So What Did We Learn Here?

[Cold Rain]

10 minutes ago, NGTim said:

i learned that jburns needs to use a stronger password to not get hacked (or if not someone nearby really needs to go check on him)

CR, the "pink radar just means slightly colder rain"  made me chuckle more than anything else.

As far as the models and met interpretations, and this is coming from a simplistic view, which i think is good sometimes, is it not pretty good that for my area the roughly .3 inches of moisture that got to my specific area at a fairly specific time, and formed snow in the correct zone, survived the column, and was able to accumulate approx 3 inches based on ground temps and other factors, is it not pretty impressive that this was not a total surprise?  I just think we are fooling ourselves to ask for more at least right now.   Now take this with a grain of salt, i admit if I had got shafted I'd probably be looking for that guy who says he is sending emails to the govt agencies saying bla bla bla and I'd be angrily picketing outside the nws office.  (actually i followed my nws closely and I'd say they did a great job for my state).

I wish there was a way to estimate a snowstorm more like a hurricane (visually on a map), with an ever expanding cone of uncertainty for timing, distance, precip type and amount. 

Tim, you bring up a really good point.  Nailing down the exact amount of precip that falls as well as the exact temp profile of the entire column, where just a tenth or two in either direction can have major implications, is just beyond what we can do right now.  In the summer, nobody cares about that level of accuracy.  But in winter, those details matter and matter a lot.

Thats why in my opinion (which is based on years of watching these things fall apart -- but it's still just an opinion, which is no better than anyone else's), it's a safer and more accurate forecast to give more weight to the things that can screw up snow accumulations.  Identify the things that usually occur (less QPF than expected, stronger warm nose lasting longer than expected, quicker end to the precip than expected, etc.) and give those a much heavier weighing in the forecast.  That might be oversimplifying, but it would probably put you closer to reality more times than not.

[jburns]

27 minutes ago, NGTim said:

i learned that jburns needs to use a stronger password to not get hacked (or if not someone nearby really needs to go check on him)

 

3

Our next lesson will be sarcasm.

[frazdaddy]

Not sure I learned as much as had it reinforced that the only way I see meaningful snow is to drive to Rappahanock county.

[NGTim]

1 minute ago, jburns said:

Our next lesson will be sarcasm.

i was trying to continue the sarcasm, but i should realize you guys may not know me well enough to be aware of my extremely razor sharp, witty, and humble sense of humor.  I can say that was the most enjoyable thing I've read in a while.   But I did add a disclaimer just in case, to cover my butt, so maybe you did get me just a little bit ;-) 

[frazdaddy]

20 minutes ago, jburns said:

Our next lesson will be sarcasm.

helpless old man my @#$!

[WeatherExperiment]

On ‎1‎/‎7‎/‎2017 at 8:56 AM, Cold Rain said:

The NW "trend" is real and almost always occurs.

Models usually vastly underestimate both the degree and duration of warm layers.  If the model is even hinting at a warm layer, cut your snowfall forecast by at least 50%.

Even with ideal conditions, forecasted snowfall totals should be trimmed by at least 25%.

Can you be more specific about the NW trend "almost always" occurring? If the percentage of storms that trend this way is extremely high -- let's say for over 90% of the storms tracked -- then why can't the models simply account for these past storm trends, since they almost always occur, and simply take those trends into consideration instead of relying on forecasters to make the adjustments as you suggest they should? The real data for all of those storms is now available. Programmers should plug that data in and beta-test it alongside the current models and see if the beta-models perform better. Maybe they already do that. I'm not sure. But it doesn't seem to be working if they do.

[alhooks13]

i'm prepared to be made fun, have darts thrown at me and so forth.  To play devil's advocate and stimulate additional conversation, was it really that big of a bust? Was the QPF significantly off? There was a major SE snowstorm predicted and it materialized albeit a bit further NW of where most thought it would occur. Is a 25-50 mile shift to the NW a material and significant variance? Some would say yes when it is the difference in snow versus rain. I would posit perhaps not if temp variances are only a degree or two from forecast. Was the actual track and outcome in the realm of possibilities shown by some of the ensemble members?

 

[NCSNOW]

4 minutes ago, alhooks13 said:

i'm prepared to be made fun, have darts thrown at me and so forth.  To play devil's advocate and stimulate additional conversation, was it really that big of a bust? Was the QPF significantly off? There was a major SE snowstorm predicted and it materialized albeit a bit further NW of where most thought it would occur. Is a 25-50 mile shift to the NW a material and significant variance? Some would say yes when it is the difference in snow versus rain. I would posit perhaps not if temp variances are only a degree or two from forecast. Was the actual track and outcome in the realm of possibilities shown by some of the ensemble members?

 

Great post. Think you hit the nail on the head. Bust to me are h5 and placement of the bigger features, not so much a 25 to 50 mile shift in the transition line. Thought qpf and transition line was in the neighborhood of where we thought it would be. It worked out great for me this storm, but I've been burned by it before by as little as 10 miles. However in both cases along with several others I knew right before start time the pendulum could swing eitheir way. 

[Cold Rain]

23 minutes ago, WeatherExperiment said:

Can you be more specific about the NW trend "almost always" occurring? If the percentage of storms that trend this way is extremely high -- let's say for over 90% of the storms tracked -- then why can't the models simply account for these past storm trends, since they almost always occur, and simply take those trends into consideration instead of relying on forecasters to make the adjustments as you suggest they should? The real data for all of those storms is now available. Programmers should plug that data in and beta-test it alongside the current models and see if the beta-models perform better. Maybe they already do that. I'm not sure. But it doesn't seem to be working if they do.

Models frequently overestimate the cold press in the east and underestimate the strength of southern stream waves.  They tend to "see" them better as zero hour approaches, this the correction NW.  These biases can be really useful in forecast.  I don't know why they exist or why they can't be corrected.  Probably has a great deal to do with how well we can sample the atmosphere at this time.

[WeatherExperiment]

2 minutes ago, alhooks13 said:

i'm prepared to be made fun, have darts thrown at me and so forth.  To play devil's advocate and stimulate additional conversation, was it really that big of a bust? Was the QPF significantly off? There was a major SE snowstorm predicted and it materialized albeit a bit further NW of where most thought it would occur. Is a 25-50 mile shift to the NW a material and significant variance? Some would say yes when it is the difference in snow versus rain. I would posit perhaps not if temp variances are only a degree or two from forecast. Was the actual track and outcome in the realm of possibilities shown by some of the ensemble members?

 

Brad Panovich made a really good point in one of his post-storm broadcasts about the storm being a bust. He said if he had just said there would be a Trace to 8-inch chance across the entire north-to-south viewing area, his forecast would have been correct. This isn't what the public wants to hear though, so that's why it's broken down to try and cover the IMBY public need. But how many gradients are really necessary? Some mets use T-1", 2-4", 4-6", 6-8", etc. Others overlap -- T-2", 1-4", 3-6", 5-8", etc. Maybe a more uniform split of the counties should be used as the boundaries every time instead of creating arbitrary lines of snowfall for every storm and then the Mets can just fill in the totals for each sectioned-off area (i.e.: Mountains, Piedmont, Foothills, Sandhills).

[NorthernUpstateSC]

13 hours ago, drfranklin said:

for those in Upstate SC: move/live "north" of the magical I-85 if you desire snow

snowbird: where specifically do you live? I was under the impression that Cashiers/Lake Toxaway/Brevard rec'd a significant amount of snowfall

So true.. we ended  up with just over 5 inches of snow close to hwy. 11 and one of my buddies in Anderson got next to nothing other then rain and sleet and a little back end snow shower.

[NGTim]

21 minutes ago, WeatherExperiment said:

Brad Panovich made a really good point in one of his post-storm broadcasts about the storm being a bust. He said if he had just said there would be a Trace to 8-inch chance across the entire north-to-south viewing area, his forecast would have been correct. This isn't what the public wants to hear though, so that's why it's broken down to try and cover the IMBY public need. But how many gradients are really necessary? Some mets use T-1", 2-4", 4-6", 6-8", etc. Others overlap -- T-2", 1-4", 3-6", 5-8", etc. Maybe a more uniform split of the counties should be used as the boundaries every time instead of creating arbitrary lines of snowfall for every storm and then the Mets can just fill in the totals for each sectioned-off area (i.e.: Mountains, Piedmont, Foothills, Sandhills).

i don't see why anyone would want to give tight gradients, when so many people are affected, there are always going to be busts.  My range from the nws was at least 0, and at most 7, with 4 most likely, and my county got 3 to 4 - good.  in atlanta, for example, they were in at least zero, and maybe 4 max?, so some got zero, which really sucks, but that was in the range. 

[griteater]

2 hours ago, snowbird1230 said:

If you will look at the radar returns for the storm, the snow in East Tennessee came up from the southwest out of Alabama and Georgia. Never did the snow build back into Tennessee out of the east from North Carolina

Just a thought here...outside of the Carolina Crusher (Jan 2000), this pretty much never happens.  Now, precip can pop on the backside of a storm as the trailing upper wave approaches and the comma head is rotating thru, especially if it is a strong and slow moving upper wave.

[jburns]

2 hours ago, frazdaddy said:

helpless old man my @#$!

I'm hurt to the quick by your callous comment. 

[Lookout]

On 1/7/2017 at 1:10 PM, palmettoweather said:

If you are in the Upper Savannah River Valley and the NAM is the only one cutting your totals, you better believe it. It sniffed out the warm nose decently well, especially the 4 km and the 3 km Para if I remember correctly.

Sent from my SM-G920I using Tapatalk
 

 

actually the gfs was sniffing it out for a LONG time...run after run it didn't waver on the warm surface temps...even though it was too cold from 800mb to 925mb. I actually thought it was too warm at the surface because of the temps it was showing aloft and i think if it had actually verified that cold, those surface temps would have been colder and the result different. But what hurt the most...by far..as the warm nose aloft. The only chance we had to evaporate the warm surface layer in northeast ga was heavy snow falling into it and cooling it from the top down. Unfortunately there was only rain falling into it thanks to that ewl. 

Another thing that hurt was there was literally zero dry air to work with and get some evap cooling. Dewpoint depressions were very low with dewpoints only in the low to mid 30s. Not going to cut it. Models do a MUCH better job with surface temps when there is not much or any potential for evap cooling. That is the most likely time the models are too warm...but absent that dry air the models surface temps are much more reliable. 

On 1/7/2017 at 1:36 PM, Queencitywx said:

We need more damn balloon launches. 

Isn't that the truth.  I know the cost and all but it would seem like there should be more of a way to get more of these balloons in the air, especially during winter threats. it's very easy to see how the models and everyone else misses something when the launches are so far apart as ffc and gso. 

13 hours ago, HWY316wx said:

Something that keeps rolling around in my mind is why I-85?  Seriously, why in the world does it seem like so many weather events use the interstate as a fault line? From SW GA all the way up through NC.  I know it's not all the time, but sure does seem like we talk about I-85 and I-40 a lot on this board.

One thing I learned to start to factor in is, IT WILL COME NW.  It's just a matter of when on the models and how far in real life.

Hats off to everyone on the board.  I personally want to thank, GRIT he does great PBP on the EURO.

I've always wondered this myself. I can't tell you how many times i've seen the changeover setup within a few tens of miles of me, one way or the other. Most times in marginal situations it always is north of here but even in colder situations where it is snow or ice where I"m at...it seems like south of highway 78 is always mostly rain. It's understandable to a degree when dealing with a wedge because there is a limit to how far south the wedge can extend subfreezing temps but what always amazes me is when it happens when dealing with temps aloft. It's harder to understand why at 5000 or 6000 feet that boundary sets up in the same 10 to 25 mile wide area the vast majority of times. 

 

3 hours ago, CentralNC said:

For areas south of I85 you have to have a pressing HP from the NE building down as the storm is forming.  I was never sold on N GA getting into this storm that much.

Also, EURO is great for sniffing things out in the long term but not as good closer in.  One of the few times I can recall sitting in the western piedmont that I was pretty comfortable telling my friends that we would stay all snow.  Of course QPF was going to be a question but ptype around where I am was not an issue.  HP was strong enough to get the job done here.

If there was a high, all it would have done was produce more sleet and freezing rain in georgia. The lack of snow was primarily due to warm air aloft where unless you have a monster high where the wedge extends up to the 850mb level, which is rather rare and reserved to the strongest of wedges,  it wouldn't have mattered as far as snow is concerned. hell in this case, the elevated warm layer was even above 850mb....at 750mb to 825mb...so even the strongest of wedges wouldn't have touched that warm nose. 

As for the euro..bleh..once again i'm not terribly impressed with it. It was way to the south for quite a while and actually had to play catch up to the gfs. 

One thing i have to say i learned is don't underestimate the hrrr on precip types...likewise with the 4km nam.  Both were very good on picking up where it was. I was right on the line of snow/sleet in hall county....with rain just east of me and all snow to the west... and it was pretty much spot on showing lake lanier as the dividing point. ..it's still normally too warm at the surface..but with respect to the elevated warm layer it did well. 

 

 

 

 

[Tony Sisk]

12 hours ago, palmettoweather said:

My two cents would be that interstates are going to be laid out at pretty consistent elevation levels for ease of construction . There will obviously be deviations from this, but I-85 is just following the topography of the land, which also often delineates climatology. I may be way off on this though.

I agree with this, spot on. I-85 basically runs at the foot of the blue ridge mountains...hence, the foothills. elevations shift pretty quickly from Atlanta, GSP, Charlotte...

so using I-85 is an excellent reference point. And sense most of our winter storm potentials move northeast, it's a perfect line.

[CaryWx]

Look at that precip-type totals map of NC again. This does not look like your typical LP "warm nose" event.  Too SW/NE oriented and fairly straight.  Realizing of course this is a 2 dimensional map and the above 0 Celsius occurred at the 850-800 layer.  

I watched the precip radar returns that night and watched the snow line approach from the NW towards Chapel Hill. It consistently advanced until about 6pm and then stopped, retreated, advanced slightly, retreated again and more or less remained there until the 850s crashed.  

This did not look warm nose-ish to me.  It looked like the cold air could not break down the SE ridge until more HP reinforcement arrived so to speak. (Crashing 850s). By that time as we all know your best precip has passed you by.   Precip onset was too early--by about 6hrs. Thoughts?

[BlunderStorm]

4 hours ago, snowbird1230 said:

If you will look at the radar returns for the storm, the snow in East Tennessee came up from the southwest out of Alabama and Georgia. Never did the snow build back into Tennessee out of the east from North Carolina

http://weather.rap.ucar.edu/radar/displayRad.php?icao=KMRX&prod=n0r&bkgr=color&endDate=20170107&endTime=10&duration=12 Although not building back necessarily I was referring to the radar returns showing development behind the main band which prolonged the event as it advanced northeast. Not really to the east of you but north of you. Though I do agree it was't blockbuster snow material.

[UpStateCAD]

17 hours ago, Cold Rain said:

The weather can surprise us, and it often does.  Unfortunately, it usually surprises to the downside around here.  It takes so many things to go right in order to get a snowstorm here, and we've become too accustomed to relying on models to tell us what to think.

I used to hear Fishel and other TV mets, for example, say things like "the models are showing this, but I'm not buying it because (whatever the reason was)."  Not much anymore.  Now the forecast is just a blend of various models and ensemble data.

I know the models are better than they used to be and they're really the primary tool for making a forecast.  But the reality is, there's a booby trap in virtually every digital snowstorm and usually more than one.  We really don't give them the proper weight they deserve.

It takes everything to go right to create a snowstorm here.  If one thing is off, it ruins the forecast...and usually by a lot.  And that potential degree to which it will ruin the forecast is usually not factored in as much as it should be.  If a forecast is for 6" of snow here, and the rule of thumb  was to cut it in half, that would turn out to be more accurate, I would bet, in 75% or more of the cases.  Just that one rule.  You don't even need to look at anything else.  Take your forecast and cut the amount by 40-50% and that would be closer to the truth in most cases.  It is just the way it works here.

one thing I remember DT saying when he used to post here was that he would look at a given storm and start asking himself what could go wrong that would prevent said storm from producing......probably a good tact to adopt.

[Avdave]

So after reading this thread, and some people are saying all the models and ensembles are useless, there should be a lot less posting when the next time there is a winter wx event in the southeast. That will make it easier for Jburns and Lookout as they wont have to be screening so many posts

[nomanslandva]

I know NC did not get the crazy lows post storm but where the heavy snow ended up really being did end up with some low numbers.  I think we got to 8 o 9+ but I was looking at Wundermap this am an a big bubble south of Richmond had some serious low numbers.  I saw lots of - 7,8,9 and one -13.  I did not believe the models showing that and that snowpack would make that big of a difference but dang that is cold for south central VA.  

     

[WeatherExperiment]

54 minutes ago, Avdave said:

So after reading this thread, and some people are saying all the models and ensembles are useless, there should be a lot less posting when the next time there is a winter wx event in the southeast. That will make it easier for Jburns and Lookout as they wont have to be screening so many posts

I think the underlying frustration has more to do with simply wanting better accuracy from the data these models provide so forecasters can interpret the model data based on other known factors that computers can't extrapolate. How that can be accomplished is the ongoing topic of discussion once the storm(s) pass. Improvement on both sides of the equation, from the actual model data to how we analyze that data (such as knowing that the majority of the time a NW trend occurs) will improve forecasts. But it doesn't help when we see clown maps that completely skew in a different direction. Of course the danger is that we underestimate the storm and it then overperforms. But more often than not, what happens here in the Southeast, anyway, is that we overestimate it.

[gman]

I am 59 years old, here long enough to remember Monty Dupree being our weatherman on WFBC-TV. It sure seems like he always got the snow forecast right. He had no computers, yet when he gave the forecast, you could usually count on it happening. Could it be we have gotten too sophisticated for our own good? Could it be climate change and global warming is real? I have learned to simply watch the radar. The radar never lies.

[jburns]

1 hour ago, Avdave said:

So after reading this thread, and some people are saying all the models and ensembles are useless, there should be a lot less posting when the next time there is a winter wx event in the southeast. That will make it easier for Jburns and Lookout as they wont have to be screening so many posts

2

Don't forget Buckeye. we only had to delete about 1400 posts from the storm thread. Of course, there were probably another 400 or so that should have been deleted that got by us.  Busy few days but at least the thread remained readable.

[griteater]

4 hours ago, CaryWx said:

Look at that precip-type totals map of NC again. This does not look like your typical LP "warm nose" event.  Too SW/NE oriented and fairly straight.  Realizing of course this is a 2 dimensional map and the above 0 Celsius occurred at the 850-800 layer.  

I watched the precip radar returns that night and watched the snow line approach from the NW towards Chapel Hill. It consistently advanced until about 6pm and then stopped, retreated, advanced slightly, retreated again and more or less remained there until the 850s crashed.  

This did not look warm nose-ish to me.  It looked like the cold air could not break down the SE ridge until more HP reinforcement arrived so to speak. (Crashing 850s). By that time as we all know your best precip has passed you by.   Precip onset was too early--by about 6hrs. Thoughts?

I agree with your statements.  This was more a case of the air mass ahead of the storm just wasn't quite cold enough for CLT To RDU and southeast as opposed to it being cold, then having a strong 850mb low track into the area with a surge of warmth aloft.  Instead, the warm air aloft was already in place pre-storm and won out thru the storm.

[jshetley]

2 hours ago, gman said:

I am 59 years old, here long enough to remember Monty Dupree being our weatherman on WFBC-TV. It sure seems like he always got the snow forecast right. He had no computers, yet when he gave the forecast, you could usually count on it happening. Could it be we have gotten too sophisticated for our own good? Could it be climate change and global warming is real? I have learned to simply watch the radar. The radar never lies.

Charlie Gertz did a great job too on that same channel. He usually got things right too. He knew how things worked around here from experience and did not hesitate to go against everyone else when he thought they were wrong. I'll never forget Jan 21-22 1987. While everyone said the I-85 corridor and south of there would see mostly, if not all rain, he went with mostly snow in that part of his viewing area and turned out to be right. It did start as rain, but went to snow within about 3 hours and did not change back. It can be hard to go against the models I guess, but sometimes it may need to be done.

[Tacoma]

4 hours ago, gman said:

I am 59 years old, here long enough to remember Monty Dupree being our weatherman on WFBC-TV. It sure seems like he always got the snow forecast right. He had no computers, yet when he gave the forecast, you could usually count on it happening. Could it be we have gotten too sophisticated for our own good? Could it be climate change and global warming is real? I have learned to simply watch the radar. The radar never lies.

I am 65 and also remember all the old weather mets including Monty Dupree, as I have said in the past back in the day these mets never seemed to have the trouble about would it be snow or rain , would it be cold enough, how many inches,,,these old mets did a great job back in the day without all these computers plus we don't seem to get the snowstorms we use to.  I remember missing the whole month of school Feb. 1962 , I believe was the year so we went to school I think every Saturday in April and May.  There was a good snow it seemed like at least once a week. 

 

[LovingGulfLows]

The GFS sucks. If it's on an island vs. every other global and mesoscale model, throw it out.

[Cold Rain]

8 minutes ago, LovingGulfLows said:

The GFS sucks.  throw it out.

Fyp!

[wh_adkins01]

Lol...that many a few aboard this train were screaming torch 10 days ago. Ol lady nature threw the family about 10" this weekend. Few down east are not to joyful right now. I gotta say though, weather or not we ever end up right, I always enjoy the ride. I've learned to not trust a word nor computer beyond just a few....