* Snow total reports from NWS | Snowfall total map *

For almost a week, we talked about today’s snow potential. We discussed the favorable ingredients in place: a cold arctic high pressure center to the north and a moisture-rich area of low pressure passing to our south. Despite being late into March, the setup had the makings of a classic D.C. snow event.

And we predicted accumulating snow. Just not enough. Not nearly enough.

Our Sunday afternoon forecast was for 2-4 inches for most of the region, except 3-6 inches in our colder suburbs to the west, and 1-3 inches around city. (Yes, we increased totals Sunday evening to 3-6″ for most of the region, 4-7″ for our western areas, and 2-4 inches for the city, but that was too little, too late.)

Updated CWG snowfall forecast as of 2:30 p.m. March 16.

In reality, a general 6-10 inches fell over most of the region (a few areas south/southeast of town were in the 4-6″ range, and a few places west and northwest hit double digits).

Actual snow totals March 16-17 (National Weather Service)

Basically, we forecast only about half the amount of snow that fell.


We underestimated the cold air

As temperatures on Saturday had hit 70, and even on Sunday climbed into the upper 40s, we were skeptical temperatures would cool quickly enough for significant accumulations during the first half of the storm (Sunday evening) – particularly around the city. (We further reasoned the warm ground would melt flakes, which it did, to an extent.)

We fully recognized the arctic air feeding the storm would eventually allow accumulation, it just happened faster than we thought.

It’s totally possible we were influenced by past events when the air did not cool as fast as models predicted. But, in this case, the cold air mass was impressive and we should’ve put more faith in all of the indicators (i.e. low dew points, temperatures in the teens in northern Pennsylvania Sunday afternoon, low atmospheric thicknesses, cold high altitude temperatures).

The heavy precipitation band shifted north slightly

In the days leading up to the storm, models were shifting their forecasts for the heaviest precipitation to fall south of D.C. On Sunday, that shift stopped and there was even a subtle shift back north in some of the models. It turned out that the heaviest snow bands set up right over the D.C. area – focused on our west and north suburbs – rather than south.

All winter long we’ve talked about how tough it is pinpoint exactly where these bands will focus. In most storms, these bands have ended up slightly north of where the models suggest. We took that into consideration with this storm, but not quite enough.

What we did well

  • Identified the potential for accumulating snow many days in advance and told a consistent story
  • More or less correctly described the timing of the storm and when the heaviest snow would fall. We were right that it would take until late in the evening in most areas for snow to begin accumulating on roads.
  • Correctly forecast that there would be slick Monday morning travel conditions and widespread school disruptions

Where we erred

  • In most areas, precipitation began as simply snow, rather than rain or a rain/snow mix
  • Snow began accumulating on grassy areas sooner than we thought, and around the city, temperatures cooled more quickly than we thought
  • Per the discussion above, heavy snow bands came a little bit farther north than we expected
  • We did not think the storm would be significant enough for the Federal government to shut down (we said less than 50 percent chance). We were wrong.

Kudos to the National Weather Service

The National Weather Service predicted 4-8 inches for the metro region (increasing the forecast to 5-10″ late last night), more than we did, and overall had one of the better forecasts in the area for this storm. It also did a nice job at highlighting the (relatively high) probabilities that snow could exceed 4 inches well in advance (starting Friday).

Our grade

Overall, I don’t think our reasons for being conservative in our forecasts for this storm were necessarily bad ones, but we should’ve done a better job at evaluating the available cold air.

Forecasting snow amounts is hard and humbling.

As always, our entire team worked hard to provide a carefully-considered forecast and comprehensive storm coverage. We correctly predicted an unusual late season accumulating snowstorm that would cause disruptions in the region, but fell short of forecasting what turned out to be a historic storm.

Grade: C

Do you agree? As always, we welcome your feedback…