Eric Tomlinson's terrier mix Jax sits on the sidewalk during the first snow of the season in Washington on Nov. 15. (Calla Kessler/The Washington Post)

An unclear snow threat continues for the Washington region Sunday into Monday.

We wish we were in a position to offer new details about whether we face a significant snow event, but computer models still diverge.

The region could receive no snow, a little snow, moderate snow or even heavy snow. Models vary on the storm timing, but if snow materializes, it should begin between Sunday morning and evening (the afternoon seems most likely) and likely end some time Monday; the early morning hours are most likely, but some models suggest the storm could linger into the evening.

(Washington Post)

Here’s how we would break down our accumulation chances at the moment:

  • Chance of no snow: 35 percent
  • Chance of at least a dusting: 65 percent
  • Chance of at least one inch: 35 percent
  • Chance of at least three inches: 25 percent
  • Chance of at least six inches: 15 percent

The snow chances increase as you head south and southwest of Washington and decrease to the north. The southwest part of Virginia, including Roanoke and Blacksburg, still seem likely to experience a major snowstorm.

Readers might question why there’s so much equivocation about the forecast in Washington. There are several reasons:

1) This forecast is very dependent on the complex interaction between the northern and southern branches of the jet stream, which models often have a hard time handling. A key to this forecast is whether the northern stream relaxes enough to allow the storm along the southern stream to come north.

2) The exact storm location is critically important for our snow prospects but is a huge wild card. By Monday night, the 51 simulations in the European modeling system show enormous spread in where the storm and its low pressure center will be. Those that have the center hugging the coast offer substantial snow. But those farther east would produce a dusting to an inch, at most. The spread in these forecasts argues strongly to keep all options open, from a significant snowstorm to a complete miss.

European modeling system simulations of the position of the storm or low pressure centers, indicated by red Ls, on Monday. (

3) With the start of the event delayed until perhaps Sunday afternoon or night, we are still in the window of time when forecast predictability is low.

The forecast uncertainty is reflected in the following range of model solutions:

  • Operational American (GFS) model: The snow shield barely edges into Washington, offering dusting around the city and keeping it dry to the north.  

Operational American (GFS) model simulation of snow Sunday night — mostly just south of Washington.
  • Experimental American (GFS-FV3) model: It projects several inches of snow from the Beltway south but decreasing amounts to the north and little or no snow in northern Maryland. The snow begins late Sunday morning or early afternoon and continues through the night. Note that it’s snowfall prediction has dropped considerably (from 15 to 20 inches to 3 to 6 inches) in the last 24 hours, when it predicted very heavy snow.

Experimental American (GFS-FV3) model simulation of snow in D.C. region on Sunday night. (
  • Canadian model: Whereas it had move away from a big storm in our region in its forecast issued Tuesday, it now again is predicting a heavy snow event, perhaps mixed with sleet or rain. Snow would begin Sunday morning and continue through the night. The massive change between these two forecasts illustrates how difficult this pattern is for the models to handle.  

Canadian model shows heavy snow in Washington region Sunday night.
  • European model: It barely brings the northern edge of the snow into the Washington region with little or no accumulation.

The American modeling system, which contains 20 simulations in addition to the main operational forecast shown above, shows a huge range of forecasts. It is pretty much evenly split between no snow, a little snow, moderate snow and heavy snow forecasts. On average, it forecasts slightly less snow than it did at the same time Tuesday.

Snowfall simulations from the American modeling system for Sunday and Monday storm. (

So what do we know? A storm will track across the South to the southeast coast by the end of the weekend. The storm will be a wet one across the South and is likely to produce heavy snow from the mountains of western North Carolina to Southwestern Virginia.

The big question remains how far north the edge of the heavy snow progresses. We’ll continue to follow the storm over the next several days and provide updates.