December 19-20 Talking Points - Part 3

[Saturn510]

It is hilarious though that when we have models that show a snow storm and the euro does not. The storm aint happening. This time we have the Euro model within 96 hours showing a major dumping almost two runs in a row and it turns out to be totally hogwash.

[Stormlover74]

Also its more forgivable for the GFS to show a big hit 5 days out, much less so for the Euro 3 days out IMO

As far as getting a storm in this type of winter I guess we go back to climatology. I'm assuming we've gotten some good snowstorms in other moderate to strong Ninas...so when and how?

No, the 00z run yesterday gave the NYC area several inches with a significant hit for Long Island.

[famartin]

Didnt the Canadian and Kma and Jma have out to sea and then the 12 z runs had closer to the 40/70 benchmark

That's not a jump. That's a nudge. The EC is the only one that jumped.

[famartin]

And that the shift happened at 12z yesterday, when most were expecting better sampling of the southern stream s/w. At least that's why I bit harder on one run than I normally do.

That was the other key. It failed.

[eduggs]

All hope not lost. Several chances for light snow next week with something to track for Christmas. And of course there's still a very small chance that the Sun/Mon storm surprises.

[Stormlover74]

That's why we can't always trust that we know what we think we know. Throw out the 06z /18z runs, the ensembles are always east of the OP, this model has a SE bias, the EURO is unbeatable within x amount of hours. All of these assumptions have proven to be at least somewhat false with this storm.

And what's weird is that typically the GFS is too far south and east on coastal storms too initially and then tend to move NW as we get within a day or two.

[famartin]

It's just so shocking that the Euro (supposedly the best model out there) would make such a terrible blunder like that, whatever the reason for it may be. And it wasn't even just the Euro that had a significant shift west yesterday, the GGEM and SREFs also did. Very strange considering that was supposedly when we were going to start getting a better sampling of the southern shortwave.

I definitely think a lot of people's faith in the Euro (and all of the models, really) will be significantly lowered because of this event...I know mine will be.

Its lower somewhat. But its a reminder to be extremely cautious whenever a model that has been consistent makes a major jump... don't jump on that change immediately especially if its not well supported. No other model (save maybe a few ensemble members) had the bomb the EC had yesterday.

[famartin]

No, the 00z run yesterday gave the NYC area several inches with a significant hit for Long Island.

What he wrote was correct. He said that other than 12Z yesterday, 6 of 7 EC runs had nothing. Which means that 0Z Thursday was the 1 outlier.

[JERSEYSNOWROB]

All hope not lost. Several chances for light snow next week with something to track for Christmas. And of course there's still a very small chance that the Sun/Mon storm surprises.

Yeah even though we are still threading the needle here, that fact remains we still have a slight chance of this thing exploding closer to the coast. All the key players are on the field. We just need everything to time up just right. Miracles do happen.

[famartin]

Also its more forgivable for the GFS to show a big hit 5 days out, much less so for the Euro 3 days out IMO

This I agree with. But regardless, the EC corrected itself.

[eduggs]

That's not a jump. That's a nudge. The EC is the only one that jumped.

True, but the NAM, GFS, and GGEM all "nudged" noticeably west after such a disappointing cycle the night before. That primed everyone with optimism as the EC blizzard run rolled in. Everything just fit together so favorably in that time period to set up for disappointment.

I'm in agreement with you. I don't think the modeling was so terrible. Too much obsession with individual runs' QPF.

[rgwp96]

This I agree with.  But regardless, the EC corrected itself.

corrected itself,its been awful even in the begging of the week when the 0z would have a hit than 12 z didnt(or vice versa) than it was out to sea for numerous runs than gives nyc 8 inches than 2ft in consecutive runs?  Dont understand why everyone defends that model so much when it does poorly

[bluewave]

The old rule comes into play here.Don't start to get too excited about snow up here until the low gets to our

latitude within 72 hrs.12z yesterday made me a little nervous when I saw the low near Hatteras

at the 72 hr mark and all the good stuff here beyond that point.

[am19psu]

corrected itself,its been awful even in the begging of the week when the 0z would have a hit than 12 z didnt(or vice versa) than it was out to sea for numerous runs than gives nyc 8 inches than 2ft in consecutive runs?  Dont understand why everyone defends that model so much when it does poorly

Because it is right more often than the other models? Bayesian statistics? I mean, that's why I give it more weight.

[famartin]

corrected itself,its been awful even in the begging of the week when the 0z would have a hit than 12 z didnt(or vice versa) than it was out to sea for numerous runs than gives nyc 8 inches than 2ft in consecutive runs? Dont understand why everyone defends that model so much when it does poorly

If you average it out, the EC willl likely turn out better than the GFS still. GFS has been consistently NW of the EC, except for a small handful of runs which I've already outlined.

[RutgersWx92]

Its lower somewhat. But its a reminder to be extremely cautious whenever a model that has been consistent makes a major jump... don't jump on that change immediately especially if its not well supported. No other model (save maybe a few ensemble members) had the bomb the EC had yesterday.

I agree with this and you're right that no other model had the extreme solution that the 12z Euro had yesterday, but it definitely appeared that the other models were trending in that direction at 12z yesterday, and the 0z Euro also seemed to indicate that the 12z wasn't a complete fluke. Even pro-forecasters at Mt. Holly were fooled; yesterday they had an 80% chance of snow for Sunday for New Brunswick.

[Saturn510]

And honestly i dont recall good model outputs when we have radio shows.

[Snowlover11]

no more radio shows!

[njrealtor84]

Been reading that the WV images show the s/w holding stronger in NM than what was shown in models. Any Mets have thoughts on this?

[am19psu]

I agree with this and you're right that no other model had the extreme solution that the 12z Euro had yesterday, but it definitely appeared that the other models were trending in that direction at 12z yesterday, and the 0z Euro also seemed to be backing up the idea of a trend. Even pro-forecasters at Mt. Holly were fooled; yesterday they had an 80% chance of snow for Sunday for New Brunswick.

And HPC. There were a number of reasons to jump out ahead of this yesterday. It just turned out to be a giant failure.

[Weathergun]

And that the shift happened at 12z yesterday, when most were expecting better sampling of the southern stream s/w. At least that's why I bit harder on one run than I normally do.

I don't think the models had good sampling of the northern stream s/w until 0z and 12z runs today. It was still off the coast. This is may be why the models don't show any phasing closer to the coast. The southern stream s/w running too far ahead to northern stream s/w.

[am19psu]

And honestly i dont recall good model outputs when we have radio shows.

Clearly, the radio waves are causing interference with shortwaves over the Southwest

[eduggs]

corrected itself,its been awful even in the begging of the week when the 0z would have a hit than 12 z didnt(or vice versa) than it was out to sea for numerous runs than gives nyc 8 inches than 2ft in consecutive runs? Dont understand why everyone defends that model so much when it does poorly

It's not uncommon for successive operational runs to fluctuate within or even slightly beyond the scope of their ensemble population. That doesn't make the model awful, rather it shows the difficulty in modeling the atmosphere. When you take a look at the global verification scores it's hard not to be impressed. But it's always easy to nitpick a small error in a single run, at a particular location, when that error is the difference between snowstorm and sunny.

[Saturn510]

And Ray buying plane tickets to come to the east coast. Everyone finds ways to jinx ourselves.

[famartin]

Clearly, the radio waves are causing interference with shortwaves over the Southwest

Especially the shortwave radios

[famartin]

And Ray buying plane tickets to come to the east coast. Everyone finds ways to jinx ourselves.

Yup. Sorry

[ArtRosen]

So I guess even Long Island won't even get a few inches?

[tmagan]

If, you think about it, the track error for tropical cyclones three days out is what, 100 miles, so this should not be the biggest surprise in the world.

[tmagan]

So I guess even Long Island won't even get a few inches?

It may not even see any flakes.

[ace0927]

0z Thursday.

yup- i just went back and checked. so 7 of 9 had nothing