The Trend May Not Be Your Friend

The Trend May Not Be Your Friend

If you want a lot of snow in Little Rock, the model trend is not your friend
There's a good reason you don't see a lot of reliable snowfall forecasts more than 48 hours out. When it comes to snowfall forecasts beyond 2 days, things can change a lot. Take for example the change in snowfall guidance over the last 24 hours for Central Arkansas. Forecasting precipitation amounts is tough enough, but when you take into account precipitation amounts near the rain/snow line the challenge is greatly compounded.

Take a look at computer model output from Saturday at Noon. The white line shows total snowfall accumulation. The time frame on these charts is right to left. The color of the bars shows precipitation type. Green is rain, red is sleet, freezing rain or mix, and blue is liquid equivalent of snow. Grey is forecast snow amounts.

 (GFS 18Z Sat Dec 22 2012 Precipitation forecast for Little Rock)


(NAM 18Z Sat Dec 22 2012 Precipitation forecast for Little Rock)

 The GFS model shows over 4 inches of snow for Little Rock, beginning after 3 PM Tuesday. The NAM model shows less snow, just about an inch of accumulation after 4PM. 
24 hours can certainly change things though, the model runs are trending very slightly warmer for late Tuesday afternoon and early evening. This makes a big difference in the timing of the change to snow and therefore even a bigger difference in snowfall amounts. Below are the same models run Sunday at Noon. 

 (GFS 18Z Sun Dec 23 2012 Precipitation forecast for Little Rock)


 (NAM 18Z Sun Dec 22 2012 Precipitation forecast for Little Rock)

The trend is obvious. Using the same set of parameters, snowfall output is way down with model output now showing less than an inch of accumulation for Little Rock, with the rain to snow transition not occurring until much later in the evening. The GFS model shows the changeover between 6pm and 9pm, while the NAM is even later showing mostly just rain in the Little Rock metro area  until almost 11PM. Of course, model trends may swing back the other way, but inside of 48 hours, the computer model runs tend to become far more consistent. All of this being said, significant snowfall is still likely for parts of the State Christmas Day, but pinpointing who gets exactly how much for a precise location is still tricky business.
Page: [[$index + 1]]
comments powered by Disqus