Long Range Lookout: Brutally Cold December Continues into January

New data I've compiled over the past few weeks leads me to believe that the end half of December and on into January will be filled with some pretty cold weather.

This data came as a result of analyzing the stratosphere, and determining the placement of temperature anomalies. Using a 2-4 week gap between warming and cooling air masses in the stratosphere, something I like to call the Stratospheric Temperature Index (STI) was created. 

For some verification, take a look at this text from a post I made on December 1st:

Between October 30 and the first few days of November, the Central and East US experienced above normal temperatures in the upper stratosphere. This is displayed well in the Climate Prediction Center's reanalysis of the 30th in the image above. These few days between the end of October and the start of November also included slightly above normal stratospheric temperatures diving south from Canada, with below normal temperatures observed in the Pacific Northwest. Because the effects of stratospheric warmings are typically seen 2-4 weeks after the warming event, we took a look at surface temperatures from November 14-28 to see how well the surface temperatures matched up with the upper stratospheric temperature anomalies (note that this is a negative correlation relationship, meaning above-normal stratospheric temperatures will lead to cold surface temperatures, and vice versa).

Ground-level temperature verification from November 14th to November 28th, approximately 2-4 weeks after the height of this US upper stratospheric warming event, is nearly identical to the temperature anomalies experienced in the stratosphere. The regions that saw above normal stratospheric temperatures, including the Central and East US, as well as Canada, saw below normal temperatures during this timeframe. Similarly, the Pacific Northwest, which was bathed in slightly below normal upper stratospheric temperatures to kick off November, was included in slightly above normal surface temperatures from November 14th to November 28th. It would seem to me that the stratosphere had a direct involvement with these surface temperature anomalies, given how well the stratospheric and surface temperature anomalies line up. (End Dec. 1 post.)

Get the gist? Above normal temperature anomalies tend to lead to below normal temperature anomalies, and vice versa. So now, let's take a look at what's happening here, with the most recent verification.

This image from the 10 millibar region in the stratosphere shows significant below normal temperature anomalies across much of Asia into the Northern Pacific, with above normal temperature anomalies stretching from the United States into Western Europe, with another swath of warm temperatures in East Asia on November 28th, 2013. Let's see how surface temperatures for December 7-10 are showing up, as the 2-4 week verification window from November 28th of December 14-28 is upon us.

It's no coincidence that we saw strong above normal temperatures across the North Pacific in response to those below normal stratospheric anomalies in late November. The warm stratosphere in the United States shows up as below normal temperature anomalies, fitting in to the STI's definition. Also, note how Western Europe shows up as below normal, and even a portion of East Asia getting in on the cold weather. Can you see how the index is verifying?

Now, take a look at this image of upper stratospheric temperature anomalies for December 7th. We see a strong body of above normal anomalies in the United States, as well as Central/Eastern Europe into East Asia. Significantly below normal anomalies are observed across Canada, the North Pacific and far northeast Asia. 

If we take that verification window from the late November warming (December 14-28), we can expect that timeframe to be colder than normal in the US. With the current warming intensifying in the image immediately above, we can now expect continued cold weather in the December 21-January 4. Depending on how long the warming in the United States lasts, we could continue to see cold weather well into January. It is tough to tell the intensity of the cold, but based on verification stats, it appears clear that the United States should see at least some cold weather.

Andrew

December 19-23 Potentially Significant Winter Storm

I am rather confident that we will see a winter storm in time for the December 19-23 period, and there is a chance it could be significant.

A pretty strong storm system is expected to move across Japan in 48 hours, after dumping multiple storm systems over that area in the past few days. This storm system is looking healthy, as defined by the GFS ensemble mid-level height anomaly forecast above over Japan. As I have mentioned on this blog, there is a 6-10 day correlation between weather that occurs in East Asia, and weather that occurs in the United States. Thus, with this forecast image, we arrive at the December 13th date. Go ahead 6-10 days, and we find the potential storm timeframe of December 19th to 23rd. The reason I believe it may be significant is due to the strength of this system on the image above. Unlike the upcoming weekend storm, which saw only a moderate system in East Asia 6-10 days prior, the current East Asian projection is for a strong storm to hit Japan, and long range model guidance senses that this system will also be a strong one for the US.

The ECMWF model delivers a very strong system to the Midwest, Great Lakes and Ohio Valley before shifting northeast to hammer Canada. As is denoted above, there could be some heavy snows north of that dark blue 32 degree temperature line. This would put the states of Illinois, Indiana, Michigan and Ohio in line for such snows, though we remain roughly 10 days away from this potential storm. I'll discuss a little more on the atmosphere surrounding this system later on in this post, but for now, you snow fans can take a look at the total snowfall projection for this storm below.

The ECMWF isn't the only model showing a potential winter storm for this timeframe. It's global model rival, the GFS, is also in on it.

GFS precipitation forecast for December 20th, 12 hours earlier than the ECMWF image above.

The GFS image shown above details 12 hour precipitation totals for this storm, and you can clearly see the wide swath of potential heavy snowfall across Illinois, Wisconsin, Michigan and Indiana. The GFS model has the low pressure system in Kentucky 12 hours after this image, whereas the ECMWF has the system in central Indiana. The GFS being on a more southern track does mean the snowfall would theoretically be suppressed south, but the wide precipitation shield on the image directly above gives away the GFS' projected snowfall, which is shown below.

GFS 48 hour snowfall totals for this snowstorm.

My thoughts on this storm system are actually of rather high confidence right now. This strong Japan storm system is more likely than not to translate to a storm system (if not a strong storm system) for the United States. Once it enters the United States, it is most likely going to drop down into the eastern Rockies (think Colorado).

This relative measure of predictability map, from the NCEP, gives us a glimpse at what atmospheric conditions are likely to be unfolding by an assortment of colors, defined in likelihood by the legend at the bottom. In this case, we see that blues have a single or low double-digit percent chance of verifying, while oranges define close to a 50% chance of verification. This map is valid on the evening of December 19th, roughly two days before this storm is projected to strike.

If we look around the map, we first notice a wide swath of oranges across the Gulf of Alaska. A glance at height contours tells us there is ridging forecasted in that area for the evening of December 19th, and the presence of widespread oranges tells us that the ensembles are at least decently confident in that atmospheric factor verifying. The effect that Gulf of Alaska ridge has on the potential winter storm is that the system should follow the contour lines and drop south into the Rockies. It is unable to drop too far south into the Southwest, as a positive North Atlantic Oscillation (NAO) will keep the pattern more progressive, and thus more west-to-east (hence the relatively flat contour lines over the United States). The second item we observe is a stripe of orange extended from southwest Canada into the central and western Rockies. This looks to be the potential storm system, as it dives southward towards the Rockies. The oranges tell us there is rather high confidence, which is a good sign for the storm prospects. The third (and possibly most important) item is the swath of oranges over the Southeast. It may not look too impressive, but there is likely to be high pressure stationed over the Southeast as a result of storminess over the West.

The image above is what is referred to as a negative Pacific-North American index phase (or -PNA for short). The negative PNA results from that ridge we discussed in the Gulf of Alaska that then produces storminess/troughiness in the Western US. In response to that West US storminess, we then see the development of high pressure in the Southeast. The jet stream in this situation takes storms from the West and runs them north and east, typically delivering snow to the same areas that were outlined by the GFS and ECMWF earlier in this post. Considering the relative measure of predictability chart has high confidence (for being nearly 10 days out) in the Gulf of Alaska ridge, the storm itself entering the Rockies, and the ridge in the Southeast, I do think that the GFS and ECMWF solutions are two very possible solutions.

Let me explain why I think we could see the GFS or ECMWF work out in terms of storm track and snowfall amounts.

1. The East Asian storm system is strong. Typically, we see weak or moderate storms in Japan translate to weak or moderate storms in the US. Now that a strong system is headed for the area, it is quite possible that this winter storm is strong as well.

2. Models appear rather confident with the environment surrounding this storm system. This negative PNA environment is looking like its gaining confidence in the GFS ensemble set, and a look at the ECMWF ensemble prediction system seems to agree on both the set up and the storm system actually occurring.

3. All put together, the environment looks ripe. This one is mainly thanks to the East Asian component, but when you look at the overall picture of the mid-levels, a storm track favoring the areas depicted by both modeling systems does appear to be a viable solution on either guidance system.

It's a bit too far out for me to give a graphical outlook, but I think you get the gist here- there is some mighty fine potential with this system.

Andrew

December 13-16 Potentially Significant Winter Storm

It's looking very likely that there will be a snowstorm for the nation between December 13 and 16, and there is a decent possibility that it will be significant.

The operational ECMWF model projects the event to be a result of two separate pieces of energy- one from Canada, and one from the Southwest. You can see the Canadian energy shown as the white streak pointing southwest-ward into Kansas, and the Southwestern energy depicted as the dip in the contour lines in far southern Missouri. It is expected that these two systems will bring the potential for anything from a winter storm, to a monster of a winter storm. But that isn't shown here because of some disagreements.

Model guidance is in disagreement on the track of this system, as well as the question to if it will phase (merge). The idea here is that if the two pieces of energy are able to emerge from Canada and the Southwest, respectively, at the right times, then they may combine to form a potentially significant winter storm system. However, some model guidance prefers to either phase the systems further to the east, giving the northern Ohio Valley some very heavy snows, or not even phase them at all. I'm not sold on this non-phasing idea, and to be honest, I think the systems will actually phase further west than what is being shown. Why? I'm glad you asked.

Teleconnection projections for the Pacific-North American index (top left), the North Atlantic Oscillation (top right), the West Pacific Oscillation (bottom left), and the East Pacific Oscillation (bottom right) are shown in the image above, composed by the Earth System Research Laboratory, or ESRL. Taking a look at the PNA forecast for the December 13-16 timeframe, we see the index moving towards positive territory, but not really getting there in time to flip the Pacific North American index from negative (unfavorable for a huge East US snowstorm) to positive (favorable for a East US snowstorm).

NCSU

Now, the negative PNA means that high pressure will be present off of the West Coast, so the mid-latitude jet stream will be able to dive south to deliver storms to the Rockies. An additional jet stream system (you weather junkies know it as the subtropical jet stream) that tends to stay around the Deep South is pushed northward due to the emergence of high pressure in the Southeast. If you look towards the middle of the nation, we see those two jet stream systems pairing up and moving northeast. You guessed right; such a development would support the earlier phasing (or at least some form of phasing) idea I illustrated earlier. But we've only scratched the surface.

Going back to the ESRL teleconnection forecast, we see that the NAO is projected to be positive throughout this timeframe. This positive NAO will keep the overall weather pattern quite progressive, meaning the probability of stagnating high and/or low pressure systems is lowered. Usually, this would go against the idea of the two pieces of energy phasing, but if the progressive flow can limit the ridging in the Southeast provoked by the negative PNA, it's possible we see a very ideal set-up for the phasing of these two storm systems.

There is also the idea from model guidance that a secondary low pressure system may form over the waters off the Northeast to produce a very snowy situation for much of the Northeast. I am skeptical of this idea for a few reasons, one of the more pronounced ones being the presence of that positive NAO. Usually, a positive North Atlantic Oscillation is progressive and does not support the formation of these coastal Nor'easters. Adding to my skepticism is how models are doing really poorly with the conditions in the Pacific right now, where the energy for the main storm and possible secondary low is. Thus, model guidance for both events isn't really being viewed as entirely trustworthy. I'd like to wait for the models to work themselves out on this initial storm first before we open the idea of a secondary low forming in the Northeast. I'm not completely shutting the door on the idea, but I'm also not exactly in favor of it just yet.

Adding to my thinking that these two pieces of energy may very well phase more east than is currently projected is because of a very prevalent model bias among global modeling systems.

It has been determined by meteorologists who carefully analyze and work with the models that there is a tendency for models to eject energy coming out of the Southwest too quickly than how it actually ends up unfolding. This means that it's very possible (some would even say likely) that the two pieces of energy I pointed out at the top of this post may look very different- the energy coming from the Southwest may end up further west than the current solution shows for the timeframe above, which is early on Saturday morning. I do think that the models are displaying this bias yet again, and the trouble in the Pacific doesn't help a bit, as I'll explain below. If we do see a slowing down in the model solutions with that Southwestern energy, expect the chances of a phasing solution to be on the rise, and the possibility of an even bigger storm rises as a result.

The issue with the Pacific is that we don't have nearly the amount of sampling in the Pacific like we do in the United States. What's sampling? It's what happens when the National Weather Service offices send up the big weather balloons twice a day, with monitors attached to them. Unfortunately, we don't have any of these NWS offices in the middle of the Pacific- we only have buoys, commercial aircraft, and satellite measurements to give us a rather mediocre depiction of what's happening in the Pacific, as is explained by Cliff Mass in his blog post. I say mediocre, because while those buoys, satellites and commercial aircraft do help in giving the models some data to work with, they are nothing compared to the radiosondes that go up with the weather balloons. That said, I expect models to be dicey with this storm for the next couple of days, before the energy gets into the network of radiosondes, which should greatly help model accuracy.

Speaking of model accuracy, take a gander at this discussion of the models from the Weather Prediction Center on this storm:

...PROGRESSIVE TROUGH ARRIVING ON THE WEST COAST DAY 2...
PREFERENCE: BLEND OF 12Z CANADIAN WITH 00Z ECMWF ENSEMBLE MEAN
CONFIDENCE: BELOW AVERAGE

A PROGRESSIVE TROUGH OF SHORT TO MEDIUM WAVELENGTH WILL ARRIVE
ALONG THE WA/OR/NORTHERN CA COASTS BY 13/00Z. THE CHARACTER OF
THIS SYSTEM...COMBINED WITH ITS INITIAL POSITION OVER THE DATA
VOID IN THE PACIFIC...IS CAUSING SOME DIFFICULTY IN THE MODELS.
THE GFS AND NAM ARE CERTAINLY FLATTER...GIVING THEIR SOLUTIONS A
LITTLE EARLIER ARRIVAL TIME ON THE COAST. THEY ALSO TAKE HEIGHT
FALLS EASTWARD INTO UTAH WITH LESSER HEIGHT FALLS GRAZING NORTHERN
ARIZONA. MEANWHILE...THE ECMWF AND CANADIAN INDICATE A MUCH MORE
PRONOUNCED SPLIT OF THE UPPER ENERGY BEGINNING LATE ON DAY
2...ALLOWING A ROBUST SHORTWAVE TO DIG FARTHER SOUTH. SOLUTIONS
FROM THE CANADIAN ARE RELATIVELY CONSISTENT...AND DIG THIS TROUGH
TOWARD WESTERN ARIZONA...AN IDEA THAT HAS SUPPORT FROM SEVERAL OF
THE 00Z ECMWF ENSEMBLE MEMBERS. THE OPERATIONAL ECMWF SHOWS MORE
RUN TO RUN VARIABILITY...THOUGH...AND HAS BEEN ONE OF THE
DEEPEST/FARTHER WEST SOLUTIONS. THE 12Z RUN...IN FACT...CLOSES A
CIRCULATION BACK OFFSHORE OF SOUTHERN CALIFORNIA BY 14/00Z. ONE
COULD SAY THAT THE 12Z UKMET SUPPORTS THE ECMWF...BUT ONLY
LOOSELY...WITH THE UK SOLUTION BEING EVEN FARTHER WEST AND COMING
ON THE HEELS OF A 00Z RUN THAT WAS LESS THAN INSPIRING/WAS A SLOW
OUTLIER/. GIVEN THE AMOUNT OF UNCERTAINTY...WE RECOMMEND A BLEND
THAT GETS TOWARD A MIDDLE GROUND SOLUTION. IN OUR FINAL PREFERENCE
WE REPLACED THE GFS WITH THE CANADIAN...AS THE GFS NOW APPEARS TOO
FLAT/PROGRESSIVE VERSUS THE FULL 12Z SUITE.

The consensus does appear to be that models are really having trouble here, and the WPC recommends going with today's GEM (Canadian) model, as well as last night's ECMWF ensemble grouping. Because I'm always a bit suspicious of the GEM model, we'll take a look at what last night's 0z ECMWF ensembles say.

The 0z ECMWF Ensemble set has this storm system traveling from Arkansas to Ohio, as I outlined with the yellow arrow. Such a set-up is very conducive to heavy snows in the Lower Great Lakes, Midwest and Ohio Valley, though the ensembles only drop a handful of inches of snow from Illinois to Indiana before the system really gets cranking further east. Again, that southwest bias is still in play here, so I'm skeptical about this solution (and all others) as well.

Here's my preliminary outlook on this system, subject to potentially drastic change.

Andrew

December 13-16 Potential Winter Storm

There are hints that there may be another winter storm across the Midwest and Northeast through the December 13-16 timeframe.

The GFS model has a storm system moving north and east across the lower Midwest and towards the Northeast, as is shown by the elevated vorticity values over the Kentucky, Illinois, Indiana and into Tennessee. The GFS prefers to develop this storm system, and moves it northeast over the ridging in the Southeast, giving the Midwest, Lower Great Lakes, Ohio Valley and Northeast another round of snow. Given that the pattern does support this storm track, I do think that this is within the realm of possibility.

The ECMWF, at the same timeframe as the GFS image above, has this energy more elongated than the GFS, and it is not developed correctly as a result. Both model systems have ridging over the West Coast, as well as ridging in the Southeast, which is why I find this solution rather peculiar. I think the ECMWF may not be correct in its analysis of the system. Let's discuss why below.

For one, the ECMWF ensemble set (ECMWF EPS) follows the same track (drawn in yellow) that the 12z GFS put up earlier today. Both model guidance sets have the storm system traveling from northern Mississippi to West Virginia. Looking at the deviation on the image above, it does look like the ensemble members agree with the GFS, with dark blues rather than light blues (which would indicate low confidence in the solution) showing up for this potential storm. If we are to go by the idea that the ECMWF EPS and GFS are pairing up for this potential winter storm, we can take a look at the overall storm projection from the GFS and see what both guidance systems may be thinking:

Note, this storm would bring snows to the northern Northeast regions, not coastal states. Subtract an inch or two from western New York, VT, NH, ME, northwestern PA to get snow from only this storm. 

As the GFS projects above, this storm would bring plowable snows from much of Illinois to Indiana, Michigan, Ohio, Pennsylvania, New York, Vermont, New Hampshire, and Maine. I don't have the figures on the ECMWF ensemble snowfall (not sure if anyone has access to that), but based on the storm track from the GFS and ECMWF EPS matching up so well, I would think that snowfall orientation would be the same, stretching from the Midwest to the Northeast.

I'd like to wait for a bit more consistency before putting in more confidence to this system, which would include the GFS and ECMWF EPS matching up again on this system, and even the ECMWF operational model jumping on board.

Andrew

Sudden Stratospheric Warming Incoming

It does appear likely that a stratospheric warming event is in the works for the next week or so.

An animation of 30 millibar temperature anomalies high up in the stratosphere suggest above normal anomalies in eastern Asia, combined with a shifting body of warmth to the east from Europe will provide a base for what is expected to be a sudden stratospheric warming event. As both masses of warmth in northern Asia and southeast Asia intensify, it is expected that they will combine, and possibly push northeast into the Bering Sea to provide a base for this SSW.

Mountain torque (MT) values around the 40N parallel are beginning to spike in coordination with this warmth in the upper stratosphere. Without getting too technical, it has been seen that above normal mountain torque anomalies come around when a sudden stratospheric warming event is immenint. I circled the spike in anomalies in black on the top panel to better highlight it. Taking a look back at the animation at the top of the post, we see this warming event is beginning around the 40N parallel, so it is no wonder we are seeing MT values on the rise in the same parallel region.

The ECMWF model projects this event to reach its peak in about five days, when the maximum strength of the Wave 2 event occurs. 

Now, there are two main types of stratospheric events that disturb or split the polar vortex. There is a Wave-1 response, which involves the polar vortex becoming elongated and/or weakened, however a split does not occur. This can result in displacement of the vortex out of the Arctic, but that prospect is not as likely as it is in a Wave-2 scenario. A Wave-2 stratospheric response involves the polar vortex being split into two main vortices, as the image from NASA shows above. While the temperature images show the Wave-1 and Wave-2 (top and bottom rows, respectfully) responses occur over Greenland and Eurasia, especially with the two split vortices where one vortex goes over Greenland and one goes into Eurasia, the split can lead the vortices into any land mass, not just those two regions. 
We can deduce a few things from that ECMWF forecast image. For one, it appears that the event will be enhanced mainly over the upper stratosphere, and with an unfavorable Quasi-Biennial Oscillation, I doubt the polar vortex will split from this anticipated SSW alone. Second, it looks like the event is projected to occur at the 60N parallel, which is located in Canada and northern Asia. This, if it were to come true, would indicate the sudden stratospheric warming event would have shifted north and impacted the upper latitudes. I don't see the warming event protruding into the 90N parallel, which is nearly the highest possible latitude parallel in the northern hemisphere. This means that the SSW shouldn't slice and dice the polar vortex, which is what we're expecting (not the polar vortex split, but a lack thereof).

As for what the future may hold for the stratosphere. 100 millibar eddy heat flux values have risen to normal levels after a nearly record-setting below normal stint in late November. Whether the rising heat flux continues remains in question, but such a trend would bode very well for the stratosphere, particularly later on in winter when the Quasi-Biennial Oscillation becomes more favorable for sudden stratospheric warmings.

Andrew