Possible Snow Gradient Explanation

[tacoman25]

Yes. Trying to run a linear correlation with the QBO won't get you far outside of the blocking relationship. You also cannot put the "NAO" in the same list as ENSO, PDO and QBO because we are talking about different time scales in general.

It has been known since the 80's what direct effects the QBO has and how it works in relation to other variables and processes. I feel like a broken record about the QBO sometimes.

Well, it depends on if you are talking about month-to-month fluctuations or general phases which are longer, but still play a significant role in individual winters.

[tacoman25]

Are you talking about your October temperature thing? I get the feeling you are just trying to be argumentative right now.

Yes, that is one recent example. Not being argumentative at all. It's a fair point.

[HM]

Yes, that is one recent example. Not being argumentative at all. It's a fair point.

Come on man. That is not a fair point.

Let's just for the sake of argument say that October temperature trends are some how this major global oscillation similar to ENSO /PDO etc. How are we to judge the forcings if you pick years before 1950? Why instead of October temperatures don't we look at the bigger players that made those October temperatures?

October temperatures are an outcome not a forcing.

[HM]

Well, it depends on if you are talking about month-to-month fluctuations or general phases which are longer, but still play a significant role in individual winters.

The NAO is not a forcing on a seasonal scale in the way that the other factors are. The bigger players are the forcings that dictate the NAO which then dictates the temperatures. So if you are talking shorter timescales, then yes the NAO is a forcing on the pattern. But if we are talking seasonal level, like we are in this thread, you cannot say the NAO is a forcing, regardless of whatever decadal trend you think is there.

[tacoman25]

Come on man. That is not a fair point.

Let's just for the sake of argument say that October temperature trends are some how this major global oscillation similar to ENSO /PDO etc. How are we to judge the forcings if you pick years before 1950? Why instead of October temperatures don't we look at the bigger players that made those October temperatures?

October temperatures are an outcome not a forcing.

Of course they are. That isn't the point. It's a correlation, just like QBO/ENSO relationships as you have demonstrated in this thread.

[BethesdaWX]

Of course they are. That isn't the point. It's a correlation, just like QBO/ENSO relationships as you have demonstrated in this thread.

No, we demonstrate how the QBO is a driver of various features, how ENSO is a driver of various features. The October pattern does not drive anything directly, it is just a correlation. There is a difference.

[tacoman25]

No, we demonstrate how the QBO is a driver of various features, how ENSO is a driver of various features. The October pattern does not drive anything directly, it is just a correlation. There is a difference.

What drives the QBO? What drives ENSO? How do we know that the correlation is due to the QBO?

[Ellinwood]

What drives the QBO? What drives ENSO? How do we know that the correlation is due to the QBO?

The seasonal temperature pattern sure is hell isn't driven by shoulder-month temperature anomalies in one part of the hemisphere.

[tacoman25]

The seasonal temperature pattern sure is hell isn't driven by shoulder-month temperature anomalies in one part of the hemisphere.

Ok. Who said it was? I don't recall anyone saying that was a driver.

[Ellinwood]

Ok. Who said it was? I don't recall anyone saying that was a driver.

Alright, you did not say it was a driver. I take that back.

You looked at October temperatures with a cold East and warm West during -ENSO years and found a correlation with winter temps, with no other discussion as to WHY there was a correlation. Teleconnections are related to synoptic features in different parts of the hemisphere, and analyzing these synoptic features are much more significant than looking at a specific temperature pattern in the U.S.

While there is some value to looking at temperature correlations (normally it's better to have a correlation rather than not), one must use the classic "correlation does not imply causation" and figure out WHY there is a correlation and why such a correlation would actually mean something significant. There are a number of ways to get the same temperature pattern in the U.S., which is why finding out how we got the pattern is so crucial. By dissecting the different synoptic patterns via QBO, NAO, etc., we can discover what caused the temperature anomalies.

While the true science behind what causes the different teleconnections to change or remain stagnant is still somewhat of a mystery, using teleconnections is certainly a step above simplistic temperature pattern correlations.

[tacoman25]

Alright, you did not say it was a driver. I take that back.

You looked at October temperatures with a cold East and warm West during -ENSO years and found a correlation with winter temps, with no other discussion as to WHY there was a correlation. Teleconnections are related to synoptic features in different parts of the hemisphere, and analyzing these synoptic features are much more significant than looking at a specific temperature pattern in the U.S.

While there is some value to looking at temperature correlations (normally it's better to have a correlation rather than not), one must use the classic "correlation does not imply causation" and figure out WHY there is a correlation and why such a correlation would actually mean something significant. There are a number of ways to get the same temperature pattern in the U.S., which is why finding out how we got the pattern is so crucial. By dissecting the different synoptic patterns via QBO, NAO, etc., we can discover what caused the temperature anomalies.

While the true science behind what causes the different teleconnections to change or remain stagnant is still somewhat of a mystery, using teleconnections is certainly a step above simplistic temperature pattern correlations.

Well, to be perfectly honest there are a lot of correlations where we don't understand the causation exactly. There are explanations offered, and I'm sure there could be some decent ones for why the October pattern in -ENSO years has some correlation to the following winter, but they are often theoretical and open to interpretation. I do think it is helpful to discuss possible reasons why certain relationships exist, but the most solid/objective evidence is often the correlation itself.

And a lot of this is chicken/egg stuff...the NAO reflects weather patterns, the QBO reflects stratospheric patterns, ENSO reflects ocean cycles which are related to weather patterns, etc. Looking at teleconnectors offers more specialized, indexed data, but it all reflects the larger patterns - which many times feature many connecting correlations.

[Ellinwood]

Well, to be perfectly honest there are a lot of correlations where we don't understand the causation exactly. There are explanations offered, and I'm sure there could be some decent ones for why the October pattern in -ENSO years has some correlation to the following winter, but they are often theoretical and open to interpretation. I do think it is helpful to discuss possible reasons why certain relationships exist, but the most solid/objective evidence is often the correlation itself.

And a lot of this is chicken/egg stuff...the NAO reflects weather patterns, the QBO reflects stratospheric patterns, ENSO reflects ocean cycles which are related to weather patterns, etc. Looking at teleconnectors offers more specialized, indexed data, but it all reflects the larger patterns - which many times feature many connecting correlations.

The difference here is that we're using synoptic features to determine temperature patterns. It does not usually go in the opposite direction where temperature patterns determine the synoptic features (though it occasionally does have some effect). Granted, you're using temperatures to determine temperatures, but certain synoptic features are implied with these temperatures.

[BethesdaWX]

What drives the QBO? What drives ENSO? How do we know that the correlation is due to the QBO?

What "drives" them? They "drive" themselves. ENSO a state of the climate, but it all comes down to maintaining energy balance at certain thresholds. Not that mother nature has a mind of it's own, but rather brute physics, and feedbacks as a result of a change in heat/energy present will lead to a physical response, which in turn will eventually lead to a return-snap, think of the rubber-band theory, it applies with ENSO only in a more complicated way, weaker/stronger, +/-. etc. Periods of weak ENSO won't last and eventually we get a response to achieve balance as best we can.

Sure, correlation doesn't equal causation, but looking at October is not the same thins as looking at the QBO, because the October pattern is over 2 months old by the time we get to winter, and there is no "memory" in chaos theory.

Back to the ENSO thing, it is all about achieving balance, and being able to understand how the system may achieve balance may help forecasting quite a bit. In El Nino, we're gaining energy, in La Nina, we're losing energy. Many will disagree with this but there is no way around that fact, and it makes sense once you dive into it.

[tacoman25]

What "drives" them? They "drive" themselves. ENSO a state of the climate, but it all comes down to maintaining energy balance at certain thresholds. Not that mother nature has a mind of it's own, but rather brute physics, and feedbacks as a result of a change in heat/energy present will lead to a physical response, which in turn will eventually lead to a return-snap, think of the rubber-band theory, it applies with ENSO only in a more complicated way, weaker/stronger, +/-. etc. Periods of weak ENSO won't last and eventually we get a response to achieve balance as best we can.

Sure, correlation doesn't equal causation, but looking at October is not the same thins as looking at the QBO, because the October pattern is over 2 months old by the time we get to winter, and there is no "memory" in chaos theory.

Back to the ENSO thing, it is all about achieving balance, and being able to understand how the system may achieve balance may help forecasting quite a bit. In El Nino, we're gaining energy, in La Nina, we're losing energy. Many will disagree with this but there is no way around that fact, and it makes sense once you dive into it.

Ok...but people look at correlations/inverse correlations for the summer months and October for AO/NAO/PDO/etc, in relation to the following winter. I don't see how that is any different than looking at correlations involving temperature. All of those things are products of atmospheric patterns, just like temperature anomalies.

[BethesdaWX]

Ok...but people look at correlations/inverse correlations for the summer months and October for AO/NAO/PDO/etc, in relation to the following winter. I don't see how that is any different than looking at correlations involving temperature. All of those things are products of atmospheric patterns, just like temperature anomalies.

I certainly don't, the summer NAO thing is in the same camp as the October thing for me, and I prefer to avoid referencing either of them since they are both a qualitative method of of statistical analysis with no causation known, and both have failed numerious times. If anyone else does use the summer -NAO theory they shouldn't have a problem with the October correlation in my view.

[ORH_wxman]

I certainly don't, the summer NAO thing is in the same camp as the October thing for me, and I prefer to avoid referencing either of them since they are both a qualitative method of of statistical analysis with no causation known, and both have failed numerious times. If anyone else does use the summer -NAO theory they shouldn't have a problem with the October correlation in my view.

Correlations are very good to use if they are strong. Its probably silly not to ever use them in those cases. The best meteorologists still do not know the physical reasons behind many of the processes in our atmosphere.

[okie333]

Summer NAO is not nearly as important as summer AO.

[SoCoWx]

Summer NAO is not nearly as important as summer AO.

And did the Summer AO average negative??

[uncle W]

And did the Summer AO average negative??

negative...

http://www.cpc.ncep.noaa.gov/products/precip/CWlink/daily_ao_index/monthly.ao.index.b50.current.ascii.table

[HM]

Ok...but people look at correlations/inverse correlations for the summer months and October for AO/NAO/PDO/etc, in relation to the following winter. I don't see how that is any different than looking at correlations involving temperature. All of those things are products of atmospheric patterns, just like temperature anomalies.

It's called specific heat/high heat capacity of the ocean. I hope you are not saying that ocean temps and land-based surface temps have the same effect on climate and patterns. And while I would lump things like summer AO to winter AO correlations in with October surface temperatures, even the former can represent something larger going on than an October temperature map.

You are arguing to argue because no one with a general understanding of climate would seriously be saying the things you have been saying throughout this thread. And I know you are better than this.

[HM]

Of course they are. That isn't the point. It's a correlation, just like QBO/ENSO relationships as you have demonstrated in this thread.

No, my points are valid because you are saying that an October temp correlation is equivalent to what I originally posted and that is borderline insulting, given what I presented and did. If you think multi-dimensional oscillations that incorporate and exist over the ocean, that have been well documented and researched, are equivalent to land-based temperature anomalies, then what more can be said.

[tacoman25]

No, my points are valid because you are saying that an October temp correlation is equivalent to what I originally posted and that is borderline insulting, given what I presented and did. If you think multi-dimensional oscillations that incorporate and exist over the ocean, that have been well documented and researched, are equivalent to land-based temperature anomalies, then what more can be said.

That's not what I said. Please read again. You are being way too defensive here. My point was simply about correlation strength.

[tacoman25]

It's called specific heat/high heat capacity of the ocean. I hope you are not saying that ocean temps and land-based surface temps have the same effect on climate and patterns. And while I would lump things like summer AO to winter AO correlations in with October surface temperatures, even the former can represent something larger going on than an October temperature map.

You are arguing to argue because no one with a general understanding of climate would seriously be saying the things you have been saying throughout this thread. And I know you are better than this.

Clearly you are missing the points I've been making, and assuming cause/effect relationships that I've never stated. If you "know I am better than this", perhaps you should reconsider your interpretation of what you think I've been saying.

The fact is I never challenged what you said in this thread. You seem to be the one looking for an argument here.

[Ellinwood]

Clearly you are missing the points I've been making, and assuming cause/effect relationships that I've never stated. If you "know I am better than this", perhaps you should reconsider your interpretation of what you think I've been saying.

The fact is I never challenged what you said in this thread. You seem to be the one looking for an argument here.

The issue comes back to the "why" part of your correlation. We use things like QBO to describe why one could favor a certain temperature pattern based on how the atmosphere is behaving. All you've done is say "given this temperature pattern at this time we can expect to see this other temperature pattern in the future" with no real reasoning. This is why the correlations you have found are being dismissed regardless of the strength of them.

[HM]

That's not what I said. Please read again. You are being way too defensive here. My point was simply about correlation strength.

Clearly you are missing the points I've been making, and assuming cause/effect relationships that I've never stated. If you "know I am better than this", perhaps you should reconsider your interpretation of what you think I've been saying.

The fact is I never challenged what you said in this thread. You seem to be the one looking for an argument here.

Maybe you're right, maybe I'm taking your words out of context. This all started over your initial comment that said "But I've seen stronger correlations in tandem with ENSO dismissed on here. Just saying" which was implying the October temperature connection (and you later verified by saying it was a fair point that your Oct connection was dismissed). Then later you said, "That isn't the point. It's a correlation, just like QBO/ENSO relationships as you have demonstrated in this thread. "

So, you can see why I would think the things I am thinking right? So please, tell me what your point is because I want to make sure I have it right. I am sorry in advance if I got it wrong.

[okie333]

And did the Summer AO average negative??

Near-historic negative... consistent too.

[tacoman25]

The issue comes back to the "why" part of your correlation. We use things like QBO to describe why one could favor a certain temperature pattern based on how the atmosphere is behaving. All you've done is say "given this temperature pattern at this time we can expect to see this other temperature pattern in the future" with no real reasoning. This is why the correlations you have found are being dismissed regardless of the strength of them.

Well, I think you do have to weigh both the strength of the correlation and the possible reasons it might exist. You can explain a correlation until you're blue in the face, but the fact is that a lot is still unknown about what drives our atmosphere, and even if you have a wonderful theory behind it, if your correlation is weak that doesn't make it a great correlation.

The other thing is that many other correlations have been presented on this forum without much other than "look at this correlation" (and that isn't necessarily a bad thing - a valid explanation could be offered for just about any correlation), and yet the same standards of evaluation are not applied across the board.

[Ellinwood]

Well, I think you do have to weigh both the strength of the correlation and the possible reasons it might exist. You can explain a correlation until you're blue in the face, but the fact is that a lot is still unknown about what drives our atmosphere, and even if you have a wonderful theory behind it, if your correlation is weak that doesn't make it a great correlation.

I agree with this, as long as you have a reason (or at the very least a plausible theory) as to the "why" behind it. Meteorology is a science, and if you don't know the "why" behind something, you try to figure it out and/or provide a plausible reasoning. Also, from this particular example, using Apr/May/Sep/Oct temperature patterns to find correlations is often frowned upon as these shoulder months are fairly unstable due to the shifts in the jet stream creating erratic, temporary patterns. As such, these months can almost be treated as "anything goes" in terms of long-term predictability, and long-term forecast information derived from these patterns need to be heavily scrutinized due to their sometimes chaotic nature.

As far as the "explaining a weak correlation" goes... if it's weak, it tends to get ignored in favor of the stronger signals, especially if the two signals contradict each other. I'm not going to use a weak -AO correlation if there's a strong MJO correlation working in opposition of it. (Instead, it would be treated as a "risk")

The other thing is that many other correlations have been presented on this forum without much other than "look at this correlation" (and that isn't necessarily a bad thing - a valid explanation could be offered for just about any correlation), and yet the same standards of evaluation are not applied across the board.

Most of us don't feel like spending the time to unfurl all of the misconceptions and poor science on the board. Usually it's just easier to ignore it and let it slip into the abyss of the archives.

[tacoman25]

I agree with this, as long as you have a reason (or at the very least a plausible theory) as to the "why" behind it. Meteorology is a science, and if you don't know the "why" behind something, you try to figure it out and/or provide a plausible reasoning. Also, from this particular example, using Apr/May/Sep/Oct temperature patterns to find correlations is often frowned upon as these shoulder months are fairly unstable due to the shifts in the jet stream creating erratic, temporary patterns. As such, these months can almost be treated as "anything goes" in terms of long-term predictability, and long-term forecast information derived from these patterns need to be heavily scrutinized due to their sometimes chaotic nature.

As far as the "explaining a weak correlation" goes... if it's weak, it tends to get ignored in favor of the stronger signals, especially if the two signals contradict each other. I'm not going to use a weak -AO correlation if there's a strong MJO correlation working in opposition of it. (Instead, it would be treated as a "risk")

I understand what you are saying, but as I pointed out, it is not hard to come up with plausible reasoning for just about any correlation. I don't think that alone validates it. Also, I think if there is a strong correlation, it makes more sense to look in to possible reasons why it might exist rather than just dismissing it with: "that sort of correlation is frowned upon."

[Ellinwood]

I understand what you are saying, but as I pointed out, it is not hard to come up with plausible reasoning for just about any correlation. I don't think that alone validates it. Also, I think if there is a strong correlation, it makes more sense to look in to possible reasons why it might exist rather than just dismissing it with: "that sort of correlation is frowned upon."

So give us a plausible reason. Tell us why we should use an October correlation in seasonal forecasting other than "because it's there." If we don't have a reason, then how would we be able to determine whether we should use it or not beyond the pure odds?