GaWx

Yep. 850s are +2 or lower for 3/4 of the qpf keeping ZR much lower than GFS; total qpf ~0.95” CMC, similar to gfs but much less than 1.65” of icon

With 850s at +4 for the majority of the qpf on the 12Z GFS, mainly ZR at RDU would agree with the past….if those 850s were to verify.

12Z GFS:

From what I experienced in ATL and SAV (and confirmed by what I’ve studied), the conversion of ZR to accrual on branches/bushes approached 50% under ideal accrual conditions (winds not too strong, rain not too heavy, temps not right at 32, etc).

I think that’s conservative at this point per model consensus, which suggests at least ~twice that much accretion is quite possible. But a long ways to go.

That (12Z Icon) could mean ~0.7” of ice accretion on trees in RDU, a huge problem that would be similar to 2002. It takes only 0.5” of accretion for widespread power outages:

12Z Icon for RDU suggests ~0.1” qpf of snow, ~0.2” qpf of sleet, and ~1.35” of ZR. Temps get to as low as 27F at end of ZR.

The preliminary AO for today per CPC is -5.26, well predicted. This would compare to the alltime record -AO for Feb, -5.29 (2/5/1978).

I’ve found based on analysis of 850s for SE winter storms going back many decades that +1 to +2 850s tend to be sweet spot for sleet and +3 to +5+ tends to be sweet spot for ZR.

That’s the point. 2002 was Miller B/quite different. Miller A (upcoming storm) typically not as associated with major ice. But model consensus still shows a lot of ZR. Maybe that will change.

For comparison, here are the sfc maps for Dec of 2002 Miller B major icestorm for NC:

The 18Z Euro AI is similarly quite suppressed to the 12Z.

18Z GFS:

It did very well for the early Jan 2018 historic coastal SE winter storm. I remember it clearly as I was following it extra closely for obvious reasons. It was the first global to move the offshore low NNE instead of NE, which resulted in much higher qpf on the coast and that verified. So, this was at least an exception.

18Z Icon 120 850 0C line further SE than 12Z Icon 126 in NC/NW SC/N GA (toward the 12Z Euro AI):

12Z Euro AI way colder/further SE 850 0C line than other models (hour 126, which is as of 18Z on Feb 19th): Fayetteville to Gainesville: 12Z Euro op: N of RDU to TN 12Z GFS: S VA to TN 12Z CMC: S VA to TN 12Z Icon: S VA to TN

0Z GEFS (see below) and other models remain insistent on a rare sub -5 AO tomorrow. Today’s preliminary is already down to -4.8. There have been only 12 such periods since 1950 (average of once every 6 years). In Feb, itself, there’s only been 4 periods of sub -5 (one every 15 Febs on average): -2/10-11/2021 (only La Niña Feb instance): major GSO icestorm 2 days later -2/6-7, 14/2010 (El Nino): 2 SE US winter storms (2/5, 2/12-13) -2/5-6/1978 (El Nino): several SE light snows 2/2-9 -2/13-16/1969 (El Nino): major winter storm 2/15-17 Charlotte to Asheville to GSP Today’s 0Z GEFS based AO forecast:

Yes, to be fair, those historic Euro AI snow maps were for when this week’s activity was 8-9 days out. Now we’re within 6 days. So, perhaps those too cold misses (including way too cold 850s) aren’t as relevant now. Also, I now realize that the S portion of StormVista snow maps have to be added to the non trusted snowfall map list. So, a portion (not all and probably not even most) of the overdoing of snow on the insane SV Euro AI maps may have been due to SV algos. But regardless, the 0C 850/snow line verified way too far S.

1) As just mentioned it had very heavy (historic in some cases) snow on several runs in a row very far to the south (as far S as NE GA, upstate SC) for earlier this week, for example. However, to be fair, those runs were further out than 5 days. 2) For the Jan 10-11 storm, it had warming S winds ahead of the N Gulf low on the backside of the high moving offshore and ahead of a Clipper. That hardly ever happens. It ended up dead wrong at the surface. 3) I’m not disputing what you’re saying, but I don’t know exactly what you’re referring to and what exactly the verification measures. It could have been an H5 verification for all I know. The main problems may lie underneath 500 mb. Do you have a link or can you post the graphics? 4) I’m not saying it couldn’t end up right with the very suppressed track. With the preceding -5 AO, it could occur despite La Niña. What I’m saying is I’m taking it with a huge grain while it is on its own. 5) Are you favoring this very suppressed track?

The Euro AI has not been good overall this winter imo as I just posted. It especially blew the most recent storm’s snowfall, way overdoing it with 850 0C line way too far south and thus having historic snow amounts way far south of the actual snow line (though Stormvista snow maps likely overdid the amounts somewhat). So, I’m taking this very SE suppressed 12Z Euro AI with a gigantic grain at this point:

This Stormvista 12Z EPS snow map is way overdone in some of the southern portion. Here’s the comparable WB 10:1 map:

And unlike some other snow maps, these WB 12Z Euro Kuchera maps look legit based on the 850 mb 0C line. There’s a short period of non-snow just N of the NC/VA border in your vicinity when 850s go >0C (centered on 6Z of Feb 20th), but otherwise it is snow.

12Z Euro ZR looks bad for many in NC, especially Triangle vicinity and S/SE of there: Low track 12Z Euro: Sleet:

12Z Euro Kuchera:

Looking at this more closely, this 12Z Pivotal UK map has 2.5” of snowfall at RDU. Based on 850s, RDU gets no snow, ~0.2” of qpf as sleet, and ~0.5” qpf as ZR. Based on that, that’s ~0.6” of sleet (and no snow) vs this Pivotal map’s 2.5”. The SE/S portions of Pivotal UK snow maps are often not a reliable indicator of what the UK is actually generating. Algorithm issue at Pivotal for the UK snowfall maps is the problem.