March 2023 Obs/Disco

[WinterWolf]

2 minutes ago, 40/70 Benchmark said:

He certainly seemed to be correctly on the mild train, but its difficult to really assess when there was never a formal outlook published.

This!
He also called for no changes, and there drastic changes a few times..:most weren’t able to cash in, but that doesn’t mean there weren’t changes as many called for and supported. 

[jbenedet]

1 minute ago, 40/70 Benchmark said:

It was clear to me this month wasn't going to be as cold as March 2018 because the PV is on the other side of the globe again, but I did think that we'd do better with respect to snow than we have. Just couldn't catch a break.

You were spot on about how active this winter was. I'll def give you that.  Nice call in that regard.

Marginal boundary layer temps prob cost me 30" in white rain this year. 

 

Snowfall accums are a bitch to forecast; vast majority don't appreciate how right you can be, and still be wrong. It's extremely stressful for that reason.

 

[40/70 Benchmark]

Just now, jbenedet said:

You were spot on about how active this winter was. I'll def give you that.  Nice call in that regard.

Marginal boundary layer temps prob cost me 30" in white rain this year. 

 

Snowfall accums are a bitch to forecast; vast majority don't appreciate how right you can be, and still be wrong. It's extremely stressful for that reason.

 

100% this.

Nice job with pattern recognition this year.

[WinterWolf]

11 minutes ago, 40/70 Benchmark said:

Well, I don't think the fact that winter was very mild was just bad luck....but do think the fable about November warmth holding up was luck.

No, I didn’t mean that, I meant that because November was warm, the winter was warm;  that’s the luck part.  Sometimes November is frigid…that doesn’t mean the winter will be frigid. Sometimes it can be, sometimes not. That’s what I meant. 

[40/70 Benchmark]

TBH, I think January being +8 is what sunk my forecast...you replace that with even pedestrian warmth, and most of region hits my forecast seasonal amounts.

[jbenedet]

6 minutes ago, 40/70 Benchmark said:

He certainly seemed to be correctly on the mild train, but its difficult to really assess when there was never a formal outlook published.

I don't do this anymore. The pay-off is not worth the risk. And by risk, I mean opportunity cost. I have learned; mostly the hard way.

Now I monitor the weather and try to improve my forecasting abilities to enhance the productivity of my outdoor hobbies; mostly gardening. Much better balance of risk/reward.

[40/70 Benchmark]

3 minutes ago, jbenedet said:

I don't do this anymore. The pay-off is not worth the risk. And by risk, I mean opportunity cost. I have learned; mostly the hard way.

Now I monitor the weather and try to improve my forecasting abilities to enhance the productivity of my outdoor hobbies; mostly gardening. Much better balance of risk/reward.

Totally get it. Wasn't meant as a criticism at all.

[WinterWolf]

5 minutes ago, jbenedet said:

I don't do this anymore. The pay-off is not worth the risk. And by risk, I mean opportunity cost. I have learned; mostly the hard way.

Now I monitor the weather and try to improve my forecasting abilities to enhance the productivity of my outdoor hobbies; mostly gardening. Much better balance of risk/reward.

Understood.  

[40/70 Benchmark]

19 minutes ago, Great Snow 1717 said:

Let me rephrase...The Pope had the BEST take on what was in store for winter from November on...

BTW, this was @raindancewxIMO....go back read his. Nailed it.

[Great Snow 1717]

16 minutes ago, WinterWolf said:

Exactly.  The crows, the squirrels, the grass…the acorns etc etc…its total and complete Bull Shit period. November  warmth or cold had nothing to do with winters outcome.  And if he believes that it wasn’t just blind luck, he’s just plain wrong.  

Debate it with Eric fisher...

https://www.cbsnews.com/boston/news/warm-november-lackluster-winter/

 

[jbenedet]

Also, I've learned in the art of forecasting, it's really much less about seeing into the future, and instead fully (and accurately) understanding the present. Models can hinder forecasting ability, if misused. We all want to see an image of the future. It's eye-candy and easily abused. Addictive. 

If you get the best picture (wholistic) view of the present, you can then better predict the future. To KNOW the present is to see the future. 

I've become a zealot for Lorenz and chaos theory. If you are after accurate forecasts: Know the initial conditions as best as you can; forget most everything else. 

[weatherwiz]

What makes this stuff (long-range) so difficult is, sometimes it's not about the pattern, it's about the processes which occur for the pattern to develop and the deviations within the pattern. Sometimes I think we get too caught up in looking at just a static ENS mean. We see a deep trough in the East on the D6-10 or D10-14 and assume that equates to "good potential", but does it really? 

[ORH_wxman]

November does have a small correlation to winter temps. But it's certainly not strong enough to use as a primary variable for forecasting winter temps.

 

[Great Snow 1717]

Just now, 40/70 Benchmark said:

BTW, this was @raindancewxIMO....go back read his. Nailed it.

Two factors work in his favor:

1) He doesn't over analysis things.

2) He doesn't live in SNE which means he doesn't go searching for an outcome that he wants. 

[40/70 Benchmark]

2 minutes ago, jbenedet said:

Also, I've learned in the art of forecasting, it's really much less about seeing into the future, and instead fully (and accurately) understanding the present. Models can hinder forecasting ability, if misused. We all want to see an image of the future. It's eye-candy and easily abused. Addictive. 

If you get the best picture (wholistic) view of the present, you can then better predict the future. The KNOW the present is to see the future. 

I've become a zealot for Lorenz and chaos theory. If you are after accurate forecasts: Know the initial conditions as best as you can; forget most everything else. 

I think I apply that in seasonal forecasting.....its about assessing current conditions, ad nauseum...but I know what you mean haha

[40/70 Benchmark]

1 minute ago, Great Snow 1717 said:

Two factors work in his favor:

1) He doesn't over analysis things.

2) He doesn't live in SNE which means he doesn't go searching for an outcome that he wants. 

I think I am guilty of the first, but not the second.

[40/70 Benchmark]

2 minutes ago, ORH_wxman said:

November does have a small correlation to winter temps. But it's certainly not strong enough to use as a primary variable for forecasting winter temps.

 

Yep. I think its also stronger if October is cool and November flips.

[Great Snow 1717]

6 minutes ago, jbenedet said:

Also, I've learned in the art of forecasting, it's really much less about seeing into the future, and instead fully (and accurately) understanding the present. Models can hinder forecasting ability, if misused. We all want to see an image of the future. It's eye-candy and easily abused. Addictive. 

If you get the best picture (wholistic) view of the present, you can then better predict the future. To KNOW the present is to see the future. 

I've become a zealot for Lorenz and chaos theory. If you are after accurate forecasts: Know the initial conditions as best as you can; forget most everything else. 

The old school mets back in the day often mentioned the bolded. They were far more reliant on pattern recognition than the models.  Same applies to scouting in sports, far too many scouts rely on projection than what a player is. 

[jbenedet]

1 minute ago, Great Snow 1717 said:

The old school mets back in the day often mentioned the bolded. The were far more reliant on pattern recognition than the models.  Same applies to scouting in sports, far too many scouts rely on projection than what a player is. 

Yes and it's also why experts, generally are terrible forecasters. They tunnel. They are a great resource but they don't understand (or value) the present outside of their very narrow focus. 

[weatherwiz]

Just now, Great Snow 1717 said:

The old school mets back in the day often mentioned the bolded. The were far more reliant on pattern recognition than the models.  Same applies to scouting in sports, far too many scouts rely on projection than what a player is. 

There are just so many products and data now and it becomes overload...not too mention laziness. This isn't applied to anyone in here, but it seems alot of forecasts out there are just based off the assessment of a few products. 

I mean take severe weather for example (I'll lay off the snow maps for once). I know this winter has actually been above-average for severe weather and tornadoes, but there have been a ton of setups which have been way too hyped. There's been several events all mentioning "potential for strong tornadoes" and some of those events barely even produced. Everyone sees cute hodographs, pretty colors on the Significant Tornado Parameter/Supercell Composite parameter charts, and you'd think there were going to be 50 EF5's

[Great Snow 1717]

Just now, jbenedet said:

Yes and it's also why experts, generally are terrible forecasters. They tunnel. They are a great resource but they don't understand (or value) the present outside of their very narrow focus. 

I see the same in sports.

[ORH_wxman]

Just now, 40/70 Benchmark said:

Yep. I think its also stronger if October is cool and November flips.

Yes, you want multiple signals compounded in the same direction which increase the correlation. Oct has a slight inverse correlation, while Nov has a slight positive correlation, but add them both together and it becomes stronger likelihood.

I'll say though that the Oct correlation has weakened in recent years, but it's still there.

 

The biggest conundrum with seasonal forecasting is trying to fit pattern projection onto sensible wx. Nobody who was ever given a composite of December's H5 pattern would predict an above average month on temps and virtually snowless for SNE. The small nuances in that longwave pattern are not predictable at more than a few days lead time so you'll never be able to accurately forecast that type of blocking while also adding the caveat "but it won't produce cold/snowy sensible wx for us".

[Great Snow 1717]

Just now, weatherwiz said:

There are just so many products and data now and it becomes overload...not too mention laziness. This isn't applied to anyone in here, but it seems alot of forecasts out there are just based off the assessment of a few products. 

I mean take severe weather for example (I'll lay off the snow maps for once). I know this winter has actually been above-average for severe weather and tornadoes, but there have been a ton of setups which have been way too hyped. There's been several events all mentioning "potential for strong tornadoes" and some of those events barely even produced. Everyone sees cute hodographs, pretty colors on the Significant Tornado Parameter/Supercell Composite parameter charts, and you'd think there were going to be 50 EF5's

Agree, far too many mets do not actually forecast weather, they just show model output. 

[ORH_wxman]

1 minute ago, weatherwiz said:

There are just so many products and data now and it becomes overload...not too mention laziness. This isn't applied to anyone in here, but it seems alot of forecasts out there are just based off the assessment of a few products. 

I mean take severe weather for example (I'll lay off the snow maps for once). I know this winter has actually been above-average for severe weather and tornadoes, but there have been a ton of setups which have been way too hyped. There's been several events all mentioning "potential for strong tornadoes" and some of those events barely even produced. Everyone sees cute hodographs, pretty colors on the Significant Tornado Parameter/Supercell Composite parameter charts, and you'd think there were going to be 50 EF5's

We should be careful in our analysis of "too much data" though....

 

I agree there's often data overload, but lets not sugarcoat the prior era of forecasting....it absolutely SUCKED compared to now. Even the best mets back then couldn't forecast 5 days out even close to the same accuracy we have now. The computer models that all of us mets rely on just were not good enough and no human can overcome that lack of computing power.

[weatherwiz]

2 minutes ago, ORH_wxman said:

We should be careful in our analysis of "too much data" though....

 

I agree there's often data overload, but lets not sugarcoat the prior era of forecasting....it absolutely SUCKED compared to now. Even the best mets back then couldn't forecast 5 days out even close to the same accuracy we have now. The computer models that all of us mets rely on just were not good enough and no human can overcome that lack of computing power.

Great point. Forecasting has come a significant way...even in the last 10-15 years. We can have a solid idea now of storm potential 7-10 days out now which was unheard of before. Take even this winter...while it didn't work out, there were some storms that were signaled well in advance. Look at California too...the stretches of active weather they had were signaled well over a week out

[40/70 Benchmark]

11 minutes ago, ORH_wxman said:

Yes, you want multiple signals compounded in the same direction which increase the correlation. Oct has a slight inverse correlation, while Nov has a slight positive correlation, but add them both together and it becomes stronger likelihood.

I'll say though that the Oct correlation has weakened in recent years, but it's still there.

 

The biggest conundrum with seasonal forecasting is trying to fit pattern projection onto sensible wx. Nobody who was ever given a composite of December's H5 pattern would predict an above average month on temps and virtually snowless for SNE. The small nuances in that longwave pattern are not predictable at more than a few days lead time so you'll never be able to accurately forecast that type of blocking while also adding the caveat "but it won't produce cold/snowy sensible wx for us".

This is what have started to incorporate this past fall.

[BrianW]

1 hour ago, weatherwiz said:

I just want to see trees start budding and get 60's for now. Then have months on end of 80's 90's with humidity. 

You need to move back to Branford. Plenty of trees budding, flowers out and lawns are nice and green...

 

 

[powderfreak]

37 minutes ago, 40/70 Benchmark said:

TBH, I think January being +8 is what sunk my forecast...you replace that with even pedestrian warmth, and most of region hits my forecast seasonal amounts.

Yeah and that’s shown by it being so warm +7 to +10 just north of you in January but with above normal snow during that time.  If that’s like +3 or +4 maybe that big CNE snow zone dips to the Pike instead of Lakes Region.

[Snowcrazed71]

TT is already calling next winter off lol 

He posted the Noah maps for the temperature outlook for December, January, February. As if that'll stick right now. I'm just ready for the warm weather for now. Until next year. Enjoy the summer you guys, let's have a nice big tropical storm come up here. See you later

 

 

 

[jbenedet]

31 minutes ago, ORH_wxman said:

We should be careful in our analysis of "too much data" though....

 

I agree there's often data overload, but lets not sugarcoat the prior era of forecasting....it absolutely SUCKED compared to now. Even the best mets back then couldn't forecast 5 days out even close to the same accuracy we have now. The computer models that all of us mets rely on just were not good enough and no human can overcome that lack of computing power.

We've seen a great improvement thanks to technological enhancements and fine-tuning the theory. Not even debatable. My basis is we're stagnating here, and it's not because of the former; but an over-reliance on it. And some take this further with an arrogance to ignore everything else, "we know everything, that's relevant to weather forecasting". No, we don't. 

In mathematical terms, we've hit a forecasting asymptote. More effort of staring at models is not creating more accurate forecasts.