Winter Storm Update: Models Diverging not Converging

qpf= amount of water that will fall

I have looked at the US 00z models which come in during the evening so this analysis includes the 00z Nam and gfs, Wednesday’s 12z Canadian, euro and several other overseas and less known models and there 12z runs. Normally as it gets closer to crunch time the models begin to arrive at a common solution however the 00z nam and gfs are far apart on qpf totals. The track of the low on the modeling is in a tight cluster from south Texas to the Florida panhandle and off the east coast and the 540 line on the modeling has been projected on the other side of the Tennessee line consistently for several days now all of which suggests snow will fall.

Let’s go over the meteorological set-up and questions I have with this storm.

A cold front will move through Wednesday night stretching from Texas to eastern New England by late Thursday. At the same time low Pressure will be forming over South Texas as cold arctic air moves into the region. So after temperatures in the upper 30’s to low 40’s for the middle of week they will fall off into the mid 20’s by the time moisture begins moving northeast. This low pressure will send moisture northward ahead of it into the Tennessee and lower Ohio valley Friday and early Saturday this moisture will be lifted along the front during the Friday to produce snowfall here. There are significant QPF differences in the modeling, more so than earlier today. The first one is how quick to bring in the precipitation Friday the nam holds it off while the gfs brings it in early Friday, both of these projections fit the typical biases of both models and both will likely need to correct with regards to this difference.  Further south across the deep south  thunderstorms will form ahead of the low these thunderstorms may act to take moisture destined for snowfall over South Central Kentucky, models may overdo this such as Wednesday’s 12z gfs or miss it, this variable is still a uncertainty and may lead to more changes in expected qpf. The third uncertainty I  see with this system is what the wrap around band will contain late Friday night into Saturday this is when I think snow ratios have the potential to rise a bit. I do see  a minor  feature in the upper levels at the 500 mb level early Saturday on the models which may serve to enhance this back period of snow but right now the qpf ranges from a few hundredths to .1 from this, but I could see this going higher as the models get a better sense of any potential shortwave that could enhance the back edge.  

Model Solutions

The 00z GFS was the most aggressive with regards to qpf with the 12z Canadian and 12z Japanese models fairly close to this solution yielding about an inch liquid from the storm from Friday morning into Saturday. The ECMWF qpf forecasts are not public but the track of the low on this model   is similar to the 00z GFS however several NWS offices have referenced the ECMWF being quite a bit lighter in terms of total qpf. The 00z NAM was quite a bit lighter with the 00z run and skipped Bowling Green with a half an inch of qpf west across paducah and east toward Pikeville future runs will determine if it is hinting at something in terms of precipitation patterns with the storm or a minor one run glitch, the 18z nam had .5 qpf with the new 00z run around .3.  The UKMET model was slightly moister would likely be around .4 or .5 at the completion of the storm and is similar to Wednesday afternoon runs of the GFS ensembles.

Forecast

Taking the model spread literally would result in a forecast of 3-10 inches using 10:1 ratios these ratios can  shift either way

A: sleet mixes in which is possible as the 540 is close this would potentially lower the ratio a bit

B: temperatures in the 10’s and 20’s Friday night can raise these ratios leading to dryer fluffier snow, the clipper earlier this month at Lexington brought 30:1 ratios with temps in the 20’s however with this being a southern storm with more moisture they will much closer to the traditional 10:1 but still do have the potential to go over 10:1 since the surface temp will be well under 32.

My thoughts are to tighten that spread of totals up a bit given we still have another set of runs to adjust the forecast. For now the forecast is for 4-8 inches of snow this is slightly up from the 3-7 inches I issued for the stormtoppers earlier tonight primary based on the 00z gfs bringing close to an inch of liquid which couldn’t be ignored but this forecast also indicates the gfs is overdone with the qpf. Like with the stormtoppers forecast I would tend to lean more to the lower end of the totals for now.

Detailed  timeline

– Wal mart and Kroger will be mobbed On Thursday with all of you getting bread and milk

– snow overspreads the region Friday morning with temps in the 20’s throughout the storm.

–  snow may become moderate to heavy at times Friday afternoon making for one of the worst evening commutes due to weather in a long time as roads will become snowcovered in many areas. If schools don’t close expect your kids to eat a 9 am lunch and to be home by noon based on this forecast. I could even see businesses and maybe even WKU closing early on Friday if the GFS solution is correct, but at the same time the NAM solution would yield an advisory level event with much less impact.

– Wind will increase Late Friday into Saturday and snow tapers off. Very cold weather will last into Sunday.

– storm total 4-8 inches

Another discussion will be posted around noon Thursday.  Just a note as I’m finishing this write-up the snowfall projection map which takes into account ratios on the 00z gfs has 14 inches in Bowling Green. However the 12z model runs will continue to adjust snow totals and remember as you may jump for joy after reading that, the 00z nam has just 4 in Bowling Green.

MG