UPDATE: I was unable to look at other models or the GFS ensembles today, I will do that tomorrow. A commenter left a comment with a good point - the snow that falls tomorrow (see Headlines at right) will be around for a while in many areas, because of this cold outbreak. I would also caution those (like me in Central PA) with ice still covering the ground, to be careful walking and driving on the fresh snow.

NOTICE: Due to time constraints, I cannot produce seasonal or point forecasts for every blog reader. Please access AccuWeather.com, especially AccuPOP, for a customized forecast from our team of mets, considering all options for your zipcode. Discuss all the upcoming storms on our Forums. Model images are from Pro.

AccuWeather.com Professional's Joe Bastardi [BIO] (PRO USERS READ NOW | 30-DAY FREE TRIAL) pointed out this morning that the GFS Forecast Models [JessePedia] has been outputting some incredible temperatures for New York City next week. I decided to make use of our Pro site to dig into this deeper, because it has a robust selection of meteorological tools to find out more. Here's a "meteogram" showing surface conditions at New York City (KOKX) next week:

That's right, you're reading it correctly. The model is predicting that the temperature in New York City will be about -13 F on the morning of January 17th.

I can tell you without even looking at historical data that a -13 degree reading would be "historic" in the Big Apple. If you do look at their history, you find that -15 was the lowest they have been there since 1869. Now, there are a few things to consider here, before we freak out.

1. The graph about is a "point and click" model forecast. If you run the numbers from the "MOS" ("Model Output Statistics") program, which refines the forecast, you come out with -20.6 C or -5 F:

2. The model is attempting to predict a forecast at 8 days out, and the GFS does have a cold bias in its longer range, so it's likely it would moderate between now and then.

3. The model could be underestimating the "heat island effect" or "regular island effect" of New York City. The model does not have the resolution to even recognize Long Island as a piece of land surrounded by water.

4. Other models are probably saying other things (check back this afternoon for more on that).

5. The GFS itself has backed off slightly from these numbers with the 06Z overnight run, and was showing nothing close to this in previous runs yesterday.

Negative 5 degrees in New York City is still impressive, but let's assume even the MOS is 5 degrees too cold and call it 0. Given all that, then the question becomes, when was the last time it was below zero in New York City? The answer, according to the National Weather Service, is 1994. Before that, you have to go back to 1948 to find a reading below -2.

Given that, where is the core of the cold air that morning? Here's a map of GFS temperatures, and they are absolutely incredible, with eastern PA reading less than -30 Celsius (closest MOS station showed -22 C, meteogram showed -30 F!)

Given that this is 8 days out on a model known to have cold bias, I can almost guarantee this will not happen. However, that's what the model said as of last night.

Report a Typo