Central PA Winter 25/26 Discussion and Obs

[Mount Joy Snowman]

3 minutes ago, canderson said:

I plan to be a first shovel at 1 p.m.  if it flips after it’ll have snow on the grown, if all snow basically cut in half. Praying to baby Jesus it’s all snow and I get 12”. That’s my goal - a foot. 

My brotha, honestly I think you're good now.  I think the NAM is overamped as per normal bias and the battleground is through Lanc/York.  Doesn't mean you won't see some sleet but I doubt it's meaningful.  I've talked to some other people and they'd be shocked if Harrisburg didn't get 12+ and honestly 15".  The NAM will get credit for sniffing out the sleet, as it should, but it will overplay its hand.  Can't wait for tomorrow!

[Jns2183]

[mention=2304]psuhoffman[/mention] just made a great post about the model situation for this storm.  

“So I’ve never seen this drastic of a split between the globals and American run CAMs at this range. What’s odd about this and gives me no past reference to draw on is typically when we see this kind of thing the euro and rgem/HRDPS kind of bridge the gap as those 3 are decent at seeing mid level warming. 
But they are all in the camp with the other globals (UK/Icon/gfs).  The impact is most drastic actually for Maryland. Around DC and south we’re talking maybe the difference between 4-5” and 6-8”. But for places NW of 95 in extreme NW VA up through central MD we’re talking the difference between 5-6” and 10-12”!   
The divergence seems simple. The globals along with the Canadian high res models have an intense WAA band over the area from 12z-18z that the American CAMs do not. That’s why they are warmer.  Less dynamic cooling to fight off the WAA at mid levels. Also less precip. That combo means 6” instead of 12” for places like Winchester-Frederick-westminster. 
What this comes down to imo is which camp is correct about the precip representation from 12-18z. Unfortunately I don’t have any great insight. Usually here is when I’d be saying “in this or that situation this is what happened” but I can’t remember a single case like this to draw upon. 
I guess I’m gonna ride with the euro camp. It would be hard to take the NAM and a bunch of experimental stuff over the highest verification tools we have. But on the other hand those CAMs were designed for this. When they score the euro a 30 mile shift in 800 mb temps and a meso scale precip band aren’t really going to impact those scores at all!  
Yes I just contradicted myself. If I had to make a forecast maybe I’d hedge and go in between even though that’s probably not the most likely outcome, one of these camps is going to win”

Look way above at the snow ratio table image. What made me confused as hell was NAM being less than half that of the SREF. The nam is one of the members of the SREF and its core engine accounts for half the ensemble members just with physics tweaked a bit. Usually SREF and nam are pretty close. This tell me the cam ensemble members are mostly much much higher than nam. It's weird

Sent from my SM-S731U using Tapatalk

[canderson]

6 minutes ago, Mount Joy Snowman said:

My brotha, honestly I think you're good now.  I think the NAM is overamped as per normal bias and the battleground is through Lanc/York.  Doesn't mean you won't see some sleet but I doubt it's meaningful.  I've talked to some other people and they'd be shocked if Harrisburg didn't get 12+ and honestly 15".  The NAM will get credit for sniffing out the sleet, as it should, but it will overplay its hand.  Can't wait for tomorrow!

Can’t wait - regardless we’re getting is heavy snow! Reason to celebrate. 

[Jns2183]

I've also never seen a sleet changeover blow through at 60mph like it wants

Sent from my SM-S731U using Tapatalk

[Blizzard of 93]

18z RGEM looks great for all of us.

[Ruin]

when does the nam start 

[Jns2183]

Same as before

I mean it also never backed down in Oklahoma and Arkansas as was and this was the result

The NAM Split: The NAM is the most interesting case. For liquid rain, it has a staggering 0.96 correlation, meaning it mapped the storm's shape almost perfectly. However, for snow, its correlation plummeted to 0.18. This proves the NAM understood where the moisture was, but failed completely at the "transition physics"—it thought the air was too warm, keeping it as rain when it should have been snow.

Every other model had snow correlation between .5 to .7

Sent from my SM-S731U using Tapatalk

[Mount Joy Snowman]

The NAM just won't quit, slightly cooler but not enough to matter, an I-80 to Route 6 jackpot haha.  Lancaster 5-8".  It's going down with the ship one way or another.  Tomorrow will be quite illuminating. 

[Jns2183]

The HRRR Consistency: The HRRR was the most balanced. It had the lowest error across both categories, making it the most reliable tool for this specific event. It handled the "Rain-Snow line" better than any other model.

Sent from my SM-S731U using Tapatalk

[canderson]

It’s NAM vs NWS 

[canderson]

Just now, Jns2183 said:

The HRRR Consistency: The HRRR was the most balanced. It had the lowest error across both categories, making it the most reliable tool for this specific event. It handled the "Rain-Snow line" better than any other model.

Sent from my SM-S731U using Tapatalk
 

It busted pretty bad in Oklahoma and ET iirc - it has Tulsa getting a ton of s kw that hasn’t happened 

[Jns2183]

The NAM just won't quit, slightly cooler but not enough to matter, an I-80 to Route 6 jackpot haha.  Lancaster 5-8".  It's going down with the ship one way or another.  Tomorrow will be quite illuminating. 

It's driving me nuts. Something is there but after how I saw it do so far, a pixel analysis correlation of .18 vs every model being 2.5x-3.5x higher because of how bad it screwed up the changeover line

Sent from my SM-S731U using Tapatalk

[Jns2183]

It busted pretty bad in Oklahoma and ET iirc - it has Tulsa getting a ton of s kw that hasn’t happened 

In Arkansas it was best model for snow. The GFS, euro were close.

Sent from my SM-S731U using Tapatalk

[Mount Joy Snowman]

5 minutes ago, Jns2183 said:

The HRRR Consistency: The HRRR was the most balanced. It had the lowest error across both categories, making it the most reliable tool for this specific event. It handled the "Rain-Snow line" better than any other model.

Sent from my SM-S731U using Tapatalk
 

God I hope you're right.  That's why I was so locked in on the 0z HRRR.  Would be a fun battle zone down here, not a pure takeover like the NAM.

[Jns2183]

Here

Sent from my SM-S731U using Tapatalk

[Ruin]

yeah was afraid the nam would nam us but the precip shield to the south west looks ugly dry 

[Ruin]

16 minutes ago, Jns2183 said:

The HRRR Consistency: The HRRR was the most balanced. It had the lowest error across both categories, making it the most reliable tool for this specific event. It handled the "Rain-Snow line" better than any other model.

Sent from my SM-S731U using Tapatalk
 

I remember several times last 3 years the hrrr has been really bad 

[GrandmasterB]

Stolen from MA and the LWX office. Interesting:

If anything, perhaps some guidance
leans slightly slower in terms of a transitioning to sleet
Sunday morning. However, not buying into that just yet, and
would like to see the 00z guidance roll in with the latest upper
air data around the country. Could be something to watch though,
as a slower onset of sleet in any one given location could make
a significant difference in snow amounts. Will be something to
watch this evening into tonight."

[Jns2183]

I remember several times last 3 years the hrrr has been really bad 

It definitely has. I'm just starting what was the results of analyzing model performance in Arkansas up to 6pm today. I'll be able to start looking at states to East soon. All the models there had about the same mean absolute error of 1.8". It's was the nam though with a root mean square deviation of almost twice that, while others were close, along with a massive dry bias, and horrible snowfall distribution correlation that stood out. Below is a metaphor

Thinking about these weather models is a bit like judging a free-throw shooting contest between five different players. Even if they all missed by the same average distance, the way they missed tells the real story.
Imagine each model is a basketball player trying to hit a specific spot on the rim:
The "Average" Score (MAE)
Every player in this contest missed their target by about 1.8 inches on average. If you just looked at that one number, you’d think they were all equally "okay" at their jobs. But once you look at the game tape, one player (the NAM) stands out for all the wrong reasons.
The NAM: The "Wild Card" (RMSE & Bias)
The NAM was the player who didn't just miss—it missed spectacularly.
The Big Misses (RMSE): While the other players missed by 1 or 2 inches consistently, the NAM would hit the backboard or miss the rim entirely on some shots. Because RMSE penalizes big mistakes much more than small ones, the NAM's "penalty score" was twice as high as the others.
The "Dry" Excuse (Bias): On top of the wild misses, the NAM had a massive dry bias. In our metaphor, this player was consistently shooting way too short. If the hoop was at 10 feet, the NAM was aimlessly throwing the ball at 8 feet.
The "Broken Rhythm" (Correlation)
Finally, there’s the snowfall distribution correlation.
In plain English, this is rhythm. A good player might miss, but if the target moves left, they move left too.
The NAM had horrible correlation. It was essentially playing a different game. When the actual storm "moved left," the NAM "moved right." It didn't just get the amounts wrong; it got the entire pattern of where the snow would fall completely backward.

Sent from my SM-S731U using Tapatalk