Monday, December 16, 2013

December 20-23 Significant Winter Storm - First Outlook & Discussion

This is my first preliminary outlook for what is shaping up to be a significant winter storm over the upcoming weekend for a large swath of the country. A model discussion is included in today's post, and that's what we will begin with.

We're going to start off with the ECMWF model's projection, since it is the one that seems to be setting the trend for this event. In the image above, valid for Sunday morning, we see a strong low pressure system emerging over the southern Midwest and Ohio Valley. This chart from Meteocentre calculates the minimum central pressure to be at 996 millibars, making this a pretty strong storm. The intensity is no surprise, as I've been illustrating since December 11th that a very strong storm over East Asia will provide a base for a strong storm in the United States. The ECMWF has this system follow a track extending from the Southern Plains through southeast Missouri before continuing on through the Ohio Valley and eventually leaving the United States. I believe we will eventually see something very similar to this track, maybe shifted a bit north or south by the time the storm arrives. My reasoning?

The Earth System Research Laboratory (ESRL) puts out almost-daily teleconnection forecasts, and I have screenshotted two of them above. We see the forecast for the Pacific-North American index (PNA) on the left, with the North Atlantic Oscillation (NAO) forecast on the right. Matching up the index with the date on the bottom legends of each panel, we see that the PNA is projected to be negative (albeit rising) during this timeframe, with a strong positive NAO.



During a negative NAO, we tend to see big high pressure form across the Gulf of Alaska. Because we have a big rise in height anomalies over that body of water, there is a response of a dropping in height anomalies over the West Coast, meaning stormy weather forms in that region, as the graphic above from NCSU illustrates. Thanks to that stormy weather in the West, high pressure likes to form in the Southeast, and the two arrows (which illustrate the wind pattern in the area) combine over the Plains to show the storm track of systems during a negative PNA. Now usually, negative PNA storms like to go up into the Plains because the Southeast ridge is usually very strong. In this case, however, the positive NAO helps to keep the overall atmospheric pattern very progressive, meaning we don't see any persistent high or low pressure systems in any given area. As a result, that Southeast ridge is not given the opportunity to develop into a massive high pressure system. So, with this particular storm, I expect the system to move north due to that Southeast ridge (which will be slightly suppressed), and then take a track favoring snowfall from the east-central Plains through the Midwest and Lower Great Lakes and into a portion of the Ohio Valley, possibly even into the Northeast (though that is TBD).


The image above, from WeatherBell, shows mean sea level pressure anomaly forecasts from the ECMWF ensemble prediction system (EPS) across the nation for 144 hours, or 6 days out. We can clearly see the low pressure system in the Ohio Valley, and this is something the ECMWF Ensemble system has been hinting at for a while now. It has preferred this sort of track, and although I do not have snowfall figures from the ensemble mean, the similarity in track to the aforementioned 12z ECMWF run would tell me snowfall figures should be relatively similar. (For anyone interested, the ECMWF EPS Control snowfall forecast is directly below)




And now, for the moment you've all been waiting for, here is that 12z ECMWF snowfall forecast, using 10:1 ratios (where 10 inches of snow equals 1 inch of liquid)

Do I believe that this snowfall forecast will work out? Well, it's hard to tell. On one hand, the ECMWF does have a pretty good handle on this event to me in terms of consistency (compared to other models), and its idea for a significant snowfall jives well with the East Asian correlation saying that this will indeed be a significant system. On the other hand, we're still a ways away, meaning forecast models will most certainly change- the big question is if they change significantly, or just a bit. Based on the atmospheric set-up associated with this storm, I would think that this solution of snowfall placement would be within the realm of possibility.


Another possibility is that the storm actually goes further north, giving heavy snow to southern Wisconsin and a stripe of Michigan, as the most recent GEM model shows above. I believe this is a possibility, considering there shouldn't be a huge high pressure system in Canada to really suppress this storm to the south, but going against this idea would be the lack of a significant ridge in the Southeast to push it north, especially with the weakening -PNA (which really propels that Southeast ridge). In order to consider this solution a bit more, I would want to see more consistency with the GEM, as well as agreement from other modeling systems. However, judging by the overall pattern, as well as some model bias' I will discuss next, I'm not so sure that it's a real possibility.

Model guidance can be great, but it's not perfect. There are known bias' that have been outlined by various weather agencies, specifically the NCEP agency here in the US. These bias' should be able to help us determine what may happen that the models aren't clueing in on yet.

1. Tendency for models to hold energy in the Southwest for too long.
This is a rather well-known model bias, in that most guidance systems will end up keeping storm systems in the Southwest US for longer than they actually end up there. This affects our situation specifically in that solutions that are going for a further south projection for this system (see the GFS) may be incorrect. Should the energy eject from the Southwest quicker, it would seem to me that a more northern solution would be probable. I'm not exactly sure if that has been shed off by models like the ECMWF, GEM and ECMWF EPS yet, though it's just something to keep in mind.

2. Some models may close off low pressure systems too readily.
This is a model bias that really affects us here. Some medium range model guidance systems have the tendency to close off a low pressure system, meaning it will be cut off from the jet stream and just meander around until it is forced away by another low or high pressure system. In this case, model guidance generally agrees that our system will be closed off when it moves into the Southwest, and that's when model discrepancies begin to arise. Some guidance systems prefer to keep the system closed off for a while, which then results in a further south storm track, whereas other model systems don't keep the system closed off for so long, resulting in a northward track. In this case, I believe the GFS has kind of shed that bias it previously showed last week, and as a result we did see (and will probably continue to see) a northward movement in the storm track. As we move closer to the event, bear in mind these bias' become less and less involved in the models as they are shed away.

3. Arctic air will plunge southward quicker than models may indicate.
This is one that may or may not affect this system, as it's pretty apparent there will be precipitation mixing issues regardless of if this bias is in play. The NCEP indicates that medium and short range models will push cold air south to the lee of the Rockies slower than how it actually happens. In this case, one could argue that the cold air on the models is being underestimated, but as the cold air source in the Pacific (you weather junkies know it as the weakening negative East Pacific Oscillation) weakens during this storm system, I'm hesitant to count this bias as one to watch.

When I account for all the bias', the model trends, and some things I've been looking at over the past few days, I come up with this preliminary outlook.

To be quite honest, I'm feeling pretty good with the ECMWF operational and ECMWF EPS agreement. I feel that they do have the best handle on this system and the atmospheric pattern surrounding it, so my forecast was modeled after it. It should be noted that I did extend the significant snow area a bit north to account for any possible northward trends in this system. It should also be noted that there will be a sharp cutoff with snow amounts on both the northern and southern fringes, regardless of its track.

Rather than ask how your location looks, please look at the map- there really isn't any additional information I can provide you if you ask for a certain location.

Andrew