February 2013 mid-long range disco thread Part 2

[usedtobe]

  On 2/14/2013 at 3:01 PM, WinterWxLuvr said:

I guess nobody cares.  Amazing.  I don't see the Euro, but I'll take what the GFS is showing right now.  We don't have to have a bombing low off the Carolina coast to get a good winter storm.  The setup on the GFS works too.

 

If I lived where you lived I'd sure be excited about the Feb 22 event as the euro shows really good looking damming. I doubt the Ji version of euro snowfall is correct especially towards DC. 

[chris87]

  On 2/14/2013 at 2:55 PM, WxUSAF said:

It is and they certainly do.  DT posted a chart about this I think yesterday or the day before.  Hell...I'm stunned at how much computing power NASA has at hand (and apparently underutilized).  For my new job here, my colleague suggested I request 400-800K hours of processor time and that it was a modest request!    

 

Showmethesnow:  I think it's even more...try DT's FB page for the chart.  If they're running 50 ensembles at 16 km resolution, then they have A LOT more computing power than NCEP.  

 

It's not just NASA...a lot of NOAA's computing power is tied up at places like ESRL (outside of the operational centers) and is climate-centric (and with a tenuous research-to-operations path). Regardless, the issues with NWP here are a lot more complicated than just needing more computing power....

[gymengineer]

  On 2/14/2013 at 2:27 PM, Ji said:

it was on a saturday evening. IAD got between 4-6 inches and northern Mont county got zero. it was a weird system.

Saturday evening? Maybe 1/28/95? Although, I lived in Potomac then and we did get accumulating snow- ~3-4".

[BlizzardNole]

  On 2/14/2013 at 2:57 PM, 87storms said:

there was a storm that bob ryan called "the stripe snow" that gave dc and some surrounding areas a good 4-8"+, but not much farther south or north.  i feel like that was in dec 98, but could be wrong.

 

there was also a storm in the 90s (maybe earlier 90s) that gave rockville about 5" and dc about an inch.  i know because i was in rockville.  i'm just pretty awful with dates.  

I live in Germantown and work in DC -- the March 99 deal was at least 8" in DC and I remember close to a foot in parts of NoVa? In Germantown we got about 5" or so, but as close as northern MD got close to nothing. The map for it is a stripe NW to SE right over the DC area. The news called in DC's private snowstorm LOL.

In 12/5/03, I got on the train in Germantown with 8" and mod sn; and got off in DC to only rain. That was the biggest difference I've seen.

[HighStakes]

  On 2/14/2013 at 2:18 PM, Ji said:

does anyone remember an event in the 1990s where  DC/Northern Virginia area got 4-6 inches but north of Rockville, they got flurries. I remember getting pounded but my parents went to their friends house up near Rockville and only saw flurries?

There was a storm on Jan. 28th 1995 that fits this description. Terps beat Duke that day and when we walked out of Cole Field House there was about 4 on the ground and it was still snowning. Traffic was a nightmare. Took an hour just to get back on 95. Once you got 5 miles north there was absolutely nothing on the ground. It was a virga storm from Columbia north. It was my birthday and I was pissed but the Terps win helped to ease the pain. Fortunately we all got something decent the next weekend with the noreaster.

[gymengineer]

  On 2/14/2013 at 3:22 PM, HighStakes said:

There was a storm on Jan. 28th 1995 that fits this description. Terps beat Duke that day and when we walked out of Cole Field House there was about 4 on the ground and it was still snowning. Traffic was a nightmare. Took an hour just to get back on 95. Once you got 5 miles north there was absolutely nothing on the ground. It was a virga storm from Columbia north. It was my birthday and I was pissed but the Terps win helped to ease the pain. Fortunately we all got something decent the next weekend with the noreaster.

That's the one I mentioned above too. The only thing is that the cutoff was not "just north of Rockville." Gaithersburg and Germantown both reported 3".

[HighStakes]

  On 2/14/2013 at 3:26 PM, gymengineer said:

That's the one I mentioned above too. The only thing is that the cutoff was not "just north of Rockville." Gaithersburg and Germantown both reported 3".

Must have been a west to east sharp cutoff as well as south to north. I also remember it being a very cold storm.

[HighStakes]

  On 2/14/2013 at 2:33 PM, psuhoffman said:

March 99

Got almost 3 in Reisterstown. LWX has a snowfall accumulation map for that storm. Even areas in southern PA received measurable.

[dtk]

  On 2/14/2013 at 2:40 PM, WxUSAF said:

Maybe you know...I've never gotten a good answer on this...are the Euro ensembles useful within 72...48...or 24 hours of an event?  I assume they must be lower resolution than the Op.  

 

Relative to the GEFS, I would say they are still useful even at short-ish ranges, given their resolution and treatment of model error/uncertainty.  The resolution is lower than the operational (just like the GEFS), but still pretty high:  T639 resolution from day 0 to day 10, and with lower (T319) resolution from day 10 to day 15.  They use semi-Lagrangian dynamics and a corresponding linear grid which means it isn't directly comparable to the GFS (which currently uses an Eulerian model and quadratic grid).  Even still, their ensemble integrations are comparable in resolution (out to day 10) to the current operational deterministic GFS.....but much coarser than their operational deterministic integration.

[usedtobe]

  On 2/14/2013 at 3:50 PM, dtk said:

Relative to the GEFS, I would say they are still useful even at short-ish ranges, given their resolution and treatment of model error/uncertainty.  The resolution is lower than the operational (just like the GEFS), but still pretty high:  T639 resolution from day 0 to day 10, and with lower (T319) resolution from day 10 to day 15.  They use semi-Lagrangian dynamics and a corresponding linear grid which means it isn't directly comparable to the GFS (which currently uses an Eulerian model and quadratic grid).  Even still, their ensemble integrations are comparable in resolution (out to day 10) to the current operational deterministic GFS.....but much coarser than their operational deterministic integration.

 

Daryl,  Thanks for the clarification about the european ensembles. 

[dtk]

  On 2/14/2013 at 3:14 PM, chris87 said:

It's not just NASA...a lot of NOAA's computing power is tied up at places like ESRL (outside of the operational centers) and is climate-centric (and with a tenuous research-to-operations path). Regardless, the issues with NWP here are a lot more complicated than just needing more computing power....

I agree in principle, but the lack of computing power not only holds back the ability to run higher resolution in production, it severely limits our capabilities in R&D mode.  However, no excuses, we have to do the best we can with what we have.  It's not a matter of what we cannot do, but a matter of what we can/should do.

 

By the way, the references in computing power are for our current operational machine (IBM P6, which is severely outdated now)....we are going through a transition process to our new supercomputer; scheduled to go live in August (I think).  This will give us some increase in computing power.

[chris87]

  On 2/14/2013 at 3:55 PM, dtk said:

I agree in principle, but the lack of computing power not only holds back the ability to run higher resolution in production, it severely limits our capabilities in R&D mode.  However, no excuses, we have to do the best we can with what we have.  It's not a matter of what we cannot do, but a matter of what we can/should do.

 

By the way, the references in computing power are for our current operational machine (IBM P6, which is severely outdated now)....we are going through a transition process to our new supercomputer; scheduled to go live in August (I think).  This will give us some increase in computing power.

 

Yeah I agree, I think a commitment from NOAA leadership to make NWP a larger priority would fix a lot of the other issues (outside of just pure upgrades to the infrastructure)....

[mitchnick]

GFS 12z looks like it will come in south of 0z with the 2/22 event

[Bob Chill]

  On 2/14/2013 at 4:19 PM, mitchnick said:

GFS 12z looks like it will come in south of 0z with the 2/22 event

 

There's just nothing to stop it from going N. Pretty strong departing the rockies. I can't remember many that track across the country and come underneath us without a massive block. 

[WxUSAF]

  On 2/14/2013 at 4:19 PM, mitchnick said:

GFS 12z looks like it will come in south of 0z with the 2/22 event

If that is south...where did 0z have it?  Manitoba?  

 

Not that it's worth discussing much at 180+ hours, but still starts as snow or ice for the 95 corridor and points N/W at least. 

[Bob Chill]

CAD/wedge sig. 850's and surface below freezing at onset. Snow to rain. .25+/- falls as snow if i had to guess. Not that is matters much. 

[North Balti Zen]

That's a nice midwest storm on 2/22. Are we tracking midwest storms now?

[mitchnick]

  On 2/14/2013 at 4:31 PM, WxUSAF said:

If that is south...where did 0z have it?  Manitoba?  

thru 153 hrs vs. 165 at 0z, it was south by a noticeable bit

then the block, if that's what we want to call it, just shriveled up and let it come north

we root for front end thump and that's the best we do

slight chance it cuts so far west that it doesn't bring so much warm air with it, but that's a long shot

[usedtobe]

  On 2/14/2013 at 4:32 PM, Bob Chill said:

CAD/wedge sig. 850's and surface below freezing at onset. Snow to rain. .25+/- falls as snow if i had to guess. Not that is matters much. 

 

If I was in Winchester or even Ji in Leesburg I'd be thinking I'd see another accumulating snow.  In DC without looking at the temps above 850, it's hard to know how we shake out. 

[WxUSAF]

Cutting to Chicago or Madison, WI, is usually better for us than cutting to Cleveland if we can get some overrunning precip into the cold wedge before the warm front scours it out.  

[mitchnick]

end of the gfs run is March

time to get out the fork on this winter

[Ian]

at least the gfs makes more sense than the euro. the euro somehow thumps us with like 10" in 10 hours.

[Bob Chill]

  On 2/14/2013 at 4:35 PM, usedtobe said:

If I was in Winchester or even Ji in Leesburg I'd be thinking I'd see another accumulating snow.  In DC without looking at the temps above 850, it's hard to know how we shake out. 

 

Yea, I was just casually pointing it out. Euro has the same idea. Comes down to details that won't be know for days and days. 

[mitchnick]

  On 2/14/2013 at 4:36 PM, Ian said:

at least the gfs makes more sense than the euro. the euro somehow thumps us with like 10" in 10 hours.

I guess I'll need to delete all my other posts re the Euro and hug it until it starts leaking PCBs

[Ian]

  On 2/14/2013 at 4:38 PM, mitchnick said:

I guess I'll need to delete all my other posts re the Euro and hug it until it starts leaking PCBs

 

tho the gfs gives us some accum on the backside .. i dont think any really on the front end dc/balt despite 850s.. tho of course cad will grow.   the high location and movement is not spectacular.

[Ian]

  On 2/14/2013 at 4:47 PM, zwyts said:

Are we really going to see such a strong cutter?...I know the euro has lost some of its luster lately, but I would use it and the gefs/euro ensembles at this range for some very very broad guidance versus the GFS

 

maybe not but the euro op is kind of on its own with that bigger snow potential. euro ens mainly suggests its a midwest to new england sno event with some down here. not a horrible idea i suppose.

[usedtobe]

  On 2/14/2013 at 4:47 PM, zwyts said:

Are we really going to see such a strong cutter?...I know the euro has lost some of its luster lately, but I would use it and the gefs/euro ensembles at this range for some very very broad guidance versus the GFS

 

At that time range I'd certainly trust the ensemble more than the operational runs. 

[Ian]

  On 2/14/2013 at 4:54 PM, zwyts said:

not sure it is relevant but we saw teh same thing with the KU....it was a solid lakes cutter a week out....the models usually trend away from the lakes with these things though not necessarily good enough for us....i'll go ahead and say the primary is weaker and and runs up the apps or just west into the eastern OV

 

yeah the storm around christmas did sort of the same thing too. i'd probably hedge on the eastern side myself for now tho i don't think it's impossible for the ops solution to verify (at least out west) either.

[psuhoffman]

a few of the individual GEFS members still have a better solution for Feb 22, the GGEM for what little it is worth looks really good, and the Euro still seems in the ballpark.  Until I see the NAO break down, I am not bailing on that one. 

[clskinsfan]

Thats the key is the block. Will it really be there? Who knows.