Wednesday Soaking

[earthlight]

we saw that happen before a few of our snowstorms last winter

Oh yeah, those were good times sifting through those posts.

"NAM HAS CUT THE PRECIP TO 0.70". WHAT DO WE DO NOW"

[Brian5671]

I'm still confused...what does spin up mean? And the NAM does have a history of chopping it's output right before ANY event....not just a big wound up storm

[TheTrials]

so the NAM has a spin up problem but the gfs, euro, cmc, rgem, ukmet don't? here is what i can find on model spin up and these are both from 2003.

http://www.oc.nps.edu/nom/modeling/initial.html

http://climate.envsci.rutgers.edu/pdf/CosgroveLDASspinup2002JD003316.pdf

[forkyfork]

given the strength of the upper level system, i'm leaning west for the heavy precip amounts... closer to the nam/sref

[Allsnow]

I'm still confused...what does spin up mean? And the NAM does have a history of chopping it's output right before ANY event....not just a big wound up storm

I believe they are referring to the fact the Nam is prone to spitting out ridiculous preciep totals in certain situations, yet wennies believe it.

[TheTrials]

I believe they are referring to the fact the Nam is prone to spitting out ridiculous preciep totals in certain situations, yet wennies believe it.

He is referring to that, but this is the second time in the last year he has come in here and said the model can't spin up, yet how come all the other models can do it? Hopefully one of the Mets or OP can chime in here and give us an explanation of this spin up problem so we can all learn from it.

[TheTrials]

And a request on the main forum for spin up to be explained is deleted, so much for trying to educate posters.

[nycsnow]

How about winds?

[Allsnow]

Euro is now close to 2 inches for the area. And says we do it again next week also.....fwiw

[CAT5ANDREW]

Euro is now close to 2 inches for the area. And says we do it again next week also.....fwiw

seasonal trend.....or trend over the last 2 3/4 months has been wet,I would say the wetter solution is the way to go.

[earthlight]

I'm still confused...what does spin up mean? And the NAM does have a history of chopping it's output right before ANY event....not just a big wound up storm

What I believe he's talking about is largely a myth. The models aren't less likely to "spin up" the systems dynamics at 6, 12 or 18 hours than they were at 24, 36, or 48 hours. However, sometimes this is what people think they are seeing (the models not having enough time to "develop the system"). Truth is, if the initialization is correct, the model should theoretically be able to capture the same dynamics and energy that it would at other ranges.

[TheTrials]

What I believe he's talking about is largely a myth. The models aren't less likely to "spin up" the systems dynamics at 6, 12 or 18 hours than they were at 24, 36, or 48 hours. However, sometimes this is what people think they are seeing (the models not having enough time to "develop the system"). Truth is, if the initialization is correct, the model should theoretically be able to capture the same dynamics and energy that it would at other ranges.

would be great if the OP would come back in and explain what he meant and teach us about this. Forky responded in the affirmative, perhaps he can teach us about spin up.

[forkyfork]

What I believe he's talking about is largely a myth. The models aren't less likely to "spin up" the systems dynamics at 6, 12 or 18 hours than they were at 24, 36, or 48 hours. However, sometimes this is what people think they are seeing (the models not having enough time to "develop the system"). Truth is, if the initialization is correct, the model should theoretically be able to capture the same dynamics and energy that it would at other ranges.

if you have a dynamic system developing over the gulf and/or gulf stream, latent heat release due to condensation in the lower tropopause acts to decrease static stability. this is a positive feedback mechanism and the models can lag a bit in developing a storm quickly enough (boxing day). baroclinic instability made a very good post on this subject

[TheTrials]

if you have a dynamic system developing over the gulf and/or gulf stream, latent heat release due to condensation in the lower tropopause acts to decrease static stability. this is a positive feedback mechanism and the models can lag a bit in developing a storm quickly enough (boxing day). baroclinic instability made a very good post on this subject

Thank you for the explanation. If you have any articles on this, it would be great if you could post them so we can read them.

[earthlight]

if you have a dynamic system developing over the gulf and/or gulf stream, latent heat release due to condensation in the lower tropopause acts to decrease static stability. this is a positive feedback mechanism and the models can lag a bit in developing a storm quickly enough (boxing day). baroclinic instability made a very good post on this subject

Yeah, I remember the extensive discussion regarding that. However, I have been hearing the term being thrown around wildly ever since---it's the point now where people legitimately believe that models aren't spinning up storms fast enough at 6 or 12 hours. So that's mainly what I was trying to clear up.

But I distinctly remember the positive feedback on the NAM and the lagging strengthening of the system...it was almost a domino type effect. For several runs it was de-amplifying the ridge ahead of the low pressure just enough so that the surface low and associated dynamics could escape east by ~100 miles. I believe you, myself, dsnow and baroclinic had a conversation regarding this during the 06z runs a few nights before the storm.

[TheTrials]

Yeah, I remember the extensive discussion regarding that. However, I have been hearing the term being thrown around wildly ever since---it's the point now where people legitimately believe that models aren't spinning up storms fast enough at 6 or 12 hours. So that's mainly what I was trying to clear up.

But I distinctly remember the positive feedback on the NAM and the lagging strengthening of the system...it was almost a domino type effect. For several runs it was de-amplifying the ridge ahead of the low pressure just enough so that the surface low and associated dynamics could escape east by ~100 miles. I believe you, myself, dsnow and baroclinic had a conversation regarding this during the 06z runs a few nights before the storm.

I remember that convo, I thought that was more convective feedback from the dynamic systems, not the time factor of getting the storm going.

[baroclinic_instability]

if you have a dynamic system developing over the gulf and/or gulf stream, latent heat release due to condensation in the lower tropopause acts to decrease static stability. this is a positive feedback mechanism and the models can lag a bit in developing a storm quickly enough (boxing day). baroclinic instability made a very good post on this subject

True and not true. Model spin-up these days has been taken care of with "hot start" initializations and other advanced schemes to solve this problem. Cold starting used to require spin up times to resolve cloud bearing features, etc. I would ask dtk though he could give a way better explanation.

But the rest here is true

[TheTrials]

If spinning up a storm in short time is a problem for models, then how does the RUC do it, and its a problem for the NAM, why doesnt the RGEM do it, or the srefs, or the hi res models, and if the other models don't do it, why can't the NAM be corrected to not do that?

[TheTrials]

True and not true. Model spin-up these days has been taken care of with "hot start" initializations and other advanced schemes to solve this problem. Cold starting used to require spin up times to resolve cloud bearing features, etc. I would ask dtk though he could give a way better explanation.

But the rest here is true

thank you. Coudl you put a little blurb into the thread I started on the main page? More people need to hear this.

[CoastalWx]

Yeah, I remember the extensive discussion regarding that. However, I have been hearing the term being thrown around wildly ever since---it's the point now where people legitimately believe that models aren't spinning up storms fast enough at 6 or 12 hours. So that's mainly what I was trying to clear up.

But I distinctly remember the positive feedback on the NAM and the lagging strengthening of the system...it was almost a domino type effect. For several runs it was de-amplifying the ridge ahead of the low pressure just enough so that the surface low and associated dynamics could escape east by ~100 miles. I believe you, myself, dsnow and baroclinic had a conversation regarding this during the 06z runs a few nights before the storm.

This is a very simple (old timer rule as well), but some wise forecasters that I know, always told me how Gulf lows sometimes track closer than modeled. We know the physical reasons behind it...but many times this is true.

[baroclinic_instability]

If spinning up a storm in short time is a problem for models, then how does the RUC do it, and its a problem for the NAM, why doesnt the RGEM do it, or the srefs, or the hi res models, and if the other models don't do it, why can't the NAM be corrected to not do that?

It really isn't a problem of the NAM, it was just that storm (boxing Day, and even more so Groundhogs Day) it was being limited significantly by negative effects of poor convective precip modeling. Remember these numerical models are just computer code, and they all have advanced convective precipitation schemes that are "activated" by the model which then simulates DMC (deep, moist, convection), but under certain circumstances extreme synoptic storms can result in overdone convective precipitation in the model (which can then alter the model phyics/performance greatly...often in a negative manner). The NAM, under some circumstances, has always seemed to have the most difficult time with the role of DMC and intense synoptic storms. The new NAM rolls out tomorrow though (12Z).

[earthlight]

This is a very simple (old timer rule as well), but some wise forecasters that I know, always told me how Gulf lows sometimes track closer than modeled. We know the physical reasons behind it...but many times this is true.

It's something I definitely have noticed over the past several years for sure. Obviously there are exceptions, but when you take the vast majority of gulf low events that are tracking just off the coast, they are either whiffs to the east completely..or they tuck in a bit farther northwest than the guidance forecast.

[TheTrials]

It really isn't a problem of the NAM, it was just that storm (boxing Day, and even more so Groundhogs Day) it was being limited significantly by negative effects of poor convective precip modeling. Remember these numerical models are just computer code, and they all have advanced convective precipitation schemes that are "activated" by the model which then simulates DMC (deep, moist, convection), but under certain circumstances extreme synoptic storms can result in overdone convective precipitation in the model (which can then alter the model phyics/performance greatly...often in a negative manner). The NAM, under some circumstances, has always seemed to have the most difficult time with the role of DMC and intense synoptic storms. The new NAM rolls out tomorrow though (12Z).

Thank you again. We have seen discussions on the convective feedback and latent heat issues of the NAM, its good to see it called out for what it is and not generalized into some kind of "the models can't get a storm going right away"

[earthlight]

It really isn't a problem of the NAM, it was just that storm (boxing Day, and even more so Groundhogs Day) it was being limited significantly by negative effects of poor convective precip modeling. Remember these numerical models are just computer code, and they all have advanced convective precipitation schemes that are "activated" by the model which then simulates DMC (deep, moist, convection), but under certain circumstances extreme synoptic storms can result in overdone convective precipitation in the model (which can then alter the model phyics/performance greatly...often in a negative manner). The NAM, under some circumstances, has always seemed to have the most difficult time with the role of DMC and intense synoptic storms. The new NAM rolls out tomorrow though (12Z).

The NAM is literally notorious for the overdone convective precipitation. In laymans terms, the QPF blob, has become a well known feature of the NAM since it began back in 2006 or 2007 (I can't remember). I'll never forget one storm in 2007-2008 (I believe it was January 2008, don't hold me to it)...where the NAM literally cemented itself in history for the completely busted "activated" convective precipitation scheme.

The general idea behind the storm was that deepening along the gulf stream and very intense forcing/lift would allow for very heavy precipitation and rapid cooling of the boundary layer. Temperatures were in the 40's during the day of the storm, but the model was insistent that the boundary layer would cool enough for very heavy snow.

A day later the storm was gone and we had a trace of snow on the ground--the heavy precipitation never developed...was completely overdone by the NAM. Now admittedly, the NAM was the only model still showing this intensity around 24-48 hours. Still, this is a fully operational NCEP model which is expected to be at least somewhat reliable.

So at the time, it was pretty shocking to me. Obviously they have improved on the model since then, and will be doing so again tomorrow, but that event will always stick in my memory.

[CoastalWx]

It's something I definitely have noticed over the past several years for sure. Obviously there are exceptions, but when you take the vast majority of gulf low events that are tracking just off the coast, they are either whiffs to the east completely..or they tuck in a bit farther northwest than the guidance forecast.

When I think back to Boxing Day..the one thing that I remember thinking to myself..is that position of the coastal front and how the H5 height field was already going almost negative to the south. We just had enough blocking to keep that storm from going Dec 2000 on us. Given that H5 position, lows love...LOVE to form along that cstl front that separates north winds to the west and moist e-se winds to te east of the front. The lows will shoot nnw right along that front...hook a bit to the left into the thermal gradient before either drifting east, or semi-occluding and a new low taking over to the east of the main low. We could see that coming from far away. Tie all that into Baroclinic's thoughts...and there ya go.

[tornadojay]

I haven't been keeping up on what changes are coming with the updated NAM, but are any of the convective feedback issues supposed to be addressed? If not, what kind of updates are being made?

[earthlight]

When I think back to Boxing Day..the one thing that I remember thinking to myself..is that position of the coastal front and how the H5 height field was already going almost negative to the south. We just had enough blocking to keep that storm from going Dec 2000 on us. Given that H5 position, lows love...LOVE to form along that cstl front that separates north winds to the west and moist e-se winds to te east of the front. The lows will shoot nnw right along that front...hook a bit to the left into the thermal gradient before either drifting east, or semi-occluding and a new low taking over from the east. We could see that coming from far away. Tie all that into Baroclinic's thoughts...and there ya go.

Agreed. I kept saying that it was a ticking time bomb ready to explode. I will say, though, I thought we were sh** out of luck when the 00z/24 runs came in and whiffed.

That storm was amazing in the way the entire thing transpired. I'll never forget how it played our almost like a snow-lovers movie. After the 00z/24 runs whiffed, I decided to stay up for the 06z runs. I was disappointed, but took at least an hour to analyze the entire synoptic setup. When I looked in the Plains, I saw a huge difference maker--which was a shortwave diving due southward from North Dakota to Oklahoma and then amplifying the trough base. This feature was literally non-existant on forecast models the day before.

You can imagine how p*ssed off I was. This feature, which could potentially save the day, was going to miss amplifying the entire trough/phase by 50 miles. I went to bed, after a long conversation in the NYC Thread with baroclinic, angry. Obviously, this feature got into the right position and the rest is history. But I will never forget how it all played out.

[usedtobe]

True and not true. Model spin-up these days has been taken care of with "hot start" initializations and other advanced schemes to solve this problem. Cold starting used to require spin up times to resolve cloud bearing features, etc. I would ask dtk though he could give a way better explanation.

But the rest here is true

That's also my understanding but as you say Daryl would know more about it.

[usedtobe]

The NAM is literally notorious for the overdone convective precipitation. In laymans terms, the QPF blob, has become a well known feature of the NAM since it began back in 2006 or 2007 (I can't remember). I'll never forget one storm in 2007-2008 (I believe it was January 2008, don't hold me to it)...where the NAM literally cemented itself in history for the completely busted "activated" convective precipitation scheme.

The general idea behind the storm was that deepening along the gulf stream and very intense forcing/lift would allow for very heavy precipitation and rapid cooling of the boundary layer. Temperatures were in the 40's during the day of the storm, but the model was insistent that the boundary layer would cool enough for very heavy snow.

A day later the storm was gone and we had a trace of snow on the ground--the heavy precipitation never developed...was completely overdone by the NAM. Now admittedly, the NAM was the only model still showing this intensity around 24-48 hours. Still, this is a fully operational NCEP model which is expected to be at least somewhat reliable.

So at the time, it was pretty shocking to me. Obviously they have improved on the model since then, and will be doing so again tomorrow, but that event will always stick in my memory.

Almost all models suffer for precipitation bombs at times due to the mismatch between the model grid size and the scale that convection occur on. The reason convective schemes are in models is to keep the grid scale instabilities from mucking up the model too badly. In certain patterns the nam seems to suffer worse but there have been times when the gfs also goes bonkers, not as often as it used to. You sometimes see that in handling bogus tropical systems in the longer time ranges though the models do not seem to forecast as many bogus tropical systems as they used to.

[earthlight]

Almost all models suffer for precipitation bombs at times due to the mismatch between the model grid size and the scale that convection occur on. The reason convective schemes are in models is to keep the grid scale instabilities from mucking up the model too badly. In certain patterns the nam seems to suffer worse but there have been times when the gfs also goes bonkers, not as often as it used to. You sometimes see that in handling bogus tropical systems in the longer time ranges though the models do not seem to forecast as many bogus tropical systems as they used to.

I've definitely seen it on almost all guidance at one point. Generally, I've noticed the NAM seems to be more prone to the qpf bombs, although it may be a bit of an illusion in a sense. It would be interesting to see some data or statistics as to which model suffers the worst from the qpf bombs/blobs.