Dec/Jan Medium/Long Range Disco

[pazzo83]

4 hours ago, NorthArlington101 said:

partially?

I've been looking to see if you can run Google's GraphCast model with snow accums but I can't find it. Have found a site that shows rough MSLP + precip

those models are pure neural network models that don't incorporate any hard-coded equations like traditional weather models, right?

[WesternFringe]

3 minutes ago, pazzo83 said:

i don't think you are quite grasping what he's saying.  all weather models are probabilistic models that factor in historical data, current observed data, and the understood mathematics that describe the atmosphere.  obviously it could be wrong because models are often wrong.  but let's look at the inputs again - unless you are arguing that one of those inputs was somehow corrupted for this particular model run, it is telling you that the most likely outcome is what is shown (with rain basically everywhere).  i think that is where the concern lies - it is outputting an outcome that is not what one would've expected in years past, meaning it is incorporating some sort of fundamental change in terms of those inputs.  I don't think the math has changed wildly (if only to improve model precision), so...

So a model that factors in historical data is outputting what would not have been expected in the past? Thanks for the insight.

Abd btw, the model isn't even showing anywhere close to what we were disagreeing about 2 runs ago at hr 192 anyways, so you are way late.

[MDScienceTeacher]

1 hour ago, fujiwara79 said:

you don't need to search too hard.  the answer is staring you in the face.  Underwater volcanos!  

Yes.  Volcanic activity (large and small scale) can lead to significant exothermic processes.  Definitely enough to have an impact on world wide climate.  I dont think people realize just how much energy it takes to skew the earths natural Glacial-Interglacial Cycles, things like volcanic activity, sun spots and astroid strikes are all capable of changing the earths weather patterns.

But honestly, we are probably just in a slump and we will wind up getting a 50 inch season sometime very soon.

[NorthArlington101]

those models are pure neural network models that don't incorporate any hard-coded equations like traditional weather models, right?

I bet someone here knows more than me, but that’s my understanding. It’s trained off historical weather data and tries to use that to predict the weather going forward. Conceptually that seems like a great application of that kind of learning to me.

[psuhoffman]

1 hour ago, Ji said:

Looks the same to me. It’s going to take a while to change our source region(Canada to cooler)

In 2010 Canada was way above normal. Quite often in a Nino Canada is torched. We survive by roughing under the NS flow (split flow) and just cold enough with a good track. Warm the whole equation and it might not work. 
 

But guidance has always said this next week would have temp issues. from 14 days ago. 

And 7

Im still disappointed even with a perfect setup we couldn’t overcome those temps on that one run. But guidance has said temps were still an issue. 

1 hour ago, TSG said:

oh come on PSU... we both know that was not a comparable setup. First, positively (Jan 3rd) vs negatively (Dec 11) tilted troughs. Second, the storm on the 11th had another piece of energy over the GL pressing that cold into the backside which gave us a couple hrs of changover, instead of the 0-15 mins we're used to with cold chasing precip. There is no such mechanism available for the storm on the 3rd, at least not as currently depicted.

On the 6z run the trough wasn’t just negative it cut off completely to our southeast!  No idea what you’re talking about. 

[pazzo83]

3 minutes ago, WesternFringe said:

So a model that factors in historical data is outputting what would not have been expected in the past? Thanks for the insight.

Abd btw, the model isn't even showing anywhere close to what we were disagreeing about 2 runs ago at hr 192 anyways, so you are way late.

again - you are missing the point.  the fact that it was a model output at any point is the concern.

[snowfan]

5 minutes ago, clskinsfan said:

Meh. It would just be white rain anyways. The following week we have some manageable temps at least. 

I believe he’s referring to those west of the Allegheny front.

[pazzo83]

2 minutes ago, NorthArlington101 said:

I bet someone here knows more than me, but that’s my understanding. It’s trained off historical weather data and tries to use that to predict the weather going forward. Conceptually that seems like a great application of that kind of learning to me.

yeah training on re-analysis is brilliant.  in this way the model can "figure out" the math on its own.

[WesternFringe]

13 minutes ago, pazzo83 said:

again - you are missing the point.  the fact that it was a model output at any point is the concern.

If every model run’s output is a concern to you, you are ridiculous and missing the point.  Hour 192 never verifies on the GFS.  But yeah, I am missing the point!? Lol

eta:  it was model output that we were discussing that never happened at 192, and hasn’t been modeled since as even close to our area at 186 or 180 in the 2 subsequent models runs, but I am missing the point? I should have been alarmed by the one run?
 

Okay, you are right.  The fact the models showed something that never happened is evidence of….what exactly?

[CAPE]

7 minutes ago, NorthArlington101 said:

Butterfly effect ends up with a substantial winter storm for GA (by their standards) up into the Carolinas for the 3-4th. Wacky run

I like the wave timing for that window with the predecessor low in a good spot and HP just west/sw of there where we want it. It's one of those simple paths to victory with a well timed discrete shortwave tracking underneath and a favorable look in the NA. Cold enough verbatim. Just need it a little further north.

[WxUSAF]

19 minutes ago, NorthArlington101 said:

Butterfly effect ends up with a substantial winter storm for GA (by their standards) up into the Carolinas for the 3-4th. Wacky run

GFS working overtime to avoid giving us digital blue after the Christmas miracle fantasy HECS.

[jayyy]

I like the wave timing for that window with the predecessor low in a good spot and HP just west/sw of there where we want it. It's one of those simple paths to victory with a well timed discrete shortwave tracking underneath and a favorable look in the NA. Cold enough verbatim. Just need it a little further north.

There are certainly worse positions to be in. Maybe it doesn’t pan out ultimately, but I’d almost rather things be where they are now. Models showing us being fringed with a low tracking to the south with a workable airmass in place… than have models show a flush hit over a week out.

[pazzo83]

14 minutes ago, WesternFringe said:

If every model run’s output is a concern to you, you are ridiculous and missing the point.  Hour 192 never verifies on the GFS.  But yeah, I am missing the point!? Lol

eta:  it was model output that we were discussing that never happened at 192, and hasn’t been modeled since as even close to our area at 186 or 180 in the 2 subsequent models runs, but I am missing the point? I should have been alarmed by the one run?
 

Okay, you are right.  The fact the models showed something that never happened is evidence of….what exactly?

just keep your head buried in the sand I guess.

[WxUSAF]

11 minutes ago, WxUSAF said:

GFS working overtime to avoid giving us digital blue after the Christmas miracle fantasy HECS.

See…shamed it into performance!

[EastTNWeatherAdmirer]

Who knows if the JMA is even worth checking out and I wish I could see the surface precip with it, but it shows a significant storm for Jan 2nd over the mid Atlantic. Probably won’t verify of course, but this seems like the most bullish model on this storm yet. 

[BristowWx]

1 minute ago, EastTNWeatherAdmirer said:

Who knows if the JMA is even worth checking out and I wish I could see the surface precip with it, but it shows a significant storm for Jan 2nd over the mid Atlantic. Probably won’t verify of course, but this seems like the most bullish model on this storm yet. 

I’ve seen the JMA score a coup hundreds of times…thousands

[nj2va]

2 minutes ago, WxUSAF said:

See…shamed it into performance!

I was about to call for a GFS upgrade to get it back into being a weenie model.  Digital blue at HH, baby!

[SnowenOutThere]

30 minutes ago, MDScienceTeacher said:

volcanic activity, sun spots and astroid strikes are all capable of changing the earths weather patterns.

I feel like we are leaving something out … must just be my imagination!  

[WesternFringe]

14 minutes ago, pazzo83 said:

just keep your head buried in the sand I guess.

If that means discounting an op run at 192 that never happened that it’s own two subsequent runs disagreed with and said would not happen, count me an ostrich!  Lol

ETA: but keep preaching to me about events that never happened.  That will prove your point.

[pazzo83]

2 minutes ago, WesternFringe said:

If that means discounting an op run at 192 that never happened that it’s own two subsequent runs disagreed with and said would not happen, count me an ostrich!  Lol

again - you are focusing on the discrete (PSA to the board - not you WF - the word is discrete, not discreet - discreet describes acting in a way so as to avoid attention) outcome relative to others in the overall model timeseries vs the fact that the model generated such an output at all - and we've seen myriad other such outputs over the past couple of years (and real outcomes that match them).

[BristowWx]

10 minutes ago, nj2va said:

I was about to call for a GFS upgrade to get it back into being a weenie model.  Digital blue at HH, baby!

Cold air draining like ground beef grease at the end of the run over Canada.  Damn.  

[WesternFringe]

19 minutes ago, pazzo83 said:

again - you are focusing on the discrete (PSA to the board - not you WF - the word is discrete, not discreet - discreet describes acting in a way so as to avoid attention) outcome relative to others in the overall model timeseries vs the fact that the model generated such an output at all - and we've seen myriad other such outputs over the past couple of years (and real outcomes that match them).

Yeah, we were all talking about a discrete threat, so you have it right about that.

The models saw rain bc the ground layers were warm in its depiction of the future.  And then we all agreed to kinda see what actually happens in 8 days instead of focusing on one model run’s output.  But I am missing the point?  Enlighten me

eta:  And it is bs if you say it is concerning when the model says this at all.  Where were you all in saying it was concerning when the models were portraying 20-30 inches a few runs ago?  Get real and accept the good with the bad. And also realize that 192 hr depictions rarely verify.

[Kevin Reilly]

31 minutes ago, EastTNWeatherAdmirer said:

Who knows if the JMA is even worth checking out and I wish I could see the surface precip with it, but it shows a significant storm for Jan 2nd over the mid Atlantic. Probably won’t verify of course, but this seems like the most bullish model on this storm yet. 

One thing it does have that all the other models have is no high to our north.

[psuhoffman]

WRT the SSW a few things I’d need answered before formulating an opinion. 
 

what does evidence say is the difference in predicted outcomes if the warming peaks just shy of a reversal and official SSW v achieving?
 

That other study linked that it has more impact if the TPV is already negative. So how much of this is cross contamination of data?  

We’ve had this debate and when the AO was hostile rarely do I remember the SSW making a significant difference. On the other hand the times it seems to have got credit like 2010 we already had a blocking regime. Dec 2009 happened before the SSW. 

[pazzo83]

11 minutes ago, WesternFringe said:

Yeah, we were all talking about a discrete threat, so you have it right about that.

The models saw rain bc the ground layers were warm in its depiction of the future.  And then we all agreed to kinda see what actually happens in 8 days instead of focusing on one model run’s output.  But I am missing the point?  Enlighten me

eta:  And it is bs if you say it is concerning when the model says this at all.  Where were you all in saying it was concerning when the models were portraying 20-30 inches a few runs ago?

yeah i mean that's what we'd expect from a slightly cold-biased model output given low position and other factors.  i think the question is, on that particular model run (and we've seen this in quite a few others over the past couple of years), why - given the low position, antecedent air mass, etc - was the boundary layer so warm?  What is having the model generate that as the highest probability outcome?  We've seen a number of such outputs - and actual events - that should have some snow that were rain for virtually everyone.  Obviously the concern here is that the base state has warmed significantly and models are starting to output scenarios - even fleeting - that reflect that.  You disagree with that premise so you try to undermine any such concern.  That's fine - but don't trivialize the concern.

[NorthArlington101]

GEFS looks slightly improved for the 2nd and 4th, FWIW. Not worth getting excited about but should reassure folks neither threat is “dead.” Also has that same window of opportunity on the 7th/8th.

Would be really funny if we did all this handwringing and somehow went on a heater.

[WEATHER53]

8 minutes ago, psuhoffman said:

WRT the SSW a few things I’d need answered before formulating an opinion. 
 

what does evidence say is the difference in predicted outcomes if the warming peaks just shy of a reversal and official SSW v achieving?
 

That other study linked that it has more impact if the TPV is already negative. So how much of this is cross contamination of data?  

We’ve had this debate and when the AO was hostile rarely do I remember the SSW making a significant difference. On the other hand the times it seems to have got credit like 2010 we already had a blocking regime. Dec 2009 happened before the SSW. 

I believe that in our righteous efforts to better predict weather, this is another sampling that’s unproven. Hell we are having trouble with long established Enso and Ao/NAO being reliable so the new indexes that seem to pop up every 3/4 years just don’t have enough outcomes yet  to be a reliable prediction package 

[WesternFringe]

11 minutes ago, pazzo83 said:

yeah i mean that's what we'd expect from a slightly cold-biased model output given low position and other factors.  i think the question is, on that particular model run (and we've seen this in quite a few others over the past couple of years), why - given the low position, antecedent air mass, etc - was the boundary layer so warm?  What is having the model generate that as the highest probability outcome?  We've seen a number of such outputs - and actual events - that should have some snow that were rain for virtually everyone.  Obviously the concern here is that the base state has warmed significantly and models are starting to output scenarios - even fleeting - that reflect that.  You disagree with that premise so you try to undermine any such concern.  That's fine - but don't trivialize the concern.

Again, people putting words in my mouth.  I don’t disagree with any premise you mentioned.  I also am not going to pretend that a 192 hr op depiction proves anything!   That is my point.  I am not talking climate change at all.  The next 2 runs of the fricking same model showed nothing like we were talking about!!!!  So why even say this shouldn’t happen when it never fucking did!!!!  Jesus Christ, that is all I am saying!!!

[eduggs]

The 18z GFS has a parade of between 6 and 10 successive shortwaves (depending on how you distinguish them) that partially interfere with each other over the next 10 days to prevent any significant local storm development. This highlights one of the problems of using LR time-blended height anomalies to try to identify favorable or active "periods." The averaged anomalies look interesting over the next week, but as usual, everything comes down to the evolution and orientation of the height fields. The actual weather could end of being quite boring depending on the fine details of wave interaction.

I prefer Walt Drag's method of threat identification mostly keeping inside of 10 days using a mid-range multi-model super-ensemble focusing on QPF and temperature distributions. To my knowledge Walt doesn't mention climate indices or height anomalies. And he doesn't frequently trigger annoyed disappointment with a lot of LR false alarms. 

[nj2va]

Can @pazzo83 and @WesternFringe take it to banter or just create a CC thread?  FFS.