December 2022

[EastonSN+]

Just now, Great Snow 1717 said:

Agree. The is far too much reliance on the models as opposed to pattern recognition.

Agree on pattern recognition focus.

However, this horrific performance by the ensembles is just outright disappointing and pathetic. They are supposed to be a tool for those in the meteorology profession. Now I just look at them past day 7 the same way I look at weeklies - entertainment.

Maybe I am overreacting but jeez.

[EastonSN+]

Ha. I guess one good thing about not trusting the ensembles past day 7 is I can ignore the new bad look

[Great Snow 1717]

19 minutes ago, EastonSN+ said:

Agree on pattern recognition focus.

However, this horrific performance by the ensembles is just outright disappointing and pathetic. They are supposed to be a tool for those in the meteorology profession. Now I just look at them past day 7 the same way I look at weeklies - entertainment.

Maybe I am overreacting but jeez.

I do not think you are overreacting at all.  The so-called upgrades to the models haven't worked all that well. Perhaps it is time to re-engineer" the models back to their previous state(s). I've known a few people who wrote software. All said doing upgrades is difficult.  If given the choice all would have elected to just write new software rather than to attempt to upgrade existing software.
 

[bluewave]

3 minutes ago, WX-PA said:

Well really past day 10 on any ensemble, OP is useless..And then  CFS put's monthly's and seasonal out which is a joke. How can any model predict the weather 3-6 months ahead. Basically it's for entertainment use, but some forecasters take it seriously 

The EPS is a little better at detecting the ridge axis out near 60-65W than the GEFS close to day 10. At least the EPS didn’t have the 50/50 low there like the GEFS which used to be common under these set ups in the past. I would apply machine learning bias correction to those maps from 192-360. Maybe that will be the next step at ECMWF now that that they have their new super computer facilities. 
 

New run

 

Old runs

[LibertyBell]

6 minutes ago, donsutherland1 said:

FYI:

https://www.ecmwf.int/sites/default/files/elibrary/2021/19877-machine-learning-ecmwf-roadmap-next-10-years.pdf

Our tax money should be going towards doing this for our models too (not to mention 4DVAR.)  Not more GOES satellites!

 

[Great Snow 1717]

6 minutes ago, donsutherland1 said:

FYI:

https://www.ecmwf.int/sites/default/files/elibrary/2021/19877-machine-learning-ecmwf-roadmap-next-10-years.pdf

The strongest chess engine in the world is Alpha Zero. The "program" was never "taught" how to play chess and understand chess. It learned how to play incredibly strong chess by playing millions of games against itself. All AI machine based learning

[EastonSN+]

Lol check out the one ensemble member that goes off the chart in 6.

[donsutherland1]

3 minutes ago, Great Snow 1717 said:

The strongest chess engine in the world is Alpha Zero. The "program" was never "taught" how to play chess and understand chess. It learned how to play incredibly strong chess by playing millions of games against itself. All AI machine based learning

It did. It even learned how to make sacrifices.

[EastonSN+]

WRT this upcoming storm, still holding out a LITTLE hope that we can get a coating to 2 inches before the changeover. Long shot but fun to track.

Season snowfall to date - 1.5.

[Great Snow 1717]

5 minutes ago, donsutherland1 said:

It did. It even learned how to make sacrifices.

Alpha Zero plays such remarkable high level chess that even grandmasters admit that humans would not be able to "find" some of Alpha Zero's moves.  

[North and West]

It did. It even learned how to make sacrifices.

.

[SnowGoose69]

17 minutes ago, bluewave said:

The EPS is a little better at detecting the ridge axis out near 60-65W than the GEFS close to day 10. At least the EPS didn’t have the 50/50 low there like the GEFS which used to be common under these set ups in the past. I would apply machine learning bias correction to those maps from 192-360. Maybe that will be the next step at ECMWF now that that they have their new super computer facilities. 
 

New run

 

Old runs

Really the GEFS only was correct about the pattern switch really happening closer to 12/15 vs 12/10, after that it was sort of owned by the GEPS/EPS,  I posted in the SNE thread I'd feel better about January right now as the GEPS/EPS show a +PNA by D14-16 while the GEFS wants to go -PNA...the GEFS failed miserably about 10-12 days ago saying the 12/20-12/25 period would be a -PNA while the EPS/GEPS had it positive 

[wdrag]

1 hour ago, EastonSN+ said:

I am in complete agreement. I posted a great post from CoastalWX in the NE forum where he pointed out the bad luck and flow nuances that hurt this opportunity. 

We have been extremely lucky 00/01 onwards as compared to the 90s where, other than 2 winters, were unlucky. Perhaps it's just evening out to average.

I personally do not feel that global warming has suddenly made it impossible to get a big snowstorm again with a -4AO. If the ridge out west was a bit further east the warmth, ridge to our east would have been pushed a little east and we would have had a good one. 

 

I haven't given up on the two hazards (front and back end) but had to give it up for the coastal cities.  Complex and the colder solutions are not yet kaput.  Already 12z/18 NAM has positive snow depth into central VA.  Could it be wrong??? yes, but I do think ice and or snow will develop in w NC and nw 1/2 VA at the very start.  Can I hope for the 12Z/18 GFS to give me a little more to grasp onto? 

[eduggs]

48 minutes ago, Great Snow 1717 said:

Agree. There is far too much reliance on the models as opposed to pattern recognition.

Disagree. 1) Our ability to recognize a future "pattern" is completely dependent on models. 2) "Patterns" only exist in a broad sense. They are not intrinsic nor defined to a fine scale. They are just descriptive and statistical human inventions. Our feeble brains try to simplify things into "patterns" for better understanding. But in truth, the weather, characterized by height fields, temperature, pressure, and all the other parameters is unique. Relying on general pattern recognition to very generally predict synoptic scale weather probabilities might be plausible. But anticipating specific weather in a local region based on "patterns" is fruitless.

[Great Snow 1717]

2 minutes ago, eduggs said:

Disagree. 1) Our ability to recognize future "pattern" is completely dependent on models. 2) "Patterns" only exist in a broad sense. They are not intrinsic nor defined to a fine scale. They are just descriptive and statistical human inventions. Our feeble brains try to simplify things into "patterns" for better understanding. But in truth, the weather, characterized by height fields, temperature, pressure, and all the other parameters is unique. Relying on general pattern recognition to very generally predict synoptic scale weather probabilities might be plausible. But anticipating specific weather in a local region based on "patterns" is fruitless.

Incorrect, back in the day meteorologists focused far more on pattern recognition than the models. Leading into the winter season it was common for on air(TV and radio) mets to mention the weather patterns of the fall leading into the winter months. A key for them was where the main storm track(s) had set up since late summer.

Also we know that certain patterns are more likely to be associated with different types of weather. For example a pattern that features a trough in the west usually isn't all that good for winter weather prospects in the east

 

[eduggs]

19 minutes ago, wdrag said:

I haven't given up on the two hazards (front and back end) but had to give it up for the coastal cities.  Complex and the colder solutions are not yet kaput.  Already 12z/18 NAM has positive snow depth into central VA.  Could it be wrong??? yes, but I do think ice and or snow will develop in w NC and nw 1/2 VA at the very start.  Can I hope for the 12Z/18 GFS to give me a little more to grasp onto? 

The ULL/PV over Hudson Bay with soaring 500mb heights to Northern Quebec and a strong shortwave diving south through the Pac NW just looks so ugly. I guess we can hope that the weak preceding coastal precipitation scrapes us before the cutter, but it feels like a depressing consolation.

[Great Snow 1717]

22 minutes ago, wdrag said:

I haven't given up on the two hazards (front and back end) but had to give it up for the coastal cities.  Complex and the colder solutions are not yet kaput.  Already 12z/18 NAM has positive snow depth into central VA.  Could it be wrong??? yes, but I do think ice and or snow will develop in w NC and nw 1/2 VA at the very start.  Can I hope for the 12Z/18 GFS to give me a little more to grasp onto? 

Back in the day the setup for late Thursday/Friday would often lead to a secondary low developing along the eastern seaboard.  That seems to be happening with less frequency than in the past. Can you offer up some reasons as to why that is happening less often?

Thank you.

[eduggs]

6 minutes ago, Great Snow 1717 said:

Incorrect, back in the day meteorologists focused far more on pattern recognition than the models. Leading into the winter season it was common for on air(TV and radio) mets to mention the weather patterns of the fall leading into the winter months. A key for them was where the main storm track(s) had set up since late summer.

Also we know that certain patterns are more likely to be associated with different types of weather. For example a pattern that features a trough in the west usually isn't all that good for winter weather prospects in the east

 

All of that is very non scientific. The use of the word "pattern" should be banned from the lexicon. 

[Tatamy]

1 hour ago, WX-PA said:

Nobody was talking about global warming from 2002-2014..when every other year was a blockbuster winter for snow. The coastal sections of the mid atlantic, NYC are not a snowy climate lets face it. Most of the 70's, 80's and 90's it hardly snowed here, especially in December. I think it's Mother Nature's way of balancing things out to normal.Man it took 10 years to get a snowfall over a foot in NYC from 83-93. And 9 years from 69-78

This is a great post.  As someone who came of age in the 70s and 80s it amazes me how people and weenies in particular have come to expect major winter events each year.  As you stated during the period from 69-93 there only three storms that produced over a foot in NYC.  Back in those days storms that were forecast to produce 4-8” were a big deal and if you had one of those in a winter that was a lot.  This region has a long term climate history and it is not one where big events (MECS/SECS/HECS) are common.  White Christmases are the exception, not the rule here.  We have had a good run in recent years with these big events however long term climatology dictates that this not likely to continue.

[HailMan06]

5 hours ago, LibertyBell said:

For me it's the strong feeling that we've experienced the most extreme weather we are ever likely to experience in our lifetimes already, so anything we get from now on is just icing on the cake.  Back in the 80s I would throw a fit if a storm busted, but after all we've seen in the past 20 years, it's all fine by me lol.  No snow season will top 95-96 and no snowstorm will top January 2016 and we (hopefully) won't get anything like Hurricane Sandy ever again and the raging extreme heat we had in 2010 and 2011 will likely never again be experienced by us in our lifetimes either.

 

 

The #1 rule in weather is never talk definitively. Lots of things are certainly possible that we haven’t seen yet.

[Dark Star]

58 minutes ago, LibertyBell said:

I think with climate change we might also be seeing a feedback effect, so with time the changes become more and more apparent.

So an analogy might be like....if steroids weren't banned, players would keep hitting more and more HR.....so McGwire with 70 and then Bonds with 73 and then a few years later someone else would hit 80, then 85, etc.

There is a physical limit to the amount of home runs, as well as pitchers giving intentional passes (even with ADHD medication, which Bonds was also taking)...

[Great Snow 1717]

2 minutes ago, eduggs said:

All of that is very non scientific. The use of the word "pattern" should be banned from the lexicon. 

"Pattern" is one of the most use meteorological terms, it is impossible to ban it from the lexicon

[Dark Star]

5 minutes ago, eduggs said:

All of that is very non scientific. The use of the word "pattern" should be banned from the lexicon. 

Disagree.  Models are tools, the meteorologist must interpret the outcome.  

[Great Snow 1717]

Just now, Dark Star said:

Disagree.  Models are tools, the meteorologist must interpret the outcome.  

Exactly!

[gravitylover]

22 minutes ago, Dark Star said:

Disagree.  Models are tools, the meteorologist must interpret the outcome.  

Yes. Unfortunately many have become modelologists rather than meteorologists. 

[MJO812]

25 minutes ago, Dark Star said:

Disagree.  Models are tools, the meteorologist must interpret the outcome.  

Agree

[eduggs]

24 minutes ago, Dark Star said:

Disagree.  Models are tools, the meteorologist must interpret the outcome.  

Humans cannot accurately predict future weather outcomes without models. We are entirely dependent on them. Meteorologists are biased by gut instinct, wish-casting, and false pattern recognition. It's why long-term forecasting has such a low success rate. Most meteorologists are not scientists.

The constant mistake that is made is assuming that a modeled future "pattern" will come to exist. And then lots of effort is spent analyzing likely weather outcomes based on that assumed pattern. This is illogical and highly problematic for forecasting. "Patterns" aren't specific enough or stable enough to be useful in most mid-latitude regions. And we can't accurately see particular combinations of features, e.g., "patterns", coming very far out. Meteorologists should stop trying to be experts in things they don't understand.

[Allsnow]

Remember when it was supposed to be cold after Xmas….

[eduggs]

37 minutes ago, Dark Star said:

Disagree.  Models are tools, the meteorologist must interpret the outcome.  

Explain what a meteorologist interprets with respect to model output? What is the reference point for that interpretation? And what does that have to do with wishing the word "pattern" went out of use?

You might say, for example, that a model outcome doesn't fit a "pattern." But how is a future model outcome separate from a future modeled "pattern." They are both unknown and both depicted by modeling. 

 

[winterwx21]

3 minutes ago, Allsnow said:

Remember when it was supposed to be cold after Xmas….

Yep we're down to maybe a 3 day cold shot after friday's cutter. Then a very warm period.