Network News

X My Profile
View More Activity
The new Washington
Post Weather website
Jump to CWG's
Latest Full Forecast
Outside now? Radar, temps
and more: Weather Wall
Follow us on Twitter (@capitalweather) and become a fan on Facebook
Posted at 12:50 PM ET, 11/17/2010

A look ahead toward Thanksgiving and beyond

By Wes Junker

* Wind advisory: Full Forecast | What caused last night's storms? *

Long-range models show potential (but no guarantee) for cold air to start spilling East by around Thanksgiving. This is represented on the left panel by the blue coloring over the East Coast indicating projections for a dip in the jet stream (or trough). The squiggly lines in the right panel represent the range of projections.

You may have noticed chatter about potential for wintry weather around or just after Thanksgiving at AccuWeather and on the Baltimore Sun's weather blog. It's too early to be speculating about forecast details more than a week away, but we can talk about how the large scale weather patterns are evolving and what kind of weather they might support.

Over the past couple of weeks, we've discussed La Nina and the North Atlantic Oscillation (NAO) and noted that one of the big uncertainties in this winter's forecast was how the two might interact.

The latest forecasts from the European Center model and NOAA's global model (the GFS from National Center for Environmental Prediction, NCEP) both now showing a high latitude blocking pattern developing over the North Atlantic - which signals the potential for cold air to settle into the region. Above, I've posted the NOAA's ensemble mean forecast for 216 hours from Tuesday evening which is valid Thanksgiving evening.

On the left panel, note all the red colors at the higher latitudes and the blue area to the southeast and east of Nova Scotia. That's a classic negative NAO look, a classic pattern which usually increases the chances of snow and cold in our area. However, it's only November and snow this early is relatively rare but not unheard of. You may recall snow around Thanksgiving in Washington in 1989.

The map posted above is from the same time range when the operational NCEP high resolution model from yesterday morning was predicting a possible snowstorm or wintry mix like pattern for Washington around or a little after Thanksgiving.

The low that was supposed to give us the storm has shifted westward and northward on each of the last two runs. So can anything be said about a potential snow storm?

The panel on the right in the image above, called a spaghetti plot, gives an estimate of how predictable the pattern is. The same model is run multiple times with slightly different initial conditions to see how much variation occurs. The more variation in the forecasts, the lower the confidence in being able to intelligently choose which solution might be the correct one . The closer the lines are clustered together, usually the more confident you can be of a model forecast. When all the lines look like a small child was scribbling on a piece of paper (like in this case), the forecast is highly questionable; not necessarily wrong, but it is almost impossible to draw any real inference from it.

For those interested in trying to assess the predictability of the pattern themselves you can view such products on Penn State's weather wall. As suggested by the spaghetti plot, at such long time ranges, most potential snowstorms are illusions

Then what can we say about the pattern?

Long-range models show the likelihood of a pattern supporting mild air (as shown in red) over the East Coast in the days prior to Thanksgiving.

The pattern looks more predictable through around next Tuesday. By that time we already have the negative NAO but also have a strong reverse (or negative) Pacific North American pattern (rPNA), characterized by a ridge over the eastern Pacific ocean and trough over the western U.S. (cold air) - which tends to build a ridge to the East Coast (warmer air).

While there still is quite a bit of spread in the lines on the right panel in the above image, the magnitude of the mean ridge (red area) over the Pacific and of the trough over the West indicates there is considerable agreement about these two critical components of the rPNA pattern. The same can be said about the negative NAO.

As we've discussed before, La Nina years like this one favor an rPNA pattern like the one shown above. That same article discussed the considerable impact of a negative NAO on the temperatures during a La Nina.

The rPNA and negative NAO will be battling for supremacy over our area for the next couple of weeks making longer range forecasts really tricky. How these two features wax or wane will have a significant influence on our weather.

Complicating the situation is that during fall, the PNA tends to have more of an impact on the weather over the U.S. than the NAO but as we head into December, the NAO's influence typically waxes. The southerly flow across our area and orange colors on the map above suggest that next Monday and Tuesday might give us well above normal temperatures unless there is cloud cover.

The forecast shift of the cold air east and the area of below normal heights (the blue area) eastward away from the west after Thanksgiving is predicated on the changes forecast to occur in the Pacific by both the GFS and ECMWF ensemble mean forecasts. Compare the position of the ridge/hill in the lines near Alaska (red area with the solid line) on the earlier 156 hr prediction to the two maps below showing the patterns on the evenings of Thanksgiving and Black Friday.

Ridge-trough patterns on the evenings of Thanksgiving and Black Friday. These images show a trough setting up in the East bringing cold air south towards the D.C. area.

Note that the ridge near and just south of Alaska is replaced by a trough (the dashed line). The strongest ridging (red area) develops farther west in the Pacific which forces changes to the pattern further east. The new trough (dashed line) just south of Alaska helps promote the development of another ridge or bump near the West Coast which in the models forces the blue area eastward as shown on the 216 and 240 hour forecast. This weak ridge near the West Coast helps force the cold air east.

If the strongest ridging in the Pacific does not shift westward as much as currently forecast by the various models, then the pattern over the U.S. may be slower to change than shown on any of the models.

Last night's runs of both the ECMWF and GFS were a little slower with the transition than on the previous day but both eventually bring the pattern change.

My guess is (and it's certainly not much more than a guess) is that the changes in the Pacific will occur enough to force changes over the U.S. leaving the mid-Atlantic states in a battle ground for how far south the colder than normal temperatures reach after Thanksgiving. Right now both model ensemble mean products are calling for below normal temperatures across our area. The individual ensemble members continue to show considerable spread suggesting there is little that can be said about any potential storms.

By Wes Junker  | November 17, 2010; 12:50 PM ET
Categories:  Latest, Winter Storms  
Save & Share:  Send E-mail   Facebook   Twitter   Digg   Yahoo Buzz   StumbleUpon   Technorati   Google Buzz   Previous: Fierce midnight storms of Nov. 17: what happened?
Next: PM Update: Gusty winds to subside


Thanks for the update, makes me even more unsure now.

Hope you guys can make a firm prediction as early as this weekend on what will happen next thursday...!?!?

Part of my job is salting and plowing snow around here so its hard for me to stick to my holiday weekend getaway plans with the family when I have a possible snowstorm looming around the same time, in which case I will have to cancel my plans :(

Posted by: KRUZ | November 17, 2010 1:10 PM | Report abuse


I don't think it likely that it will snow thursday as it probably will not be cold enough as if anything, it's more likely that the cold be delayed than come more quickly than forecast. I'm less sure of what might happen once the cold air gets here. I think that is where the uncertainty rises. Wes CWG

Posted by: wjunker | November 17, 2010 1:22 PM | Report abuse

The Thanksgiving 1989 snow started Wed. evening and by the next a.m. there was a nice snow cover, which stayed in place because of the cold (and resulted in Camden's being snowbound in Culpeper, according to his post a few days ago).

A month later, before Christmas, there was more snow, and the Potomac was frozen solid at the 14th St. bridges. A few days after Christmas the weather broke and pretty much put paid to that winter.

First, though, Wilmington, NC had 16" of snow and an all-time record low of 0 degrees and there were flurries as far south as Melbourne, FL. (I was at the Greenbrier at Christmas that year, sleigh riding, and having a blast in the snow and severe cold.)

I'm so ready for some Cold Turkey this year.

Posted by: JerryFloyd1 | November 17, 2010 1:25 PM | Report abuse


My wife just pointed out how badly my blog reply was written. Sorry about the grammar. I also looked at the 7Am run of the GFS which was not available when I wrote the piece. It's jumped back towards the front coming through quicker than the runs last night again emphasizing the uncertainty of the pattern and the weather even for Thanksgiving. I still think the odds are against snow in dc for Thanksgiving. However, that's just a wild guess. Wes CWG

Posted by: wjunker | November 17, 2010 1:43 PM | Report abuse

I am just waiting for the models to come around to bringing us our annual December 6th DC snow event :) Great post!

Posted by: curtmccormick | November 17, 2010 3:22 PM | Report abuse

I must say that Wes is a fine addition to the CWG team!

Posted by: iammrben1 | November 17, 2010 4:00 PM | Report abuse

Great post Mr. Junker!

However, looking at the index, the AO is also supposed to tank around this time period as well. I know that whenever the NAO and AO are both negative at the same time, our chances of seeing snow greatly increases.

So, I was wondering why you didn't mention the AO in your post. Does it not play a significant role in this situation? Thanks!

Posted by: Yellowboy | November 17, 2010 4:13 PM | Report abuse

Agreed iammrben1; Wes is a fine addition and this was a well written and interesting piece. I'll continue to follow the gang through t-day.

Posted by: parksndc | November 17, 2010 4:38 PM | Report abuse


The reason I didn't also mention the AO is that the NAO and AO are so closely related that I didn't think it was necessary and in our neck of the woods, the more important of the two is the NAO. If you go back to the second post about last year's winter you can look at the pressure correlations for the two indexes and see how similar they look. The NAO can lock in confluence across the northeast and at times the AO won't. That is probably why I consider the NAO the more important feature. You can reread about the AO and NAO below

Posted by: wjunker | November 17, 2010 5:22 PM | Report abuse

is it too early to do a snow dance?

Posted by: bachaney | November 17, 2010 6:11 PM | Report abuse

Yes. But you can keep your fingers crossed!

Posted by: pjdunn1 | November 17, 2010 6:50 PM | Report abuse


Thanks so much for all the info! I'll keep checking CWG for any updates! But I'll be keeping my fingers crossed for a snow-free Thanksgiving day, black friday and throughout the weekend.

I really appreciate the quick responses and look forward to more info from everyone at CWG!

Posted by: KRUZ | November 17, 2010 7:01 PM | Report abuse


The 18Z gfs (comes out at around 2:00PM took the potential storm for Thanksgiving back north and west. There is still uncertainty but my guess still is that the storm will not give us snow in the dc area Thursday or Friday, but again, that's a guess. The CWG guys will try to stay on top of it.

Posted by: wjunker | November 17, 2010 7:26 PM | Report abuse

odds are we wont know for sure until we wake up thanksgiving morning, that dos int stop me from doing the snow dance.
o and when does the next model run come out?

Posted by: snowlover31 | November 17, 2010 7:42 PM | Report abuse

The GFS model runs and the NAM model runs are posted every 6 hours. The 12 midnight (00z) is posted sometime between 10PM and midnight.

Posted by: pjdunn1 | November 17, 2010 8:09 PM | Report abuse

WES back on NOV 14 -- 3DAYS AGO -- I made the VERY SAME POINT you made above ... in the weekly audio THIS WEEK IN WEATHER NOV 14 EDITION

Posted by: wxrisk | November 17, 2010 8:45 PM | Report abuse

the problem is that the usual group of snow mongerers (JB et al) are out there saying that the Low that develops over the Ohio valley on NOV 25-26 ...COULD track further south because of the Block/ -NAO.

Let me explain why that sort of reasoning is well silly and really BAD forecasting.

First this isnt Jan or feb. We are In a MILD pattern and the SEridge is BUILDING as the trough over the western US comes into the west coast / Rockies.

Second we need the cold air in Place FIRST. This isnt a marginal situation at all. Check the weather Books... I am pretty sure it has to be "cold" for it to snow.

Third ALL the Models today have the Main Low MUCH further N and W on 11/25-26. The 12z euro has the Low over ORD at 989 as does the CMC and now the 18z GFS in well to the N and W.

This means ahead of the Low... we will have SOUTH winds and serious WAA (warm air advection) and RAIN/ storms. Even for Most of PA NYC and southern New England this looks to be a rain event.

Posted by: wxrisk | November 17, 2010 8:56 PM | Report abuse


I didn't decide to write anything on the situation until last night because to me any forecast beyond day 10 is just smoke and mirrors. Then Jason mentioned there were blogs mentioning a potential storm so I thought I should write something.

Posted by: wjunker | November 17, 2010 9:11 PM | Report abuse

OK, I know it is too early to know for sure, BUT I am trying to learn to read all of this accurately, and I am looking south as I am headed there next week- from what I see here, IF the trough settles in, the cold air will not penetrate into the midline of Florida.

And, BTW, it's never too early to do a snow dance - I've been doing them intermittently since July.

Posted by: Snowlover2 | November 17, 2010 9:40 PM | Report abuse

Wes, GREAT post. It hits the nail on the head of how to garner, interpret and convey the maximum information from model runs, especially in addressing the varying, situation dependent levels of uncertainty derived from ensembles.

I recall you being one of the first true believers amongst forecasters at NCEP on the merits and value of ensemble prediction back in the 90s. At that time most were reluctant at best to trade in the "deterministic" approach to forecasting {"model of the day" and single best forecast} for recognizing the importance of dealing with the unavoidable uncertainties in ALL forecasts

Posted by: SteveT-CapitalWeatherGang | November 17, 2010 10:41 PM | Report abuse


Thats cool... I hope my posts where didnt come across as a case of my bitching that you " stole my thunder " etc etc...

That is why I made the THIS WEEK IN WEATHER post when I did... if you get the chance take a listen.

I heard the same MONGERING as you did... "the trough all NOV will be in the east" .

I know if YOU and I are on the same page then its a pretty darn good forecast .

Posted by: wxrisk | November 17, 2010 10:43 PM | Report abuse

The "magic" snow date is Dec. 5th, with significant snows on something like 4 out of the past 6 winters on that date; incredible! White Thanksgivings are even rarer than White Christmases in DC; I can only remember one (1989) in the past 50 years or so! In any event, Wes and I should be golfing early next week; fore!

Posted by: buzzburek | November 17, 2010 11:49 PM | Report abuse


Thanks, I enjoyed writing it and was thinking of you some as I wrote it.


It does look like some good golf weather


you came across fine.


Posted by: wjunker | November 18, 2010 8:38 AM | Report abuse

Thanks for this posting.

I'm curious about if/how you see this particular event relating to the warm arctic/cold continent weather pattern [that is linked backward to loss of sea ice and linked forward to negative AO/NAO] recently identified by NOAA.

Discussed here:

and here:

and related research discussed here:


Hunter Cutting

Posted by: huntercutting | November 18, 2010 7:36 PM | Report abuse


Let me preface by saying I'm not a climatologist so I may not be the best person to judge the ideas mentioned in those papers. That said, I think it dangerous to ascribe last winter's negative AO to any one cause.

There were other things besides the melting sea ice that have been shown to correlate to a negative AO. One is the solar cycle, if you do a simple plot of solar flux and correlate it to the pressure pattern you end up with a negative NAO look when the soalr flux is low like last year. The correlations are not tremendously high but do exist. The same holds for the easterly phase of the Quasi-biennial oscillation. Work by Baldwin an Dunkerton have shown that the QBO has an impact on the northern annular mode and arctic oscillation (see link below). They have also argued that stratospheric warming events can weaken the vortex and aid in developing a negative AO. While it's controversial to ascribe last years negative AO to the stratopheric warming, we had two such stratospheric warming events and during both the AO tanked. Last year the Brewer Dobson circulation seemed to be really strong with lots of ozone at the higher latitudes which was probably related to the QBO but that two suggested the potential for the AO to tank. I guess I'm saying there are lots of factors that may have contributed to the development of the AO and we don't completely understand what drives the different oscillations of the AO or NAO. Last years negative AO and NAO might also be due partly to serendipity.

There is also research that suggests that the fall snow cover over the NH can help produce a negative AO in winter. Heck Cohen has authored an article on that as well as at least one article on how the Madden Julian Oscillation could play a role in changes in the NAO from positive to negative.

Posted by: wjunker | November 18, 2010 10:47 PM | Report abuse

The comments to this entry are closed.

RSS Feed
Subscribe to The Post

© 2012 The Washington Post Company