Jason Sandford
Jason Sandford is a reporter, writer, blogger and photographer interested in all things Asheville.
Asheville got a snowy surprise Friday as a weather system that forecasters initially said would bring a couple of inches of snow ended up dumping half a foot or more of snow on the mountain metropolis.
There was more snow overnight into Saturday morning, with folks reporting a total of 7 to 10 inches of snow or more. Forecasters were predicting bitter cold temperatures Saturday night, with lows around 19 degrees, and the chance of yet another dusting overnight into Sunday.
The winter storm has a name, Benji, and its potential triggered schools to shut down early. Benji’s snow took down tree limbs and power lines while creating slick road conditions that slowed traffic and caused some wrecks. The biggest event in town this weekend is the Warren Haynes Christmas Jam, and the wintry conditions forced organizers to cancel the Pre-Jam at The Orange Peel.
But most folks seemed to take the snow in stride. It helped that the snow storm arrived on a Friday.
What went wrong with the forecast: Meteorologist Jason Boyer of WLOS was contrite on the Friday evening newscasts. He posted a graphic of an ostrich hiding its head in the sand and said that’s the way most meteorologists feel about the bad forecast.
Boyer explained that the computer models that meteorologists from the National Weather Service on down depend on for their forecasts failed to account for a significant move in the jet stream that allowed a big flow of moisture move in over cold air that arrived in the mountains earlier in the week.
Thus the “Skywatch” forecasts and “Futurecast” predictions were calling first for an inch or two. Forecasters gradually increased their predictions for snow Thursday, but none of them accurately predicted the dumping Asheville actually got. With a final tally of 9 inches to a foot of snow in some places, winter storm Benji will go down as one of the bigger ones on record for early December in Asheville in recent years.
So that was the explanation: the computer models got it wrong. That still begs the question of why the computer models messed up, and makes me wonder when the human experts should step with their knowledge. A bad computer model doesn’t absolve a weather forecaster.
Feeding the frenzy: We joke a lot in Asheville about how everyone freaks out about predictions of snow, even an inch or two. The bad forecasting in this case will only serve to reinforce that freak out, as residents come to distrust the weather calls and go all out to prepare for the worst-case scenario – every single time.
Saturday update from the American Red Cross: The Asheville-Mountain Area Red Cross is opening its office at 100 Edgewood Road in north Asheville to anyone seeking to get out of the cold. Snacks, warm beverages, games, and books will be provided, according to a press release.
We salute those restless spirits who enjoy spending snow day outdoors! There are so many outdoors opportunities and adventures every day of the year in our area.
“A bad computer model doesn’t absolve a weather forecaster.” This is a misleading statement. Meteorology absolutely revolves around the use of computer modeling – accurately predicting weather could never be accomplished to the degree we have today without so. Humans can choose which model they expect to be the most accurate but you cant expect them to just conjure up something entirely unrelated from any models they have.
“Wx South” guy on Facebook predicted “a significant snow event” as early as Monday, so we were fully prepared. I recommend him to anybody who’ll listen.
Anyone who’s ever taken a basic stats class can tell you that predictions made with Model X are only ever useful for predicting events that are generated from conditions that fall within the range of Dataset X, the data used to “train” the model.
As climate change increasingly makes the weather more volatile and extreme, the current models will have more and more ‘fails’ as the actual conditions that go into creating that weather fall further and further outside the range of historical data used to train it. Retraining the model with new data can help but it’s expensive and time consuming, and since the changes are happening at an accelerated pace, it’s always going to be a game of catch-up.