Yes, the models are consistently inconsistent and it’s enough to drive me nuts… or it was.
Do I miss forecasting the weather on TV? Absolutely! Do I miss snowfall forecasts? Nope.
Of all the different types of weather I forecast snowfall was the type most consistently misforecast! I’m not just talking about me. No snow forecast by anyone is ever completely right!
Maybe I’m being too hard on myself and my colleagues? There’s a lot of utility in snow forecasts. One reason schools are closed more now is because decisions can be made with confidence. We mostly hit the target. We never get the bullseye.
When snow forecasts are really wrong there’s nowhere to hide!
Here’s what we do reasonably well:
- precipitation type
- start time
- wind
- temperature
Here’s what we do poorly:
“You said six to ten inches and and I got four!” Trust me, I’ve heard that angry sentence more than once.
Shoot me. It happens.
A snowfall forecast is actually a forecast on top of a forecast on top of a forecast. If any of the intermediate calculations are off everything else is off too. That’s bad news because forecasts can change radically over a short period of time.
The good news is we don’t see cars stranded like they were in 1978 because even if we don’t get totals right there are few storms (and no large storms) that appear as surprises.
To most of us there’s no difference in our lives between one inch and four inches. There is no difference between three inches and eight inches. Sure there’s more to plow, but whatever stops for eight inches has mostly stopped for two.
I used to say all the slippery’s in the first quarter inch. It’s true.
At the top of this entry is a time-series plot courtesy of Dr. Robert Hart from FSU and his coolwx.com website. It’s a breakdown of the prediction from Monday’s 2:00 PM EDT (1800Z) NAM computer run. The NAM says New Haven will start seeing snow Wednesday afternoon with the mainly light snow continuing through the night. The scale on the left side shows water equivalent as opposed to snow totals.
By the time the snow ends Thursday this model says we’ll see about .7″ of water converted to snow. At a typical 10:1 that’s seven inches of snow.
Of course no snow is really typical and warmer ground and light snow at the onset will probably melt the first flakes that fall. Beyond that when snow falls for any length of time it tends to settle, so it takes more than five inches of snow to get five inches of snow! That makes the .7″ academic, not practical.
I mentioned this is from the 1800Z run. As I was typing the next run, the 0000Z came in with around 1/5 the snow!
Yes, the models are consistently inconsistent and it’s enough to drive me nuts… or it was.
Here’s the takeaway. There will probably be some snow on the ground when you wake Thursday. There will be delays if schools think they can wait it out… and they probably can.
The last word on this system hasn’t been written. In some ways I’m happy to be an outside observer.