Dissecting The Bust

[rygar]

All I know is that something needs to be done with the inconsistency of these models.... However for this storm in particular only yet NAM and EURO were ever thrilled about it... All of the other models more or less continued to show a moderate snowstorm for the area, but the majority of us just kept discounting them because the EURO kept showing a bomb for 3 straight days. It never changed only until the 0z run last night. In fact it was even a bomb at 12z when the clipper was coming through the area. It obviously was a major bust for the NYC area but certainly not for Eastern LI and Boston.

Something needs to be done with the inconsistency of humans.  Personally, I find it absurd to think the models are the problem and not us.  The atmosphere is incredibly complex and to want the models to accurately portray that - well, maybe in 50 years.  We all need to stop blaming the models and simply give a confidence level to each forecast.

[ag3]

NYC never really got into the heavy snow area that eastern LI and New England got into, unlike what was predicted. So NYC did get shafted. That isn't the same as saying NYC didn't get a sizable snow event. It just wasn't historic/crippling. 9-13" in NYC is a pain but manageable.

Yes we did. LGA reported 2" in 1 hour at around midnight.

It was a whiteout with very heavy snow for 2+ hours.

Manhattan got into it as well but the band dissipated quickly over them.

[codfishsnowman]

a good chunk of what is considered sne got shafted as well, to be sure some central and all eastern areas got clobbered but hfd spfd to the berks were ALL forecasted ridiculous amounts that went against climo and the pattern. the goods for me were even closer than for many in uptons viewing area (20 to 30 miles )!! even the normally conservative local mets around here went crazy high (15 to 20 plus to 24 plus)....end result in my area 5 to 10. worst bust i can remember since feb 87!!......this one stings moreso since if for us the capture happened an hour earlier we wouldve been smoked up here!! there are only so many chances in life for a storm with this kind of potential. when i didnt see precip exploding up through jersey and epa i knew there was going to be some trouble.

[TomWH]

This is a fools party, hindsight's 20-20. That being said still fun to talk about it. A few things came to mind. First it's science they should have scaled back those totals. At no time, ever should the words historical and crippling be used together in a professional forecast. If you forecast 12-18" people get the idea. Raise the totals as needed to at most 18-24". Most people won't be able to tell the  deference between 20"or 30" of wind blow snow. Sounds good until the poor bastard that forecasts 10-12" and ends up with 3' of snow with people trapped in cars on the LIE and I95. People would die with a blown forecast like that! 

So now, "Really?" These forecasters screwed up yes but it could have been worse, A lot worse. I'm pretty sure the first concern was for peoples lives, then their reputations.  I say no problem I forgive you .

The scientific side of this offers a lot to learn too.

[bluewave]

This is just one of those times that the GFS beat the Euro. The Monday 1-26 GFS 12z total

precip forecast was right on the money.

 

NYC verification.......0.84

ISP verification........1.63

 

1-26 GFS 12z total precip forecast

 

[PB GFI]

 

This is just one of those times that the GFS beat the Euro. The Monday 1-26 GFS 12z total

precip forecast was right on the money.

 

NYC verification.......0.84

ISP verification........1.63

 

1-26 GFS 12z total precip forecast

 

f54.jpg

 

I will reply here . This was so close Uniondale 20 miles east of Brooklyn started the 20 inch line . I went to Laurel Hollow last nite 30 miles east of NYC and up on the shore there was close to 24 .

 

This was that close to being a top 5 storm for KNYC . 20 miles that's it .

[BxEngine]

The real "bust" was probably the NWS not putting enough uncertainty in their wording (something Mr. Uccellini acknowledged himself), much more than any model bust. Yeah the Euro was only off by 20-30 miles from giving nyc a much larger storm, but it was off by over a foot just west of that. Everyone knew the cutoff would be dramatic, and 50 miles a few days out is a pretty damn good job.

[Guest Pamela]

I will reply here . This was so close Uniondale 20 miles east of Brooklyn started the 20 inch line . I went to Laurel Hollow last nite 30 miles east of NYC and up on the shore there was close to 24 .

 

This was that close to being a top 5 storm for KNYC . 20 miles that's it .

 

 

Third time in just over 5 years they miss a two foot storm by 30 miles or so to the east (12/19/2009 & 2/9/2013). 

[bluewave]

I will reply here . This was so close Uniondale 20 miles east of Brooklyn started the 20 inch line . I went to Laurel Hollow last nite 30 miles east of NYC and up on the shore there was close to 24 .

 

This was that close to being a top 5 storm for KNYC . 20 miles that's it .

 

The only reason this made news is we are in the heaviest populated part of the country.

Had we seen a 40 mile forecast miss  in a lightly populated  area with the predicted heaviest 

snow bands  no one would even be talking about this. And consider how many times this

size of an error in rainfall forecasts happens around our region during the year.

 

CENTRAL PARK 9.8 100 PM 1/27 CENTRAL PARK ZOO

DEER PARK 20.0 1230 PM 1/27 PUBLIC

[wolfsheepsheadbay]

The real "bust" was probably the NWS not putting enough uncertainty in their wording (something Mr. Uccellini acknowledged himself), much more than any model bust. Yeah the Euro was only off by 20-30 miles from giving nyc a much larger storm, but it was off by over a foot just west of that. Everyone knew the cutoff would be dramatic, and 50 miles a few days out is a pretty damn good job.

Bingo. They can 'review the modeling' as much as they want, but without the funds, nothing will change any time soon in that regard.

 

With more and more of the mainstream now 'model watching', on social media especially, I think the NWS will begin being more specific and cautious in their wording. They may still select the snowier solution for their call, but they will share the other options and maybe assign a percentage of liklihood to each.

 

Bottom line is, they aren't prophets, but they can certainly cover their a$$es better in the future by sharing more information.

[TheManWithNoFace]

This was not only " over " forecasted by UPTON and MT HOLLY .

If someone can find me 1 PRIVATE forecaster that did not have 20 to 30 out I would be shocked .

I am taking about guys that make their money in the private sector and who are paid by clients based on performance .They all bought the bigger totals , at least the ones I saw .

NJTV never went bigger than 18-24 for NYC and NENJ.

[Guest Pamela]

The only reason this made news is we are in the heaviest populated part of the country.

Had we seen a 40 mile forecast miss  in a lightly populated  area with the predicted heaviest 

snow bands  no one would even be talking about this.

 

When millions of people and millions of dollars are affected, what else could you expect...it can't be swept under the rug...

[donsutherland1]

The only reason this made news is we are in the heaviest populated part of the country.

Had we seen a 40 mile forecast miss  in a lightly populated  area with the predicted heaviest 

snow bands  no one would even be talking about this. And consider how many times this

size of an error in rainfall forecasts happens around our region during the year.

 

CENTRAL PARK 9.8 100 PM 1/27 CENTRAL PARK ZOO

DEER PARK 20.0 1230 PM 1/27 PUBLIC

I agree with a lot of this. Moreover, I believe this outcome could, if utilized effectively, lead to greater emphasis on probabilistic forecasting relative to deterministic numbers, especially by those who use such forecasts.

 

IMO, the high-profile forecasting failure should not be permitted to detract for the NWS's large and outstanding body of work. Certainly, my confidence in the NWS and appreciation of all the NWS does remains undiminished. Political leaders who made extreme decisions e.g., the unprecedented decision to shut down mass transit in NYC, should not pass all accountability onto the NWS. Moreover, IMO, a snowstorm or blizzard is not reason to shut down the entire mass transit system. The costs of doing so vastly exceed the benefits. 

[coastalplainsnowman]

Third time in just over 5 years they miss a two foot storm by 30 miles or so to the east (12/19/2009 & 2/9/2013). 

 

Enthusiasts question here - looking at the snowfall map for 2/9/2013.  To me, that's just the way these storms go.  You would never see those huge amounts that occurred in the central LI northshore and up into central connecticut because you just simply don't get those amounts on the south shore of LI (other than some one-off lollipops like the 30' somewhere on suffolk south shore in 96.)  If it ever happened that south shore LI got widespread 24-30, then all other things being equal, wouldn't the northshore amounts and the central connecticut amounts be 50"(!)

 

That was one of the things I wondered about this past storm.  IF LI was 24-36, wouldn't connecticut/RI have been forecast for 36-48?

 

Lastly, I agree with the points about the population here and also the fact that, again, as a layman here, if this was the middle of kansas, the difference in 30 miles might be more due to random things we don't yet grasp rather than than all the other obvious factors that come into play here with the proximity of the atlantic ocean, the long island sound, the variations in elevation, etc.

 

I also agree that forecasters were in a tough spot.  I know it boils down to better safe than sorry, but really -- what situation would they rather have -- that they overforecast and people were prepared, or they underforecast and people are stuck in snowbanks.  Imagine the political response then.  Who can rule out some idiot politician trying to criminalize the mistake?

[MJO812]

Snowfall totals from this past storm. Courtesy of SNE.

[RU848789]

The only reason this made news is we are in the heaviest populated part of the country.

Had we seen a 40 mile forecast miss  in a lightly populated  area with the predicted heaviest 

snow bands  no one would even be talking about this. And consider how many times this

size of an error in rainfall forecasts happens around our region during the year.

 

CENTRAL PARK 9.8 100 PM 1/27 CENTRAL PARK ZOO

DEER PARK 20.0 1230 PM 1/27 PUBLIC

Bluewave - great point on rainfall. Let me elaborate, as I don't know that much about meteorology, but I know a lot about energy, thermodynamics and mixing (chem engineer).  I've posted about this several times before, but the atmosphere should follow the "Arrhenius rule" of a doubling of energy for every 10C temperature rise (although I've never seen this published; Arrhenius's "rule" is technically about doubling the reaction rate of simple chemical reactions with every 10C temp rise).  

 

If it does, that would lead to roughly an 8X increase in atmospheric energy for a 30C day vs. a 0C day.  The only thing I've found that talks about a related topic is that the capacity of the atmosphere to hold water increases by 8% per 1C temperature rise: this would mean there's a 240% or 2.4X increase in atmospheric water capacity at 30C vs. 0C (using those temps as my surrogates for summer-thunderstorms and winter-snow).  I know there are other variables that lead to max precip at 30C vs. 0C, but I imagine it's somewhere between the 2.4X and 8X above: often record rainfalls from summer convective storms are in the 10" range vs. record snowfalls being in the 2" range (24"+ of snow) and that's a factor of 5X.  

 

The reason why I've always thought the Arrhenius 8X factor would apply is this.  We've all seen how we can have an isolated thunderstorm convectively putting down an inch of rain in 15 minutes, while the most amazing convective snowfall rates are around 0.5" LE/hour or 0.12"/15 min, which is a factor of 8X vs. 1"/15 min.  I don't think that's coincidence.  If I didn't have a day job and had a meteorology background, I might even want to research this a bit.  

 

So, that's all interesting as background, but my take home point was going to be elaborating further on your comment.  We get these isolated 1" t-storms in one town and nada one town over on occasion.  Can you imagine the uproar if that kind of difference (12" snow) were observed over maybe a 2 mile gradient instead of the 20 mile gradient we just witnessed, roughly?  But nobody cares in the summer, since there's no lasting impact beyond the 15-minutes, unless you're in the middle of an outdoor activity - people just accept that there's a chance of t-storms and you better hope to get lucky.  

 

The other reason this temperature vs. max precip potential relationship is useful is as a teaching tool - when I point this out to people, as a way of explaining why snowfall forecasts are actually usually way better than rainfall forecasts, due to the energy of the systems involved, and that predicting these convective bands in a snowstorm is impossible, just like it is for t-storms, they seem to get it.  I think the meteorological community needs to hammer away at the incredible uncertainty in precip amounts when any convection is involved, in particular, and, obviously, as a function of track/placement, too.  This is not easy to do, as much of the public is either incapable (don't think so) or unwilling (mostly) with regard to understanding this.  But I know that simply showing a map with only one set of accumulations is as far away from incorporating uncertainty as one can get.  

[LongBeachSurfFreak]

I was probably the most directly effected person on this board. I can say that this storm did happen just the way it was supposed to. Just 50 miles east. So on some level the nws was right. Despite being extremely pissed at the time I can't say I blame them for what they did. What billx said the day before was right on the money these storms tend to send the best band NW of what's modeled. You just have to look back to Boxing Day to see that happen.

At home in wantagh it was a legit blizzard with 16" and good drifting. We would have had 20 but we missed the band during the day. Pretty much right on par with feb 13 so not bad. Top 20 in my life

Good news is it's still January. I think we have a shot at building record snow pack. (Especially out east as they are almost there)

[bluewave]

When millions of people and millions of dollars are affected, what else could you expect...it can't be swept under the rug...

 

 

I agree with a lot of this. Moreover, I believe this outcome could, if utilized effectively, lead to greater emphasis on probabilistic forecasting relative to deterministic numbers, especially by those who use such forecasts.

 

IMO, the high-profile forecasting failure should not be permitted to detract for the NWS's large and outstanding body of work. Certainly, my confidence in the NWS and appreciation of all the NWS does remains undiminished. Political leaders who made extreme decisions e.g., the unprecedented decision to shut down mass transit in NYC, should not pass all accountability onto the NWS. Moreover, IMO, a snowstorm or blizzard is not reason to shut down the entire mass transit system. The costs of doing so vastly exceed the benefits. 

 

 

Bluewave - great point on rainfall. Let me elaborate, as I don't know that much about meteorology, but I know a lot about energy, thermodynamics and mixing (chem engineer).  I've posted about this several times before, but the atmosphere should follow the "Arrhenius rule" of a doubling of energy for every 10C temperature rise (although I've never seen this published; Arrhenius's "rule" is technically about doubling the reaction rate of simple chemical reactions with every 10C temp rise).  

 

If it does, that would lead to roughly an 8X increase in atmospheric energy for a 30C day vs. a 0C day.  The only thing I've found that talks about a related topic is that the capacity of the atmosphere to hold water increases by 8% per 1C temperature rise: this would mean there's a 240% or 2.4X increase in atmospheric water capacity at 30C vs. 0C (using those temps as my surrogates for summer-thunderstorms and winter-snow).  I know there are other variables that lead to max precip at 30C vs. 0C, but I imagine it's somewhere between the 2.4X and 8X above: often record rainfalls from summer convective storms are in the 10" range vs. record snowfalls being in the 2" range (24"+ of snow) and that's a factor of 5X.  

 

The reason why I've always thought the Arrhenius 8X factor would apply is this.  We've all seen how we can have an isolated thunderstorm convectively putting down an inch of rain in 15 minutes, while the most amazing convective snowfall rates are around 0.5" LE/hour or 0.12"/15 min, which is a factor of 8X vs. 1"/15 min.  I don't think that's coincidence.  If I didn't have a day job and had a meteorology background, I might even want to research this a bit.  

 

So, that's all interesting as background, but my take home point was going to be elaborating further on your comment.  We get these isolated 1" t-storms in one town and nada one town over on occasion.  Can you imagine the uproar if that kind of difference (12" snow) were observed over maybe a 2 mile gradient instead of the 20 mile gradient we just witnessed, roughly?  But nobody cares in the summer, since there's no lasting impact beyond the 15-minutes, unless you're in the middle of an outdoor activity - people just accept that there's a chance of t-storms and you better hope to get lucky.  

 

The other reason this temperature vs. max precip potential relationship is useful is as a teaching tool - when I point this out to people, as a way of explaining why snowfall forecasts are actually usually way better than rainfall forecasts, due to the energy of the systems involved, and that predicting these convective bands in a snowstorm is impossible, just like it is for t-storms, they seem to get it.  I think the meteorological community needs to hammer away at the incredible uncertainty in precip amounts when any convection is involved, in particular, and, obviously, as a function of track/placement, too.  This is not easy to do, as much of the public is either incapable (don't think so) or unwilling (mostly) with regard to understanding this.  But I know that simply showing a map with only one set of accumulations is as far away from incorporating uncertainty as one can get.  

 

This is the challenge of forecasting winter storms in the most densely populated part of the country. My guess

is that if we saw a greater percentage of our annual precipitation as snow, then there would potentially be

more 20 to 40 mile errors in snowfall forecasts. NYC generally sees on average only about 2.00" to 3.00"

of the nearly 50.00" annual precipitation as snowfall. So many of the rain event totals  that are off by 20 or 40 miles

really never get noticed since the impact isn't as great or as noticeable as snow. 

[Dark Star]

First, I think we all need to agree this was a bust, at least regarding my area in central Union county, NJ.  I disagree with educating the general public about percentages.  With a degree in meteorology (no longer practicing), I believe you should be able to put your cards on the table and issue a forecast 24 hours before a storm begins.  I used to enjoy Jeff Berradelli's forecast discussions on his website.  He said the devil is in the details. 

 

I was incorrect when I saw some nice echoes off of Hatteras Monday morning.  This was not the storm, nor the area it was supposed to form.  This area quickly died out as the one offshore of Maryland did in the afternoon.  I failed to remember that storms which form this close to the NY metro area need time to strengthen, which rarely happens.  This time was no different.

In the afternoon, WCBS radio said all the ingredients were on the table and the least snow scenario was if the storm jogged 50 miles further east, which would result in 12-18 inches.  (Later Mr. Quinn would say that the storm was 50 miles too far east).

 

Unfortunately, the NWS seems to always be the slowest in getting their changes to a forecast out to the public.  Of all the forecasters, Dave Curren of NJ12 began to question the models around 8 PM Monday night.  Watching the radar, I agreed that we would see less than 12 inches.  Garwood NJ received less than 4 inches, and was mostly from the clipper on Monday.  As usual, there were some members on this site scoffing at the bloggers who were jumping off the boat . 

 

Too bad Alan Kasper retired.

[ForestHillWx]

Reading through the posts, and not to overly simplify the miss to the east, but was the late capture involving the ULL the primary driver in allowing the storm to get going to far away from the coast? If so, what are some of the factors that prevented the earlier capture? The lack of any substantial downstream blocking?

 

Side note, I agree with the poster on the travel ban for NYC, issuing $300 fines and misdemeanors. Seems a bit excessive.

[SnoSki14]

I was skeptical of the forecasts when I saw the huge precip shield far to the east and it struggled moving westward. The fact that all their chips were on the Nam was worrying considering how critical everyone is of that model.

[Guest Pamela]

Enthusiasts question here - looking at the snowfall map for 2/9/2013.  To me, that's just the way these storms go.  You would never see those huge amounts that occurred in the central LI northshore and up into central connecticut because you just simply don't get those amounts on the south shore of LI (other than some one-off lollipops like the 30' somewhere on suffolk south shore in 96.)  If it ever happened that south shore LI got widespread 24-30, then all other things being equal, wouldn't the northshore amounts and the central connecticut amounts be 50"(!)

 

That was one of the things I wondered about this past storm.  IF LI was 24-36, wouldn't connecticut/RI have been forecast for 36-48?

 

 

 

Every storm is unique...but yes, that's one of the reasons snow records are kept; to help recognize patterns of snowfall with various systems.  This is especially useful near the Great Lakes...where areas north of Buffalo average 40 or 50 inches of snow per year...and a few miles to the south the average is at least twice that.  When a lake effect snow event is headed towards Buffalo...that climatological record provides significant guidance to the western NY forecasters as to how much snow to predict in their assigned CWA.  This is why climatology is such a useful science and is far more than just going over a bunch of numbers just for the fun of it.  Same with the situation you cite...the records over the last 50 years (and especially the voluminous record gathered since the internet age arrived) has provided forecasters with an extraordinary repository of data to aid them in providing the public with a more precise forecast regarding snow, rain and all the other aspects of the weather. 

[Clinch Leatherwood]

The real "bust" was probably the NWS not putting enough uncertainty in their wording (something Mr. Uccellini acknowledged himself), much more than any model bust. Yeah the Euro was only off by 20-30 miles from giving nyc a much larger storm, but it was off by over a foot just west of that. Everyone knew the cutoff would be dramatic, and 50 miles a few days out is a pretty damn good job.

 

 

True on being so definitive but let's not minimize the multi-run error of the Euro.  The 1/24 12z forecast had what 3 feet through all of NJ and almost into parts of PA and DE?  It wasn't just wrong once, it was wrong every single run the entire time.

 

 

Forecasters are over explaining what happened.   The #1 model for more than a decade, the model that in many instances differed from other models during significant events and proved to be right in the last 10 years,  was very wrong.  This is the first time it's been this wrong in the modern era.   Going forward we now know we can no longer rely on it exclusively.  The end. 

 

Because ultimately that's what happened here.  I don't really even want to hear about the NAM being similar.  We are lucky the NAM is able to forecast 6 hours out, every forecaster, amateur and hobbyist has known that for the last 3 years.   That's an excuse and almost makes the reasoning worse.

 

If Warren Buffet tells you where to put your money you do it.   If the Euro tells you it's going to be a blizzard over the last 10 years at that range a blizzard was coming.

 

We can argue within the science of why in the last year that isn't the case but that's irrelevant to the general public.

[Guest Pamela]

 

 Imagine the political response then.  Who can rule out some idiot politician trying to criminalize the mistake?

 

As a meteorologist, one is not likely to overtly enter into a contractual agreement with an individual or small group where a failure to satisfy the requirements of the contract (i.e. a proper weather forecast) could potentially lead to civil or criminal liability...you might hire someone to predict the weather for you...but you would probably be unable to collect damages in civil court if they failed to provide you with a "satisfactory" forecast. You might be able to sue them if you hired them to forecast for you, paid them, and they never provided you with any forecast...but failing to provide the "right" forecast does not sound like grounds for an actionable lawsuit simply based on the notion that a "correct" weather forecast is a concept utterly rife with ambiguity. 

[TheManWithNoFace]

Again, I think the NJTV weather people put together a pretty thoughtful post storm article. It touches on some of the things discussed here- area of impact, psychology of forecasting, There isn't a lot of science in the article, so I don't really get the title. "The Science Behind a Monumental Bust" etc. http://www.njtvonline.org/news/uncategorized/the-science-behind-a-monumental-bust-of-a-forecast/

[chietanen]

Short discussion from MSNBC with a former head of the AMS discussing the storm. He brings up the possibility of using something  for Noreasters similar to the cone of uncertainty that they use for tropical systems. Also mentions using a probabilistic forecast for snowfall totals (which I believe they already do in snowfall maps, but that information is not widely disseminated to the public).

 

Start at 3 minutes if you want to avoid the Joe Scarborough chatter

 

http://www.msnbc.com/morning-joe/watch/what-weather-models-got-right-in-blizzard-390087235837

[chietanen]

True on being so definitive but let's not minimize the multi-run error of the Euro.  The 1/24 12z forecast had what 3 feet through all of NJ and almost into parts of PA and DE?  It wasn't just wrong once, it was wrong every single run the entire time.

 

 

Forecasters are over explaining what happened.   The #1 model for more than a decade, the model that in many instances differed from other models during significant events and proved to be right in the last 10 years,  was very wrong.  This is the first time it's been this wrong in the modern era.   Going forward we now know we can no longer rely on it exclusively.  The end. 

 

Because ultimately that's what happened here.  I don't really even want to hear about the NAM being similar.  We are lucky the NAM is able to forecast 6 hours out, every forecaster, amateur and hobbyist has known that for the last 3 years.   That's an excuse and almost makes the reasoning worse.

 

If Warren Buffet tells you where to put your money you do it.   If the Euro tells you it's going to be a blizzard over the last 10 years at that range a blizzard was coming.

 

We can argue within the science of why in the last year that isn't the case but that's irrelevant to the general public.

 

Have to give you props. You were on top of this for a couple of days almost expecting that the Euro would come east, and when it didn't, explaining why you thought it off just a touch. I guess being burned by a couple of previous Cape blizzards that didn't materialize was a good learning experience for you. Congrats on a sweet storm for you guys.

[donsutherland1]

Short discussion from MSNBC with a former head of the AMS discussing the storm. He brings up the possibility of using something  for Noreasters similar to the cone of uncertainty that they use for tropical systems. Also mentions using a probabilistic forecast for snowfall totals (which I believe they already do in snowfall maps, but that information is not widely disseminated to the public).

 

Start at 3 minutes if you want to avoid the Joe Scarborough chatter

 

http://www.msnbc.com/morning-joe/watch/what-weather-models-got-right-in-blizzard-390087235837

The probabilistic tool is still experimental. Nevertheless, I believe placing the emphasis probabilistic information when working with public officials is probably far better than sticking with largely or wholly deterministic information.

 

At the same time, I believe the Mayor of NYC, Governor of NY, etc., should develop a protocol on how to address various storm scenarios (actions + lead time + communication). The process by which probabilistic information is translated into actions should be fairly automatic and objective, and it should take into consideration the major economic and social costs related to such measures. Such an approach would lead to greater consistency, more rapid implementation, greater advance notice, etc. , as well as mitigation of costs in cases of error.

 

Moreover, shutting down the underground portion of the subway system made little sense. In cases such as hurricanes, it would make sense, as parts of the subway system could be flooded by storm surge. In cases of snowstorms, the underground system proved capable of even handling the 26.9" snowfall in 2006. The above-ground system can be impacted and that's where the focus should have been. Shutting down the entire system essentially maximized economic cost exposure in the case of an error and given an absence of empirical data, had an uncertain impact on  advancing safety.

[Juliancolton]

If only Cuomo's new weather detection network were already in place, this could have all been avoided...

[chietanen]

I've been over it, I think you need to accept that fact that this was a bust all around. Even Boston, which ended up with 24.4 inches of snow, only had a liquid equivalent of 1.05", a significant amount less than all the models were forecasting. They were "saved" by very high ratios, just as NYC was saved by the fairly intense and stationary band that formed Monday afternoon.