The model giveth and the model taketh away. Whether you want a big storm or 70-degrees and sunshine, when we're peering into the future more than a few days ahead, computer model data is all we have to go on. And while weather modeling has becoming increasingly accurate over the past few decades, we sometimes tend to forget that the solutions they give us aren't real. The pretty graphics they generate are based on many approximations, and regardless of how great a storm might look several days away, any one model is only as good as its initial conditions. If it fails to resolve a feature in the atmosphere when it initializes, it will give an incorrect solution for the future. That's why we often approach the extended forecasts with due caution, knowing good and well that things can and will change, even if every model out there shows the same thing. We've seen some pretty big changes in the model forecasts in the past 24 hours for the potential snow this Sunday, and it now looks less likely we'll see much of a storm. But before you write this one off completely, understand that we could just as easily see a move back in the other direction over the next day or two. The system is just now coming on shore near Seattle and Portland, and the models should have a better handle on what it will do in subsequent runs.
Here are a couple of things I like and don't like about the potential for snow this weekend: I like the fact that this is a pattern that has been depicted by the GFS model for over week now. It looked like something could happen around the 17th-19th of the month, and the system is due on the 20th. I also like that while the models are in agreement with a more southerly track that doesn't effect Virginia much, at least all of the models are showing a storm in the region. You can't have a storm shift it's course north if it's not there to begin with. I don't like that the main vort max becomes nearly cut off across the southwest U.S. on Friday. Cut off lows can be a nightmare for forecasting. I also don't like the lack of a Greenland block. Even the most aggressive model takes the storm on a SW to NE track, which has a much better chance of completely missing us to the south. If we had more of a blocking pattern (indicated by a negative phase of the North Atlantic Oscillation (NAO)), this would allow a more southern extent to the cold air, and would also cause the storm to track more S to N, or right along the coast. And the last red flag, and a very important one, was the lack of near-freezing surface air during the time of the event. Our in-house model generated a whopping 0.7" of snow yesterday when the Euro model looked its most impressive. Similar to what happened here on Saturday, it can pour snow, but if temps are in the mid to upper 30s, it won't amount to much. But again, it's a model, and it too will change.
So all this leads to me making no changes to my forecast for Sunday. I'm leaving us at 40% for snow, and won't move it until we get a little closer. As I type this, the midday GFS run has completed through the weekend, and is more favorable for snow. But I don't like the midday and overnight model runs. Confused yet? Be ready for snow on Sunday, and a plan for some decent weather as well. Mother Nature will take care of the details. -ZD