Since the high res. went operational on 15th I have noticed a tendency to
deepen lows all over the place compared to ECM/UK/JMA. Has anyone else
noticed this? Jon? It is important as such behaviour can cuase big
differences in forecast evolution. E.g. compare T+72s on
http://www.weatheronline.co.uk/cgi-b...=-&WMO=&ZOOM=0

Will
--

On Dec 19, 5:47*pm, "Will Hand" wrote:
Since the high res. went operational on 15th I have noticed a tendency to
deepen lows all over the place compared to ECM/UK/JMA. Has anyone else
noticed this? Jon? It is important as such behaviour can cuase big
differences in forecast evolution. E.g. compare T+72s onhttp://www.weatheronline.co.uk/cgi-bin/expertcharts?LANG=en&MENU=0000...

Will
--

Keep your eyes on these for the outcomes to 5/6 days Will.

http://www.emc.ncep.noaa.gov/gmb/STATS/html/acz5.html

The latest verifications will still be from the old model, but the new
res should kick into the verification stats in a couple of days time.
As usual, the ECM has recently been outperforming all the other models
but I imagine there will be a few American eyes looking at those
verification stats over the next month.

"Dawlish" wrote in message
...
On Dec 19, 5:47 pm, "Will Hand" wrote:
Since the high res. went operational on 15th I have noticed a tendency to
deepen lows all over the place compared to ECM/UK/JMA. Has anyone else
noticed this? Jon? It is important as such behaviour can cuase big
differences in forecast evolution. E.g. compare T+72s
onhttp://www.weatheronline.co.uk/cgi-bin/expertcharts?LANG=en&MENU=0000...

Will
--

Keep your eyes on these for the outcomes to 5/6 days Will.

http://www.emc.ncep.noaa.gov/gmb/STATS/html/acz5.html

The latest verifications will still be from the old model, but the new
res should kick into the verification stats in a couple of days time.
As usual, the ECM has recently been outperforming all the other models
but I imagine there will be a few American eyes looking at those
verification stats over the next month.
=====================

Excellent, many many thanks for that link Paul, I'll have a look now.

Another thought, I have seen models with similar problems before and the
cause was linked to over enthusiastic latent heat feedback loops.

PS I'm not trying to be over critical of GFS, which is a good model, these
things happen sometimes!

Will
--

Will Hand wrote:


"Dawlish" wrote in message
...
On Dec 19, 5:47 pm, "Will Hand" wrote:
Since the high res. went operational on 15th I have noticed a tendency to
deepen lows all over the place compared to ECM/UK/JMA. Has anyone else
noticed this? Jon? It is important as such behaviour can cuase big
differences in forecast evolution. E.g. compare T+72s
onhttp://www.weatheronline.co.uk/cgi-
bin/expertcharts?LANG=en&MENU=0000...

Will
--

Keep your eyes on these for the outcomes to 5/6 days Will.

http://www.emc.ncep.noaa.gov/gmb/STATS/html/acz5.html

The latest verifications will still be from the old model, but the new
res should kick into the verification stats in a couple of days time.
As usual, the ECM has recently been outperforming all the other models
but I imagine there will be a few American eyes looking at those
verification stats over the next month.
=====================

Excellent, many many thanks for that link Paul, I'll have a look now.

Another thought, I have seen models with similar problems before and the
cause was linked to over enthusiastic latent heat feedback loops.

PS I'm not trying to be over critical of GFS, which is a good model, these
things happen sometimes!


I think it was the UM's first appearance when it forecast a low of 920. We
thought it was overegging it but it was proved right and so we then thought
what a marvellous model it was. Unfortunately, it then went for a succession
of forecasts of 920 lows, none of which materialised.

--
Graham P Davis, Bracknell, Berks., UK. E-mail: newsman not newsboy
"I wear the cheese. It does not wear me."

"Will Hand" wrote in message
...
Since the high res. went operational on 15th I have noticed a tendency to
deepen lows all over the place compared to ECM/UK/JMA. Has anyone else
noticed this? Jon? It is important as such behaviour can cuase big
differences in forecast evolution. E.g. compare T+72s on
http://www.weatheronline.co.uk/cgi-b...=-&WMO=&ZOOM=0

Will

Has the model resolution actually changed with this new formulation, Will ?
I ask as Martin's post about it at 1743Z on the 17th didn't appear to
mention any change in resolution.

When Martin first mentioned the parallel suite's availability on Meteociel
some weeks ago I was comparing both outputs as often as I could and I did
wonder if it was more stable from run to run but that's a highly subjective
assessment. I wasn't particularly aware of it overcooking things.

Incidentally, EC's planned resolution upgrade is now scheduled for the 26th
January.

Jon.

"Dawlish" wrote in message
...
On Dec 19, 5:47 pm, "Will Hand" wrote:
Since the high res. went operational on 15th I have noticed a tendency to
deepen lows all over the place compared to ECM/UK/JMA. Has anyone else
noticed this? Jon? It is important as such behaviour can cuase big
differences in forecast evolution. E.g. compare T+72s
onhttp://www.weatheronline.co.uk/cgi-bin/expertcharts?LANG=en&MENU=0000...

Will
--

Keep your eyes on these for the outcomes to 5/6 days Will.

http://www.emc.ncep.noaa.gov/gmb/STATS/html/acz5.html

The latest verifications will still be from the old model, but the new
res should kick into the verification stats in a couple of days time.
As usual, the ECM has recently been outperforming all the other models
but I imagine there will be a few American eyes looking at those
verification stats over the next month.


Will clearly asked the opinion of the other "experts" ! ;-)

"Lawrence Jenkins" wrote in message
...
Will clearly asked the opinion of the other "experts" ! ;-)

But he is an "expert; An expert at conning and bullying others and at
pretending he is an expert.
--
Ned

"Jon O'Rourke" wrote in message
...
"Will Hand" wrote in message
...
Since the high res. went operational on 15th I have noticed a tendency to
deepen lows all over the place compared to ECM/UK/JMA. Has anyone else
noticed this? Jon? It is important as such behaviour can cuase big
differences in forecast evolution. E.g. compare T+72s on
http://www.weatheronline.co.uk/cgi-b...=-&WMO=&ZOOM=0

Will

Has the model resolution actually changed with this new formulation, Will
?
I ask as Martin's post about it at 1743Z on the 17th didn't appear to
mention any change in resolution.


Perhaps I was mistaken about a resolution change Jon? Anwhow changes were
made to the model on the 15th and since then it has looked a bit "weird" IMO
cf JMA/UKMO/ECM.

When Martin first mentioned the parallel suite's availability on Meteociel
some weeks ago I was comparing both outputs as often as I could and I did
wonder if it was more stable from run to run but that's a highly
subjective assessment. I wasn't particularly aware of it overcooking
things.


I agree about the stability, almost too stable at times. Did you compare the
charts I suggested from DT12Z today?

Incidentally, EC's planned resolution upgrade is now scheduled for the
26th January.

Ah! Thanks Jon.

Will
--

"Will Hand" wrote in message
...

I agree about the stability, almost too stable at times. Did you compare
the charts I suggested from DT12Z today?

I did and there's some pretty big differences between EC, the GFS and GM
before T+144 too..

I don't feel confident enough at this stage to say if this new formulation
is any worse (or better !) than the previous one - although you'd expect an
improvement due to the improved assimilation that Martin mentioned.
Certainly in recent days it does seem to have been more developmental to the
south of the UK next week compared to the operational output from EC and the
GM; whether that's a result of these recent changes I just don't know.

Either way there's clearly a lot of uncertainty from about Tuesday onwards,
more so than you would normally
expect in the T+72 to T+144 range. Spare a thought for the poor s*ds
currently scratching their heads on the Medium Range bench..

Jon.