Total snowfall accumulations at my house (6 miles east of Bowling Green) were 2.5″, with 2″ of that falling between 11 pm and 7 am this morning. Up to 4.5″ fell in western parts of Bowling Green where a nice convective band set up for several hours around midnight. Thus my forecast of 2-4″ total storm accumulations for Bowling Green with higher amounts in convective bands worked out well. The convective nature of the snowfall also illustrates why this type of storm is prone to major bust potential (ie., Paducah, Lexington).
This map here shows the problem. The way that synoptic scale models like the GFS and NAM handle a convective snow band winter storm (which is a mesoscale feature) is to average across the mesoscale-synoptic scales and print out way too much precipitation. In reality, the final snowfall contour map from this storm (and any convective snowfall storm) will look more like a map of topography in a ridge-valley terrain with tight snowfall gradients. This tight snowfall gradient cannot be captured by a synoptic scale model, which means that forecasters tend to forecast only the high-end parts of the gradient. This is evidenced by what happened in Louisville. The forecast was for 10-15″ across north-central KY including Louisville. The end result was that while the Louisville NWS office was directly under a convective band (11″) the airport only got 6.5″. A similar thing happened over Warren County where the western part of the county got 4.5″ and I got 2.5″ while areas near Mammoth Cave got around 8″. Lexington and Paducah just happened to be located in places where convective bands never set up while Memphis and NE Arkansas got hammered. All in all, this is a very difficult type of storm to forecast and whether or not snow-lovers are happy or not depends on just where you happened to live.
So what is the best way to forecast this type of storm? My philosophy is to forecast to accurately predict the lower amounts but add that some areas could get much more (ie., most people in Bowling Green will get 2-4″ but a few spots will get 6″+). Or for Louisville, the forecast could have been “most people will get 5-10″ but a few spots will get 15 inches”. That way no one gets disappointed on the low end, but the people who do get under a convective band and get hammered aren’t surprised. Just my two cents worth.