Over the past couple of days, the European weather model (ECMWF) has been advertising the potential for snow in the Washington area Saturday, especially for areas west of the city with elevation. As Ian Livingston noted , accumulating snow is extremely rare in Washington during October. It has only been recorded five times in a total of 4,340 days. Because of its rarity, any model forecast of snow should be looked at with some skepticism, especially when the forecast is a projection that is four days from now.
The most recent significant October snowstorm in the Washington area occurred on October 10th, 1979. That storm produced up to 17 inches of snow in the mountains of Virginia but even locations as far east as Prince George’s County recorded a couple of inches.
Rob Olinger who lived in Warren County (his elevations was 1,200 ft.) recently posted his recollections of the storm on the AmericanWx forums:
I remember being woken up by thunder .. Someone quickly ran and turned on the outside flood lights and I was shocked to see giant wet snowflakes literally pouring from the sky, imagine a summer T-storm deluge only with snow and every one of us looking in disbelief when we saw that everything outside was white...
He remembers measuring 14 inches, losing power for several days and having a large tree limb fall on their roof. October snowstorms while extremely rare can be destructive as the trees still have most of their leaves. Therefore, despite my skepticism of the European model solution, the potential storm is worth looking at a little closer.
The main reason that snowfall is so rare in October is that average temperatures are so warm. The pattern has to be very unusual to get any snow. The average maximum temperature remains in the middle 60s and the average low temperature is 46 through the end of the month. The coldest maximum temperature recorded in October was 41 degrees so any snow that fall needs to fall late at night into the early morning hours before the sun starts warming the atmosphere.
In addition, for snow not to melt into rain on its descent and then accumulate on the ground, the precipitation needs to fall rather heavily. The timing of the European model forecast from last night and today meets the first criterion having the snow start towards dawn and then continue through at least 11 a.m. Note the snow accumulation maps (over three hours) ending Saturday at 11 a.m. and 2 p.m below from last night’s run (the accumulating snow in the second image after 11 a.m. is somewhat dubious due to warming surface temperatures).
These European snow maps are a snow lover’s dream and the model typically verifies a little better than the other medium range models (such as the GFS). Unfortunately for snow lovers, this time around it’s probably fool’s gold.
Last year, the European model overdeveloped some of the waves early in the season. It may be doing this again.
Among the suite of models, the European is the only one model currently simulating the perfect “snow” track with the low deepening rapidly just off the coast with the mid-level center tracking just south of DC. Such a track/scenario keeps the cold air in place and favors the necessary heavy precipitation.
For sake of comparison, let’s look at the GFS model forecast from this morning valid Saturday morning at 8 a.m. (below, left panel). The black Ls on the map are the positions of a frontal wave well off the southeast coast. The other low near Florida is the model’s forecast of Hurricane Rina. The red Ls I’ve painted in (they’re somewhat faint, look closely) depict the track of the European surface low starting over Georgia at 8 p.m. Friday evening and then progressing to a position just off the Delmarva coast on Saturday at 8 a.m. with the low then slowing and moving to a position just off of Long Island by 8 p.m. Saturday evening.
In short, the GFS and European models are worlds apart in their forecasts of the surface low.
The NCEP ensemble output (valid Saturday morning, shown in the right panel) from last night - which shows a range of simulations - do not have deep enough dip (trough) in any of the green lines (jet stream simulations) or a big enough bump (ridge) to the east of the dips to support a surface low track similar to the European model. Each suggests a flatter solution with the low tracking farther south and east than the European model. Most would keep the D.C. area dry.
Right now — even though the new European model coming in continues to show a wintry Saturday morning — it looks like the European is long shot at best.