December 2017 Mid-Long Range Disco

[Cobalt]

3 minutes ago, eurojosh said:

That reminds me. How was the Siberian snow cover in October?

I think it was still above average, but not as good as recent years. 

[cae]

I apologize in advance for the length of this post.  I'm going to defend the GGEM, so yeah, I suspect I'm going to get trolled pretty hard for this one.

We often hear the GGEM referred to as a JV model.  In the past I've argued that it actually does pretty well in the winter, based on years of following verification scores and looking back after the events to see how different models performed.  This recent storm is a good example.

To start with:  On average, the GGEM is clearly the #4 global model.  The Euro, UKMET, and GFS are better, in that order.  There's lots of verification data to back that up. but maybe the simplest thing to look at is the long term day-5 H5 verification scores, which can be found here. 

But there is also a strong seasonal trend to the peformance of the GGEM, at least relative to the GFS.  We can see this by looking at the 3-month average of the day-5 northern hemisphere H5 scores over time.  In the below image, the GGEM's performance relative to the GFS is the blue line.  (The Euro is red and the NAVGEM is orange). 

Since 2010 the 3-month trailing average of the GGEM's peformance relative to the GFS has peaked around March of each year and bottomed out around August.  The GGEM is a relatively week summer model, especially for tropical systems.  But in the winter, it regularly matches or exceeds the GFS.

The common response to this point is that day-5 northern hemisphere H5 verification scores aren't everything.  That's true.  But you see something similar if you click around on different verification metrics.  For example, the below figure shows the North American sea-level-pressure verification scores for different models relative to the GFS for the last month.  Red means the GFS was better, and green means the other model was better.

It's about what we'd expect.  The Euro is almost always better than the GFS.  The UKMET is usually better.  The JMA and the NAVGEM (FNO) are usually worse.  But the GGEM and GFS are closely matched.  If you look at other metrics, you'll see similar results.

At this point, the response is usually that these verification scores aren't everything, and the GGEM doesn't actually do well for winter storms in our region.  (I'm presenting both sides of the conversation here, to move things along.)  That's when I mention that after big events, when I look back to see how the different models performed, I often find that the GGEM did surprisingly well.  That brings us to this most recent event.  Below are the GGEM (left) and GFS (right) forecasts for 18z on Saturday, starting from 126 hours out. 

Starting about 5 days out, the GGEM picked up the event when the GFS had effectively nothing.  At 90 hours out, the GGEM showed something that looked a lot like what we actually got.  The GFS (and Euro) had no snow for our region at that time.  The GFS started showing the event once we got within 60 hours.

You might remember that the UKMET also picked up on this event well ahead of the GFS (and, I believe, the Euro).  As evidence that the verification scores actually do reflect actual things happening in our backyard, including snow events, here are the 120-hour North American sea-level pressure verification scores for last Saturday.  The GGEM posted the highest score, followed very closely by the UKMET, then the Euro, then the GFS.

Of course that isn't a common occurence.  On any given day, the Euro usually posts the best score.  (See the red and green plots above.)  But between the GFS and GGEM at this time of year, it's generally a toss-up.

This brings me to the last counter-argument I've heard:  It's not that the GGEM is any good, it's that the GFS is pretty bad.  This is somewhat a matter of semantics, but I disagree.  The GFS is a very good weather model.  All models have their good days and bad days, good events and bad events, strengths and weaknesses.  There will surely be future events when the GFS shines and the GGEM stinks.  Even in this event, by the end, the GFS did a good job predicting the distribution of snowfall, correctly foreseeing the Westminster deathband.  (They should just hard-code that into the other models.)

OK, I've said my piece.  I'm not sure if this changed any minds, but I hope at least it was helpful for those of you who are starting out.  If you've read this far, thanks. 

 

[Cobalt]

@cae Great post! It seems like trusting the Canadian on the Wednesday before the storm was the right idea, but may not be the next time around

[cae]

2 minutes ago, Cobalt said:

@cae Great post! It seems like trusting the Canadian on the Wednesday before the storm was the right idea, but may not be the next time around

Thanks.  It's probably not a good idea to trust any one model 3-4 days before a storm.  I find it's best to treat the ops as their own ensemble, with the Euro of course having the highest weight.

[frd]

1 hour ago, StormchaserChuck said:

The disappointing thing about this recent blocking pattern is that there is no Stratospheric warming

What is very interesting is that despite the favorable QBO, we have not had a major SSW event I believe for four winters  in a row, I believe the average is once every other year. 

  

[poolz1]

Dont know about a major SSW event but current obs show a weak and displaced vortex with warming at least to 10mb....GEFS continue the weakening and displacement through the end of the period.

I dont claim to fully understand this stuff but I dont think the strat is giving us any issues and is not forecasted to do so either.

[das]

3 hours ago, leesburg 04 said:

Possible to add a few more years to the blend?

This made me LOL.  Get out of my head.

[AfewUniversesBelowNormal]

1 hour ago, frd said:

What is very interesting is that despite the favorable QBO, we have not had a major SSW event I believe for four winters  in a row, I believe the average is once every other year. 

  

They seem to happen in bunches, at the same time of year for many years in a row. This is more true in October-November-December, and mixes up more January-February-March

[Amped]

591DM on the Alaska coast. Won't verify but it's insane.

 

 

[AfewUniversesBelowNormal]

I've seen it 600dm
More likely this verifies -PNA, La Nina is really strengthening 

[showmethesnow]

@cae If J.B. 2.0 swears by it during the winter it can't be all that bad. Right? Nice post. 

With weatherbell output marginal at best with the CMC I have found I have fallen to the wayside on using the CMC for anything more then verification for the other models. After reading your post I may have to give more weight to the CMC in the future. The one thing I have noticed over the years though, is that quite often a model may score higher overall during a given period of time and yet a lower verifying model would actually perform much better through our general region with the setup. Now this was due to the broad brush of the verification to that of the N hemi/N American (maybe a little smaller scale?) and not a more regional look. Now it has been some years since I really looked into the hows and whys of the verification process so maybe things have changed. So I guess I am asking, if you know, have they narrowed verifications to a smaller scale with a more regional look to give a better representation for our neck of the woods?

Again, nice write up.

 

 

 

[AfewUniversesBelowNormal]

Nice Stratospheric warming happening now

 

 

[andyhb]

4 hours ago, StormchaserChuck said:

I've seen it 600dm
More likely this verifies -PNA, La Nina is really strengthening 

600 dm in Alaska? 500 mb height records in those parts aren't even close to that.

[showmethesnow]

After looking over the weeklies all I can say is lets enjoy the next 3+ weeks or so (if they are correct) and not sweat the details that come afterwards. I will take what looks to be a below average period of temps, both on the model projections and the general overall look at 500's through that period. The dry anomalies are not extreme. And when I see the heaviest of the precip, excepting the Pacific NW, confined to the east coast (especially through the gulf states) that leads me to believe there is a decent probability that we are seeing the southern jet coming into play. Add in what looks to be a fairly active period of time and the chances are good we will see multiple opportunities to score snow/ice in the coming 3 weeks or so. 

And who knows, half way through this period we may look up and see that the extended on the weeklies has once again changed and all that worry was for nothing.

eta: Also wanted to mention that with the pv located throughout the extended around the Hudson Bay region I find the temp anomalies we are seeing highly suspect. Possible yes, but probable no. At least IMO.

[AfewUniversesBelowNormal]

7 minutes ago, andyhb said:

600 dm in Alaska? 500 mb height records in those parts aren't even close to that.

Models have a tendency to go nuts over -EPO

[BristowWx]

49 minutes ago, showmethesnow said:

After looking over the weeklies all I can say is lets enjoy the next 3+ weeks or so (if they are correct) and not sweat the details that come afterwards. I will take what looks to be a below average period of temps, both on the model projections and the general overall look at 500's through that period. The dry anomalies are not extreme. And when I see the heaviest of the precip, excepting the Pacific NW, confined to the east coast (especially through the gulf states) that leads me to believe there is a decent probability that we are seeing the southern jet coming into play. Add in what looks to be a fairly active period of time and the chances are good we will see multiple opportunities to score snow/ice in the coming 3 weeks or so. 

And who knows, half way through this period we may look up and see that the extended on the weeklies has once again changed and all that worry was for nothing.

eta: Also wanted to mention that with the pv located throughout the extended around the Hudson Bay region I find the temp anomalies we are seeing highly suspect. Possible yes, but probable no. At least IMO.

I noticed the GFS op runs keep switching from torch Xmas to cold Xmas.  GEFS remains somewhat cold but nothing silly.  All about where that front sets up I suppose. 

[showmethesnow]

13 minutes ago, BristowWx said:

I noticed the GFS op runs keep switching from torch Xmas to cold Xmas.  GEFS remains somewhat cold but nothing silly.  All about where that front sets up I suppose. 

Been seeing the back and forth with the snow/ice/torch ect... but really have no thoughts at this time on what to expect. Been mostly focusing on the coming week which looks to be somewhat active. I am half expecting to see a somewhat significant storm blow up somewhere in the eastern CONUS but that is more a gut feeling then anything else. And if that is the case that would probably change the dynamics of what to expect heading into Christmas anyway. 

[WxUSAF]

EPS basically caved to the GEFS as far as Xmas goes with the trough staying in the southwest and the SE ridge/WAR backing up into our area.  It's been bouncing around, so I certainly wouldn't say this is locked in, but definitely not the look I wanted to see.  

[CAPE]

13 minutes ago, WxUSAF said:

EPS basically caved to the GEFS as far as Xmas goes with the trough staying in the southwest and the SE ridge/WAR backing up into our area.  It's been bouncing around, so I certainly wouldn't say this is locked in, but definitely not the look I wanted to see.  

Yup. Its really too bad we still cant get a legit -NAO. For all the "advertising" in the guidance, the most that ever verifies is a transient ridge near GL.

[BristowWx]

4 minutes ago, C.A.P.E. said:

Yup. Its really too bad we still cant get a legit -NAO. For all the "advertising" in the guidance, the most that ever verifies is a transient ridge near GL.

Maybe it will rain or be cloudy.  Or some sort of weak wedge.  Full sun and 65 would be just..

[aldie 22]

38 minutes ago, WxUSAF said:

EPS basically caved to the GEFS as far as Xmas goes with the trough staying in the southwest and the SE ridge/WAR backing up into our area.  It's been bouncing around, so I certainly wouldn't say this is locked in, but definitely not the look I wanted to see.  

Good post

[CAPE]

Nice -EPO. Pretty crappy Atlantic. Mild look and any storm would likely cut. If this does verify, lets hope its transient.

I have gotten used to a mild Christmas. Really doesn't bother me. 

[psuhoffman]

22 minutes ago, C.A.P.E. said:

Nice -EPO. Pretty crappy Atlantic. Mild look and any storm would likely cut. If this does verify, lets hope its transient.

I have gotten used to a mild Christmas. Really doesn't bother me. 

We need that epo ridge to push a bit east. Everything has been trending west on the last 24-36 hours of runs and that allows the whole pattern to retrograde placing us on the wrong side of the boundary. 

[CAPE]

Just now, psuhoffman said:

We need that epo ridge to push a bit east. Everything has been trending west on the last 24-36 hours of runs and that allows the whole pattern to retrograde placing us on the wrong side of the boundary. 

Yeah and we saw this sort of a retro show up in the guidance a couple weeks back and it only lasted a few runs, and never really materialized- I think it was showing up on the GEFS more than the EPS.

If it does happen, hopefully it's temporary. It would really be nice to get a little more help in the NA too.

[BristowWx]

12 minutes ago, C.A.P.E. said:

Yeah and we saw this sort of a retro show up in the guidance a couple weeks back and it only lasted a few runs, and never really materialized- I think it was showing up on the GEFS more than the EPS.

If it does happen, hopefully it's temporary. It would really be nice to get a little more help in the NA too.

My optimistic side could see a scenario where a back door cold front keeps us seasonal and damp.  But the warmth would lurk just south with a large temp gradient north to south until the front pushed through.  Might be just dreaming that up

[Cobalt]

I'm just hoping by that that time near the week before Christmas, we have identified when the pattern is gonna go back into cold. Trends aren't in our favor, but there's still time for them to move back to favorable for Christmas.

[BristowWx]

3 minutes ago, Cobalt said:

I'm just hoping by that that time near the week before Christmas, we have identified when the pattern is gonna go back into cold. Trends aren't in our favor, but there's still time for them to move back to favorable for Christmas.

I’m sure it will look different.  Maybe not better different but it’s not like a massive ridge in the eastern half of the US..just over us. 

[psuhoffman]

9 minutes ago, BristowWx said:

My optimistic side could see a scenario where a back door cold front keeps us seasonal and damp.  But the warmth would lurk just south with a large temp gradient north to south until the front pushed through.  Might be just dreaming that up

I guess... yes I would rather it be cold then warm on christmas day.  However, its not that big a deal to me that I am going to spend much time or lose any sleep on that.  If we have a shot at snow on the ground for christmas now that is something I will spend some mental energy on, but just simply worrying about if it is 45 or 55 on christmas day...nah.  Just praying for chilly air seems such a low bar.  Give me a white christmas or honestly I dont care much what the weather is.   

[Cobalt]

Just now, BristowWx said:

I’m sure it will look different.  Maybe not better different but it’s not like a massive ridge in the eastern half of the US..just over us. 

Well, the potential storm during that time would ride the ridge wouldn't it? It looks like once the storm leaves, the cold sinks it behind it, at least from the OP and the ensembles

[PhineasC]

Please god no warm Christmas.