Model Output Discussion - cold spell to end Autumn

[northwestsnow]

6 minutes ago, chionomaniac said:

Looks like the GFS 06Z puts us in the freezer from T+168 until the end of the run.

Remember that wave 1 strat disturbances are quite often followed by wave 2 splits lower down, so the chances of something popping up cold wise second third of December will certainly be increased this year. Therefore I would not get too despondent at the volatility of each run.

In chino we trust

Truth is the ex has just phoned (gfs6z) to say it might not be over she'll ring you back at 5pm..(but really you know shes been dating EC behind your back)...

Joking aside6z is a fabulous run- guess we just to to hope the latest data (which 6z uses), is a flip of the coin in coldies favour..

[karlos1983]

3 minutes ago, Seasonality said:

I deserve a double slap on the wrist for posting this, being 384h and a GFS snow accumulation chart, but just for fun. A snow covered UK.

someone buy this member a beer!!  definitely going to happen 

That was fun to watch nonetheless, suggest perhaps adjusting expectations for the 12z however LOL

Edited November 28, 2017 by karlos1983

[Met4Cast]

Yesterday and last night the wheels on our cold train were in severe dis-repair and on the verge of falling off. The train passengers had turned bi-polar and were in need of some help.

Today, the 6z has called in the mechanic and we've started to repair some of the damage, but we're certainly not out of the woods yet. Hopefully the 6z with it's latest data is onto something and we see the ECM begin to pick up that same signal now too. 

Just a reminder, the ensembles will be absolutely useless at trying to work out what's happening with the split flow and where the energy is going, if the higher resolution OP can't do it, you can forget the ensembles.

My advice would be to pretty much bin the entire ensemble suite until we get a clearer picture.

[bobbydog]

last 3 runs of the GFS in the latter stages-

all roads lead to....

[Tom Jarvis]

I think the models today will be getting and actual sample of how strong the block is and they may have underestimated how strong it is so we could exspect some wild swings in the models in short timeframes 

[johnholmes]

What everyone should realise is that GFS in winter is little different to how it runs in summer. Thatis is run to run comparison at  extended time frames is really nothing more than an amusing exercise and not, usually, to be taken seriously.

Many years ago I used to try to get folk to compare like with like. So 00z yesterday and today, etc. It is less liable to huge swings than watching each run-honest.

But I do not want to upset any of you, if watching each page come out is what you want, fine, just try not to get to upset when your longed for weather disappears on the next run.

Cold with sleet and snow in places with a coastal aspect for the next 48 hours then becoming less mild into the weekend. Beyond that and who knows. I did post the anomaly charts which can be a good guide, used carefully, and they are not all showing the same. If the NOAA output this evening goes along the other two from this morning then you really have to accept cold not coming back in a hurry. If NOAA keeps its meridional pattern then there is a case to be made for a return.

[karlos1983]

Just now, bobbydog said:

last 3 runs of the GFS in the latter stages-

all roads lead to....

The Pub ...

 

[Ali1977]

This has to be the best run so far, for 850s and snowfall. This if it happened would be a very noteworthy week or 2 of snowplough and grit requirement ?

[feb1991blizzard]

7 minutes ago, Daniel Smith said:

Yesterday and last night the wheels on our cold train were in severe dis-repair and on the verge of falling off. The train passengers had turned bi-polar and were in need of some help.

Today, the 6z has called in the mechanic and we've started to repair some of the damage, but we're certainly not out of the woods yet. Hopefully the 6z with it's latest data is onto something and we see the ECM begin to pick up that same signal now too. 

Just a reminder, the ensembles will be absolutely useless at trying to work out what's happening with the split flow and where the energy is going, if the higher resolution OP can't do it, you can forget the ensembles.

My advice would be to pretty much bin the entire ensemble suite until we get a clearer picture.

Disagree with this, yes they are lower resolution but the software is programmed to account for these situations, that's why you have 72 members over 2 models, its a probability forecast which also is advisable with human interpretation in the form of qualified meteorologists, its a percentage forecast - a guide.

[bluearmy]

So we are all aware that the gfs post days 7/8 is just pretty patterns that our smal kids could draw if they had some crayons

HOWEVER, the evolution days 9-12 are not out of sync with the ecm mean anomoly from the 00z suite. It’s clearly heading towards a high risk/high reward scenario but we are still in the raffle, despite the big old bird standing on stage having nicked nearly all the tickets !!

[claret047]

Good Morning everyone,

As a lot less knowledgeable on here than a number of other posters I find it very puzzling how the models can be so very different from one run to the next.

I would like to ask a question as to how these model outputs are produced. Mods if I have posted this in the incorrect thread then please move it to a more appropriate one.

Does for example the 0Z one then have computer feedback from readings onto the previous 0z one to perpetuate any bias, or is each run completely independent  of the previous OZ  run. I am afraid that I have not worded it very well, but I trust you know what I mean.

Kind regards

Dave

[snowbob]

2 minutes ago, bluearmy said:

So we are all aware that the gfs post days 7/8 is just pretty patterns that our smal kids could draw if they had some crayons

HOWEVER, the evolution days 9-12 are not out of sync with the ecm mean anomoly from the 00z suite. It’s clearly heading towards a high risk/high reward scenario but we are still in the raffle, despite the big old bird standing on stage having nicked nearly all the tickets !!

Love it

At least we have a ticket this year

[Met4Cast]

5 minutes ago, feb1991blizzard said:

Disagree with this, yes they are lower resolution but the software is programmed to account for these situations, that's why you have 72 members over 2 models, its a probability forecast which also is advisable with human interpretation in the form of qualified meteorologists, its a percentage forecast - a guide.

When an entire ensemble suite can flip backwards and forwards as much as the ops do, what use are they in determining a probability forecast when the next run they show something completely different? 

[A.J]

12 minutes ago, karlos1983 said:

someone buy this member a beer!!  definitely going to happen 

That was fun to watch nonetheless, suggest perhaps adjusting expectations for the 12z however LOL

Nah, forget what I posted earlier about FI being a waste of space et al......FI is a sure bet, it will happen.....in fact I'm so excited I need to change my underwearand I 'm starting to stutter....b-b-b-b-b-b-b-b-bank 

mouthwatering charts in FI for sure but I'll believe it if it's white stuff falling out of the sky at T0!

[Methuselah]

3 minutes ago, feb1991blizzard said:

Disagree with this, yes they are lower resolution but the software is programmed to account for these situations, that's why you have 72 members over 2 models, its a probability forecast which also is advisable with human interpretation in the form of qualified meteorologists, its a percentage forecast - a guide.

Therein lies one of weather forecasters' (and models') greatest bugbears: probability might make perfect sense to those of us who've been fortunate enough to study it; but, to those who haven't studied it, it's a bloody nightmare...

One reason for that is, I guess, that the probability of an event vanishes the moment said 'event' occurs - It cannot, for instance, '19% snow' on Christmas Day; it can only snow or not snow? 

So no wonder peeps get confused...?

[feb1991blizzard]

1 minute ago, Daniel Smith said:

When an entire ensemble suite can flip backwards and forwards as much as the ops do, what use are they in determining a probability forecast when the next run they show something completely different? 

When has an entire suite flipped during the last week? it does happen but rarely IMO, the spread of solutions has remained pretty similar just changes in a number of ensemble members to tip the balance slightly in favour of one or the other, FWIW my method of reading the GEFS has been pretty decent recently wrt knowing when to call time on a potential cold spell we are getting near the critical point now, they are just coming into the 192 range, if they get worse once they are inside 192 range (even slightly over the next 24 hours) then its game over - they wont flip back again - you can say its a crude method but ive called time on many a potential cold spell since the glory years and been ridiculed for calling it early on the back of 2 or 3 suites but ive been right evey time, so we need them to improve, if we don't have a really good number (well over half) flatlining at around -8c or below in a couple of days time, then its over.

[Mike Poole]

Who needs model output?

Dr. Cohen’s latest AO blog is hilarious.  In it he presents the winter forecasts from various models (NMME, CFS, ECMWF and UKMO), which are all screaming mild – junks the lot and in his AER forecast goes for cold right across Eurasia, to quote: ‘multiple weeks of severe winter weather across the NH’.

It’s quite a call – either he’s wrong or the models are.

Well worth a read after the rather mixed model output this morning:

http://www.aer.com/science-research/climate-weather/arctic-oscillation 

 

[feb1991blizzard]

5 minutes ago, Ed Stone said:

Therein lies one of weather forecasters' (and models') greatest bugbears: probability might make perfect sense to those of us who've been fortunate enough to study it; but, to those who haven't studied it, it's a bloody nightmare...

One reason for that is, I guess, that the probability of an event vanishes the moment said 'event' occurs - It cannot, for instance, '19% snow' on Christmas Day; it can only snow or not snow? 

So no wonder peeps get confused...?

But that's the only method!!  I am not one of those politically correct people who just agree with the Met Office top forecasters just because they are university educated to a high standard but I would be happy to have anything beyond the week ahead forecast in probability and even short range severe weather events like snow / rain marginal's in certain circumstances.

[Frosty.]

Definitely the best run of the winter..errr..I mean autumn so far!!

Here is Frosty's highlights of the Gfs 6z..there's plenty of them but a snow shovel isn't included!!..BOOM

[Weathizard]

Control run looks to have followed the OP in pattern up to day 6/7, after that it goes on a different tangent but good to see it doesn't blow that's low pressure up like the ECM.

Looking at the ENS it looks like a real lottery, there's a lot of options on the table and you can see why the models are flip flopping at the moment. Renders the 'mean' charts pretty useless to be honest.

[Summer Sun]

A quick look at the short ens for the 06z shows the Op going a fair bit below the mean around the 6th

We need to await the full ens to see if it is a rouge run or not

Edited November 28, 2017 by Summer Sun

[johnholmes]

10 minutes ago, feb1991blizzard said:

When has an entire suite flipped during the last week? it does happen but rarely IMO, the spread of solutions has remained pretty similar just changes in a number of ensemble members to tip the balance slightly in favour of one or the other, FWIW my method of reading the GEFS has been pretty decent recently wrt knowing when to call time on a potential cold spell we are getting near the critical point now, they are just coming into the 192 range, if they get worse once they are inside 192 range (even slightly over the next 24 hours) then its game over - they wont flip back again - you can say its a crude method but ive called time on many a potential cold spell since the glory years and been ridiculed for calling it early on the back of 2 or 3 suites but ive been right evey time, so we need them to improve, if we don't have a really good number (well over half) flatlining at around -8c or below in a couple of days time, then its over.

An interesting observation that, I might even be tempted to add that idea into my own look at different models, ensembles etc, thank you.

Is there sometimes a difference looking at ensembles for different places in the UK or am I talking rubbish?

[ukpaul]

1 minute ago, Summer Sun said:

A quick look at the short ens for the 06z shows the Op going a fair bit below the mean around the 6th

We need to await the full ens to see if it is a rouge run or not

It's at that point that the ensembles become useless as a predictor; in any look at probabilities weight must be given to the ops higher resolution. It's the one that is focused on for a very good reason, akin to experts who know what they are talking about (the op) rather than someone with limited knowledge posting on here (each of the ensembles).

Anyway....a red run?

[Methuselah]

3 minutes ago, feb1991blizzard said:

But that's the only method!!  I am not one of those politically correct people who just agree with the Met Office top forecasters just because they are university educated to a high standard but I would be happy to have anything beyond the week ahead forecast in probability and even short range severe weather events like snow / rain marginal's in certain circumstances.

Me too; but we both understand at least the rudiments of probability; the average Joe who takes on faith that whatever gets printed on the front-pages of a couple of nameless newspapers, probably does not...

BTW, I do not aim my criticism at the 'average Joe', for being ignorant, but at an education-system that lets so many through, without them ever learning something that is becoming ever more important - probability theory.

That said, the model-outputs hardly look probabilistic, do they?

[bluearmy]

5 minutes ago, Weathizard said:

Control run looks to have followed the OP in pattern up to day 6/7, after that it goes on a different tangent but good to see it doesn't blow that's low pressure up like the ECM.

Looking at the ENS it looks like a real lottery, there's a lot of options on the table and you can see why the models are flip flopping at the moment. Renders the 'mean' charts pretty useless to be honest.

We had this debate the other day - how is the ens mean any different to having a solution verify that is a blend between two differing op solutions - this seems to happen a lot of the time. 

I’m not someone who think means are a waste of time.