Jump to content
  • Member Statistics

    17,606
    Total Members
    7,904
    Most Online
    ArlyDude
    Newest Member
    ArlyDude
    Joined

Bust Case Study -Philly/Nyc and beyond-GFS coup? Euro fail?


stormtracker

Recommended Posts

While it's way too early to break out the champagne for the new GFS based on one storm, I think we could mostly agree that it did better than the Euro.  Could this be the beginning of a legit challenge or one fluke win?   Was the GFS that good or the Euro just had an errant fluke?

 

On the PR/Media side up there...man oh man...the MIC in Philly is apologizing on social media, people are tossing the Euro under the bus in the media....what happened?  Discuss...

Link to comment
Share on other sites

  • Replies 85
  • Created
  • Last Reply

At the least, I would think Mets would want complete model agreement for an area before pushing 2-3 foot button. From the moment the poss of this popped up three days ago, it was mass/eastern ct/Boston area that continually scored. At varying times nyc and certainly philly were on edge. Too many bullish calls for those areas too close to the edge of such a sharp cutoff.

Link to comment
Share on other sites

IMO, there's no reason to have your all-time record as part of the forecast range unless there's complete model unanimity. When there's not, a compromise forecast will still convey urgency-- "1-2 ft with locally more" is a fairly typical phrasing from Taunton for their big ones, which I think works. It's a huge range, but no one would be upset at them if 32" verifies somewhere given that forecast. The public is much more forgiving if the forecast is for a huge storm that ends up over-performing to reach historic levels than vice-versa.

Overall, I thought the RGEM proved its worth again as a model you have to factor in when making a forecast (not hug). No single model should be regurgitated as the verbatim forecast.

Link to comment
Share on other sites

IMO, there's no reason to have your all-time record as part of the forecast range unless there's complete model unanimity. When there's not, a compromise forecast will still convey urgency-- "1-2 ft with locally more" is a fairly typical phrasing from Taunton for their big ones, which I think works. It's a huge range, but no one would be upset at them if 32" verifies somewhere given that forecast. The public is much more forgiving if the forecast is for a huge storm that ends up over-performing to reach historic levels than vice-versa.

Overall, I thought the RGEM proved its worth again as a model you have to factor in when making a forecast (not hug). No single model should be regurgitated as the verbatim forecast.

 

Very good point. When you put out a forecast for say 10-20 from the beginning, the general public knows their getting a major storm so if you have to upgrade at some point during the storm it's understandable and may even go unnoticed by the casual observer. If the original forecast is reasonable then upgrading becomes very acceptable.

Link to comment
Share on other sites

Haha. DT hugged the Euro to an epic bust. He deserves every second of this bust with his ridiculous rant the other night.

The American models were insistent on a band setting up over nova and into central Virginia. They ended up being right.

 When he called TWC wrong for going 12"-20" he meant they were fooking idiots for being too high.  DOUBLE E RULE!!!!

Link to comment
Share on other sites

What I find fascinating is how much I have learned from the relative Philly/NYC bust on this one - seems very similar to the bust around DC and Balt from Boxing Day 2010. I don't know enough meteorology (duh) to begin to understand why, but know enough weenieology now that if the western edge of an area expecting big snow in a bombing nor'easter has massive drop offs in totals right behind it, to be very wary.  I don't know what had to happen for the Euro to have been right (H5 had to close off earlier?) but, whatever it is, maybe helpful to understanding future set-ups for this particular weenie. Was nice to be able to follow this one without caring much about my backyard (unlike the Boxing day fiasco).

 

I know some people love to see banding in the winter - but I mostly seem to get missed in those situations. Give me a wall of moisture covering the whole area into a cold dome 10 times out of 10 over hoping for stuff to come together right in a coastal.

Link to comment
Share on other sites

From my perspective it seems every year every model has it's day at some point even the chit ones. So far the nam and gfs and euro and Canadian and RGEM have all had a win this year to some degree....it happens every year...I say UKMET gets the Monday storm right....I sound like tenman but anecdotally it's not that far off.msure when you combine the overall performance over a period of time it all shakes out into a ranking but the term every dog has it's day certainly applies to these here models

My opinion of course

Link to comment
Share on other sites

I said yesterday morning in the banter thread that I had already felt like it was going down in flames for NYC. We joke about this here with our events, but they were just hugging the model that gave them the most snow (i.e. Euro) and tossing the GFS and everything else that didn't fit. As to which model has better verification scores, I don't know enough about it, but I do know that most forecasters on here would NOT feel confident in a HECS unless all major models were on board- or at the very least, the Euro and GFS (and even then, its not a lock). I think some of it is just the bias up there- NYC has had so many HECS in the last 12-15 years that I think many just lump themselves in with SNE and forget that their climo isn't much better than BWI.

 

As far as the actual advisories from NWS, I keep thinking of the Tony Pann quote before 2/5-6/10- "After a foot, who counts?" I think it would have been better to just say "potential for at least 12 inches or more of snow" than actually say 2-3 feet, etc. Thats just setting yourself up for a March 2001 bust. Though I'm sure some weenies would have called them out for "downplaying" it or whatever. But it was a complicated setup so I do feel for everyone in the NWS offices trying to forecast this system.

Link to comment
Share on other sites

I do think a large part of the issue is in how the various mets (gov't and private) communicate the uncertainty with that type of forecast.  There def should have been more uncertainty woven into the predictions for Philly to NYC, especially given that they were on western edge of the predicted big snows. I know the public hates uncertainty from the weather service, but, that might be the way to go in the future with this kind of set-up that is laced with potential failure.

Link to comment
Share on other sites

GFS proved that it needs to get more respect from now on. Euro will never be discounted. The issue lies with riding one model and ignoring what a whole hosts of other models were indicating. The failure was a human one, not the models.

I strongly agree with this.  Truthfully, the Euro (and a few runs of the NAM) were the only models bullish on big totals for KNYC.  I looked through the archives of this storm at the hi-res GFS, RGEM, GGEM, NAM.

 

Of the 28 runs of those 4 models over the 48 hours leading up to the storm- The following snowfall predictions for KNYC were taken from the models (assuming a 12 to 1 ratio)- I made sure each model encompassed the entire storm.

 

5 out of 28- Showed  4-8 inches

 

12 out of 28- Showed 8-12 inches 

 

5 out of 28- Showed 12-16 inches

 

4 out of 28- Showed 16-20 inches

 

2 out of 28-Showed 20+

 

 

We all understand that mesoscale banding often sets up NW of modeled, but I just don't believe that was a compelling enough story to ignore these selective model runs.

Link to comment
Share on other sites

I strongly agree with this.  Truthfully, the Euro (and a few runs of the NAM) were the only models bullish on big totals for KNYC.  I looked through the archives of this storm at the hi-res GFS, RGEM, GGEM, NAM.

 

Of the 28 runs of those 4 models over the 48 hours leading up to the storm- The following snowfall predictions for KNYC were taken from the models (assuming a 12 to 1 ratio)- I made sure each model encompassed the entire storm.

 

5 out of 28- Showed  4-8 inches

 

12 out of 28- Showed 8-12 inches 

 

5 out of 28- Showed 12-16 inches

 

4 out of 28- Showed 16-20 inches

 

2 out of 28-Showed 20+

 

 

We all understand that mesoscale banding often sets up NW of modeled, but I just don't believe that was a compelling enough story to ignore these selective model runs.

And the 8-12 verified for the NYC area, more or less.  

Link to comment
Share on other sites

I strongly agree with this. Truthfully, the Euro (and a few runs of the NAM) were the only models bullish on big totals for KNYC. I looked through the archives of this storm at the hi-res GFS, RGEM, GGEM, NAM.

Of the 28 runs of those 4 models over the 48 hours leading up to the storm- The following snowfall predictions for KNYC were taken from the models (assuming a 12 to 1 ratio)- I made sure each model encompassed the entire storm.

5 out of 28- Showed 4-8 inches

12 out of 28- Showed 8-12 inches

5 out of 28- Showed 12-16 inches

4 out of 28- Showed 16-20 inches

2 out of 28-Showed 20+

We all understand that mesoscale banding often sets up NW of modeled, but I just don't believe that was a compelling enough story to ignore these selective model runs.

Good post, I agree with this methodology. It eliminates forecasters weenie bias. Gefs might also have helped suggest this.

Link to comment
Share on other sites

Good post, I agree with this methodology. It eliminates forecasters weenie bias. Gefs might also have helped suggest this.

Agreed.  The only issue with the GEFS is the poor resolution might have made it tough to see the nuances.  The SREF might have shown this, but I have admittedly not looked at that.

Link to comment
Share on other sites

Some areas that were to receive 18-24" got an inch. This meets or exceeds our March 2001 debacle

you know...its a mixed bag feelings. It would of been Euphoria to go to bed 2-3 nights in a row thinking your getting a HECS. ALot of what makes this fun is the speculation and the model runs and the feeling you get when there is a good run for you. Then it has to play out. March 2001 was great till the first rain drop fell lol.

 

anyway...talk about mountain and valley. We didnt experience that with this storm. It was just Valley lol

Link to comment
Share on other sites

I do think a large part of the issue is in how the various mets (gov't and private) communicate the uncertainty with that type of forecast.  There def should have been more uncertainty woven into the predictions for Philly to NYC, especially given that they were on western edge of the predicted big snows. I know the public hates uncertainty from the weather service, but, that might be the way to go in the future with this kind of set-up that is laced with potential failure.

That's if the forecasters are actually uncertain. There was a really interesting interview between Uccellini and Wolf Blitzer yesterday at around 5 pm. (It was terrible how Wolf kept on trying to get him to sensationalize the event as a huge risk-to-life event). Uccellini was already hedging about NYC while remaining very confident about Boston. You could tell he had looked at the model trends for NYC and wasn't so sure it was indeed going to be a historic storm for them. 

 

However, released right around that time, here are snippets from OKX's forecast discussion:

DEPENDING ON EXACTLY WHERE THE STORM SETS UP...THERE IS THE

POTENTIAL THAT THE DRY SLOT COULD WORK INTO THE REGION FROM THE

N...CUTTING OFF THE HEAVIER PCPN WHERE THIS OCCURS. W OF THIS AREA

WILL BE WHERE THE BEST DEFORMATION ZONE WILL BE. THE MODELS DIFFER

ON THIS...WITH THE 18Z NAM PLACING THIS BAND OVER THE CITY.

THE SNOWFALL FCST REMAINS IN THE 20-30 INCH RANGE WITH LOCALLY

HIGHER AMTS. A LITTLE LESS FCST ACROSS NWRN ZONES...BUT THE LATEST

NAM SUGGESTS EVEN OVER 2 FT INTO ORANGE COUNTY. THE MODEL PRODUCES

2.7 INCH LIQUID OVER THE BRONX...SO THERE IS DEFINITE 3 FT

POTENTIAL WITH THIS SYS. WHERE EXACTLY THAT ULTIMATELY PLAYS OUT

REMAINS TO BE SEEN.

 

So, OKX had dropped the amount to 18-24" in the overnight shift, but then *raised* it again to 20-30" even after the Euro had moved the heaviest band east. What does this forecaster do? Stick with the 18Z NAM! 

Link to comment
Share on other sites

GFS wasn't perfect but no question it smacked the Euro around on this one.

 

https://twitter.com/islivingston/status/559925829342220289

http://www.erh.noaa.gov/hydromet/

 

NWS should go back to focusing on making the best forecast rather than being buddies with people on Facebook. A large chunk of the media should find a tall building in NYC and jump. 

Link to comment
Share on other sites

This was a tough storm for numerical prediction because rapid development of something that hasn't happened yet is inherently complex. Historic storm calls should be reserved for miller A's imo. Much less chaos to deal with with those. There was potential for a really big storm but there was never strong consensus for one. Even inside of 48 hours. That in itself is a big flag (great post btw nflwx). The biggest consensus was for parts of eastern mass and long island. That verified quite well. 

 

I really enjoyed tracking the potential norlun-like feature for our area. There was a lot of consensus for that. DT ripped it apart but low and behold it happened. For some reason, there seems to have been a trend on the forum  away from the standard practice of blending all guidance. Too much toss just cuz it aint the euro going on. That should be over for a while. 

 

I personally weight the euro above other global guidance inside of 72 hours but never ignore the overall trends of all models. If I see the majority of globals and short range guidance painting another picture, I'll definitely not euro hug. 

 

One thing is for certain, my respect for the new gfs has jumped quite a bit. It seems to do very well with northern stream storms. Like the old one but better. Good stuff. 

Link to comment
Share on other sites

GFS wasn't perfect but no question it smacked the Euro around on this one.

 

https://twitter.com/islivingston/status/559925829342220289

http://www.erh.noaa.gov/hydromet/

 

NWS should go back to focusing on making the best forecast rather than being buddies with people on Facebook. A large chunk of the media should find a tall building in NYC and jump.

Is your left side your best profile pic side?

Link to comment
Share on other sites

this isn't a scientific post, but looking at how things developed before the coastal took over, everything seemed a bit chaotic.  light snow to the north, energetic clipper diving from the midwest, rain to start here...maybe there was too much real estate to work with (lack of a real strong high to the north) before the coastal took over.  i'm also not sure if the trailing clipper may have helped nudge this a little further along as well.

Link to comment
Share on other sites

The problem isn't the Euro. The problem is the media/NWS who didn't do enough to convey the uncertainty about being on the edge.  There is always going to be a sharp gradient in these type of storms...always.  It just so happens that one of the biggest cities in the world was on the periphery.  Nobody would care about getting the border wrong if it happened in some suburbs.  The GFS was not exactly a world beater.  It had like 0.75"- 1.25" QPF for some areas that are going to get 20-30".  Euro definitely beat it in New England.  Yesterday's 12z Euro had the 0.75" contour like 40 miles west of Manhattan, and the 1.5" contour bisecting it.  Even if that was the only piece of guidance available, and you didn't have other guidance screaming caution, there should have been reasonable concern that a sharp cutoff might set up right over or even east of the city.  Plus the fact that they were getting most of this epic snow from a coastal backing in and not an advance thump.  I don't think this storm told us a whole lot other than both major models still have biases.  Euro overamped, and GFS was probably too progressive and too dry in the jackpot areas, even though it did well on the western flank.

 

While I think this is a good thread, we have so many awful model busts down here, I don't totally get why we should give 2 sh-its about NYC and why it should be some springboard for model discussion.  This **** happens all the time down here.  Welcome to 40S, b-itches.

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

  • Recently Browsing   0 members

    • No registered users viewing this page.

×
×
  • Create New...