NYC/PHL December 24-27 Potential - Part 2

[Rainshadow]

The 0z euro ensembles (which seem to have the highest verification scores over the long term) had this thing 150 miles SE of the BM-- so your call for sne would work out well. We'd just be on the fringe of the moderate snow here in southwestern long island.

If I were keeping a December scoreboard for what has done well recently for PHL it would be

1) Can ggem

2) Euro ensembles

3) most of the rest have been inconsistent

4) ukmet

5) nogaps, et al

[Rainshadow]

Tony, what about the probabilistic occurrence of 1.0" or greater snowfall events-- I guess thats the minimum definition of "snow cover."

BTW was 2000-01 a weak or mod la nina winter? I see the Millenium storm analog being mentioned now and then (also *gulp* March 2001 lol.)

I think it would universally be looked upon as weak. (00-01). The millennium storm has been popping up on the CIPs site.

[cnjraider]

Excellent combination of history and statistics! Thanks. So if I understand you correctly, 3.5% chance of seeing measurable snow and approaching 0% of seeing significant snow. A snow lover would not like those odds but that is as close to reality as you can get.

[A-L-E-X]

Long range is looking ugly again, looks like the Nina is roaring back once more. I wouldn't be surprised if this was our last big threat for the season. How amazing that despite the setup aloft looking very different from last weekend, the track is nearly identical to that big bust. But what are you going to do, at least I won't have to do any shoveling or watch out for black ice.

How is it that Will and Don are talking about the return of major blocking in the first week of Jan?

[Bobby]

Long range is looking ugly again, looks like the Nina is roaring back once more. I wouldn't be surprised if this was our last big threat for the season. How amazing that despite the setup aloft looking very different from last weekend, the track is nearly identical to that big bust. But what are you going to do, at least I won't have to do any shoveling or watch out for black ice.

Long range is just that, nothing more than a prediction. As faras last big threat, I think that may be a LITTLE extreme Lol. We're attempting to predict the unpredictable and looking to do it 4-5 days out. Worry if the current scenarios are here on Friday/Sat afternoon.

[A-L-E-X]

Tony, what you said about mod-strong la ninas applies to our "snowy" mod-strong la ninas also-- 1916-17 is the posterchild for a 50" inch snowfall winter with a bunch of moderate events lol.

1909-10, 1916-17, 1917-18, 1955-56 were our "good" mod-strong la nina snowfall winters, and as a matter of fact 1916-17 and 1917-18 was the only occurrence of back to back 50" snowfall winters at NYC (with identical numbers lol-- although 1917-18 was much colder, with a -13 in Dec and several well below zero lows.)

[cnjraider]

Excellent combination of history and statistics! Thanks. So if I understand you correctly, 3.5% chance of seeing measurable snow and approaching 0% of seeing significant snow. A snow lover would not like those odds but that is as close to reality as you can get.

One question though, do the odds go up based on where we are at currently with all the models? Does the percentage go up? Again, I wouldn't want to have to forecast this to keep the public informed and yet not to incite histeria among them or the media.

[Rainshadow]

Excellent combination of history and statistics! Thanks. So if I understand you correctly, 3.5% chance of seeing measurable snow and approaching 0% of seeing significant snow. A snow lover would not like those odds but that is as close to reality as you can get.

Well its not zero percent. Once (will it ever?) this event gets within 72 hours, climatological tendencies get tossed and the models will converge on a solution. There are exceptions to every rule, you have to question everything be as cynical about the ecmwf solution as one is about the gfs solution.

[A-L-E-X]

Where do you see that because I don't see it? The west based -NAO is gone by Day 8.

New England subforum and also Don's thread on the AO

They all talk about the west based nao being gone in the last week of the year but returning in the first week of the new year-- Will even said that the down time is getting less and less and the euro is getting more and more bullish about a -2 SD neg nao in January.

[A-L-E-X]

Well its not zero percent. Once (will it ever?) this event gets within 72 hours, climatological tendencies get tossed and the models will converge on a solution. There are exceptions to every rule, you have to question everything be as cynical about the ecmwf solution as one is about the gfs solution.

The models are holding the carrot in front of our noses and like donkeys we are chasing after it, not knowing as we move forward, so does the carrot

[Rainshadow]

One question though, do the odds go up based on where we are at currently with all the models? Does the percentage go up? Again, I wouldn't want to have to forecast this to keep the public informed and yet not to incite histeria among them or the media.

No, that's just a summary of what has occurred during the past 13 moderate and strong la nina winters at PHL. Given the synoptic situation I think the chances are greater than 3.5%. Its beyond me why there would be any public hysteria at all. Didn't we have three or four of these last winter?

[A-L-E-X]

Well its not zero percent. Once (will it ever?) this event gets within 72 hours, climatological tendencies get tossed and the models will converge on a solution. There are exceptions to every rule, you have to question everything be as cynical about the ecmwf solution as one is about the gfs solution.

Just wish it was a weak la nina then you could have brought in all your wonderful weak la nina stats, especially weak la nina after el nino lol. Those would look much better!

[cnjraider]

Well its not zero percent. Once (will it ever?) this event gets within 72 hours, climatological tendencies get tossed and the models will converge on a solution. There are exceptions to every rule, you have to question everything be as cynical about the ecmwf solution as one is about the gfs solution.

We just crossed paths in messages. I do believe you have to increase the odds based on where the models are today versus random day selection.

[BethlehemBlizzard]

The probability approaching zero is true for any given day in any given winter such as this. But, if you are given the current set-up, the probabiility certainly increases.

[_AR_]

So what is the current overall analysis?

06Z GFS I see brings the low out of the GOM and OTS.

What are we missing that is needed to keep this in to the coast?

[cnjraider]

We just crossed paths in messages. I do believe you have to increase the odds based on where the models are today versus random day selection.

Is there a way to look at the situations where you had similar threats and see where they led to statistically?

[SnowGoose69]

Well its not zero percent. Once (will it ever?) this event gets within 72 hours, climatological tendencies get tossed and the models will converge on a solution. There are exceptions to every rule, you have to question everything be as cynical about the ecmwf solution as one is about the gfs solution.

I'm not sure what the average model error on track at 72 hours for East Coast storms is, but I'd say around 100-150 miles or so? It seems rare we see a shift much more than that.

[Rainshadow]

The probability approaching zero is true for any given day in any given winter such as this. But, if you are given the current set-up, the probabiility certainly increases.

Which has been stated.

[Rainshadow]

I'm not sure what the average model error on track at 72 hours for East Coast storms is, but I'd say around 100-150 miles or so? It seems rare we see a shift much more than that.

For the 2008-9 winter season (last stats I have seen) the average model error was about 150 miles.

[Rainshadow]

Is there a way to look at the situations where you had similar threats and see where they led to statistically?

That would be tough, because all near misses can't be documented.

[SnowGoose69]

For the 2008-9 winter season (last stats I have seen) the average model error was about 150 miles.

Considering in 1988-89 it was probably 150 miles at 24 hours thats a major improvement...of course most average citizens don't realize how good a number 150 miles in 3 days is.

[Rainshadow]

We just crossed paths in messages. I do believe you have to increase the odds based on where the models are today versus random day selection.

Yes. Strongly agree.

[A-L-E-X]

Considering in 1988-89 it was probably 150 miles at 24 hours thats a major improvement...of course most average citizens don't realize how good a number 150 miles in 3 days is.

Especially that with the population density around here, 150 miles means much more than if it were, say, central nebraska lol.

[Rainshadow]

Considering in 1988-89 it was probably 150 miles at 24 hours thats a major improvement...of course most average citizens don't realize how good a number 150 miles in 3 days is.

I can't prove it, but I would think the average error for stronger systems is less.

[A-L-E-X]

Considering in 1988-89 it was probably 150 miles at 24 hours thats a major improvement...of course most average citizens don't realize how good a number 150 miles in 3 days is.

Looks like a 3 fold increase in accuracy.

[Rainshadow]

Especially that with the population density around here, 150 miles means much more than if it were, say, central nebraska lol.

Alex some of that error could be temporal (slower/faster) just as much as it could be latitudinal or longitudinal. They are not separated.

[A-L-E-X]

That would be tough, because all near misses can't be documented.

Plus, you'd have to qualify storms that passed over us or just inland as near misses also, not just barely OTS ones.

[SnowGoose69]

Here comes the NAM...not that it will tell us much really, maybe it will push the storm back another 12 hours

[A-L-E-X]

Alex some of that error could be temporal (slower/faster) just as much as it could be latitudinal or longitudinal. They are not separated.

Ah ok-- so it actually means the improvements could be even better than 3x the accuracy, if we account for the temporal aspect.

[SnowGoose69]

Alex some of that error could be temporal (slower/faster) just as much as it could be latitudinal or longitudinal. They are not separated.

I swear storms seem to start later than modeled the last decade or so, before that they always seemed to be a bit early..almost like the models overcompensated a bit for the fact they were always late.