Model Output Discussion

[Pennine Ten Foot Drifts]

Some people have a strange idea of 'downgrades'. Looking at (GFS) charts from Sunday night onwards, combining DPs, uppers and surface temps, there's a fair chance somewhere will see snow falling every night next week. Now for the 2nd week in December, if thats a 'downgrade' I'll take it every time.........

[Pennine Ten Foot Drifts]

Wasn't yesterday's 00z ECM a bit disappointing as well I seem to remember. How does today's compare with that ? (remembering the old adage which seems to have been ditched that you should only be comparing the same runs)

[londonsnow]

Sometimes I think that seeing so many charts a day gives us worries that years ago would not have existed, we all know that 00 will differ from 06 and so on -we should just be pleased that overall the promised spell has at this stage stayed very much as it was 2-3 days ago and fine tuning may well occur until maybe 12 hours prior to the main event. As TM said in a post earlier " in the 9 years of chart watching he cannot remember such synoptics", I fully concur.

I totally concur. these are some of the best charts for years, and so early in the season. we are spoilt by the internet, or maybe just stressed. in dec 1981 it just happened, admittedly from the north and was more of a surprise than anything else. anyway, decent models, mostly agreeing, now its all in the detail

[Neilsouth]

Nice GEM!

http://www.wzkarten.de/pics/Rgem1441.gif

http://www.wzkarten.de/pics/Rgem1681.gif

http://www.wzkarten.de/pics/Rgem1921.gif

http://www.wzkarten.de/pics/Rgem2161.gif

http://www.wzkarten.de/pics/Rgem2401.gif

[weather eater]

I had a feeling there would be a little consternation on here this morning after looking through the 00z model runs. UKMO is good, the ECM, is let’s say, not the best this morning, however its 00z run has been different for the last 4 in a row, ECM 12z run has been much better and much more consistent. As for the GFS, it’s just doing what it always does and toying with different evolutions, another one will be along shortly. The GFS parallel remains pretty solid, although I would say at lot of its runs are now indicating that this spell will only last for about a week before milder air moves in from the south.

[liltbrockie]

GFS looks like a downgrade so far... that high is not developing as quick

[Barometer Obama]

Sure the charts are a little of a worry, but yesterdays were truly amazing. I think we'd all have been happy a month ago if we knew that we'd be viewing charts like this in December!

http://91.121.94.83/modeles/gfsp/run/gfs-0-180.png?0

[Southender]

06z starting to roll.

What fantastic evolution will it throw up this morning?

Or like the ECM, will it start to wobble...

Edited December 10, 2009 by LeighShrimper

[stewfox]

I totally concur. these are some of the best charts for years, and so early in the season. we are spoilt by the internet, or maybe just stressed. in dec 1981 it just happened, admittedly from the north and was more of a surprise than anything else. anyway, decent models, mostly agreeing, now its all in the detail

Yes this is what I said although the post was removed

The idea you need inbedded cold before snow comes is not supported by historic set ups and is relevant to the model discussion.

[Geordiesnow]

I hope this morning's runs are a little bit of a wake up call to some on here and why you should never get excited about snow potential at such a long range.

Having saying that, we still don't know what will happen so the models could still upgrade once again in this respect.

[weather eater]

Nice GEM!

http://www.wzkarten.de/pics/Rgem1441.gif

http://www.wzkarten.de/pics/Rgem1681.gif

http://www.wzkarten.de/pics/Rgem1921.gif

http://www.wzkarten.de/pics/Rgem2161.gif

http://www.wzkarten.de/pics/Rgem2401.gif

Indeed it is, that’s a fantastic block.

I think we are just going to have to except that the models are really going to struggle with how this evolves, this will be crucial on our tiny isle. Over the last few runs the South has been spoiled for snowy synoptics, this morning’s runs are less so, but when some lose out, others gain. Let’s not lose sight of the fact, that this is potentially, a wonderful spell of weather for mid December, I can’t remember anything like it for this time of year, in my 47 years.

[cooling climate]

This from NOAA last night.

FOR SUNDAY DECEMBER 20 - WEDNESDAY DECEMBER 23: NO SPECIFIC HAZARDS CAN BE DEFINED AT THIS TIME RANGE DUE TO A LACK OF PREDICTABILITY. THE 0Z GFS ENSEMBLE MEAN INDICATES THAT THE NORTH ATLANTIC OSCILLATION (NAO) REMAINS NEGATIVE DURING THE next two weeks. THEREFORE, BELOW NORMAL TEMPERATURES WILL LIKELY PERSIST FOR THE EASTERN CONUS. THE LONGWAVE PATTERN ALSO TENDS TO FAVOR STORMY WEATHER ALONG THE EAST COAST. WITH A STRENGTHENING RIDGE OVER THE PACIFIC NORTHWEST, THE PACIFIC STORM TRACK MAY SHIFT SOUTHWARD.

FORECASTER: BRAD PUGH

The ECM has verified the poorest model in the last few days and it does seem to want to severe links

to the Siberian high which does not gel with what the other models are showing.

Never the less we are just talking detail the basic pattern is still there just differences with regard

to degree of cold and snow.

This is the 6-10 day pattern from the cpc last night.

Edited December 10, 2009 by cooling climate

[TEITS]

I hope this morning's runs are a little bit of a wake up call to some on here and why you should never get excited about snow potential at such a long range.

Having saying that, we still don't know what will happen so the models could still upgrade once again in this respect.

Sorry if I sound harsh (have a headache!) but I really wish people would stop referring to snow potential being in F.I.

For starters the snow potential begins within the reliable timeframe via the E,ly. The next possible stage to the cold spell which remains uncertain is only just outside the reliable timeframe i.e next Wed when we might see a cold front moving S bringing the risk of more snow.

This cold spell needs to be viewed in two stages and the first stage i.e the E,ly is almost agreed upon. Its the 2nd stage that remains far more uncertain.

[Tom Montalbano]

I think most sensible post of the morning goes to phil n.w. Looks to be quite solid agreement at around t96 now and I think we all would've banked those charts a few days ago. Strangely, there was pretty good agreement at around t168/t192 last night, that has now gone this morning and we see different solutions at that time range. Lets worry about mid/late next week in a few days time, for the moment I'm happy with the way Sun/Mon seems to be developing.

Right back to the 06z.

Regards,

Tom.

[Iceberg]

This might help, it might not though.Feel free to correct if required.

As a general rule of thumb the METO forecasting will be(maybe) looking at the following.

0-72Z.

Meto first of all and GFS, Do either of the runs differ much from the Ensemble mean ?, if not then do they differ from each other, if not use the METO, if they do differ from each other, why ?. Generally they would go with the METO, with a dose of GFS if they though it was more realistic.

72Z to 144Z

Again check to see if the ops runs differ much from the ENS mean.

If not then do they differ from each other.

If they do they would take the start point from above and perform a best fit with the 3 models, looking also at the trend in ECM and the ECM EPS and GFS ENS.

144Z to 240Z

For obvious reasons this would be made up mainly from the ENS, EPS and super ENS unless the ops closely followed them, in which case they would also include the ops run.

240Z >

Beyond 240Z they would use the extended ECM that goes out 30 days, CFS and GFS ensembles (probably not the GFS ops out to 360Z.)

As well as judgement, teleconnections etc.

Many might recall years ago something called MRF, before GFS. This was the Medium range forecasting model from NWS.

As well as MRF, the US NWS(national weather service) had a short range model(can't remember its name tbh). GFS replaced this with as an all singing all dancing model that could do short, medium and long term forecasting). GFS.

Europe/UK stayed with the route of short term(national models) and medium term(ECMWF).

There are pros and cons with both routes, GFS should be more accurate, however as it tries to model short term, smaller features more successfully its more liable to be temperamental in it's output and throw some widely varying outcomes.

ECM should be more stable, but can lose accuracy at the 72-120Z.(ECM is above all and everything a trend forecasting model)

The above isn't always the case, but is useful to remember as a rule of thumb.

Anyway the above is my thoughts and should always be used with experience, feel free to use or discard.

*Forgot to add this is why by and large NOAA will go with the ECM EPS for trends further out, (not at the moment as CC as just pointed out, but if not ECM EPS then GFS ENS).

[Geordiesnow]

Sorry if I sound harsh (have a headache!) but I really wish people would stop referring to snow potential being in F.I.

For starters the snow potential begins within the reliable timeframe via the E,ly. The next possible stage to the cold spell which remains uncertain is only just outside the reliable timeframe i.e next Wed when we might see a cold front moving S bringing the risk of more snow.

This cold spell needs to be viewed in two stages and the first stage i.e the E,ly is almost agreed upon. Its the 2nd stage that remains far more uncertain.

But for the UK hopes, 168 hours IS FI and whether you like or not, it is always possible that change, for better or worst. What i don't like seeing is people saying a cold snowy spell nailed on at 168 hours just because the models agree on it, because as you can see from this morning's output, it can change.

Regarding the end of the weekend easterly, it's an upgrade depending on what model you look at, GFS has lower heights hence more showery activity whilst the ecm does not. Hopefully this one will keep upgrading and surprise a few people.

[Guest North Sea Snow Convection]

This might help, it might not though.Feel free to correct if required.

As a general rule of thumb the METO forecasting will be(maybe) looking at the following.

0-72Z.

Meto first of all and GFS, Do either of the runs differ much from the Ensemble mean ?, if not then do they differ from each other, if not use the METO, if they do differ from each other, why ?. Generally they would go with the METO, with a dose of GFS if they though it was more realistic.

72Z to 144Z

Again check to see if the ops runs differ much from the ENS mean.

If not then do they differ from each other.

If they do they would take the start point from above and perform a best fit with the 3 models, looking also at the trend in ECM and the ECM EPS and GFS ENS.

144Z to 240Z

For obvious reasons this would be made up mainly from the ENS, EPS and super ENS unless the ops closely followed them, in which case they would also include the ops run.

240Z >

Beyond 240Z they would use the extended ECM that goes out 30 days, CFS and GFS ensembles (probably not the GFS ops out to 360Z.)

As well as judgement, teleconnections etc.

Many might recall years ago something called MRF, before GFS. This was the Medium range forecasting model from NWS.

As well as MRF, the US NWS(national weather service) had a short range model(can't remember its name tbh). GFS replaced this with as an all singing all dancing model that could do short, medium and long term forecasting). GFS.

Europe/UK stayed with the route of short term(national models) and medium term(ECMWF).

There are pros and cons with both routes, GFS should be more accurate, however as it tries to model short term, smaller features more successfully its more liable to be temperamental in it's output and throw some widely varying outcomes.

ECM should be more stable, but can lose accuracy at the 72-120Z.(ECM is above all and everything a trend forecasting model)

The above isn't always the case, but is useful to remember as a rule of thumb.

Anyway the above is my thoughts and should always be used with experience, feel free to use or discard.

*Forgot to add this is why by and large NOAA will go with the ECM EPS for trends further out, (not at the moment as CC as just pointed out, but if not ECM EPS then GFS ENS).

I'm delighted with the overall modelling this morning - I think that expectations have got far too high on the basis of seeing some exceptional charts lately. As others have said, people would have dropped a jaw seeing this output in the midst of the rain and gales a week ago and more.

It looks to me like a pretty lengthy cold spell is in the offing. Details will follow, but lets deal with the next four of five days first and not dwell on the further outlook all the time

[MR EXTREMES]

I had a feeling there would be a little consternation on here this morning after looking through the 00z model runs. UKMO is good, the ECM, is let's say, not the best this morning, however its 00z run has been different for the last 4 in a row, ECM 12z run has been much better and much more consistent. As for the GFS, it's just doing what it always does and toying with different evolutions, another one will be along shortly. The GFS parallel remains pretty solid, although I would say at lot of its runs are now indicating that this spell will only last for about a week before milder air moves in from the south.

perhapes you could give me some more detail on why this is only a 1 week event,

i thought there was a reload.

from what ive lernt and seen over the years is pressure being high can last weeks.

with this being a omega block and gl high building surely this inplace cold is more likely in our neck of the woods.

some people on here are saying gfs is brillant the only model which is the holy gral ecm being not so good.

so please a expert opion would be welcome.

[cooling climate]

GFS 06z looks rock solid in bring in the cold as early as Sunday and Monday also

a slight upgrade in potential for snow for east Anglia and the southeast.

I must admit I like where this run is going.

[Neilsouth]

GFS 06z looks rock solid in bring in the cold as early as Sunday and Monday also

a slight upgrade in potential for snow for east Anglia and the southeast.

I must admit I like where this run is going.

http://91.121.94.83/modeles/gfs/run/gfs-0-114.png?6

Agreed , sunday-monday snow risk in the east/south east

[Eugene]

What is shocking is the possible snow showers blowing in off the north sea for people near eastern coasts for monday night has been ignored

http://www.wetterzen...s/Rtavn1141.png

Strange people moan the charts are too far away for cold but when they show decent cold weather in a more reliable timeframe they suddenly look further ahead for something to go wrong.

Edited December 10, 2009 by Eugene

[Lewis]

http://91.121.94.83/modeles/gfs/run/gfs-0-114.png?6

Agreed , sunday-monday snow risk in the east/south east

I would not get your hopes up, it's certainly not nailed on.

Dont forget it's a much slacker flow so they will be a lot more modification off the North sea, so it's marginal.

Although the risk is there.

[Tom D]

GFS 06z looks rock solid in bring in the cold as early as Sunday and Monday also

a slight upgrade in potential for snow for east Anglia and the southeast.

I must admit I like where this run is going.

Indeed, although compared to the 00z things have been pushed slightly north-east, we have -10 uppers over East Anglia on Monday night into Tuesday, something that brings back good memories of February this year!

[Snowaddict]

There is no point in looking past T+144 in my opinion because we don't even have uniformity at T+120 - T+144 as shown on the 06Z GFS evolution...

It has upgraded the initial easterly significantly with potent cold 850's, of almost -12 into SE England at T+120. -8's across the rest of the British Isles..

That would easily lead to snowfall late next Monday & on Tuesday...

However, we don't have the same agreement from other models entirely..

Overall, prolonged cold looks like being highly possible, a lot more uncertainty regarding any potential snowfall.

If T+126 on this morning's GFS 06Z were to verify, it would be bitterly cold within 5 days time..

[TEITS]

perhapes you could give me some more detail on why this is only a 1 week event,

i thought there was a reload.

I think we are in danger of failing to appreciate what we have in the reliable timeframe rather than worrying about the medium/longer term.

Im so used to seeing fantastic charts in F.I and here we are with some fantastic charts between +72 & +144.

http://www.wetterzentrale.de/pics/Rtavn1201.png