November Medium/Long Range Discussion

[WEATHER53]

22 hours ago, WxUSAF said:

Pattern changes are often rushed on guidance. That’s just typical? Especially -EPO driven changes. I have been posting (along with lots of better Mets) that guidance has been overdoing mid-long range warmth for us for months now. And I personally don’t want our peak pattern on December 1. Our average high is like 52 still! I’m vey good with Thanksgiving through December 10-15 being a transitional period that builds up cold air and snowcover in our source regions. 

If pattern changes are almost built in rushed in guidance then why not develop a system that does not do that?

Over last couple winters we’ve seen gigantic failure anywhere from 36 hours to 15 days out.   Very few if any positive such surprises .  

 

[Maestrobjwa]

3 hours ago, WEATHER53 said:

If pattern changes are almost built in rushed in guidance then why not develop a system that does not do that?

Over last couple winters we’ve seen gigantic failure anywhere from 36 hours to 15 days out.   Very few if any positive such surprises .  

 

I think you're expecting things that the computers are simply not advanced enough for those things not to happen sometimes. Why not just...accept the current limitations/biases and set forecasts/expectations from there instead of railing against the failures as if it's something easy to fix or some conspiracy? Technology simply isn't there yet. NWP (if I'm using that term correctly) is leaps and bounds better than it was 25-30 years ago but it still isn't a crystal ball! (and someone can correct me if I'm off base on any of that).

[eduggs]

1 hour ago, WEATHER53 said:

If pattern changes are almost built in rushed in guidance then why not develop a system that does not do that?

Over last couple winters we’ve seen gigantic failure anywhere from 36 hours to 15 days out.   Very few if any positive such surprises .  

 

People just expect too much accuracy from model guidance, especially past day 10. The model doesn't know it's showing a "pattern change." That's just human interpretation, which is also subjective and biased. When models are inevitably less accurate in the long-range than we wish, particularly after we've identified a perceived regional "pattern change," we feel like the model failed or delayed it. But the failing is our false interpretation and unrealistic expectation.

We really should stop anthropomorphizing weather models.

[Interstate]

42 minutes ago, eduggs said:

People just expect too much accuracy from model guidance, especially past day 10. The model doesn't know it's showing a "pattern change." That's just human interpretation, which is also subjective and biased. When models are inevitably less accurate in the long-range than we wish, particularly after we've identified a perceived regional "pattern change," we feel like the model failed or delayed it. But the failing is our false interpretation and unrealistic expectation.

We really should stop anthropomorphizing weather models.

You mean 3 days 

[Weather Will]

WB latest 12Z EPS: temps average near normal/ slightly above through Dec 5. Precipitation looks like it will pick up toward the end of the period but nothing extreme. Looks like any big storms will track NW of the MA during this time frame with the NAO slightly positive and the PNA slightly negative.

[Weather Will]

WB latest EPS weeklies centers the coldest period in December around the holidays now with normal precip. Maybe we will be able to do more than dream about a White Christmas this year.  Definitely does not look like a torch right now around the holidays.

[Eskimo Joe]

The fly in the ointment right now for the holidays is a lack of a -NAO. It's tough to get a big snowstorm without that feature to hold a storm closer to the coast. I'm not unhappy by any stretch of the imagination though...it beats Pacific Puke with 580dm heights across the entire CONUS!

[TSSN+]

9 minutes ago, Eskimo Joe said:

The fly in the ointment right now for the holidays is a lack of a -NAO. It's tough to get a big snowstorm without that feature to hold a storm closer to the coast. I'm not unhappy by any stretch of the imagination though...it beats Pacific Puke with 580dm heights across the entire CONUS!

Till we actually don’t crap the bed, I expect us to crap the bed lol 

[Eskimo Joe]

15 minutes ago, TSSN+ said:

Till we actually don’t crap the bed, I expect us to crap the bed lol 

My goals before Christmas:

• Outstanding: multiple snow events, whether all advisory/warning/combination of these
• Great: One warning level snowfall event.
• Good: One advisory level snowfall event
• Bad: No snow
• Shut the Blinds: patterns fails and we get back to zonal puke

[TSSN+]

17 minutes ago, Eskimo Joe said:

My goals before Christmas:

• Outstanding: multiple snow events, whether all advisory/warning/combination of these
• Great: One warning level snowfall event.
• Good: One advisory level snowfall event
• Bad: No snow
• Shut the Blinds: patterns fails and we get back to zonal puke

Bottom 3 most likely 

[TSSN+]

1 minute ago, TSSN+ said:

Bottom 3 most likely 

I used to be hopeful but after the last 9 years I can say unlike Justin Berk, I do not have faith in the flakes haha. 

[WEATHER53]

1 hour ago, eduggs said:

People just expect too much accuracy from model guidance, especially past day 10. The model doesn't know it's showing a "pattern change." That's just human interpretation, which is also subjective and biased. When models are inevitably less accurate in the long-range than we wish, particularly after we've identified a perceived regional "pattern change," we feel like the model failed or delayed it. But the failing is our false interpretation and unrealistic expectation.

We really should stop anthropomorphizing weather models.

I would say both posts suggesting  I cool my jets are mostly just excuse making about the accuracy and statements that we should just accept it and don’t complain. 
Models  cover all the bases.  Thats how they are scored for funding. So the status quo is content. I see no serious improvement over the last 20 years. I believe they try to do too much,  like using a microscope for very close up inspection rather than a more  backed off binoculars approach.  Between now and Christmas Day we will see everything from 30 and snow to 60 and sun for Christmas Day. That’s  not science, that’s cover all bases guesswork 

[eduggs]

Weather models have improved a lot in 20 years. We have data to prove it.

[stormy]

3 hours ago, eduggs said:

People just expect too much accuracy from model guidance, especially past day 10. The model doesn't know it's showing a "pattern change." That's just human interpretation, which is also subjective and biased. When models are inevitably less accurate in the long-range than we wish, particularly after we've identified a perceived regional "pattern change," we feel like the model failed or delayed it. But the failing is our false interpretation and unrealistic expectation.

We really should stop anthropomorphizing weather models.

How about models being off 10 degrees for thermal highs in 12 hours. Is that acceptable???

[Kevin Reilly]

2 hours ago, Eskimo Joe said:

The fly in the ointment right now for the holidays is a lack of a -NAO. It's tough to get a big snowstorm without that feature to hold a storm closer to the coast. I'm not unhappy by any stretch of the imagination though...it beats Pacific Puke with 580dm heights across the entire CONUS!

Well the -NAO locks in the cold and this has been the biggest issue with the Pacific Jet ripping across the country and a flat high south and southeast.  I see more of the same from last year with hints here and there that it could possibly change but zero is locking in. 

[wxdude64]

13 minutes ago, stormy said:

How about models being off 10 degrees for thermal highs in 12 hours. Is that acceptable???

Yep, on the 18th NWS called for a 55 high here. Stuck in the upper 30's all day, high was 41.5 at midnight and then locked between 40-41 overnight until 7 am. 

[WxUSAF]

 

[Eskimo Joe]

18z GFS/GEFS shows a legitimate fail mode for us, and one that I'm always concerned about happening. The trough dumps out west, flexes the southeast ridge, and we just torch while the central and western US score. Hope this is just a fluke.

[bncho]

14 minutes ago, Eskimo Joe said:

18z GFS/GEFS shows a legitimate fail mode for us, and one that I'm always concerned about happening. The trough dumps out west, flexes the southeast ridge, and we just torch while the central and western US score. Hope this is just a fluke.

GFS maybe, GEFS not so much. SE ridge only really flexes around December 2-4 before the cold air starts pushing east. Definitely something to watch but I'm not too concerned right now.

[Ji]

GFS maybe, GEFS not so much. SE ridge only really flexes around December 2-4 before the cold air starts pushing east. Definitely something to watch but I'm not too concerned right now.

You guys need to relax. Look at euro ensembles by day 15. Even a cold Dec 1-5 is still probably cold rain. Just be patient

[WxUSAF]

22 minutes ago, Ji said:

You guys need to relax. Look at euro ensembles by day 15. Even a cold Dec 1-5 is still probably cold rain. Just be patient

My man, the sage voice of reason. 

[wxdude64]

1 hour ago, Ji said:

You guys need to relax. Look at euro ensembles by day 15. Even a cold Dec 1-5 is still probably cold rain. Just be patient

Who hacked this account?

[frd]

 

Eric Webb @webberweather   

 

It’s easy to spot Convectively Coupled Kelvin waves (blue dashed lines) propagating quickly thru the mjo envelope (black dashed line) the next few weeks on the forecast VP200a Hovmöller (left) This is largely why you’re seeing these loop-de-loops on the RMM phase plots (right)

 

 

 

 

 

[stormy]

18 hours ago, wxdude64 said:

Yep, on the 18th NWS called for a 55 high here. Stuck in the upper 30's all day, high was 41.5 at midnight and then locked between 40-41 overnight until 7 am. 

Thank You!  I could have had similar input for my location but you beat me to it.

My predicted high for the 18th 24 hrs prior at 6 am 11-17 for my newsletter was 43.  I missed it by 2 degrees 24 hrs prior. The NWS at Sterling missed it by 9 degrees, 12 hrs. prior and in real time because of buying in to  model error. The models almost always scour the cold from entrenched CAD  out of valleys too quickly on a southeast or east breeze for Augusta.

The Sterling forecaster obviously didn't understand or consider this on Tuesday even after midday red flags.

[WEATHER53]

19 hours ago, eduggs said:

Weather models have improved a lot in 20 years. We have data to prove it.

And many have direct personal experiences to Disprove it.  A lot here were 8 years old 20 years ago and some of us already adults and already familiar with model performance 

[eduggs]

2 minutes ago, WEATHER53 said:

And many have direct personal experiences to Disprove it.  Many here were 8 years old 20 years ago and some already adults and already familiar with model performance 

You cannot disprove a theory with feelings. Human quality-of-life has improved tremendously over the past few centuries (medicine [germ theory], electricity, water treatment, agriculture etc...) because (some) humans used science (data, evidence, repeatability) to solve problems instead of emotion, superstition, gut feeling etc... Human perceptions are biased and must be validated with observable data to draw reliable conclusions. Fortunately people like you (non-fact-based) do not steer social and technological policies and developments. You undoubtedly have other ways to contribute to society, but assessing the state of meteorological modeling isn't one of them.

[WesternFringe]

Anomaly Correlation Coefficient (ACC) has increased by about 20% over the last 20 years for 5 day lead times and nearly 30% for 7 day lead times.   It has increased about 60% for 10 day leads!!  Root Mean Square Error (RMSE) is down by 23% at 5 day leads.   These numbers are all from the GFS and ECMWF.

[Eskimo Joe]

Am I missing something, or do the 12z Euro/ENS and GEFS just dump all the cold air out west and just torch the east coast.

[Weather Will]

8 minutes ago, Eskimo Joe said:

Am I missing something, or do the 12z Euro/ENS and GEFS just dump all the cold air out west and just torch the east coast.

Correct.  Looks like a chilly Thanksgiving weekend with warmth on either side.

[bncho]

9 minutes ago, Eskimo Joe said:

Am I missing something, or do the 12z Euro/ENS and GEFS just dump all the cold air out west and just torch the east coast.

let it torch in early-december, IDC too much. just let it be cold mid-late december