way I'm ever going to believe a model
at T+144 not inter-run consistency in the model itself, but agreement
in the other models. In which case this is the only website you'll
ever need!
http://meteocentre.com/models/compar_models.php?area=eur&lang=fr&run=00&type=z&hour=144&range=glb
> The outcomes, in terms of accuracy, at 10 days plus,
> shout to me that the "tools" being used, including the use of
> ensembles, are very crude and of limited use. The Met Office do know
> this. That's why they group their longer period forecasts under the
> 6-15 day banner and don't track the forecast accuracy well, in the
> public domain.
I had the fortune of seeing Met Office internal guidance when doing
some post-doc work a few years ago. Believe me, they look at more than
ensembles in their 6-15 day range and look at all the other models
available. The phrase "Poor man's ensemble" used when comparing models
I think is a misnomer: agreement between models I believe is much more
powerful than the differences between ensembles in one model. They're
not daft at the Met Office, much as many like to shoot them down.
> What it says to me is that forecast accuracy at 10 days is
> frustratingly poor and we should celebrate any kind of accuracy. To
> Richard's earlier question of "Do we need another forecasting site?",
> I'd ask; "how accurate is anyone's forecasting at 10 days plus?" If
> you are happy with forecasting at that distance, we don't need
> anything more, I agree.
My point is whether we need people all interpreting the same outputs
and regurgitating forecasts. The Met Office has much more model output
to hand than any of the "Amateur" (and I use the term kindly)
forecasting websites, so really the Met Office website should be a
instant go-to, especially now as they issue wordy 6-15 day forecasts.
> "If you don't like what's available, don't just sit and complain, or
> put up with what's there, but go and do something about it!". So I did.
Fair enough then and hopefully you won't be repeating everything you
can already get them from the Met Office, UKWeatherWorld,
TheWeatherOutlook, netweather.tv, Metcheque......
Richard
Steve Willington
Why would you be better off following the changes of the operational
run and what do you mean by "After all, the computer model selects
this as the most likely option"?
I can only offer my interpretation of operational models and ensemble
use if I understand what you are getting at.
Steve.
Steve Willington
> On 7 Oct, 11:53, Dawlish <pjg...@hotmail.com> wrote:
>
> > I do feel that there is a terrific
> > amount of percieved wisdom about the use of the ensembles:
> > professional weather forecasters, especially the MetO use them, so by
> > association, they must therefore be useful.
>
> Equally the blind faith put in a single model (especially the GFS) and
> inter-run differences. The only way I'm ever going to believe a model
> at T+144 not inter-run consistency in the model itself, but agreement
> in the other models. In which case this is the only website you'll
> ever need!
>
Hi Richard,
How many years ago did you have access to the internal guidance? We
also have access to initial perturbations, difference tracking etc,
etc - the ensembles are hard to interpret and to give weight to
without an idea of these and other factors.
Dawlish
>
> > I do feel that there is a terrific
> > amount of percieved wisdom about the use of the ensembles:
> > professional weather forecasters, especially the MetO use them, so by
> > association, they must therefore be useful.
>
> Equally the blind faith put in a single model (especially the GFS) and
> inter-run differences. The only way I'm ever going to believe a model
> at T+144 not inter-run consistency in the model itself, but agreement
> in the other models. In which case this is the only website you'll
> ever need!
>
Yes, but Richard; you miss my point. Why with all this data,
ensembles, models etc. is the forecast accuracy so frustratingly low?
All the other organisations you mention do no better and don't even
keep accuracy figures, so how can their accuracy possibly be judged?
Every one of those organisations that you mention in your last
sentence, conveniently forgets almost every 10-day forecast they ever
make and doesn't return to them on a regular basis. Guess when they
remember one and remind their punters of it! *>)). I honestly don't
repeat anything seen on any other organisation's site and you'll
quickly see that if you look at my forecasts.
I'm a great supporter of the MetO, not a detractor, but this is an
area in which I have felt depserately frustrated at forecasting
accuracy progress for almost 30 years. I'm not alone, by a long way.
It is a frustration to anyone who has been involved in forecasting.
Again, because of my frustration, I'm doing something different and
finding some success. I've also been prepared to publish every single
one of 64 forecasts and allow scrutiny and judgement by others.
Wouldn't it be great, if good old Piers did that? There are many
occasions when I feel that forecasting at T240+ is impossible with
accuracy, when the gfs has major inconsistencies between runs and the
gfs and the ECM don't agree, but there are other times where I feel it
is possible to forecast, 10 days hence, with an accuracy of 75%+. I
don't use ensembles at all in those forecasts, as I feel they simply
muddy the waters. I do use the EMS for comparison, but, of course,
that goes no further than T240 and I agree with you that agreement
between the ECM and the gfs is important at T240.
The meteocentre stuff is for 5 days only, not 10 days, but thanks.
NOAA produce accuracy comparison figures to that distance and the ECM
regularly comes out as the most accurate model, but at 10 days, I
wouldn't trust looking at a single run of any model. Neither would I
trust the Meto 6-15 day forecast, when it mentions that distance. It's
often 4 gfs runs and 2 EMS runs out of date anyway.
Best wishes, Paul
Dawlish
>
> > I do feel that there is a terrific
> > amount of percieved wisdom about the use of the ensembles:
> > professional weather forecasters, especially the MetO use them, so by
> > association, they must therefore be useful.
>
> Equally the blind faith put in a single model (especially the GFS) and
> inter-run differences. The only way I'm ever going to believe a model
> at T+144 not inter-run consistency in the model itself, but agreement
> in the other models. In which case this is the only website you'll
> ever need!
>
Yes, but Richard; you miss my point. Why with all this data,
wouldn't trust looking at a single run of any model, or any one model
more than another. Neither would I trust the Meto 6-15 day forecast,
when it mentions that distance. The one on the site is often 4 gfs
Pressure In Berne
No, you didnt do "something about it", you made your own web site
because your banned from all the others!
PS Your site is rubbish.
Dawlish wrote:
Dawlish
>
> - Show quoted text -
8>)) Flattery will get you nowhere.
Richard Dixon
> How many years ago did you have access to the internal guidance? We
> also have access to initial perturbations, difference tracking etc,
> etc - the ensembles are hard to interpret and to give weight to
> without an idea of these and other factors.
1996 - 2000, whilst affiliated to the JCMM at Reading. Happy days of
all manner of data at your fingertips.
Richard
Richard Dixon
> Yes, but Richard; you miss my point. Why with all this data,
> ensembles, models etc. is the forecast accuracy so frustratingly low?
Because of the inherent uncertainties in
a) the initial conditions of the models
b) simulating the atmosphere numerically
and
c) chaos !
I would have thought you might have grasped this given the verbosity
of a number your emails I've noticed whilst lurking on here.
> All the other organisations you mention do no better and don't even
> keep accuracy figures, so how can their accuracy possibly be judged?
They do keep accuracy figures. Internally. Don't ask me why they're
not released, though. Maybe they are, somewhere.
> I'm a great supporter of the MetO, not a detractor, but this is an
> area in which I have felt depserately frustrated at forecasting
> accuracy progress for almost 30 years.
From figures I've seen, I seem to recall 7-day forecast accuracy is
similar to that of 3-day about 35 years ago. There is a great amount
of "I remember that forecasts used to be better in the past" in
general chatter which is plain b*ll*cks and nostalgia if you ask me.
From the rest of your post, I think you're asking a lot of weather
forecasts. As someone once reminded me: "It's a model". Get the speed
of a weather front wrong by 5 mph, 3-days ahead and that's 360 miles
difference: not far off the width of mainland UK - the difference
between a wet morning and a dry afternoon and a dry morning and a wet
afternoon. Forecasters have my sympathy !
Richard
Steve Willington
Too much data at times - although I'd soon moan if some were
missing....
1996 - 2000 is a little before my time, I moved to the Guidance Unit
in 2005 as Deputy Chief and hence generated the medium range guidance
during that period before moving to Chief 18 months ago.
Dawlish
Really I do know the answer to the question I posed! The implication
of my question is that the use of ensembles etc. is only as useful as
the outcome accuracy shows it is. It may well be that 7-day accuracy
is similar to 3-day accuracy 35 years ago, but it's 10-day accuracy
that I'm concerned with and that has not increased by anything like
the degree of inprovement that has been achieved in 3-7 day accuracy -
shown by the figures the met office does produce for the wider
public.
Should the fact that "I ask a lot of weather forecasts" stop me asking
the question then and we should all sit back and accept the few words
that are written about the weather in 10days+ time, in the 6-15-day
update; never been told how accurate it is; never even easily figuring
out exactly which dates are actually being referred to? If 10-day
forecasting is so poor (the implication of your statement about
"asking too much of weather forecasts"), there should be a health
warning to that effect on that 6-15 day forecast. I know the accuracy
is low. I suspect that you do too. There is little wonder, then, that
the typical reaction of the layman reading that forecast is along the
lines of; "they never get the forecast right, these weather people".
At 10 days+, they are almost correct - for all the good reasons you
give and that frustrate me to hell!
Paul
PS You do a good line in verbosity yourself - and that's not a
criticism! It's good to raise the subject and talk, but a newsgroup is
a wonderful anachronism and is really an awful structure for
discussion. It's not organised for that purpose. totally! is.
PPs Steve, again; any chance of explaining; why? You did raise the
problem.
Richard Dixon
> It may well be that 7-day accuracy
> is similar to 3-day accuracy 35 years ago, but it's 10-day accuracy
> that I'm concerned with and that has not increased by anything like
> the degree of inprovement that has been achieved in 3-7 day accuracy -
I refer the honorable gentleman to my "chaos" comment that will limit
improvements in accuracy the further ahead in time we go. The Met
Office have got it write - wordy stuff beyond 6-15 days not going into
daily detail.
> Should the fact that "I ask a lot of weather forecasts" stop me asking
> the question then and we should all sit back and accept the few words
> that are written about the weather in 10days+ time, in the 6-15-day
> update; never been told how accurate it is; never even easily figuring
> out exactly which dates are actually being referred to? If 10-day
> forecasting is so poor (the implication of your statement about
> "asking too much of weather forecasts"), there should be a health
> warning to that effect on that 6-15 day forecast.
Agreed. But then again if you look at the wording of the 6-15 day
forecasts they don't imply it's set in stone. But this could be
improved yes and part of this is down to the education of the public
that the forecasts have inaccuracy.
> I know the accuracy
> is low. I suspect that you do too. There is little wonder, then, that
> the typical reaction of the layman reading that forecast is along the
> lines of; "they never get the forecast right, these weather people".
Yes, of course I suspect the accuracy is low - I realise that weather
forecasts are based on attempts to solve the equations of motion and
thermodynamics by numerical *approximations*. Errors will creep into
the system from the approximation method, the initial conditions, the
parameterization of sub-grid scale processes. I often am amazed that
we can forecast 2 days ahead on occasions, and not surprised when we
get a forecast mess-up in knife-edge extreme events.
> a newsgroup is
> a wonderful anachronism and is really an awful structure for
> discussion.
I've been posting on and off for 10 years or so on here and don't see
any particular problems, just sometimes there are periods of, let's
say, "saturation"....
Richard
Richard Dixon
> Too much data at times - although I'd soon moan if some were
> missing....
>
> 1996 - 2000 is a little before my time, I moved to the Guidance Unit
> in 2005 as Deputy Chief and hence generated the medium range guidance
> during that period before moving to Chief 18 months ago.
Nice that we have a "chief" taking note of stuff on here !
Pity we don't get to see the internal guidance available on the main
Met Office site - but I can just see plagiarism happening (as I recall
from a certain poster many moons ago on here from his "personal
sources").
Richard
Steve Willington
>
> PPs Steve, again; any chance of explaining; why? You did raise the
> problem.
Morning Paul,
I think you are misinterpreting me, I'm not raising a problem, simply
asking for the your rationale behind the two statements you made.
Jon O'Rourke
news:e2d62f4b-398d-4bde...@k30g2000hse.googlegroups.com...
> Nice that we have a "chief" taking note of stuff on here !
.. there's a number of chiefs (and many other forecasters) that take an
active interest in what goes on in the ng even if we don't here from them
all.
>
> Pity we don't get to see the internal guidance available on the main
> Met Office site - but I can just see plagiarism happening (as I recall
> from a certain poster many moons ago on here from his "personal
> sources").
>
> Richard
Can't imagine who you mean :-)
Jon.
Jon O'Rourke
news:e2d62f4b-398d-4bde...@k30g2000hse.googlegroups.com...
> Nice that we have a "chief" taking note of stuff on here !
.. there's a number of chiefs (and many other forecasters) that take an
active interest in what goes on in the ng even if we don't hear from them
Mike Tullett
wrote in <news:6l3g2nF...@mid.individual.net>
>> Pity we don't get to see the internal guidance available on the main
>> Met Office site - but I can just see plagiarism happening (as I recall
>> from a certain poster many moons ago on here from his "personal
>> sources").
>>
>> Richard
>
> Can't imagine who you mean :-)
Nor can I :-)
--
Mike Tullett - Coleraine 55.13°N 6.69°W posted 08/10/2008 13:08:29 GMT
Dawlish
wrote:
And I asked why you have a problem with what I said.
Steve Willington
You made two statements - I'm asking for you to explain your thinking.
For one who expects others to justify their statements I think it only
fair that you do the same.
Dawlish
Thanks Richard. I feel we are quite close in our assessment of the
current state of medium distance forecasting, but I am frustrated with
it still. The problems of lack of computer power, the fineness of the
data grid and problems of data collection were there 25 years ago. but
I can remember the optimism that many of the problems would be solved
by the turn of the century, Sadly, the optimism was misplaced. It is
hard to see where improvements will come from.
More powerful computers may help - but the law of diminishing returns
comes in here and vastly increased computer capacity has not produced
commensurate increases in 10-day plus forecasting. A finer data grid
would help and that could come. I sort of favour the use of nanobots
here, each recording the conditions in every 100m3 section of the
atmosphere, or far better satellite monitoring, but both may be too
far fetched for the near future. Better models may help, but these are
dependent upon the data inputs. The forecasters themselves interpret
the data as well as possible. MetO forecasters are trained extremely
well and I'm not sure there's much more they can do. So how exactly
are we going to improve the forecasts into the future? Mr McCawber may
help, but the confidence to improve has to be judged on improvements
made so far and at T240+, the improvements, over the last quarter
century, have been frustratingly small.
Paul
John Hall
<dc3ec366-c3a3-4f68...@a18g2000pra.googlegroups.com>,
Agreed. I think that newsgroups offer a much better structure for
discussion than most web forums do, because of the threading. I see that
Dawlish is using Google, though, so may not be getting the benefit of
that.
Richard Dixon
> Sadly, the optimism was misplaced. It is
> hard to see where improvements will come from.
Little by little. But whilst you're solving the equations of motion
and thermodynamics by *numerical approximation* then I guess we'll
always be limited. Plus our friend chaos thrown in for good measure.
In NW Europe we'll always be hampered by lack of surface observations
upstream, no matter how many micro-planes we can guide across the the
Atlantic in targetting data-sensitive regions.
> More powerful computers may help - but the law of diminishing returns
> comes in here and vastly increased computer capacity has not produced
> commensurate increases in 10-day plus forecasting. A finer data grid
> would help and that could come.
I am most interested to see some every-day weather forecasting on a
global grid at "proper" mesoscale resolutions like that which is used
on the Earth Simulator in Japan (10km)? I believe the Met Office are
heading in this direction by 2010 or so. Again, I have no idea however
how increases to this resolution will help out to T+240: maybe there
are some results someone on here is able to dig out. I'd be happy to
see a notable 6-day forecast improvement before worrying about 10
days.
> but the confidence to improve has to be judged on improvements
> made so far and at T240+, the improvements, over the last quarter
> century, have been frustratingly small.
And I'm sure will remain (relatively) small by owing to numerical
approximation/simulation and dear old chaos.
Richard
Richard Dixon
> Agreed. I think that newsgroups offer a much better structure for
> discussion than most web forums do
Largely because of the inability to add animated "smileys" that should
be made a hanging offence.
Yours,
Frustrated of Blackheath
Steve Willington
So how exactly
> are we going to improve the forecasts into the future?
>
Paul,
Ensembles are still very much in their 'earlier years' (to avoid
saying infancy), there is much work going on into research into how
the perturbations are identified, generated (and from what), grown and
then how to interpret and best use the ensembles obtained etc, etc.
Improvements will be gradual but they will come. By day 10 one is
usually looking at predictability and weather type (esp changes in
type).
Will Hand
I hate web fora. Always having to use mouse clicks makes it so wearisome,
but they have their place and some are very good.
:-) :-( :-> :-O
Will
--
Will Hand
others I prefer those with observations on them (like the superb one run by
Roger Brugge) rather than one devoted to discussions on Climate Change and
T+240 (chaotic-ridden) GFS. But that is my choice and each to their own.
Will
--
"Dawlish" <pjg...@hotmail.com> wrote in message
news:2839cad5-e440-43f6...@s50g2000hsb.googlegroups.com...
Will Hand
"Mike Tullett" <use...@virgin.net> wrote in message
news:vmh7kpm4twv4$.dlg@miketullett.net...
"Come on you Reds", there's a clue peeps!
:-)
Will
--
Richard Dixon
> :-) :-( :-> :-O
Stop it you bugger !
Richard
Richard Dixon
> "Mike Tullett" <use....@virgin.net> wrote in message
>
> news:vmh7kpm4twv4$.dlg@miketullett.net...
>
>
>
> > On Wed, 8 Oct 2008 11:24:24 +0100, Jon O'Rourke <jon_orou...@hotmail.com>
> > wrote in <news:6l3g2nF...@mid.individual.net>
>
> >>> Pity we don't get to see the internal guidance available on the main
> >>> Met Office site - but I can just see plagiarism happening (as I recall
> >>> from a certain poster many moons ago on here from his "personal
> >>> sources").
>
> >>> Richard
>
> >> Can't imagine who you mean :-)
>
> > Nor can I :-)
>
> > --
> > Mike Tullett - Coleraine 55.13°N 6.69°W posted 08/10/2008 13:08:29 GMT
>
> "Come on you Reds", there's a clue peeps!
>
> :-)
>
> Will
> --
Shudder. On all three counts.