Spring Banter & General Discussion/Observations

[40/70 Benchmark]

4 minutes ago, codfishsnowman said:

do you have pets too?

Negative.

[CoastalWx]

5 minutes ago, Damage In Tolland said:

Can a 280lb man walk on it?

A 225 man can. Certainly a withered 120lb man such as yourself could.

[Baroclinic Zone]

14 minutes ago, CoastalWx said:

This snowpack is nuts. I mean it's completely bulletproof. It's going to take a lot of sun to melt it.  We don't have much..maybe 5" OTG after the rain, but you can drive a sherman tank over it.

Only 1-2" at home.

[CapturedNature]

It looks like Live Sap Cam made the Top 10 list at EarthCam:

http://www.earthcam.com/topten.php

It's more snow than sap now after the storm but things should pick up again next week.  I was thinking I should take bets on when it starts flowing again...lol.  Maybe next year I could do a "Joe's Pond" kind of deal...lol.

[Sugarloaf1989]

Some observations that I have never seen before in Mid-March are pancake ice on the Connecticut River at Hartford and the West River near Brattleboro being 90% frozen over after being ice free two weeks earlier. Also the Connecticut River in Parts of Massachusetts has new ice jams on it.

The temperature at Mount Snow yesterday held Around 17F all afternoon, which must be close to -25F below normal.

[tamarack]

2 hours ago, Damage In Tolland said:

Can a 280lb man walk on it?

Probably.  I'm almost up to that weight, and I easily stayed atop the 22" pack on our lawn this past Saturday. 

The temperature at Mount Snow yesterday held Around 17F all afternoon, which must be close to -25F below normal.

The afternoon high was zero while I was testing the pack's glacialness  That's 37F below my avg for the date.

[wxsniss]

 

5 hours ago, CT Rain said:

I also think the big heat the NWS is getting is partially based on their statement on the bust... which was effectively, "we thought the storm would be less but we didn't want to tell people in case they wouldn't take it seriously." 

That's ridiculous. I can't imagine being confident a forecast is too high and just say... wth let me just let it ride.

I do think that one problems humans have with forecasts is that we tend to ignore some computer guidance when it goes against group think (board and social media) and some kind of cognitive dissonance. We need to be careful about this because we're not too far off from having machines making these decisions for us - clealry the raw model blends and the raw WPC guidance was better than what the WFOs came up with. 

One thing we try to do on the broadcast side is to stay consistent and avoid a windshield wiper kind of forecast... and sometimes I think we do this to a fault. Going back and looking at some of the model data for NYC and coastal CT is that there was a clear signal for mixing issues as early as the 00z run ealry Monday AM. 

 

This and other gems of 3/14/17 post-mortem today, great stuff guys. Copied CT Rain's post from other thread.

"Machines making these decisions for us..."

As a regular lurker, the quality of analysis here and elsewhere has declined steeply in the past 2 years with the wide accessibility of clown maps. Part of why I don't post as much as I used to. Mets (here, on TV and social media) are increasingly rip-reading rather than relying on knowledge and experience. Lay public has access to them and uses them as a replacement of human expertise. And it's not just the profession of meteorology that is feeling this threat.

 

Other pearls from 3/14/17:

- low SLR and crappy snowgrowth should have been expected from soundings: drying in the column + best omega 850-700 rather than in DGZ. I think Wiz raised this repeatedly (as well as frontolysis) but there was a group "blindspot" / momentum that WAA would overproduce.

- overdone WAA qpf... there was skepticism voiced, but others countered that it can be done and this was the perfect 6-hour setup for it AND just about every piece of guidance was showing ~ 2+ qpf

- H5 / H7 lows too far west, very tilted system... we saw it on guidance, but again we also saw ~ 2+ qpf over SNE, so "machines making these decisions for us" was big here. Ultimately mid level lows tracked even farther west than guidance predicted

 

All in all a fascinating study in group think blindspots. And as CT Rain and other said, even for NWS, wanting to avoid flip-flopping on a forecast where a few uncertain but hugely impactful variables were "right on the edge".

[CT Rain]

33 minutes ago, wxsniss said:

 

All in all a fascinating study in group think blindspots. And as CT Rain and other said, even for NWS, wanting to avoid flip-flopping on a forecast where a few uncertain but hugely impactful variables were "right on the edge".

Yeah there's a lot in this whole bust to digest. One of the most frustrating things I've seen mets do is say something like, "well, even though the I-95 corridor busted it was a really bad snowstorm in middle of nowhere NY." What an incredible stupid and silly argument. People make decisions about what will happen in their backyard. If I forecast a tornado to hit one town and it hits another no one is going to say, "well, the tornado happened but it just hit a town that didn't have a warning about it." 

I think all of us can spend more time trying to challenge the "groupthink" forecasts and coming up with meteorological arguments for why something will or won't happen. If we keep seeing busts like this where machines > humans the higher ups at the NWS are going to take more and more control away from the WFOs in really short order. 

[Ginx snewx]

5 minutes ago, CT Rain said:

Yeah there's a lot in this whole bust to digest. One of the most frustrating things I've seen mets do is say something like, "well, even though the I-95 corridor busted it was a really bad snowstorm in middle of nowhere NY." What an incredible stupid and silly argument. People make decisions about what will happen in their backyard. If I forecast a tornado to hit one town and it hits another no one is going to say, "well, the tornado happened but it just hit a town that didn't have a warning about it." 

I think all of us can spend more time trying to challenge the "groupthink" forecasts and coming up with meteorological arguments for why something will or won't happen. If we keep seeing busts like this where machines > humans the higher ups at the NWS are going to take more and more control away from the WFOs in really short order. 

I wouldn't call this storm a bust just because the I 95 totals were low though. The societal impact for those 3-4 hrs was as I expected. I think the warnings were warranted even if the snow totals were not impressive.

[Juliancolton]

One of the worst things that happens is some politician invariably cracks a joke about the NWS or "weathermen" and then it becomes an irreversible us vs. them thing where the wx-oriented people blindly defend their own side. It would be nice if the governors could avoid the inflammatory one-liners so introspection can run its course...

[OceanStWx]

Well this will come as a disappointment to weenies out there, but it looks like this storm sparked a lightbulb for Maue at WeatherBell. The Euro has explicit pytpes as output (the big four, but also a "mix" category). However the Euro, as we know, sums sleet and snow the same 10:1, and we've also proven the Kuchera method runs hot in those situations too.

So Maue has stated he plans to pull the QPF by ptype and apply his own ratio (i.e. 1:1 or 2:1 for sleet) to create WeatherBell snowfall maps. 

Because as sad as it may be, I've heard forecasters reference those 10:1 snowfall maps as reasoning to go with headlines. 

[OceanStWx]

To jump in on the NWS discussion, with my non-NWS hat on (my opinion, does not represent my employer, yada yada):

A lot of the issue is with the interview that was given stating that we knew the storm was trending west and decided to hold the forecast as is. That came from WPC. I truly think it was just a poor choice of words, and I hope it was not throwing local WFOs under the bus because WPC pushed for lower amounts Monday night.

If WPC had just said the storm took a last minute shift west, therefore the heaviest snow was shifted west too, this story is no big deal. 

You always try and give warnings a chance to verify before taking them down, unless it is plainly obvious it is not going to happen. Because nobody likes a windshield wiper forecast of warning>advisory>warning (which GYX has done at least once this year). But you don't explain it like we knew the forecast wasn't going to verify and let it go out as is anyway.

[OceanStWx]

This is my gripe about watches and warnings coming out earlier than ever before.

Once we make the decision to go with headlines, snowfall ranges are out there for the world to see. And we can't walk those back. The first range is the one most people remember. It is why I still advocate for watches being general and saying 6 or more inches possible, like the good old days. 

OKX had a high profile bust in 2015. I have no idea how it played out, but I hope they weren't gun shy to drop amounts because they didn't want to look bad again.

Time after time, our snowfall forecasts are at the high end of the probability ranges on our experimental website. That should tell us something. You can't take the ARW and 4km NAM QPF verbatim. Convection is local by nature, and not a good expression of QPF over a large area. Or conversely, we think we are better at forecasting ratios than we are. I see a lot of internal discussions about going "at least 15:1" for an event. Well 15:1 is a rarity for most locations in New England. And at 70+ hours lead time, why are we assuming we'll have any confidence in a ratio forecast?

Say CON was forecast on day 3 to get 1.5" QPF, and we decided 15:1 was the way to go because of mid level magic. That's 22.5" of snow. 18-24" range. Now say day 1 rolls along and now things trend west, warmer temps move in aloft, CON never mixes but the ratio could be closer to climo 11:1. That's 16.5" of snow and a 12-18" range. The forecast looks like a massive change, but the QPF stayed the same, we just made a bold call about ratios that isn't rational that far out.

Sometimes we do it to ourselves, whether it's the internal weenie in us all, or something else.

[OceanStWx]

14 hours ago, CT Rain said:

I also think the big heat the NWS is getting is partially based on their statement on the bust... which was effectively, "we thought the storm would be less but we didn't want to tell people in case they wouldn't take it seriously." 

That's ridiculous. I can't imagine being confident a forecast is too high and just say... wth let me just let it ride.

I do think that one problems humans have with forecasts is that we tend to ignore some computer guidance when it goes against group think (board and social media) and some kind of cognitive dissonance. We need to be careful about this because we're not too far off from having machines making these decisions for us - clealry the raw model blends and the raw WPC guidance was better than what the WFOs came up with. 

One thing we try to do on the broadcast side is to stay consistent and avoid a windshield wiper kind of forecast... and sometimes I think we do this to a fault. Going back and looking at some of the model data for NYC and coastal CT is that there was a clear signal for mixing issues as early as the 00z run ealry Monday AM. 

Incredibly frustrating, as the WFOs seemingly took just enough rope to hang themselves on this one. 

WPC was right on the final call. But if you check their forecasts starting Sunday NYC went from 12-15" (00z Sun), to 15-18" (12z Sun), to 18-21" (00z Mon), to 8-10" (12z Mon), to 6-8" (00z Tue). So they trended up and then down too. Interior locations really looked worse in my opinion. A place like CON going from 8-10, to 15-18, to 15-18, to 18-21, to 21-24 over that same time period. Just up, up, and away, despite there being just as many signals of a dry slot working into the area. And I've seen them be guilty of the windshield wiper, when you get the dueling 00z/12z runs that flip flop on storm track.

That's where WPC has the advantage of not really having to answer to anyone. I know that's not totally true, but they don't really have the same pressures on them as the local WFOs do. When WPC shifts heavy snow out of NYC, they aren't the ones answering the phones. OKX does.   

[OceanStWx]

14 hours ago, CT Rain said:

Yes, our ensembles are underdispersive. That said there were EPS members that did show a far western solution - a handful on Monday (both 00z and 12z) did bring the sfc low west of BID/FOK. 

Once again Colle and Stony Brook doing some great work in that regard. They perform fuzzy clustering (I'm not even going to try and explain it) on the ensembles to separate them into groups.

The 12.00z run (Sat night) had an overall mean outside the benchmark. They broke that down into four camps essentially. Roughly mean positions of the clusters are shown below (sorry password protected otherwise I'd share images).

Right away there are two means low positions that look enticing. What the the two western lows (red/orange) have in common? They were dominated by ECMWF EPS members, and featured no NCEP members (a couple Canadians sneaked in there though). 

The navy was dominated by NCEP members, no EPS. And the purple had a mix of all three, but fewest Canadian.

The GEFS struggles to break from its own pack, but the EPS and CMC are at least showing an envelope of solutions. 

Keep in mind, this is Saturday night, showing us there was definitely at least a couple camps W of the overall mean low position (which as you might image was in the center of those 4 lows on the map). 

[CT Rain]

5 hours ago, OceanStWx said:

To jump in on the NWS discussion, with my non-NWS hat on (my opinion, does not represent my employer, yada yada):

A lot of the issue is with the interview that was given stating that we knew the storm was trending west and decided to hold the forecast as is. That came from WPC. I truly think it was just a poor choice of words, and I hope it was not throwing local WFOs under the bus because WPC pushed for lower amounts Monday night.

If WPC had just said the storm took a last minute shift west, therefore the heaviest snow was shifted west too, this story is no big deal. 

You always try and give warnings a chance to verify before taking them down, unless it is plainly obvious it is not going to happen. Because nobody likes a windshield wiper forecast of warning>advisory>warning (which GYX has done at least once this year). But you don't explain it like we knew the forecast wasn't going to verify and let it go out as is anyway.

I think the strangest thing about all of this is that the NWS didn't say in that AP article that the storm moved west at the last minute and we always strive to do our best etc. Why they would say what they said about wanting to leave a bad forecast in place is bizarre - though probably accurate to some extent (you want to give a forecast a chance to verify).

[CT Rain]

1 hour ago, OceanStWx said:

Once again Colle and Stony Brook doing some great work in that regard. They perform fuzzy clustering (I'm not even going to try and explain it) on the ensembles to separate them into groups.

The 12.00z run (Sat night) had an overall mean outside the benchmark. They broke that down into four camps essentially. Roughly mean positions of the clusters are shown below (sorry password protected otherwise I'd share images).

Right away there are two means low positions that look enticing. What the the two western lows (red/orange) have in common? They were dominated by ECMWF EPS members, and featured no NCEP members (a couple Canadians sneaked in there though). 

The navy was dominated by NCEP members, no EPS. And the purple had a mix of all three, but fewest Canadian.

The GEFS struggles to break from its own pack, but the EPS and CMC are at least showing an envelope of solutions. 

Keep in mind, this is Saturday night, showing us there was definitely at least a couple camps W of the overall mean low position (which as you might image was in the center of those 4 lows on the map). 

Interesting - how do those look for later Sunday into Monday? Clearly based on the GEFS spread we were talking about earlier giant caution flags for a tucked in low.

[OceanStWx]

Just now, CT Rain said:

I think the strangest thing about all of this is that the NWS didn't say in that AP article that the storm moved west at the last minute and we always strive to do our best etc. Why they would say what they said about wanting to leave a bad forecast in place is bizarre - though probably accurate to some extent (you want to give a forecast a chance to verify).

Totally mystifying. I mean I have a lot of respect for Carbin and the work he did at SPC, but I'm not sure why he would go and say what he said. 

[CT Rain]

1 minute ago, OceanStWx said:

Totally mystifying. I mean I have a lot of respect for Carbin and the work he did at SPC, but I'm not sure why he would go and say what he said. 

It's easy when you're giving an interview to answer something honestly and with good intentions (social science!) only to have it blow up in your face when it's the one quote that goes "viral". 

[OceanStWx]

8 minutes ago, CT Rain said:

Interesting - how do those look for later Sunday into Monday? Clearly based on the GEFS spread we were talking about earlier giant caution flags for a tucked in low.

For this upcoming thing?

There's a couple clusters east and northeast of the ensemble mean, dominated by GEFS. A southwest cluster dominated by EPS. And a north cluster dominated by EPS, with a GEFS and CMC member each thrown in for good measure.

Not surprisingly, that northern group has a farthest north 546 dm contour. 

[OceanStWx]

5 minutes ago, CT Rain said:

It's easy when you're giving an interview to answer something honestly and with good intentions (social science!) only to have it blow up in your face when it's the one quote that goes "viral". 

There has to be a learning curve there. I mean relatively speaking Oklahoma is pretty weather savvy. They know the difference between a watch and a warning. It really isn't the case around here. The two sort of blend into one another. 

There's also a big difference between a 30 min tornado warning that keeps people sheltering versus a blizzard warning that spans 12+ hours. The societal impacts are much more significant in the latter case. 

[CT Rain]

21 minutes ago, OceanStWx said:

For this upcoming thing?

There's a couple clusters east and northeast of the ensemble mean, dominated by GEFS. A southwest cluster dominated by EPS. And a north cluster dominated by EPS, with a GEFS and CMC member each thrown in for good measure.

Not surprisingly, that northern group has a farthest north 546 dm contour. 

Oh I meant the later runs leading up to the blizzard. Do you see a big signal for the western outliers?

[OceanStWx]

4 minutes ago, CT Rain said:

Oh I meant the later runs leading up to the blizzard. Do you see a big signal for the western outliers?

Oh yeah, 24 hours later (14.00z) from what I posted and 3 of the 5 clusters were western solutions. Spoiler alert, nearly all of those members were EPS.

[Cold Miser]

21 hours ago, CoastalWx said:

This snowpack is nuts. I mean it's completely bulletproof. It's going to take a lot of sun to melt it.  We don't have much..maybe 5" OTG after the rain, but you can drive a sherman tank over it.

The concrete snow that just keeps getting stronger.  A bit of melting yesterday, and then this morning it was right back to solid again.

[Cold Miser]

12 hours ago, Ginx snewx said:

I wouldn't call this storm a bust just because the I 95 totals were low though. The societal impact for those 3-4 hrs was as I expected. I think the warnings were warranted even if the snow totals were not impressive.

This is a great map. I love the small weenie zones like the ones in New Hampshire, then that best of a spot near Bennington. 

[ORH_wxman]

ORH hasn't hit freezing since a week ago today....might manage to tickle above freezing today. That's pretty impressive of a stretch for mid-March. Go figure, after record warmth in the final week of February....ORH has set two record low maxes this month so far.

[Damage In Tolland]

12 minutes ago, ORH_wxman said:

ORH hasn't hit freezing since a week ago today....might manage to tickle above freezing today. That's pretty impressive of a stretch for mid-March. Go figure, after record warmth in the final week of February....ORH has set two record low maxes this month so far.

Same here though today will def do it both places 

[tamarack]

12 hours ago, OceanStWx said:

 I see a lot of internal discussions about going "at least 15:1" for an event. Well 15:1 is a rarity for most locations in New England. And at 70+ hours lead time, why are we assuming we'll have any confidence in a ratio forecast?

Sometimes we do it to ourselves, whether it's the internal weenie in us all, or something else.

Excerpting from the original...

Only one location, but my records bear that out.  Of 35 snowfalls 10" or greater since we moved here in May 1998, only 3 have had ratios 15:1 or higher and the highest was Dec. 29-30, 2013 with 18.0 (11.0" from 0.61".)  Dec. 6-7, 2003 was close, 24.0" from 1.63" for 14.7-to-one, and 4 other 10+ events topped 14:1.  Of those 8 events with ratios 14:1 or higher, 6 came at temps mid teens to low 20s, one milder and one colder.  At the other end, I've had 4 taint-free 10+ storms at under 9:1, with this week's 7.3 (15.5" from 2.12") leading the way and Jan. 2015 blizzard just missing at 9.2 (20.0" from 2.17".)  Temps for those 5 low-ratio storms are all over - singles to near freezing.

[Juliancolton]

We got an elusive post-snowstorm Landsat pass, awesome. Have some fun zooming in to see where the snow is and isn't: https://re.ssec.wisc.edu/s/hp2g3

 

[OceanStWx]

15 hours ago, jbenedet said:

Thanks for your response.

I agree, verification for this is still not set in stone; I'm just noting the trend that so far they are winning...

Synoptically speaking this event certainly isn't a strong analog--definitely not an apples to apples comparison--but I suppose I'm trying to isolate the common variable in this case, which is the robust 500 mb low occlusion near the Great Lakes. 

To me--again this is another hypothesis-- in the case of the blizzard bust I think the mesos keying in on the deepest convection and how that related to the location of best surface pressure falls is what allowed them to see the track closer to the coast. This is particularly important when H5 and H7 became cut off --losing baroclinity--and the atmosphere transitioned quickly to a quasi equivalent barotropic state. At that point perhaps convective drivers trumped UL synoptic features? There's obviously a lot more to this (e.g phasing) I just find it fascinating. In a perfect world Mother Nature would provide us with several more cases before this season is out to test and/or fine tune this hypothesis...

To avoid cluttering the model thread, I pulled this over here.

It looks like the early reviews on the blizzard verification where that the northern stream wave came in stronger, forced a larger ridge out ahead of it, and kept the southern stream from escaping east too quickly. 

Eventually at the short lead times (inside 24 hours) the significant convection with the southern stream helped enhance the ridge ahead of the northern stream wave even more.