Welcome to American Weather
Sign in to follow this  
Followers 0
BTRWx

March 13/14th Post Storm Undertaking (psu) Storm

63 posts in this topic

My first Amwx thread!

I can already see this particular event as a good case study!  Maybe we can copy posts (not just I) to this thread that can help improve forecasts for future events. :)

My personal notes are as follows.

- having no blocking high in place matters (especially early and late season)

- snow maps only support reality when ratios match (32 degree snowflakes are not the same as 20 degree snowflakes and sleet never has as high ratios as pure snowflakes)

- model trends are important (globals were trending badly (warmer) late in the game within the last 36 hours or so, even though short range guidance was improving (colder) with their late frames (later frames shouldn't be trusted as often)

 

What are everyone's thoughts on writing similar post-storm analysis threads for anomalous events? Booms/busts, surprise squalls, early/late season stuff, hecs (when they happen)...

Share this post


Link to post
Share on other sites

19 minutes ago, vinylfreak89 said:

As I promised in the banter thread, I would do a quick after action write up. So, here's what I got right. QPF was way underpredicted by the models, as expected things phased sooner, and we were certainly not too far south for this Miller B. Had the low been just 75 miles east, this would've been a much different storm for everyone. I expected the low to develop more along the gulf stream, since that seemed to be the usual location that lows want to jump considering all other factors being equal. Here is what I missed. The lack of a full phase and instead an energy transfer really hurt us. By staying disconnected from the main h5 low, the shortwave was not given enough room to amplify. The storm was and continues to be warm air advection dominant. A strong CCB never really got going (good prediction by the models until well in to NE PA/NY). Even there, it is not as impressive a CCB as many other large storms. The other thing that both myself and the models missed was the lack of blocking contributing to the fast movement. This is going to significantly cut down on totals and I will be surprised to see any reports over 20 inches. As far as what could've saved us in the mid levels is a stronger, more consolidated vort. Probably the biggest lesson I've learned from this storm is a good dynamic snow is entirely driven by h5. A 990 low off OC is great, but when you don't have a sharp vort, there will be a lot of warm air intrusion at the mid levels. A stronger 850 low could've helped here too.

Sent from my HTC 10 using Tapatalk
 

Share this post


Link to post
Share on other sites

6 minutes ago, high risk said:

I had mentioned Sunday that the GFS ran too cold in NC in the early January storm and voiced concerns that it would be too cold again.   It often misses those warm layers and shows too much of a rain/snow transition without enough sleet and/or freezing rain in between.     The 3 km NAM parallel nest ptype forecast from 12z yesterday was pretty decent.   I dismissed the idea of some freezing rain, but it was right.

gfs_mslp_pcpn_frzn_neus_4.png

nam3km_mslp_pcpn_frzn_neus_24.png

Share this post


Link to post
Share on other sites

I had resigned to flipping to sleet during the heart of the storm yesterday before it started but was worried the day before. This panel shows some key features that were the cause of the problem.

OaiMp0.jpg

 

 

The area of HP to the NW was useless. The NS energy was pumping SE flow west of the apps. It was 60 degrees yesterday in Charleston, WV. But we can survive with that no problem if there was a HP to the north feeding cold midlevel and surface winds into the developing cyclone. Instead there was HP over the Atlantic to the NE of the storm. Cyclonic flow around the storm and anti-cyclonic flow pumping around the HP to the NE creating very strong midlevel flow off the Atlantic. The storm had already wrapped in warm air from the SE and the HP to the NE over the Atlantic couldn't offset it.

The high and low placement allowed a much stronger push of warmer midlevel air westward over our area than was modeled leading in. Some of this was certainly cause by the track being on the left side of the envelope but this wasn't a classic CAD setup for our region. The insitu-cad overperformed as surface temps we far lower in most places than modeled. But there was no stopping the warm mid-level push westward. 

QPF verified through the corridor no problem. But the mix line pushing all the way to almost FDK did us no favors. The corridor almost always flips to sleet and mixed precip earlier than forecast. That's usually a given. Last night once it was confirmed that the mix line was going to push much further west than consensus forecast I knew the corridor was not coming back with snow. Irreversible damage was done. 

I knew Philly and NYC were in big trouble as well. There was no stopping warm air intrusion in the mid levels even to our NE. If it was going to push far north and west in MD, it was going to do the same in PA/NY. 

Share this post


Link to post
Share on other sites

2 minutes ago, ers-wxman1 said:

My gut feeling for several days was low amounts and mix for DC and points east. I never thought I would have a changeover in Ashburn at 1 am. Not only to sleet but freezing rain as well. Freezing rain is so rare climatologically in this area for Mid March let alone out here in Ashburn during a coastal storm. Sleet all the way though the Mason Dixon line equally shocking. I was expecting at least 8" of snow here and wound up with 2-3" of sleet/snow/ice slop. A colleague in Frederick county MD who lives near the MARC train POR station did receive 9.5" but well north and west. Snow forecasting can really humble  us all. 

Share this post


Link to post
Share on other sites
43 minutes ago, Bob Chill said:

I had resigned to flipping to sleet during the heart of the storm yesterday before it started but was worried the day before. This panel shows some key features that were the cause of the problem.

OaiMp0.jpg

 

 

The area of HP to the NW was useless. The NS energy was pumping SE flow west of the apps. It was 60 degrees yesterday in Charleston, WV. But we can survive with that no problem if there was a HP to the north feeding cold midlevel and surface winds into the developing cyclone. Instead there was HP over the Atlantic to the NE of the storm. Cyclonic flow around the storm and anti-cyclonic flow pumping around the HP to the NE creating very strong midlevel flow off the Atlantic. The storm had already wrapped in warm air from the SE and the HP to the NE over the Atlantic couldn't offset it.

The high and low placement allowed a much stronger push of warmer midlevel air westward over our area than was modeled leading in. Some of this was certainly cause by the track being on the left side of the envelope but this wasn't a classic CAD setup for our region. The insitu-cad overperformed as surface temps we far lower in most places than modeled. But there was no stopping the warm mid-level push westward. 

QPF verified through the corridor no problem. But the mix line pushing all the way to almost FDK did us no favors. The corridor almost always flips to sleet and mixed precip earlier than forecast. That's usually a given. Last night once it was confirmed that the mix line was going to push much further west than consensus forecast I knew the corridor was not coming back with snow. Irreversible damage was done. 

I knew Philly and NYC were in big trouble as well. There was no stopping warm air intrusion in the mid levels even to our NE. If it was going to push far north and west in MD, it was going to do the same in PA/NY. 

This is a great write up Bob Chill! Excellent analysis. Origin of the low and rapid deepening was also pulling high octane air/moisture up from the Carribean and with strong easterly winds of 40-50kts off the deck, cold air didn't stand a chance off the surface. Our air temperatures have been much above normal February, some January so the Atlantic at our latitude was also milder than normal for this time of the year. 

1 person likes this

Share this post


Link to post
Share on other sites

 

Except that is a 6-hour forecast.  No serious hobbyist or professional forecaster would have predicted all snow for "us" with a high in that location. The better question is why the models misplaced the high in earlier forecast cycles. 

The models read in SST data and forecast the winds.  The warm Atlantic and Bay should already be accounted for in the simulations. Vertical resolution issue? Treatment of atmosphere/water interface? 

I resigned myself to mostly sleet Sunday morning when the GFS "caved" and joined the Euro in predicting rising 850 temperatures for Monday evening and when the forecast lows started hugging the coast. So the forecast for the DC area verified.  Would have been very disappointed if I lived in the far northwestern suburbs. 

Was impressed that the global ensembles sniffed out the storms at least 7 days ahead and that PSU sniffed it out a day before that. 

I thought the NWS did a good job of locating watches. When PG County was not included in the Winter Storm Watch it was a red flag to me as I assume (with no inside information) that the NWS pads watches by adding counties along the edges and shift them to weather advisories at game time.  

 

 

Share this post


Link to post
Share on other sites
5 minutes ago, dallen7908 said:

 

Except that is a 6-hour forecast.  No serious hobbyist or professional forecaster would have predicted all snow for "us" with a high in that location. The better question is why the models misplaced the high in earlier forecast cycles. 

The models read in SST data and forecast the winds.  The warm Atlantic and Bay should already be accounted for in the simulations. Vertical resolution issue? Treatment of atmosphere/water interface? 

I resigned myself to mostly sleet Sunday morning when the GFS "caved" and joined the Euro in predicting rising 850 temperatures for Monday evening and when the forecast lows started hugging the coast. So the forecast for the DC area verified.  Would have been very disappointed if I lived in the far northwestern suburbs. 

Was impressed that the global ensembles sniffed out the storms at least 7 days ahead and that PSU sniffed it out a day before that. 

I thought the NWS did a good job of locating watches. When PG County was not included in the Winter Storm Watch it was a red flag to me as I assume (with no inside information) that the NWS pads watches by adding counties along the edges and shift them to weather advisories at game time.  

 

 

 

I think where all models failed in the medium range was how far south the vort dug. The trend for the 3 days leading in kept digging the vort further south and slowing it down. This kept raising and amplifying heights in the east. Initially all guidance pointed towards the southern vort staying in front of the northern vort. But it actually ended up trailing it.

This is the GFS @ d4 lead. Vort is in front of the energy digging down. Confluence to our NE. Looks pretty good.

gfs_z500_vort_atl_17.png 

 

Here is verification. Vort is trailing the NS energy and confluence is gone. And with the confluence went our better chances at maintaining a colder solution:

gfs_z500_vort_atl_1.png

If the vort was 6 hours quicker and not as far south then the precip totals would have been less but the inherent temp problems wouldn't have been nearly as bad. We ended up having "too much of a good thing" in the upper levels. That's another one of the problems other than not having HP directly to the north feeding in. A little means a lot in nearly all coastal events here. The models did a really good job identifying the threat at fairly long leads and also did a good job with the general progression. But the nitty gritty details got in the way here as they often do unfortunately. 

Share this post


Link to post
Share on other sites

Mount Holly gave kudos the mesos in their AFD-

The mesoscale models (NAM, RGEM) far outperformed the global models (GFS-ECMWF). The GGEM of the global models did a nice job showing ice near 195.

 

Share this post


Link to post
Share on other sites

When I put up the NAM three days ago with it suggesting flurries to drizzle to dry slot, I was castigated.

The NAM does not have such poor verification scores that it has an inverse correlation with reality.

Maybe I was right for the wrong reasons but still...when everyone is thinking the same thing, no one is thinking.

Share this post


Link to post
Share on other sites
13 minutes ago, winterymix said:

When I put up the NAM three days ago with it suggesting flurries to drizzle to dry slot, I was castigated.

That is literally the opposite of how the storm played out.

Share this post


Link to post
Share on other sites
1 hour ago, C.A.P.E. said:

Mount Holly gave kudos the mesos in their AFD-

The mesoscale models (NAM, RGEM) far outperformed the global models (GFS-ECMWF). The GGEM of the global models did a nice job showing ice near 195.

 

As bad as those models are with amounts they are better at detecting warm layers than globals.   NAM also handled the shape of the Precip shield very well.   

Oh the value that could be added if we had a trustable mesoscale model and snowfall algorithms that didn't ridiculously exaggerate marginal events.

 

1 person likes this

Share this post


Link to post
Share on other sites
1 hour ago, winterymix said:

When I put up the NAM three days ago with it suggesting flurries to drizzle to dry slot, I was castigated.

The NAM does not have such poor verification scores that it has an inverse correlation with reality.

Maybe I was right for the wrong reasons but still...when everyone is thinking the same thing, no one is thinking.

Who the heck had flurries drizzle dryslot? I'm confused with your post. 

The nam did very well with warm air intrusion inside of 24 hours. That's undeniable and that also when I pretty much started expecting mixing problems. But leading in the nam was all over the place. So much so that it probably covered every possible outcome at one point or another. 

The two euro runs before onset did really well after jumping around. It did a good job with the midlevels and qpf distribution. The gfs did a good job in general but was obviously too cold. It was the steadiest global with track and progression for days leading in. I'll never forget the gfs cold bias though. That's an imortant lesson for future storm. 

Imho- the ukie was the most accurate in the midlevels for the globals. It was screaming sleet to almost the catoctins and across the md pa border. Good job there for sure. 

A blend of all guidance leading in to game time left me thinking I was in for a lot of precip with a chunk not being the snow variety and that ended up being pretty accurate. I was surprised (just like every single other pro forecaster and enthusiast up and down the coast) with the extent of the mixing problems. But big storms have tricks up there sleeves. Happens almost every time. This one broke warm. Won't be the last time. That's for sure 

Share this post


Link to post
Share on other sites
1 hour ago, North Balti Zen said:

That is literally the opposite of how the storm played out.

 

41 minutes ago, Bob Chill said:

Who the heck had flurries drizzle dryslot? I'm confused with your post. 

The nam did very well with warm air intrusion inside of 24 hours. That's undeniable and that also when I pretty much started expecting mixing problems. But leading in the nam was all over the place. So much so that it probably covered every possible outcome at one point or another. 

The two euro runs before onset did really well after jumping around. It did a good job with the midlevels and qpf distribution. The gfs did a good job in general but was obviously too cold. It was the steadiest global with track and progression for days leading in. I'll never forget the gfs cold bias though. That's an imortant lesson for future storm. 

Imho- the ukie was the most accurate in the midlevels for the globals. It was screaming sleet to almost the catoctins and across the md pa border. Good job there for sure. 

A blend of all guidance leading in to game time left me thinking I was in for a lot of precip with a chunk not being the snow variety and that ended up being pretty accurate. I was surprised (just like every single other pro forecaster and enthusiast up and down the coast) with the extent of the mixing problems. But big storms have tricks up there sleeves. Happens almost every time. This one broke warm. Won't be the last time. That's for sure 

I see the points that both of you are making and agree for sure.  I'm not here to argue with anyone, this is my interest.

No one got flurries/drizzle/dry slot.  Many got 25% of the snow they thought they would get and then a cut off of strong rates before the middle layers would get cold again.  Right for the wrong reason, if you will.   I never looked at the UKIE, but will next time, thanks.

The NAM did bounce around but it did tend to consistently model strong vertical velocities overnight accompanied by warm air advection at 850 mb intruding almost to Gettyburg and then all of that out of here at sunrise and then quickly to the NE.  The NAM as well as a number of other models depicted the snow showery stuff we've been having on and off as the 500 mb low tries to race after the surface low to catch up with it.    In some NAM runs, dry slots did show up and this was modeled incorrectly.  However, we did get a sense from NAM that the intense mixed precipitation would cut off regarding intensity before the heights would crash in warp around.

Most of our areas mixed between 10 pm and 2 am and then crashed to frozen around 9 am.  The NAM thickness charts modeled this fairly well.

So let us get into specifics for a moment.  We have seen other storms such as this one, storms that try to phase but come through in pieces.

With any of the models, the elements I like to look at are,

-500 mb pattern

The NAM seemed to hint that the southern stream was scooting out faster than the northern and I think this kept the surface low close to the

coast and allowed so much warm air advection at 850 mb.

1000-500 mb thickness as well as other thickness depictions.

So many people in this forum saw what was happening and ignored warning signs and red flags about thickness. 

Warm air advection was modeled to at least Gettysburg and that is exactly what transpired.

If everyone discounted the model snow maps and put more emphasis upon 850 mb warm air advection as depicted in

the NWS mesoscale analysis real time runs as well as thickness models and real time depictions, the skill would/will improve.

I got it wrong too but the red flags were there in the NAM three days ago.

 

700 mb  not only shows the intense rates but also shows just how lacking in intensity the wrap around stuff and 500 mg stuff has

been.

Long story short, the NAM of three days ago threw out a red flag signal.

1 person likes this

Share this post


Link to post
Share on other sites

Two things I noticed:

First, the ZR in southeast parts of the area that did tree damage showed that the surface verified colder compared to all/almost all of the modeling. Even right on the Tidal Basin, there are those stunning pics of ice covered cherry blossom buds. So we had warmer than modeled mid-levels (except UKMet) and colder than modeled surface temps which made the forecast particularly difficult. But this reinforces the idea to go with the meso-models for surface temps, *especially* when ZR shows up as a precip type. 

Second---So this is now three splashy, headline making busts for the New York City office in three straight winters. Two were much less than forecast and one much more. For this storm, I did pay attention to the modeling for areas northeast of us, and could never figure out why they settled on 17-20" (20" at Central Park) as the "Most likely" number for the boroughs yesterday.  20" is a top seven event for the city.  Verbatim, several models showed the warm air intrusion getting awfully close to the city. 

And there's this article: http://hosted.ap.org/dynamic/stories/U/US_SCI_WINTER_WEATHER_FORECAST?SITE=AP&SECTION=HOME&TEMPLATE=DEFAULT&CTIME=2017-03-14-17-17-34

What do you all think about this approach of keeping the forecast high even when they didn't buy it themselves for the sake of public safety? That doesn't seem like a good approach and I can't see how this being out in the news is going to inspire any public confidence. 

1 person likes this

Share this post


Link to post
Share on other sites
51 minutes ago, gymengineer said:

Two things I noticed:

First, the ZR in southeast parts of the area that did tree damage showed that the surface verified colder compared to all/almost all of the modeling. Even right on the Tidal Basin, there are those stunning pics of ice covered cherry blossom buds. So we had warmer than modeled mid-levels (except UKMet) and colder than modeled surface temps which made the forecast particularly difficult. But this reinforces the idea to go with the meso-models for surface temps, *especially* when ZR shows up as a precip type. 

Second---So this is now three splashy, headline making busts for the New York City office in three straight winters. Two were much less than forecast and one much more. For this storm, I did pay attention to the modeling for areas northeast of us, and could never figure out why they settled on 17-20" (20" at Central Park) as the "Most likely" number for the boroughs yesterday.  20" is a top seven event for the city.  Verbatim, several models showed the warm air intrusion getting awfully close to the city. 

And there's this article: http://hosted.ap.org/dynamic/stories/U/US_SCI_WINTER_WEATHER_FORECAST?SITE=AP&SECTION=HOME&TEMPLATE=DEFAULT&CTIME=2017-03-14-17-17-34

What do you all think about this approach of keeping the forecast high even when they didn't buy it themselves for the sake of public safety? That doesn't seem like a good approach and I can't see how this being out in the news is going to inspire any public confidence. 

Shows how far we've come in 2 decades.   I remember March 93 had similar hype and some were making 18-24 forecast for I95.  We ended up with 10 inches with a lot of sleet and nobody complained. Now it's a huge bust.

Share this post


Link to post
Share on other sites

I hope this storm is a lesson to everyone who looks at those god awful snow maps. 

Share this post


Link to post
Share on other sites
11 minutes ago, mattie g said:

It's often the only hope we have...

well if anyone was hanging on to those snow maps were let down for sure. 

Share this post


Link to post
Share on other sites

Great write up Bob as usual!

I busted badly on my 81 Corridor forecast. There were two main reasons.

1) The QPF forecasts on the models were wrong for my area. Mixing is rarely an issue in my area. Although some of the late model runs brought mixing very close to Winchester. The GFS kept lowering QPF numbers for my area leading up to the storm and I should have paid attention to that. The Euro stayed rock steady at around an inch of QPF the last 2 days of runs. The NAM went ballistic with 1.6 and 1.8 inches of QPF leading up to the storm. I took a blend and relied on knowledge of my climo and figured I would end up with AT LEAST an inch of QPF with lift and other factors. It was close but fell just short of that number.

2) I counted on too high of ratio's for what did fall. This snow was WET. Extremely wet and the ratios paid a price for it. Even with surface temps in the mid 20's during the heaviest snow. That warm air intrusion made it somewhere in the upper levels even out this way. Not enough to melt the falling snow all of the way. Just enough to make it a very wet snow. 

Live and learn. On to the next one!

 

Share this post


Link to post
Share on other sites
1 hour ago, mappy said:

I hope this storm is a lesson to everyone who looks at those god awful snow maps. 

We said this after every syorm. You can't avoid them. They're all over every thread, and posts with actual meteorology get buried.

1 person likes this

Share this post


Link to post
Share on other sites
38 minutes ago, clskinsfan said:

Great write up Bob as usual!

I busted badly on my 81 Corridor forecast. There were two main reasons.

1) The QPF forecasts on the models were wrong for my area. Mixing is rarely an issue in my area. Although some of the late model runs brought mixing very close to Winchester. The GFS kept lowering QPF numbers for my area leading up to the storm and I should have paid attention to that. The Euro stayed rock steady at around an inch of QPF the last 2 days of runs. The NAM went ballistic with 1.6 and 1.8 inches of QPF leading up to the storm. I took a blend and relied on knowledge of my climo and figured I would end up with AT LEAST an inch of QPF with lift and other factors. It was close but fell just short of that number.

2) I counted on too high of ratio's for what did fall. This snow was WET. Extremely wet and the ratios paid a price for it. Even with surface temps in the mid 20's during the heaviest snow. That warm air intrusion made it somewhere in the upper levels even out this way. Not enough to melt the falling snow all of the way. Just enough to make it a very wet snow. 

Live and learn. On to the next one!

 

Even at the "height" of the storm around 4 AM I noticed the flake size was crappy-small.  No dendrites for sure, so the temps in that zone were too warm probably.  Add in some wind, and that gave us poorer-than-expected ratios.  Temp was 29 - 32F here for the entirety, so it was a warm event also.  Colder up your way, but that layer must've been shallow.  Models backing off QPF as a storm rolls in IS ALWAYS a red flag.  When that started happening Monday morning, I knew I was toast.

1 person likes this

Share this post


Link to post
Share on other sites
2 hours ago, mappy said:

well if anyone was hanging on to those snow maps were let down for sure. 

 

1 hour ago, Amped said:

We said this after every syorm. You can't avoid them. They're all over every thread, and posts with actual meteorology get buried.

I think the snow maps are useful in that not everyone has time to look through soundings for their backyards, and I think a lot of people might not know how to interpret a sounding well if they saw one.  There are a lot of readers for these forums; some are here to discuss the finer points of meteorology, some are here to reminisce about the great hailstorm of 1932, some are here because they enjoy the company, and some are here because they just want to know how much snow they're going to get.  On every thread there are posts some readers will love and others will want to ignore. 

In the end, the snow maps weren't too far off, providing you were looking at the right snow map for the right model.  Here are the latest snowfall totals from the storm:

F9ftASJ.jpg

And here's a snow map of the run of the 3k NAM just before the storm.  This map doesn't count sleet as snow, and it uses Kuchera ratios:

fr2UiL0.png

Not too far off.  And I think the differences are about as much due to errors in modeling as errors in snow mappery.

The problem is that there are some sites out there that don't have good algorithms for making snow maps, especially when temps and p-type are a concern.  Here's the TT snowmap for that same run of the 3k NAM.

0sJrAmD.png

This map includes sleet as snow at 10:1 ratios.  To their credit, TT is upfront about this, but "*includes sleet*" is a pretty important qualifier for a storm like this.  There's a big difference between getting 10" of snow and 2" of sleet.

It turns out that the quality of the snowmaps also depends on the model.  I didn't realize until this storm that apparently the HRRR includes sleet as snow in the model output, so even the maps on pivotalweather that normally exclude sleet will include it for the HRRR.

The good news is that maps are getting better, and forum members are getting more sophisticated about posting and interpreting them.  Just as we're getting better about interpreting the output of the various weather models.  While people will always occasionally post a bad snowmap or a questionable interpretation of model output, folks around here will generally be pretty quick to point out the problems.  It's how we all learn and how this forum keeps getting better.

1 person likes this

Share this post


Link to post
Share on other sites

Sure, if you were watching HRRR come in all night long and discounted a large amount of their totals as sleet, then you saw the change in totals. Other than the NAM at the 11th hour cutting back totals, not a single model nailed my backyard. I was expecting a foot. I got a third of that, most of it sleet. 

My point was, yes snow maps are pretty to look at, and I suppose one should look at them, but they shouldn't be used to forecast. And many here were hugging the **** out of some of these maps. 

Share this post


Link to post
Share on other sites
38 minutes ago, mappy said:

Sure, if you were watching HRRR come in all night long and discounted a large amount of their totals as sleet, then you saw the change in totals. Other than the NAM at the 11th hour cutting back totals, not a single model nailed my backyard. I was expecting a foot. I got a third of that, most of it sleet. 

My point was, yes snow maps are pretty to look at, and I suppose one should look at them, but they shouldn't be used to forecast. And many here were hugging the **** out of some of these maps. 

I've decided I hate snow maps in the short range. A pure clean storm with no chance of mixing is the only time snow maps are accurate but even then using soundings, surface, and midlevel temp plots and thinking it through is an exercise that is far more intellectual and fun. When I started seriously tracking weather back in 06 there was no such thing as a snow map. They are a recent development. There is no standardized consistency in the algorithm each vendor uses. Kuchera is relatively new and I like those the best if I had to pick one. 

My favorite pre-storm discussions are when Wes, Matt, Ian and I look at the fine details of all the different layers. Breaking down the atmosphere at the different levels and envisioning what sort of ground truth to expect is the art of this hobby. Snow maps are a shortcut, often misleading, and require no skill to understand. 

With that being said, individual ensemble member snow maps at range are pretty useful for setting confidence at longer leads. You get a fast glance at how much support there is for frozen precip in general within the ensemble means. There is value there. 

Share this post


Link to post
Share on other sites
55 minutes ago, mappy said:

Sure, if you were watching HRRR come in all night long and discounted a large amount of their totals as sleet, then you saw the change in totals. Other than the NAM at the 11th hour cutting back totals, not a single model nailed my backyard. I was expecting a foot. I got a third of that, most of it sleet. 

My point was, yes snow maps are pretty to look at, and I suppose one should look at them, but they shouldn't be used to forecast. And many here were hugging the **** out of some of these maps. 

Another point on the HRRR though, it really went off-the-rails at one point with the QPF amounts.  I think it is having some trouble with too much precipitation in banded features.

Share this post


Link to post
Share on other sites
2 hours ago, EastCoast NPZ said:

Even at the "height" of the storm around 4 AM I noticed the flake size was crappy-small.  No dendrites for sure, so the temps in that zone were too warm probably.  Add in some wind, and that gave us poorer-than-expected ratios.  Temp was 29 - 32F here for the entirety, so it was a warm event also.  Colder up your way, but that layer must've been shallow.  Models backing off QPF as a storm rolls in IS ALWAYS a red flag.  When that started happening Monday morning, I knew I was toast.

Good points all around. I need to actually do some more reading up on types of flakes and how the column temps and humidity corresponds to their formation. I didnt see any dendrites at all. The majority of the storm was definitely rimed flakes from what I can tell.

1 person likes this

Share this post


Link to post
Share on other sites
4 minutes ago, clskinsfan said:

Good points all around. I need to actually do some more reading up on types of flakes and how the column temps and humidity corresponds to their formation. I didnt see any dendrites at all. The majority of the storm was definitely rimed flakes from what I can tell.

Here's a scholarly article with tons of info on snow structures and the atmospheric conditions to see different types of snowflakes. There's some pretty intense sections, but plenty of really solid info with what you're looking for enjoy!

https://www.its.caltech.edu/~atomic/publist/rpp5_4_R03.pdf

1 person likes this

Share this post


Link to post
Share on other sites
1 minute ago, MillvilleWx said:

Here's a scholarly article with tons of info on snow structures and the atmospheric conditions to see different types of snowflakes. There's some pretty intense sections, but plenty of really solid info with what you're looking for enjoy!

https://www.its.caltech.edu/~atomic/publist/rpp5_4_R03.pdf

Thanks for this! I will knock it out this weekend.

1 person likes this

Share this post


Link to post
Share on other sites
1 hour ago, mappy said:

Sure, if you were watching HRRR come in all night long and discounted a large amount of their totals as sleet, then you saw the change in totals. Other than the NAM at the 11th hour cutting back totals, not a single model nailed my backyard. I was expecting a foot. I got a third of that, most of it sleet. 

My point was, yes snow maps are pretty to look at, and I suppose one should look at them, but they shouldn't be used to forecast. And many here were hugging the **** out of some of these maps. 

I get what you're saying, but some of that has more to do with the models than the maps.  A lot of models missed the warm layer until the last couple of days (and even then some of the globals struggled with it).  It's frustrating, but overall I still think the models are pretty impressive.  There's clearly still room for improvement when it comes to getting precipitation type right, which is understandable considering that a small error in temperature can have a big effect on what falls out of the sky.  If the models ever do get good enough to be able to nail p-type well in advance, I wonder if I'll miss these days.  In a way it's more interesting when we know there's a reasonably good chance of a last-minute surprise.

Regarding maps, another innovation that I've seen more use of is the "snow depth" maps.  I'm not sure what algorithm pivotalweather uses for theirs and I'm not sure how accurate they are, but I like the idea.  Snow depth gives a more complete picture of what to expect out of a storm.  I think yesterday was a good example.  It was snowing here at different rates through much of the day, and I'm pretty sure some of it accumulated, as areas that had been trampled sleet in the morning had a layer of snow on them by the evening.  But snow depth in my yard was roughly steady or even decreasing throughout the day, as the new accumulations were offset by melting/compacting of what was already on the ground.

1 person likes this

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!


Register a new account

Sign in

Already have an account? Sign in here.


Sign In Now
Sign in to follow this  
Followers 0

  • Recently Browsing   0 members

    No registered users viewing this page.