This experimental presentation of snowfall probabilities fro...

This experimental presentation of snowfall probabilities from the NWS is outstanding. No sensationalism, no over-simplification, just probabilities. Basically, this weekend's storm is a "we're not sure" for snow accumulation, but that's under-specific. These forecasts show it's more like 50% chance of nothing much (under an inch) and 20% chance of over 6 inches.

Winter Weather Page

Joe Carnahan - 2016-01-22 10:26:11-0500

That is cool! I notice there's a feedback link on that page - I made a point to go fill it out and rave about this.

Matthew Gray - 2016-01-22 12:03:33-0500

They also provide, CDFs and PDFs of the accumulation by town, but they're not consistent :-/

I left glittering and slightly confused feedback.

Glenn Willen - 2016-01-22 12:35:15-0500

+Matthew Gray Are you sure? I double-checked the CDF and PDF for Boston and they appeared to match. That is, I summed e.g. the PDF entries for Boston for 0", 0-0.1", 1-2", 2-4", which gives 61%; the CDF entry for >=4" is 39%, which adds up to 100% as expected if you assume that >=4" is basically the same as >4" because we're talking about the real numbers, so the PDF at exactly 4 is 0.

Matthew Gray - 2016-01-22 14:15:23-0500

For me (earlier today) the 8-12, 12-18 and 18+ all summed to a much lower number than the 8+ exceeds number did. I left feedback, but the data now is consistent, so maybe it was a transient glitch when the refreshed the estimates?