Pattern January 2020 - Operation Thaw Alaska

[KyloG]

The OP GFS shouldn't run past day 10...all that does is give people the excuse that models flip/flop or have no use past day 10 and to come on here and preach about how models are useless.

The EPS and GEFS for that matter have use in the day 10+ with the large hemispheric pattern. As I posted above half of the EPS members have a -EPO. So if the EPS starts to show a -EPO pattern in the next couple of runs then we shouldn't be surprised. But, the same people that trash models will come in and say...'see the models flip flop'. No, the models haven't flip flopped, the EPS is hinting there could be a -EPO that develops after day 10. Most on here know how to use ensembles but some sure don't.

 

[SD]

That is the problem right there. I've mentioned this before. It makes no sense whatsoever for this, or any model for that matter, to be run past 10 days, because it can't even get the first 10 days right. It is useless and a complete fantasy, and we are better off throwing darts at a dart board to predict the weather at that point. I am a fan of the Euro but I admit that if they ran it past 10 days, it would probably be wrong a majority of the time also. I think they are smart for cutting it off at that point.

If no models ran past D10 I think a lot of people would wither away. It really feels like we as a region spend 75% of our time looking at d10+ in really all seasons. I'm not sure if that is a testament to uninteresting weather or a desire to see the next big thing.

Sent from my SM-G975U using Tapatalk

 

[Metwannabe]

If no models ran past D10 I think a lot of people would wither away. It really feels like we as a region spend 75% of our time looking at d10+ in really all seasons. I'm not sure if that is a testament to uninteresting weather or a desire to see the next big thing.

Sent from my SM-G975U using Tapatalk

Bingo... imho. And I don't think it's just weather either, way off topic I know but seems we (society as a whole) are always looking at that next big thing, to the point we often miss what's happening at the moment. On weather boards though it seems it's been this way for a while, I've witnessed disappointment in upcoming patterns while snow was still on the ground. Anywho...

 

[Broken024]

I get my best snows past 10 days. I'm all for it running 10+ days.

 

[Arcc]

The OP GFS shouldn't run past day 10...all that does is give people the excuse that models flip/flop or have no use past day 10 and to come on here and preach about how models are useless.

The EPS and GEFS for that matter have use in the day 10+ with the large hemispheric pattern. As I posted above half of the EPS members have a -EPO. So if the EPS starts to show a -EPO pattern in the next couple of runs then we shouldn't be surprised. But, the same people that trash models will come in and say...'see the models flip flop'. No, the models haven't flip flopped, the EPS is hinting there could be a -EPO that develops after day 10. Most on here know how to use ensembles but some sure don't.

Part of the problem is people make the "Mean' more useful than it is intended to be in the long range. The only time it's really useful is in the closer range to see if the members are on the same page. In the longer range, it just smooths everything.

 

[SD]

I like the idea of the ensembles running past 10 days though and the op stopping at D10, at least from a pattern recognition standpoint or an idea of the direction we may be headed.

EPS showing atleast half the members with a -EPO at end of it's run.

View attachment 29074

These 2 posts sort of go hand in hand with me. I agree about the ensembles running past D10 but as kylo shows when you break the members out from the mean you get some wildly different solutions. I'd love to be able to get more "cluster" data from the ensembles. Give me the ability to click on D11 and show me the most common 500mb patterns from the members and how many common members there are.

Sent from my SM-G975U using Tapatalk

 

[KyloG]

These 2 posts sort of go hand in hand with me. I agree about the ensembles running past D10 but as kylo shows when you break the members out from the mean you get some wildly different solutions. I'd love to be able to get more "cluster" data from the ensembles. Give me the ability to click on D11 and show me the most common 500mb patterns from the members and how many common members there are.

Sent from my SM-G975U using Tapatalk

Agreed, that would be far more useful. You have to infer when looking at the teleconnections. Based on the below...we are going to be in a
+AO/-PNA pattern the next 2 weeks, odds of that flipping are very low. But with a -EPO we could fluke/time something.

 

[Six Mile Wx]

These 2 posts sort of go hand in hand with me. I agree about the ensembles running past D10 but as kylo shows when you break the members out from the mean you get some wildly different solutions. I'd love to be able to get more "cluster" data from the ensembles. Give me the ability to click on D11 and show me the most common 500mb patterns from the members and how many common members there are.

Sent from my SM-G975U using Tapatalk

The clusters are exactly what I’ve been wishing for as well!

 

[superjames1992]

Part of the problem is people make the "Mean' more useful than it is intended to be in the long range. The only time it's really useful is in the closer range to see if the members are on the same page. In the longer range, it just smooths everything.

There is definitely truth to this. You can have a situation where in a 20-member ensemble model, 2 members are pumping out a 40” once in a millennium storm and the other 18 members are giving you zilch. The mean is thus 2”, but this is no realistic expectation, it’s oversmoothed. The median is probably more useful, and the median in this case is zip/zilch/nada.

Medians are really what we should use for our yearly snowfall “averages”, too, since the mean will be inherently right-skewed given how we tend to get snowfall around here. In the “typical” winter, we won’t reach the mean.

 

[Metwannabe]

Well fwiw the 12z GFS from a temp standpoint isn't horrible after the 1st, IF it were to verify I think we could live with that but what I really like to see and hope it is heading in the right direction, is that cold finally building in Canada. But that's past D10 and the GFS....

 

[Metwannabe]

Of course the GOA ridge is still there but we're somewhat saved by some blocking starting to show up.... and don't @ me, it's a weather discussion board.

 

[SD]

There is definitely truth to this. You can have a situation where in a 20-member ensemble model, 2 members are pumping out a 40” once in a millennium storm and the other 18 members are giving you zilch. The mean is thus 2”, but this is no realistic expectation, it’s oversmoothed. The median is probably more useful, and the median in this case is zip/zilch/nada.

Medians are really what we should use for our yearly snowfall “averages”, too, since the mean will be inherently right-skewed given how we tend to get snowfall around here. In the “typical” winter, we won’t reach the mean.

I'd actually be curious to do mean v median for RDU. I used to have all seasons snow total but that computer decided to no longer compute so I don't have it anymore.

Sent from my SM-G975U using Tapatalk

 

[SD]

Of course the GOA ridge is still there but we're somewhat saved by some blocking starting to show up.... and don't @ me, it's a weather discussion board.

View attachment 29078

Even has a flizzard chance at 264

Sent from my SM-G975U using Tapatalk

 

[Metwannabe]

Even has a flizzard chance at 264

Sent from my SM-G975U using Tapatalk

I mean I could live with what the GFS just vomited out, other than the next 3 days it's really not bad.... but but but

 

[GaWx]

There's a full-blown torch even per the 12Z GEFS at hour 180 (as well as surrounding hours):



Remember when the GEFS had normal and even cold for early January?

 

[Metwannabe]

There's a full-blown torch even per the 12Z GEFS at hour 180 (as well as surrounding hours):

View attachment 29079

Remember when the GEFS had normal and even cold for early January?

Yuck

 

[superjames1992]

I'd actually be curious to do mean v median for RDU. I used to have all seasons snow total but that computer decided to no longer compute so I don't have it anymore.

Sent from my SM-G975U using Tapatalk

That would be interesting. I’ve never found the data, though. I recall reading one time that Portland, OR, for example, averages 3” of snow per year but has a median of 0”, for example (I may be wrong on that, though). RDU’s obviously isn’t 0”, but I wouldn’t be surprised if it’s something like 4-5” as opposed to the mean of 6-7” (or whatever it is these days).

 

[GaWx]

Lol at how much warmer the 12Z GEFS is vs the 0Z run in the 11-15 day:

 

[GaWx]

Here's the change map for the entire 11-15 12Z GEFS averaged out vs the 0Z GEFS: this warming brings the 12Z GEFS 11-15 to N in the SE vs BN in most of the recent runs. #GEFSSucksButJBWillRideItTilTheEndOfTime

 

[Metwannabe]

Couple of cold shots on the GEFS and that's about it..... and those are long range, which I'm sure will moderate as we get closer