Long-range forecasts are "Snow Joke"

There has been a LOT of buzz on social media and around the watercooler over the past couple days about the potential for a major storm next weekend. No, not Super Bowl weekend like the Farmer's Almanac predicted, but the following weekend. I hope that this talk raised your eyebrows and your level of suspicion about the accuracy of such a forecast. 

The biggest problem with this forecast is that it is way too far in advance. The rumor mill began on January 29th, which is 12 days prior to the supposed arrival of this storm. On that date, the European computer model's long-range output showed a massive storm for the Mid-Atlantic region. But that is just one run from one computer model; it's not nearly enough information to make any sort of prediction! It's like seeing a photograph from a city block away, through a frosted pane of glass, and saying that your view of the photo is crystal-clear. I'm going to show you a few maps so you can really visualize how much the forecast can change, based on small changes in the model output!

First of all, the forecast will change from one model run to the next. A model run is each instance that the forecast model receives new data ("initial conditions" from satellite and radar data, weather balloons, and other sources). Most forecast models are run 4 times per day. 

The GFS, or Global Forecast System, is the primary model used and maintained by NOAA. Last year, the GFS received the first of two upgrades, which will boost its computing power tremendously. (You can learn more about the upgrades here.) Even with the upgraded computing power, the model still has wide variations from one model run to the next! Take a look at the different forecasts from the GFS model, valid for 12Z Sunday. The map shows the GFS's prediction for precipitation during the 6-hour period leading into 7am on Sunday, February 9th. 

These two model runs happened 6 hours apart from each other, and check out the huge difference in the placement of the precipitation! The same computer forecast model came up with two very different ideas for Sunday morning's precipitation potential. 

Now, let's compare two forecast models against each other: the GFS and the ECMWF (the European model). This time, I'll use the same model initialization time, but I'll still show you the same forecast time of 7am on Sunday, February 9th. 

Here, the difference is even bigger! The European model keeps us dry during the early morning hours... a huge contrast to the approximately 1.00" of liquid precipitation over the same 6-hour period that the GFS model is predicting. This model variability is extremely common in the long-range, which is why we professional meteorologists don't put any stock in any extreme weather predictions that they might produce! 

I want to make something clear... I'm not saying that a big storm is impossible in the 2nd weekend of February. All I'm saying is that it's way too early to make any sort of prediction. Hopefully, now that you've looked at the highly variable model data with me, you can see why it's too early to make such a forecast, and you understand why there aren't any reputable weather sources that are making such a bold prediction!