Pretty much every conceivable opportunity for snow this month has overachieved; so I suppose it shouldn’t come as a shock Sunday’s grand finale exceeded expectations as well.
After forecasting at most some conversational snowflakes late Sunday, a coating to a couple sloppy inches fell in the immediate area.
Let’s review how and why the snow showed up while our forecasts fell flat.
What we actually forecast
The forecast wasn’t a total whiff as we advertised the possibility for some snow Sunday in our Friday evening PM Update as well as in Saturday’s forecast .
The Friday PM Update opening paragraph said: “Winter may try to throw a parting egg in our face late Sunday, with falling temps, wind, and rain (and maybe, gulp, some conversational flakes?).”
Clearly, we jinxed winter’s death. New model info suggests small chance snow mixes w/ rain late Sun: http://t.co/ki0Rbk4hnN— Capital Weather Gang (@capitalweather) March 28, 2014
And the detailed outlook for Sunday in that Friday forecast added: “…periods of rain are likely (60 percent chance) as temperatures hold steady – mostly in the raw 40s. Into the afternoon and evening some 30s are even possible. If the rain persists late in the day, it could even mix with or even change briefly to conversational (non-accumulating) snow, especially in our colder suburbs. ”
Our forecast on Saturday also noted the possibility of wintry precipitation Sunday. Ian Livingston described bands of rain “perhaps mixing with conversational snow by late day north and west” and a chance of light snow a night, with grassy accumulation possible in the hills of northern Maryland.
The snow, of course, arrived earlier, was more widespread, and accumulated more than any of these forecasts indicated.
While the forecasts for Sunday posted late Friday and Saturday were on the right track, even if flawed, our forecast for Sunday posted Sunday itself completely derailed. The forecast published at 6 a.m. made no mention of frozen precipitation during the day on Sunday. It simply called for a cold, wind-swept rain, with temperatures falling through the 40s. The mention of snow was a passing one, for Sunday night: “We’ve been watching for the possibility of some back-end snowflakes with this system,” the forecast says. “I wouldn’t rule it out completely, but it’s more likely out west toward the Blue Ridge than in the D.C. metro area.”
Our critical error
Why did we get it so wrong in the short run? Without getting too technical, the model forecasts overnight Saturday suggested it would be too warm for snow accumulation during the day Sunday. They suggested temperatures would hold in the 40s – too warm for snow.
But there was reason to be skeptical of these temperature forecasts – which regrettably – was overlooked. While the models showed temperatures in the 40s on the ground, they showed a deep pool of cold air at high altitudes (see graphic below, first frame). As the precipitation became heavy Sunday morning and afternoon, some of that cold air was drawn down to the surface through dynamic (and evaporational) cooling. That caused the rain to mix with and change to sleet and snow, first in the mountains and eventually to the I-95 corridor and even briefly in parts of southern Maryland.
Then, the storm threw another curveball. Once the snow really got going and it appeared the region might be in for a more serious surprise snow event, the snow transitioned back to rain towards Sunday evening from west to east. Even though it was precipitating heavily and darkness was approaching – which ordinarily would support dynamic cooling and additional snow – the cold air pool drifted away to the east and northeast (see second frame, above) and temperatures rose through the 30s. The erosion of the cold air spared much of the region from a potentially disruptive snow event. The exception to this was a sliver of northern Maryland in Carroll County, where enough cold air lingered (and elevations were high enough) for the snow to persist and 8 to 9 inches fell near Manchester.
This was an extremely complex and challenging storm to forecast due to model fluctuations and subtle arrival and departure of cold air barely supportive of snow. Although no forecaster in the D.C. area got it right (to my knowledge), we should not have abandoned the idea that there was just enough cold air around for some snow to occur Sunday, especially in our colder suburbs. As our forecast posted 6 a.m. Sunday morning included nary a mention of frozen precipitation during the day and we were on our heels attempting- somewhat futilely – to get the forecast right the rest of the day, I have to self-grade us a D for this event – maybe a C if grading on the curve.