January Med/Long Range Disco Part 2

[Ji]

next thing to watch is how much of our little snow event dries up. Ukmet and GGEM has nothing. Euro is coming in drier. We lost the ICON...nam had nothing. My guess is that we dont even get an inch of snow this week

[Ji]

12z NAVGEM looks the best lol

[Bob Chill]

9 minutes ago, Ji said:

next thing to watch is how much of our little snow event dries up. Ukmet and GGEM has nothing. Euro is coming in drier. We lost the ICON...nam had nothing. My guess is that we dont even get an inch of snow this week

It's all playing out perfectly with the front. First it looked like nothing. Then 1-3. Then backed off from there. Now it's going to come back strong before going completely away. 

[CAPE]

1 minute ago, WxWatcher007 said:

I called for a reshuffle too. Have to gamble if we want to win around here. The UL pattern was excellent for cold and bad for coastals.

ETA: in mby 

It certainly is not a shut out look on the ensembles, especially since the NA looks pretty decent. It is conceivable things could briefly align at some point over the next couple weeks where we can get something to track underneath, especially if the advertised look up top is real. But overall the Pac look is crap from the end of this week forward on the ensembles, so it appears there will be a major change in the pattern. I would have rather not gone that route and kept an overall cold look in the east, so we will have to see how things progress going forward. Hopefully by the very end of Jan or early Feb we can get the Pac to flip back to a more favorable look.

[Ji]

2 minutes ago, Bob Chill said:

It's all playing out perfectly with the front. First it looked like nothing. Then 1-3. Then backed off from there. Now it's going to come back strong before going completely away. 

hopefully it goes away Tuesday at 00z after giving us 2-4 Monday at 18z for maximum affect

[Jandurin]

It's all playing out perfectly with the front. First it looked like nothing. Then 1-3. Then backed off from there. Now it's going to come back strong before going completely away. 

My body is ready

[Cobalt]

2 minutes ago, losetoa6 said:

H5 goes negative tilt on the normally progressive  Navgem   . It's actually done decent this year from what ive seen. The Blizzard it was on the west side of guidance in the medium range and the rest followed . And thats just 1 off the top of my head 

It did lead the way during the Dec 8-9 event IIRC. Still, it's just the Navgem, but I'd say we have about 24 hours of guidance to see what the models do with the costal part before giving up. 

[showmethesnow]

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

It certainly is not a shut out look on the ensembles, especially since the NA looks pretty decent. It is conceivable things could briefly align at some point over the next couple weeks where we can get something to track underneath, especially if the advertised look up top is real. But overall the Pac look is crap from the end of this week forward on the ensembles, so it appears there will be a major change in the pattern. I would have rather not gone that route and kept an overall cold look in the east, so we will have to see how things progress going forward. Hopefully by the very end of Jan or early Feb we can get the Pac to flip back to a more favorable look.

It is what it is. As you, I don't think it is a shut out pattern whatsoever so we will probably will have other fails to track in the coming weeks. I am still not sold that the PAC is going to be bad as the models are now advertising anyway.  

[Bob Chill]

17 minutes ago, Ji said:

hopefully it goes away Tuesday at 00z after giving us 2-4 Monday at 18z for maximum affect

Joking aside... the majority of eps memebers are in the 1-3" range for the front. Over 30. Some 2-4's mixed in. Mean precip is a good bit more than the op and more widespread as well so the euro ens run supports much better outcome than the op. I know you probably don't give a crap but my gut isn't wavering from a 1-3" event next week. 

[mitchnick]

Nam'ed. More the panel before at 81 hrs

https://www.tropicaltidbits.com/analysis/models/?model=namconus&region=neus&pkg=ref_frzn&runtime=2018011318&fh=84

[MD Snow]

Just now, mitchnick said:

Nam'ed. More the panel before at 81 hrs

https://www.tropicaltidbits.com/analysis/models/?model=namconus&region=neus&pkg=ref_frzn&runtime=2018011318&fh=84

Yeah, I'm hugging this look until 0z when we get un-Nam'ed. 

All joking aside, I'd be totally fine with this depiction from the NAM. 

[poolz1]

I thought we were about to get reverse nam'd but ended up pretty good... 

[Bob Chill]

Just now, mitchnick said:

Nam'ed. More the panel before at 81 hrs

https://www.tropicaltidbits.com/analysis/models/?model=namconus&region=neus&pkg=ref_frzn&runtime=2018011318&fh=84

I was just getting ready to post that the nam would be 2-4" if it went 90 hours. 

[mitchnick]

1 minute ago, Bob Chill said:

I was just getting ready to post that the nam would be 2-4" if it went 90 hours. 

Yep. Look at those 250 mb winds about to move over us.

https://www.tropicaltidbits.com/analysis/models/?model=namconus&region=neus&pkg=uv250&runtime=2018011318&fh=84

[AfewUniversesBelowNormal]

Do you think models are making too much of that Atlantic SSTs are warm? The 18z NAM pretty much stalls a front and 500mb vort completely to monster Atlantic ridge. This seems suspicious. 

[Bob Chill]

2 minutes ago, mitchnick said:

Yep. Look at those 250 mb winds about to move over us.

https://www.tropicaltidbits.com/analysis/models/?model=namconus&region=neus&pkg=uv250&runtime=2018011318&fh=84

I said this a couple days ago. Southwest flow works ok east of the mountains. If there is decent moisture and lift embedded in the jet support then we won't have to worry nearly as much about the apps like we do with clippers. West or northwest flow sucks. Especially for Ji's yard. SW flow is much better for us piedmont folks

[Ralph Wiggum]

Not a thaw. Legit pattern change. Really bad looking Pac is going to run the show for a while. This should make Mitch happy. The "reshuffle". See where we end up sometime in early(or mid) Feb.

Crud pattern change wants to lock in on ens for weeks. Oh well, everyone was saying we were DOA till sometime in Feb anyway.

[cae]

Yeah, tough 12z suite.  It's too bad to see the changes in the Euro / EPS, and interesting discussion about the Euro's performance.  Fwiw, here are my thoughts:

1)  It hasn't snowed yet.  We have 3 1/2 days before the GFS says the first flakes fall on the cities.  I think sometimes we jump the gun with assessing model performance.

2)  After storms, we tend to remember model performance based on their last few forecasts, not the ones that occur 3-6 days out.  I've read several posts here that suggest the GFS did better than the Euro on the January 4th storm.  Really?  I put together a gif after that storm comparing the performance of the Euro, GFS, GGEM, and ICON.  I'll re-post it here.  It ends with the 00z forecast just before snow fell, and starts 96 hours before that.  Top left is Euro, top right is GFS, bottom left is GGEM, and bottom right is ICON. 

The Euro messed up its final forecast, but it was clearly better than the GFS at 96, 84, 72, 60, and 48 hours out.  It might have been better at 36, 24, and 12 hours out as well - I'd have to see verification data to know for sure.  It was around 60 hours out that Bob Chill had this prescient post about the first GFS run to get a little precip across the bay:

Quote

Got precip west of the bay. Looks like a lower qpf version of the euro. I'm telling you, the gfs is going to look like the smartest model when this is all said and done even though it was too far east for like 50 runs. 

Good call. 

3)  All of the models are getting better, and they might have closed the gap with the Euro (but not by much, based on verification scores).  At the very least, we're taking 4 and 5 day forecasts more seriously than we used to.  But the models have always been a lot closer to each other in performance than we sometimes seem to believe.  The difference in skill between the Euro and NAVGEM is about 1 day.  That is, today's NAVGEM is about as good as yesterday's Euro, with the other models somewhere in between.  It's possible that recent changes have made the Euro worse for storms in our area, but the Euro has always had its bad events.  People tend to overlook them when its reputation is good and overemphasize them when its reputation is bad.  It's confirmation bias.  We all do it, and it's really hard to avoid.  And yeah, there's recency bias as well.  But when you look at things objectively and in the long run, the Euro is still the best model by good measure. 

fwiw, the Canadian ensemble still doesn't like this storm, but it finally has one big hit in there which had been lacking earlier.

 

[Ralph Wiggum]

H5 goes negative tilt on the normally progressive  Navgem   . It's actually done decent this year from what ive seen. The Blizzard it was on the west side of guidance in the medium range and the rest followed . And thats just 1 off the top of my head 

It is a function of it's progressive bias and phasing with the follow-up sw. In a land all by itself, no red flags.

[CAPE]

12 minutes ago, showmethesnow said:

It is what it is. As you, I don't think it is a shut out pattern whatsoever so we will probably will have other fails to track in the coming weeks. I am still not sold that the PAC is going to be bad as the models are now advertising anyway.  

Yeah I have no idea whether the advertised PAC look is going to be persistent. Fits boilerplate Nina I suppose, but you and I had a lot of discussion about this early on and we did end up flipping to a pretty amazing and extended -EPO period. Hopefully it is not that bad and temporary. Something will surely save us by Feb though. SSW and -AO and stuff. lol.

[Ralph Wiggum]

Yeah, tough 12z suite.  It's too bad to see the changes in the Euro / EPS, and interesting discussion about the Euro's performance.  Fwiw, here are my thoughts:
1)  It hasn't snowed yet.  We have 3 1/2 days before the GFS says the first flakes fall on the cities.  I think sometimes we jump the gun with assessing model performance.
2)  After storms, we tend to remember model performance based on their last few forecasts, not the ones that occur 3-6 days out.  I've read several posts here that suggest the GFS did better than the Euro on the January 4th storm.  Really?  I put together a gif after that storm comparing the performance of the Euro, GFS, GGEM, and ICON.  I'll re-post it here.  It ends with the 00z forecast just before snow fell, and starts 96 hours before that.  Top left is Euro, top right is GFS, bottom left is GGEM, and bottom right is ICON. 

The Euro messed up its final forecast, but it was clearly better than the GFS at 96, 84, 72, 60, and 48 hours out.  It might have been better at 36, 24, and 12 hours out as well - I'd have to see verification data to know for sure.  It was around 60 hours out that Bob Chill had this prescient post about the first GFS run to get a little precip across the bay:

Got precip west of the bay. Looks like a lower qpf version of the euro. I'm telling you, the gfs is going to look like the smartest model when this is all said and done even though it was too far east for like 50 runs. 

Good call. 
3)  All of the models are getting better, and they might have closed the gap with the Euro (but not by much, based on verification scores).  At the very least, we're taking 4 and 5 day forecasts more seriously than we used to.  But the models have always been a lot closer to each other in performance than people believe.  The difference in skill between the Euro and NAVGEM is about 1 day.  That is, today's NAVGEM is about as good as yesterday's Euro, with the other models somewhere in between.  It's possible that recent changes have made the Euro worse for storms in our area, but the Euro has always had its bad events.  People tend to overlook them when its reputation is good and overemphasize them when its reputation is bad.  It's confirmation bias.  We all do it, and it's really hard to avoid.  And yeah, there's recency bias as well.  But when you look at things objectively and in the long run, the Euro is still the best model by good measure. 
fwiw, the Canadian ensemble still doesn't like this storm, but it finally has one big hit in there which had been lacking earlier.
 

NAVGEM is being uber progressive with the trailing sw so much so it puts it in play with the trof. The surface isn't a red flag in that the NAVGEM is apples to apples with other guidance at H5.....it isn't. Its being so progressive upstream it is keying on a completely different sw. Its doing what is does best.....losing the battle.

[mitchnick]

2 minutes ago, cae said:

Yeah, tough 12z suite.  It's too bad to see the changes in the Euro / EPS, and interesting discussion about the Euro's performance.  Fwiw, here are my thoughts:

1)  It hasn't snowed yet.  We have 3 1/2 days before the GFS says the first flakes fall on the cities.  I think sometimes we jump the gun with assessing model performance.

2)  After storms, we tend to remember model performance based on their last few forecasts, not the ones that occur 3-6 days out.  I've read several posts here that suggest the GFS did better than the Euro on the January 4th storm.  Really?  I put together a gif after that storm comparing the performance of the Euro, GFS, GGEM, and ICON.  I'll re-post it here.  It ends with the 00z forecast just before snow fell, and starts 96 hours before that.  Top left is Euro, top right is GFS, bottom left is GGEM, and bottom right is ICON. 

The Euro messed up its final forecast, but it was clearly better than the GFS at 96, 84, 72, 60, and 48 hours out.  It might have been better at 36, 24, and 12 hours out as well - I'd have to see verification data to know for sure.  It was around 60 hours out that Bob Chill had this prescient post about the first GFS run to get a little precip across the bay:

Good call. 

3)  All of the models are getting better, and they might have closed the gap with the Euro (but not by much, based on verification scores).  At the very least, we're taking 4 and 5 day forecasts more seriously than we used to.  But the models have always been a lot closer to each other in performance than people believe.  The difference in skill between the Euro and NAVGEM is about 1 day.  That is, today's NAVGEM is about as good as yesterday's Euro, with the other models somewhere in between.  It's possible that recent changes have made the Euro worse for storms in our area, but the Euro has always had its bad events.  People tend to overlook them when its reputation is good and overemphasize them when its reputation is bad.  It's confirmation bias.  We all do it, and it's really hard to avoid.  And yeah, there's recency bias as well.  But when you look at things objectively and in the long run, the Euro is still the best model by good measure. 

fwiw, the Canadian ensemble still doesn't like this storm, but it finally has one big hit in there which had been lacking earlier.

 

You forget that only the Euro showed a major snowstorm on back-to-back runs 6 days out and making light of the fact that the 0z run the night before is a mistake.  In fact, I never posted it because I was too disgusted, but the morning of the storm the 12z run dropped .08" qpf imby. The storm was over except for light snow and flurries with no accumulations. In short, even though there were periods before the storm when it beat other guidance,  it was all over the place and the only model to drop 5-9" on us. Major fail in my book.

[Cobalt]

1 minute ago, mitchnick said:

You forget that only the Euro showed a major snowstorm on back-to-back runs 6 days out and making light of the fact that the 0z run the night before is a mistake.  In fact, I never posted it because I was too disgusted, but the morning of the storm the 12z run dropped .08" qpf imby. The storm was over except for light snow and flurries with no accumulations. In short, even though there were periods before the storm when it beat other guidance,  it was all over the place and the only model to drop 5-9" on us. Major fail in my book.

Canadian had back to back big hits just before the Euro IIRC

[cae]

2 minutes ago, mitchnick said:

You forget that only the Euro showed a major snowstorm on back-to-back runs 6 days out and making light of the fact that the 0z run the night before is a mistake.  In fact, I never posted it because I was too disgusted, but the morning of the storm the 12z run dropped .08" qpf imby. The storm was over except for light snow and flurries with no accumulations. In short, even though there were periods before the storm when it beat other guidance,  it was all over the place and the only model to drop 5-9" on us. Major fail in my book.

Next time I put together a similar gif I might go back farther in time.  I didn't this time because I usually don't give the ops much weight 100+ hours out.  I didn't mean to make light of it messing up the last run - sorry if it came across that way.  I agree that run was a clear fail.

[Bob Chill]

Great post Cae. I appreciate the look back. Post mortem isn't done often enough. 

 

Global h5 verification is a different metric than imby verification. The gfs was always pretty good with northern stream systems. It ruled 13-14 unless I'm forgetting a lot (very possible). The upgrade messed with us because it changed what we knew was usually wrong with the gfs midrange runs. 

I can say now that I believe once again that the gfs is a better med range global that the euro with the northern stream. Euro has always destroyed the gfs with southern stream stuff. We haven't had a chance to test the new gfs with that becuase we haven't had a good southern stream winter since 09-10. Lol

One thing the euro has done multiple times recently is went off the rails with amplification in the med range. I was skeptical with this most recent event becuase it was out on its own. Now that leads shortened those amplified runs are gone. I do think the euro is the best qpf model in the 24-48 hour range. It threw out a goofy run at ground zero but in the end, the 12z run before the bomb is exactly what everyone around here got. It underdid precip to the SE but it did very well on the margins before the jacked up run. 

The short story (and something we all need to do every time) is enjoy good runs beyond 96 hours but never take anything seriously until inside of 72. If you do that with every event, models will look like they do a great job. And they should because d3 is when a global should have it mostly right. Beyond day 3 big swings can always happen. Especially beyond d4. 

[SnowGoose69]

46 minutes ago, Cobalt said:

It did lead the way during the Dec 8-9 event IIRC. Still, it's just the Navgem, but I'd say we have about 24 hours of guidance to see what the models do with the costal part before giving up. 

It’s been ridiculously good.  To the point it didn’t flop one bit inside Day 6 on either that event or the big coastal last week 

[MD Snow]

7 minutes ago, SnowGoose69 said:

It’s been ridiculously good.  To the point it didn’t flop one bit inside Day 6 on either that event or the big coastal last week 

I meant to post it's trend GIF after the Jan 4 storm. I remember it literally bringing precip to just west of the cities run after run after run when every other model was back and forth. 

 

[clskinsfan]

Interesting that the 18Z Nam looks almost identical to the 12Z Navgem during next weeks event. They both are not far off from what the GFS is showing . Just slightly west with the coastal development and hence wetter. 

And if you like the regular NAM take a look at the 12K. Would be a very nice event.

[Bob Chill]

18z gfs is a big shift at h5 over top of the shortwave at really short leads. Not saying it's a good or bad shift. Just that it's not just a little bump and we're only talking 60 hours out in time. Don't usually see run over run shifts like that at 2.5 day leads. 

[usedtobe]

The NAM has a sharper 500h than the GFS at the same time.  Big differences. The NAM looks sort of like last night's Euro before ti completely folded.  Too bad we're probably being NAMed