Carvers Gap

You are not alone in middle and west TN. Even in Knoxville, the over performed early as the cold air made it that far and stopped. Precip shield expanded northward. Again, the NAM had it - but for the wrong reason (over amped). All modeling missed it until the RAP and HRRR came on board. The model that correctly modeled the cold air intrusion to Knoxville and then stopped it....that is the one that is correct...and the model which correctly backfilled the precip track over middle and west TN without having track issues over E TN. The ICON is the winner in my book. It got both scenarios right with the first and second wave it appears.

I definitely recommend at least trying a dedicated observation thread next time. The problem is that we have model talk and observations buried in the same thread. My two cents, and my opinion is probably worth exactly that. We(NE TN) are about 12-14 hours behind the onset of precip in western areas. While observations are (and we HUGELY appreciate those as it makes it easier to see which model is actually initializing correctly) rolling, the storm is still being modeled in far NE areas.

For several days, the NAM busted badly in the eastern valley as it didn't have any snow east of the Plateau. Globals did much better over here. The final run of the GFS before the event is probably going to verify for the most part with the exception of the foothills which mostmodels basically failed w/ until the last minute.. The Canadian sniffed out the problems in the eastern foothills first. Once the NAM came on board yesterday, it also had it. The HRRR also had it to some extent when it came into range. It is also worth pointing out that the NAM doing well in middle and west TN while not doing well here is easily explainable. It was closer in time to the event and those areas got cold quickly(better ratios). On this end, we were still several hours behind from a modeling standpoint...and it overdid accums on the Plateau and Cumberlands which was feedback. I do think the NC lee side low caused it problems. Once it lost that feedback(and we got closer in time), it did much better. It didn't dial-in here until about 12-16 hours to go. The other problem is that as the system backed and slowed, the NAM tended to feedback. The ICON is probably gonna be the big winner in all of this statewide, and that is probably a shock to most - me for sure. The 3k NAM was particularly good. The 12k NAM was terrible. The RAP and HRRR have been really good as well. I don't like the RAP, but it has done decently this go around. As MRX noted, no model had the warm air pushing to I-81 in NE TN.....personally, I think the front stalled at dawn at the Plateau and then daytime temps caused issues. The front stalling on the Plateau is not unusual. As Tellico noted the other day, Knoxville has a nice pathway for cold air from the Plateau. Chattanooga and TRI do not. The big red flag was when temps in NE TN yesterday beat forecast highs by 8 degrees. That produced a lot of warm air to scour as we got cloudy at dark, and warm air couldn't scour. Then it go banked against the Apps this AM. The NAM itself is I think a false positive. It got some things right, but for the wrong reason IMO. It got incorrectly amped and missed the track, but its snow totals worked out. Why? The cold air moved aggressively into middle and west TN and allowed rates to be crazy good. I have roughy 4" of snow right now. My ratio might be 8:1. If it was colder, I would likely have double that. So, whichever model moved the cold in most aggressively is correct...and I don't which one that is....probably the RGEM? The other thing is that the over-amping of the NAM gave it some semblance of expanding the precip shield. Its precip shield was juiced due to feedback and had a bad track in the East - plenty of snow fell east of the Apps which it didn't have. In reality, that precip shield was going to fill in as the rush of cold air squeezed every drop of moisture out of the air - it was impressive. The entry point of heavy precip was decently modeled and didn't change much. Still looking through some of the reports from middle and west TN which are incredibly helpful. I think the main bust(and let's be honest, nobody cares if a model busts and they get more snow than shown), is the foothills...and that may take some time to figure out.

Yes.

Yeah, man, you all are under that storm prediction center disco from 3:00-5:00 for 1" rates. Have fun.

Humongous slug of moisture about to head through the eastern valley.

Running about 3"+ here as well. Beware, the rain snow line is at John B Dennis. It moved north along I-26 from Eastern Start to John B in about two hours. We still have snow here, but hoping maybe we can hold on until temps start going down as the sun drops lower in the sky.

The wooly warm was right, Holston...and the elk. Go feed that elk, man. Keep it around and keep it happy.

It is right on the line. I just drove KPT to Gray. I-26 is awful from John B to Eastern Star and then it flipped to heavy mix and then rain by Gray. Main snow axis appear to be just to the left of I-81. Colonial Heights should be getting blitzed. Has changed back to moderate to heavy snow now at Gray.

Haha. Yeah, go for it. I think my spell check and lack of proofing got me there...but it isn't all wrong! LOL.

Scientific method. Let's see which one does the best w/ this set-up: Short range models vs NBM:

Probably wise.

One other note on using deterministic global models once inside of 36 hours, they don't have the ability to dial things in like short range models as they lack the hi-res ability to do it. But somehow, they still outperformed the mess scale models IMHO. The ultra hi-res, radar based models did well... So far(and this is like keeping score at halftime), I would rank globals in this order: 1. ICON 2. GFS/Euro 3. CMC(though the CMC picked up on mixing along the mountains first...) Short range modeling(1-2 days out): ....No winners so far. They were all bad. I will probably use global deterministic runs in the future and then switch to RAP/HRRR at the last minute. GFS trends were money at the last minute. Models inside of 24 hours: 1. RAP - for middle and west TN. Not great E TN. But it picked up the expansion of the precip shield northward. 2. HRRR ***The NBM will win this. I don't have it ranked as it is technically not a model, but blends other models. It will be my go-to next time around.***. Each model had its failing and some surprise biases, but the NBM was able to fish out the best outcome IMO.

So far..........The victor will be the later runs of the NBM in my opinion, the ICON in middle and west TN, and track goes to Euro but missed on northern extent of precip(does it have a dry bias???). The GFS absolutely nailed this with its later runs. The Canadian/RGEM package looks sketchy over NE TN. The NAM was terrible - woeful in terms of an amped track. It was bad enough that I won't use it any more. The RGEM probably will win track also but lose on northern precip shields, and lose for Chattanooga. Systems w/ two pulses of snow are notoriously difficult to predict, especially the second. Onsets(bulk fell after dark) of precip helped middle and west TN as precip arrived at night. E TN will be fighting daylight today. In the end, the NBM which blended the modeling probably is going to score the easy win. And it is important to note...Most of this is preliminary as the second half of this has just reached areas north of 40 in the eastern valley. But short range model runs are gutting the west side of this second line - and that is where the NAM likely has met its Waterloo. Knoxville should do really well, and that was on pretty much any model that had breath in its lungs. El Nino systems tend to over perform on the northern precip shield, and this one did yesterday evening and last night as well.

...a WIN for the Euro in E TN for sure w/ the second line. The middle and west precip shield was the error on that model and not the track. The models which were amped with this second wave(E TN) are gonna be wrong, and that includes the NAM. This second wave is very suppressed. The initial wave was about right, but the precip shield was underdone to the north. The NAM whiffed. The HRRR is shifting east with each run. Chattanooga is an exception as the Euro was too SE w/ it.

Snowing in Kingsport pretty good now - high side of moderate.

Very fine flakes of snow are now falling here.

Not trying to one-up Tellico. Here is the entire run. Is it right? IDK. But modeling at 0z has been bumping up numbers, and that gets my attention.

It is even more than that.

Seriously, go look at the GFS.

The 0z GFS is big.

Totals on models at 0z are so far trending upward at 0z for E TN - 3k/12 NAM, ARW, ICON, FV3

For posterity. WSWs for portions of 11 states. That is one cold map. Not to mention the snow in Buffalo and like the fourth coldest NFL playoff game ever in KC yesterday. What a cold snap.

The 0z ARW just went nuts.