February 2013 Blizzard NESIS Speculation

[andyhb]

Well apparently almost all of the 30" totals in CT were bogus...

 

[sankaty]

Well apparently almost all of the 30" totals in CT were bogus...

 

This map is bizarre. Not only does it ignore virtually all the 30 plus reports from central CT, it excludes the 30 plus report from the OKX forecast office. If that's not reputable, what is?

[nzucker]

Well, I busted high for sure.  Wonder if the NYC metro was within or outside of the 10" contour - would make a significant difference.

 

Central Park had 11.3".

 

I measured 17" 25 miles to their north. 

 

This is a very fair rating...of course this sub-forum wants a higher rating but this storm did not have the extent of the greats...Feb '03, Jan '96, Feb '83, Feb '78 etc all hit much larger areas of the Northeast. 

[nzucker]

Why is January 1964 so high on NESIS?

[gymengineer]

This map is bizarre. Not only does it ignore virtually all the 30 plus reports from central CT, it excludes the 30 plus report from the OKX forecast office. If that's not reputable, what is?

The central CT spotter reports had no chance of making it into the map based on the criteria. I don't know about an actual forecast office's total-- I guess if that location doesn't count as a COOP or first-order station, it still gets excluded.

Remember, the April Fool's storm map did not include ORH's >30" total.

[SR Airglow]

Well apparently almost all of the 30" totals in CT were bogus...

 

Got to say, if that map is what's going to go down as the official map, that's a pretty horrible representation. Most of the 20-30 area in CT is 30"+, and if Upton cannot be trusted, what can be? It's absurd to exclude the NWS offices, they are the experts and yet they still cannot be trusted according to NESIS? Bizzare...

-skisheep

[dan11295]

Guess the OKX office gets excluded due to a technicality its not a "first order station" or "co-op". Because of their strict criteria, the maps tend to exclude the pockets of heaviest snowfall. Some of the porton of CT with the highest totals is devoid of qualifying stations. But you have to figure that effects every storm on the list, not just this one. So the numbes may be biases low but they are biases low for all storms. so the ranking is probably close to correct.  

 

Was looking back at the OKX/BOX totals for the event. In the OKX data. there isnt a Single report from New Haven County, where the heaviest totals were observed, that meets their standards (Skywarn doesnt count). Doesnt matter how much snow fell there it would not have counted for that map. so even though 18/23 reports indicate 30"+ nothing gets counted.

[ORH_wxman]

The central CT spotter reports had no chance of making it into the map based on the criteria. I don't know about an actual forecast office's total-- I guess if that location doesn't count as a COOP or first-order station, it still gets excluded.

Remember, the April Fool's storm map did not include ORH's >30" total.

 

 

NCDC doesn't use the ORH total from 1997 because the airport wasn't having F6 data recorded for some reason. Its actually a pretty poor oversight on NCDC's part...almost like a red tape technicality. I understand not using potentially weenie spotter reports that aren't coops or whatever, but to ignore a total that NWS BOX recognizes as the #1 snowfall in that city's records is pretty silly.

 

 

That said, this rating is exactly what I thought it would be. Cat 3 sounds right.

[CoastalWx]

NCDC is horrible...what an awful map for CT.

[sankaty]

I don't have any objection to the NESIS score, and naturally I don't expect them to accept every random spotter report, but it's just too bad the map is so nonrepresentative of the historic snowfall axis from LI through CT to ME.  I think this was the likely the most prolific snowfall in CT since 1888, but you'd never know it from this map.  If it's a consistent bias for these maps, that's good to know, but it makes them so much less useful, which is a shame as they are so well done otherwise.

 

On a related note, it seems there is remarkably little data collection in central CT.  I would have expected kmmk to have an official observer.  Also, according to the BOX PNS, Staffordville in Tolland County is a COOP location reporting 31.4 inches, which I don't see reflected on the map.

[donsutherland1]

NCDC is horrible...what an awful map for CT.

 

I agree about the map. It is not representative of the storm, as it ignores New Haven County, which saw the storm's heaviest snows.

 

If the methodology doesn't produce maps that are representative of a storm's impact, the methodology should be changed. Had the map's methodology been used, some of the most prolific amounts recorded in the Blizzard of 1888, for example, would not have been registered.

 

IMO, the maps put out by NWS Taunton and Raleigh following the storm provided a more reasonable depiction of the storm's accumulations. Having visited parts of Connecticut following the storm, it is my strong opinion that the 30" accumulations were real, even if the map suggests otherwise.

 

A better approach would entail using a robust statistical framework to all snowfall reports, not simply including only those from first order stations or co-ops unless NCDC were to dramatically expand the number of such stations to deal with the representativeness issue (unlikely given funding constraints). Outliers would be flagged and excluded. Even as such a report would lead to a degree of smoothing, it would at least produce an outcome that is more representative of what happened than the above NCDC map.

[Ginx snewx]

Map does not rep text either, the line of 49 k seeing 30 is off by a magnitude of 10x

A massive blizzard hit the Northeast February 9–11, 2013 bringing more than 2.5 feet of snow to some areas, which left thousands of people without power and stranded countless airline passengers. This blizzard rated 3 out of 5 on the Regional Snowfall Index (RSI) scale according to preliminary observations through 7:00 AM Sunday. (We will continue to update these statistics as more station data become available.) While the area impacted by this storm was less than many other major storms, the heaviest snowfall landed in more densely populated areas, making it a “major “storm in the RSI categories. Over 49,000 people across 192 square miles saw 30 inches of snow or more as a result of this storm.

[jm1220]

This map is bizarre. Not only does it ignore virtually all the 30 plus reports from central CT, it excludes the 30 plus report from the OKX forecast office. If that's not reputable, what is?

I'd say a low-end 3 is fair, actually a slightly higher rank that I would have thought given the limited area of significant impact. But I agree, WTF is up with the lack of 30" totals being counted? Much of Suffolk County on LI came near it and some spots exceeded it. And since Upton measured over 30", you know they were legit.

[andyhb]

I agree about the map. It is not representative of the storm, as it ignores New Haven County, which saw the storm's heaviest snows.

 

If the methodology doesn't produce maps that are representative of a storm's impact, the methodology should be changed. Had the map's methodology been used, some of the most prolific amounts recorded in the Blizzard of 1888, for example, would not have been registered.

 

IMO, the maps put out by NWS Taunton and Raleigh following the storm provided a more reasonable depiction of the storm's accumulations. Having visited parts of Connecticut following the storm, it is my strong opinion that the 30" accumulations were real, even if the map suggests otherwise.

 

A better approach would entail using a robust statistical framework to all snowfall reports, not simply including only those from first order stations or co-ops unless NCDC were to dramatically expand the number of such stations to deal with the representativeness issue (unlikely given funding constraints). Outliers would be flagged and excluded. Even as such a report would lead to a degree of smoothing, it would at least produce an outcome that is more representative of what happened than the above NCDC map.

 

This post, right here, agreed times a million.

[skierinvermont]

low end 3 / high 2.  Map's not great but not terrible either. Probably should be a splotch of 30" in CT but not everybody in the first map posted that was in the 32+ shading actually got 32"+. Somewhere between the two maps should be representative.

 

And of course, most of the ncdc maps have the same issue of not showing localized higher totals. So at least the methodology is consistent. In terms of the NESIS scale though, half of a county or two getting 30" doesn't make much difference (unless it's Manhattan). 

[tamarack]

Maybe it's just a function of a small-scale map, but it seems odd that the larger of Maine's two tiny spots lies west of PWM. (The smaller one is even farther west.) Doesn't PWM's 31.9" count? And if so, when did the city's population (66,000+ in the 2010 census) drop low enough to allow a total 30"+ pop of 49K?

And I'm still puzzled by the conflating of two separate events in Jan 1978 to create all that 10-20" area in N.Maine.

[Quincy]

You can say the 40" totals or even the 36"+ totals are suspect, but there was most certainly a 30"+ zone through much of south-central CT. I've driven all around. You cannot tell me there was less than 30" of snow.

[ORH_wxman]

You can say the 40" totals or even the 36"+ totals are suspect, but there was most certainly a 30"+ zone through much of south-central CT. I've driven all around. You cannot tell me there was less than 30" of snow.

 

 

The lack of official obs down there almost certainly hurts. Though it seems pretty silly to put a little bullseye of 30" at BDR and then no 30" amounts to the north-northeast of that in the state. Fairly rudimentary analysis (just look at the darn radar that night) could have determined that 30"+ amounts were definitely present to the northeast of BDR...a task you would have hoped NCDC could handle, but clearly do not. I definitely understand ommitting suspect totals from spotters and unreliable coops, but in this case, it seems blanket regional omissions were made without any real logical grounds for doing so.

[snowman21]

I agree about the map. It is not representative of the storm, as it ignores New Haven County, which saw the storm's heaviest snows.

 

If the methodology doesn't produce maps that are representative of a storm's impact, the methodology should be changed. Had the map's methodology been used, some of the most prolific amounts recorded in the Blizzard of 1888, for example, would not have been registered.

 

IMO, the maps put out by NWS Taunton and Raleigh following the storm provided a more reasonable depiction of the storm's accumulations. Having visited parts of Connecticut following the storm, it is my strong opinion that the 30" accumulations were real, even if the map suggests otherwise.

 

A better approach would entail using a robust statistical framework to all snowfall reports, not simply including only those from first order stations or co-ops unless NCDC were to dramatically expand the number of such stations to deal with the representativeness issue (unlikely given funding constraints). Outliers would be flagged and excluded. Even as such a report would lead to a degree of smoothing, it would at least produce an outcome that is more representative of what happened than the above NCDC map.

I would have applied some kind of grid point averaging. There is not a real good reason to throw out public measurements especially when they provide a higher density of observations. Take all the PNS reports, plug in the latitude and longitude for all the locations, then one by one create an average based on the surrounding reports within 10 or 15 miles to smooth out and crudely QC the numbers a bit. Even doing that I think would still yield a sizable 30+ area through central LI & CT.

[andyhb]

low end 3 / high 2.  Map's not great but not terrible either. Probably should be a splotch of 30" in CT but not everybody in the first map posted that was in the 32+ shading actually got 32"+. Somewhere between the two maps should be representative.

 

And of course, most of the ncdc maps have the same issue of not showing localized higher totals. So at least the methodology is consistent. In terms of the NESIS scale though, half of a county or two getting 30" doesn't make much difference (unless it's Manhattan). 

 

New Haven County, CT is not just some dot on the countryside, it is quite a heavily populated area...

[JBG]

This happened in the 1960's, 1970s, now the 2000's.   The only thing common to those eras (other than PDO if so...) is the solar cycling had impressive negative phase. 

 

perhaps that's more than mere coincidence - I don't know. 

PDO and negative solar phase are likely big factors and not mere coincidence.