The European model, the model that did such an outstanding job predicting the track of Sandy, had for several consecutive simulations, forecast snow, late Wednesday afternoon into the night in the Washington, D.C. area. But - in a stunning turn of events this afternoon - has backed off its snowy forecast.
The latest simulation develops the storm too far to the east to produce significant heavy precipitation in the D.C. area. Rather, it shows a glancing blow. Temperatures may be cold enough for some snowflakes Wednesday night, but - based on all available current information - snow is unlikely to fall heavily enough to accumulate.
Of course, small shifts in the storm track could still deliver more (or less) snow than the current European and other models indicate. So this forecast is still very difficult and uncertain.
The European model has now joined the ranks of the Canadian , GFS, and SREF models which tend to graze the D.C. area with a period of light to moderate precipitation rather than bringing a big storm. The European’s about face shows the even the best models may not be consistent and reliable. No model is gospel.
Last night (and in runs prior), the European model and its ensemble mean (the average of a number of model runs where the initial conditions have been tweaked) both offered D.C. an early season snowstorm.
While the Euro was bullish for snow, the other operational models (the GFS, UKMET, Canadian) kept the low farther offshore giving the D.C. area either a quick glancing blow or no precipitation at all. The differences in the model solutions stemmed from how they handled the interaction of two different streams of flow and atmospheric waves within those streams.
In the European model, the northern wave captured the southern southern wave forming an impressive storm closer to the coast compared to the other models. That merging or phasing of the waves allowed the model to spread precipitation well west of other models.
SREF model: snow a long shot
The Short Range Ensemble Forecasting System (SREF) offered one clue the European model *may* have been too aggressive with its snow forecast.
The SREF’s plume diagram for Reagan National Airport (see below) showed little snow potential.
A plume diagram displays each ensemble member’s precipitation during the event showing the timing of when the precipitation starts and displays how it accumulates with time. The slight variations between the different members (green lines) give meteorologists insight into the predictability of a storm and variability in the predictions. For this storm, these simulations were warmer than the European model and favored rain over snow as evident by all of the green (rather than blue, which indicate snow). Furthermore, more members showed light precipitation compared to heavy precipitation.
While the SREF favors rain rather than snow and lighter amounts rather than heavier, the broad spectrum it simulates summarizes the difficulty in confidently predicting how much precipitation will fall from the system. The SREF guidance is indicating forecasting this storm is much harder than the it was for Sandy.
The latest Canadian and GFS model runs are pretty similar to the European, showing a glancing blow for the D.C. area. But note, because the low is well offshore (meaning wind flow is from the north), temperatures are cold so we cannot rule out some snow flakes. If the models shift a bit west on subsequent runs and bring steadier precipitation into the region, snow chances might go up. Therefore, the chance of some snow still cannot be completely discounted.
In summary, there are still four possible scenarios:
1) a heavy precipitation event that may have snow falling at times and possibly accumulating. The best chance of snow in this scenario would be in the colder suburbs to the northwest and west (with more elevation); that is, if the precipitation extends that far inland. [possible]
2) a light-moderate rainfall event with perhaps some snow mixing in but no accumulation [most likely]
3) a light rainfall event across the eastern suburbs [possible]
4) a complete miss [least likely]
My gut feeling is that a complete miss is the least likely of the bunch. I noted Sunday that we were probably 24 to 48 hours from having the models reach a consensus allowing us to make a definitive call about whether snow would fall in the D.C. area. A day later and we still may be 24 hours away from an answer.
Wes Junker is Capital Weather Gang’s winter weather expert. Jason Samenow contributed to this post.