Winter 2013 - 2014 Banter Thread

[Stormlover74]

this current storm might even be a bust for dc

No way they get a foot. Maybe half that

[Superstorm93]

HPC from March 2001 storm

 

Ahh...the good ole days :

 

 

[MJO812]

No way they get a foot. Maybe half that

Agree. A lot of places have a few inches.

[IsentropicLift]

I think we're done with warning criteria snowfall events for the season unless we get some high latitude blocking over the Davis Strait. The overall progressive nature of the pattern has prevented this winter from possibly becoming number one overall.

[BxEngine]

Id like some normal temps and dry weather for a few weeks...these golf courses are gonna be in bad shape.

[IsentropicLift]

Id like some normal temps and dry weather for a few weeks...these golf courses are gonna be in bad shape.

You look at the extended range of the GEFS ensembles and you see no signs of the PV beginning to retreat northward until mid March at the earliest.

[SnoSki14]

You look at the extended range of the GEFS ensembles and you see no signs of the PV beginning to retreat northward until mid March at the earliest.

Temperatures will struggle to get close to normal rather than going above. Maybe we get above normal Saturday but then it's back into the colder weather.

 

And while I'm sure people will say it feels alright/nice in the upper 30s and low 40s outside due to the sun angle or longer daylight, there's no doubt that there will still a very noticeable chill in the air especially in the shade, when the sun fades within the clouds, or when the wind picks up. It certainly won't feel very springlike. 

[GD0815]

In two weeks, in two weeks, in two weeks...that is the story of this winter lol

 

No the story of this winter has been persistent cold/snow.   

[GD0815]

You look at the extended range of the GEFS ensembles and you see no signs of the PV beginning to retreat northward until mid March at the earliest.

I think the relaxing of the PV will coincide with our opportunity for an EC storm.  

[IsentropicLift]

I think the relaxing of the PV will coincide with our opportunity for an EC storm.  

Whenever the PV does finally retreat it will take the cold air with it. If you want continued chances you'd like to see the PV remain nearby and hope for some phasing.

[pazzo83]

Whenever the PV does finally retreat it will take the cold air with it. If you want continued chances you'd like to see the PV remain nearby and hope for some phasing.

 

You'd need it timed well, the PV starting to retreat with a storm headed our way.

[IsentropicLift]

Going back to the end of last week the surface low today was forecasted to move off the southern NJ coastline. Today in actuality it's over southern NC and South Carolina. That's what you call a southward shift.

[IsentropicLift]

BTW you can see the polar vortex start to relax today and move a bit further north. Today's event or lack their of had a lot to do with poor timing of the features over Canada.

[forkyfork]

these dewpoints suck. my contact lenses almost never irritate me yet i want to claw them out today

[Rjay]

these dewpoints suck. my contact lenses almost never irritate me yet i want to claw them out today

19/-5 at 230pm...Impressive for March

[n504jb]

But it shouldn't have to be that way anymore with the rise in technology. What's the point in even issuing watches and warnings if we don't know what's to happen until the day of or the day before the storm. There is still so much more work to be done to improve these models. The GGEM caught on faster than the others in the southward shift, but it still showed a significant event for us before that. 

 

This definitely ranks up there in terms of busts for me, good or bad. However, at least there was a progression where the snow amounts starting fading away gradually instead of the day of the storm like March 2001. Oh well, I can't complain in what was a great MET winter. Things are looking up to it seems as the gfs has been trending warmer this week. I'm done with following storms, it's been a great run but it's time to move forward I think. 

 

The perceived "rise in technology". As an example, consider the advances in the diagnosis and treatment of breast cancer. Women are said to be living longer now due to advances in tomography and treatment, etc. Wrong. what has hapened is that a tumor once  discovered by self-examination or another means when it had already metastasized can now be discovered--even if not much can be done about it--years earlier because of advanced tomography. Consequently, women seem to be living longer when, in fact, they are merely being diagnosed much earlier.

 

At the same time, there are fundamental problems with ALL models, indifferant to whether they are models of cancer, 757's, the cosmos, the economy, human behavior, or meteorology. These are deletion, distortion, and generalization. Models CAN NOT include

all variables, nor do the algorhythms upon which they are based. Genralization suggests taking data and inferring from the particular to the general and the inherent problem with generalizations are they lead to false assumptions. "Deletion" implies oversimplification--omitting important, significant details to prevent overburdening the model. Third, every model can produce distortions in the observer; that is we unconsciously change what is real, into what we don't know to be real but assume is real.

 

Most importantly, since most models are based on mathematics there is the not insignificant problem that despite the belief of most people, mathematics cannot tell us what is right or wrong. This seeming paradox has baffled mathematicians for centuries. "How do you know what is right or wrong?"--mathematically speaking. Bertrand Russell tried to solve the problem but failed. So, if no mathematical system is airtight--and can't be--as Alan Turing demonstrated, how can mathematically-based models be unfailingly precise and predictive?

 

They can't be. Models of the weather are similar to alarm systems. They have become extremely sensitive--a potential hurricane can be seen in an embryonic state two weeks away off the coast of Africa; a potential blizzard 10 days away. The problem is--the unsolvable problem--the more sensitive an alarm system, the more prone it is to false alarms.

 

That is why better technology is not always the answer. New technology often means greater problems and in some cases, less predictability rather than more. It wasn't very long ago PSA tests for men over 40 seemed a no-brainer. Today, it is highly controversial

and by many physicians--including it's inventor--worthless. 

[IsentropicLift]

Forecasts even busted pretty badly in DC proper. Looks like lots of reports in the 4-6" range. More southwest of the area.

[NJHurricane]

A word about internet snow hype-casting sites....

When you follow internet snow hypecasting sites you see out of context computer model maps that make you think its a lock for two feet of snow more than 5 days away.

When you think its a lock for two feet of snow more than 5 days away you panic

When you panic you buy a pallet of rock salt from a sketchy east european web site and cancel your date with the woman of your dreams.

When you buy a pallet of rock salt from a sketchy web site and cancel a date with the woman of your dreams, your identity gets stolen by Russian mobsters who assign you a new identity as "Olaf" from Lithuainia and date a woman NOT of your dreams (See picture...she makes the rock salt look good)

And when your identity is changed and you date the woman not of your dreams you get deported to Lithuainia and live with a troll woman.

Don't get deported to Lithuainia to live with a troll woman....stop following internet snow hypecasting sites.

(Feel free to share this....spread the word, unlike, unfollow, and don't share the absurd!!)

[pazzo83]

The perceived "rise in technology". As an example, consider the advances in the diagnosis and treatment of breast cancer. Women are said to be living longer now due to advances in tomography and treatment, etc. Wrong. what has hapened is that a tumor once  discovered by self-examination or another means when it had already metastasized can now be discovered--even if not much can be done about it--years earlier because of advanced tomography. Consequently, women seem to be living longer when, in fact, they are merely being diagnosed much earlier.

 

At the same time, there are fundamental problems with ALL models, indifferant to whether they are models of cancer, 757's, the cosmos, the economy, human behavior, or meteorology. These are deletion, distortion, and generalization. Models CAN NOT include

all variables, nor do the algorhythms upon which they are based. Genralization suggests taking data and inferring from the particular to the general and the inherent problem with generalizations are they lead to false assumptions. "Deletion" implies oversimplification--omitting important, significant details to prevent overburdening the model. Third, every model can produce distortions in the observer; that is we unconsciously change what is real, into what we don't know to be real but assume is real.

 

Most importantly, since most models are based on mathematics there is the not insignificant problem that despite the belief of most people, mathematics cannot tell us what is right or wrong. This seeming paradox has baffled mathematicians for centuries. "How do you know what is right or wrong?"--mathematically speaking. Bertrand Russell tried to solve the problem but failed. So, if no mathematical system is airtight--and can't be--as Alan Turing demonstrated, how can mathematically-based models be unfailingly precise and predictive?

 

They can't be. Models of the weather are similar to alarm systems. They have become extremely sensitive--a potential hurricane can be seen in an embryonic state two weeks away off the coast of Africa; a potential blizzard 10 days away. The problem is--the unsolvable problem--the more sensitive an alarm system, the more prone it is to false alarms.

 

That is why better technology is not always the answer. New technology often means greater problems and in some cases, less predictability rather than more. It wasn't very long ago PSA tests for men over 40 seemed a no-brainer. Today, it is highly controversial

and by many physicians--including it's inventor--worthless. 

 

Uh, this isn't quite right.  The atmosphere can be modeled exactly in theory.  We simply don't have the computational power nor the ability to get all of the initial data to do such a thing at present.

[patrick05]

Uh, this isn't quite right.  The atmosphere can be modeled exactly in theory.  We simply don't have the computational power nor the ability to get all of the initial data to do such a thing at present.

 

Laplace's Demon??

[TWCCraig]

Uh, this isn't quite right.  The atmosphere can be modeled exactly in theory.  We simply don't have the computational power nor the ability to get all of the initial data to do such a thing at present.

 

Agree, I think it's more of problem of getting the correct and complete initialization data rather than the model's physics. If we wan't to predict the weather with much more accuracy, I'd focus more getting more data. We would need to sample every single particle for of air for nearly 100% accuracy and right now we do not have the tools nor the computing power to do so.

[forkyfork]

Agree, I think it's more of problem of getting the correct and complete initialization data rather than the model's physics. If we wan't to predict the weather with much more accuracy, I'd focus more getting more data. We would need to sample every single particle for of air for nearly 100% accuracy and right now we do not have the tools nor the computing power to do so.

no, that's not true. wikipedia does a good job explaining it succinctly

http://en.wikipedia.org/wiki/Numerical_weather_prediction

"Although post-processing techniques such as model output statistics (MOS) have been developed to improve the handling of errors in numerical predictions, a more fundamental problem lies in the chaotic nature of the partial differential equations used to simulate the atmosphere. It is impossible to solve these equations exactly, and small errors grow with time (doubling about every five days). In addition, the partial differential equations used in the model need to be supplemented with parameterizations for solar radiation, moist processes (clouds and precipitation), heat exchange, soil, vegetation, surface water, and the effects of terrain."

[TWCCraig]

no, that's not true. wikipedia does a good job explaining it succinctly

http://en.wikipedia.org/wiki/Numerical_weather_prediction

"Although post-processing techniques such as model output statistics (MOS) have been developed to improve the handling of errors in numerical predictions, a more fundamental problem lies in the chaotic nature of the partial differential equations used to simulate the atmosphere. It is impossible to solve these equations exactly, and small errors grow with time (doubling about every five days). In addition, the partial differential equations used in the model need to be supplemented with parameterizations for solar radiation, moist processes (clouds and precipitation), heat exchange, soil, vegetation, surface water, and the effects of terrain."

 

I never said model physics were perfect, they still need to be worked on. But I was thinking that if we'd want to achieve 100% accuracy for any forecast period, you're also going to need to sample every single particle of air, no matter how good the equations are.

 

Known as the Butterfly effect

 

 

 

In chaos theory, the butterfly effect is the sensitive dependency on initial conditions in which a small change at one place in a deterministic nonlinear system can result in large differences in a later state.

[n504jb]

Uh, this isn't quite right.  The atmosphere can be modeled exactly in theory.  We simply don't have the computational power nor the ability to get all of the initial data to do such a thing at present.

 

Uh, this isn't quite right.  The atmosphere can be modeled exactly in theory.  We simply don't have the computational power nor the ability to get all of the initial data to do such a thing at present.

"In theory" are the operative words.  And, as I suggested, the equally important liability of ALL models resides in the limitations of mathematics. Any mathematical system--and all systems based on mathematics, and the mathematics that are the underpinning of atmospheric models, must satisfy three requirements: consistency, completeness, and decidability. Consistency means you can'y have a contradiction in your system; completeness means that if any statement is true, there must be a way of proving it, and decidability means there must be some method to prove any assertion is true. Godel proved all mathematical systems were flawed insofar that they could not be both consistent and complete. He did this with the mathematical equivalent of a paradox: "This statement cannot be proved." If it can there's a contradiction in the system and it is inconsistent. If it can't be proved, the system is incomplete. Consequently, math--the "spine" of all technology and science is either inconsistent or incomplete.

 

Of course, this hasn't prevented lunar landings or fairly useful predictions of mortality with certain disease processes but in a chaotic environment like the atmosphere which is constantly changing (as opposed to the speed it takes a 757 to stall under certain flight profiles) no computer will be able to overcome the mathematical problems of modeling weather because as Alan Turing demonstrated, no one method can work for all questions. Doctors know this, as do attorneys, physicists, and engineers. The people who create these models, however, do seem to believe the same method--with slight variations in their algorithms, can work for, if not all, almost all meteorological scenarios.

 

Weather models--all models--are based on assertions. Therefore, to be able to model the atmosphere exactly, all assertions, past, present, and future, would have to be examined for their validity. "In theory", there are an infinite number of atmospheric conditions so any computer would have to have the ability to weigh in on whether an assertion about any specific complex of atmospheric conditions was valid. 

 

The fact that we are not doing very well in our "war against cancer", are routinely sucker-punched by the economy,  (despite MASSIVE computer capability), and often cannot see threats in war, would seem to suggest that the 

 problem is not developing better computers but surrendering to the fact that there will quite likely always be almost imperceptible variables in the atmosphere, as in the human body, which many times makes prediction little more than a guessing game.

[forkyfork]

i know about the butterfly effect. i don't think you understand what i'm saying... you can't solve the partial differential equations used to simulate the atmosphere exactly and that puts an absolute limit on accuracy even if you sampled every particle. there's also the fact that other factors like terrain are modeled using parameterizations.

[TWCCraig]

i know about the butterfly effect. i don't think you understand what i'm saying... you can't solve the partial differential equations used to simulate the atmosphere exactly and that puts an absolute limit on accuracy even if you sampled every particle. there's also the fact that other factors like terrain are modeled using parameterizations.

 

It's ok. I understand now, I kinda skipped over that point. 

 

Good little discussion anyway.

[pazzo83]

no, that's not true. wikipedia does a good job explaining it succinctly

http://en.wikipedia.org/wiki/Numerical_weather_prediction

"Although post-processing techniques such as model output statistics (MOS) have been developed to improve the handling of errors in numerical predictions, a more fundamental problem lies in the chaotic nature of the partial differential equations used to simulate the atmosphere. It is impossible to solve these equations exactly, and small errors grow with time (doubling about every five days). In addition, the partial differential equations used in the model need to be supplemented with parameterizations for solar radiation, moist processes (clouds and precipitation), heat exchange, soil, vegetation, surface water, and the effects of terrain."

 

Eventually computing power will allow for the solving of those equations exactly, probably by 2030.

[JetsPens87]

Eventually computing power will allow for the solving of those equations exactly, probably by 2030.

It's much less about that than the amount of initial data that can be ingested into the models in the first place.

[pazzo83]

It's much less about that than the amount of initial data that can be ingested into the models in the first place.

 

Well yea, that is the more difficult part.  How could we even conceivably sample every molecule in the atmosphere?  That technology won't exist for a while yet.

[n504jb]

no, that's not true. wikipedia does a good job explaining it succinctly

http://en.wikipedia.org/wiki/Numerical_weather_prediction

"Although post-processing techniques such as model output statistics (MOS) have been developed to improve the handling of errors in numerical predictions, a more fundamental problem lies in the chaotic nature of the partial differential equations used to simulate the atmosphere. It is impossible to solve these equations exactly, and small errors grow with time (doubling about every five days). In addition, the partial differential equations used in the model need to be supplemented with parameterizations for solar radiation, moist processes (clouds and precipitation), heat exchange, soil, vegetation, surface water, and the effects of terrain."

"It is impossible to solve these equations exactly"--Which is what Kurt Godel hypothecized when he suggested no mathematical model could be both complete and consistant.