December and a Hint of January Pattern Discussion

[Ian]

its only real when it warms us up and takes snow away. What happened to the -50 below normal in WV?

20 above normals ate it -- think it was snow induced anyway 

 

[Bob Chill]

It is so much better than the last 2....Dec 2004/2006/2011 etc were painful...

 

No doubt. We've talked about getting nickled and dimed to a semi-respectable year total. Lets bag an inch or 2 this weekend and see what kind of mess we can get into closer to the end of the month. I having a lot of fun so far this year. Wasting an entire month in the djf stretch is vomitworthy. 

[Ji]

20 above normals ate it -- think it was snow induced anyway 

 

ecmwf_t2m_anom_ky_41.png

wow what a flip. As DT says...when a model flips that much in one run....toss it. #Dt101

[HM]

I didn't know where to post these, but here are some papers on freezing rain/sleet (for some of you, this will be boring):

http://journals.ametsoc.org/doi/pdf/10.1175/1520-0434(2004)019%3C0377:AAOFRF%3E2.0.CO%3B2

http://journals.ametsoc.org/doi/pdf/10.1175/1520-0450(2003)042%3C1302:TASVOF%3E2.0.CO%3B2

http://journals.ametsoc.org/doi/pdf/10.1175/1520-0450%282001%29040%3C1724%3AASWPAS%3E2.0.CO%3B2

http://journals.ametsoc.org/doi/pdf/10.1175/1520-0434%282002%29017%3C0047%3ALASEAW%3E2.0.CO%3B2

[mitchnick]

I didn't know where to post these, but here are some papers on freezing rain/sleet (for some of you, this will be boring):

http://journals.ametsoc.org/doi/pdf/10.1175/1520-0434(2004)019%3C0377:AAOFRF%3E2.0.CO%3B2

http://journals.ametsoc.org/doi/pdf/10.1175/1520-0450(2003)042%3C1302:TASVOF%3E2.0.CO%3B2

http://journals.ametsoc.org/doi/pdf/10.1175/1520-0450%282001%29040%3C1724%3AASWPAS%3E2.0.CO%3B2

http://journals.ametsoc.org/doi/pdf/10.1175/1520-0434%282002%29017%3C0047%3ALASEAW%3E2.0.CO%3B2

thanks HM....are there any pictures?          

 

j/k

[HM]

 

thanks HM....are there any pictures?          

 

j/k

ha...these are definitely easy on the brain, as far as research papers go, and they have nice statistics you guys might find interesting. The common temperatures/rh etc. are pretty interesting and are not as cold as you would think. Heck, there's even a diurnal cycle-correlation with the ZR. A lot of the stuff is intuitive, though.

[snowfan]

Tower Falls Ranger Station in Yellowstone had a low of -20 and is currently at -1. Lots of WY and eastern MT will struggle to see highs >single digits for the next 5-6 days.

[Bob Chill]

I'm starting to think the 12-13th timeframe is something worth watching. It will be probably the last breath of the cold spell and various looks have been showing up. There appears that there could be a vort exiting the rockies and the gulf looks to be open for business. Long way off but throwing it out there to be watched. 

[WinterWxLuvr]

Are analogs and teleconnections programmed into weather models? Hope that's not a remarkably dumb question. I really have no earthly idea of how the models function from a nuts and bolts standpoint.

[WxUSAF]

Are analogs and teleconnections programmed into weather models? Hope that's not a remarkably dumb question. I really have no earthly idea of how the models function from a nuts and bolts standpoint.

No.  There's probably some good links to explain this at a basic level, but the gist is this:

1.  The planet or region of interest is divided up into a 3D grid at a particular horizontal and vertical resolution.

2.  The model has a particular "time step" that it uses to move forward in time.

3.  The model "assimilates" observational data and an earlier model projection to produce the initial conditions for the model.

4.  For each time step, the model then solves simplified versions (but still very complicated) of the differential equations that govern atmospheric motion and radiative transfer. 

[WinterWxLuvr]

No. There's probably some good links to explain this at a basic level, but the gist is this:

1. The planet or region of interest is divided up into a 3D grid at a particular horizontal and vertical resolution.

2. The model has a particular "time step" that it uses to move forward in time.

3. The model "assimilates" observational data and an earlier model projection to produce the initial conditions for the model.

4. For each time step, the model then solves simplified versions (but still very complicated) of the differential equations that govern atmospheric motion and radiative transfer.

Thank you.

The reason I ask is that it seems like the models at long range seem to gravitate toward the same solution over and over. It may just seem that way. In years past it seemed they always went toward cold which never seemed to materialize. This year it seems they want to go warm and it isn't materializing, other than a day or two. Probably just perception on my part and not reality.

[Ian]

I don't know model physics but I thought they at least factor in climo to some degree at range? There is some reason 384 always has an EC storm in winter.

[WinterWxLuvr]

I don't know model physics but I thought they at least factor in climo to some degree at range? There is some reason 384 always has an EC storm in winter.

You would think that if models don't use prior instances (analogs) that they would miss an opportunity to improve over time (evolution through a larger set of possibilities). Again, I have no idea if a computer program can be sophisticated enough to "learn". Maybe in the future.

[WxUSAF]

I don't know model physics but I thought they at least factor in climo to some degree at range? There is some reason 384 always has an EC storm in winter.

MOS does, but I don't think the deterministic model does.  I'm sure dtk could answer these questions in detail. 

 

I'll use the example I know best.  The model I use is sort of a global low-resolution NAM for Mars.  It actually is the core WRF physics.  It doesn't "know" it's on Mars or what temperature it's "supposed" to be.  Of course some things had to be tweaked for a different planet (e.g., solar luminosity, planet size, gravity, etc...), but the code just solves differential equations as best as it can. 

[Ian]

You would think that if models don't use prior instances (analogs) that they would miss an opportunity to improve over time (evolution through a larger set of possibilities). Again, I have no idea if a computer program can be sophisticated enough to "learn". Maybe in the future.

Well their basis of understanding has to be tied to past events in some fashion at a base level it would seem, whether or not it's eating specific analogs for a run.  Then again I still haven't the foggiest idea how radio waves travel through the air and carry data with them. 

[Ian]

MOS does, but I don't think the deterministic model does.  I'm sure dtk could answer these questions in detail. 

 

I'll use the example I know best.  The model I use is sort of a global low-resolution NAM for Mars.  It actually is the core WRF physics.  It doesn't "know" it's on Mars or what temperature it's "supposed" to be.  Of course some things had to be tweaked for a different planet (e.g., solar luminosity, planet size, gravity, etc...), but the code just solves differential equations as best as it can. 

Yeah I know MOS does. I'm probably wrong or thinking about it wrong.  I thought long ago someone said that out in range you often run to a climatological type of solution.

[WxUSAF]

  I thought long ago someone said that out in range you often run to a climatological type of solution.

Well, if that's true, I don't believe it's by design.  Of course, if you compare a GFS 16 day forecast to climatology for the same day, there's probably not much difference in accuracy. 

[Ian]

Well, if that's true, I don't believe it's by design. Of course, if you compare a GFS 16 day forecast to climatology for the same day, there's probably not much difference in accuracy.

I'm probably just remembering wrong. DCAlexandria ftw.

[clueless]

MOS does, but I don't think the deterministic model does.  I'm sure dtk could answer these questions in detail. 

 

I'll use the example I know best.  The model I use is sort of a global low-resolution NAM for Mars.  It actually is the core WRF physics.  It doesn't "know" it's on Mars or what temperature it's "supposed" to be.  Of course some things had to be tweaked for a different planet (e.g., solar luminosity, planet size, gravity, etc...), but the code just solves differential equations as best as it can.

Yah.

[Snowproblem]

You would think that if models don't use prior instances (analogs) that they would miss an opportunity to improve over time (evolution through a larger set of possibilities). Again, I have no idea if a computer program can be sophisticated enough to "learn". Maybe in the future.

Even if they don't "learn" couldn't the algorithms be manually changed? For instance, the models always seem to have problems handling CAD. Couldn't a computer programmer go in and make adjustments to compensate for issues that always seem to appear like the inability to forecast CAD?

[GEOS5ftw]

In principle yes but remember the atmosphere behaves on a continuum of scales from planetary down to molecular while the models basically have 2 scales they care about, the "resolved grid" and sub-grid processes. So with finite computing power you make the resolved grid as fine as you can but you still need to make a ton of assumptions about what happens on the sub-grid scale. For example, precipitation is usually just represented as total condensed liquid mass to save memory and computations, but in reality there are a huge variety of raindrop and snowflake sizes and these differences do matter when it comes to things like evap cooling (smaller particles evaporate faster because of larger surface area to volume ratio), which has feedbacks to density and CAD. So the mean relationships may not represent every situation. There are research models that handle these things more explicitly but there is not enough power to run them in real time generally. Someday...

[WxUSAF]

12z GFS has the Bob Chill storm on the 12-13th, but it's a fish.  Not surprising with the progressive pattern in the Atlantic, but certainly a timeframe worth monitoring.  

[Bob Chill]

12z GFS has the Bob Chill storm on the 12-13th, but it's a fish.  Not surprising with the progressive pattern in the Atlantic, but certainly a timeframe worth monitoring.  

 

oh jeeeze.... My name is to never be associated with a storm again. i still haven't recovered from last December's debacle. Yea, winwxluvr still touts it as a great event. For the rest of us 99%'ers...barf...

 

But it's worth casually watching for sure. Vort is there but it's flat. I'm fairly certain a future model run or 2 will give us hope before the inevitable. 

[Bob Chill]

Well, the euro forms the storm down in the gulf much better than the gfs but the hp and and strong confluence to the north forces it well below us. Definitely a bit more interesting but not a threat....yet...

[snow.]

Euro gives it the old college try but an open contoured 1020mb low running into a 1040mb suppressive high is like a fat guy running uphill into the wind.  he gets blown back.  And eventually gives up and turns around (and presumably goes and eats something)

[Bob Chill]

Euro gives it the old college try but an open contoured 1020mb low running into a 1040mb suppressive high is like a fat guy running uphill into the wind.  he gets blown back.  And eventually gives up and turns around (and presumably goes and eats something)

 

Luckily we have a whole week ahead to figure it out. It has some of the pieces to have a more interesting solution. you know, like a widespread 6" snowfall in central NC. 

[ravensrule]

Euro gives it the old college try but an open contoured 1020mb low running into a 1040mb suppressive high is like a fat guy running uphill into the wind.  he gets blown back.  And eventually gives up and turns around (and presumably goes and eats something)

At least Atlanta gets snow.

[Bob Chill]

GEFS continuing to trend towards a better look down the road. AK ridge in a much better spot than some recent runs and possibly another big cold dump out west. Reload/repeat? I'm good with that. PAC zonal fears knocked down a notch. 

[mitchnick]

GEFS continuing to trend towards a better look down the road. AK ridge in a much better spot than some recent runs and possibly another big cold dump out west. Reload/repeat? I'm good with that. PAC zonal fears knocked down a notch. 

this year is the year to forget the usual rules on lr forecasting methinks

for once in many years, the cold wants to stick around vs. run away and having the pv on our side of the Pole is likely one of the reasons (why it's there, idk)

[WxUSAF]

GEFS continuing to trend towards a better look down the road. AK ridge in a much better spot than some recent runs and possibly another big cold dump out west. Reload/repeat? I'm good with that. PAC zonal fears knocked down a notch. 

Any good analogs in here?