The models seem to be converging on a solution offering us some snow next Tuesday. But how much and whether it accumulates is still very much up in the air.  Most models now track a low to our south while we have much colder than normal air over us.  These two factors don’t guarantee snow, especially accumulating snow this late in season, but are needed to give us any chance.

This year, the models have done an unusually good job at ferreting out patterns conducive to winter weather.  They are again showing the evolution of such a pattern as another shot of cooler than normal  air should start feeding into the region during the latter half of the weekend.   Some models like the GFS even hint that period of wet light conversational snow could fall during the day on Sunday as weak low pressure passes to the south.   The NAM model, on the other hand, favors that the bulk of the precipitation with that system would fall as rain.  None of the models give that system much potential for accumulating snow.  That system is just a harbinger for the more significant threat on Tuesday-Tuesday night.

Late March snowstorm threats are always tenuous because the sun is so strong and the average maximum temperatures even at Dulles Airport are in the upper 50s.  Daytime temperatures need to be around 27 degrees below normal for snow to accumulate during the afternoon.  Therefore,  most accumulating snow in late March occurs at night.  Even then, you need abnormally cold air. For better or for worse, the models certainly are advertising such cold.

GFS model at 8 p.m. Tuesday shows snow over the region, but the heaviest precipitation is to our southeast and offshore. (

Technical discussion

Today’s GEFS (Global Ensemble Forecast System) mean forecast (left) and spaghetti diagram (right) – which display the upper level pattern (at 500 mb) – do a good job of illustrating why it is expected to be colder than normal next week.    Both panels below are valid at 7 a.m. Tuesday morning.

The panel on the left reveals the Arctic origin of the air.  Note how upper level winds, approximated by the contoured lines,  extend from the near the North Pole southward into the U.S.  Such a pattern usually pulls strong  Arctic high pressure system into the U.S.

The spaghetti diagram (on the right) shows how similar all the ensemble members (different simulations) are in handling the pattern.  The blue, red, and green lines represent the various ensemble model forecasts.  When the various sets of colored lines are clustered together, the various ensemble members are in close agreement.  (Remember ensemble members are simply models run with the same physics of the main model but slightly different initial conditions in hopes of assessing the predictability of the pattern.)  They all show that same flow from the north and northwest into the U.S.  That flow is what feeds the cold air towards us and gives a low passing to our south the potential to produce snow.

Today’s GEFS forecast of the mean sea level pressure and 850 mb (5,000 feet) temperature forecasts valid at 7 p.m. Tuesday (see below) show how strong the signal is for a storm but also illustrates some of the uncertainty in the forecast.  Note that most, but not all, have a storm near the East Coast.  A few suppress the storm to our south but they are in the distinct minority.

The latest operational GFS (top image in this post) is similar to the top left panel.  It gives us some precipitation – probably in the form of snow – but generally keeps the heaviest precipitation to our south.

Today’s European model (below) is much more bullish about a storm.  It is more similar to the wound-up second member in the second row (of the GEFS) but is farther to the east with its wound-up low.

For snow haters, there is good news.   Like many of the GEFS members, the European model  valid at 7 p.m. Tuesday (see above) has the bulk of its precipitation falling during the day on Tuesday with surface temperatures above freezing.   One cautionary note is that the storm is still more than 5 days from now and at that time range the timing of the storm evolution could end up either faster or slower than forecast by 6 to 12 hours.

While the models are suggesting that more snow is possible, a snowstorm still is not a certainty.   A few ensemble members keep the low too far south to give us much, other models and ensemble members only give us light precipitation as the low skirts to our south.  None of those solutions would offer accumulating snow. And light precipitation during the day might end up rain rather than snow.  However,  two of the last three European model runs have predicted the development of a strong low pressure system off the North Carolina coast.   Today’s European run would offer snow too but, with the snow falling during the day, it probably would not be that big a deal.  However, wrap the low a little closer to the coast and change the timing a little so that the bulk of the snow falls at night and and some accumulations could occur.  Stay tuned.