With all of the uncertainty between which model is right, which run, will it go east or west of the Appalachians, etc., I think it is important to step back and think about the synoptic pattern. The overall pattern idea of a two-part storm with frontogenetic cyclogenesis followed by phasing of the jet streams and subsequent deepening of the surface low still looks good. So why are the models trending colder when all winter the models have trended warmer as winter storms have approached the region? I’ll tell you why.
Remember that last night I wrote about how there was going to be three distinct shortwaves associated with the southern stream storm (precipitation batch #1). The main difference between the models from early this morning to this afternoon is the strength of the easternmost shortwave compared to the northernmost shortwave. During the 12z Thursday run of the GFS, the 24 hr forecast for 12z Friday morning had a strong shortwave over Paducah, KY. Notice that the 00z Friday run for 12z Friday morning has a much weaker shortwave in the same region. A weaker shortwave over western KY on Friday morning means less warm air advection, which is why the models have been trending colder. This is also allowing the shortwave over GA, (the surface low is a reflection of this shortwave and the right rear flank of the jet streak), to trend eastward.
By 00z Saturday, note how the initial surface low has moved off the Atlantic coast and weakens while a secondary low develops as the stronger southern shortwave over MS moves under the favorable right rear quadrant of the jet streak (which is now at >150 kts). This all coincides with the phasing of the northern stream vorticity which allows the system at all levels to rapidly deepen. By 12z Saturday morning, the storm has developed a closed contour at 500 mb and has become negatively tilted, which allows for rapid surface intensification and blizzard conditions from Ohio into western New England.
So what does this mean for the Mid-South? I think the NWS may be too optimistic with the heavy snowfall accumulations over Bowling Green (currently they have 6″ forecasted by Saturday afternoon). While the NAM is printing out monumental snowfall totals, the GFS still has Bowling Green in a 2-4″ band, which higher snowfall amounts to the NW and to the east. Again, I think Bowling Green will be caught in a relative snowfall minima between what should be 5-10″ to the NW and what could be very heavy snow over the mountains of eastern KY.
Here is the caveat: It is not uncommon for a rapidly deepening system like this to have convective snowfall bands set up to the NW of the surface low. Since this is a mesoscale phenomena, it will not show up on a synoptic scale model like the GFS, therefore one can only speculate where it might set up. Somewhere in central KY, it is likely that a band of convective snowfall will set up and accumulations could be on the order of 10+. But there is no way to know this until you see it set up on radar.
In general, I will not change my forecast much from yesterday expect to bump up snow amounts for the Friday overnight period.
Friday: A mix of rain/sleet/snow should begin early Friday morning. This may briefly changeover to all rain before noon Friday but should mix back to rain/sleet/snow and turn back to light snow Friday afternoon. Lack of arctic air, marginal boundary layer conditions, and an early March sun angle will conspire to limit snowfall accumulations to less than an inch by sunset.
Friday night into Saturday afternoon: After a lull in the evening, snow should intensify overnight into early Saturday before tapering to flurries Saturday afternoon. Accumulations of 1-3 inches seem likely.
Total storm accumulations should be in the 2-4 inch range for Bowling Green although if convective banding sets up over the region, there is the potential for higher snow totals.
Check out the tight snowfall gradient over Warren county from the 18Z NAM. This model has 12+ inches over far NW Warren county and 2 inches over far SW Warren county.