Yesterday we introduced our White Christmas tracker. We’ll run this every week day through Christmas Eve. It’s a daily update on the potential for snow on the ground Christmas Day using computer models.

It’s important everyone understand that model forecasts more than a week into the future are very unreliable. They have not demonstrated the ability to accurately predict the weather this far out in time.

That said, we are showing you the models so we can track together how they evolve (i.e., do they shift snowier, less snowy, or constantly flip-flop?) and then, in hindsight, evaluate how well or poorly they simulated reality.

Predicted snow depth Christmas Day at 7 p.m. EST from GFS model (

It continues to show substantial snow cover in places you’d expect a lot of snow: near the Canadian border, in the Rockies and around the Great Lakes. Same as we saw in yesterday’s simulation, the I-95 corridor remains snow-free, including Washington, D.C. and New York City, as does the South and most of the Pacific coast. But snow depth has increased to 6”+ in the mountains of West Virginia, while the stripe of snow seen in yesterday’s forecast across the north-central U.S. from Colorado to Michigan has mostly disappeared.

Next up, we have the CFS version 2 - this is a long-range or seasonal forecasting model. It has a product which shows you the 5-day maximum snow depth over a given time period. It includes four simulations that are generated using slightly different inputs.

For the 5 days leading up to and including Christmas, the two runs on the left have a similar look to the GFS, but the bottom one pushes the snow farther south and east. Notice that three of the four show a snow hole over the north-central U.S. similar to the GFS, and the top-right model run shows the West Virginia bull’s-eye like the GFS.

Predicted maximum snow depth between late December 20 and late December 25 by four runs of the CFS version 2 model. (

Also note that the CFS is showing maximum snow depth over a 5-day period (not the amount on the ground per the GFS above). So consider that some if not all (where totals are low) snow shown may melt before Christmas Day.

Probability of a White Christmas (defined as 1” of snow on the ground Christmas Day) based on historical data. (NOAA)

Considering available information now and the possibility (but not a probability at this point) of a storm next week, I would place the odds of having snow on the ground Christmas Day in D.C. just slightly higher than the historical average (see above) of just over 10 percent. (Stay tuned for a post by Ian Livingston that will provide many more details about the traditional odds of a White Christmas in D.C.)

Previous White Christmas Trackers:
December 10, 2012