December Med/Long Range Disco

[yoda]

Few flakes N and W of town next Thursday morning? 

 

ETA:  Hrs 165 to 168

[Ji]

There definitely seems to be multi-year dominant phases with both the AO and NAO. The hows and whys are way above my pay grade. But with any multi-year trend there are exceptions. We're supposed to be in a -PDO phase but last year into this year have proved otherwise. 

 

My confidence in the SAI/SCE is at an all time low. 

judah cohen followed me on twitter. Wont be long till i start trolling him

[AvantHiatus]

There definitely seems to be multi-year dominant phases with both the AO and NAO. The hows and whys are way above my pay grade. But with any multi-year trend there are exceptions. We're supposed to be in a -PDO phase but last year into this year have proved otherwise. 

 

My confidence in the SAI/SCE is at an all time low. 

Agreed, caution is recommended going forward, particularly when using older analogs.

[BristowWx]

Few flakes N and W of town next Thursday morning? 

 

ETA:  Hrs 165 to 168

Darn you Yoda.  I wanted to post that.  few flakes that day perhaps. 

[yoda]

12z GFS is at Day 12

[WinterWxLuvr]

Few flakes N and W of town next Thursday morning? 

 

ETA:  Hrs 165 to 168

That one seems close to be something bigger IMO.  Need a bit more depth in the middle I guess.

[Ji]

That one seems close to be something bigger IMO.  Need a bit more depth in the middle I guess.

well after the 18-21...i did notice that heights started to rise in alaska and pull north to the pole...a few days later,  our ridge becomes a trough and it snows lol

 

http://www.tropicaltidbits.com/analysis/models/gfs/2015121112/gfs_z500a_namer_41.png

[usedtobe]

well after the 18-21...i did notice that heights started to rise in alaska and pull north to the pole...a few days later,  our ridge becomes a trough and it snows lol

 

http://www.tropicaltidbits.com/analysis/models/gfs/2015121112/gfs_z500a_namer_41.png

That solution sure is at odds with last night's ensemble mean forecast from all the models.  They have the 850 temps warmer than normal on the 24th and 25th while the latest GFS has it possibly snowing.  I'd sure bet on the ensemble mean forecasts at that time range. It will be interesting to see what today's ens mean temp forecasts look like.  For the frist wave on the 18-21,  last night's GEFS were cold enough to suggest the GFS isn't crazy.   The 00Z Euro ensemble,  dropped from 8 of 50 showing snow during that window to 3 out of 50.  Of course the next run could double or triple that number. 

[snowman19]

Another day another totally different operational GFS solution in the long range. That model flips and flops like a fish out of water. It will show something totally different at 0z. SMH...

[Ian]

Increased LR chaos on the ops might be a sign of some sort of shakeup ahead but I'd imagine they are rushing it. There is a slow move in the ensembles to something that might not want us to slit our wrists. I think we're still bleeding into Jan.. how far, not sure... for anything other than a fluke or something like snow tv+.

[Heisy]

That event still would be a rainer, there's no HP to lock in cold air at the surface. 

[usedtobe]

I agree wrt the ensembles--I'm weighing those more in the LR. The period next weekend is of interest with colder air nearby, but it still has a thread the needle look. Would a stronger low make it more likely we see a colder system? Track dependent of course.

I think it is track dependent and heavier precipitation might help providing the low stays right along the coast.  Still, the airmass over the CONUS is not very cold so no matter what away from the higher terrain I think snow is unlikely.   I don't see anything to make me change my mind from yesterday's outlook.

[usedtobe]

Increased LR chaos on the ops might be a sign of some sort of shakeup ahead but I'd imagine they are rushing it. There is a slow move in the ensembles to something that might not want us to slit our wrists. I think we're still bleeding into Jan.. how far, not sure... for anything other than a fluke or something like snow tv+.

Maybe but until the AO goes negative or we get high heights over AK, it will take a fluke.  The ensembles show a tad of ridging over Ak but heights remain below normal making me wonder how much ridging there will actually be;  One the bright side, the pattern can't get any worse than it is this week.  Any changes will be positive.

[dtk]

And it's almost guaranteed to show something different by this time tomorrow.

 

Excitement and despair are both futile exercises.

Why are you so obsessed with commenting on lack of continuity in long range deterministic forecasting?  Do you understand the (lack of) skill of deterministic prediction at these lead times and how it is fundamentally related to predictability?

[Ian]

Maybe but until the AO goes negative or we get high heights over AK, it will take a fluke.  The ensembles show a tad of ridging over Ak but heights remain below normal making me wonder how much ridging there will actually be;  One the bright side, the pattern can't get any worse than it is this week.  Any changes will be positive.

Both Euro and GFS ensemble analogs are hinting at some attempted -NAO in the d16-30 range. Though it's a rollover so who knows. Plus if you change the variables it's not as noticeable. Perhaps a reach. No major signs of such on the actual runs or the monthlies etc. We are of course getting into a time of year it can snow without perfect conditions. Getting a real southern stream going -- which does seem to be hinted at more and more and is historically favored -- might at least give us some more ops.

[dtk]

Another day another totally different operational GFS solution in the long range. That model flips and flops like a fish out of water. It will show something totally different at 0z. SMH...

Ugh.

[WinterWxLuvr]

Why are you so obsessed with commenting on lack of continuity in long range deterministic forecasting?  Do you understand the (lack of) skill of deterministic prediction at these lead times and how it is fundamentally related to predictability?

I was actually responding to another post.

 

Why are you so obsessed and defensive about computer models?

 

I'm usually a defender of the GFS.  I've never really considered day 6 or 7 as long range.  And yes, I think the lack of skill is plainly evident.  Maybe I'll have to redefine my ranges.  Go with 72 hours as the max.

[dtk]

I was actually responding to another post.

 

Why are you so obsessed and defensive about computer models?

 

I'm usually a defender of the GFS.  I've never really considered day 6 or 7 as long range.  And yes, I think the lack of skill is plainly evident.  Maybe I'll have to redefine my ranges.  Go with 72 hours as the max.

You (and others) have made many posts in the past week+ about model solutions changing at longer ranges.  

 

I'm obsessed with computer models because this is what I do for a living.  More precisely, I am quite interested in data assimilation, model initialization, and inherit predictability.  I have several peer-reviewed papers on the current scheme used in the NCEP GDAS, and several research projects trying to continue to move things forward...including the upcoming implementation of hybrid 4D EnVar.

 

For what it's worth, this has nothing to do with the GFS and I'm not defending anything...the same comments apply to the deterministic ECMWF, Canadian, etc.  The mean, hemispheric ACC skill for wavenumbers 4-9 (large scale down to high end synoptic scale) is only <0.65 at day 7, and for wavenumbers 10-20 the monthly skill is < 0.3 (for all models).  Anything below 0.6 is almost useless.  I'll leave my feelings about the current version of the GFS aside, since it leaves a lot to be desired.

 

This is why the operational centers run ensemble systems.  There are things that exhibit predictability at ranges beyond 6-7 days, but individual events are not high on the list.  You absolutely need to understand predictability at various time ranges in order to understand how to use the tools (the models).  Bob had some good comments on this the other day so I won't reiterate, but using ensembles at longer ranges to identify patterns is much more useful than worry about what a deterministic model shows for a highly nonlinear cyclogenesis event at 7+ days.  

[WinterWxLuvr]

You (and others) have made many posts in the past week+ about model solutions changing at longer ranges.  

 

I'm obsessed with computer models because this is what I do for a living.  More precisely, I am quite interested in data assimilation, model initialization, and inherit predictability.  I have several peer-reviewed papers on the current scheme used in the NCEP GDAS, and several research projects trying to continue to move things forward...including the upcoming implementation of hybrid 4D EnVar.

 

For what it's worth, this has nothing to do with the GFS and I'm not defending anything...the same comments apply to the deterministic ECMWF, Canadian, etc.  The mean, hemispheric ACC skill for wavenumbers 4-9 (large scale down to high end synoptic scale) is only <0.65 at day 7, and for wavenumbers 10-20 the monthly skill is < 0.3 (for all models).  Anything below 0.6 is almost useless.  I'll leave my feelings about the current version of the GFS aside, since it leaves a lot to be desired.

 

This is why the operational centers run ensemble systems.  There are things that exhibit predictability at ranges beyond 6-7 days, but individual events are not high on the list.  You absolutely need to understand predictability at various time ranges in order to understand how to use the tools (the models).  Bob had some good comments on this the other day so I won't reiterate, but using ensembles at longer ranges to identify patterns is much more useful than worry about what a deterministic model shows for a highly nonlinear cyclogenesis event at 7+ days.  

dkt, I've always respected your knowledge and your posts.

 

Perhaps my posts should have been worded better.  I usually am impressed with the GFS.  Others will back me up when I say that I defend it.

 

I just think lately that it has had a hard time when it gets out to about 6 or 7 days.  I think it's usually more consistent than that.

 

I think my comment that despair and excitement being futile exercises backs up what you are saying.  We shouldn't be getting excited or upset over a 7 day model forecast.  If we do, we are expecting more from them than they can deliver.

[WinterWxLuvr]

Euro is a lot deeper with the trough than the GFS next Friday.  Stronger storm, but still looks warm......way too warm.

 

But I don't have much access.

[dtk]

dkt, I've always respected your knowledge and your posts.

 

Perhaps my posts should have been worded better.  I usually am impressed with the GFS.  Others will back me up when I say that I defend it.

 

I just think lately that it has had a hard time when it gets out to about 6 or 7 days.  I think it's usually more consistent than that.

 

I think my comment that despair and excitement being futile exercises backs up what you are saying.  We shouldn't be getting excited or upset over a 7 day model forecast.  If we do, we are expecting more from them than they can deliver.

Fair enough...I wasn't trying to be a jerk.

 

Your comment about lack of consistency, in a relative sense, may have some merit.  As you point out, however, people shouldn't get excited about 7 day deterministic model forecasts unless predictability is higher than normal (and there is consistency among models, and within the same model as lead times get shorter).

 

There are some signs from the CDAS that perhaps this early winter is somewhat less predictable than in recent years.  If I compare the current skill of the GFS and ECMWF with recent December performances in 2013/2014 (different models, so not clean), there are some signs that the skill is somewhat lower than expected consistent with the slightly lower predictability.  Even more interesting, I don't think this was the case for October/November....

 

I also happen to think that the current GFS -- 13km Semi-Lagrangian, which was implemented back in January, is more susceptible to jumps and less continuity when looking cycle to cycle.  Some of this is likely due to the increased resolution and inherit error growth therein, but I wonder about some of the other changes in the model as well.  The 4D parallel package, noted as GFSX on some of the current verification, does seem to help with this somewhat, but it is certainly not solving all of the ills in the GFS.

[WinterWxLuvr]

Fair enough...I wasn't trying to be a jerk.

 

Your comment about lack of consistency, in a relative sense, may have some merit.  As you point out, however, people shouldn't get excited about 7 day deterministic model forecasts unless predictability is higher than normal (and there is consistency among models, and within the same model as lead times get shorter).

 

There are some signs from the CDAS that perhaps this early winter is somewhat less predictable than in recent years.  If I compare the current skill of the GFS and ECMWF with recent December performances in 2013/2014 (different models, so not clean), there are some signs that the skill is somewhat lower than expected consistent with the slightly lower predictability.  Even more interesting, I don't think this was the case for October/November....

 

I also happen to think that the current GFS -- 13km Semi-Lagrangian, which was implemented back in January, is more susceptible to jumps and less continuity when looking cycle to cycle.  Some of this is likely due to the increased resolution and inherit error growth therein, but I wonder about some of the other changes in the model as well.  The 4D parallel package, noted as GFSX on some of the current verification, does seem to help with this somewhat, but it is certainly not solving all of the ills in the GFS.

LOL, your knowledge of these models is impressive to say the least, but way over my head.  But, as always, we all certainly appreciate your posts.

[Ji]

Euro is a lot deeper with the trough than the GFS next Friday.  Stronger storm, but still looks warm......way too warm.

 

But I don't have much access.

you need a strung out low...not a strong low to have a chance it seems

[WxUSAF]

Next Friday looks like something worth watching. Trailing wave on gfs, euro, and Canadian. Temps obviously a major issue but it's not like there's something else to watch. Snowtv?

[Bob Chill]

you need a strung out low...not a strong low to have a chance it seems

 

More importantly (irt the euro solution) is you need the front behind the cutter to clear our area quicker. GFS and Euro are handling the storm and associated front differently. Euro is pretty slow and weak with the front in comparison to the GFS. Even with the gfs being a much better solution for getting the cold in here, it's really marginal when the wave rides up. It's another one of those things where everything has break our way with virtually no wiggle room. 

 

If we're going to go from a blast furnace to a snow storm it usually requires an arctic front and not some barely seasonal air mass pushing in. At least for areas east of the mountains anyways. 

[BTRWx]

I had this posted in the wrong forum earlier, but I would like to know if we can post these ecmwf images here from the main ecmwf.int source? These are from last night's 0Z euro that look rather promising to me, though near the end of the model's range.

 

http://www.ecmwf.int/en/forecasts/charts/medium/mean-sea-level-pressure-wind-speed-850-hpa-and-geopotential-500-hpa?area=North%20America&step=48&relative_archive_date=2015101612&parameter=Geopotential%20500%20hPa%20and%20Temperature%20at%20850%20hPa&time=2015121100,216,2015122000

 

 

http://www.ecmwf.int/en/forecasts/charts/medium/mean-sea-level-pressure-wind-speed-850-hpa-and-geopotential-500-hpa?area=North%20America&step=48&relative_archive_date=2015101612&parameter=Wind%20850%20and%20mslp&time=2015121100,216,2015122000

[WinterWxLuvr]

you need a strung out low...not a strong low to have a chance it seems

 

Seems the deeper trough slows the front.  I think a strong low would be better if we can get a colder airmass in place first.

 

Next Friday looks like something worth watching. Trailing wave on gfs, euro, and Canadian. Temps obviously a major issue but it's not like there's something else to watch. Snowtv?

 

I think so too.  It's the first thing that we've seen that's even interesting enough to follow.

 

More importantly (irt the euro solution) is you need the front behind the cutter to clear our area quicker. GFS and Euro are handling the storm and associated front differently. Euro is pretty slow and weak with the front in comparison to the GFS. Even with the gfs being a much better solution for getting the cold in here, it's really marginal when the wave rides up. It's another one of those things where everything has break our way with virtually no wiggle room. 

 

If we're going to go from a blast furnace to a snow storm it usually requires an arctic front and not some barely seasonal air mass pushing in. At least for areas east of the mountains anyways. 

 

I think we need a strong front that just barely clears us.

 

Wasn't Dec 5, '09 that very thing?  Seems like I remember a good front coming through and then stalling with a wave riding it.  Maybe I'm mistaken, but that's my memory.

[Ji]

JB is hyping the 12z GFS run with Arctic Hammer and eastern storm words

[Bob Chill]

Both Euro and GFS ensemble analogs are hinting at some attempted -NAO in the d16-30 range. Though it's a rollover so who knows. Plus if you change the variables it's not as noticeable. Perhaps a reach. No major signs of such on the actual runs or the monthlies etc. We are of course getting into a time of year it can snow without perfect conditions. Getting a real southern stream going -- which does seem to be hinted at more and more and is historically favored -- might at least give us some more ops.

 

I've been watching the strat plots the last couple weeks. There is distinct warming over eastern siberia at 10-30-50hpa that isn't moving back in time. It starts in 5 days now. I'm not saying it's anything related to a SSW but the PV is getting bumped a little. Just having it disturbed and elongated can shake things up. Last year had multiple elongations that did help displace cold polar air into the continents even with a persistent +AO/NAO. It's nothing exciting right now but worth watching. If the signal holds and gets stronger and ensembles start showing a real regime shift then we can start getting a little excited. 

 

Also, looking at the 12z GEFS mean it seems pretty benign but looking at the members shows a near equal split of amplification and fairly deep troughs east and west. So the d10-15 period looks weak on the means, sensible wx will probably be more interesting. Flow is moving along with plenty of potential storminess across the country. 

 

Overall it still looks really rough for us to get snow but I'm envisioning things a good bit more active coming up. For better or worse at least it's not what were going through over the next week. 

[WinterWxLuvr]

SnowTV would be a huge win. Temps on the euro don't get all that decent until precip is shutting off.

I'd rather have today than that.