Tuesday, December 11, 2012

December 18-21 Potential Significant Winter Storm (Updated 12/11)

The potential is rising for a major winter storm to affect the Central and East US. Let's see how this could all unfold.


This is the Hour 168 forecast from today's 12z (6:00 AM) run of the National Weather Service's GFS model. This is the sea level pressure (white lines) and 500 millibar level (colors) forecast for the morning of December 18th. In this forecast, we see a large blue circle in the southern Plains, as well as a low pressure system near the Gulf Coast. The aligning of white lines (isobars) in a circular formation around the 'L' near the Gulf Coast, as well as the blue area in the South Plains suggests the presence of a storm system. This does look to be the storm system we are watching for this time period.

Going ahead a bit to the morning of December 19th, we now see that our storm system is predicted to move due east and become a strong coastal storm in an event that would be all too familiar to Superstorm Sandy, which I will discuss further later on. In a few words, I don't believe this forecast. The atmospheric pattern is much more conductive for a storm track either through the Plains or into the Ohio Valley/Midwest, the latter of which I think is more likely.

We now move on to the European weather service's ECMWF model, known as the most reliable forecasting model man has made to date, although recently this model has been slipping. The same rules apply for colors and white lines. This is the forecast from the ECMWF, valid on the morning of the 18th. Here, we see our low pressure system in the South Plains just like the GFS. Pay attention to what happens in this next forecast frame for the morning of the 24th.
The ECMWF takes this storm into the Southeast and off the coast! The reason I am so shocked is because this is a huge change from what the ECMWF was previously showing. Before, we had multiple forecasts of this system going into the Ohio Valley and Midwest, bringing rain and possibly good snows to those areas. When I come to situations like these, where the models suddenly change their forecast, I say 'Look at the trends'. So, I don't think I'll take this forecast too seriously unless tonight's new ECMWF forecast shows a similar idea. I will describe more why I don't think this forecast is correct further down.

Now here's where things get serious. This is the forecast from the 12z ECMWF ensemble prediction system (also known as the ECMWF EPS). To think of this the right way, imagine that forecast of the ECMWF above, but now there are 51 of those forecasts, combined into one, and that's the ECMWF EPS. Given, each different forecast has at least some change in their parameters so there is a difference from the actual ECMWF model, but when it comes to the ECMWF EPS, there's no model that can match that skill, so take that into account when seeing this forecast.
The ECMWF EPS starts out the morning of December 18th with our storm system in the southern Midwest, rather than the South Plains. Additionally, the system is not as strong as the ECMWF/GFS are showing. Why is this significant? Because if the ECMWF EPS goes up against the ECMWF/GFS, who do you think will win: 2 models with very high skill, or 51 models with the same high skill?


We now move ahead to the morning of December 19th. We now see the ECMWF EPS has taken this system into the waters just offshore the New England/Mid Atlantic area. It is stronger than the past forecast image. To me, this seems like the more obvious solution, because this is essentially what the ECMWF model itself was showing for quite a while before today's sudden change.

Now, let's discuss just why I'm so confident in the ECMWF EPS, and am disregarding the ECMWF and GFS.

(Originally posted Dec. 9, NEW INFORMATION ADDED on 12/11) The Lezak Recurring Cycle, or LRC is involved in this situation. Using the 53 day LRC timetable, I traced back December 18th to October 28th, and here is the 500mb pattern observed that day.


Look at that! We see a strong system in the Ohio Valley on this day, as well as a weak high pressure system in the Southwest indicating a possible positive PNA. There is a negative NAO in place, though we can't tell whether it is west or east based. This is our storm system that we could see hit on Dec. 18-21, and is the same system that the ECMWF was doing so well with before it got lost.

There is a method to the ECMWF and GFS' madness, and it is shown below.


Our storm system has now moved to the east, wobbling a bit as it did so, and combined with Sandy. Notice we still had our positive PNA showing itself with high pressure in the West US. If we were to follow this system and discount Sandy, I still believe an eastward movement would have prevailed, likely not as fast if Sandy were not present. The lack of Sandy in this potential storm system, combined with a positive PNA and east-based negative NAO leads me to believe that this forecast system should go a little more to the north and east than it did in October. 
The ECMWF and GFS are pulling the idea of a Sandy repeat, not in strength but in track. Personally, we all know Sandy isn't going to happen every 53 days, so we won't see this system ride the South US and shoot up the coast. If it DOES end up happening, it won't be a HUGE surprise, but it will be something I was not initially predicting.

Preferred Forecast: ECMWF EPS
Confidence: 45%

Andrew