Here are some representative totals from the National Weather Service:
Brandywine (PG co.): 1”
BWI Airport: 1.05”
Reagan National Airport: 1.39” (just short of record, 1.52” from 1932)
Laurel (PG co.): 1.5”
Quantico (Prince William co.) 1.92”
Fairfax Station (Fairfax co.) 2.04”
Dulles Airport: 2.3” (new record for the date)
Gaithersburg (Montgomery co): 2.38”
Sterling Park (Loudoun co): 3.2”
Where our forecast went wrong was in the snowfall department. The rain changed over to sleet and then snow from west to east between between about 8:00 p.m. (in western Loudoun county) and midnight. No snow accumulated inside the beltway. The closest accumulating snow (mainly on grassy areas) occurred 20-40 miles to the west and northwest of D.C., with a coating to 0.5”.
While we called for temperatures to drop late yesterday afternoon, we thought the precipitation would cutoff by the time it became cold enough for snow. However, a strong upper level disturbance kept the precipitation going several hours longer than anticipated which allowed for the surprise wet flakes.
We had seen this possibility of snow earlier in the week and some commenters specifically brought it up, but we never chose to include the chance in our forecast because 1) the warm ground temperatures would make it hard for any snow to stick if it snowed 2) temperatures would be at or above freezing further reducing the chance of accumulation 3) the overall pattern was the opposite of what’s typically required for accumulating snow in the metro area late in the season and 4) the models went back and forth in simulating any snow at all, not lending much confidence in the scenario.
Many people pointed out the irony of the Sunday’s Snow Potential Index (SPI) reading “0 (→) - Nada, naught, zip, zilch, goose egg, bupkis, need I go on?” when snow flakes were falling. I reminded readers that the SPI is a tool to assess the odds of <em>accumulating</em> snow, which did not occur inside the beltway (which is where the SPI most directly applies; because the area is so big, a single number can’t do justice to snow odds everywhere in every situation). In hindsight, a better SPI for yesterday would’ve been 1 or 2.
The bottom line is that we just didn’t want to hype up snow chances when the odds were so stacked against any impactful accumulation. Were we totally surprised to see a little conversational snow on the backside of a dynamic March weather system? No. Were the snowflakes consequential? No - in most areas (save western Loudoun and Frederick counties where it was enough to cause some issues). Did we miss it and should we have better factored it in to our forecast? You bet.